Embed Size (px)
Citation preview
Science, In
novation
, and
Sustainab
ility: Investin
g in
British
Colum
bia’s K
now
ledg
e-based
Natural R
esource Sector
Cover and text
2 0 0 9
F O R R E X S E R I E S 2 5
Cover and text
ISSN 1495-964X
TM
Review of Hydrologic Models for Forest Management and Climate Change Applications in British Columbia and Alberta
Forrex Forum for Research and Extension in Natural Resources
Jos Beckers, Brian Smerdon, and Matt Wilson
Review of Hydrologic Models for Forest Management and Climate Change Applications in British Columbia and Alberta
TM
Forrex Forum for Research and Extension in Natural Resources
ii
© 2009 Forrex Forum for Research and Extension in Natural Resources Society, Kamloops, British Columbia, Canada
ISSN 1495-9658. Articles or contributions in this publication may be reproduced in electronic or print form for use free of charge to the recipient in educational, training, and not-for-profit activities provided that their source and authorship are fully acknowledged. However, reproduction, adaptation, translation, application to other forms or media, or any other use of these works, in whole or in part, for commercial use, resale, or redistribution, requires the written consent of Forrex Forum for Research and Extension in Natural Resources Society and of all contributing copyright owners. This publication and the articles and contributions herein may not be made accessible to the public over the Internet without the written consent of FORREX. For consents, contact: Managing Editor, Forrex, Suite 702, 235 1st Avenue, Kamloops, BC V2C 3J4, or email [email protected]
While this is a peer-reviewed publication, the information and opinions expressed in this publication are those of the respective authors and Forrex does not warrant their accuracy or reliability, and expressly disclaims any liability in relation thereto.
For more information, visit the Forrex website: www.forrex.org
This report is published by: Forrex Forum for Research and Extension in Natural Resources Suite 702, 235–1st Avenue Kamloops, BC V2C 3J4
Production of this report is funded, in part, by the BC Forest Investment Account–Forest Science Program through the Provincial Forest Extension Program, the Forest Management Branch of Alberta Sustainable Resource Development (ASRD), and the BC Ministry of Forests and Range, Future Forest Ecosystems Initiative (BC MFR).
Library and Archives Canada Cataloguing in Publication
Beckers, Jos, 1965- Review of hydrologic models for forest management and climate change applications in British Columbia and Alberta [electronic resource] / Jos Beckers, Brian Smerdon, and Matt Wilson.
(Forrex series, ISSN 1495-9658 ; 25) Includes bibliographical references.
Electronic monograph in PDF format.Also available in print format.ISBN 978-1-894822-60-2
1. Hydrologic models. 2. Forest management--Mathematical models. 3. Climatic changes--Mathematical models. 4. Watershed management--Mathematical models. 5. Forest management--British Columbia. 6. Forestmanagement--Alberta. I. Smerdon, Brian, 1974- II. Wilson, Matt, 1982- III. FORREX IV. Title. V. Series: FORREX series (Online) ; 25
SD567.B42 2009a 333.7509711 C2009-905982-7
iii
AbstrAct
This review summarizes the capabilities and limitations of existing hydrologic models for use in an operational forest management context in British Columbia (BC) and Alberta (AB). The review brings together relevant information contained in user manuals, technical model documentation, and in published materials that describes model applications, and emphasizes studies conducted in the Pacific Northwest and in physical and climatic settings similar to those encountered in BC and AB.
One outcome of this review is to provide guidance (decision support) for resource managers and other practitioners to help them identify which hydrologic models are most appropriate for addressing their forest management questions. To do this, the review identifies trade-offs between model complex-ity and model functionality for addressing forest management questions and makes recommendations for advancing the routine and consistent use of watershed models. These recommendations include improving interdisciplinary education; performing model inter-comparisons at data-rich (experimen-tal) and data-poor (ungauged) watersheds; enhancing data availability; communicating uncertainty in results; developing better models, graphical user interfaces (GUIs), commercial software, and model support; and establishing regulatory guidance and professional precedence.
The review also considers the suitability of select models for exploring the potential effects of climate change on future watershed processes that are relevant to forest management. Emphasis is placed on shifts in site water balances (evapotranspiration); changes in snow accumulation and melt rates; melting of permafrost, river, and lake ice processes; adjustments in glacier mass balance; changes in stream-flow generation; and the increased risk of disturbances such as wildfire, pest outbreaks (e.g., mountain pine beetle), flood events, windthrow, and landsliding. The barriers and challenges to using hydrologic models for answering climate change questions are discussed, and areas for model improvement are identified.
keywords: Alberta, British Columbia, climate change, decision support, forest management, model review, model selection, watershed hydrology.
.Authors And AffiliAtions
Jos Beckers*, PhD., P.Geo., Senior Technical Specialist, WorleyParsons Canada, 8658 Commerce Court, Burnaby BC V5A 4N6. Email: [email protected] Smerdon, PhD., Research Scientist, Department of Earth Sciences, Simon Fraser University. Burnaby, BC V5A 1S6. Email: [email protected] Wilson, BSc., E.I.T., LEED AP, Environmental Engineer, WorleyParsons Canada, 8658 Commerce Court, Burnaby BC V5A 4N6. Email: [email protected]* Corresponding author
Citation—Beckers, J., B. Smerdon, and M. Wilson. 2009. Review of hydrologic models for forest management and climate change applications in British Columbia and Alberta. forrex Forum for Research and Extension in Natural Resources, Kamloops, BC forrex Series 25. www.forrex.org/publications/forrexseries/fs25.pdf
iv
Acknowledgements
Funding for this hydrologic model review was provided by the BC Forest Investment Account–Forest Science Program through the Provincial Forest Extension Program, the Forest Management Branch of Alberta Sustainable Resource Development (ASRD), and the BC Ministry of Forests and Range, Future Forest Ecosystems Initiative (BC MFR). Discussions with, and input from, the project steering com-mittee of Todd Redding (FORREX), Axel Anderson (ASRD), Robin Pike (BC MFR), and Arelia Werner (PCIC) are gratefully acknowledged. The comments of two external reviewers greatly improved this report. Arelia Werner and Katrina Bennett provided constructive comments and valuable materials for the review of PREVAH, VIC, WaSiM-ETH, and Watflood models, and for the discussion of downscaling techniques.
v
tAble of contents
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Authors and Affiliations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
List of Acronyms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .viii
List of Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .x
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11.1 Report Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
2 Background and Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22.1 Models Considered . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22.2 Model Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42.3 Model Rankings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
2.3.1 Model Functionality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52.3.2 Model Complexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92.3.3 Model Applicability to Climatic and Physiographic Settings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112.3.4 Model Application Scale and Grid Size Selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122.3.5 Model Outputs Relevant to Forest Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .132.3.6 Model Adaptability to Represent Future Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14
2.4 Model Calibration and Prediction Confidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142.5 Limitations of Model Review and Ranking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
3 Decision Support for Selecting Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173.1 A Step-wise Approach to Selecting Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173.2 Model Advantages and Disadvantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
3.2.1 Low-complexity Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .243.2.2 Medium-complexity Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .243.2.3 High-complexity Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
4 Barriers to the Operational Use of Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .264.1 User Knowledge and Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .264.2 Model Inter-Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .274.3 Data Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .284.4 Communicating Model Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .284.5 Need for Better Models, GUIs, and Model Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .284.6 Policy and Professional Precedence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29
5 Using Models in a Climate Change Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .295.1 Hydrologic Implications of Climate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .295.2 Model Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
5.2.1 Low-complexity Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .355.2.2 Medium-complexity Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .355.2.3 High-complexity Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .365.2.4 Macroscale Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
5.3 Barriers to the Operational Use of Models in a Climate Change Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
6 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43
7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
vi
Appendix 1 Model Review and Ranking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60A1.1 ACRU. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60A1.2 BROOK90. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62A1.3 CRHM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63A1.4 DHSVM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65A1.5 ForHyM and ForWaDy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68A1.6 HBV-EC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .69A1.7 HEC-HMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71A1.8 HELP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .72A1.9 HSPF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74A1.10 InHM and HydroGeoSphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75A1.11 MIKE-SHE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .77A1.12 MODHMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .79A1.13 PREVAH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80A1.14 PRMS (MMS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82A1.15 RHESSys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .84A1.16 SSARR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87A1.17 SWAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88A1.18 UBC-UF Peak Flow Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90A1.19 UBCWM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92A1.20 VIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94A1.21 WaSiM-ETH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96A1.22 Water Balance Model by QUALHYMO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98A1.23 WATFLOOD, CLASS, and MESH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98A1.24 WEPP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100A1.25 WRENSS (WinWrnsHyd and ECA-AB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103A1.26 WRMM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105A1.27 WUAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Appendix 2 Model Review Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
tAbles
1 Models reviewed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
2 Model functionality evaluation criteria.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
3 Model complexity evaluation criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
4 Model functionality ranking; refer to Table 2 and Section 2 for evaluation criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . .16
5 Model complexity ranking; refer to Table 3 and Section 2 for evaluation criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
6 Model applicability to forest management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
7 Model applicability to climatic and physiographic settings; refer to Section 2 for evaluation criteria. 21
8 Model outputs for forest planning; refer to Appendix 1 for model descriptions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
9 Model advantages and disadvantages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
10 Climate change model evaluation criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
11 Climate change model ranking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
A2.1 Agricultural Catchments Research Unit (ACRU) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
vii
A2.2 BROOK90 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
A2.3 Cold Regions Hydrologic Model (CRHM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
A2.4 Distributed Hydrology Soil Vegetation Model (DHSVM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
A2.5 Forest Hydrology Model (ForHyM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
A2.6 Forest Water Dynamics (ForWaDy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
A2.7 Hydrologiska Bryåns Vattenbalansavdelning–Environment Canada (HBV-EC). . . . . . . . . . . . . . . . . . . . . . . . . . 119
A2.8 Hydrologic Engineering Center's Hydrologic Modelling System (HEC-HMS) . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
A2.9 Hydraulic Evaluation of Landfill Performance (HELP). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
A2.10 Hydrologic Simulation Program–Fortran (HSPF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
A2.11 HydroGeoSphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
A2.12 Integrated Hydrology Model (InHM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
A2.13 MIKE-SHE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
A2.14 MODHMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
A2.15 Precipitation-Runoff-Evapotranspiration-Hydrotope model (PREVAH). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
A2.16 Precipitation Runoff Modelling System/Modular modelling system (PRMS/MMS) . . . . . . . . . . . . . . . . . 137
A2.17 Regional Hydro-Ecologic Simulation System (RHESSys). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
A2.18 Streamflow Synthesis and Reservoir Regulation (SSARR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
A2.19 Soil and Water Assessment Tool (SWAT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
A2.20 UBC–UF Peak Flow Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
A2.21 University of British Columbia Watershed Model (UBCWM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
A2.22 Wasserhaushalts-Simulations-Modell (WaSiM-ETH). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
A2.23 Water Balance Model for BC (based on QUALHYMO model). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
A2.24 Variable Infiltration Capacity (VIC) model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
A2.25 WATFLOOD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
A2.26 Water Erosion Protection Project (WEPP). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
A2.27 Water Resources Evaluation of Non-Point Silvicultural Source (WRENSS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
A2.28 Equivalent Cut Area–Alberta (ECA-AB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
A2.29 Water Resources Management Model (WRMM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
A2.30 Water Use Analysis Model (WUAM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
figures
1 Models organized according to overall functionality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
2 Model functionality for forest management and climate change. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
3 Model complexity and functionality for forest management and climate change. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
viii
list of Acronyms
AB Alberta
ACE Army Corps of Engineers (United States)
ACRU Agricultural Catchments Research Unit
AET Actual Evapotranspiration
ASRD Alberta Sustainable Resource Development
BC British Columbia
BC MFR BC Ministry of Forests and Range
CN Curve Number (method)
CRHM Cold Regions Hydrologic Model
DEM Digital Elevation Model
DHSVM Distributed Hydrology Soil Vegetation Model
EC Environment Canada
ECA-AB Equivalent Clearcut Area–Alberta (model)
EPA Environmental Protection Agency (United States)
ET Evapotranspiration
ETH Eidgenössische Technische Hochschule (Switserland)
ForHyM Forest Hydrology Model
ForWaDy Forest Water Dynamics (model)
GIS Geographic Information System
GRU Grouped Response Unit
HBV Hydrologiska Byråns Vattenbalansavdelning (model)
HEC-HMS Hydrologic Engineering Center’s Hydrologic Modelling System
HELP Hydraulic Evaluation of Landfill Performance (model)
HRU Hydrologic Response Unit
HSPF Hydrologic Simulation Program–Fortran
InHM Integrated Hydrology Model
LAI Leaf Area Index
MMS Modular Modelling System
MODHMS MODFLOW-Hydrologic Modelling System
PCIC Pacific Climate Impacts Consortium
PET Potential Evapotranspiration
PNW Pacific Northwest
PREVAH Precipitation-Runoff-Evapotranspiration-Hydrotope (model)
PRMS Precipitation Runoff Modelling System
RHESSys Regional Hydro-Ecologic Simulation System
ix
ROS Rain-on-Snow
SCS Soil Conservation Service (United States)
SHE Système Hydrologique Européen
SMHI Swedish Meteorological and Hydrological Institute
SSARR Streamflow Synthesis and Reservoir Regulation (model)
SVAT Soil Vegetation Atmosphere Transfer
SWAT Soil and Water Assessment Tool (model)
SWE Snow Water Equivalent
UBC University of British Columbia
UBCWM UBC Watershed Model
UF University of Freiburg (Germany)
UL University of Lethbridge (Alberta)
UNP University of Natal in Pietermaritzburg (South Africa)
URL Uniform Resource Locator (web page locator)
US United States
USGS United States Geological Survey
UW University of Washington (United States)
VIC Variable Infiltration Capacity (model)
WaSiM Wasserhaushalts-Simulations-Modell
WEPP Water Erosion Protection Project (model)
WRENSS Water Resources Evaluation of Non-Point Silvicultural Source (procedure)
WRMM Water Resources Management Model
WUAM Water Use Analysis Model
x
list of definitions
The following list provides definitions of selected technical terms used in this report. Definitions are adapted from Kampf and Burges (2007) and the Environment Canada freshwater glossary (http://www.ec.gc.ca/water/en/info/gloss/e_gloss.htm).
Actual Evapotranspiration (AET) – The quantity of water removed from a surface by the combined processes of evaporation and transpiration.
Analytical equation (model) – A model that uses simplifying assumptions to derive solutions of the governing mass and/or energy conservation and transport equations. Also see empirical and physically based equation (model).
Application scale – The watershed size that a model can be applied to (small, medium, or large). In this report, small watersheds have been defined as having an area less than 100 km2, medium watersheds are defined as being between 100 and 10 000 km2 in area, and large watersheds as being greater than 10 000 km2 in area.
Calibration – The process whereby certain model parameters are adjusted until simulation results match a set of measurements such as streamflow or snow water equivalent.
Distributed model – A model that explicitly accounts for spatial variability of input variables, typically by dividing the watershed into equally sized grid cells.
Empirical equation (model) – A model based on simplified, experimentally derived relationships such as linear regressions. Also see analytical and physically based equation (model).
Equifinality – The situation in which multiple model parameter combinations can lead to the same streamflow calibration (i.e., non-uniqueness) but may result in different answers regarding the effects of land use or climate change. This situation is typically caused by either an incorrect process representa-tion or a lack of data regarding internal watershed processes and may lead to errors in model predictions regarding future hydrologic conditions in a watershed.
Equivalence – See equifinality.
Forest stand – An area of the forest that is relatively uniform in species composition or age and can be managed as a single unit.
Fully distributed model – See distributed model.
Grouped Response Unit (GRUs) – Grid cells lumped into bins having similar land cover, elevation, slope, and/or aspect, in order to maintain computational efficiency.
Grid cell – A discrete (often square) unit that represents a portion of a watershed.
Grid scale – The size of grid cells in semi-distributed and distributed models.
Hydrologic Response Unit (HRUs) – Grid cells with similar hydrologic characteristics lumped into bins.
Internal watershed processes – See watershed processes.
Leaf Area Index (LAI) – The one-sided, green-leaf area per unit ground area in broadleaf canopies, or as the projected needle-leaf area per unit ground area in needle canopies.
Lumped model – A model that does not account for the spatial distribution of input variables or param-eters to represent heterogeneity in vegetation, soils, or other watershed characteristics.
Mixed model – A model that combines physically based, analytical, and/or empirical equations.
Model complexity – Defined in this report by the estimated data, resources, and time (cost) required to initialize and calibrate a model.
xi
Model discretization – See watershed discretization.
Model functionality – Defined in this report by the range of watershed processes considered in a model, the equations used to simulate these processes (physical, analytical, or empirical) and model discretization.
Parameterization – The process of defining or deciding the parameters of a model. Decisions include defining parameter values and (any) spatial variability in these values to reflect hetererogeneity of a watershed (e.g., soil types and land cover).
Parsimonious model – A model having a small number of “free” parameters (i.e., parameters that are not known from data).
Physically based equation (model) – A model derived from equations describing conservation of mass, momentum, and/or energy. Also see analytical and empirical equation (model).
Planning scale – The minimum landscape unit within a model for which forest cover characteristics can be modified. The planning scale is closely tied to the watershed discretization of a model (fully distrib-uted, semi-distributed, or lumped).
Runoff processes – The processes that direct rainfall and/or snowmelt from a hillslope towards the stream channel. The processes may include infiltration, percolation through the soil column, interflow (lateral subsurface flow above the water table), saturated subsurface flow (below the water table), infil-tration excess (Horton) overland flow and saturation excess (Dunne) overland flow.
Semi-distributed model – A model that divides the watershed into areas with common hydrologic properties. Examples include elevation bands, hillslopes, sub-basins, Grouped Response Units (GRUs) or Hydrologic Response Units (HRUs).
Snow water equivalent (SWE) – The amount of water contained within the snowpack. It can be thought of as the depth of water that would result if you melted the snowpack.
Soil-vegetation atmosphere transfer processes (SVAT) – Soil-vegetation-atmosphere transfer (SVAT) processes include canopy rain interception and evaporation, fog drip, stemflow, canopy snow intercep-tion, snowmelt, snowdrift/ blowing snow, vegetation transpiration, and soil evaporation.
Stand level – At the scale (size) of a forest stand.
Strahler stream order – See stream order.
Stream order – An algorithm used to define stream size based on a hierarchy of tributaries. Streams with no tributaries are first-order streams. When two first-order streams come together, they form a second-order stream. When two second-order streams come together, they form a third-order stream, etc.
Watershed processes – The processes that occur on hillslopes and in stream channels, and that inter-act to generate streamflow at the watershed outlet, encompassing soil-vegetation atmosphere transfer (SVAT) processes, runoff processes, channel routing, lake or wetland storage, and other stream processes.
1
introduction 1
Predicting the impact of forestry activities on watershed hydrologic processes and the associated effects on streamflow is complicated (e.g., Alila and Beckers 2001; Pike et al. 2007). Often, there are intricate linkages between the disturbance activities and the consequences for an affected resource, such as flood risk to human life and property, water availability, or aquatic habitat (e.g., Pike and Scherer 2003; Scherer and Pike 2003; Moore and Wondzell 2005). It is difficult to draw conclusions about the effects of forestry on watershed hydrology from data alone because field experiments have practical limita-tions such as cost, duration, sparse monitoring locations, and transferability of findings (Ward 1971; Dunne 2001). Furthermore, climatic variability obscures cause-and-effect relationships, and many other factors can contribute to the effects of forest management on watershed hydrology (Alila and Beckers 2001). For example, the chosen silvicultural system (e.g., clearcut, partial cut) and logging method (e.g., cable yarding, ground skidding), the location of harvesting within a watershed, and road management (construction, use, maintenance, and decommissioning) can produce overlapping or cumulative effects.
Field measurements must be supplemented with numerical modelling to better understand the cumulative watershed effects of forest management (Ziemer et al. 1991; Dunne 2001; Alila and Beckers 2001). From a scientific perspective, models are valuable as tools to assist in interpreting data and in linking physical processes that are measured from the stand level to the watershed scale (Leaf 1975; US-Environmental Protection Agency 1980; Hillel 1986; Alila and Beckers 2001). From an operational perspective, models are valued for their scenario analysis and predictive capability that enable resource managers in industry, government, First Nations, and other sectors to predict what future watershed conditions might be under varying management practices. Models can also be used to account for com-plicating factors such as the effects of climate variability and climate change, or other disturbances in long-term planning.
There is a move in British Columbia (BC) and Alberta (AB) towards increased use of numerical hydrologic models to answer questions related to the water resource impacts of forest management (Hudson and Quick 1997; Hudson 2000; Swanson and Rothwell 2001; Whitaker et al. 2002; Schnorbus and Alila 2004a; Luo and Alila 2006; Alila and Luo 2007; Moore et al. 2007) and climate change (Hut-chinson et al. 1999; Scibek and Allen 2006a, 2006b; Toews 2007; Stahl et al. 2008). However, the use of models for routine operational planning by resource managers and their consultants is not widespread. Currently, resource managers lack the tools to identify which model is the most appropriate to answer their specific forest management-related questions, taking into account site-specific conditions and practical constraints such as time (budget), expertise, and data availability. Prospective model users or reviewers of hydrologic modelling studies may also not have the specialized technical knowledge to understand the possible assumptions and limitations of a given model, or the implications of model parameterization. This model review provides both the relevant tools and knowledge and is intended to advance the operational use of models and to ensure that model outputs (i.e., the information on which management decisions are based) are credible and reliable.
With climate change comes many adjustments in watershed hydrology in British Columbia and Alberta, including shifts in site water balances (evapotranspiration); changes in snow accumulation and melt rates; changes in the melting of permafrost, river and lake ice processes; and adjustments in gla-cier balances and in the frequency of disturbances such as flood events, wildfire, pests (e.g., mountain pine beetle), windthrow and landsliding (Whitfield et al. 2002; Rodenhuis et al. 2007; Pike et al. 2008b; Sauchyn and Kulshreshtha 2008; Stahl et al. 2008; Walker and Sydneysmith 2008). The hydrologic im-plications of climate change pose questions that could potentially be answered by applying appropriate watershed models. Thus, the suitability of the reviewed models for exploring potential climate change effects on future watershed processes and outputs relevant to forest management is also considered in this synthesis.
2
Previous hydrologic model reviews with an emphasis on forest watershed management or climate change have been conducted (e.g., Pike 1995; Pike 2003; Whitaker et al. 1998; Alila and Beckers 2001; Hutchinson 2007; Pike et al. 2007; Werner and Bennett 2009). However, these reviews were mostly limited to a few select models. Therefore, the objective of this synthesis is to comprehensively review and describe a broad range of hydrologic models and to discuss their applicability to the forested water-sheds of BC and AB. This review is the basis for developing a decision-support framework that will help resource managers select hydrologic models that are most appropriate to their circumstances (watershed conditions), needs (forest management or climate change questions), and constraints (data, expertise, and funding). This synthesis will also help improve the ability of forest managers to understand the as-sumptions and limitations that underlie model results.
report organization1.1
This report is organized as follows: Section 2 introduces 30 hydrologic models, summarizes the methodology and criteria used to review and describe the models and their applications, and discusses the underlying considerations for selecting and applying a model. Because of the volume of technical material reviewed, the hydrologic models and their applications are discussed and ranked in Appendix 1. The objective of this organization is to limit the length and technical content of the main report, thus emphasizing the decision-support aspect of model selection. Section 3 provides this decision support, including a step-wise approach to model selection, and an overview of the main advantages and disadvantages of the nine hydrologic models identified as being the most appropriate for answering forest management questions. Current barriers to the operational use of models in a forest manage-ment context are discussed in Section 4, while Section 5 researches the suitability of select models for exploring potential climate change effects on future watershed processes and outputs relevant to forest management. A summary of the main findings of this synthesis is provided in Section 6.
The purpose of Appendix 1 (model descriptions and model application reviews) is to establish links between technical model documentation and published model applications, while the review tables are used to provide decision support for model selection. Appendix 1 also serves to provide the reader with comprehensive background information and literature sources for each model. Appendix 2 provides an abbreviated tabular version of the review, based on a consistent set of criteria, to facilitate inter-compari-son of individual model components and features.
bAckground And methods2
models considered 2.1
The first step of the model evaluation and synthesis was to identify and compile a list of models. This list is presented in 1, which first classifies a total of 30 models according to two categories (those that were specifically developed for use in a forest management context and all other models) and then identifies their development purpose (e.g., simulation of changes in annual yield). This classification is retained in tables throughout the report. Full model names, model developers, model development websites (URLs), and other reference materials are provided in Appendix 1.
The intent of the synthesis was to consider a comprehensive list of models that could potentially be applied in BC and AB. As such, the synthesis does include models that are known to be unsuitable for use in a forest management context, such as the Water Balance Model for BC (a widely used model for urban stormwater management applications) and the Environment Canada Water Use Analysis Model (WUAM), which is intended for analyzing water-use scenarios in river basins.
3
Category Purpose Abbreviated model name1
Forest hydrology models
Annual yield WRENSS (WinWrnsHyd/ECA-AB)
Peak flow UBC-UF Peak Flow Model2
Water balance BROOK90
ForHyM
ForWaDy
Watershed hydrology
DHSVM
RHESSys
Other models
Water balance HELP
Water Balance Model by QUALHYMO
Watershed hydrology
ACRU
CRHM
HBV-EC
HEC-HMS
HSPF
PREVAH
PRMS/MMS
SWAT
UBCWM
WaSiM-ETH
Watflood
Groundwater-surface water models
InHM (HydroGeoSphere)
Mike-SHE
MODHMS
Soil erosion WEPP
River basin models SSARR
VIC
WRMM
WUAM
1 Full model names and model developers are provided in Appendix 1.2 Tentative model name for purpose of this review.
table 1 Models reviewed.
Closely related models with nearly identical capability were lumped together in the review (Appendix 1; Table 1). This includes the following:
• TheForestHydrologyModel(ForHyM)andtheForestWaterDynamics(ForWaDy)model• TwoversionsoftheWaterResourcesEvaluationofNon-PointSilviculturalSources(WRENSS)
procedure: WinWrnsHyd and Equivalent Clearcut Area-Alberta (ECA-AB) • TheIntegratedHydrologyModel(InHM)andHydroGeoSphere
Of these models, only ForHyM and ForWaDy are listed separately in tables and figures to reflect enhancements with respect to the simulation of evapotranspiration and snowmelt made in the latter model.
4
Two promising models (not listed in Table 1) were not reviewed because they are still in the early stages of development: a “low-flow” prediction model being developed as a collaborative effort between the University of British Columbia (UBC) and the University of Freiburg (UF) in Germany (Dr. M. Weiler, pers. comm., Jan. 2009) and a rewrite of the Environment Canada (EC) version of the Hydrologiska Byråns Vattenbalansavdelning (HBV) model. This latter model, which will be completed by mid-2009, will have a target complexity that is intermediate between the Distributed Hydrology Soil Vegetation Model (DHSVM) and the current HBV-EC model (Dr. R. D. Moore, pers. comm., Jan. 2009), and will be developed as a collaborative effort between UBC, EC, and Alberta Sustainable Resource Development (ASRD) for the purpose of addressing forest management questions. These two models have the poten-tial to address some of the challenges and trade-offs between model complexity and model functionality that exist with current hydrologic models. A discussion of these challenges and trade-offs is provided.
The model review (Appendices 1 and 2) was based on peer-reviewed publications and other refer-ence materials (e.g., technical manuals and/or user guides) listed on the model websites. In specific cases, where information was limited or in the case of new models, information requests were sent to the model developers directly.
model Applications 2.2
A comprehensive review of published forest management and climate change applications was complet-ed for each model (Appendix 1). While this synthesis report is focused on BC and AB, the application review also considered studies in similar physiographic settings, with an emphasis on the Pacific North-west (PNW). The search was based on publications and reference materials listed on the model websites, as well as an on-line search of key peer-reviewed journals. In specific cases where information was limited, requests were sent to the model developers directly for publications and/or synopses of model applications.
This section of the review focused on model applicability for predicting the hydrologic effects of for-est management and climate change in BC and AB, documenting the following:
• Temporalandspatialscalesoftheapplication• Conditionsunderwhichthemodelwassuccessfullyorunsuccessfullyapplied• Referencestopublications(e.g.,climatechange,forestry-related,orotherrelevantreferences)• Abriefsynopsisoftheapplicabilityofthemodeltosimulateforestwatershedhydrologyin
BC and AB
model rankings 2.3
A model is a perception of how a system works. It is a hypothesis of the real world’s functioning, codi-fied in quantitative terms (Savenije 2009). The models considered in this review have been developed to describe and predict an array of watershed hydrologic characteristics, including, but not limited to: soil-vegetation-atmosphere transfer (SVAT) processes, rainfall-runoff relationships, glacial melt, and groundwater-surface water interactions. Hydrologic models can divide (discretize) watersheds either in a lumped, semi-distributed, or fully distributed fashion and may simulate hydrologic processes using either physically based mass and energy-conservation equations, analytical expressions that simplify these conservation equations, empirical (conceptual) expressions, or a combination of these approaches (e.g., Kampf and Burges 2007). Some models have the capability to address specific hydrologic pro-cesses that are important in forested watersheds (e.g., precipitation interception from a forest canopy or melt from a snowpack) in great detail, while others may lack the necessary hydrologic-process rep-resentation to accurately simulate forest hydrologic regimes. Some models may only be suitable for use in large watersheds or in relatively flat terrain, while others are more applicable to small watersheds or
5
mountainous terrain. The models reviewed also differ in data and resource requirements, ranging from relatively simple procedures that can be applied in a matter of days to highly complex models that may take several weeks or months to set up and calibrate. Our model review and ranking took these and other considerations into account.
For the purposes of developing a decision-support tool, the models were reviewed and ranked against five main criteria:
• Modelfunctionality(Section2.3.1),whichisdefinedastherangeofprocessesconsideredbythemodel, the equations adopted to simulate these processes, and by model discretization.
• Modelcomplexity(Section2.3.2),whichisdefinedbytheestimateddata,resources,andtime(cost) required to initialize and calibrate the model.
• Modelapplicabilitytoparticularclimaticandphysiographicsettings,includingwatershedscale(Section 2.3.3).
• Modelabilitytoprovideoutputsrelevanttoforestmanagement,includingconsiderationforspatialand temporal scale at which these outputs can be provided (Section 2.3.5).
• Modeladaptabilitytorepresentfuturewatershedconditionsinalong-termplanningand/orcli-mate change context (Section 2.3.6).
To facilitate these rankings, each model was described based on a consistent set of criteria, allowing for direct comparison between models for different components or features. The complete list of evalua-tion criteria for each model is provided in Appendix 2, and also includes generic model qualities such as availability of documentation, technical support, computing requirements, and purchase cost.
Model Functionality2.3.1
When choosing a model for a certain application, it is important to consider its ability to simulate the desired land use or climate change scenarios (Whitaker et al. 1998; Alila and Beckers 2001) and the accuracy of the predictions. This functionality aspect of model selection is affected by the hydrologic processes represented in the model, the equations adopted to simulate these processes, and by model discretization.
Hydrologic Process Simulation
When using numerical modelling to quantify the complex linkages between water-related concerns and timber harvesting and roads, it is important that the chosen model accounts for the watershed processes that bring about these linkages. Chang (2006) provides a general overview of forest hydrology. For the purpose of this review, model functionality for addressing forest management questions was assessed against the following hydrologic processes and factors (Table 2):
• Soil-vegetation-atmospheretransfer(SVAT)processes,includingcanopyraininterceptionandevaporation, fog drip, stemflow, canopy snow interception, snowmelt, snowdrift/ blowing snow, vegetation transpiration, and soil evaporation
• Soilmoisturestorageandrunoffgenerationprocesses,includinginfiltration,depressionstorage,shallow subsurface runoff, preferential flow, Horton (infiltration excess) overland flow, and Dunne (saturation excess) overland flow
• Channelrouting• Roadprocessesandstructures,includinghillsloperunoffinterception,precipitationinterception,
flow diversion, and stream-crossing structures such as culverts• “Other”processesandfactors,includinggroundwaterflow,glaciermelt,frozensoil/permafrost,
lakes, wetlands, and structures that control water
6
Criterion Evaluation measure Score
Space discretization Lumped L
Semi-distributed (GRUs, HRUs, elevation bands, hillslopes, sub-basins, etc.)1 M
Distributed H
Soil layers Not represented
Conceptual representation (unsaturated/saturated zones) L
Explicit soil depths M
Finite difference of finite element discretization (subsurface grid cells) H
Vegetation layers Not represented
Single layer L
Two layers M
Two or more layers H
Watershed processes, road hydrology, and other features
Not simulated
Empirical approaches L
Analytical approaches M
Physical approaches H
1 GRU represents grouped response unit; HRU represents hydrologic response unit.
table 2 Model functionality evaluation criteria.
The approaches (equations and models) used to simulate these hydrologic processes have been defined as follows (e.g., Kampf and Burges 2007; Table 2):
• Empiricalapproaches(“low”rank)arebasedonsimplified,experimentallyderivedrelation-ships such as linear regressions. Examples include the US Soil Conservation Service (SCS) curve number method for calculating runoff, the representation of groundwater flow systems as linear reservoirs, and temperature index-based snowmelt calculations. Empirical methods for calculating evapotranspiration (ET) include radiation-based equations (e.g., Priestly-Taylor) or temperature-based methods (Maidment 1993).
• Analyticalmodels(“medium”rank)usesimplifyingassumptionstoderivesolutionsofthegoverning equations describing conservation of mass, momentum, and/or energy. Examples include the Green-Ampt equation, which describes the infiltration of rainfall or snowmelt into the subsurface; and hybrid temperature-radiation, energy-based snowmelt methods (e.g., Hock 2003). Analytical approaches to calculate evapotranspiration include the Kristensen and Jensen (1975) method, which was developed for Nordic climates.
• Physicallybasedmodels(“high”rank)arederivedfromtheequationsdescribingconservationofmass, momentum, and/or energy. Examples include energy balance-driven snowmelt calculations (Anderson 1973) or the use of the three-dimensional form of the Richards’ equation for solv-ing subsurface flow. Physically based evapotranspiration equations include the widely accepted Penman-Monteith method, which is recommended for calculating reference evapotranspiration by the Food and Agricultural Organization (FAO) of the United Nations and by the American Society of Civil Engineers (ASCE) (Allen et al. 2005).
No rank was assigned for processes not being simulated (Table 2). In practice, models may represent some processes in an empirical fashion while analytical or physically based methods may be used for other processes. An example is a model that simulates evapotranspiration in a physically based manner,
7
but represents snowmelt using an empirical degree-day method. In this review, these models will be ranked as having a “mixed” representation for simulating SVAT processes that are important for pre-dicting the hydrologic effects of forest harvesting and/or climate change.
Data availability may often be one of the primary drivers behind the choice for a particular model ap-proach. Empirical models typically (but not always) have relatively low data requirements, but may also be less accurate when applied outside the conditions for which the empirical relationships were deter-mined (e.g., using a temperature-index snowmelt method calibrated to forested conditions to predict melt under clearcut conditions). This limited accuracy may be adequate for many watershed applica-tions. However, when using numerical modelling in high-value or high-risk situations (e.g., watersheds with high property values, community watersheds, ecologically sensitive watersheds, and streams with significant fisheries values), it may be important that the chosen model accurately simulates the gov-erning hydrologic processes. Generally, physically based models are characterized by a higher intrinsic accuracy for predicting the effects of land use disturbance or climate change, but they also suffer from high data requirements (e.g., Abbott et al. 1986a; Beven 1989; Grayson et al. 1992a, 1992b). As such, in selecting a model for a particular application, trade-offs between model accuracy and data requirements should be expected. Therefore, site-specific, tailor-made model approaches are needed that take into ac-count the questions to be addressed and the data available to answer these questions (Savenije 2009).
Model Spatial Discretization
The model functionality ranking also took into account the model’s ability to spatially discretize the watershed, soil, and vegetation (Table 2). The purpose of dividing (discretizing) a watershed into dif-ferent units is to account for the spatial variability and pathways of water movement, and to describe geologic and land cover variability and the effects of shade, slope, and aspect on hydrologic response (Kampf and Burges 2007). Vegetation discretization allows for representation of the vertical pathways of water through multi-layered forest canopies, while soil vertical discretization is used to distinguish between separate subsurface flow pathways (shallow and deep), and to account for soil moisture storage and rooting depth of different vegetation species.
Watershed discretization was distinguished as follows (e.g., Kampf and Burges 2007):
• Lumpedmodels(“low”rank)donotaccountforthespatialdistributionofinputvariablesorparameters that represent heterogeneity of, for example, vegetation and soils within a watershed. Lumped models typically do not require a digital elevation model (DEM) as input.
• Semi-distributedmodels(“medium”rank)dividethewatershedintoareaswithcommonhydro-logic properties (reflected by uniform input variables or parameters) due to similarities in terrain, land cover, and/or soils. Examples include elevation bands, hillslopes, sub-basins, grouped response units (GRUs), or hydrologic response units (HRUs). GRUs and HRUs are grid cells with similar land cover, elevation, slope, and/or aspect lumped into bins to maintain computational efficiency. Semi-distributed models typically do require a DEM as a basis for discretizing the watershed.
• Distributedmodels(“high”rank)explicitlyaccountforspatialvariabilityofinputvariables,typi-cally by dividing the watershed into equally sized grid cells. Within distributed models, HRUs can be implicitly achieved by, for example, assigning common soil classes or vegetation classes to groups of cells. Distributed models require a DEM as input.
The choice for a lumped, semi-distributed, or (fully) distributed model may affect the ability of a user to analyze specific land use or climate change scenarios. When considering complex forest manage-ment plans, the lumped structure is the least flexible, while fully distributed models are the most flexible. Lumped models have difficulty simulating the effects of spatial patterns of forest management because the location of individual cutblocks within a watershed cannot be represented. For example, in complex
8
mountainous terrain, hydrologic processes are strongly controlled by terrain elevation, slope, and aspect. In this situation, many of the questions regarding climate and land use change on watershed hydrology cannot be fully addressed because lumped models do not account for these variations (Whitaker et al. 2003). If only a fraction of the land area within an elevation band of a mountainous watershed was har-vested, average parameter values would have to be set over this entire area to account for a mix of forest and clearcut conditions, which is not physically realistic (Whitaker et al. 1998). However, lumped mod-els may still be useful for investigating the implications of various percentage cut levels on watershed hydrology without considering the details of the location of the cutblocks. Lumped models may also be useful in gently sloped terrain, where variations in terrain elevation, slope, and aspect are less important for determining the watershed hydrologic response to forest harvesting.
Semi-distributed models offer intermediate flexibility for representing complex forest management plans between the capabilities of lumped and fully distributed models. Furthermore, nuances exist between semi-distributed models, with the GRU and HRU approaches offering greater flexibility for representing harvesting plans compared to the relatively rigid watershed divisions such as elevation bands or sub-basins. Additional differences exist between GRUs and HRUs, with the latter concept be-ing somewhat more flexible for calculating runoff within a watershed. HRUs (Pomeroy et al. 2007) or “patches” (e.g., Band et al. 2001) allow mass (water) and energy to be transported between HRUs (i.e., lateral connectivity between HRUs is allowed for), while the GRU concept (e.g., Stahl et al. 2008) typ-ically does not allow for this interconnectivity.
The use of Digital Elevation Models (DEM) in semi-distributed and distributed models is an im-portant advantage because they can be used efficiently to calculate topographic factors such as slope, contributing area, aspect, and radiation shading in steep and complex terrain. These factors may be critical in determining the spatial distribution of snowmelt and evapotranspiration processes within a watershed. In addition, DEMs can be used with precipitation models and temperature lapse rates to estimate the climatic conditions across a watershed.
Soil (subsurface) vertical discretization was distinguished as (Kampf and Burges 2007) the following:
• Aconceptualseparationofdepthintounsaturated(abovewatertable)andsaturated(belowwatertable) zones (“low” rank)
• Anexplicitdiscretizationofdepth(“medium”rank)• Afinitedifferencediscretizationofthesubsurface(“high”rank)
No rank was applied for those models that lacked an explicit representation of soils. Subsurface verti-cal discretization may affect the ability of a model to simulate soil moisture conditions or subsurface runoff processes. Finite-difference discretizations of soil depths are typically reserved to groundwater models or integrated groundwater-surface water models that employ Richards’ equation for simulating subsurface flow. Most of the hydrologic models reviewed either define soil depths or divide the sub-surface conceptually into unsaturated and saturated zones, with infiltration and percolation occurring above the water table and lateral runoff occurring below the water table.
Vegetation vertical discretization was distinguished as follows (Table 2):
• Singlevegetationlayer(“low”rank)• Two-layervegetation(“medium”rank)• Multiplevegetationlayers(“high”rank)
No rank was applied for those models lacking an explicit representation of vegetation. Vegetation vertical discretization may affect the ability of a model to represent precipitation interception, evapo-transpiration, and snow accumulation and melt processes in multi-layered forest environments (e.g., overstorey canopy and understorey shrub) and may thus influence the accuracy of predictions regarding the hydrologic effects of forest harvesting or wildfire.
9
Model Temporal Discretization
Apart from spatial discretization, the time step at which model simulations are performed (i.e., temporal discretization) is also important. Some models can only run on a specific time interval (e.g., sub-daily, daily, or monthly). Time discretization may affect the ability of a model to provide outputs relevant to forest management (Section 2.3.5). In addition, temporal discretization has important implications for data availability and preparation (model complexity). For example, it should be considered that most climate stations report daily meteorological variables such as temperature and precipitation, while physi-cally based models are often best run at sub-daily time steps.
Model Complexity2.3.2
Choosing a model of appropriate complexity is equally important in considering the ability of the model to perform the desired land use or climate change scenarios. Generally, there is a trade-off between these two selection criteria, which has made it difficult to find suitable models that can make reliable predic-tions in operational applications, given the data and financial constraints of most planners. In many forested watersheds, data may be lacking, and this may force forest managers to select a relatively low-complexity lumped and/or empirical model. This choice may be adequate in many forest management situations. However, when more detailed results are needed, a fully distributed and/or physically based approach may be required, and it may be necessary to collect the detailed data to apply the model. In practice, medium-complexity models may often provide the best trade-off between data availability and functionality for addressing forest management questions. However, the medium-complexity models reviewed mostly employ a lumped or semi-distributed watershed discretization, and this will affect the ability to analyze complex forest management plans (Section 2.3.1).
For this review, model complexity was defined by the estimated data, resources, and time (which is a proxy for cost) that are required to parameterize and calibrate a model, as well as the professional judg-ment and experience required to operate these models (based on local experience with a sub-set of the reviewed models [Dr. Y. Alila, pers. comm., Jan. 2009]), (Table 3):
• Low-complexitymodelsaretypicallycharacterizedbymodestrequirementsregardingmeteo-rological-forcing data (monthly temperature and precipitation only), low overall model input requirements (typically less than 25 distinct input data), and can be applied by a single person to ungauged watersheds in a matter of days to a couple of weeks. These models are typically useful for screening-level studies that seek to assess the potential effects of forest management in relative terms and with limited accuracy. This could include assessing potential percent changes to annual yield due to various proposed cut levels for a watershed without defining streamflow volumes in absolute terms.
• Medium-complexitymodelsarecharacterizedbysomewhathigherrequirementsregardingmeteo-rological-forcing data (typically daily temperature and precipitation), medium overall model input requirements (about 25 to 75 distinct input data), and can be applied by a single person to water-sheds with only modest at-site data in a matter of a couple of weeks to a couple of months. These models are typically useful for intermediate-level planning studies that seek to assess the potential
effects of forest management in absolute terms (but still withlimited accuracy) and therefore re-quire some calibration (streamflow data only).
• High-complexitymodelsarecharacterizedbyhavingthegreatestrequirementsonmeteorologi-cal forcing data (hourly to daily temperature and precipitation at a minimum), high overall model input requirements (typically over 75 distinct input data), and may require substantial at-site data and calibration to streamflow data as well as internal watershed processes. Model setup and execu-tion may require a team of experts over several months, up to half a year of time. The use of these
10
Low
com
plex
ity
Med
ium
com
plex
ity
Hig
h c
ompl
exit
y
Dat
a re
quir
emen
tsM
onth
ly p
reci
pita
tion
an
d te
mp
erat
ure
Dai
ly p
reci
pita
tion
an
d te
mp
erat
ure
Hou
rly
to d
aily
pre
cipi
tati
on
and
tem
per
atu
re
No
addi
tion
al m
eteo
rolo
gica
l for
cin
gs r
equ
ired
Add
itio
nal
met
eoro
logi
cal f
orci
ngs
may
be
requ
ired
Add
itio
nal
met
eoro
logi
cal
forc
ings
may
be
requ
ired
No
nee
d fo
r m
ap d
ata
Req
uir
es m
ap d
ata
(DE
M, s
oils
, an
d fo
rest
cov
er)
Req
uir
es m
ap d
ata
(DE
M,
soils
, an
d fo
rest
cov
er)
Less
th
an 2
5 pa
ram
eter
sA
bou
t 25
to 7
5 in
put
para
met
ers
Typi
cally
ove
r 75
inpu
t pa
ram
eter
s
Para
met
ers
are
exp
erim
enta
lly b
ased
(n
o ca
libra
tion
re
quir
ed);
mod
els
suit
able
for
un
gau
ged
basi
ns
Min
imal
nu
mbe
r of
cal
ibra
tion
par
amet
ers
Som
e m
odel
app
licab
le to
un
gau
ged
basi
ns
Med
ium
to h
igh
nu
mbe
r of
ca
libra
tion
par
amet
ers;
mod
els
appl
icab
le to
gau
ged
basi
ns
only
Res
ourc
e re
quir
emen
tsLo
w d
ata
prep
roce
ssin
g ef
fort
Med
ium
dat
a pr
epro
cess
ing
effo
rtH
igh
dat
a pr
epro
cess
ing
effo
rt
Doe
s n
ot r
equ
ire
GIS
an
alys
isG
IS a
nal
ysis
req
uir
ed fo
r so
me
mod
els
GIS
an
alys
is r
equ
ired
for
mos
t m
odel
s
Can
be
com
plet
ed b
y on
e p
erso
nC
an m
ostl
y be
com
plet
ed b
y on
e p
erso
nM
ay r
equ
ire
proj
ect
team
Tim
e re
quir
emen
ts (
cost
1 )Le
ss t
han
2 w
eeks
Abo
ut
2 w
eeks
to 2
mon
ths
Abo
ut
2 to
6 m
onth
s
Less
th
an $
10 0
00A
bou
t $1
0 00
0–$4
0 00
0Fr
om a
bou
t $4
0 00
0 to
gr
eate
r th
an $
100
000
1 C
ost
esim
ates
bas
ed o
n q
uot
ed t
ime
requ
irem
ents
, 40
per
son
hou
rs p
er w
eek
and
a $1
00-h
ourl
y ra
te. T
he
hou
rly
rate
is t
ypic
al o
f
inte
rmed
iate
-lev
el c
onsu
ltan
t.
ta
bl
e 3
Mod
el c
omp
lexi
ty e
valu
atio
n cr
iteria
.
11
models is typically reserved for complex high-value or high-risk planning studies (e.g., optimization of cutblock and road locations to minimize the environmental consequences of forest management).
The above generalizations regarding relative model complexity provide reasonable guidance for most situations and are adequate for the purposes of providing decision support for model selection in Section 3. However, the guidelines are subjective and, within each of the three categories, specific re-quirements of individual models on data, resources, time, and calibration may vary. As such, the reader should refer to individual model descriptions (Appendix 1) for additional information.
Model Applicability to Climatic and Physiographic Settings2.3.3
The review of model applicability to particular climatic and physiographic settings considered the following factors:
• ApplicabilityofmodelstooneormoreofthefourprimaryhydrologicregimesinBritishColum-bia: rain-dominated, snowmelt-dominated, mixed/hybrid, and glacier-augmented (Eaton and Moore 2007). A similar categorization applies to Alberta.
• Terrainsetting,classifiedassteep(mountainousterrain),gradual(undulatingorflatterrain),orboth (indicated as “flexible”).
• Modelapplicabilitytoparticularwatershedsizes(i.e.,applicationscale)asdefinedinSection2.3.4.• Modelabilitytosimulateparticularprocessesthatmaybeimportantincertainwatersheds,includ-
ing groundwater flow, frozen soils (permafrost), lakes, and/or wetlands.
Rain Regimes
All of the models reviewed are able to handle rain-dominated conditions in principle. Model accuracy (functionality) in pluvial settings will be principally determined by the approaches (equations) adopted to simulate precipitation interception, evapotranspiration, and runoff processes, which was considered in Section 2.3.1. Models that employ the Penman-Monteith equation for calculating evapotranspiration should be preferred when high accuracy on site water-balance calculations is required, such as when accounting for the role of vegetation density expressed as leaf area index (LAI), or the dependence of canopy stomatal resistance on air vapour pressure deficit (relative humidity). This latter dependence simulates the closing of stomata under relatively dry conditions, decreasing canopy conductance to limit plant evapotranspiration. The accuracy of other model outputs, such as peak flows, will also depend on the simulation of runoff-generating processes, which in forested settings may often include preferential flow (e.g., Beckers and Alila 2004) and can be limited by the availability of reservoirs in a given model. For example, HBV-EC has only one fast and one slow reservoir, making it difficult to calibrate flashy response peaks from single rainfall events, seasonal peaks, and baseflow conditions equally.
Snow Regimes
Model applicability to snow regimes was based on the absence or presence of snow accumulation and melt equations. The different snowmelt calculation methods provide an example of the trade-offs between data requirements and model accuracy, which need consideration in an operational setting. Generally, temperature-index methods have been found to work well in predicting snowmelt in a for-ested environment (due to the strong correlation between longwave radiation and temperature), but not as well in open areas (where shortwave radiation may dominate the snowpack energy balance). However, this latter limitation can be overcome by calibrating the snowmelt indices to at-site data (e.g., Spittle-house and Winkler 2004). If this at-site data is lacking, using a degree-day method may be problematic for predicting the effects of harvesting. Energy-based snowmelt models are considered intrinsically more
12
accurate in predicting the effects of forest removal. However, in a data-limited mountainous setting, it is relatively straightforward to distribute temperature across a watershed to drive the empirical snowmelt model; more assumptions are needed to distribute solar radiation, air humidity, and wind speed over the same watershed, potentially leading to decreased accuracy from the physically based model. In some situations, analytical models may provide a good trade-off between data availability and accuracy. Hock (2003) has shown that a combined radiation–temperature-index snowmelt model can perform almost as well as an energy balance model when using meteorological data measured within the basin.
The physics of radiation transmission and scattering through a forest canopy is complex, and this may bias predicted irradiance at the snow surface (e.g., Thyer et al. 2004) and predicted snowmelt rates. As such, it is important to ensure that a physically based model does represent the physics correctly. This observation applies not only to snowmelt calculations, but also to other watershed processes.
Mixed Regimes
Model applicability to mixed regimes was based on the presence or absence of simulation of rain-on-snow (ROS) energy transfer. ROS events also occur in snowmelt regimes, but the inability to simulate these events may not be critical. ROS events are considerably more important in the rain/snow transi-tion zones (e.g., mid-elevation watersheds in coastal BC) and model ability to simulate these events should be a critical decision factor. The ability to simulate ROS events is limited to physically based, en-ergy balance snowmelt models, analytical models (mixed radiation-temperature approaches), and some temperature-index models than contain an empirical correction for rain events.
Glacier Regimes
Model applicability to represent glacier-augmented streamflow regimes was determined by the presence or absence of a glacial melt component. An assessment of the types of approaches taken or equations used to simulate glacier regimes was beyond the scope of this synthesis.
Model Application Scale and Grid Size Selection2.3.4
For many models, explicit guidance on model applicability to particular watershed sizes and an under-lying rationale is absent. Some models claim to have great versatility in possible scales of application. Therefore, scale recommendations made in this synthesis are approximate and interpretive and are based largely on a review of historical model applications. For the purposes of assessing scales at which models can be applied, small watersheds have been defined subjectively (based on the reviews) as having an area less than 100 km2, medium watersheds as being between 100 and 10 000 km2 in area, and large water-sheds as being greater than 10 000 km2 in area.
The absence of channel routing may dictate an upper limit for scale of model application. Models without a channel-routing component should normally be applied only to stand-level, water bal-ance questions or small watersheds with first-order streams (Strahler 1957), for which streamflow can be calculated as the sum of all runoff components. Using these models in watersheds that contain higher-order streams will misrepresent the length of surface and subsurface flow pathways (i.e., times of flow concentration at the outlet) and may not represent the hydrologic interaction between differ-ent sub-basins (e.g., de-synchronization of snowmelt runoff). Associated simulation errors will become progressively larger as the stream order increases. Models that do not contain explicit channel routing are referred to as “water balance models” in this report (Table 1).
For models that do incorporate channel-routing, factors determining an upper limit to model ap-plication scale could be the equations (i.e., empirical versus physically based) used to simulate channel
13
routing. The accuracy of these equations will become more important for predicting streamflows with increasing stream order. Another factor could be the lack of representation of groundwater flow, which, as scale increases, generally becomes an increasingly important factor in the overall water balance of a watershed. This general trend notwithstanding, groundwater can also be a dominant process in some headwater catchments (e.g., Whyte 2004); for example, in watersheds characterized by karst or other relatively permeable geologic materials. Groundwater processes are also relatively important in gradual terrain (Smerdon and Redding 2007).
For semi-distributed and distributed models, grid size selection (i.e., watershed discretization; Section 2.3.1) is intricately linked to application scale. When applying these models to large watersheds, import-ant considerations should include maintaining computational efficiency (smaller grid sizes result in more model cells causing longer run times and higher computer memory requirements), preserving ad-equate representation of spatial variability in the physiographic characteristics of the watershed, and the equations used to represent physical processes (for the latter two aspects, smaller grid sizes are better). For example, kilometre-scale grids are typically not appropriate for physically based models given that the validity of the governing mass or energy balance equations breaks down when spatial variability in watershed physical properties (e.g., hydraulic conductivity) is being overly lumped. This consideration, when combined with available computing power, may impose a practical upper limit to possible scales of application of physically based models.
Model suitability for application to small watersheds is typically also linked to grid size and process considerations. Centimetre-scale grids are typically only warranted for detailed process simulation, such as the use of the Richards’ equation for simulating variably saturated flow. In practice, the lower end of model grid size and application scale is often determined by data resolution (e.g., DEM, soil, or forest cover maps). The adequacy of the DEM resolution will be a function of the nature and scale of ter-rain features in the project area. In BC, the Terrain Resource Information Management (TRIM) DEMs are characterized by 25-m resolution1. Higher-resolution data is available in some cases. For example, Alberta is in the process of purchasing LiDAR data for most of the forested areas that are likely to be af-fected by mountain pine beetle.
In summary, determining model applicability to particular watershed sizes and the interrelated issue of grid size selection are model, data, and computing-power dependent, and require a considerable amount of professional judgement.
Model Outputs Relevant to Forest Management2.3.5
The model output assessment criteria considered in this review were based on the information required to inform forest management decisions, including, but not limited to, flood hazards, aquatic habitat, water availability, and potential for wetting up of sites. The model outputs needed to address these con-cerns (e.g., peak flows, low flows, water yield, soil moisture, snow water equivalent, runoff, and roads) were considered. The complete set of model output criteria is provided in Appendix 2.
The spatial and temporal scale at which model outputs are provided was also taken into account. The temporal scale of outputs depends on the time step(s) at which a model can be run. For some models, these time steps are fixed (hourly, daily, and monthly), but for others, time-step selection may be flexible. The spatial scale of model output (hereafter referred to as “planning scale”) is defined by the minimum landscape unit within a model for which forest-cover characteristics can be modified to represent harvesting or other disturbances. Thus, planning scale is different from application scale, but is closely tied to the watershed discretization in a model (Section 2.3.1).
1 http://ilmbwww.gov.bc.ca/bmgs/products/imagery/gridded.htm
14
For fully distributed models, the minimum landscape unit within a model for which forest-cover characteristics can be modified to represent harvesting or other disturbances is flexible and could theor-etically correspond to the size of a single grid cell. As such, the need for a certain minimum planning scale may be one of the considerations for model grid-size selection (Section 2.3.4). For semi-distributed models, the planning scale is defined by the landscape units used to represent the watershed such as sub-basins, hillslopes, elevation bands, GRUs, or HRUs. GRU or HRU selection may also depend on the forest management plan that needs to be analyzed. For lumped models, the planning scale is the entire watershed, given that spatial variability in forest cover within the watershed is not defined. The lumped and semi-distributed models are therefore the least flexible for analyzing forest management plans, while fully-distributed model approaches are the most flexible.
Model Adaptability to Represent Future Conditions2.3.6
For the purpose of short-term forest planning (e.g., to assess the immediate consequences of proposed harvesting and road construction), a model user might want to compare watershed hydrologic condi-tions with and without cutblocks and/or with and without roads using short-term data records that span a couple of years. In such a situation, a computer simulation will typically be characterized by constant (i.e., time invariant) land cover characteristics; model adaptability to represent changes in these characteristics during the course of the simulation (i.e., within the time period considered) is not critical. However, when conducting long-term simulation and harvest planning, it may become impor-tant whether model input can be easily adapted to represent the gradual or abrupt changes in climate or land use that might occur over the course of a simulation because of natural or human disturbances. This issue of model adaptability to represent future conditions was explored against the following model criteria (Appendix 2):
• Abilitytosimulateforestgrowth• Abilitytosimulateforestmortality• Abilitytoeasilyalterlandcoverdetailsduringthecourseofasimulationthroughtemporalinput
control• Adaptabilitytoincorporatefutureclimateconditionsandabilitytolinkthemodeltooutputfrom
regional climate models
These review criteria are explored further in the climate change section of this report (Section 5). In the context of forest growth, only the intrinsic ability of the model to simulate plant growth was con-sidered. An assessment of the types of equations (e.g., for simulating biomass accumulation based on water and nutrient availability), or the growth and yield curves used, was beyond the scope of this syn-thesis. The reader should therefore consider whether a model was developed for biogeoclimatic settings similar to or distinct from those encountered in BC or AB.
model calibration and Prediction confidence2.4
Considerations regarding model calibration requirements were incorporated into the evaluation of model complexity (Section 2.3.2). Calibration (i.e., the process of adjusting parameters such that model results match field measurements such as streamflow) and model predictive uncertainty (i.e., the level of confidence in model predictions) are important concerns in the application of hydrologic models (Abbott et al. 1986a; Beven 1989; Grayson et al. 1992a, 1992b; Beven and Binley 1992; Refsgaard 1997). A brief discussion on these topics is therefore warranted.
15
In general, a weakness of fully distributed, physically based models is the large number of parameters involved, often leading to high calibration requirements and issues such as parameter equivalence or equifinality (e.g., Beven and Binley 1992; Refsgaard 1997). Equifinality refers to when multiple model parameter combinations lead to the same streamflow calibration, but may result in different answers re-garding the effects of forest management. Even if a calibration to streamflow data is deemed satisfactory, based on an established set of criteria expressing fit between model and data, the model may provide the right answer for the wrong reasons, with the interaction of several errors leading to an apparently correct outcome (e.g., Seibert and McDonnell 2002). For example, an error in simulating precipitation inter-ception by the forest canopy could be offset by another error in simulating rainfall runoff to the stream channel. These equifinality errors might not be evident during calibration, but will provide invalid predictions when forecasting the impact of disturbance. Thus, the higher intrinsic accuracy of physically based models may not be fully realized if the watershed application is not supported by adequate data, or if insufficient attention is being paid to the consequences of parameterization and calibration.
Lumped or semi-distributed models will have lower data and calibration requirements than fully distributed models and this may become an important model selection criterion in watersheds with little data. Empirical models generally also have lower data and calibration requirements compared to physically based models. However, this is not always the case. Some empirical watersheds models (e.g., Hydrologic Simulation Program–Fortran [HSPF]) may be quite parameter intensive and will offer little advantage with respect to parameterization and calibration over physically based models.
The presence or absence of internal algorithms for model calibration and prediction-confidence analysis was not considered a separate criterion in model review and selection. Models such as HSPF, the Soil Water Assessment Tool (SWAT; Gassman et al. 2007) and the Hydrologic Engineering Center’s Hydrologic Modelling System (HEC-HMS; US-ACE 2000) contain internal components for auto-mated sensitivity, calibration (parameter estimation), and uncertainty (prediction confidence) analysis. However, external procedures such as Monte Carlo simulation (Demaria et al. 2007), the generalized likelihood uncertainty estimation (GLUE; Beven and Binley 1992), and the shuffled complex evolution (SCE; Kuczera 1994) have also been developed. These external procedures allow any model (i.e., hydro-logic or other) to be calibrated and subjected to prediction-confidence analyses.
The availability of sophisticated tools for calibration and uncertainty analysis does not eliminate inherent issues such as parameter equivalence or equifinality. Therefore, these tools cannot eliminate the need for collecting relevant data regarding internal watershed processes (Alila and Beckers 2001). This data can be used to confirm that watershed sub-processes are being modelled accurately, will help nar-row down parameter limits (reducing model uncertainty), and/or may eliminate some parameters from the calibration process (addressing the issue of equifinality).
limitations of model review and ranking2.5
The authors have exercised reasonable skill, care, and diligence to assess the information acquired during the preparation of this model synthesis, but make no guarantees as to its accuracy or completeness. The information contained in this report is based on a review of publicly available information contained in user manuals, published model applications, and other reference materials. The review did not include hands-on testing of the models and, as such, is solely based on information generated by others. Some inferences regarding model capabilities and limitations may therefore be interpretive and/or approxi-mate. Some models are better documented than others and this may be reflected in this synthesis. The review did not specifically address accuracy of the information generated by others; while the informa-tion is believed to be accurate, it cannot be guaranteed. The reader should refer to the original model theory and user guides whenever possible and/or whenever in doubt.
16
Cat
egor
yP
urp
ose
Mod
el n
ame
Dis
cret
izat
ion
Wat
ersh
ed p
roce
sses
Roa
dsO
ther
feat
ure
s
Watershed
Soil layers
Vegetation layers
Canopy rain interception
Fog drip
Stemflow
Canopy snow interception
Snowmelt
Snowdrift/ blowing snow
Vegetation transpiration
Soil evaporation
Infiltration
Depression storage
Shallow subsurface runoff
Preferential flow/macropores
Horton overland flow
Saturation overland flow
Channel routing
Hillslope runoff interception
Precipitation interception
Flow diversion
Culverts
Groundwater
Glacier melt
Frozen soil (permafrost)
Lakes
Wetlands
Water control structures
Fore
st
hydr
olog
y m
odel
s
An
nu
al y
ield
WR
EN
SS
(Win
Wrn
sHyd
/E
CA
-AB
)
LL
LM
M
Peak
flow
UB
C-U
F Pe
ak
Flow
Mod
elH
LL
LL
LL
LL
LL
L
Wat
er b
alan
ceB
RO
OK
90L
LL
HL
LH
HL
LL
LL
ForH
yML
ML
HL
LL
LL
L
ForW
aDy
LM
MH
LM
HH
LL
Wat
ersh
ed
hydr
olog
yD
HSV
MH
MM
HH
HH
HM
MM
MM
MM
HH
RH
ESS
ysM
LH
HL
MH
HM
MM
MM
MM
MH
HL
Oth
er m
odel
sW
ater
bal
ance
HE
LPL
ML
LL
MM
LL
L
Wat
ersh
ed
hydr
olog
yA
CR
UM
ML
HL
LM
ML
LL
ML
ML
M
CR
HM
HL
LH
HH
LH
HL
LL
LM
HB
V-E
CM
LL
LL
LL
LL
LL
LL
MM
HE
C-H
MS
ML
LL
LL
LL
LL
LL
HL
HM
M
HSP
FL
LL
LM
LL
LL
LL
LH
LM
L
PR
EV
AH
HL
LH
MH
HL
LM
LM
M
PR
MS/
MM
SM
LL
LL
MM
ML
LL
LL
L
SWA
TM
ML
LL
HH
MM
MM
MM
L
UB
CW
MM
LL
LL
ML
LL
LL
LM
LM
MM
WaS
iM-E
TH
HM
MH
MH
HH
MM
MM
HM
MM
Wat
floo
dM
LL
ML
LL
ML
LM
ML
LM
MM
Gro
un
dwat
er-
surf
ace
wat
er
mod
els
InH
M
(Hyd
roG
eoSp
her
e)H
HL
MM
MH
MH
HH
HM
HH
ML
HH
M
Mik
e-SH
EH
HL
ML
MM
MH
HH
HH
HL
M
MO
DH
MS
HH
LM
MM
HH
HH
HM
HH
ML
Soil
eros
ion
WE
PP
MM
LL
LL
LM
MM
LL
LL
Riv
er b
asin
m
odel
sSS
AR
RM
LL
LL
LL
LL
LL
M
VIC
MM
LH
HH
LH
HL
ML
LL
LL
L
ta
bl
e 4
Mod
el fu
nctio
nalit
y ra
nkin
g; r
efer
to
Tabl
e 2
and
Sect
ion
2 fo
r ev
alua
tion
crite
ria.
17
decision suPPort for selecting models3
The outcomes of the model review and ranking (Appendices 1 and 2) are summarized in Tables 4 to 8. Figure 1 organizes the models according to overall functionality (based on the detailed review of in-dividual processes and factors in Table 4) and complexity (Table 5). Tables 6 to 8 are used as decision support for selecting models. The sections below provide a step-wise approach to selecting models (Sec-tion 3.1) and offer the main advantages and disadvantages of models interpreted to be most suitable for addressing forest management questions (Section 3.2).
WRENSS
UBC-UF Peak Flow Model
BROOK90
DHSVM
ForHyM
ForWaDy
RHESSys
HELP
ACRU (with snow routines)
CRHM
HBV-EC
HEC-HMS
HSPF
PREVAH
PRMS/MMS
UBCWM
WaSiM-ETH
Watf lood
InHM (HydroGeoSphere)Mike-SHE
MOD-HMS
SSARR
SWAT
WEPP
Mod
el F
unct
iona
lity
for F
ores
t M
anag
emen
t
Model Complexity Low
(Screening Level Studies) High
(Complex Studies)
Low
est
Hig
hest
Medium
Inte
rmed
iate
f igure 1 Models organized according to overall functionality (based on detailed review of individual processes and factors in Table 4).
A step-wise Approach to selecting models 3.1
One of the outcomes of this review is that there currently is no “best” model for use in an operational forest management context, i.e., an easy-to-use model with low data requirements that is very accu-rate and that can be applied under all circumstances. Site-specific, tailor-made model approaches are therefore needed. Selecting an appropriate model for a particular forest management application is a complicated process, and needs to take into account a variety of considerations, as discussed in Section 2. Allowing for proper lead time in planning a model study is therefore extremely important.
18
The first step in selecting a model for a particular project is to assess data, time, and resource con-straints. These constraints then need to be balanced with expectations for the model study. For example, if there is an inherent conflict between study constraints and expectations (e.g., a model study for a high-value watershed is being initiated with little or no data available), then identifying a suitable model approach cannot be expected. As such, it is important to identify expected outcomes of the study early (e.g., through stakeholder consultation) and to make efforts to align the data, time, and resources that are available to the model component accordingly. Alternatively, if constraints on the model study can-not be alleviated, it may be necessary to scale back expectations on study’s outcomes.
table 5 Model complexity ranking; refer to Table 3 and Section 2 for evaluation criteria.
Category Purpose Model name Model complexity
Forest hydrology models
Annual yield WRENSS (WinWrnsHyd/ECA-AB) Low
Peak flow UBC-UF Peak Flow Model Medium
Water balance BROOK90 Medium
ForHyM Medium
ForWaDy Medium
Watershed hydrology
DHSVM High
RHESSys High
Other models Water balance HELP Medium
Watershed hydrology
ACRU High
CRHM High
HBV-EC Medium
HEC-HMS High
HSPF High
PREVAH High
PRMS/MMS High
SWAT High
UBCWM Medium
WaSiM-ETH High
Watflood High
Groundwater-surface water models
InHM (HydroGeoSphere) High
Mike-SHE High
MODHMS High
Soil erosion WEPP High
River basin models SSARR High
VIC High
Proceeding from the assessment of data support and resources, a model of suitable functionality would be evaluated against expectations of the modelling study. A relatively straightforward, stepped ap-proach towards model selection is:
Step 1: Use data, time, and resource constraints to determine an appropriate model complexity, using Tables 3 and 5. For example, if the budget for a study is between about $10 000 and $40 000 and the subject watershed is ungauged, or if only streamflow data are available, the logical choice would be to select a model of either low or intermediate complexity (Table 5). If the study needs to be completed within 2 weeks, a low-complexity model should be selected; however, if more time is
19
available, an intermediate-complexity model may be better. The most stringent constraint (either data, time, budget, or resources) will typically dictate the selection of a model-complexity category, although some of the restrictions (e.g., need for GIS analysis) may be relatively straightforward to overcome.
Step 2: The normal first choice would be to select the top-ranked model in the chosen model com-plexity category using Figure 1. In the medium-complexity category, this is the UBC Watershed Model (UBCWM), while for high-complexity models, it is DHSVM.
Step 3: Assess whether the model can address the forest management question(s) of interest, using Table 6. This step addresses questions regarding model capability to analyze the forest management plan being considered, including whether the spatial layout of disturbed areas in the watershed needs to be represented, whether road construction and management should be simulated, and/or whether forest-growth simulation capability is needed.
Step 4: Confirm that the model can be applied to the climatic and physiographic region of interest, using Table 7. This step assesses whether snowmelt, rain-on-snow (mixed regimes), and or glacial melt capability is needed, and/or whether groundwater, lakes, or wetlands are important factors to consider. Whether the model under consideration is applicable to the terrain setting (steep or gradual) and the size of the watershed (small, medium, or large) should also be considered.
Step 5: Determine if the model will generate the required outputs to support assessment at the appropriate planning scale and time scale using Table 8. For example, if the spatial layout of cut-blocks in a complex forest management plan needs to be analyzed, it is important that the chosen model has a flexible planning scale (i.e., the model should be fully distributed). When considering instantaneous peak flows or situations where short-duration, high-intensity precipitation events or diurnal fluctuations in meteorological conditions are important, it becomes essential that the model be able to run at an hourly time step. On the other hand, if only a site water balance needs to be determined, it may be adequate if the model can provide output at monthly time scales. It is important that these kind of expected outcomes of the model study are recognized early to inden-tify the required model inputs and outputs.
Step 6: Consider the main advantages and disadvantages of the selected model(s) (Table 9 and Section 3.2) and conduct a detailed review of the model (Appendices 1 and 2) to ensure that the expecta-tions of the modelling study will be met.
If the model does not appear to satisfy the requirements and constraints of the study, the normal next step would be to consider the next highest-ranked model (e.g., BROOK90 in the medium-complexity category; Figure 1) and repeat the above model-evaluation steps. This model-selection process would be repeated until a satisfactory model is identified. In a case where no appropriate model can be identified within the chosen model-complexity category, managers should consider whether either a more sim-plistic (low-complexity) or advanced (high-complexity) approach to the model study is warranted, or whether expectations on the model study should be scaled back (i.e., perhaps the model cannot be used to provide certain outputs).
model Advantages and disadvantages 3.2
While model selection will often be a site-specific process, this review (Appendices 1 and 2) indicates the general advantages and disadvantages of each model for use in an operational forest management con-text (Table 9). This can help resource managers narrow down their choice of an appropriate model. The discussion below focuses on WRENSS in the low-complexity category; UBCWM, BROOK90, ForWaDy, and the UBC-UF Peak Flow Model in the medium-complexity cateogry; and DHSVM, the Regional Hydro-Ecologic Simulation System (RHESSys), the Wasserhaushalts-Simulations-Modell (WaSiM-ETH), and the Cold Regions Hydrologic Model (CRHM) in the high-complexity category. Jointly,
20
Cat
egor
yP
urp
ose
Mod
el n
ame
Wat
ersh
ed
disc
reti
zati
on1
Sim
ula
tion
of
fore
st
harv
esti
ng1,
2
Fore
st
grow
th3
Roa
d co
nst
ruct
ion
an
d m
anag
emen
t4
Fore
st
hydr
olog
y m
odel
s
An
nu
al y
ield
WR
EN
SS (
Win
Wrn
sHyd
/EC
A-A
B)
Lum
ped
Mix
edN
o5N
o
Peak
flow
UB
C-U
F Pe
ak F
low
Mod
elD
istr
ibu
ted
Em
piri
cal
No
As
Hor
ton
ian
ove
rlan
d fl
ow a
reas
Wat
er b
alan
ceB
RO
OK
90Lu
mp
edM
ixed
No
No
ForH
yMLu
mp
edE
mpi
rica
lN
oN
o
ForW
aDy
Lum
ped
Mix
edN
o5N
o
Wat
ersh
ed
hydr
olog
yD
HSV
MD
istr
ibu
ted
Phy
sica
lN
oR
oad
drai
nag
e n
etw
ork
and
stru
ctu
res
RH
ESS
ysH
iera
rch
ical
An
alyt
ical
Yes
Roa
d dr
ain
age
net
wor
k an
d st
ruct
ure
s
Oth
er
mod
els
Wat
er b
alan
ceH
ELP
Soil
colu
mn
Mix
edYe
sN
o
Wat
ersh
ed
hydr
olog
yA
CR
USu
b-ba
sin
sM
ixed
No5
No
CR
HM
HR
Us
Phy
sica
lN
oN
o
HB
V-E
CG
RU
sE
mpi
rica
lN
oN
o
HE
C-H
MS
Sub-
basi
ns
Em
piri
cal
No
No
HSP
FLu
mp
edM
ixed
No
No
PR
EV
AH
GR
Us
An
alyt
ical
No
No
PR
MS/
MM
SG
RU
sM
ixed
Yes
No
SWA
TH
iera
rch
ical
Mix
edYe
sN
o
UB
CW
ME
leva
tion
ban
dsM
ixed
No
No
WaS
iM-E
TH
Dis
trib
ute
dA
nal
ytic
alN
oN
o
Wat
floo
dG
RU
sM
ixed
No
No
Gro
un
dwat
er-
surf
ace
wat
er
mod
els
InH
M (
Hyd
roG
eoSp
her
e)D
istr
ibu
ted
N/A
No
1-D
ove
rlan
d fl
ow s
egm
ents
Mik
e-SH
ED
istr
ibu
ted
Mix
edN
oN
o
MO
DH
MS
Dis
trib
ute
dE
mpi
rica
lN
oN
o
Soil
eros
ion
WE
PP
Hill
slop
esM
ixed
Yes
No
Riv
er b
asin
mod
els
SSA
RR
Ele
vati
on b
ands
Em
piri
cal
No
No
VIC
Stat
isti
cal
Phy
sica
lN
oN
o
1 R
efer
to S
ecti
on 2
for
eval
uat
ion
cri
teri
a.
2
Sim
ula
tion
of
soil-
vege
tati
on-a
tmos
pher
e tr
ansf
er p
roce
sses
con
side
red.
3
“No”
impl
ies
that
mod
el d
oes
not
dir
ectl
y si
mu
late
fore
st g
row
th. I
n m
ost
case
s, fo
rest
gro
wth
can
be
repr
esen
ted
as m
odel
inpu
t. 4
On
ly m
odel
s w
ith
exp
licit
pro
visi
on fo
r si
mu
lati
ng
the
effe
ct o
f ro
ads
con
side
red.
In
mos
t m
odel
s ro
ads
can
be
appr
oxim
ated
as
imp
ervi
ous
or o
verl
and
flow
are
as.
5
Gro
wth
equ
atio
ns
can
be
prog
ram
med
into
WR
EN
SS. F
orW
AD
y ca
n b
e in
terf
aced
wit
h F
OR
EC
AST
or
FOR
CE
E m
odel
s; A
CR
U h
as a
cro
p
yi
eld
com
pon
ent.
ta
bl
e 6
Mod
el a
pp
licab
ility
to
fore
st m
anag
emen
t.
21
Cat
egor
yP
urp
ose
Mod
el n
ame
Clim
atic
reg
ime
Terr
ain
se
ttin
gW
ater
shed
sca
leG
rou
ndw
ater
Gla
cial
mel
tFr
ozen
soi
l (p
erm
afro
st)
Lake
sW
etla
nds
Fore
st
hydr
olog
y m
odel
s
An
nu
al y
ield
WR
EN
SS
(Win
Wrn
sHyd
/E
CA
-AB
)
Rai
n o
r sn
owFl
exib
leSm
all t
o M
ediu
mN
oN
oN
oN
oN
o
Peak
flow
UB
C-U
F Pe
ak
Flow
Mod
elR
ain
or
snow
Flex
ible
Smal
l to
Larg
eN
oN
oN
oN
oN
o
Wat
er b
alan
ceB
RO
OK
90R
ain
or
snow
Gra
dual
Smal
lE
mpi
rica
lN
oN
oN
oN
o
ForH
yMR
ain
or
snow
Gra
dual
Smal
lN
oN
oN
oN
oN
o
ForW
aDy
Rai
n/s
now
/mix
edG
radu
alSm
all
No
No
No
No
No
Wat
ersh
ed
hydr
olog
yD
HSV
MR
ain
/sn
ow/m
ixed
Stee
pSm
all t
o M
ediu
mN
oN
oN
oN
oN
o
RH
ESS
ysR
ain
/sn
ow/m
ixed
Flex
ible
Smal
l to
Med
ium
Em
piri
cal
No
No
No
No
Oth
er m
odel
sW
ater
bal
ance
HE
LPR
ain
/sn
ow/m
ixed
Gra
dual
Smal
lN
oN
oYe
sN
oN
o
Wat
ersh
ed
hydr
olog
yA
CR
UR
ain
(sn
ow/m
ixed
1 )Fl
exib
leSm
all t
o M
ediu
mE
mpi
rica
lN
oN
oYe
sYe
s
CR
HM
Rai
n/s
now
/mix
edG
radu
alSm
all t
o M
ediu
mE
mpi
rica
lN
oYe
sN
oN
o
HB
V-E
CR
ain
or
snow
Stee
pSm
all t
o M
ediu
m
Em
piri
cal
Yes
No
Yes
No
HE
C-H
MS
Rai
n o
r sn
owG
radu
alSm
all t
o La
rge
Em
piri
cal
No
No
Yes
Yes
HSP
FR
ain
/sn
ow/m
ixed
Gra
dual
Smal
l to
Larg
eE
mpi
rica
lN
oN
oYe
sN
o
PR
EV
AH
Rai
n/s
now
/mix
edSt
eep
Smal
l to
Med
ium
Em
piri
cal
Yes
No
No
No
PR
MS/
MM
SR
ain
/sn
ow/m
ixed
Flex
ible
Smal
l to
Larg
eE
mpi
rica
lN
oN
oN
oN
o
SWA
TR
ain
or
snow
Gra
dual
Smal
l to
Larg
eP
hysi
cal
No
No
Yes
No
UB
CW
MR
ain
/sn
ow/m
ixed
Stee
pSm
all t
o M
ediu
mE
mpi
rica
lYe
sN
oYe
sN
o
WaS
iM-E
TH
Rai
n/s
now
/mix
edFl
exib
leSm
all t
o La
rge
Phy
sica
lYe
sN
oYe
sN
o
Wat
floo
dR
ain
or
snow
Gra
dual
Smal
l to
Larg
eE
mpi
rica
lYe
sN
oYe
sYe
s
Gro
un
dwat
er-
surf
ace
wat
er
mod
els
InH
M
(Hyd
roG
eoSp
her
e)R
ain
Flex
ible
Smal
l to
Larg
eP
hysi
cal
No
No
Yes
Yes
Mik
e-SH
ER
ain
or
snow
Gra
dual
Smal
l to
Larg
eP
hysi
cal
No
No
Yes
Yes
MO
DH
MS
Rai
n
Gra
dual
Smal
l to
Larg
eP
hysi
cal
No
No
Yes
Yes
Soil
eros
ion
WE
PP
Rai
n/s
now
/mix
edFl
exib
leSm
all
No
No
No
No
No
Riv
er b
asin
m
odel
sSS
AR
RR
ain
/sn
ow/m
ixed
Flex
ible
Med
ium
to L
arge
Em
piri
cal
No
No
Yes
No
VIC
Rai
n/s
now
/mix
edFl
exib
leM
ediu
m to
Lar
geN
oN
oYe
sYe
sYe
s
1
Snow
rou
tin
es a
dded
at
the
Un
iver
sity
of
Leth
brid
ge; n
ot p
art
of t
he
orig
inal
mod
el d
istr
ibu
ted
by t
he
Un
iver
sity
of
Nat
al.
ta
bl
e 7
Mod
el a
pp
licab
ility
to
clim
atic
and
phy
siog
rap
hic
sett
ings
; ref
er t
o Se
ctio
n 2
for
eval
uatio
n cr
iteria
.
22
Cat
egor
yP
urp
ose
Mod
el n
ame
Pla
nn
ing
scal
e1Te
mpo
ral s
cale
Mod
el o
utp
uts
2
FHA
YP
FLF
SWE
TW
BSM
IFW
TO
FSF
MF
GF
RO
RF
SEM
WN
FW
Q
Fore
st h
ydro
logy
m
odel
sA
nn
ual
yi
eld
WR
EN
SS
(Win
Wrn
sHyd
/ E
CA
-AB
)
Wat
ersh
edA
nn
ual
X3
XX
Peak
flow
UB
C-U
F Pe
ak F
low
Mod
elFl
exib
leD
aily
33
XX
XX
XX
XX
Wat
er
bala
nce
BR
OO
K90
Wat
ersh
edD
aily
44
44
XX
XX
XX
XX
XX
X
ForH
yMW
ater
shed
Mon
thly
44
44
XX
XX
XX
X
ForW
aDy
Wat
ersh
edD
aily
44
44
XX
XX
XX
X
Wat
ersh
ed
hydr
olog
yD
HSV
MFl
exib
leH
ourl
yX
XX
XX
XX
XX
XX
XX
XX
X
RH
ESS
ysPa
tch
es (
HR
Us)
Dai
lyX
XX
XX
XX
XX
XX
XX
XX
X
Oth
er m
odel
sW
ater
bal
ance
HE
LPSo
il co
lum
nD
aily
XX
XX
X
Wat
ersh
ed
hydr
olog
yA
CR
USu
b-ba
sin
Dai
lyX
XX
X5
XX
XX
XX
XX
XX
X
CR
HM
HR
UH
ourl
y6
66
6X
XX
XX
XX
XX
HB
V-E
CG
RU
Dai
ly6
66
6X
XX
XX
XX
XX
HE
C-H
MS
Sub-
basi
nSu
b-da
ilyX
XX
XX
XX
XX
XX
XX
HSP
FW
ater
shed
Sub-
daily
XX
XX
XX
XX
XX
XX
XX
X
PR
EV
AH
GR
UH
ourl
yX
XX
XX
XX
XX
XX
XX
PR
MS/
MM
SG
RU
Dai
ly6
66
6X
XX
XX
XX
XX
XX
SWA
TG
RU
Dai
lyX
XX
XX
XX
XX
XX
XX
XX
XX
UB
CW
ME
leva
tion
ban
dD
aily
XX
XX
XX
XX
XX
XX
X
WaS
iM-E
TH
Flex
ible
Sub-
daily
XX
XX
XX
XX
XX
XX
XX
X
Wat
floo
dG
RU
Hou
rly
XX
XX
XX
XX
XX
XX
X
Gro
un
dwat
er-
surf
ace
wat
er
mod
els
InH
M
(Hyd
roG
eoSp
here
)Fl
exib
leFl
exib
leX
XX
XX
XX
XX
XX
XX
XX
X
Mik
e-SH
EFl
exib
leFl
exib
leX
XX
XX
XX
XX
XX
XX
XX
MO
DH
MS
Flex
ible
Flex
ible
XX
XX
XX
XX
XX
XX
XX
Soil
eros
ion
WE
PP
Hill
slop
eD
aily
66
66
XX
XX
XX
XX
XX
Riv
er b
asin
m
odel
sSS
AR
RE
leva
tion
ban
dSu
b-da
ilyX
XX
XX
XX
XX
XX
XX
VIC
25 k
m2
Sub-
daily
XX
XX
XX
XX
XX
XX
1 Sm
alle
st p
ossi
ble
plan
nin
g sc
ale
liste
d. F
lexi
ble
indi
cate
s th
at s
ilvic
ult
ura
l pra
ctic
es c
an b
e si
mu
late
d u
nde
r a
ran
ge o
f co
nfi
gura
tion
s (d
istr
ibu
ted
mod
els)
. 2
Abb
revi
atio
ns:
FY
= fu
ll hy
drog
raph
, AY
= a
nn
ual
yie
ld, P
F =
pea
k fl
ow, L
F =
low
flow
, SW
= s
now
wat
er e
quiv
alen
t (s
now
cov
er),
ET
= e
vapo
tran
spir
atio
n, W
B =
wat
er b
alan
ce (
for
soil
colu
mn
an
d/
or b
asin
), S
M =
soil
moi
stu
re, I
F =
infi
ltra
tion
, WT
= w
ater
tabl
e, O
F =
ove
rlan
d fl
ow, S
F =
shal
low
subs
urf
ace
flow
, MF
= m
acro
pore
(pr
efer
enti
al)
flow
, GF
= g
rou
ndw
ater
flow
(ba
sefl
ow),
RO
= b
asin
tota
l ru
nof
f, R
F =
roa
d fl
ow, S
E =
sed
imen
t So
il E
rosi
on, M
W =
mas
s w
asti
ng,
NF
= n
utr
ien
t fl
uxe
s,W
Q =
wat
er q
ual
ity.
3
Cla
imed
cap
abili
ty b
ut
un
test
ed.
4 St
ream
flow
sim
ula
tion
cap
abili
ty fo
r fi
rst
orde
r w
ater
shed
s on
ly (
sum
of
all r
un
off
com
pon
ents
); m
odel
s do
not
inco
rpor
ate
chan
nel
rou
tin
g.5
Snow
rou
tin
es h
ave
been
add
ed b
y D
r. St
efan
Kie
nzl
e (U
niv
ersi
ty o
f Le
thbr
idge
) bu
t ar
e n
ot p
art
of t
he
orig
inal
sof
twar
e di
stri
bute
d by
Sou
th A
fric
an d
evel
oper
s.
6 N
o ex
plic
it c
han
nel
rou
tin
g; s
impl
ified
pro
cedu
res
are
use
d. S
ee n
arra
tive
des
crip
tion
of
each
mod
el (
App
endi
x 1)
.
ta
bl
e 8
Mod
el o
utp
uts
for
fore
st p
lann
ing;
ref
er t
o A
pp
endi
x 1
for
mod
el d
escr
iptio
ns.
23
Cat
egor
yP
urp
ose
Mod
el n
ame
Mai
n a
dvan
tage
sM
ain
dis
adva
nta
ges
Fore
st
hydr
olog
y m
odel
s
An
nu
al y
ield
WR
EN
SS
(Win
Wrn
sHyd
/EC
A-A
B)
Eas
y to
use
pro
cedu
res
wit
h lo
w d
ata
requ
irem
ents
Lim
ited
fun
ctio
nal
ity
(an
nu
al y
ield
, hyd
rolo
gic
reco
very
)
Peak
flow
UB
C-U
F Pe
ak F
low
Mod
elV
alu
able
pro
vin
ce-w
ide
(BC
) da
taba
ses
prov
ided
Lim
ited
fun
ctio
nal
ity
(pea
k fl
ows)
; lim
ited
mod
el te
stin
g to
dat
e
Wat
er b
alan
ceB
RO
OK
90U
sefu
l for
sit
e w
ater
bal
ance
s; m
oder
ate
com
plex
ity
Em
piri
cal d
egre
e-da
y m
eth
od fo
r sn
owm
elt;
no
chan
nel
rou
tin
g
ForH
yMU
sefu
l for
sit
e w
ater
bal
ance
s; m
oder
ate
com
plex
ity
Em
piri
cal s
now
mel
t an
d ev
apot
ran
spir
atio
n
met
hod
s; n
o ch
ann
el r
outi
ng
ForW
aDy
Lin
kage
wit
h fo
rest
gro
wth
mod
els;
mod
erat
e co
mpl
exit
yN
o ch
ann
el r
outi
ng;
lim
ited
mod
el d
ocu
men
tati
on a
nd
test
ing
Wat
ersh
ed
hydr
olog
yD
HSV
MPo
wer
ful f
or a
wid
e ra
nge
of
wat
ersh
ed h
ydro
logy
app
licat
ion
s D
iffi
cult
to u
se; f
un
ctio
nal
ity
may
be
limit
ed to
mou
nta
inou
s w
ater
shed
s
RH
ESS
ysPo
ten
tial
for
ecoh
ydro
logy
app
licat
ion
(fo
rest
gro
wth
, mor
talit
y)D
iffi
cult
to u
se w
ith
hig
h m
odel
par
amet
eriz
atio
n r
equ
irem
ents
Oth
er m
odel
sW
ater
bal
ance
HE
LPR
elat
ivel
y ea
sy to
use
mod
el fo
r si
te w
ater
bal
ance
sN
ot d
evel
oped
for
fore
st m
anag
emen
t or
wat
ersh
ed s
cale
app
licat
ion
s
Wat
er B
alan
ce M
odel
by
QU
ALH
YM
OU
rban
sto
rmw
ater
man
agem
ent
mod
el (
best
man
agem
ent
prac
tice
s)N
ot s
uit
able
for
fore
st m
anag
emen
t ap
plic
atio
ns
Wat
ersh
ed
hydr
olog
yA
CR
U
(wit
h s
now
rou
tin
es)
Flex
ible
mod
el; u
ser-
frie
nd
ly m
enu
s; b
ein
g u
sed
at U
. of
Leth
brid
geC
ompl
ex m
odel
; ori
gin
ally
dev
elop
ed fo
r So
uth
Afr
ican
con
diti
ons
CR
HM
Stro
ng
focu
s on
col
d re
gion
s hy
drol
ogy
(blo
win
g sn
ow/f
roze
n s
oils
)Si
mpl
ified
ch
ann
el r
outi
ng;
fun
ctio
nal
ity
limit
ed to
bor
eal f
ores
ts
HB
V-E
CG
reen
Ken
ue
GU
I of
fers
eas
e of
use
; can
han
dle
gla
cial
mel
tE
mpi
rica
l wat
ersh
ed p
roce
ss a
nd
fore
st c
over
rep
rese
nta
tion
HE
C-H
MS
Focu
s on
ru
nof
f, ch
ann
el r
outi
ng
and
wat
er c
ontr
ol s
tru
ctu
res
Em
piri
cal w
ater
shed
pro
cess
sim
ula
tion
; fu
nct
ion
alit
y lim
ited
to r
iver
pro
cess
es
HSP
FFo
cus
on w
ater
qu
alit
y in
nat
ura
l an
d m
an-m
ade
wat
er s
yste
ms
Em
piri
cal w
ater
shed
pro
cess
an
d fo
rest
cov
er
repr
esen
tati
on; d
iffi
cult
to u
se
PR
EV
AH
Win
PR
EV
AH
GU
I of
fers
eas
t of
use
; can
han
dle
gla
cial
mel
tE
mpi
rica
l sim
ula
tion
of
run
off
proc
esse
s; n
o te
stin
g in
fore
sted
env
iron
men
ts
PR
MS/
MM
SFo
cus
on w
ater
qu
alit
y in
nat
ura
l an
d m
an-m
ade
wat
er s
yste
ms
Em
piri
cal w
ater
shed
pro
cess
rep
rese
nta
tion
; no
expl
icit
ch
ann
el r
outi
ng
SWA
TFo
cus
on w
ater
qu
anti
ty a
nd
qual
ity
and
repr
esen
atio
n o
f gr
oun
dwat
erSn
ow p
roce
ss r
epre
sen
tati
on r
equ
ires
impr
ovem
ent
UB
CW
ME
xten
sive
ly u
sed;
mod
est
data
req
uir
emen
ts; g
laci
ers
and
upl
and
lake
sE
leva
tion
ban
ds a
nd
sim
plifi
ed fo
rest
cov
er li
mit
fo
rest
man
agem
ent
fun
ctio
nal
ity
WaS
iM-E
TH
Can
han
dle
gla
cial
mel
t, gr
oun
dwat
er, l
akes
an
d re
serv
oirs
Com
plex
mod
el; n
o te
stin
g to
dat
e in
fore
st m
anag
emen
t co
nte
xt
Wat
floo
dC
an b
e lin
ked
to G
reen
Ken
ue
GU
I; c
an h
and
le w
etla
nds
an
d la
kes
Com
plex
mod
el; s
impl
ified
sn
owm
elt;
not
su
itab
le fo
r co
mpl
ex te
rrai
n
Gro
un
dwat
er-
surf
ace
wat
er
mod
els
InH
M (
Hyd
roG
eoSp
her
e)G
rou
ndw
ater
-su
rfac
e w
ater
inte
ract
ion
focu
s an
d ro
ads
No
expl
icit
rep
rese
nta
tion
of
clim
atic
forc
ings
or
fore
st c
over
Mik
e-SH
ESi
mu
late
s co
mpl
ete
lan
d ph
ase
of h
ydro
logi
c cy
cle
Sim
plifi
ed r
epre
sen
tati
on o
f fo
rest
cov
er; h
igh
pu
rch
ase
cost
MO
DH
MS
Sim
ula
tes
com
plet
e la
nd
phas
e of
hyd
rolo
gic
cycl
eN
o sn
ow r
epre
sen
tati
on; s
impl
ified
fore
st c
over
; pro
prie
tary
mod
el
Soil
eros
ion
WE
PP
Sim
ula
tion
of
soil
eros
ion
, in
clu
din
g ro
ads
Em
piri
cal a
ppro
ach
es w
ith
res
pec
ts to
wat
ersh
ed h
ydro
logy
pro
cess
es
Riv
er b
asin
m
odel
sSS
AR
RM
odel
dev
elop
ed fo
r ri
ver
fore
cast
ing
and
wat
er m
anag
emen
tE
mp
eric
al; n
ot r
ecom
men
ded
for
fore
st m
anag
emen
t ap
plic
atio
ns
VIC
Mac
rosc
ale
mod
el; l
arge
-sca
le e
ffec
ts (
e.g.
mou
nta
in p
ine
beet
le)
Not
su
itab
le fo
r ty
pica
l for
est
man
agem
ent
appl
icat
ion
s du
e to
sca
le
WR
MM
Op
erat
ion
al w
ater
use
pla
nn
ing
mod
elD
oes
not
sim
ula
te w
ater
shed
pro
cess
es;
un
suit
able
for
fore
st m
anag
emen
t
WU
AM
Op
erat
ion
al w
ater
use
pla
nn
ing
mod
elD
oes
not
sim
ula
te w
ater
shed
pro
cess
es;
un
suit
able
for
fore
st m
anag
emen
t
ta
bl
e 9
Mod
el a
dvan
tage
s an
d di
sadv
anta
ges.
Cat
egor
yP
urp
ose
Mod
el n
ame
Pla
nn
ing
scal
e1Te
mpo
ral s
cale
Mod
el o
utp
uts
2
FHA
YP
FLF
SWE
TW
BSM
IFW
TO
FSF
MF
GF
RO
RF
SEM
WN
FW
Q
Fore
st h
ydro
logy
m
odel
sA
nn
ual
yi
eld
WR
EN
SS
(Win
Wrn
sHyd
/ E
CA
-AB
)
Wat
ersh
edA
nn
ual
X3
XX
Peak
flow
UB
C-U
F Pe
ak F
low
Mod
elFl
exib
leD
aily
33
XX
XX
XX
XX
Wat
er
bala
nce
BR
OO
K90
Wat
ersh
edD
aily
44
44
XX
XX
XX
XX
XX
X
ForH
yMW
ater
shed
Mon
thly
44
44
XX
XX
XX
X
ForW
aDy
Wat
ersh
edD
aily
44
44
XX
XX
XX
X
Wat
ersh
ed
hydr
olog
yD
HSV
MFl
exib
leH
ourl
yX
XX
XX
XX
XX
XX
XX
XX
X
RH
ESS
ysPa
tch
es (
HR
Us)
Dai
lyX
XX
XX
XX
XX
XX
XX
XX
X
Oth
er m
odel
sW
ater
bal
ance
HE
LPSo
il co
lum
nD
aily
XX
XX
X
Wat
ersh
ed
hydr
olog
yA
CR
USu
b-ba
sin
Dai
lyX
XX
X5
XX
XX
XX
XX
XX
X
CR
HM
HR
UH
ourl
y6
66
6X
XX
XX
XX
XX
HB
V-E
CG
RU
Dai
ly6
66
6X
XX
XX
XX
XX
HE
C-H
MS
Sub-
basi
nSu
b-da
ilyX
XX
XX
XX
XX
XX
XX
HSP
FW
ater
shed
Sub-
daily
XX
XX
XX
XX
XX
XX
XX
X
PR
EV
AH
GR
UH
ourl
yX
XX
XX
XX
XX
XX
XX
PR
MS/
MM
SG
RU
Dai
ly6
66
6X
XX
XX
XX
XX
XX
SWA
TG
RU
Dai
lyX
XX
XX
XX
XX
XX
XX
XX
XX
UB
CW
ME
leva
tion
ban
dD
aily
XX
XX
XX
XX
XX
XX
X
WaS
iM-E
TH
Flex
ible
Sub-
daily
XX
XX
XX
XX
XX
XX
XX
X
Wat
floo
dG
RU
Hou
rly
XX
XX
XX
XX
XX
XX
X
Gro
un
dwat
er-
surf
ace
wat
er
mod
els
InH
M
(Hyd
roG
eoSp
here
)Fl
exib
leFl
exib
leX
XX
XX
XX
XX
XX
XX
XX
X
Mik
e-SH
EFl
exib
leFl
exib
leX
XX
XX
XX
XX
XX
XX
XX
MO
DH
MS
Flex
ible
Flex
ible
XX
XX
XX
XX
XX
XX
XX
Soil
eros
ion
WE
PP
Hill
slop
eD
aily
66
66
XX
XX
XX
XX
XX
Riv
er b
asin
m
odel
sSS
AR
RE
leva
tion
ban
dSu
b-da
ilyX
XX
XX
XX
XX
XX
XX
VIC
25 k
m2
Sub-
daily
XX
XX
XX
XX
XX
XX
1 Sm
alle
st p
ossi
ble
plan
nin
g sc
ale
liste
d. F
lexi
ble
indi
cate
s th
at s
ilvic
ult
ura
l pra
ctic
es c
an b
e si
mu
late
d u
nde
r a
ran
ge o
f co
nfi
gura
tion
s (d
istr
ibu
ted
mod
els)
. 2
Abb
revi
atio
ns:
FY
= fu
ll hy
drog
raph
, AY
= a
nn
ual
yie
ld, P
F =
pea
k fl
ow, L
F =
low
flow
, SW
= s
now
wat
er e
quiv
alen
t (s
now
cov
er),
ET
= e
vapo
tran
spir
atio
n, W
B =
wat
er b
alan
ce (
for
soil
colu
mn
an
d/
or b
asin
), S
M =
soil
moi
stu
re, I
F =
infi
ltra
tion
, WT
= w
ater
tabl
e, O
F =
ove
rlan
d fl
ow, S
F =
shal
low
subs
urf
ace
flow
, MF
= m
acro
pore
(pr
efer
enti
al)
flow
, GF
= g
rou
ndw
ater
flow
(ba
sefl
ow),
RO
= b
asin
tota
l ru
nof
f, R
F =
roa
d fl
ow, S
E =
sed
imen
t So
il E
rosi
on, M
W =
mas
s w
asti
ng,
NF
= n
utr
ien
t fl
uxe
s,W
Q =
wat
er q
ual
ity.
3
Cla
imed
cap
abili
ty b
ut
un
test
ed.
4 St
ream
flow
sim
ula
tion
cap
abili
ty fo
r fi
rst
orde
r w
ater
shed
s on
ly (
sum
of
all r
un
off
com
pon
ents
); m
odel
s do
not
inco
rpor
ate
chan
nel
rou
tin
g.5
Snow
rou
tin
es h
ave
been
add
ed b
y D
r. St
efan
Kie
nzl
e (U
niv
ersi
ty o
f Le
thbr
idge
) bu
t ar
e n
ot p
art
of t
he
orig
inal
sof
twar
e di
stri
bute
d by
Sou
th A
fric
an d
evel
oper
s.
6 N
o ex
plic
it c
han
nel
rou
tin
g; s
impl
ified
pro
cedu
res
are
use
d. S
ee n
arra
tive
des
crip
tion
of
each
mod
el (
App
endi
x 1)
.
24
these nine models encompass the forest watershed-modelling capabilities identified in Tables 6 to 8 and provide the highest functionality in the respective complexity categories (Figure 1). Thus, it should be possible, in principle (i.e., data, time, and resource limitations notwithstanding) to address forest watershed-management questions in AB or BC by selecting one of these nine models. The remaining 21 models reviewed may offer similar features and capabilities, but are characterized by lower functionality for answering questions about the hydrologic aspects of forest management (Figure 1). Thus, these 21 models should generally not be selected for use in a forest watershed-management context. This obser-vation applies to water quantity only; water quality simulation was not considered in this review.
Low-complexity Models3.2.1
In the low-complexity category, only one option for model studies is currently available—WRENSS and its companion programs, WinWrnsHyd and ECA-AB (Figure 1). Overall, the WRENSS procedure has proven useful for evaluating existing harvests and for planning future harvests. The model allows for quick evaluation of changes in average annual streamflows (yield) under different forest manage-ment regimes, and includes forest regrowth. Advantages of WinWrnsHyd and ECA-AB are low overall data requirements and ease of use (Table 9). The main drawback is that overall model functionality is limited, with model output restricted to annual yield, watershed water balances, and evapotranspiration, although WinWrnsHyd may have some untested use for assessing peak flows (Table 8). The models do not simulate absolute streamflows, but rather look at the relative changes in streamflows due to harvest regimes (Section A1.25). WinWrnsHyd and ECA-AB are commonly used in AB (Dr. A. Anderson, ASRD, pers. comm., Jan. 2009).
Medium-complexity Models3.2.2
Overall, forest watershed modelling capabilities are encompassed by UBCWM, BROOK90, ForWaDy, and the UBC-UF Peak Flow model, with the remaining medium-complexity models having overlap-ping features but offering lower functionality for answering forest management questions (Figure 1). The suitability of these four models for answering forest management questions can be summarized as follows:
• UBCWM(SectionA1.19)shouldbethepreferredmodelforuseinmountainousterrain,andin settings where glacial melt or upland lakes are important. However, its simplified forest cover representation, and the use of elevation bands limit flexibility in simulating forest management scenarios. Furthermore, with the retirement of the developer (Dr. M. Quick, UBC), the model may no longer be actively maintained.
• BROOK90(SectionA1.2)shouldbethepreferredmodelforuseingraduallyslopedterrainun-less rain-on-snow processes are important or forest growth needs to be considered; in these cases. ForWaDy (Section A1.5) may provide a viable alternative. The main limitation of both models is the lack of a channel-routing routine, which limits model applicability to the stand/site level or to small watersheds with no sub-basins. Peer-reviewed publications that test the ForWaDy model against field data are currently lacking.
• TheUBC-UFPeakFlowmodel(SectionA1.18)canbeusedtoassesstheeffectsofroadsinasim-plified fashion, is fully distributed, and can be applied to large watersheds, which are all capabilities not contained in any of the other models in this category. However, functionality of the model is mainly limited to peak flows and it is still under development, with only limited testing at this point.
25
Overall, the medium-complexity model category has limited capability with respect to answering forest management questions under the possible settings encountered in AB and BC (Tables 6 to 8). With the exception of the UBC-UF Peak Flow model, all of the models in this category are lumped or semi-distributed, restricting the ability to evaluate the effects of complex forest management plans or the ability to handle the intricacies of snowmelt processes in complex terrain (i.e., representation of slope-aspect-vegetation-elevation co-variability is coarse; representing this co-variability, and its changes following harvesting, may be critical for certain watersheds). None of the models can provide output at a sub-daily time-step level, which may be important for simulating peak flows. There is inadequate capability for addressing the effects of road construction and management, with the UBC-UF Peak Flow model offering a simplified and untested approach. None of the models can be applied to medium-sized watersheds in gradual terrain, and none of the models has the ability to represent multi-layered forest vegetation (Tables 4 and 7). These numerous limitations may be important in certain operational set-tings and can only be overcome by applying suitable high complexity models.
A development worth noting is that a rewrite of the HBV-EC model is currently underway and is intended for application to forest management scenarios (Dr. R. D Moore, pers. comm., Jan. 2009). Experience with the current version of the HBV-EC model has shown that functionality in a forest management context is limited by an “overly simplistic representation of canopy influences on snow deposition” (Moore et al. 2007). The new model will have a target conceptualization between the current HBV-EC model and DHSVM. This development is promising and should be closely monitored. A fea-ture of the HBV-EC that sets it apart from other models is the availability of a Graphical User Interface (GUI) in the form of Green-Kenue (Section A1.6).
High-complexity Models3.2.3
Overall, forest watershed modelling capabilities are encompassed by DHSVM, RHESSys, WaSiM-ETH, and CRHM, with the remaining high-complexity models having overlapping features, but offering lower functionality for answering forest management questions (Figure 1). The suitability of these four models for answering forest management questions can be summarized as follows:
• DHSVM(SectionA1.4)shouldbethepreferredmodelforuseinmountainousterrain.However,only limited efforts have been paid to making the model user friendly. Additionally, the model has limited functionality in gently sloped terrain and in settings with a substantial glacial melt or groundwater component to the hydrological budget.
• RHESSys(SectionA1.15)hascapabilitiesnotofferedbyDHSVMineco-hydrologicalareassuchas forest growth (Table 6) and a rudimentary groundwater component has also been added to the model (Table 7). However, with its daily time step, RHESSys is not as suitable as DHSVM for simulating (instantaneous) peak flows or in situations where short-duration, high-intensity precipitation events (i.e., rainfall-dominated settings) or diurnal fluctuations in meteorological conditions are important.
• WaSiM-ETH(SectionA1.21)offersanumberofadvantagesoverbothDHSVMandRHESSys, including the possibility of rigorously treating groundwater processes, a glacier model (accu-mulation of snow, melt of snow, ice, and firn), a lake model component and a channel-routing component that accounts for artificial and natural reservoirs and lakes. These capabilities should allow WaSiM-ETH to simulate virtually any watershed hydrologic regime encountered in AB and BC (Table 7). The model’s main drawback is that its forest hydrology-specific components (e.g., forest canopy interactions with precipitation) appear not to have been tested. The WaSiM-ETH developer (Dr. J. Schulla) began working for the Pacific Climate Impacts Consortium (PCIC) in July 2009, bringing expertise with the model to Western Canada.
26
• TheCRHM(SectionA1.3)wasspecificallydevelopedforprairie,tundra,andborealforestsettings,with corresponding consideration for watershed processes (e.g., blowing snow and frozen soils). As such, within a forest management context, CRHM may be applicable to boreal and high-elevation forest settings in AB and BC. The main limitation appears to the rudimentary streamflow-routing routine, which constrains model applicability to small to medium watersheds (Tables 7 and 9).
Jointly, when compared to medium-complexity models (Tables 6 to 8), the above models offer a considerably broader range of capabilities for answering forest management questions in AB and BC. As such, the four high-complexity models address the limitations of medium-complexity models, provided that the higher demands on data (in particular meteorological variables such as temperature, precipitation, shortwave and longwave radiation, relative humidity, and wind speed), time (budget), and resources (GIS, model calibration) can be overcome.
Other models with lower functionality (Figure 1) that are worth noting include ACRU (Section A1.1), which has recently been applied in AB (Dr S. Kienzle, University of Lethbridge, pers. comm., Jan. 2009); MIKE-SHE (Section A1.11), an integrated groundwater surface model with a well-developed GUI and user-support; and InHM (Section A1.10), which has been used to simulate the effects of roads on watershed hydrology. SWAT (Section A1.17) is being adapted to the boreal forest environment for the purpose of water quality simulations by researchers at Lakehead University (Ontario), as part of the Forest Watershed and Riparian Disturbance (FORWARD) project. The model review did not specifically consider water quality aspects.
bArriers to the oPerAtionAl use of models4
One of the important findings of this model review is that low- and medium-complexity models of-fer limited functionality for answering forest management questions (Section 3.2). As such, it would appear that many of the more intricate forest management questions can only be addressed with high-complexity models that require considerable data, time, and resources to set up and operate. Developing models that better recognize and minimize trade-offs between functionality and complexity is needed, or perhaps more importantly, developing flexible model approaches that can be tailored to site-specific situations (Savenije 2009). Until such models are developed and tested, it may be necessary to create an environment in which high-complexity models can be routinely, reliably, and consistently applied. The sections below address some of the barriers that will need to be overcome to create such an environment.
user knowledge and education 4.1
To confidently apply most of the models reviewed, an intermediate- to senior-level professional with an advanced academic degree (MSc or PhD) emphasizing quantitative hydrology is likely needed. Current-ly, only limited education in watershed hydrology and modelling is available at university-level forestry schools and, as such, there are few forestry professionals and practitioners who are trained to properly apply hydrologic models or who have the necessary expertise to adequately interpret the model out-put, including associated assumptions and limitations. Conversely, many hydrologic and groundwater modellers have backgrounds as engineers, geographers, or earth scientists, which often does not provide a strong understanding of forest hydrology and management. Therefore, interdisciplinary training in all of the above subjects is needed.
27
model inter-comparison4.2
A wide variety of models have been developed to describe and predict an array of watershed hydro-logic characteristics. Some models have the capability to address specific forest management-related hydrologic conditions in great detail, while others lack the necessary hydrologic process representation to accurately simulate forest hydrologic regimes. Cross-comparison of models in relatively data-rich environments (experimental watersheds) will help better identify model functionality in a forest man-agement context by calibrating and testing each model against detailed stand-level and watershed-scale data. The general usefulness of model intercomparisons has, for example, been shown in studies of macroscale models (Lohmann et al. 1998) and for streamflow forecasting (e.g., Smith et al. 2004; Reed et al. 2004).
While a detailed discussion about the types of tests that would be required to assess the ability of a model to make accurate predictions about hydrological changes associated with forest management is beyond the scope of this review, Klemes (1986) argued that model tests should be designed to simulate the ways in which a model would run operationally. As such, at the basin scale, data from paired-water-shed studies provide a context for assessing not only how well a model can simulate streamflow, but also how well it can reproduce the magnitude of streamflow change associated with harvesting (Waichler et al. 2005). Furthermore, in snow-dominated watersheds, a prerequisite to accurately simulating the hydrologic effects of forest management is the ability to represent the gradients of snow accumulation and melt with elevation, slope/aspect, and forest cover, and to represent how all these elements interact. This requirement suggests that models should be tested against snow survey data that is designed to rep-resent the full range of combinations of these factors. Data from paired watershed experiments are better suited to conduct such detailed model tests (e.g., Upper Penticton Creek: Thyer et al. 2004; Carnation Creek: Beckers and Alila 2004; Cotton Creek: Jost et al. 2007).
Only through careful testing with high-quality experimental data that describes multiple concur-rent internal watershed processes under a range of hydrologic regimes can one determine if a model is a reliable tool for simulating watershed hydrology (and more importantly, the changes due to land use or climate change) or whether improvements to the model structure are needed (e.g., Burges 2003). Once model performance has been compared at experimental watersheds and suitable model param-eterizations have been determined, additional model testing could be undertaken in relatively data-poor watersheds (e.g., locations with only streamflow data available). This would provide insight into the transferability of adopted model parameters and into the potential decrease in accuracy that might be expected when applying models in areas with little data. For example, model testing at experimental watersheds would involve the use of comprehensive weather data (including solar radiation, humidity, and wind speed) measured at one or more locations. However, using a model in a data-poor watershed would involve weather data extrapolated from nearby climate stations (daily air temperature and pre-cipitation only). An important question in this context is whether it is possible to estimate variables such as solar radiation, and thus to run an energy balance melt model, or whether the propagation of errors in this approach causes model performance to deteriorate to levels below that of simpler models (e.g., Walter et al. 2005).
Experience and expertise obtained through testing is required to successfully use hydrologic mod-els to answer forest management questions in watersheds with low data resources (Alila and Beckers 2001). Such experience would improve the consistent use of models at sites with lesser data, including ungauged basins. The resulting knowledge of appropriate model parameterization and transferability of these parameters might be incorporated into databases that industry and consultants could readily access. Compiling and maintaining these databases may require an effort at the federal- and/or provin-cial-government level.
28
data Availability4.3
The lack of readily available DEM, forest cover, soils, and climate data hinders the widespread opera-tional use of models. Hamilton (2007) discusses, in some length, the lack of data to operate models and the problems this has created with respect to making decisions from model outputs. Improvements are being made, notably in compiling climate data for BC (Spittlehouse 2006) and for water portals in AB2, and in other efforts to assemble existing databases on forest cover and disturbances (UBC-UF Peak Flow Model; Section A1.18). Nonetheless, additional data is needed to support the use of physic-ally based models (e.g., LAI, soil cover, and soil depth) and data at high resolution is needed for use in small watersheds. Compiling and maintaining this data will likely require an effort at the federal- and/or provincial-government level.
Historical databases of temperature and precipitation could be supplemented with algorithms that generate additional meteorological variables, such as shortwave and longwave radiation and relative hu-midity, which are required to run physically based models. Such algorithms already exist either internal to hydrologic models (e.g., MTN-Clim model in RHESSys; Section A1.15) or as external procedures (e.g., Waichler and Wigmosta 2003; Schnorbus and Alila 2004b), and have been tested and applied in BC (Schnorbus and Alila 2004a). These algorithms could be made user friendly and readily available to industry and consultants for use in an operational context. Testing of meteorological-forcing algorithms to identify any performance issues in replacing actual climate conditions could take place as part of the above-mentioned model-intercomparison experiments.
communicating model uncertainty4.4
When using models to guide forest management decisions, there is an inherent danger that managers or clients may view the model outputs as absolute. To avoid this potential misperception, it is critical for study proponents to provide an estimate of the uncertainty in model outputs and to communicate this uncertainty to end-users of the model results (Ivanovic and Freer 2009). Better, more-informed decisions can be made when model uncertainties and limitations are known. Several methods exist for quantifying model uncertainty, such as Monte Carlo simulations that take into account confidence bounds on key model parameters (see Section 2.4; Model Calibration and Prediction Confidence). However, the use of such techniques in an operational context is not widespread. To maximize the value of their results, model users should therefore make greater use of available methods to calculate uncertainty.
need for better models, guis, and model support4.5
Few hydrology models exist that have been developed specifically with forestry applications in mind. Most of these forest hydrology models (e.g., Table 9) have been developed in an academic environment. Consequently, limited efforts have been made to make the models user-friendly, and technical support is typically lacking unless special arrangement has been made. The development of commercial software and the associated increased availability of model support is therefore needed.
Trade-offs exist between model complexity and forest management functionality (Figure 1). The difficulties of obtaining enough data for model parameterization and calibration limit the operational application of fully distributed, physically based models that would normally be most suitable (i.e., based on process representation) for addressing forest management questions. At the same time, rela-tively easy-to-use models with low data requirements may include assumptions regarding hydrologic processes that limit their ability to provide satisfactory results. There is no “best” model (Savenije 2009),
2 www.albertawater.com/
29
i.e., an easy-to-use model with low data requirements that is very accurate and that can be applied under all circumstances does not exist. However, “best practices” can be achieved by developing new flexible modelling approaches that a user can tailor on a site-specific basis to better recognize and minimize trade-offs between functionality and complexity.
Only a few of the models reviewed are linked to GUIs to facilitate model setup. This capability needs to be expanded to more models. What is also needed is for GUIs to provide access to common databases regarding climate, forest cover soils, and topography (Section 4.3) and databases that could guide model parameterization for watersheds with limited data (Section 4.2). The use of common databases will lead to a more unified approach to hydrologic modelling and will allow users to compare results from differ-ent watersheds. With such developments, GUIs might help improve the consistent use of models.
Policy and Professional Precedence4.6
Forest hydrology modelling is still very much in the realm of academic institutions and there is a lack of policy and professional precedence (i.e., a case history of operational watershed hydrologic modelling studies). No information exists that provides direction to forest managers and other resource practition-ers about which models are, or are not, acceptable for use in a forest management context. This review provides information that, to some extent, alleviates this knowledge gap; model intercomparisons at ex-perimental watersheds would provide further insight. This information could be used to provide either prescriptive (policy) or soft guidance for forest watershed-management modelling studies. This guid-ance could spell out expectations regarding model selection, application, and calibration, and associated documentation requirements (reporting). Providing this guidance for modelling and other watershed management studies will likely require involvement at the federal- and/or provincial-government level.
using models in A climAte chAnge context5
With climate change comes many adjustments in watershed hydrology, which in turn may affect aspects of long-term forest management planning. From an operational perspective, hydrologic models could be used to address a range of questions, including, but not limited to, assessing future growing conditions, the survival of replanted trees in a changing climate, the permanence of current watershed features such as upland lakes and wetlands, and the potential changes in flood risks and other disturbance factors. Using hydrologic models to address such complex questions in a practical manner is expected to pose a number of challenges. The purpose of this section is to highlight these challenges and to identify which of the models reviewed may be best suitable for answering climate change questions.
hydrologic implications of climate change5.1
Several detailed studies of past trends and future predictions regarding climate and hydrology in AB and BC have been conducted to date (e.g., BC Ministry of Environment, Land and Parks 2002; Rodenhuis et al. 2007; Pike et al. 2008a, 2008b; Sauchyn and Kulshreshtha 2008; Walker and Sydneysmith 2008). Pike et al. (2008b) discussed eight high-level hydrologic implications of climate change. These implica-tions, as summarized below, reflect questions that could potentially be answered by applying appropriate hydrologic models.
Associated with each of the eight broad hydrologic implications of climate change, are specific pro-cesses (e.g., evapotranspiration), watershed outputs (e.g., timing and magnitude of peak and low flows and stream temperature), and other factors (e.g., growing conditions for trees and wildfire risk) that could be affected by anticipated shifts in meteorological and hydrological conditions. These processes
30
(and associated model inputs), outputs, and other factors must be present in watershed models to an-swer climate change questions in a forest management context. The sections below build on the review and ranking of model functionality, and therefore put specific emphasis on those processes, outputs, and factors whose interactions with forestry activities may be exacerbated by climate change.
Increased Atmospheric Evaporative Demand
Increases in atmospheric evaporative demand are predicted with climate change and this may signifi-cantly affect water resources through greater evaporative losses from streams, lakes, and reservoirs, and changing water demands (Pike et al. 2008b). Evaporative demand is a function of air temperature, solar radiation, humidity, and wind speed (Moore et al. 2008). As such, incorporating these weather variables into calculations of reference evapotranspiration (e.g., through the Penman-Monteith equation) is criti-cal for assessing the potential consequences of increased evaporative demand due to climate change (Table 10). Thus, models with physically based evapotranspiration routines are inherently better suited to estimate evapotranspiration under changing climates than are other models.
Altered Vegetation Composition Affecting Evaporation and Interception
Changes in atmospheric evaporative demand are sufficient to reduce water availability (soil moisture), which in turn may affect forest productivity (growth), species survival (mortality), and changes in age-class distribution and in the form of vegetation (e.g., Pike et al. 2008b; Gayton 2008). These issues must be considered carefully in hydrologic models that are used for planning purposes (Table 10), as they affect financial aspects of forest management, and may also influence hydrologic recovery and decisions regarding tree species selection following harvesting. Furthermore, when conducting long-term model simulations and harvest planning in the context of climate change, it may become important whether or not the model input can be easily adapted to represent the gradual or abrupt changes in vegetation composition that might occur during the time period being considered (simulated). This ability of a model to allow for time-varying vegetation properties within the course of a single simulation (i.e., the ability of a user to change these properties without having to re-start the model) is referred to as tempo-ral input control (Table 10).
The amount of plant material and physiological characteristics of the altered vegetation may have an important effect on a site’s water balance (Pike et al. 2008b). The interaction between plants and the atmosphere (i.e., evapotranspiration and precipitation interception) is substantially determined by vegetation surface area (Monteith and Unsworth 1990), typically described as leaf area index (LAI). LAI is also a primary reference parameter for plant growth. As such, with climate change, LAI is a critical model variable to describe forest characteristics and their potential episodic or long-term changes.
Stomatal resistance (or its inverse, stomatal conductance) is another crucial parameter (Table 10) as it is used to calculate the vegetation transpiration rate from humidity (vapour pressure) gradients (Monte-ith and Unsworth 1990). Stomatal resistances vary between vegetation species and are therefore an important physiological parameter to assess the effect that vegetation may have on a site’s water balance. Furthermore, the ability of models to simulate the closing of stomata (i.e., an increase in stomatal resist-ance) when atmospheric water demand exceeds water availability is the primary mechanism to assess plant response to this stress condition.
Decreased Snow Accumulation and Accelerated Melt
Increased temperatures as a result of climate change will lead to a decrease in snow accumulation, earlier melt, and less water storage for spring freshet and/or to release to groundwater storage (Whit-field et al. 2002; Merritt et al. 2006; Rodenhuis et al. 2007; Sauchyn and Kulshreshtha 2008; Walker and
31
Sydneysmith 2008). Watersheds that would be the most sensitive to temperature shifts are those in mixed regimes and that occupy the interface between rainfall and snow deposition in the winter (Pike et al. 2008b). For the purpose of long-term simulations, models may have to represent initially predomin-antly nival conditions, which may become hybrid (mixed), or perhaps even pluvial dominated over time. Another consideration is that alternating precipitation in the form of rain or snow in the early and late parts of winter (late fall and early spring) may become increasingly important for streamflow generation as temperatures continue to increase. These trends have, for example, been observed at Redfish Creek in southeastern BC (R. Pike, pers. comm., Jan. 2009). As such, the ability of hydrologic models to accurately represent mixed regimes (rain-on-snow energy transfer) may be a crucial factor in a climate change con-text. Futhermore, models with physically based or analytical (temperature-radiation) snowmelt routines are inherently better suited to predict the potential for accelerated melt under changing climates than are empirical models. (Table 10).
Criterion Evaluation measures
Atmospheric evaporative demand Solar radiation, humidity, and wind speed
Altered vegetation composition affecting evaporation and interception
Leaf Area Index
Stomatal resistance
Forest growth (productivity)
Forest survival (mortality)
Temporal input control
Snow accumulation and melt Physical/analytical snowmelt equations
Rain-on-snow simulation
Permafrost, river, and lake ice Frozen soil influence on water movement
River and lake ice model component
Glacier mass balance adjustments Glacier melt model component
Altered streamflow Groundwater
Lakes
Wetlands
Water consumption (water supply systems)
Stream and lake temperatures Water temperature model component
Increased frequency/magnitude of disturbances
Channel routing (floods)
Multiple vegetation layers (wildfires, pests)
Vegetation albedo (wildfires, pests)
Soil albedo (fires)
Hydrophobicity (wildfires)
Landslide simulation
table 10 Climate change model evaluation criteria.
Accelerated Melting of Permafrost, Lake Ice, and River Ice
Rising temperatures will affect ice-related hydrologic features. Projections of milder winter temperatures mean that river and lake ice could develop later and disappear earlier than normal. Permafrost can also be expected to respond to changes in temperature and precipitation (Pike et al. 2008b). These hydrologic changes will have implications on forest harvest scheduling (operable ground, seasonal water tables, and timing) and transportation (e.g., ice bridges). Permafrost thaw may also lead to altered soil nutrient cycling and changes in vegetation distribution (Jorgenson et al. 2001). Depending on model application,
32
the ability to simulate some of the above processes may be an important consideration when selecting a hydrologic model (Table 10).
River and lake ice formation and break-up processes are the focus of specialized kinematic models (e.g., Beltaos 2007) and are not typically incorporated into hydrologic models. Soil temperatures are more widely accounted for in hydrologic models, typically to calculate the ground heat-flux component of the snowpack energy balance (e.g., Wigmosta et al. 1994) and in some cases to assess frozen soil con-ditions and associated effects on water movement (e.g., Pomeroy et al. 2007). The information required to simulate soil temperatures and frozen soil conditions includes meteorological data and soil thermal properties.
Glacier Mass Balance Adjustments
Recent studies have shown that glaciers throughout BC and contiguous parts of Alaska are dominantly losing mass (Moore et al. 2009 and references therein). It is expected that most glaciers in British Col-umbia will continue to recede, except those at the coldest locations (Rodenhuis et al. 2007). Negative annual glacier mass balances should result in increased summer streamflows for some years or even decades as glacial melt accelerates. This effect will be followed by a larger decrease when the glaciers eventually disappear or decrease in size to some small proportion of the watershed area, which may already be occurring in some areas (Stahl et al. 2008). In the long term, the reduction or elimination of the glacial melt water component in the summer/early fall would increase the frequency and duration of low-flow days in these systems, which may affect aquatic habitat. For AB, a particular concern is more frequent reductions in water availability on the eastern slopes of the Rocky Mountains (Byrne et al. 1989; Demuth and Pietroniro 2002).Thus, for some watersheds in BC and AB, the ability to simulate such changes in glacial melt contributions may be an important consideration when selecting a hydro-logic model (Table 10).
Altered Surface Water Flows
Depending on the region and the season, all or some of the above-mentioned hydrologic implications of climate change may interact to alter the timing and magnitude of streamflows (peak and low flows). Most models will calculate associated changes in overland flow, infiltration, soil moisture conditions, and shallow subsurface runoff, and the subsequent discharge of water to the channel environment without the need for modifying how the model works. Nonetheless, in certain settings, specific questions regarding watershed management in a climate change context may revolve around the permanence of lakes or wetlands. In other settings, the changes in groundwater recharge rates associated with climate change (e.g., Scibec and Allen 2006a, 2006b) may have consequences for baseflow contributions to main-taining low flows. The capability to account for the anticipated increased competition between human water use and water availability for ecological receptors may be another important feature in selecting a model (Table 10).
Increased Stream and Lake Temperatures
Stream and lake temperatures are projected to increase due to climate change, which can result in a number of specific concerns for water and fish species. The effects of increased water temperatures are likely to be compounded when hydrologic regime changes result in reduced seasonal flows (Pike et al. 2008b) and may also be exacerbated by forest harvesting. As such, stream temperature may be a crucial watershed output when considering forest management in a climate change context (Table 10).
Models to predict stream temperatures fall into two general classes (Sridhar et al. 2004): (1) empirical relationships based on observations of stream temperature and stream properties (such as discharge,
33
channel geometry, and streamside vegetation characteristics) and (2) models that represent the energy balance of the stream. Using physically based models to predict stream temperature has become feas-ible through interfacing with GIS methods. Furthermore, the data requirements for predicting stream temperatures simulation are such that clear synergies exist with physically based watershed hydrologic modelling.
Increased Frequency/Magnitude of Disturbances
Storm frequency and intensity are projected to increase (Rodenhuis et al. 2007), likely raising the fre-quency of flooding (Table 10). Watershed modelling can be used to assess the suitability of current infrastructure (e.g., stream crossings) under a future climate and/or to determine the need to adjust en-gineering design criteria accordingly. In some rain-dominated settings, the ability of hydrologic models to answer such questions may critically depend on accurately simulating preferential runoff mechanisms as triggered by storm intensity (e.g., Carnation Creek on Vancouver Island; Beckers and Alila 2004). In snow or mixed regimes, accurately simulating melt rates is important for predicting peak flows (e.g., Redfish Creek in southeast BC; Schnorbus and Alila 2004a) with rain or rain-on-snow events likely be-coming increasingly important for flood generation.
Other disturbances that are projected to increase include wildfires, pests (jnsects), the frequency of windthrow, breakage of trees, and landslides (Pike et al. 2008b). Of these disturbances, modelling of landslides (specifically, shallow slides and debris flows) is the only factor that provides clear synergies with watershed simulation (Table 10). Landslide modelling has been the focus of specialized physically based slope-stability models such as dSLAM (Wu and Sidle 1995) and IDSSM (Dhakal and Sidle 2003) and has also recently been incorporated in DHSVM (Doten et al. 2006).
Windthrow and its associated consequences are the focus of specialized risk models (e.g., Lanquaye and Mitchell 2005), thereby offering little synergy with watershed hydrologic modelling. This lack of synergy with watershed modelling also holds true for predicting the occurrence of pests. However, it may still be important for a hydrologic model to incorporate (as input) the changes that may occur as a result of these disturbances, and to assess potential consequences for hydrologic processes, watershed outputs, and associated resources. An important aspect related to the occurrence of pests (tree mortality) is the change in the canopy albedo of the dead trees (Table 10), which in turn affects the radiation energy bal-ance of these stands, causing increases in snowmelt rates. Such changes due to mountain pine beetle and associated effects on streamflows were investigated by the Forest Practices Board (2007).
Forest fires also cause vegetation changes, which, depending on the nature of the fire, may include re-moval of the understorey without canopy disruption or full consumption of the overstorey resulting in standing dead timber. These complex changes can only be represented in a straightforward fashion with models that allow for multiple vegetation layers (Table 10). Fires can also cause changes in soil properties that affect the hydrologic response of a watershed, including a change in soil albedo and, under certain conditions, formation of a hydrophobic layer (Agee 1993) that limits the infiltration capacity of the soil (Table 10; altered soil properties). Soil hydrophobicity is expected to decline over time, but the process is poorly understood (DeBano 2000).
model review 5.2
This section addresses which of the models can currently be used for climate change scenarios and what processes or outputs they can provide, including the relative level of confidence in the results. The review of model functionality for use in a climate change context is limited to the nine models short-listed as being most suitable for addressing forest management questions (Section 3). The low-complexity category includes WRENSS, and the medium-complexity category includes BROOK90, ForWaDy, the
34
Mod
el c
ompl
exit
yM
odel
nam
e
Radiation, humidity, wind speed
Leaf area index
Stomatal resistance
Forest growth (productivtity)
Forest survival (mortality)
Temporal input control
Physical/analytical snowmelt
Mixed rain/snow processes
Frozen soil/permafrost
River and lake ice
Glacier melt
Stream temperature
Groundwater
Lakes
Wetlands
Water consumption
Channel routing (floods)
Multiple vegetation layers
Vegetation albedo
Soil albedo
Hydrophobicity (wildfires)
Landslide simulation
Low
WR
EN
SSX
X
Med
ium
BR
OO
K90
XX
XX
X
ForW
aDy
XX
XX
XX
UB
C-U
F Pe
ak F
low
Mod
elX
X
UB
CW
MX
XX
XX
XX
CR
HM
XX
XX
XX
XX
X
DH
SVM
XX
XX
XX
XX
XX
RH
ESS
ysX
XX
XX
XX
XX
XX
XX
WaS
iM-E
TH
XX
XX
XX
XX
XX
XX
X
Mac
rosc
ale
VIC
XX
XX
XX
XX
XX
XX
ta
bl
e 1
1 C
limat
e ch
ange
mod
el r
anki
ng.
35
UBC-UF Peak flow model, and UBCWM. The high-complexity category encompasses CRHM, DHSVM, RHESSys, and WaSiM-ETH (Table 11). As a macroscale model, separate consideration is given for VIC.
Below, the models are reviewed against the criteria identified in the previous section (Table 10). Figure 2 presents overall model climate change functionality based on the number of “scores” in Table 11, with forest management functionality on the vertical axis reproduced from Figure 1. Figure 2 indi-cates that those models that rank high for forest management functionality generally also rank high with respect to ability for answering climate change questions.
Low-complexity Models5.2.1
WRENSS
WRENSS offers little climate change functionality, meeting two criteria in Table 11. The main limita-tions of using WRENSS for climate change applications are the lack of climatic input variables to assess potential changes in atmospheric evaporative demand, the empirical representation of vegetation effects on evaporative processes (through water use modifier factors), and an inability to simulate mixed cli-matic regimes (the model is split into rain- and snow-dominated procedures; Section A1.25). Thus, the model is not recommended for climate change applications.
Medium-complexity Models5.2.2
BROOK90
BROOK90 (Section A1.2) offers low to intermediate climate change functionality (Figure 2), meeting five criteria in Table 11. The main limitations of using BROOK90 for climate change applications are an inability to simulate mixed regimes (the temperature-index method does not take rain-on-snow energy transfer into account), a lack of a multi-storey canopy representation that may be useful for simulat-ing the effects of disturbances such a wildfire, and an inability to account in a straightforward manner (through simulation or as temporal input control) for forest growth and mortality (Table 11). The model may have use in examining shifts in site water balances due to climate change in rain-dominated regimes (the evaporative equations are physically based) or in regions which are expected to remain snowmelt dominated under climate change.
ForWaDy
ForWaDy (Section A1.5) offers low to intermediate climate change functionality (Figure 2), meeting five criteria in Table 11. The main limitation of using ForWaDy for climate change applications is an inability to account directly (through simulation) for forest growth and mortality (Table 11). However, the capabilities of ForWaDy in this latter respect may be enhanced considerably by linking it with stand-level forest growth models such as FORCEE and FORECAST. Therefore, the main use of ForWaDy in a climate change context may be in assessing future site water balances and forest growth potential.
UBC-UF Peak Flow Model
The UBC-UF peak flow model (Section A1.18) offers little climate change functionality, meeting two of the criteria in Table 11. The model offers only a rudimentary representation of evaporative processes through interception coefficients and water-use modifiers, while snow processes are also simulated em-pirically. Thus, the model is not recommended for climate change applications.
36
UBCWM
The UBCWM (Section A1.19) offers intermediate climate change functionality (Figure 2), meeting six criteria in Table 11. The main limitations of using UBCWM for climate change applications are a lack of representation of the climatic variables needed to calculate changes in atmospheric evaporative demand, a lack of a multi-storey canopy representation, and an inability to account in a straightforward manner (through simulation or as temporal input control) for forest growth and mortality (Table 11). The model may have use in examining hydrologic implications of climate change in watersheds with a substantial glacial melt contribution and in settings where upland lakes or water supply considerations are important. However, overall confidence in modelled annual yield and particularly low flows would be reduced due to empirical representation of forest cover characteristics and evaporative processes.
High-complexity Models5.2.3
CRHM
CRHM (Section A1.3) offers intermediate to high climate change functionality (Figure 2) meeting eight criteria in Table 11. Its main limitations are an inability to account directly (through simulation) for forest growth and mortality (Table 11) and a lack of a multi-storey canopy representation. CRHM’s main use is in investigating the hydrologic implications of climate change on site water balances (snow accumulation and melt, evapotranspiration and precipitation interception processes) and streamflow variables in boreal forest environments, including possible use for water movement in (partially) frozen soils.
DHSVM
DHSVM (Section A1.4) offers intermediate to high climate change functionality (Figure 2), meeting nine criteria in Table 11. DHSVM functionality is limited to steep mountainous settings without sub-stantial glacial melt, groundwater, lake, or wetland contributions to the hydrologic budget. An inability to account in a straightforward manner (through simulation or as temporal input control) for forest growth and mortality limits model functionality for long-term simulations (Table 11). Taking these limitations into account, the model is expected to be a useful tool for analyzing the effects of climate change on site water balances (evaporation and precipitation interception processes, snow accumulation and melt) and streamflow variables, and for analyzing possible changes in disturbance events, including flood flows and mass wasting. Another advantage of DHSVM is that it can be readily linked to gridded output from climate models such as MM5.
RHESSys
RHESSys (Section A1.15) offers high climate change functionality (Figure 2), meeting 12 criteria in Table 11. The main limitation of RHESSys is it does not represent setting-specific hydrologic processes that may affect the response of a particular watershed to climate change. RHESSys is expected to be useful for assessing the effects of climate change on site water balances (evapotranspiration, snow ac-cumulation and melt processes) and streamflow variables, for simulating forest growth and mortality, and for situations in which temporal input control is required to represent gradual or abrupt changes in vegetation due to disturbance factors such as insect, fire, and harvesting. RHESSys is the only model with an ecology component (BIOME-BGC). This ecology model component could be used to investigate the effects of hydrological processes on the distribution, structure, and function of ecosystems; the effects of biotic processes on elements of the water cycle; and how these two-way interactions are affected by climate change.
37
WaSiM-ETH
WaSiM-ETH (Section A1.21) offers high climate change functionality (Figure 2), meeting 13 criteria in Table 11. The main limitation of WaSiM-ETH is it does not simulate forest growth and mortality or provide temporal input control to represent gradual or abrupt changes in vegetation. WaSiM-ETH may be useful for assessing the effects of climate change on site water balances (evapotranspiration, snow accumulation and melt processes) and streamflow variables. The model also offers advantages over most other models (except UBCWM) for simulating a variety of setting-specific hydrologic processes that may affect the response of a particular watershed to climate change. Another advantage of WaSiM-ETH is that it can be readily linked to gridded output from climate models.
Macroscale Models5.2.4
The VIC model (Section A1.20) has high climate change functionality, meeting 11 criteria in Table 11. The model is a useful tool for analyzing the effects of climate change on evaporation and precipitation interception processes, and on snow accumulation and melt for large-scale watersheds, and for assess-ing the cumulative effects of large-scale disturbances. Currently, the VIC model is being applied in an mountain pine beetle context for the Fraser River, and in a climate change context in the Columbia Riv-er, the Peace River, and the Campbell River by the Pacific Climate Impacts Consortium, with the support of the University of Washington and the MOE River Forecast Centre. As mentioned in the Section 5.2
WRENSS
UBC-UF Peak Flow Model
BROOK90
DHSVM
ForWaDy
RHESSys
CRHM
UBCWM
WaSiM-ETH
Mod
el F
unct
iona
lity
for F
ores
t M
anag
emen
t
Model Functionality for Climate Change
Lowest Highest
Low
est
Hig
hest
Intermediate
Inte
rmed
iate
f igure 2 Model functionality for forest management and climate change.
38
introduction, VIC is a macroscale model. This means that VIC is designed for watersheds scales greater than 500 km2, which is larger than most forest management applications. However, because of the recent application of VIC in BC watersheds and in the Columbia Basin, a discussion and an understanding of its functionality are warranted.
5.2.5 Model Selection
Figure 3 compares model complexity (reproduced from Figure 1) and simultaneous model functionality for answering forest management and climate change questions (obtained by combining the rank-ings from each axis in Figures 2). This figure indicates that RHESSys has highest overall functionality, followed by WaSiM-ETH, DHSVM, and CRHM. A model-selection procedure within the context of climate change would normally proceed along similar lines as are outlined in Section 3.1, with addi-tional consideration given for advantages and disadvantages of each model for answering climate change specific questions (as discussed in Section 5.2).
barriers to the operational use of models in a climate change context5.3
This section seeks to address the following general questions related to using watershed hydrologic mod-els in a forest management and climate change context:
• Whatimprovementsareneededtoexistingmodelsand/oristhedevelopmentofnewmodelsrequired?
WRENSS
UBC-UF Peak Flow Model
BROOK90
DHSVM
ForWaDy
RHESSys
CRHM
UBCWM
WaSiM-ETH
Mod
el F
unct
iona
lity
for F
ores
t M
anag
emen
t an
d Cl
imat
e Ch
ange
Model Complexity
Low(Screening Level Studies)
High(Complex Studies)
Low
est
Hig
hest
Medium
Inte
rmed
iate
f igure 3 Model complexity and functionality for forest management and climate change.
39
• Whatarecurrentbarriersand/orchallengestousinghydrologicalmodelsforclimatechangesce-narios in a forest management context, i.e., what are the next steps required?
• Whatadditionaldata,resources,and/ortoolsarerequiredtomoveforwardinthisarea?
The discussion is organized along the eight topic areas identified in Section 5.1, with additional con-sideration for the challenges of linking hydrologic models to climate change predictions.
Increased Atmospheric Evaporative Demand
The relative level of confidence in the results of more physically based approaches to calculating evapo-transpiration that use the necessary climate variables, LAI, and stomatal resistance, as employed in BROOK90, ForWaDy, CRHM, DHSVM, RHESSys and WaSiM-ETH (Table 11), is greater when climate change is introduced as a disturbance factor, provided that the required meteorological-forcing data are available. This is because physically based equations are not hard-coded in historical data (as is the case with empirical methods) and are thus inherently better suited for predicting possible shifts in hydrologic responses outside historical ranges. The physically based Penman-Monteith equation is recommended by the Food and Agricultural Organization (FAO) of the United Nations and the American Society of Civil Engineers (ASCE) to determine reference evapotranspiration (Allen et al. 2005). Overall, the theoretical understanding of suitable equations for calculating evapotranspiration is in a mature state, and the main challenge in anticipating future increases in evaporative demand comes from a lack of quantitative understanding regarding possible shifts in temperature in different portions of BC and AB.
Altered Vegetation Composition Affecting Evaporation and Interception
Of the models reviewed, only RHESSys is able to simply account for forest mortality and forest growth (Table 11) by including BIOME-BGC to allocate carbon and nitrogen to the various tissues (leaves, roots, and stems) that make up plant biomass (Section A1.15). Temporal input control, to allow for dynamic vegetation changes during the course of a single model run, is also generally lacking in most models reviewed (Table 11). To move forward in this topic area, improvements might include the following:
• Adaptinghydrologicmodelstoincludeforestgrowthandmortalityequationsorlinkingthesemodels to forest productivity and growth models.
• Addingorimprovingtemporalinputcontrolinmodels.• Overcomingthedocumenteddifficultythatmostexistinghydrologicalmodelsdevelopedforrela-
tively humid or sub-humid conditions have in producing acceptable results in more arid climates (e.g., Pike 1998; Tague et al. 2004). The inability to accurately account for semi-arid conditions may lead to a bias in evapotranspiration estimates under conditions of high atmospheric water demand and limited soil moisture availability, and should therefore be an area of continued re-search and improvement.
Further to these model improvements, there is a need to do the following:
• ProducespatiallyexplicitvegetationdatasetsforBCandABthatincludeup-to-dateestimatesofLAI and stomatal resistance.
• ConductresearchtopredicttheexpectedfuturevegetationcompositionduetoclimatechangeinBC and AB. Some efforts have already been made in this direction (e.g., Pike et al. 2008b; Gayton 2008 and references therein). Nonetheless, physiological characteristics of future vegetation, such as LAI and stomatal resistance, should also be researched and catalogued in databases for use in physically based models.
40
Decreased Snow Accumulation and Accelerated Melt
Most of the models reviewed should be able to account for shifts in snow accumulation and melt pro-cesses (i.e., potential smaller snowpacks and more rapid snowmelt) due to a changing climate, including simulation of mixed regimes (Table 11). The physically based or analytical approaches to calculating snowmelt, as employed in CRHM, DHSVM, ForWaDy, RHESSys, UBCWM, and WaSiM-ETH, are gen-erally better suited to addressing climate change effects compared to the empirical approaches employed in BROOK90, the UBC-UF peak flow model, and WRENSS. For the purpose of long-term simulations, models have to work well when presented with systems that are initially predominantly nival, and which may, over time as temperature increases, shift to mixed or perhaps even pluvial conditions. The physical-ly based approaches will be able to better deal with such shifts in hydrologic regimes over the modelling period, whereas empirical relationships developed for current conditions may not hold under a shifting climate.
It appears then that a theoretical understanding of suitable methods for snowmelt calculations is in a relatively mature state. The main challenge in this topic area appears to be improving the quantita-tive understanding regarding the magnitude of temperature and precipitation shifts that may occur and associated implications for snow accumulation and melt patterns across the region (i.e., climate change research).
Accelerated Melting of Permafrost, Lake Ice, and River Ice
River and lake ice formation and break-up processes are often the focus of specialized kinematic models (e.g., Beltaos 2007) that are not typically incorporated into watershed-scale hydrologic models (Table 2). Soil temperatures, however, are more widely accounted for in hydrologic models, typically to calcu-late the ground heat-flux component of the snowpack energy balance (e.g., Wigmosta and Lettenmaier 1994). Only the CRHM (Pomeroy et al. 2007) has the ability to assess frozen soil conditions (via soil temperatures) and associated effects on water movement (Table 2). Several general modelling improve-ments are therefore suggested that include the following:
• Givinghydrologicmodelsagreaterabilitytosimulatetheeffectsofpermafrostthawonhydrologi-cal processes applicable to the northern portions of BC and AB and other areas where permafrost occurs. Frozen soil conditions may also be important to model in non-permafrost areas (i.e., effects on infiltration).
• Improvingourunderstandingofhowclimatechangewillalterthethree-wayinteractionbetweenstreamflow generation, water temperatures, and river and lake ice formation and break-up.
• Developingtoolsthatallowresourcemanagerstoassesstheimportanceoftheseinteractions(andhow they may change in the future) for forest management.
Glacier Mass Balance Adjustments
Glacial processes are represented in WaSiM-ETH and UBCWM (Table 11). These models, therefore, can handle calculation of increases in melt rates due to climate change. However, for the purposes of long-term simulations, it is also necessary that glacier mass balances are calculated and that glacier areas can be adjusted internally based on these calculations (i.e., to simulate glacial retreat). This capability was specifically developed in the HBV-EC version used by Stahl et al. (2008) and should be expanded to oth-er models. Alternatively, stand-alone, glacier mass balance models can be used to estimate future glacier volume, which can be used as an input to hydrologic models with glacier processes, such as WaSiM-ETH and UBCWM.
41
Altered Surface Water Flows
Improvements with respect to simulating altered surface water flows (peak and low flows) in a climate change context are contingent on advances in the above topic areas. In some regions, dominant water-shed hydrologic processes may change entirely in the next 100 years to the point that what starts out as a snowmelt-dominated regime may change to a mixed (hybrid) or potentially even to an entirely rain-dominated regime (Pike et al. 2008b). If a specific model is incapable of modelling hybrid processes, such as rain-on-snow (Table 11), it will generally have a lower level of accuracy in predicting the ef-fects of climate change than those models that can. Furthermore, model accuracy for predicting future streamflow conditions may decline if a model was developed and calibrated for simulating snowmelt-dominated watershed conditions and is subsequently used for assessing the consequences of a regime shift to mixed or rainfall-dominated conditions. Improvements in this topic area should include allow-ances in more models for processes related to the presence of groundwater, wetlands, and lakes, and factors such as human water consumption.
Increased Stream and Lake Temperatures
While numerous models have been developed to predict stream temperature (e.g.,Webb et al. 2008), none of the hydrologic models reviewed here has this capability (Table 11). This limits the ability of resource managers to account in a straightforward manner for possible interactions between shifts in surface water flows and stream temperature in a changing climate. To make advances in this topic area, existing watershed models need to be adapted to calculate water temperature and stream temperature. Given that the more physically based hydrologic models already incorporate energy-balance calculations, this ought to be a relatively straightforward extrapolation of existing capabilities.
Increased Frequency/Magnitude of Disturbances
Much emphasis has been placed on predicting climate change shifts in average meteorological condi-tions, but understanding the potential future changes in the frequency and magnitude of extreme events is considerably poorer (Rodenhuis et al. 2007). As such, a quantitative understanding of the expected increase in the occurrence of extreme events (temperature, precipitation, and wind) must be improved to move forward in this topic area.
Increased operational use of models for analyzing disturbances such as landslide hazards, mass wast-ing, fire hazards, pests (insects), and windthrow is also needed and the output from these models could be used to parameterize hydrologic models for watershed-planning purposes in a climate change con-text. The more complex, physically based models are generally better suited to parameterize the effects of these disturbances because they include multi-layered vegetation and parameters such as soil and vegetation albedo (Table 11). Representing the potential effect of soil hydrophobicity on infiltration can be problematic because, while it is possible to alter soil physical properties, none of the models reviewed allows soil properties to be changed temporally within a single simulation to account for a decrease in hydrophobicity over time.
Linking Hydrologic Models to Climate Change Predictions
Climate change projections are generated from global-scale, general circulation models (GCMs). GCMs provide output at a resolution too coarse (typically grid cells are several 100 km a side or several 1000 km2) to be of use in forest management studies. Statistical downscaling techniques or dynamical tech-niques such as regional climate models (RCMs) are therefore required to downscale GCM outputs to regional and local scales to estimate climate impacts and response within hydrological models (Hutchin-son and Roche 2008). Linking GCM climate change projections to hydrological models is particularly
42
onerous in complex mountainous terrain, a challenge that has been well documented (e.g., Wood et al. 2004; Merritt et al. 2006; Stahl et al. 2008).
Statistical downscaling techniques are relatively efficient to run and therefore can be used to down-scale several GCMs to explore the range in future scenarios. Four techniques that have been applied in western North America include the following:
• Thedeltamethod,whereahistoricaltimeseriesisadjustedbythedifferencebetweenthecurrent(often 1961–1990) and the future (i.e., 2041–2070), (e.g., Loukas et al. 2002; Toth et al. 2006). Al-though this method maintains the observed sequence of weather and natural variability, the trend from the GCM is not preserved in the projection, and changes to variance and skewness will be missed (Mote and Salathe 2009).
• ClimateBC(Spittlehouse2006),developedbyWangetal.(2006),downscalesGCMdatausinga combination of bilinear interpolation and elevation adjustments. Monthly temperature and precipitation data from PRISM, a gridded 4-km2 dataset that has been corrected for distance from the ocean, elevation, aspect, wind direction, and slope, is adjusted to produce scale-free data over BC. Anomalies from the GCM are added to the adjusted PRISM dataset. Where the PRISM dataset is not very accurate, ClimateBC results will not be robust. This includes areas in northern BC and areas at high latitidudes where there is a lack of observational data, and areas affected by cold-air drainage.
• Bias-correctionstatisticaldownscaling(BCSD),producesadailytimeseriesbybiascorrectingmonthly GCM data to match the statistical properties of a 1/16th degree gridded record (observed), locally scaling these values to give a representative spatial distribution and disaggregating the resulting monthly values to daily by resampling the historic dataset (Wood et al. 2002; Widmann et al. 2003; Salathé 2005). Potential drawbacks of this method include not adequately reflecting changes to the diurnal temperature range, which has consequences for simulating future snowmelt and evapotranspiration.
• Tree-GEN,amethoddevelopedbyEnvironmentCanada(A.Cannon,pers.comm.,Dec.2009),includes components from multiple statistical techniques to achieve optimal results (Stahl et al. 2008), but has been applied at only a few sites in BC.
Climate change impacts in BC and AB are currently being investigated by the Pacific Climate Im-pacts Consortium (PCIC) with each of the tools listed above. In particular, PCIC is collaborating with Tongli Wang (UBC) to improve ClimateBC by creating additional components and extending the tool to western North America. PCIC is also applying the BCSD technique for use with the VIC model at several watersheds in BC and is currently learning Tree-GEN.
Dynamic downscaling approaches, such as RCMs, are also available for translating course resolution GCM results to finer scales. In contrast to statistical techniques, RCMs are computationally expensive to run. However, they are representative of the physical processes and therefore conserve energy and water balances, which statistical methods do not. PCIC is working to conduct diagnostics of the Canadian Regional Climate Model (RCM) over Pacific North America. Initial results from the RCM will be provided at a grid scale resolution of 45 x 45 km for BC and beyond. At this resolution, it is advisable to statistically downscale results prior to applying them in hydrologic models to better reflect local factors such as topography. In the future, RCM results may be available for BC over a 15 x 15-km grid.
Continued development of such tools and associated data is underway at research centers in BC and AB. Statistical and dynamical techniques are being applied to downscale using NARR data (G. Clark and C. Reuten, UBC); a three-dimensional, mesoscale numerical atmospheric model (RAMS) is being run at an 8 x 8-km resolution (P. Jackson, UNBC); and synoptic typing is being applied in statistical
43
downscaling (J. Shea, UBC and S. Marshall, U of C) (A. Werner, pers. comm., Jan. 2009). However, due to the challenge of this work in complex terrain, further research is required. PCIC will continue to learn and apply new tools to downscale GCMs and to improve information on how downscaling techniques compare and how they might be best applied (Murdock 2009).
summAry And conclusions6
This review and synthesis has summarized the capabilities and limitations of existing hydrologic mod-els for answering forest management and climate change questions in an operational context in British Columbia and Alberta. The review considered a total of 30 models: ACRU, BROOK90, CRHM, DHSVM, ForHyM, ForWaDy, HBV-EC, HEC-HMS, HELP, HSPF, HydroGeoSphere, InHM, Mike-SHE, MOD-HMS, PREVAH, PRMS/MMS, RHESSys, SSARR, SWAT, UBC-UF Peak Flow Model, UBCWM, VIC, WaSiM-ETH, Water Balance Model by QUALHYMO, Watflood, WEPP, WRENSS (as WinWrnsHyd and ECA-AB spreadsheet programs), WRMM, and WUAM. The review brought together relevant infor-mation contained in user manuals, technical model documentation, and published model studies in settings with physiographic and climatic characteristics similar to those encountered in AB and BC, and emphasized model applications in the Pacific Northwest.
One of the main outcomes of this review is that there currently is no “best” model for use in an operational forest management context. Instead, nine models were identified that could potentially be used for addressing forest management and climate change questions in BC and AB. Each of these nine models is characterized by advantages and disadvantages for operational use, and is only applicable to particular physiographic or climatic settings.
Only one low-complexity model exists, the WRENSS (WinWrnsHyd, ECA-AB), with functionality limited to simulating changes in annual yield due to forest harvesting and subsequent recovery.
Among the medium-complexity models, UBCWM and BROOK90 rank as the models with greatest functionality in a forest management context, with the former being applicable to mountainous settings, including glaciarized terrain and watersheds with upland lakes, and the latter being applicable to small first-order watersheds in more gradually sloped terrain. ForWaDy is a possible alternative to BROOK90 when forest growth needs to be considered, as it can be linked to the FORECAST and FORCEE growth models. The UBC-UF peak flow model, which is being developed under the lead of Dr. M. Weiler at the University of Freiburg, promises to integrate a novel model approach for simulating forest management effects on peak flows using a provincial (BC) database containing the information on watershed physical characteristics that is required to run the model. Overall, the medium-complexity model category has limited capability with respect to answering forest management questions under possible settings en-countered in AB and BC. These limitations can only be overcome by applying suitable high-complexity models.
Among high-complexity models, DHSVM and RHESSys rank as the models with the greatest forest management functionality. However, in certain settings, WaSiM-ETH offers a number of advantages over both DHSVM and RHESSys, including a rigorous treatment of groundwater processes, a glacier component, a lake model, and a river (channel) routing model that accounts for artificial and natural reservoirs. The WaSiM-ETH developer (Dr. J. Schulla) now works for the Pacific Climate Impacts Consortium (PCIC), bringing expertise with the model to Western Canada. CRHM was specifically developed to consider watershed processes that are applicable to boreal forest settings, including how blowing snow and frozen soils influence water movement. Jointly, when compared to
44
medium-complexity models, these four high-complexity models offer a considerably broader range of functionality for answering forest management questions in AB and BC. These models address the limitations of medium-complexity models, provided that their higher demands on data, time (budget), and resources (GIS and model calibration) can be overcome.
Trade-offs between model functionality (accuracy) and model complexity were identified in this review and were found to complicate the operational application of hydrologic models for answering forest management questions. Recommendations for advancing the routine and consistent use of models were made, including the need to improve interdisciplinary education and training and model intercomparisons at data-rich (experimental) and data-poor (ungauged) watersheds; to enhance data availability; and to create better (more flexible) models that can be adapted to handle a range of circum-stances, GUIs, commercial software, and model support. Finally, there is a need to provide regulatory guidance and professional precedence.
While not currently suitable for forest management applications, the HBV-EC is being rewritten, under the lead of Dr. R. D. Moore at UBC, to make this model more suitable for answering forest water-shed management questions. The new model, which is expected to be completed by late 2009, has the potential to alleviate some of the challenges and trade-offs that exist with current hydrologic models. This development is promising and should be closely monitored. A recent version of the HBV-EC also contains a component to internally calculate glacier mass balances (i.e., to simulate glacial retreat), a capability that is useful for climate change applications.
The suitability of selected models for exploring potential climate change effects on future water-shed processes and outputs relevant to forest management was also considered. Of the nine models determined most suitable for addressing forest management questions, BROOK90, CRHM, DHSVM, ForWaDy, RHESSys, and WaSiM-ETH should be able to best handle issues regarding changes in evapo-transpiration and plant stress. CRHM, DHSVM, ForWaDy, RHESSys, UBCWM, and WaSiM-ETH are best suited to account for shifts in snow accumulation and melt processes. Frozen soils and associated consequences for water movement are only considered in CRHM, while glacial processes can be repre-sented in UBCWM and WaSiM-ETH. DHSVM may be best suited to answer the watershed management questions associated with the potential for increased flood flows and mass wasting due to climate change. RHESSys is the only model with an ecology component (BIOME-BGC). This ecology model component could be used to investigate the effects of hydrological processes on the distribution, struc-ture, and function of ecosystems; the effects of biotic processes on elements of the water cycle; and how these two-way interactions are affected by climate change.
Research and model development needs were identified to move forward the interdisciplinary area of forest management and climate change. Overall, it appears that developing new models is not necessarily required, and that incrementally enhancing the capabilities of existing hydrologic models should suffice to address hydrologic climate change questions. Data resources need to be aligned to allow application of the complex physically based models that are inherently better suited for answering climate change questions. Fundamental climate change research is also needed to better quantify the understanding of possible future shifts in temperature and precipitation, the occurance of extreme events, and the changes in glacier mass balances. Continued development of tools such as statistical downscaling to link hydro-logic models to climate change predictions is required, with steps in this direction being taken by the PCIC and researchers at several universities in BC, AB, and Washington.
45
references7
Abbott, M.B., J.C. Bathurst, J.A. Cunge, P.E. O’Connel, and J. Rasmussen. 1986a. An introduction to the European Hydrological Systems-Systeme Hydrologique Europeen, ‘SHE’. 1. History and philosophy of a physically based distributed modelling system. Journal of Hydrology 87:45–59.
Abbott, M.B., J.C. Bathurst, J.A. Cunge, P.E. O’Connel, and J. Rasmussen. 1986b. An introduction to the European Hydrological Systems-Systeme Hydrologique Europeen, ‘SHE’. 2. Structure of a physically based distributed modelling system. Journal of Hydrology 87:61–77.
Agee, J.K. 1993. Fire ecology of Pacific Northwest forests. Island Press: Washington, DC.
Ahl, R.S., S.W. Woods, and H.R. Zuuring. 2008. Hydrologic calibration and validation of SWAT in a snow dominated rocky mountain watershed, Montana, U.S.A. Journal of the American Water Resources Association 44(4):1–21.
Alila, Y. and C.Q. Luo. 2007. Delinieating the limits on peak flow and water yield responses to clearcut salvage logging in large watersheds. Mountain Pine Beetle and Watershed Hydrology Workshop: Preliminary Results of Research from BC, Alberta and Colorado, Kelowna, BC.
Alila, Y. and J. Beckers. 2001. Using numerical modelling to address hydrologic forest management issues in British Columbia. Hydrological Processes 15:3371–3387.
Allen, R.G., L.S. Pereira, M. Smith, D. Raes, and J.L. Wright. 2005. FAO-56 dual crop coefficient method for estimating evaporation from soil and application extensions. Journal of Irrigation and Drainage Engineering-ASCE 131(1):2–13.
Anderson, E.A. 1973. National Weather Service River Forecast System—snow accumulation and ablation model. NOAA Technical Memo. NWS. U.S. Department of Commerce, Silver Spring.
Armbuster, M., J. Seegert, and K.H. Feger. 2004. Effects of changes in tree species composition on water flow dynamics–Model applications and their limitations. Plant and Soil 264:13–24.
Arnold, J.G., J.R. Williams, A.D. Nicks, and N.B. Sammons.1989. SWRRB, a simulator for water resources in rural basins, Agricultural Research Service, USDA, Texas A&M University Press, College Station, TX.
Arp, P.A. and X. Yin. 1992. Predicting water fluxes through forests from monthly precipitation and mean monthly air temperature records. Canadian Journal of Forest Research 22:864–877.
Assaf, H. 2007. Development of an energy-budget snowmelt updating model for incorporating feedback from snow course survey measurements. Journal of Engineering, Computing and Architecture 1(1).
Baird, A.J. and R.L. Wilby. 1999. Eco-hydrology: Plants and water in terrestrial and aquatic environments. Routledge, London.
Band, L.E., C.L. Tague, S.E. Brun, D.E. Tenenbaum, and D.A. Fernandes. 2000. Modelling watersheds as spatial object hierarchies: Structure and dynamics. Trans GIS 4:181–196.
Band, L.E., C.L. Tague, P. Groffman, and K. Belt. 2001. Forest ecosystem processes at the watershed scale: Hydrological and ecological controls of nitrogen export. Hydrological Processes 15:2013–2028.
Beckers, J. and Y. Alila. 2004. A model of rapid preferential hillslope runoff contributions to peak flow generation in a temperate rain forest watershed. Water Resources Research 40, W03501, DOI:10.1029/2003WR002582.
Beeson, P., C. Duffy, E. Springer, and S. Panday. 2004. Integrated hydrologic models for closing the water budget. Whitewater River Basin, Kansas. Eos Trans. AGU, 85(47), Fall Meeting Supplement, Abstract H34B-05.
46
Beltaos, S. 2007. River ice breakup processes: Recent advances and future directions. Canadian Journal of Civil Engineering. 34(6):703–716.
Berris, S.N. and R.D. Harr. 1987. Comparative snow accumulation and melt during rainfall in forested and clear cut plots in the western Cascades of Oregon. Water Resources Research 23:135–142.
Beven, K.J. and A. Binley. 1992. The future of distributed models: Model calibration and uncertainty prediction. Hydrological Processes 6:279–298. Beven, K. and M. Kirkby. 1979. A physically based variable contributing area model of basin hydrology. Hydrololgy Science Bulletin 24:43–69.
Beven, K. 1989. Changing ideas in hydrology—the case of physically-based models. Journal of Hydrology 105:157–172.
Bicknell, B.R., J.C. Imhoff, J.L. Kittle Jr., T.H. Jobes, and A.S. Donigian, Jr. 2001. Hydrological Simulation Program–Fortran (HSPF). User’s Manual for Release 12. U.S. EPA National Exposure Research Laboratory, Athens, GA, in cooperation with U.S. Geological Survey, Water Resources Division, Reston, VA.
Bingeman, A.K., N. Kouwen, and E.D. Soulis. 2006. Validation of the hydrological processes in a hydrological model.” Journal of Hydrologic Engineering 451–462.
Bowling, L.C., J.W. Pomeroy, and D.P. Lettenmaier. 2004. Parameterization of blowing snow sublimation in a macroscale hydrology model. Journal of Hydrometeorology. 5(5):745–762.
Bowling, L.C., P. Storck, and D.P. Lettenmaier. 2000. Hydrologic effects of logging in Western Washington, United States, Water Resources Research. 36:3223–3240.
British Columbia Ministry of Water, Land and Air Protection. 2002. Indicators of climate change for British Columbia. Government of British Columbia, Victoria, BC.
Burges, S.J. 2003. Process representation, measurements, data quality, and criteria for parameter estimation of watershed models. In Calibration of Watershed Models. Q. Duan, H. Gupta, S. Sorooshian, A. N. Rousseau, and R. Turcotte (editors). American Geophysical Union Monograph, Water Science and Application 6:283–299
Burges, S.J., M.S. Wigmosta, and J.M. Meena. 1998. Hydrological effects of land-use change in a zero-order catchment. ASCE Journal of Hydrological Engineering 3:86–97.
Byrne, J. M., R. Barendregt, and D. Schaffer. 1989. Assessing potential climate change impacts on water supply and demand in southern Alberta. Canadian Water Resources Journal 14(4):18–29.
Calvert, P.P. 2003. Modelling the effects of logging roads on the streamflow of a mountainous, snow-dominated watershed. MSc. Thesis. The University of British Columbia, Vancouver, BC.
Chang, M. 2006. Forest Hydrology, 2nd Edition, CRC Press, Boca Raton.
Cheng, J.D. 1989. Streamflow changes after clear-cut logging of a pine beetle-infested watershed in southern British Columbia, Canada. Water Resources Research 25(3):449–456.
Cherkauer, K.A. and D.P. Lettenmaier. 1999. Hydrologic effects of frozen soils in the Upper Mississippi River Basin. Journal of Geophysical Research 104(D16):19,599–19,610.
Cherkauer, K.A., L.C. Bowling, and D.P. Lettenmaier. 2003. Variable infiltration capacity (VIC) cold land process model updates. Global and Planetary Change 38(1-2):151–159.
Christensen, L., C.L. Tague, and J.S. Baron. 2008. Spatial patterns of simulated transpiration response to climate variability in a snow dominated mountain ecosystem. Hydrological Processes 22(18):3576–3588.
Christensen, N.S., A.W. Wood, N. Voisin, D.P. Lettenmaier, and R.N. Palmer. 2004. The effects of climate change on the hydrology and water resources of the Colorado River Basin. Climatic Change 62: 337–363.
Clark, C.O. 1945. Storage and the unit hydrograph. Proceedings of the American Society of Civil Engineering 69:1419–1447.
47
Conroy, W.J., R.H. Hotchkiss, and W.J. Elliot. 2006. A coupled upland erosion and instream hydrodynamic-sediment transport model for evaluating sediment transport in forested watersheds. American Society of Agricultural and Biological Engineers 49(6):1713–1722.
Coughlan, J.C. and S.W. Running. 1997. Regional ecosystem simulation: A general model for simulating snow accumulation and melt in mountainous terrain. Landscape Ecology 12:119–136.
Covert, S.A., P.R. Robichaud, W.J. Elliot, and T.E. Link. 2005. Evaluation of runoff prediction from WEPP-based erosion models for harvested and burned forested watersheds. Soil and Water Division of ASAE 48(3):1091–1100.
Daly, S.F., R. Davis, E. Ochs, and T. Pangburn. 2000. An approach to spatially distributed snow modelling of the Sacramento and San Joaquin basins, California. Hydrological Processes 14:3257–3271.
DeBano, L.F. 2000. Water repellency in soils: A historical overview. Journal of Hydrology 232:4–32.
Demaria, E.M., B. Nijssen, and T. Wagener. 2007. Monte Carlo sensitivity analysis of land surface parameters using the Variable Infiltration Capacity model, Journal of Geophysical Research 112, D11113, DOI:10.1029/2006JD007534.
Demuth, M.N. and A. Pietroniro. 2002. The impact of climate change on the glaciers of the Canadian Rocky Mountain eastern slopes and implications for water resourcesrelated adaptation in the Canadian prairies. In Praire Adaptation Research Collaborative (PARC). Synthesis of research projects (Draft).
Dhakal, A.S. and R.C. Sidle. 2003. Long-term modelling of landslides for different forest management practices. Earth Surface Processes and Landforms 28, DOI:10.1002/esp.499.
Doten, C.O., L.C. Bowling, J.S. Lanini, E.P. Maurer, and D.P. Lettenmaier. 2006. A spatially distributed model for the dynamic prediction of sediment erosion and transport in mountainous forested watersheds. Water Resources Research 42, W04417, DOI:10.1029/2004WR003829.
Dunne, T. 2001. Introduction to Section 2—Problems in measuring and modeling the influence of forest management on hydrologic and geomorphic processes. In Land use and watersheds: Human influence on hydrology and geomorphology in urban and forest areas. AGU Water Science and Application 2:77–83.
Eaton, B. and R.D. Moore. 2007. Chapter 4—Regional hydrology [draft]. In Compendium of forest hydrology and geomorphology in British Columbia. R.G. Pike et al. (editors). B.C. Ministry of Forests and Range, Research Branch, Victoria, B.C. and FORREX Forest Research Extension Partnership, Kamloops, B.C. Land Management Handbook. In preparation. www.forrex.org/program/water/compendium.asp
Ebel, B.A., K. Loague, J.E. Vanderkwaak, W.E. Dietrich, D.R. Montgomery, R. Torres, and S.P. Anderson. 2007. Near-surface hydrologic response for a steep unchanneled catchment near Coos Bay, Oregon: 2. Physics based simulations. American Journal of Science 907:709–748.
Elliot, W.E., S.M. Graves, D.E. Hall, and J.E. Moll. 1998. The X-Drain cross-drain spacing and sediment yield model. United States Department of Agriculture and Forest Service Technology and Development Program. San Dimas Technology and Development Center, San Dimas, CA.
Federer, C.A., C. Vörösmarty, and B. Fekete. 2003. Sensitivity of annual evaporation to soil and root properties in two models of contrasting complexity. Journal of Hydrometeorology 4:1276–1290.
Feyen, L., R. Vazquez, K. Christiaens, O. Sels, and J. Feyen. 2000. Application of a distributed physically-based hydrological model to a medium sized catchment. Hydrology and Earth System Sciences 4: 47–63.
Fischenich, C. 1999. Overview of stream restoration technology. State of the Science, EMRRP Vol. 2, No. 3, U.S. Army Engineer Research and Development Center, Vicksburg, MS.
Flanagan, D.C. and S.J. Livingston. 1995. WEPP user summary. National Soil Erosion Research Laboratory (NSERL) Report no 11. W. Lafayette, IN.
48
Forest Practices Board. 2007. The effect of mountain pine beetle attack and salvage harvesting on streamflows, special investigation, FPB/SIR/16.
Gassman, P.W., M.R. Reyes, C.H. Green, and J.G. Arnold. 2007. The soil and water assessment tool: historical development, applications, and future research directions. American Society of Agricultural and Biological Engineers 50(4):1211–1250.
Gayton, D. 2008. Impacts of climate change on British Columbia’s biodiversity: A literature review. Forrex Forest Research Extension Partnership, Kamloops, BC. Forrex Series 23. www.forrex.org/publications/forrexseries/fs23.pdf
Gluns, D.R. and R.D. Eremko. 1988. Opportunistic use of hydrologic data to assess CFS-WRENNS for Forestry Interpretations in South Central B.C. In Proceedings, Canadian Hydrology Symposium, NRC Canada, Associate Committee on Hydrology, May 9-11, 1988, Banff, AB.
Golding, D.L. 1986. Hydrological implications of salvage harvesting lodgepole pine in hydraulic creek watershed. Unpublished report for B.C. Ministry of Forests, Kamloops Forest Region.
Graham, D.N. and M.B. Butts. 2005. Chapter 10: Flexible integrated watershed modeling with MIKE SHE. In Watershed Models. V.P. Singh and D.K. Frevert (editors). CRC Press, Boca Raton, FL, USA.
Granger, R.J., D.M. Gray, and G.E. Dyck. 1984. Snowmelt infiltration to frozen prairie soils. Canadian Journal of Earth Sciences 21(6):669–677.
Gray, D.M., B. Toth, J.W. Pomeroy, L. Zhao, and R.J. Granger. 2001. Estimating area snowmelt infiltration into frozen soils. Hydrological Processes 15:3095–3111.
Grayson, R.B., I.D. Moore, and T.A. McMahon. 1992a. Physically based hydrologic modelling 1: A terrain-based model for investigative purposes. Water Resources Research 28:2639–2658.
__________. 1992b. Physically based hydrologic modelling 2: Is the concept realistic. Water Resources Research 28:2659–2666.
Gurtz, J., A. Baltensweiler, and H. Lang. 1999. Spatially distributed hydrotope-based modelling of evapotranspiration and runoff in mountainous basins. Hydrological Processes 13:2751–2768.
Gurtz, J. and M. Zappa. 2004. Application training on hydrological models ETH Zürich–Swiss Federal Institute of Technology Institute for Atmospheric and Climate Science (IACETH), Zürich, Switzerland.
Hamilton, A.S., D.G. Hutchinson, and R.D. Moore. 2000. Estimating winter streamflow using a conceptual streamflow model. Journal of Colds Region Engineering 14:158–175.
Hamilton, S. 2007. Completing the loop: From data to decisions and back to data. Hydrological Processes 21(22):3105–3106.
Hamlet, A.F. and D.P. Lettenmaier. 1999. Effects of climate change on hydrology and water resources in the columbia river basin. Journal of the American Water Resource Association, 35(6):1597–1624.
__________. 2000. Long-range climate forecasting and its use for water management in the Pacific Northwest Region of North America. Journal of Hydroinformatics 02.3: 163–182.
__________. 2005. Production of temporally consistent gridded precipitation and temperature fields for the Continental United States. Journal of Hydrometeorology 6:330–336.
Hanson, C.L., K.A. Cumming, D.A. Woolhiser, and C.W. Richardson. 1994. Microcomputer program for daily weather simulation in the contiguous United States. USDA Agricultural Research Service, Publication ARS-114.
Hetherington, E.D., B. Walsh, and M. Leytham. 1995. Calibration of HSPF hydrologic simulation model using Carnation Creek Experimental Watershed data. In Mountain hydrology; peaks and valleys in research and application. B.T. Guy and J. Barnard (editors.) Proceedings of a Conference May 16-19, 1995, Vancouver BC. Canadian Water Resources Association, ON.
49
Hillel, D. 1986. Modelling in soil physics: A critical review. In Future developments in soil science research. Soil Science Society of America, Madison, WI.
Hock, R. 2003. Temperature index melt modelling in mountain areas. Journal of Hydrology 282:104–115.
Hudson, R. 2000. Assessing snowpack recovery of watersheds in the Vancouver Forest Region. BC Ministry of Forests. Nanaimo, BC.
Hudson, R.O. and M.C. Quick. 1997. A component based water chemistry simulator for small subalpine watersheds. Canadian Water Resources Journal 22(3):299–325.
Hutchinson, D. 2007. Modelling hydrologic effects of landscape disturbance—an introduction. In Mountain pine beetle and watershed hydrology workshop: Preliminary results of research from BC, Alberta and Colorado, Ramada Inn, Kelowna, BC, July 10, 2007.
Hutchinson, D., B. Taylor, and S. Hamilton. 1999. Impacts of climate change on snow pack and runoff in the Capilano Watershed using the UBC Watershed Model. Environment Canada.
Hutchinson, D. and D. Roche. 2008. Changing climate,uncertain futures, and evolving practices, symposium summary. Streamline 12(1):39–42.
Ilich, N., P. Simonovic, and M. Amron. 2000. The benefits of computerized real-time river basin management in the Malahayu reservoir system. Canadian Journal of Civil Engineering 27:55–64.
Ivanovic, R.F. and J.E. Freer. 2009. Science versus politics: Truth and uncertainty in predictive modelling. Hydrological Processes 23:2549–2554.
Jasper, K., P. Calanca, D.S. Gyalistras, and J. Fuhrer. 2004. Differential impacts of climate change on the hydrology of two alpine river basins. Climate Research 26:113–129.
Jensen, M.E., D.C.N. Robb, and C.E. Franzoy. 1969. Scheduling irrigation using climate-crop-soil data. In Proceedings of the American Society of Engineers National Conference on Water Resoures Engineering: New Orleans, LA.
Jorgenson, M.T., C.H. Racine, J.C. Walters, and T.E. Osterkamp. 2001. Permafrost degradation and ecological changes associated with a warming climate in central Alaska. Climate Change 48: 551–579.
Jost, G., M. Weiler, D.R. Gluns, and Y. Alila. 2007. The influence of forest and topography on snow accumulation and melt at the watershed-scale. Journal of Hydrology 347:101–115.
Kampf, S.K. and S.J. Burges. 2007. A framework for classifying and comparing distributed hillslope and catchment hydrologic models. Water Resoures Research 43: W05423, DOI:10.1029/2006WR005370.
Kassem, A. 1992. The Water Use Analysis Model (WUAM) program documentation and reference manual. Economics and Conservation Branch, Ecosystem Sciences and Evaluation Directorate, Environment Canada, Ottawa, ON.
Kassem, A.M., D.M. Tate, and P.A. Dossett. 1994. Water Use Analysis Model (WUAM) demonstration, Social Sciences Series No. 28, Environmental Conservation Service, Environment Canada, Ottawa, ON.
Kimmins, J.P., D. Mailly, and B. Seely. 1999. Modelling forest ecosystem net primary production: the hybrid simulation approach used in FORECAST. Ecological Model 122:195–224.
Kiniry, J.R., J.D. MacDonald, A.R. Kemanian, B. Watson, G. Putz, and E.E. Prepas. 2008. Plant growth simulation for landscape-scale hydrological modelling. Hydrological Sciences Journal 53:1030-1042. Special Issue: Advances in Ecohydrological Modelling with SWAT.
Kleinn, J., C. Frei, J. Gurtz, D. Lüthi, P.L. Vidale, and C. Schär. 2005. Hydrologic simulations in the Rhine basin driven by a regional climate model. Journal of Geophysical Research 110(D0), 4102, DOI:10.1029/2004JD005143.
50
Klemes, V. 1986. Operational testing of hydrological simulation models. Hydrological Sciences Journal 31:13–24.
Klok, E.J., K. Jasper, K.P. Roelofsma, A. Badoux, and J. Gurtz, J. 2001. Distributed hydrological modelling of a heavily glaciated alpine river basin.” Hydrological Sciences Journal 46(4):553–570.
Knisel, W.G., D.C. Moffitt, and T.A. Dumper. 1985. Representing seasonally frozen soil with the CREAMS model, American Society of Agricultural Engineering 28(September–October):1487–1492.
Koboltschnig, G.R., W. Schöner, M. Zappa, and H. Holzmann. 2007. Contribution of glacier melt to stream runoff: If the climatically extreme summer of 2003 happened in 1979. Annals of Glaciology 46:303–308.
Kouwen, N. 2008. WATFLOOD/WATROUTE Hydrological model routing & flow forecasting system. Since 1972. Developed for Surveys and Information Branch, Ecosystem Science and Evaluation Directorate, Environment Canada. Department of Civil Engineering, University of Waterloo, ON.
Krause, S., J. Jacobs, and A. Bronstert. 2007. Modelling the impacts of land-use and drainage density on the water balance of a lowland–floodplain landscape in northeast Germany. Ecological Modelling 200(3–4):475–492
Krezek, C., I. Creed, C. Tague, and D. Stanford, D. 2007. Impacts of timber harvesting on the flow regime of a coastal stream; from the headwaters to the entire watershed. In Air, ocean, earth and ice on the Rock. CMOS, CGU, AMS Congress 2007, St.John’s, NL.
Kristensen, K.J. and S.E. Jensen. 1975. A model for estimating actual evapotranspiration from potential transpiration. Nordic Hydrology, 6:170–188.
Krysanova, V. and J. Arnold. 2008. Advances in ecohydrological modelling with SWAT—a review. Hydrological Sciences–Journal–des Sciences Hydrologiques, 53(5) Special issue: Advances in Ecohydrological Modelling with SWAT.
Kuczera, G. 1994. NLFIT: A Bayesian nonlinear regression program suite. Department of Civil Engineering and Surveying, University of Newcastle, Newcastle, Australia.
Kuusisto, E. 1980. On the values and variablility of degree-day melting factor in finland. Nordic Hydrology 11:235–242.
Landrum, L., J. Baron, and C. Tague. 2002. Effects of projected climate change on ecosystem and hydrologic properties in the Big Thompson Watershed, Colorado: Model results from the Regional HydroEcological Simulation System (RHESSys). Americal Geophysical Union 83(47), Fall Meeting Supplement, Abstract H61B-0766.
Lanini, J.S., E.A. Clark, and D.P. Lettenmaier. 2009. Effects of fire-precipitation timing and regime on post-fire sediment delivery in Pacific Northwest forests. Geophysical Research Letters 36, L01402, DOI:10.1029/2008GL034588.
Lanquaye, C.O. and S.J. Mitchell. 2005. Portability of stand-level empirical windthrow risk models. Forest Ecology and Management 216:134–148.
Lamarche, J. and D.P. Lettenmaier. 2001. Effects of forest roads on flood flows in the Deschutes River basin, Washington. Earth Surface Processes and Landforms 26:115–134.
Leaf, C.F. 1975. Watershed management in the Rocky Subalpine zone: The status of our knowledge. USDA Forest Service Research paper RM-137.
Leavesley, G.H., S.L. Markstrom, R.J. Viger, and L.E. Hay. 2005. USGS Modular Modeling System (MMS) –Precipitation-Runoff Modeling System (PRMS) MMS-PRMS. In Watershed Models. V. Singh and D. Frevert (editors), Boca Raton, FL.
51
Lee, K.K. and J.C. Risley. 2002. Estimates of ground-water recharge, base flow and stream reach gains and losses in the Willamette River Basin, Oregon, U.S. Geological Survey Water-Resources Investigations.
Leung, L.R. and M.S. Wigmosta. 1999. Potential climate change impacts on mountain watersheds in the Pacific Northwest. Journal of the American Water Resources Association 35:1463–1471.
Li, Q., A.J.A. Unger, E.A. Sudicky, D. Kassenaar, E.J. Wexler, and S. Shikaze. 2008. Simulating the multi-seasonal response of a large-scale watershed with a 3D physically-based hydrologic model. Journal of Hydrology 357(3-4):317–336.
Li, X., M. Nour, D.W. Smith, E.E. Prepas, G. Putz, and B. Watson. 2008. Incorporating water quantity and quality modelling into forest management. Forestry Chronicle 84:338–348.
Liang, X., D.P. Lettenmaier, E.F. Wood, and S.J. Burges. 1994. A simple hydrologically-based model of land surface water and energy fluxes for general circulation models. Journal of Geophysical Research 99(D7):14,415–14,428.
Liggett, J.E. 2008. Comparison of approaches for aquifer vulnerability mapping and recharge modelling at regional and local scales, Okanagan Basin, British Columbia. Msc thesis. Department of Earth Sciences. Simon Fraser University, Burnaby, BC.
Liggett, J.E. and D.M. Allen. 2009. Comparing approaches for modelling spatially-distributed direct recharge in a semi-arid region (Okanagan Basin, Canada). Hydrogeology Journal. DOI:10.1007/s10040-009-0512-5.
Lindström, G., B. Johansson, M. Persson, M. Gardelin, and S. Bergstrom. 1997. Development and testing of the distributed HBV-96 hydrological model. Journal of Hydrology 201:272–288.
Lohmann, D., D.P. Lettenmaier, X. Liang, E.F. Wood, A. Boone, S. Chang, F. Chen, D. Yongjiu, C. Desborough, R.E. Dickinson, Q. Duan, M. ek, Y.M. Gusev, F. Habets, P. Irannejad, R. Koster, K.E. Mitchell, O.N. Nasonova, J. Noilhan, J. Schaake, A. Schlosser, S. Yaping, A.B. Shmakin, D. Verseghy, K. Warrach, P. Wetzel., X. Yongkand, Y. Zong-Liang, and Z. Qing-cun. 1998. The project for intercomparison of Land-surface Parameterization Schemes (PILPS) phase 2(c) Red-Arkansas River Basin Experiment: 3. Spatial and temporal analysis of water fluxes.” Global and Planetary Change 19:161–179.
Lopes, V.L and A.L. Vogi. 2008. Integrating modeling and field experiments to evaluate impacts of vegetative practices on ponderosa pine watersheds. Journal of the American Water Resources Association 44(5):1284–1294.
Loukas, A., L. Vasiliades, and N.R. Dalezios. 2002. Climate change implications on flood producing mechanisms in southern British Columbia, Canada using the CGCMA1 simulation results. Journal of Hydrology 259:163–188.
Luo, C.Q. and Y. Alila. 2006. Hydrologic recovery of the peak flow regime in two snow dominated meso-scale watersheds in the Southern Interior of British Columbia: an investigation using a conceptual hydrologic model. The University of British Columbia, Vancouver, BC.
Luo, C., Y. Alila, and S. Chatwin. 2006. A distributed hydrological model for a large mountain pine beetle attacked watershed. Proceedings of the CWRA BC Branch Conference, October 25–27, 2006, Vancouver, BC.
Maidment, D.1993. Handbook of Hydrology, McGraw-Hill, New York, NY.
Maloney, D., B. Bennett, A. de Groot, and A. Banner. 2002. Canopy interception in a hypermaritime forest on the north coast of British Columbia. In Extension Note No. 49. Forest Sciences Prince Rupert Forest Region, Prince Rupert, BC.
52
Markstrom, S.L., R.G. Niswonger, R.S. Regan, D.E. Prudic, and P.M. Barlow. 2008. GSFLOW Coupled ground water and surface water flow model based on the integration of the Precipitation-Runoff Modelling System (PRMS) and the Modular Ground Water Flow Model (MODFLOW-2005). U.S. Geological Survey Techniques and Methods, 6-D1.
Mattheussen, B., R.L. Kirschbaum, I.A. Goodman, G.M. O’Donnell, and D.P. Lettenmaier. 2000. Effects of land cover change on streamflow in the Interior Columbia River basin (USA and Canada), Hydrological Processes 14:867–885.
Merritt, W.S., Y. Alila, M. Barton, B. Taylor, and S. Cohen. 2006. Hydrologic response to scenarios of climate change in sub watersheds of the Okanagan basin, British Columbia. Journal of Hydrology 326:79–108.
Micovic, Z. 2006. Application of GIS in the determination of probable maximum flood. BC Hydro, Burnaby, BC.
Micovic, Z. and M.C. Quick. 1999. A rainfall and snowmelt runoff modelling approach to flow estimation at ungauged sites in British Columbia. Journal of Hydrology 226:101–120.
Mirus, B.B., B.A. Ebel, K. Loague, and B.C. Wemple. 2007. Simulated effect of a forest road on near-surface hydrologic response: Redux. Earth Surface Processes and Landforms 32:126–142.
Monteith, J.L and M.H. Unsworth. 1990. Principles of environmental physics. Edward Arnold, London, UK.
Moore, R.D. 1993. Application of a conceptual streamflow model in a glacierized drainage basin. Journal of Hydrology 150:151–168.
Moore, R.D. and S.M. Wondzell. 2005. Physical hydrology and the effects of forest harvesting in the Pacific Northwest: A review. Journal of The American Water Resources Association 41(4):763 –784.
Moore, R.D., D.G. Hutchinson, and M. Weiler. 2007. Predicting the effects of post-mpb salvage harvesting using a conceptual streamflow model (HBV-EC): Initial evaluation using a paired-catchment approach. In Mountain pine beetle and watershed hydrology workshop: Preliminary results of research from BC, Alberta and Colorado, July 10, 2007, Kelowna, BC.
Moore, R.D., D.L. Spittlehouse, P. Whitfield, and K. Stahl. 2008. Chapter 3 – Weather and climate [draft]. In Compendium of forest hydrology and geomorphology in British Columbia. R.G. Pike et al. (editors). B.C. Ministry of Forests and Range, Research Branch, Victoria, B.C. and FORREX Forest Research Extension Partnership, Kamloops, B.C. Land Management Handbook. In prep. www.forrex.org/program/water/compendium.asp
Moore, R.D., S.W. Fleming, B. Menounos, R. Wheate, A. Fountain, K. Stahl, K. Holm, and M. Jakob. 2009. Glacier change in western North America: Implications for hydrology, geomorphic hazards and water quality. Hydrological Processes 23:42–61.
Mote, P.W. and E.P. Salathé, Jr. 2009. Future climate in the Pacific Northwest. Joint Institute for the Study of Atmosphere and Ocean, Climate Impacts Group, University of Washington, Seattle, WA.
Murdock, T.Q. 2009. Research plan for regional climate impacts. Program definition: 2009–2011. Pacific Climate Impacts Consortium, Victoria, BC.
Neitsch, S.L., J.G. Arnold, J.R. Kiniry, and J.R. Williams. 2005. Soil and water assessment tool theoretical documentation, Version 2005. Grassland, Soil, and Water Research Laboratory, Agriculture Research Service, Temple, TX.
Nijssen, B., D.P. Lettenmaier, X. Liang, S.W. Wetzel, and E.F. Wood. 1997. Streamflow simulation for continental-scale river basins. Water Resources Research 33(4):711–724.
Nijssen, B., G.M. O’Donnell, A.F. Hamlet, and D.P. Lettenmaier. 2001. Hydrologic sensitivities of global ricers to climate change. Climatic Change 50:143–175.
53
Nijssen, B., I. Haddeland, and D.P. Lettenmaier. 1997. Point evaluation of a surface hydrology model for BOREAS. Journal of Geophysical Research 102:29,367–29,378.
Oja, T., Y. Xiwei, and P.A. Arp. 1995. The forest modelling series ForM-S: Applications to the Soiling spruce site. Ecological Modelling 83:207–217.
Payton, E.A. and C.M. Brendecke. 1985. Rainfall and snowmelt frequency in an alpine watershed. Paper presented at the 53rd Western Snow Conference, April 16–18, Boulder, CO.
Pike, R.G., D.L. Spittlehouse, K.E. Bennett, V.N. Egginton, P.J. Tschaplinski, T.Q. Murdock, and A.T. Werner. 2008a. Climate change and watershed hydrology: Part I – Recent and projected changes in British Columbia. Streamline Watershed Management Bulletin 11(2).
__________ 2008b. Climate change and watershed hydrology: Part II – Hydrologic implications for British Columbia. Streamline Watershed Management Bulletin 11(2).
Pike, R.G., T. Redding, D. Wilford, R.D. Moore, G. Ice, M. Reiter, and D.A.A. Toews. 2007. Chapter 16 – Detecting and predicting changes in watersheds. In Compendium of forest hydrology and geomorphology in British Columbia, R.G. Pike et al. (editors). B.C. Ministry of Forests and Range Research Branch, Victoria, B.C. and FORREX Forest Research Extension Partnership, Kamloops, B.C. Land Management Handbook (TBD). www.forrex.org/program/water/compendium.asp
Pike, R. 1995. Current limitations of hydrologic modeling in B.C.: An examination of the HSPF, TOPMODEL, UBCWM, and DHSVM hydrologic simulation models, B.C. data resources and hydrologic–wildfire impact modeling. MSc. thesis. University of Victoria, Victoria, BC.
Pike, R.G. 2003. Approaching watershed modelling. Streamline Watershed Management Bulletin 7(2). Pomeroy, J.W. and L. Li. 2000. Prairie and Arctic areal snow cover mass balance using a blowing snow model. Journal of Geophysical Research 105(D21): 26,619–26,634.
Pike, R.G. and R. Scherer. 2003. Overview of the potential effects of forest management on low flows in snowmelt-dominated hydrologic regimes. BC Journal of Ecosystems and Management 3(1):1–17.
Pomeroy, J. W. and L. Li (2000), Prairie and arctic areal snow cover mass balance using a blowing snow model. Journal of Geophysical Research 105(D21): 26,619–26,634.
Pomeroy, J.W., D.M. Gray, T. Brown, N.R. Hedstrom, W.L. Quinton, R.J. Granger, and S.K. Carey. 2007. The Cold Regions hydrological model: A platform for basing process representation and model structure on physical evidence. Hydrological Processes 21.
Power, J.M. 1985. Techniques for prediction of runoff from glacierized areas. IAHS Publication No. 149: 59–71.
Quick, M.C. and A. Pipes. 1977. UBC Watershed Model. Hydrological Sciences Bulletin XXI(1,3): 258–295.
Quick, M.C., A. Pipes, D. Nixon, E. Yu, A. Loukas, R. Millar, H. Assaf, and B. Start. 1995. U.B.C. Watershed Model manual, Version 4.0. Mountain Hydrology Group, Department of Civil Engineering, University of British Columbia, Vancouver, BC.
Quinton, W.L. and D.M. Gray. 2001. Estimating subsurface drainage from organic-covered hillslopes underlain by permafrost: Toward a combined heat mass flux model. Sixth IAHS Scientific Assembly Symposium, Maastricht, The Netherlands.
Rango, A., and J. Martinec. 1995. Revisiting the degree-day method for snowmelt computations. Journal of the American Water Resources Association 31:657–669.
Reed, S., V. Koren, M. Smith, Z. Zhang, F. Moreda, and D.-J. Seo. 2004. Overall distributed model intercomparison project results. Journal of Hydrology 298(1–4):27–60.
54
Refsgaard, J.C. 1997. Parameterisation, calibration and validation of distributed hydrologic models. Journal of Hydrology 198:69–97.
Rodenhuis, D., K.E. Bennett, A. Werner, T.Q. Murdock, and D. Bronaugh. 2007. Hydro-climatology and future climate impacts in British Columbia. Pacific Climate Impacts Consortium. www.pacificclimate.org/publications/PCIC.ClimateOverview.pdf
Rothwell, R.L. and R.H. Swanson. 2007. Hydrologic effects of mountain pine beetle infestations in western Alberta. Mountain pine beetle and watershed hydrology workshop: Preliminary results of research from BC, Alberta and Colorado, Kelowna, BC.
Runni ng, S. and E. Hunt. 1993. Generalization of a forest ecosystem process model for other biomes, BIOME-BGC and an application for global-scale models. Scaling Physiological Processes: Leaf to Globe, J. R. Ehleringer and C. Field (editors). Academic Press 141–158.
Running, S., R. Nemani, and R. Hungerfored. 1987. Extrapolation of synoptic meteorological data in mountainous terrain and its use for simulating forest evapotranspiration and photosynthesis. Canadian Journal of Forest Research (17)472–483.
Salathé, E.P. Jr., 2005. Downscaling simulations of future global climate with application to hydrologic modeling. International Journal of Climatology, 25:419–436.
Sanford, S.E., I.F. Creed, C.L. Tague, F.D. Beall, and J.M. Buttle. 2007. Scale-dependence of natural variability of flow regimes in a forested landscape. Water Resources Research 43 W08414, DOI:10.1029/2006WR005299.
Sauchyn, D. and S. Kulshreshtha. 2008. Prairies. In From impacts to adaptation: Canada in a changing climate. D.S. Lemmen, F.J. Warren, J. Lacroix and E. Bush (editors). Government of Canada, Ottawa, ON.
Savenije, H.H.G. 2009. HESS Opinions: The art of hydrology. Hydrololgy and Earth System Sciences 13:157–161.
Scherer, R. and R. Pike. 2003. Effects of forest management activities on streamflow in the Okanagan Basin, British Columbia: Outcomes of a literature review and a workshop. FORREX Series 9. FORREX Research and Extension Partnership, Kamloops, BC
Schnorbus, M.A. and Y. Alila. 2004a: Forest harvesting impacts on the peak flow regime in the Columbia Mountains of southeastern British Columbia: An investigation using long-term numerical modeling. Water Resources Research 40, W05205, DOI:10.1029/2003WR002918.
__________. 2004b: Generation of an hourly meteorological time series for an alpine basin in British Columbia for use in numerical hydrologic modeling. Journal of Hydrometeorology 5:862–888.
Schroeder, P.R. 1996. The HELP program: The military engineer.Society of American Military Engineers 88(579):28–29.
Schroeder, P.R., T.S. Dozier, P.A. Zappi, B.M. McEnroe, J.W. Sjostrom, and R.L. Peyton. 1994. The hydrologic evaluation of landfill performance (HELP) model: Engineering documentation for Version 3, EPA/600/R-94/168b, September 1994. U.S. Environmental Protection Agency Office of Research and Development, Washington, DC.
Schulla, J. and K. Jasper. 2000. Model description WaSiM-ETH (Water balance simulation model ETH), Institute of Geography ETH CH 8049 Zürich ETH, Switzerland.
Schulze, R.E. 1995. Hydrology and agrohydrology, a text to accompany the ACRU agrohydrological modelling system.Water Resource Commission. WRC Report No. TT69/95.
Scibek, J. and D.M. Allen. 2006a. Comparing modelled responses of two high-permeability unconfined aquifers to predicted climate change.” Global and Planetary Change 50:50–62.
55
Scibek, J. and D.M. Allen. 2006b. Modeled impacts of predicted climate change on recharge and groundwater levels. Water Resources Research 42: W11405, DOI:10.1029/2005WR004742.
Seibert, J. 1997. Estimation of parameter uncertainty in the HBV Model. Nordic Hydrology, 28(4/5):247–262.
Seibert, J. and J. McDonnell. 2002. On the dialog between experimentalist and modeler in catchment hydrology: Use of soft data for multicriteria model calibration.” WRR 38(11): DOI:10.1029/2001WR000978.
Shuttleworth, W.J. and J.S. Wallace. 1985. Evaporation from sparse crops—an energy combination theory. Quarterly J Royal Meteorological Society 111:839–855.
Silins, U. 2002. Streamflow models in integrated forest-watershed planning: How much complexity is warranted? 2002 Interior Watershed Conference, March 13–14, 2002, Kamloops, BC.
__________. 2003. An integrated forest-watershed planning & assessment model: ECA-Alberta. University of Alberta, Centre for Enhanced Forest Management, EFM Research Note 07/2003. www.cefm.rr.ualberta.ca/Research_Notes/Research%20Note%2007-03.pdf
Smerdon, B. and T. Redding. 2007. Groundwater:More than water below the ground! Streamline Watershed Management Bulletin.10(2):1-6.
Smerdon, B.D., C.A. Mendoza, and K.J. Devito. 2007. Simulations of fully coupled lake-groundwater exchange in a sub-humid climate with an integrated hydrologic model. Water Resources Research 43: W01416, DOI:10.1029/2006WR005137.
___________. 2008. The influence of subhumid climate and water table depth on groundwater recharge in shallow outwash aquifers. Water Resources Research. In press. DOI:10.1029/2007WR005950.
Smerdon, B.D., D.M. Allen, S.E. Grasby, and M.A. Berg. 2009. An approach for predicting groundwater recharge in mountainous watersheds. Journal of Hydrology 365(3–4):156–172.
Smith, M.B., D.-J Seo, V.I. Koren, S. Reed, Z. Zhang, Q.-Y Duan, S. Cong, F. Moreda, and R. Anderson. 2004. The Distributed model intercomparison project (DMIP): Motivation and experiment design. Journal of Hydrology 298(1–4):4–26.
Smithers, J. and R.E. Schulze. 1995. ACRU agrohydrological modelling system user manual version 3.00. Water Resarch Commission, Pretoria, RSA Report TT70/95.
Soulis, E.D., K.R. Snelgrove, N. Kouwen, F. Seglenieks, and D.L. Verseghy. 2000. Toward closing the vertical water balance in canadian atmospheric models: Coupling of the land surface 2317 Scheme CLASS with the distributed hydrological model WATFLOOD. Atmosphere-Ocean 38:251–269.
Spittlehouse, D.L. 2006. Your access to interpolated climate data for BC. Streamline Watershed Management Bulletin 9:16–21.
Spittlehouse, D.L. and R.D. Winkler. 2004. Snowmelt in a forest and clearcut. In Proceedings 72nd Western Snow Conference, Vancouver, BC.
Sridhar, V., A.L. Sansone, J. LaMarche, T. Dubin, and D.P. Lettenmaier. 2004. Prediction of stream temperature in forested watersheds. Journal of the American Water Resources Association 40(1): 197, DOI:10.1111/j.1752-1688.2004.tb01019.x.
Stadnyk, T., N.A. St. Amour, N. Kouwen, T.W.D. Edwards, A. Pietroniro, and J. J. Gibson. 2005. A groundwater separation study in boreal wetland terrain: The WATFLOOD hydrological model compared with stable isotope tracers. Isotopes in Environmental and Health Studies 41(1):49–68.
Stahl, K., R.D. Moore, J.M. Shea, D. Hutchinson, and A.J. Cannon. 2008. Coupled modelling of glacier and streamflow response to future climate scenarios. Water Resources Research 44:W02422, DOI:10.1029/2007WR005956.
56
Storck, P. 2000. Trees, snow and flooding: An investigation of forest canopy effects on snow accumulation and melt at the plot and watershed scales in the Pacific Northwest, Department of Civil Engineering. University of Washington, Seattle, WA.
Storck, P., L. Bowling, P. Wetherbee, and D. Lettenmaier.1998. Application of a GIS-based distributed hydrology model for prediction of forest harvest effects on peak stream flow in the Pacific Northwest. Hydrololgical Processes 12:889–904.
Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology, Transactions American Geophysical Union 38(6):913–920
Surfleet, C.G. 2008. Uncertainty in forest road hydrologic modeling and catchment scale assessment of forest road sediment yield. PhD. Dissertation. Department of Forest Engineering, State University, OR.
Swanson, R.H. 1994. Using the Wrnshyd Procedure to estimate long-term cumulative effects of aspen clearcutting on water yield in Alberta. Presented at Western Snow Conference 1994, Sante Fe, NM.
Swanson, R.H. and R.L. Rothwell. 2001. Hydrologic recovery of aspen clearcuts in Northwestern Alberta. USDA Forest Service Proceedings RMRS-P-18.
Swanson, R.H. 2005. Using the WinWrnsHyd program as a didactic tool. Paper presented Western Snow Conference 2005.
Tague, C. and L.E. Band. 2004. RHESSys: regional hydro-ecologic simulation system: an object-oriented approach to spatially distributed modeling of carbon, water and nutrient cycling. Earth Interact 8(19): 1–42.
___________. 2001a. Evaluating explicit and implicit routing for watershed hydro-ecological models of forest hydrology at the small catchment scale. Hydrological Processes 15:1414–1439.
___________. 2001b. Simulating the impact of road construction and forest harvesting on hydrologic response Earth Surface Processes And Landforms 26(2):135–151.
Tague, C., C. McMichael, A. Hope, J. Choate, and R. Clark. 2004. Application of the RHESSys model to a California semiarid shrubland watershed. Journal of The American Water Resources Association 40:575–589.
Tague, C., G. Grant, M. Farrell, J. Choate, and A. Jefferson. 2008. Deep groundwater mediates streamflow response to climate warming in the Oregon Cascades. Climatic Change 86:189–210 DOI:10.1007/s10584-007-9294-8.
Therrien, R., E.A. Sudicky, R.G. McLaren, and S.M. Panday. 2008. HydroGeoSphere, A three-dimensional numerical model describing fully-integrated subsurface and surface flow and solute transport. University of Waterloo Groundwater Simulations Group, DRAFT.
Thyer, M., J. Beckers, D. Spittlehouse, Y. Alila, and R. Winkler. 2004. Diagnosing a distributed hydrologic model for two high-elevation forested catchments based on detailed stand- and basin-scale data, Water Resources Research 40, W01103, DOI:10.1029/2003WR002414.
Toews, M.W. 2007. Modelling climate change impacts on groundwater recharge in a semi-arid region, Southern Okanagan, BC. MSc Thesis. Simon Fraser University, BC.
Toth, B., A. Pietroniro, F.M. Conly, and N. Kouwen. 2006. Modelling Climate change impacts in the Peace and Athabasca Catchment and Delta: I – hydrological model application. Hydrological Processes 20:4197–4214.
Troendle, C.A. and C.F. Leaf. 1980. Hydrology. In An approach to water resource evaluation of non-pointsilvicultural sources. EPA 60018-90-12. Environmental Research Laboratory, Athens, GA.
Tysdal, L., W.J. Elliot, C. Luce, and T. Black. 1997. Modeling erosion from insloping roads with the WEPP watershed model. ASAE Pap. 97-5014. American Society of Agriculture Engineers, St. Joseph, MI.
57
US Army Corps of Engineers. 1991. Streamflow Synthesis and Reservoir Regulation Model (SSARR), user manual, CPD-74B, North Pacific Division, Portland, OR.
__________. 2000. Hydrologic Modeling System HEC-HMS, technical reference manual, CPD-74B, Hydrologic Engineering Centre, Davis, CA.
US Environmental Protection Agency. 1980. An approach to water resources evaluation on non-point silvicultural sources (a procedural handbook). United States Environmental Protection Agency, Environmental Research laboratory, Athens, GA. EPA-600/8-80-012.
Valeo, C. and Z. Xiang. 2005. Climate change impacts on flooding in the Elbow River watershed. Geophysical Research Abstracts 7(04910).
VanShaar, J.R., I. Haddeland, and D.P. Lettenmaier. 2002. Effects of land cover changes on the hydrologic response of interior Columbia River Basin forested catchments. Hydrological Processes 16:2499–2520.
Verbunt, M., J. Gurtz, K. Jaspera, H. Langa, P. Warmerdamb, and M. Zappa. 2003. The hydrological role of snow and glaciers in alpine river basins and their distributed modeling. Journal of Hydrology 282: 36–55.
Verbunt, M., M. Groot Zwaaftink, and J. Gurtz. 2005. The hydrologic impact of land cover changes and hydropower stations in the Alpine Rhine basin. Ecological Modelling 187(1):71–84.
Verseghy, D.L. 1991. CLASS—a Canadian land surface scheme for GCMs, I. Soil model. International Journal of Climatology 11:111–133.
Verseghy, D.L., N.A. McFarlane, and M. Lazare. 1993. CLASS—a Canadian land surface scheme for GCMs, II. Vegetation model and coupled runs. International Journal of Climatology 13:347–370.
Viviroli, D., Zappa, M., Gurtz, J. and Weingartner, R. 2009. An introduction to the hydrological modelling system PREVAH and its pre- and post-processing tools. Environmental Modelling and Software 24:1209–1222.
Waichler, S.R. and M.S. Wigmosta. 2003. Simulation of hourly meteorology from daily data and significance to hydrology at H.J. Andrews Experimental Forest. Journal of Hydrometeorology 4(2):251–263.
Waichler, S.R., B.C. Wemple, and M.S. Wigmosta. 2005. Simulation of water balance and forest treatment effects at the H.J. Andrews Experimental Forest. Hydrological Processes 19:3177–3199.
Walker, I.J. and R. Sydneysmith. 2008. British Columbia. In From impacts to adaptation: Canada in a changing climate. D.S. Lemmen, F.J.Warren, J. Lacroix, and E. Bush (editors). Government of Canada, Ottawa, ON.
Walter, M.T., E.S. Brooks, D.K. McCool, L.G. King, M. Molnaub, and J. Boll. 2005. Process-based snowmelt modeling: Does it require more input data than temperature-index modeling? Journal of Hydrology 300:65–75.
Wang, T.L., A. Hamman, D.L. Spittlehouse, and S.N. Aitken. 2006. Development of scale-free climate data for western Canada for use in resource management. International Journal of Climatology 26:383–397.
Ward, R.C. 1971. Small experimental watersheds: An appraisal of concepts and research developments. University of Hull, Hull Printers Ltd. Hull, UK.
Webb, B.W., D.M. Hannah, R.D. Moore, L.E. Brown, and F. Nobilis. 2008. Recent advances in stream and river temperature research. Hydrological Processes 22:902–918, DOI:10.1002/hyp.6994.
Weiler, M., K. Rosin, and C. Scheffler. 2008. Large scale assessment of land cover changes on peak flow in ungauged watersheds. In prep.
58
Werner, A.D., M.R. Gallagher, and S.W. Weeks. 2006. Regional-scale, fully coupled modelling of stream–aquifer interaction in a tropical catchment. Journal of Hydrology 328:497–510.
Werner, A.T. and K.E. Bennett. 2009. A review of hydrologic models for application in climate change studies for watersheds in the cold mountainous terrain of western North America.” PCIC Report, January 15th, 2009, Pacific Climate Impacts Consortium, University of Victoria, Victoria, BC.
Westrick, K.J., P. Storck, and C.F. Mass. 2002: Description and evaluation of a hydrometeorological forecast system for mountainous watersheds. Weather and Forecasting 17:250–262.
Whitaker, A., Y. Alila, J. Beckers, and D. Toews. 2002. Evaluating peak flow sensitivity to clearcutting in different elevation bands of a snowmelt-dominated mountainous catchment. Water Resources Research 38(9): 1172. DOI:10.1029/2001WR000514.
__________. 2003. Application of the distributed hydrology soil vegetation model to Redfish Creek, British Columbia: Model evaluation using internal catchment data. Hydrological Processes 17:199–224.
Whitaker A., Y. Alila, P. Calvert, D. Toews. 1998. Evaluation of existing hydrologic models for use in assessing the impact of forest management on peak flows. In Proceedings of the Canadian Water Resources Association 51st Annual Conference: Mountains to Sea: Human Interaction with the Hydrologic Cycle, June 10–12,1998, Victoria, BC.
Whitfield, P.H., C.J. Reynolds, and A.J. Cannon. 2002. Modelling streamflows in present and future climates—examples from Georgia Basin, British Columbia. Canadian Water Resources Journal 27(4):427–456.
Whyte, D.M. 2004. Hydrologic response during snowmelt in three steep headwater catchments: Ringrose Slope, Slocan Valley, British Columbia. Unpublished MSc. thesis. The University of British Columbia, Vancouver, BC.
Whyte, D.M., R.D. Moore, A.S. Dhakal, and D. Toews. 2004. Snowmelt runoff processes and modelling for three small catchments draining a glaciated valley wall (or, When good models go bad). Presented at the Annual Meeting of the Western Snow Conference, Richmond, BC.
Widmann, M., C.S. Bretherton, and E.P. Salathé. 2003. Statistical precipitation downscaling over the northwestern United States using numerically simulated precipitation as a predictor. Journal of Climate 16:799–816.
Wigmosta, M.S., B. Nijssen, and P. Storck. 2002. The distributed hydrology soil vegetation model. In mathematical models of small watershed hydrology and applications V. P. Singh and D. K. Frevert (editors) Water Resources Publications Highlands Ranch, CO.
Wigmosta, M.S., L.W. Vail, and D.P. Lettenmaier. 1994. A distributed hydrology-vegetation model for complex terrain. Water Resources Research 30(6):1665–1679.
Winkler, R. 2001. The effects of forest structure on snow accumulation and melt in south-central British Columbia, PhD thesis. University of British Columbia, Vancouver, BC. www.for.gov.bc.ca/hfd/pubs/Docs/Mr/Mr114.htm
Wood, A.W., E.P. Maurer, A. Kumar, and D.P. Lettenmaier. 2002. Long-range experimental hydrologic forecasting for the eastern United States. Journal of Geophysical Research-Atmospheres 107:4429–4443.
Wood, A.W., L.R. Leung, V. Sridhar, and D.P. Lettenmaier. 2004. Hydrologic implications of dynamical and statistical approaches to downscaling climate model outputs. Climate Change 62(1–3):189–216.
Wu, K. and C.A. Johnston. 2008. Hydrologic comparison between a forested and a wetland/lake dominated watershed using SWAT. Hydrological Processes 22:1431–1442.
Wu, W. and R.C. Sidle. 1995. A distributed slope stability model for steep forested hillslopes. Water Resources Research 31(8):2097–2110.
59
Yin, X. and P.A. Arp. 1993. Predicting forest soil temperatures from monthly air temperature and precipitation records. Canadian Journal of Forest Research 23:2521–2536.
Zappa, M., F. Pos, U. Strasser, P. Warmerdam, and J. Gurtz. 2003. Seasonal water balance of an alpine catchment as evaluated by different methods for spatially distributed snowmelt modelling. Nordic Hydrology 34(3):179–202.
Ziemer, R.R., J. Lewis, R.M. Rice, and T.E. Lisle. 1991. Modelling the cumulative watershed effects of forest management strategies. Journal of Environmental Quality 20:36–42.
Zierl, B., H. Bugmann, and C.L. Tague. 2007. Water and carbon fluxes of European ecosystems: an evaluation of the ecohydrological model RHESSys. Hydrological Processes 21(24):3328–3339.
60
APPendix 1 model review And rAnking
The sections below describe the models considered and reviews examples of how these models were applied, with an emphasis on applications in the Pacific Northwest. Other regions with similar biogeocli-matic settings were considered for those models that could not be applied to the Pacific Northwest.
The model descriptions are supplemented by Tables 4 to 8, which form the basis for supporting deci-sions on model selection (Section 3). Table 4 assesses model functionality (as defined in Section 2.3.1 and Table 2) and offers a tool for screening what processes a model simulates. Table 5 ranks the models according to perceived overall complexity (as defined in Section 2.3.2 and Table 3). Table 6 summar-izes model applicability for simulating the effects of forest management. Table 7 provides an overviews the climatic and physiographic settings in which the models are best applied (see Section 2.3.3). Table 8 lists model outputs for forest planning (based on a review of model functionality; Table 4), including planning scale (determined by watershed discretization; Table 6) and temporal scale of model output (determined by model time step). Referencing to tables in the descriptions below serves to document the basis for these tables. Primary source materials are also referenced for each model. Materials for review of the PREVAH, VIC, WaSiM-ETH, and Watflood models were in part provided by PCIC (Werner and Bennett 2009).
The model descriptions are provided in alphabetical order.
AcruA1.1
The ACRU model3 is a multi-purpose conceptual physical model (Schulze 1995). The acronym ACRU is derived from the Agricultural Catchments Research Unit within the Department of Agricultural Engi-neering of the University of Natal in Pietermaritzburg, South Africa (UNP).
The ACRU model “integrates the various water budgeting and runoff producing components of the terrestrial hydrological system with risk analysis” (Schulze 1995). The ACRU model can be used for hy-drological design purposes, reservoir simulation, and water resource assessment for agriculture, as well as other water resource assessments.
Model DescriptionA1.1.1
The ACRU model operates as either a point or lumped small watershed model. The model can func-tion as a distributed model for large watersheds, with sub-watersheds (ideally less than 30 km2) (Schulze 1995). The model includes multiple soil horizons and a single vegetation layer and operates at a daily time step (Tables 4, 6, and 8).
Evapotranspiration can be simulated using methods that range from empirical (reference evaporation based on pan evaporation data) to analytical (a simplication of the Penman equation in which energy components are related to temperature variables) (Chapter 4 in Schulze 1995; Table 4). The original model (as available from UNP) does not include snow processes. However, the ACRU model was adapt-ed at the University of Lethbridge (UL) to include canopy snow interception and a degree day-based snowmelt method (Dr. S. Kienzle4, pers. comm., Dec. 2009) with an empirical correction for rain-on-snow events based on Kuusisto (1980). A GIS-based method was developed to adjust daily temperatures based on the ratio between slope radiation and flat radiation to account for significant differences in temperatures on, for example, north- and south-facing slopes. Therefore, the degree-day model will result in different snowmelt rates on south- and north-facing slopes. Overall, ACRU ranks as a mixed model for simulating SVAT processes (Table 6).
3 www.beeh.unp.ac.za/acru/4 http://people.uleth.ca/~stefan.kienzle/
61
Infiltration, percolation, and runoff components in the ACRU model are empirical (Table 4). Rain-fall not lost through interception or routed as stormflow (either rapid response or delayed), enters the soil through the surface layer and is stored in the topsoil horizon. Downward drainage to the subsoil horizon(s) and into intermediate and groundwater stores is based on field capacity. Runoff components in the ACRU model are empirical. “Generated streamflow comprises baseflow and stormflow, with the stormflow component consisting of a quickflow response, (i.e., stormflow released into the stream on the same day as the rainfall event), and a delayed stormflow response, (i.e., a surrogate for post-storm in-terflow)” (Schulze 1995). Baseflow is derived from the intermediate and so-called “groundwater” stores, and a number of routing methods are available, including the Muskingum method. Road hydrology is not included in the ACRU model (Table 6).
The ACRU model (Tables 4, 6, and 7) includes a crop yield model (not a growth model), temporal input control (i.e., allowing for gradual changes in forest cover characteristics over time), and allows for the simulation of irrigation, lakes, wetland hydrology, reservoirs, stream water abstraction, nutrient loading, and sediment yield.
Climate input requirements are daily precipitation and minimum and maximum temperature. The ACRU model has been designed “as a multi-level model, with either multiple options or alternative pathways (or a hierarchy of pathways) available in many of its routines, depending on the level of input data available or the detail of output required” (Schulze 1995). ACRU operates in conjunction with the ACRU Utilities software, which assists with the preparation of input and output data. A further utility, the ACRU Menubuilder, prompts the user with unambiguous questions to assist in simplifying the entry of complex distributed watershed information (Smithers and Schulze 1995). Overall the ACRU model appears to be of relatively high complexity (Table 5) and requires substantial data collection, pre-pro-cessing, and GIS analysis (Dr. S. Kienzle, pers. comm., Dec 2009).
The South African developers charge a fee for model use to compensate for the many years it took to develop (a few 100 US dollars). Dr. Kienzle is willing to provide training for a fee. Original training, without the snow modelling, can be provided in South Africa (Pietermaritzburg).
The model was originally developed for South African conditions. The snow adaptations have not been included in the UNP version of the ACRU model. With the addition of the snow routines, the model can be applied in rain, snow, and mixed regimes. To date, the model has been applied in small to medium watersheds in Alberta, in both steeply sloped and more gradual terrain (Table 7).
Model Applications A1.1.2
The following ACRU published model applications were identified (Dr. S. Kienzle, pers. comm., Dec. 2009):
St. Mary Watershed, Montana-Alberta Rocky Mountain region•Beaver Creek, Porcupine Hills, southwest Alberta•Upper North Saskatchewan watershed, upstream of Edmonton (ongoing work)•
Unfortunately, peer-reviewed articles on these model applications were lacking at the time of model review. Outside the PNW, the ACRU model has also been applied in similar settings in New Zealand.
62
brook90A1.2
The BROOK905 model was developed to “simulate daily evaporation and soil-water movement using a process-oriented approach for a single forest stand/site or for a small watershed, with some provision for runoff (streamflow) generation by different flow paths” (Federer et al. 2003).
Model DescriptionA1.2.1
BROOK90 is a lumped parameter hydrologic model that incorporates multiple soil layers and a single vegetation layer (i.e., overstorey and understorey are not distinguished). The model is basically a stand- or site-level water balance model, and operates at a daily time step (Tables 4, 6, and 8).
Evaporation is modelled as the sum of five components (Federer et al. 2003): evaporation of inter-cepted rain and snow, evaporation of snow on the ground, soil evaporation, and transpiration. The model uses the physical Shuttleworth and Wallace (1985) method for separating transpiration and soil evaporation from sparse canopies, and for simulating evaporation of interception (Table 4). Snow accumulation and melt are controlled by a degree-day method with no apparent correction for rain-on-snow events. Overall, BROOK90 ranks as a mixed model for simulating SVAT processes (Table 6).
Infiltration, percolation, and runoff components in the BROOK90 model are empirical (Table 4). Soil water movement occurs vertically, according to Darcy’s Law for unsaturated or saturated flow. Streamflow (runoff) is generated by four simplified processes: stormflow by source area flow, subsur-face pipe-flow (macropores), delayed flow from vertical or downslope soil drainage, and a first-order groundwater storage. The model does not include a channel-routing component and does not account for road hydrology (Table 6). The BROOK90 model can simulate preferential flow (Table 4).
Required daily meteorological inputs are daily precipitation and maximum and minimum temper-atures, while daily solar radiation, vapour pressure, and wind speed are desirable. Overall, the model appears to be of medium complexity (Table 5). BROOK90 has an interactive user interface that allows parameter changes and output variable selection. This program is distributed as freeware.
The main model limitation appears to be lack of a channel-routing routine, which limits model application to small watersheds with no sub-watersheds. Other limitations include the inability to incorporate field capacity for soil water drainage and a lack of an upper limit for water availability for evapotranspiration. The functionality of BROOK90 in steep mountainous settings may be limited. The model can be applied in rain and snow regimes but is not recommended for mixed regimes where rain-on-snow events are significant (Table 7).
Model Applications A1.2.2
The following BROOK90 published model applications were reviewed:
• “HydrologicrecoveryofaspenclearcutsinnorthwesternAlberta”(SwansonandRothwell2001).Several deciduous forest sites in Northwestern Alberta ranging from 20 to 40 ha in size were modelled over 60 years to determine the change in generated runoff from original harvest. The objective of the study was to identify the period during which aspen clearcuts could significantly affect water yield and the potential for flooding.
• DetectinghydrologicdisturbancefromharvestBeaverCreekwatersheds,USDAExperimentalWatershed Area, Arizona (Lopes and Vogi 2008). WS13 (control), WS12 (clearcut), and WS14
5 http://home.roadrunner.com/~stfederer/brook/compassb.htm
63
(stripcut) watersheds were set up as a paired watershed experiment. BROOK90 was used to simu-late water yield changes for pre- and post-harvest periods, and compared to field data.
• SilvicultureandwaterflowdynamicsRotherbachandSchluchseeCatchments,Germany(Arm-buster et al. 2004). The Rotherdbach watershed (9.4 ha, 93-year-old Norway spruce) is situated in the Eastern Ore Mountains and the Schluchsee watershed (11 ha, 55-year-old Norway spruce) is located in the higher altitudes of the Black Forest. BROOK90 was used to evaluate the potential hydrological effects of the silvicultural conversion from monocultural conifer stands into mixed or pure deciduous stands. Simulation timeframe was 1987 to 1998.
The first study indicated that the BROOK90 model was successful at simulating generated runoff increases between clearcut and mature forests in northwestern Alberta, and also provided estimates for how long it takes until the decrease in evapotranspiration from regenerating aspen clearcuts becomes insignificant. The second study showed BROOK90 to be capable of replicating the temporal variation of streamflow and portions of the individual flow components (e.g., evapotranspiration and snow water equivalent) but also indicated that the model is designed for small, uniform watersheds. There is no provision for spatial distribution of parameters. In the third study, simulated water yields for the pre-treatment period were within the range of observed water yields and were similar to those obtained with the paired-watershed approach. The model also replicated field observations for the post-harvest per-iod. However, the model tended to overestimate water yields during years of low flows and there was a significant difference between the two approaches in estimating water yield changes during the first four years after treatment for both treated watersheds. This indicates that the model approach may be limited in its ability to detect small changes because it tends to overestimate changes that occur immediately fol-lowing treatment.
The Arizona study further showed the BROOK90 model to be designed for single-species stands, so transition to mixed stands is questionable. This is a limitation that is common to almost all models (“complex stand” column in Table 4).
crhmA1.3
The Cold Region Hydrology Model (CRHM)6, developed at the University of Saskatchewan (Pomeroy et al. 2007), is based on integrated field measurements and modelling research. A number of cold-regions hydrological processes are simulated using physically based equations. Processes simulated include: snow redistribution by wind, snow interception, sublimation, snowmelt, infiltration into frozen soils, hillslope water movement over permafrost, actual evaporation, and radiation exchange to complex surfaces.
Model DescriptionA1.3.1
CRHM uses a Hydrologic Response Unit (HRU) architecture for discretizing a watershed (Table 6), a two-layer soil model, and a single vegetation layer (i.e., overstorey and understorey are not distin-guished). The model normally operates at an hourly time step (Tables 4, 6, and 8).
Soil evaporation, plant transpiration, and evaporation of rainfall intercepted by vegetation are simu-lated using physically based methods (Table 4). Canopy snow interception and ablation and snowmelt (energy balance approach) are also simulated using physically based methods. Overall, CRHM ranks as a physical model for simulating SVAT processes (Table 6).
6 www.usask.ca/hydrology/crhm.htm
64
Soil moisture balances are calculated using an empirical (Table 4) “multiple flow path 3-layer linear reservoir model with options for fill and spill, saturation overland flow, shallow subsurface drainage and groundwater drainage” (Pomeroy et al. 2007). The model does not include explicit channel routing, but instead calculates the independent timing and storage control of overland, interflow, groundwater flow, and streamflow using the lag-and-route hydrograph method of Clark (1945). Road hydrology is not incorporated in CRHM (Table 6).
The CRHM model (Tables 4 and 7) includes blowing snow transport and sublimation following Pomeroy and Li (2000) and a variety of infiltration routines for frozen soils (Granger et al. 1984; Gray et al. 2001), including frost depth calculation.
“CRHM uses a modular object-oriented structure to develop, support, and apply dynamic model routines” (Pomeroy et al. 2007). Within CRHM, landscape elements can be linked episodically in pro-cess-specific sequences via blowing snow transport, overland flow, organic layer subsurface flow, mineral interflow, groundwater flow, and streamflow. The CRHM has a simple user interface; however, there is no provision for calibration. Model parameters and structure are selected based on understanding the hydrological system so the model can be used both for predicting and diagnosing the adequacy of hydrological understanding (Pomeroy et al. 2007). Overall, the model appears to be of high complexity (Table 5).
The main model limitation appears to be its rudimentary streamflow-routing, routine-limiting model applicability to small to medium watersheds. Within the models, the watershed HRU subdivisions must be characterized by the model user based upon an understanding of the hydrological processes, terrain, and land use.(Pomeroy et al. 2007). This lack of an automated routine does not make it easy to apply the model. While the model has been applied in boreal forest environments associated with the BERMS and Fluxnet sites (Pomeroy et al. 2007), to date the model has only been applied in snowmelt regimes. How-ever, with physically based SVAT routines, it could also be applied to rain and mixed regimes (Table 7).
Model ApplicationsA1.3.2
The review below is limited to watershed-scale applications of CRHM. Numerous additional investiga-tions, referenced in Pomeroy et al. (2007) have focused on testing individual components of the model, such as the blowing snow model, the canopy interception model, the canopy radiation model, the snow-melt model, the infiltration model, and the organic-layer flow model.
The following CRHM published model applications were reviewed:
• “TheColdRegionsHydrologicalModel,aplatformforbasingprocessrepresentationandmodelstructure on physical evidence” (Pomeroy et al. 2007). The CRHM model was tested in a small clearcut watershed of the Prince Albert Model Forest in Saskatchewan. The model incorporated blowing snow, overland flow, subsurface flow, groundwater flow, and streamflow modules to simu-late snow accumulation, melt, and runoff in a substantially clearcut boreal forest.
• “Estimatingsubsurfacedrainagefromorganic-coveredhillslopesunderlainbypermafrost:Towarda combined heat mass flux model” (Quinton and Gray 2001). The flow modules within CRHM were developed and tested using data from Wolf Creek and Scotty Creek, two alpine watersheds in Northern Canada with little vegetation and hydrological processes driven by blowing snow, dy-namic melting, and the presence of permafrost.
65
The examples listed above are outside of the Pacific Northwest as CRHM is primarily a prairie/tundra model. However, they provide insight into how the model may be applied in the boreal forests of north-ern Alberta or BC. Hillslope-scale flow was simulated well with CRHM, and accounted for soil freezing/thawing. The model has not been applied in conditions similar to the steep, forested regions of AB and BC.
dhsvmA1.4
The Distributed Hydrology Soil Vegetation Model (DHSVM) is a physically based model developed at the University of Washington7 to represent the effects of topography and vegetation on water fluxes through the landscape (Wigmosta et al.1994; Wigmosta et al. 2002).
Model DescriptionA1.4.1
The DHSVM is a fully distributed hydrologic model that subdivides the model domain (typically the watershed or group of watersheds under investigation) into small computational grid elements using the spatial resolution of an underlying digital elevation model (DEM). It is typically applied at high spatial resolutions on the order of 10 to 100 m (Beckers and Alila 2004) for watersheds of up to 10 000 km2 and at sub-daily time scales (best results are obtained at hourly time steps) for multi-year simulations. The model includes two vegetation layers (forest canopy overstorey and understorey) and multiple soil layers (Tables 4, 6, and 8).
Soil evaporation, plant transpiration, and evaporation of rainfall intercepted by vegetation are simu-lated using physically based methods (Table 4). Canopy snow interception and ablation and snowmelt (energy balance approach) are also simulated using physically based methods. DHSVM therefore ranks as a physically based model for simulating SVAT processes (Table 6).
Vertical unsaturated water movement through the soil layers is calculated using an analytical ap-proach (Table 4) based on the one-dimensional form of Darcy’s law which recharges the grid cell water table. “Subsurface lateral flow is calculated using a transient, three-dimensional representation of satur-ated subsurface flow” (Wigmosta et al. 1994). “Return flow and saturation overland flow are generated in locations where grid cell water tables intersect the ground surface. Open-channel routing uses explicit information on the location of stream channels” (Wigmosta et al. 2002).
The DHSVM model allows the user to simulate road hydrology, including interception of surface and shallow subsurface runoff, precipitation interception, and routing of road flow by using explicit infor-mation on road network geometry and interconnectivity, and the location of stream crossing structures such as culverts (Tables 4 and 6). DHSVM has also been modified to include a sediment model with four primary components: mass wasting (stochastic), hillslope erosion, erosion from forest roads, and a channel-routing algorithm (Doten et al. 2006).
Required meteorological inputs to DHSVM include precipitation, air temperature, air humidity, shortwave and longwave radiation, and wind speed. Meteorological variables are distributed in the mod-el based on elevation, slope, and aspect, making DHSVM useful in dealing with the effects of complex terrain. DHSVM can also use gridded output from regional climate models such as MM5. Substantial calibration requirements should be expected, depending on the level of modelling detail (e.g., Whitaker et al. 2003; Thyer et al. 2004; Beckers and Alila 2004). Overall, the model is of high complexity (Table 5),
7 www.hydro.washington.edu/Lettenmaier/Models/DHSVM/index.shtml
66
requires substantial data collection, pre-processing, and GIS analysis. DHSVM is freeware. Only limited efforts have been made to incorporate a user-friendly interface. No technical support is available, unless specific arrangements have been made. In BC to date, all applications of DHSVM have been conducted through the research group headed by Dr. Y. Alila at UBC.
The lack of a groundwater component in DHSVM reflects the model’s specific development for steep-ly sloped terrain that has a thin soil veneer overlying relatively impermeable bedrock. While the model has been applied in mountain plateau settings in BC, DHSVM’s overall functionality in more gently sloped terrain and in settings with a substantial groundwater component to the hydrological budget is likely limited. To date, the model has been applied in rain, snow, and mixed regimes, and in small to medium watersheds up to about 10 000 km2 (Table 7).
Model ApplicationsA1.4.2
The DHSVM has been widely applied in the PNW in a forest hydrology and management context, for assessing the effects of forest roads on flood flows, for hydrologic forecasting, and for assessing the po-tential effects of climate change, including but not limited to the following studies:
Forest Hydrology and Management
• “PointevaluationofasurfacehydrologymodelforBOREAS”(Nijssenetal.1997).• “Hydrologicaleffectsofland-usechangeinazero-ordercatchment(Burgesetal.1998).• “ApplicationofaGIS-baseddistributedhydrologymodelforpredictionofforestharvesteffectson
peak stream flow in the Pacific Northwest” (Storck et al.1998).• “HydrologiceffectsoflogginginWesternWashington,UnitedStates”(Bowlingetal.2000).• “Trees,snowandflooding:Aninvestigationofforestcanopyeffectsonsnowaccumulationand
melt at the plot and watershed scales in the Pacific Northwest” (Storck 2000).• “EffectsoflandcoverchangesonthehydrologicresponseofinteriorColumbiaRiverBasinfor-
ested catchments” (VanShaar et al. 2002).• “Evaluatingpeakflowsensitivitytoclear-cuttingindifferentelevationbandsofasnowmelt-dom-
inated mountainous catchment” (Whitaker et al. 2002); “Application of the distributed hydrology soil vegetation model to Redfish Creek, British Columbia: Model evaluation using internal catch-ment data” (Whitaker et al. 2003); and “Forest harvesting impacts on the peak flow regime in the Columbia Mountains of southeastern British Columbia: An investigation using long-term numerical modelling” (Schnorbus and Alila 2004a). The model was applied to the snowmelt-dominated Redfish Creek watershed (25.8 km2), which is part of the WADF-paired watershed experiment in the Kootenay Mountains near Nelson, BC.
• “Modellingtheeffectsofloggingroadsonthestreamflowofamountainous,snow-dominatedwatershed.” (Calvert 2003).
• “Hydrologicresponseduringsnowmeltinthreesteepheadwatercatchments:RingroseSlope,Slocan Valley, British Columbia” (Whyte 2004) and “Snowmelt runoff processes and modelling for three small catchments draining a glaciated valley wall (or, When good models go bad)” (Whyte et al. 2004).
• “SimulationofhourlymeteorologyfromdailydataandsignificancetohydrologyatH.J.AndrewsExperimental Forest” (Waichler and Wigmosta 2003) and “Simulation of water balance and forest treatment effects at the H.J. Andrews Experimental Forest” (Waichler et al. 2005).
• “Diagnosingadistributedhydrologicmodelfortwohigh-elevationforestedcatchmentsbasedondetailed stand- and basin-scale data” (Thyer et al. 2004). The paired watersheds, part of the Upper
67
Penticton Creek (UPC) Experimental Watersheds, considered in this study were each approximate-ly 5 km2 in area, with plateau-type topography and spring snowmelt-driven streamflow generation.
• “Amodelofrapidpreferentialhillsloperunoffcontributionstopeakflowgenerationinatemperate rain forest watershed” (Beckers and Alila 2004). A model for the Carnation Creek watershed on Vancouver Island, British Columbia, was used to assess preferential hillslope runoff contributions to peak flow generation. Stream gauges are located on the main channel near the mouth (B weir, drainage area 9.6 km2) and in the upper watershed (E weir, 2.7 km2) and on three tributary streams: C (1.4 km2), H (0.12 km2), and J (0.24 km2).
Forest Roads and Sediment Erosion
• “EffectsofforestroadsonfloodflowsintheDeschutesRiverBasin,Washington”(LamarcheandLettenmaier 2001).
• “Aspatiallydistributedmodelforthedynamicpredictionofsedimenterosionandtransportinmountainous forested watersheds” (Doten et al. 2006).
• “Uncertaintyinforestroadhydrologicmodelingandcatchmentscaleassessmentofforestroadsediment yield” (Surfleet 2008).
• “Effectsoffire-precipitationtimingandregimeonpost-firesedimentdeliveryinPacificNorthwestforests” (Lanini et al. 2009)
Climate Change and Hydrological Forecasting
• “PotentialclimatechangeimpactsonmountainwatershedsinthePacificNorthwest”(LeungandWigmosta 1999).
• “Descriptionandevaluationofahydrometeorologicalforecastsystemformountainouswater-sheds” (Westrick et al. 2002).
These publications indicate that DHSVM is a useful model for addressing complex resource manage-ment questions in the Pacific Northwest (e.g., Bowling et al. 2000; VanShaar et al. 2002; Whitaker et al. 2002; Schnorbus and Alila 2004a), and that the model is able to successfully reproduce a wide range of watershed hydrologic processes in highly instrumented watersheds (Nijssen et al. 1997; Storck 2000; Whitaker et al. 2003; Thyer et al. 2004; Beckers and Alila 2004). A limitation of the model may be its inability to represent preferential hillslope runoff in the context of simulating peak flow generation in rain-dominated regimes (Beckers and Alila 2004). It is also uncertain as to how accurate the role of forest roads in the overall watershed hydrology is represented (Surfleet 2008). Waichler et al. (2005) assessed the effectiveness of the DHSVM in reproducing observed water balances in several watersheds of the H.J. Andrews Experimental Forest in Oregon. The model accurately predicted streamflow from 1958 to 1998. However, it underestimated low flows, probably due to inadequate storage and groundwater baseflow, and slightly under-predicted high flows, probably because of under-measuring snowmelt and downslope water movement. In general, this study confirmed the effectiveness of DHSVM of simulating watershed processes in the snow-rain transition zone.
DHSVM was also used in a special investigation to assess the effect of mountain pine beetle (MPB) and salvage harvesting on streamflows (Luo et al. 2006; Forest Practices Board 2007) at Baker Creek (1570 km2). This study has been influential in MPB-related hydrology decisions. The lack of comprehen-sive watershed data at Baker Creek does not currently permit a detailed assessment of the performance of DHSVM in simulating hydrologic processes for this medium-sized tributary watershed, which drains into the Fraser River at Quesnel, including the role of selected grid size. The model performance was compared at 200- and 500-m grids (similar results were obtained) and was subsequently used for long-term simulations with the coarser grid to alleviate computational requirements.
68
In a hybrid consulting–research environment, the DHSVM model is being used for short-, medium- and long-term flow forecasting and hydropower assessments (Westrick et al. 2002).
forhym and forwadyA1.5
The Forest Hydrology Model (ForHyM) was developed at the University of New Brunswick to predict water fluxes through forests from mean monthly precipitation and temperature records (Arp and Yin 1992). Companion models include a forest soil temperature model (FORSTEM; Yin and Arp 1993) and a model for nutrient cycling (Oja et al. 1995).
The ForHyM model was modified at the University of British Columbia (the Forest Water Dynamics model [ForWaDy] 8) and can be integrated with the forest ecosystem (growth) models, FORCEE and FORECAST (Kimmins et al. 1999), for the purpose of simulating the effects of forest water dynamics on stand growth and development.
The section below first describes ForHyM features and subsequently discusses the modifications made with respect to simulating evapotranspiration processes in ForWaDy.
Model DescriptionA1.5.1
ForHyM is a lumped watershed model that operates at a monthly timescale. The model includes a single vegetation layer and two soil layers (Tables 4, 6, and 8).
Soil evaporation, plant transpiration, and evaporation of rainfall intercepted by vegetation are simulated using empirical potential evapotranspiration (PET)-based methods (Table 4). Snowmelt is simulated using a degree-day method with no apparent correction for rain-on-snow events. Overall, ForHyM ranks as an empirical model for simulating SVAT processes (Table 6).
Vertical unsaturated water movement through the soil layers is calculated empirically (Table 4), pro-portional to soil water content above field capacity. Streamflow is set equal to subsoil (deepest soil layer) percolate. ForHyM does not incorporate explicit routines for simulating runoff, channel routing, or road hydrology (Table 6).
ForHyM uses mean monthly precipitation and temperature records as meteorological inputs. The ForHyM model is designed to minimize the number of calibrated parameters. Overall, the model ap-pears to be of medium complexity (Table 5).
The major difference between ForHyM and ForWaDy is in the potential evapotranspiration (PET) algorithms (Table 4). While ForHyM uses a PET equation based on relationships with air temperature, ForWaDy drives PET using an energy-budget approach. The simulation of snowfall and snowpack dynamics in ForWaDy is based on the analytical RHESSys Snow Model (Coughlan and Running 1997), which accounts for temperature melt and radiation melt. ForWaDy ranks as a mixed-process model for simulating SVAT processes (Table 6). ForWaDy uses a daily time step (Table 8) and has higher meteoro-logical input requirements compared to ForHyM (daily precipitation, daily minimum and maximum temperature, and solar radiation).
The ForWaDy model can be integrated with the forest ecosystem models FORCEE and FORECAST (Kimmins et al. 1999) for forest growth modelling (Table 6).
A limitation of ForHyM and ForWaDy is the lack of channel-routing simulation, which limits ap-plying the models to water balance calculations for forest stands or small watersheds. The models also do not incorporate seepage losses to groundwater or the effects of aspect and slope (Arp and Yin 1992). Model functionality in simulating stand-level or small-watershed hydrology may therefore be limited to
8 www.forestry.ubc.ca/ecomodels/moddev/forwady/forwady.htm
69
gently sloped regions with shallow bedrock, which are similar to settings in which the model has been previously applied in Ontario and Quebec (see below). ForHyM can be applied in rain or snow regimes, but does not handle rain-on-snow events, while the modified snow routines in ForWaDy are applicable to mixed regimes (Table 7).
Model ApplicationsA1.5.2
Model applications for ForHyM appear to be limited to Ontario (Turkey Lakes: deciduous forest, model application to Basin 31 comprising 4.62 ha) and Quebec (Lac LaFlamme: coniferous forest comprising 62.3 ha). ForHyM was able to successfully reproduce seasonal variations in hydrologic characteristics (SWE, soil moisture, and streamflow) over multi-year periods in these gently sloped watersheds with shallow bedrock (Arp and Yin 1992). No applications of ForHyM in the Pacific Northwest have been found; ForWaDy applications that provide insight into simulating hydrologic processes could also not be found .
hbv-ecA1.6
The HBV (Hydrologiska Byråns Vattenbalansavdelning) model is a conceptual hydrological model designed for use in mountainous environments. It was originally developed in the early 1970s at the Swedish Meteorological and Hydrological Institute (SMHI), and has been used extensively for hydrological forecasting, engineering design, and climate change studies (Lindström et al.1997). The HBV-EC model, reviewed here, was adapted by Environment Canada and UBC (Moore 1993) to better represent glacier processes. Several additional improvements have been made since, as described below.
Model DescriptionA1.6.1
The HBV-EC model uses the Grouped Response Unit (GRU) concept to group DEM/GIS grid cells into bins having similar land cover, elevation, slope, and aspect (Stahl et al. 2008). The model incorporates two soil layers, no explicit vegetation layer, and operates at a daily time step (Tables 4, 6, and 8).
The HBV model uses simplistic fixed throughfall fractions to simulate forest canopy influences on rainfall and snow deposition (Moore et al. 2007). Evapotranspiration from soil moisture storage is cal-culated in an empirical fashion for non-glaciated terrain based on potential evaporation (PET) estimates (Table 4). Snowmelt is calculated using the degree-day method, which varies by aspect and slope, but with no apparent correction for rain-on-snow events. Overall, HBV-EC ranks as an empirical model for simulating SVAT processes (Table 6).
Infiltration, percolation, and runoff simulation in HBV-EC are empirical (Table 4). Soil moisture is modelled separately for non-glaciated, forested, and open areas in each elevation zone, but the same parameter values are used for all zones. The amount of water release that percolates through the soil moisture zone to become runoff is calculated based on field capacity (Moore 1993; Hamilton et al. 2000). Runoff routing is computed separately for glacierized and non-glacierized GRUs. For non-glacierized GRUs, water draining from the soil enters a set of lumped fast- and slow-draining reservoirs (Hamilton et al. 2000). Each glacier GRU has a separate reservoir for runoff routing. Streamflow at the watershed outlet is the sum of the outflow from the fast, slow, and glacier reservoirs. Channel routing appears to be calculated by a simple triangular weighting function (Seibert 1997) or a lumped reservoir routing approach. The model does not account for road hydrology (Tables 4 and 6).
70
Required climate inputs are monthly average temperature and evaporation-rate values, as well as daily temperature, rainfall, and snowfall measurements for the time period being simulated. Overall, the model appears to be of medium complexity (Table 5). A distinguishing feature of the HBV-EC model is that it can be operated using the Green Kenue (formerly EnSim Hydrologic) graphical user interface (GUI), which considerably facilitates model setup and simulation. The HBV-EC and Green Kenue pro-grams are distributed as freeware by the Canadian Hydraulics Centre9. Another distinguishing features is the model’s ability to represent lakes and glacial processes (Tables 4 and 7).
The main model limitation appears to be the simplified channel-routing routine. The model has been applied in small to medium watersheds, and only in mountainous settings (Table 7) with a predominant snowmelt- (and glacial melt) driven runoff component. The temperature-index snowmelt method is not suitable for mixed rain–snow regimes.
Model ApplicationsA1.6.2
The following HBV-EC published model applications were reviewed:
• “Coupledmodellingofglacierandstreamflowresponsetofutureclimatescenarios”(Stahletal.2008). This study investigated the sensitivity of streamflow to changes in climate and glacier cover for the Bridge River watershed, British Columbia, using HBV-EC coupled with a glacier response model. The Bridge River watershed, located in the Southern Chilcotin Mountains (a transition zone from wet coastal mountains to dry interior climate), drains an area of 152.4 km2, of which the glacier cover is approximately 61.8%. • “Applicationofaconceptualstreamflowmodelinaglacierizeddrainagebasin”(Moore1993).Inthis study, HBV-EC was applied to the Lillooet River watershed, a moderately glacierized watershed (total area of 2160 km2, 17% glacierized).
• “Estimatingwinterstreamflowusingconceptualstreamflowmodel”(Hamiltonetal.2000).Themodel was applied to the M’Clintock River drainage area, which covers 1700 km2 of pine and spruce boreal forest and is relatively pristine.
• “Predictingtheeffectsofforestharvestingusingaconceptualstreamflowmodel:Evaluationusinga paired-catchment approach” (Moore et al. 2007). The objective of this study was to assess the ability of the conceptual HBV-EC model to predict the hydrologic effects of salvage harvesting. The study focused on Camp Creek, a 34-km2 watershed in south-central BC, where approximately 30% of the watershed area was salvage harvested following a mountain pine beetle outbreak in the 1970s.
Overall, with the addition of ice melt effects and glacial routing, HBV-EC was able to effectively mod-el streamflow response and glacial processes (Moore 1993; Stahl et al. 2008). The Moore et al. (2007) study illustrated how the HBV-EC model can be used as a tool for predicting streamflow changes due to forest management in ungauged watersheds, but this functionality is currently limited by an “overly simplistic representation of canopy influences on snow deposition” and requires further development to represent the effects of complex stand conditions (e.g., dead trees and regenerating forest) on water and energy fluxes. A rewrite of the model to better represent forest hydrologic processes is ongoing as a col-laborative effort between UBC, EC, and Alberta Sustainable Resource Development.
9 http://chc.nrc-cnrc.gc.ca/Numerical/Downloads/Green_Kenue_e.html
71
hec-hmsA1.7
The Hydrologic Engineering Center’s Hydrologic Modelling System (HEC-HMS) developed by the US Army Corps of Engineers (US-Army Corps of Engineers 2000)10 simulates precipitation-runoff process-es and has capabilities for reservoir operations.
Model DescriptionA1.7.1
Physical representation of watersheds is configured in the model, with hydrologic elements connected in a dendritic network to simulate runoff processes (US-ACE 2000). Available watershed and network ele-ments are: sub-basin, reach, junction, reservoir, diversion, source, and sink, with calculations proceeding in a downstream direction (US-Army Corps of Engineers 2000). The model uses a single soil layer and one vegetation layer and can operate at a sub-daily time step (Tables 4, 6, and 8).
Evaporation is modelled in a soil moisture-accounting (SMA) module, and includes the vaporization of water directly from the soil and vegetative surface, and transpiration through plant leaves. Evapora-tion and transpiration are combined and are calculated empirically (Table 4) using monthly varying evapotranspiration values (US-Army Corps of Engineers 2000). HEC-HMS originally did not account for snow processes, but in 2007 a temperature index-based snowmelt model was added11. Overall, HEC-HMS ranks as an empirical model for simulating SVAT processes (Table 6).
Soil water movement (percolation) is simulated in an empirical fashion (Table 4) using several storage reservoirs. Infiltrated water enters soil storage, with the tension zone filling first, and soil water not in the tension zone percolating to the first groundwater layer (US-Army Corps of Engineers 2000). Ground-water flow is routed from the first groundwater layer, and then any remaining water may percolate to a second deep groundwater layer, which is lost to the model. Groundwater flow is the sum of the volumes of groundwater flow from each groundwater layer at the end of the time interval. Several methods are included for transforming excess precipitation into surface runoff (US-Army Corps of Engineers 2000); unit hydrograph methods include the Clark, Snyder, and SCS technique, an implementation of the kine-matic wave method is also included, and methods are available to simulate infiltration losses (US-Army Corps of Engineers 2000). Event (storm) runoff is modelled using a SCS curve, a gridded SCS curve number, and Green and Ampt.
Open channel flow routing may be accomplished through a variety of methods, including: rout-ing with no attenuation with the lag method, the Muskingum method, and the modified Puls method, which can be used to model a reach as a series of cascading, level pools with a user-specified, storage-outflow relationship (US-Army Corps of Engineers 2000). Channel flow is modelled either with the kinematic wave or Muskingum-Cunge methods. Features of the model include representation of lakes and reservoir operations, and in-stream vegetation/wetlands (Tables 4 and 7) (US-Army Corps of Engi-neers 2000).
HEC-HMS features an integrated work environment that includes a database, data-entry utilities, computation engine, and results reporting tools (US-Army Corps of Engineers 2000). A GUI allows seamless movement between the different parts of the program. The Geospatial Hydrologic Modelling Extension (HEC-GeoHMS) is a software package for use with the ArcView GIS to develop a number of hydrologic modelling inputs. Overall, the model is of relatively high complexity (Table 5) due to its broad functionality in representing both watershed and river processes.
10 www.hec.usace.army.mil/software/hec-hms/11 HEC-HMS Research and Development Review, unpublished US-ACE powerpoint presentation, 2007
72
The main limitation of HEC-HMS appears to be the empirical manner in which evapotranspiration and snowmelt are handled. The model is not recommended for forest management applications, given that its development has emphasized simulating river processes (as reflected by the case study below) rather than watershed processes. The model can be applied to small to large watersheds, with gradual topography and rain or snow regimes (Table 7).
Model ApplicationsA1.7.2
No model applications that highlight HEC-HMS functionality for simulating watershed hydrologic pro-cesses in the forest land base of AB, BC or similar settings have been found. However, Fischenich (1999) illustrated the use of HEC-HMS in the context of stream restoration. The model was used to simulate peak flows for Stirling Branch in the western US. In the contributing watershed, the average elevation of was 1300 feet, while the total watershed area was 3.2 km2.. The effects of no restoration, restoration with trimmed trees in the floodplain, and restoration with trees and shrubs on peak flows were simulated. While this case study does not directly address forest management hydrology, it does examine the effects of increasing channel vegetation.
helPA1.8
The Hydrologic Evaluation of Landfill Performance (HELP) model is a user-friendly model developed by the US-EPA12 originally to compute estimates of water balances for municipal landfills (Schroeder 1996). However, the model has also been applied to compute water balances in a variety of other set-tings, including making groundwater recharge calculations in south-central and southeast BC (see model applications, below).
Model DescriptionA1.8.1
HELP is a quasi two-dimensional (i.e., lumped) parameter hydrologic model that incorporates multiple soil layers and a single vegetation layer (Schroeder et al. 1994). The model operates at a daily time step (Tables 4, 6, and 8).
The model simulates rainfall interception and evaporation, plant transpiration, surface evapora-tion, and soil water evaporation, all of which can be simulated using several alternatives, with the most rigorous alternative being analytical (Table 4). Snow accumulation and melt are controlled by a degree-day method, with a correction for rain-on-snow conditions. Overall, HELP ranks as a mixed model for simulating SVAT processes (Table 6).
Infiltration is computed as the sum of rainfall and snowmelt, minus the sum of interception (evap-oration of surface moisture) and runoff. “The rate of unsaturated vertical drainage is a function of soil moisture storage according to Darcy’s Law and assumed to occur by gravity drainage whenever the soil moisture is greater than the field capacity (Brooks Corey model)” (Schroeder et al. 1994). Runoff is modelled using the U.S. Department of Agriculture (USDA) Soil Conservation Service curve number (CN) method. The model also includes an empirical method for routing subsurface flow. The model does not include a channel-routing component and does not account for road hydrology (Table 6).
The HELP model accounts for seasonal variation in leaf area index through a general vegetative growth model (Table 6) based on SWRRB (Simulator for Water Resources in Rural Basins) and de-veloped by the USDA (Arnold et al. 1989). The growth model assumes that the vegetative species are perennial and that the vegetation is not harvested. In addition, a frozen soil component (Table 4) based
12 www.wes.army.mil/el/elmodels/helpinfo.html
73
on the CREAMS model (Knisel et al. 1985) has been added to improve infiltration and runoff predic-tions in cold regions.
HELP requires daily values of precipitation, temperature, and solar radiation, which can be gener-ated using the WGEN synthetic weather generator developed by the USDA Agricultural Research Service (ARS) (Hanson et al. 1994). Overall, the model appears to be of medium complexity (Table 5). Input and editing have been simplified with interactive, full-screen, menu-driven input techniques. The pro-gram is distributed as freeware. A commercial version of the model (Visual HELP) is also available for a cost of approximately $1000.
The main model limitation appears to be lack of explicit consideration of lateral runoff (i.e., the model is quasi two-dimensional), limiting functionality to water balance calculations for forest stands or small watersheds (Table 7). Model functionality in steep mountainous settings may also be limited. The model can be applied in rain, snow, and mixed regimes.
Model ApplicationsA1.8.2
The following HELP published model applications were reviewed:
• “Modelledimpactsofpredictedclimatechangeonrechargeandgroundwaterlevels”(ScibekandAllen 2006b) and “Comparing modelled responses of two high-permeability unconfined aquifers to predicted climate change” (Scibek and Allen 2006a). In these studies, a methodology was devel-oped for linking climate models and groundwater models to investigate future impacts of climate change on groundwater resources for an unconfined aquifer near Grand Forks in south-central British Columbia and for the Abbotsford-Sumas aquifer in the lower Fraser valley. The HELP model, which was linked to GIS, provided recharge estimates (from infiltration of precipitation) to the groundwater model.
• “Comparingapproachesformodellingspatiallydistributeddirectrechargeinasemi-aridregion(Okanagan Basin, Canada)” (Liggett and Allen 2009), “Comparison of approaches for aquifer vulnerability mapping and recharge modelling at regional and local Scales, Okanagan Basin, BC.” (Liggett 2008), and “Modelling climate change impacts on groundwater recharge in a semi-arid region, Southern Okanagan, BC” (Toews 2007). These studies used the HELP model to predict broad-scale recharge in the valley bottom portions of the 8000 km2 Okanagan Basin. The water-shed was divided into 50-m grid cells, incorporating 841 unique profile models, and was simulated for 30 years of climatic data.
In the first set of studies, the HELP hydrologic model was found to be sensitive to depth of water table (percolation layer depth), soil type, and saturated hydraulic conductivity of the unsaturated zone. There-fore, to achieve accurate results for recharge, it is important to capture the spatial variability of these key variables (Scibek and Allen 2006a, 2006b). Results from the second group of studies suggest that for semi-arid regions, soil water balance processes are not adequately represented. Also, it appears that HELP under-predicts actual evapotranspiration, thereby over-predicting recharge rates (percolation). This is a known drawback, not only of HELP, but also of most other models, including but not limited to HSPF, UBCWM, DHSVM (e.g., Pike 1995), and RHESSys (Tague et al. 2004).
The above studies suggest that HELP is useful for simulating stand-level (soil column) SVAT processes and the associated effects of forest harvesting, but users must carefully consider whether to use HELP in forest management applications, especially in semi-arid environments and terrain with steep slopes.
74
hsPfA1.9
The Hydrologic Simulation Program-Fortran (HSPF13) was designed by the US Envionmental Protec-tion Agency (EPA) to simulate a broad range of surface and sub-surface hydrologic and water quality processes in watersheds (Bicknell et al. 2001). The HSPS model is based on the Stanford Watershed Model and can simulate one or many pervious or impervious unit areas discharging to one or many river reaches or reservoirs.
Model DescriptionA1.9.1
HSPF is a lumped parameter hydrologic model that can simulate all primary natural hydrological pro-cesses (Bicknell et al. 2001). The unsaturated zone is approximated using a single storage reservoir, while explicit representation of vegetation in the model is limited. The model runs at any time step, from 1 minute to 1 day, that can be divided equally into 1 day (Tables 4, 6, and 8).
The model empirically simulates evapotranspiration from interception storage, upper and lower zone storages, active groundwater storage, and directly from baseflow (Table 4) (Bicknell et al. 2001). Snow-melt can be calculated using either a temperature-index method or an energy balance method. Overall, HSPF ranks as a mixed model for simulating SVAT processes (Table 6).
Infiltration, percolation, and runoff are simulated using empirical methods (Table 4). Unsaturated zone storage is is approximated using a single storage reservoir with inflow. Inflow to the unsaturated zone is controlled by the ratio of actual storage to a specified nominal storage (Bicknell et al. 2001). Unsaturated zone water content can be decreased by reducing the user-defined fraction of rainfall that enters the unsaturated zone. Groundwater for each surface water sub-basin is simulated as two storage reservoirs (active and deep) and linear routing of groundwater in the active groundwater storage res-ervoir can be simulated and controlled with an empirical routing parameter (Bicknell et al. 2001). The kinematic wave approximation is used to route overland flow to river reaches (Bicknell et al. 2001).
The kinematic wave approximation is also used to route water in the river reaches and surface water structures can be simulated using fixed stage-discharge-storage relationships (Bicknell et al. 2001). HSPF does not explicitly consider road hydrology (Table 6).
Distinguishing features of HSPF include representation of lakes and water control structures (Tables 4 and 7).
At a minimum, the model requires temperature and precipitation as meteorological inputs, while ad-ditional inputs are required for the energy balance snowmelt module. An interactive windows interface (GUI) to HSPF is provided by the winHSPF environment14 (Bicknell et al. 2001). HSPF has also been integrated into a GIS environment with the development of Better Assessment Science Integrating point and Non-point Sources (BASINS15), which also supports several other models. WinHSPF is currently distributed with BASINS, and both are freeware. The primary drawback of using the HSPF model for operational forestry purposes is that the model is complex (Pike 1995; Table 5).
13 http://water.usgs.gov/software/HSPF/14 http://www.aquaterra.com/resources/hspfsupport/index.php15 www.epa.gov/waterscience/basins/
75
HSPF’s complexity is compounded by its user-unfriendliness, which makes it difficult to use without direct guidance from an experienced HSPF model user. Furthermore, many parameters that control hydrologic processes are empirical and can only be determined through calibration. The HSPF is also limited in that it was created for terrain quite different from the mountainous terrain of BC and por-tions of Alberta. The model can be applied to small to large watersheds in gradual terrain in rain, snow, or mixed climatic regimes (Table 7). However, based on the above-mentioned limitations, HSPF is not recommended as a tool for assessing the hydrologic effects of forest management.
Model ApplicationsA1.9.2
Many of the published applications of the HSPF model have been in the central and east portions of the USA (Pike 1995). In British Columbia, data from Carnation Creek, on the eastern coast of Vancouver Island, have been used in a test calibration for using HSPF in forested watersheds (Hetherington et al. 1995). An important finding of this study was that forest cover in the HSPF is represented as a single index, making the representation of differing forest cover types very limited.(Pike 1995).
In AB, HSPF is widely used for simulating water quantity and quality in environmental-impact assess-ments associated with oil sands operations. However, these consulting applications do not provide much insight into how HSPF performs in simulating watershed hydrologic processes in forest management because the model is mostly used in predictive mode (to assess the effects of mine developments) with a set of standardized parameters and with limited comparisons to measured streamflow.
Many applications of the HSPF to wetlands in the past have proven to be less than satisfactory. How-ever, the USGS in Washington has apparently been able to obtain acceptable results from the HSPF for these types of landscapes (Pike 1995).
inhm and hydrogeosphereA1.10
The Integrated Hydrology Model (InHM16) and HydroGeoSphere, both developed at the University of Waterloo, are fully coupled, comprehensive, physically based, spatially distributed, integrated surface wa-ter and groundwater models. The subsurface module of both models is based on the three-dimensional (3-D) subsurface flow and transport code FRAC3DVS.
Model DescriptionA1.10.1
In InHM and HydroGeoSphere, the surface and subsurface domains are discretized using finite ele-ments. The models are therefore fully distributed in both horizontal and vertical direction. The models have no explicit representation of vegetation. A flexible time-stepping algorithm to reduce run times is available in both models (Tables 4, 6, and 8).
InHM does not explicitly consider SVAT processes. Actual evapotranspiration can be calculated by specifying potential evapotranspiration, which is then limited by soil moisture availability. Interception and evapotranspiration calculations are included in HydroGeoSphere. Evapotranspiration is calculated using the analytical Kristensen and Jensen (1975) method (Li, Q. et al. 2008a; Therrien et al. 2008). Snowmelt is not included in either InHM or HydroGeoSphere (Table 4).
16 http://inhm.org/
76
Variably saturated flow is simulated physically (Table 4) using Richards equation with a compre-hensive representation of unsaturated zone soil properties. First-order coupling equations (essentially one-dimensional Darcy’s equations) are used between surface and subsurface computational domains. To simulate channel and overland flow, the model employs a two-dimensional, diffusive-wave ap-proximation based on the Surface Water Flow Package of the MODHMS simulator. Both models use a two-dimensional representation to simulate the flow field and the channel cross-section. Road hydrol-ogy can also be simulated in this fashion, including the interception of precipitation, surface runoff, and subsurface flow, and the redirection of these flows. However, stream crossing structures are not explicitly incorporated in the models (Tables 4 and 6). Lakes or wetlands can be included (Table 7) either as boundary conditions (by specifying water levels and the physical properties of lake bottom or wetland sediments) or as part of the simulation (water levels are calculated).
Distinguishing features of InHM and HydroGeoSphere are the ability to simulate subsurface flow processes in a physically based manner, including porous media, fractures, subsurface conduits, macro-pores, and perched water tables (Tables 4 and 7). This makes these models ideal tools to simulate complexities of the subsurface movement of water and solutes (contaminants) within watersheds.
InHM and HydroGeoSphere only require precipitation as a meteorological input for event simu-lations and potential evapotranspiration for continuous simulations. A GUI is not available for the models, nor are InHM or HydroGeoSphere set up for integration in a GIS environment. Overall, these models rank as being of high complexity (Table 5) due to the detailed nature in which subsurface flow processes are simulated.
As predominantly groundwater-focused models, InHM and HydroGeoSphere are not set up for simulating the effects of forest harvesting. In a forest management context, their main use is in studying groundwater surface water interactions and in simulating the effects of roads on watershed hydrology, with an emphasis on the interaction of roads with subsurface flow processes. The models can be applied in both gradual and steep terrain, and in small to large watersheds (Table 7). However, in a forest man-agement context, applications to date have been limited to research applications in small watersheds.
Model ApplicationsA1.10.2
The following InHM/HydroGeoSphere published model applications were reviewed:
• “Simulatedeffectofforestroadsonnearsurfacehydrologicresponse:Redux”(Mirusetal.2007).InHM was employed to conduct hydrologic-response simulations for the small upland watershed known as C3 (located within the H. J. Andrews Experimental Forest in Oregon) with concave topography, sparse understorey vegetation, and a dissecting road. The InHM simulation of hydro-logic response was continuous in this study, using hourly rainfall over a 91-day period.
• “Near-surfacehydrologicresponseforasteep,unchanneledcatchmentnearCoosBay,Oregon:2.Physics-based simulations” (Ebel et al. 2007). InHM was used to simulate the hydrologic response from three sprinkling experiments on the intensively studied CB1 watershed. The 860-m2 Coos Bay experimental watershed located in the Oregon Coast Range is steep with a maximum elevation of 300 m above sea level. The InHM-simulated hydrologic response was evaluated against observed discharge, pressure head, total head, soil water content, and deuterium concentration records.
• “Simulationsoffullycoupledlake-groundwaterexchangeinasubhumidclimatewithanintegrated hydrologic model” (Smerdon et al. 2007), “The influence of sub-humid climate and water table depth on groundwater recharge in shallow outwash aquifers” (Smerdon et al. 2008). InHM was successfully used to simulate transient groundwater flow and interaction with a shallow lake on glacial outwash terrain, in the Boreal Plains of Canada. The site is located
77
370-km north of Edmonton, AB, and is characterized by a subhumid climate, thick heterogeneous glacial sediments, overlying sedimentary rock, and subdued topographic relief. Simulations included one-dimensional profile models for groundwater recharge (70-year climate record), and a three-dimensional model of lake-groundwater interaction (40-ha site, 2-year transient analysis).
The weathered bedrock piezometric response and runoff contribution were not simulated well with InHM in the Coos Bay study. This was most likely as a result of the uncertainty in the weathered bedrock layer geometry and fractured-rock hydraulic properties that preclude accurate fracture flow representa-tion (Ebel et al. 2007). Such limitations are expected to affect the performance of any model, and are therefore not specific to InHM or HydroGeoSphere.
Jointly, the above model applications indicate that InHM (and by extension HydroGeoSphere) are most useful for catchment-scale, event-based rainfall–runoff simulation, for characterizing associated hydrologic responses, and for developing process understanding (i.e., groundwater surface water inter-actions and the effects of roads). Use of the InHM and HydroGeoSphere models is therefore applicable in this context. However, the models require high-level subsurface (geologic and hydrogeologic) data for model setup and calibration. Model use in a continuous simulation context (i.e., periods of 1 year or longer) is limited due to rudimentary incorporation of SVAT processes.
mike-sheA1.11
The integrated hydrological modelling system, MIKE-SHE17, simulates the major processes in the hy-drologic cycle and includes process models for evapotranspiration, overland flow, unsaturated flow, groundwater flow, channel flow, as well as modelling how all these interact. The MIKE-SHE model was developed, starting in 1977, as consortium of three European organizations (Système Hydrologique Européen, [SHE]) (Abbott et al. 1986a, 1986b) and the full model is currently distributed by the Danish Hydrological Institute (DHI).
Model DescriptionA1.11.1
MIKE-SHE is a fully distributed, grid-based hydrologic modelling system that incorporates multiple soil layers and a single vegetation layer (i.e., overstorey and understorey are not distinguished) and flexible model time step (Tables 4, 6, and 8) (Abbott et al. 1986a, 1986b).
MIKE-SHE considers canopy interception of precipitation, evapotranspiration (from vegetation and soil), and snow accumulation and melt. Evapotranspiration is calculated using the analytical Kristensen and Jensen (1975) method, while snowmelt is calculated using an empirical degree-day method with no apparent correction for rain-on-snow events (Table 4). Overall, MIKE-SHE therefore ranks as a mixed model for simulating the effects of forest harvesting (Table 6).
Precipitation (throughfall or snowmelt) reaching the ground surface can either infiltrate or runoff as overland flow. MIKE-SHE includes three methods to simulate flow in the unsaturated zone, but as-sumes that flow is vertical in all three methods (Abbott et al. 1986a, 1986b). Two of the unsaturated zone methods are based on Richards’ equation, while the third, simplified module uses a linear relationship between depth-to-the-water table and average soil moisture content, and a linear infiltration equation (Abbott et al. 1986a, 1986b). MIKE-SHE includes a physical three-dimensional, saturated zone model, similar to InHM (HydroGeoSphere). MIKE-SHE does not have explicit provisions for simulating road hydrology (Tables 4 and 6).
17 www.dhigroup.com/Software/WaterResources/MIKESHE.aspx
78
Overland flow is simulated using the diffusive wave approximation. MIKE-SHE’s river modelling component is the MIKE-11 modelling system for river hydraulics, which supports any level of complex-ity and offers simulation engines that cover the entire range from simple Muskingum routing to the Higher Order Dynamic Wave formulation of the Saint-Venant equations (Abbott et al. 1986a, b). MIKE-SHE also has a full range of reservoir operation capabilities (Table 4) (Abbott et al. 1986a, 1986b).
A distinguishing feature of MIKE-SHE is that it is one of the few models that has been designed and developed to fully integrate surface water and groundwater flow, including SVAT processes and sophisti-cated water management utilities (e.g., reservoir operation). This sets MIKE-SHE apart from InHM and HydroGeoSphere, although these latter two models in turn are more rigorous with respect to simulating groundwater flow.
MIKE-SHE requires precipitation and temperature as model inputs. MIKE-SHE includes a full suite of pre- and post-processing tools, plus a flexible mix of advanced and simple solution techniques for each of the hydrologic processes. Each process can be represented at different levels of spatial distribu-tion and complexity, according to the goals of the modelling study, the availability of field data, and the modeller’s choices. The MIKE-SHE user interface allows the user to intuitively build the model descrip-tion based on the user’s conceptual model of the watershed. Nonetheless, overall, MIKE-SHE ranks a complex model (Table 5) and the software is expensive (about 16000 EURO for the Enterprise Version in December 2008).
Model applications to date do not indicate whether the model is able to handle steeply sloped terrain. MIKE-SHE is therefore recommended for gradual terrain, and can be applied to small to large water-sheds in rain or snow regimes, but is not recommended for mixed regimes (Table 7). Given its cost and development purpose, the MIKE-SHE’s main use would appear to be in forest management situations that also involve a surface water management or groundwater component.
Model ApplicationsA1.11.2
MIKE-SHE has proven valuable in hundreds of research and consultancy projects covering a wide range of climatic and hydrologic regimes, many of which are referenced in Graham and Butts (2005). The fol-lowing MIKE-SHE published model applications were reviewed:
• “Anapproachforpredictinggroundwaterrechargeinmountainouswatersheds”(Smerdonetal.2009). This study simulated overland flow and unsaturated flow in MIKE-SHE, using 100-m grids and 30 years of daily climate data. The study area was Swan Lake Valley, a 50-km2 watershed located in the northern Okanagan region of British Columbia.
• SimulationofcompletehydrologiccycleinhumidwatershedKleineandtheGroteGeteCatchment(Belgium), and Karup Catchment (Denmark) (Refsgaard 1997; Feyen et al. 2000). The model was applied to two watersheds of 440 to 600 km2, with generally low topographic relief, and 2 to 4 years of daily climate/hydrologic data. Stream discharge and groundwater levels were replicated well in the model.
In the Okanagan study, simulated actual evapotranspiration was found to generally agree with field measurements. This study also illustrated the importance of model selection, compared to other com-mon models (e.g., HELP), for estimating recharge in semi-arid regions.
The European studies, which have become “reference” watersheds for the MIKE-SHE simulator, offer an illustrative example of how MIKE-SHE could be used for watershed-scale assessment for medium watersheds with relatively modest topography.
79
An Alberta Environment (AENV)-sponsored MIKE-SHE project for the Lesser Slave Lake watershed was recently completed to assess the potential effects on streamflows of a planned increase in lake level to increase water storage capacity. Overall, MIKE-SHE was found to adequately reproduce annual flows for the watershed (including for gauged tributary streams) from the 1970s, although some questions do remain regarding the simulation of internal watershed processes (which is mostly a function of lack of data; Dr. M. Shome, WorleyParsons, pers. comm., Dec. 2009). MIKE-SHE is also being considered for simulating future watershed hydrology, including groundwater regimes, in reclaimed landscapes in the Alberta Oil Sands region (Dr. J. Fennell, WorleyParsons, pers. comm., Dec. 2009). In BC, the Okanagan Basin Water Board (OBWB)18 has recently selected MIKE-SHE for hydrologic modelling of the water-shed in the context of addressing local water management challenges.
modhmsA1.12
MODHMS combines the capabilities of the models MODFLOW and HEC-HMS. The modular finite-difference groundwater flow model (MODFLOW) developed by the U.S. Geological Survey19 is the most widely used program in the world for simulating groundwater flow. MODFLOW has been accepted by regulators and has been cited in many court cases in the United States as a legitimate approach to ana-lyzing groundwater systems. MODHMS was developed by HydroGeoLogic Inc20.
Model DescriptionA1.12.1
In MODHMS, the surface and subsurface domains are discretized using a finite difference scheme (grid blocks). The model is therefore fully distributed in both horizontal and vertical direction. MODHMS has no explicit representation of vegetation. MODHMS has an adaptive time-step algorithm that tailors time steps to convergence of the iterative solver and user-specified parameters (Tables 4, 6, and 8).
Interception and evapotranspiration are calculated in the model, along with a number of potential-evapotranspirationformulas (from bare ground or vegetated surfaces) that may be applied in different regions of a model domain. Evapotranspiration is calculated using the analytical Kristensen and Jensen (1975) method. Snowmelt is not included in the model (Table 4). Overall, MODHMS ranks as an em-pirical model for simulating SVAT process and the effects of forest harvesting (Table 6).
The subsurface flow component of MODHMS solves the physical three-dimensional Richards’ equa-tion for variably saturated flow, and overland flow is simulated using the diffusive-wave approximation (Table 4). MODHMS includes all of the standard MODFLOW functionality. MODHMS is capable of modelling open channel flow and closed pipe flow using the diffusive wave approximation, and similar to HEC-HMS, simulates lakes, wetlands and structures, including dams, weirs, culverts, and gates (Tables 4 and 6).
Similar to MIKE-SHE, a distinguishing feature of MODHMS is that it one of the few models that has been designed and developed to fully integrate surface water and groundwater flow, including SVAT processes and sophisticated water management utilities (e.g., reservoir operation). However, contrary to MIKE-SHE, snow processes are not included in MODHMS.
The model requires precipitation as input. Overall, the model is of high complexity (Table 5) due to the detailed nature in which subsurface flow processes are simulated. MODHMS is proprietary software and not typically available for distribution, except by special arrangement.
18 www.obwb.ca/19 http://water.usgs.gov/nrp/gwsoftware/modflow.html20 www.hglsoftware.com/
80
Apart from being proprietary software, the main drawback of MODHMS, from the perspective of forest management applications in BC and AB, is its lack of accounting for snow accumulation and melt processes. The utility of MODHMS would appear to be limited to rainfall-dominated, small to large watersheds in gradual terrain (Table 7), with forest management situations that also involve a surface water management or groundwater component.
Model ApplicationsA1.12.2
MODHMS applications were identified in Australia and Kansas, with conference proceedings for the latter application (Beeson et al. 2004) lacking sufficient detail to assess model performance. In the Australian study (Werner et al. 2006), the MODHMS model replicated transient groundwater levels and streamflow observations for a 420-km2 tropical watershed (1500 mm/yr precipitation). The results illus-trate how MODHMS can be used to simulate internal watershed processes. The study also suggests that finer-scale grid resolution would help better represent baseflow groundwater interaction, but would be prohibitive in terms of computational efficiency. Computational requirements are a general constraint of fully integrated hydrological models such as MODHMS, InHM (HydroGeoSphere), and MIKE-SHE. The referenced studies do not provide insight regarding MODHMS applicability to the forest land base in AB or BC.
PrevAhA1.13
The Precipitation-Runoff-Evapotranspiration-Hydrotope (PREVAH21) and WaSiM-ETH models (see Section A1.21) have been developed by the Eidgenössische Technische Hochschule (ETH) in Zürich (Switzerland), with the intent of improving the understanding of the spatial and temporal variability of hydrological processes in watersheds with complex topography and to evaluate the hydrological im-pacts for future climatic scenarios (Gurtz and Zappa 2004; Viviroli et al. 2009). PREVAH is extensively documented through user manuals and includes a GUI to simplify data input and paramterization. The models appear to be similar with respect to simulating SVAT processes, but differ in how they repre-sent runoff and groundwater processes, with PREVAH employing a relatively simple conceptualization similar to the HBV model, and WaSiM-ETH incorporating an option to use the three-dimensional form of the Richards’ equation similar to models such as InHM (HydroGeoSphere), MODHMS, and MIKE-SHE. An overview of the PREVAH model has been published by Viviroli et al. (2009) and the model has been reviewed by Werner and Bennett (2009) for climate change applications in cold mountainous watersheds.
Model DescriptionA1.13.1
PREVAH employs Hydrologic Response Units (HRUs) to characterize a watershed, conceptual layering for soil moisture storage and runoff, and a single vegetation layer (i.e., overstorey and understorey are not distinguished). The model operates at a sub-daily time step. To represent runoff generation process-es in mountainous regions, a 1-hour time interval is recommended (Tables 4, 6, and 8), while for water balance calculations, a 1-day time interval may be sufficient (Gurtz and Zappa 2004).
21 www.hydrologie.unibe.ch/PREVAH/
81
PREVAH uses physically based methods to simulate canopy interaction with precipitation, evapora-tion of intercepted water, and evapotranspiration from plants and soil (Table 4). The snow accumulation and melt relies on the combination of a temperature-index and an energy balance approach (analytical method), with a distinction between radiation-dependent melt in periods without precipitation and advectively-induced ablation (Gurtz and Zappa 2004). Overall, PREVAH ranks as an analytical model for simulating SVAT processes (Table 6).
Infiltration and soil water-storage capacity are determined empirically (Table 4) by the depth available for the roots and the plant-available field capacity of the soil (Werner and Bennett 2009). The model contains specific rules for runoff from surfaces with rocks, wet areas, non-vegetated surfaces, and urban areas. Inputs to the soil water reservoir and to the runoff storages are calculated as a function of soil moisture content and soil characteristics of different HRUs (Gurtz and Zappa 2004). The model com-ponents for runoff formation originate from an ETH-version of the HBV-model and have been adapted for the HRU-based simulation of runoff generation (Gurtz and Zappa 2004). In the runoff reservoirs of the model, three runoff components are generated: quick surface runoff, interflow (delayed), and groundwater flow as baseflow (Gurtz and Zappa 2004). There are two groundwater storages with fast groundwater flow component and a more delayed baseflow component. PREVAH does not explicitly ac-count for road hydrology (Tables 4 and 6).
Flood routing is calculated based on the combination of linear storages and translation components. The model can include inputs/outputs to/from other watersheds that are not considered in the actual model run (Gurtz and Zappa 2004). The influence of water control structures (Table 4) can be simulated indirectly by including rules that control the input and output to a reservoir, in an additional module (Gurtz and Zappa 2004).
A distinguishing component of PREVAH is the incorporation of glacial melt, based on a distributed temperature-index, ice-melt model, including potential direct solar radiation, similar to that used for snowmelt modelling (Tables 4 and 7).
Required meteorological inputs to PREVAH include precipitation, air temperature, air humidity, global radiation, relative sunshine duration, and wind speed. Meteorological variables are distributed in the model based on elevation, slope and aspect, rendering PREVAH useful in dealing with the effects of complex terrain. WINPREVAH is a GUI interface used to operate the PREVAH-Modelling System22. Overall, the model ranks as having relatively high complexity (Table 5), requiring substantial data pro-cessing and GIS.
PREVAH has mostly been applied in steeply sloped terrain and in small to medium watersheds (Table 7). The main drawbacks of the model would appear to be its relative complexity and associated high meteorological input requirements, and the fact that, to date, experience with the model resides only in Europe.
Model ApplicationsA1.13.2
So far, the PREVAH model has been used solely to study mountainous environments in Switzerland, which are similar in setting to the mountainous regions and glacierized areas of Alberta and BC. In addi-tion to the applications reviewed below, further applications have been reviewed by Viviroli et al.
• “Spatiallydistributedhydrotope-basedmodellingofevapotranspirationandrunoffinmountainous basins” (Gurtz et al. 1999). PREVAH was applied for the hourly simulation of evapo-transpiration, soil moisture, water balance, and the runoff components in 12 sub-watersheds of the alpine/pre-alpine portions of the River Thur (1703 km2). The 12 sub-watersheds were modelled
22 www.hydrologie.unibe.ch/PREVAH/
82
over the 1993 –1994 period (Gurtz et al. 1999).• “Seasonalwaterbalanceofanalpinecatchmentasevaluatedbydifferentmethodsforspatially
distributed snowmelt modelling” (Zappa et al. 2003). The alpine watershed of the Dischmabach in Switzerland was analyzed in the period from 1982–2000 with PREVAH.
• “Contributionofglaciermelttostreamrunoff:Iftheclimaticallyextremesummerof2003hadhappened in 1979.” (Koboltschnig et al. 2007). This paper presents a comparative PREVAH study at a small and highly glacierized 2.72-km2 watershed area in the Austrian Alps, where runoff was simulated based on two different glacier extents: the 2003 glacier extent (52%) and the larger 1979 extent (71%).
• “Distributedhydrologicalmodellingofaheavilyglaciatedalpineriverbasin”(Kloketal.2001).Inthis study, a glacier model component was successfully integrated into two distributed hydrologi-cal models (WaSiM-ETH and PREVAH) to simulate the discharge of a heavily glaciated watershed. The model was tested on a high-alpine, sub-watershed of the Rhone river (central Switzerland), of which 48% is glaciated. Continuous discharge simulations were performed for the period of 1990–1996 and compared with hourly discharge observations. Slightly better simulation results of the water balance and the runoff hydrographs for the Rhone/Gletsch watershed were obtained with PREVAH compared with WaSiM-ETH.
Overall, these studies suggest that PREVAH can be successfully applied in complex mountainous terrain, including watersheds with glacial melt contributions. However, forest hydrology-specific as-sessments of the model (e.g., canopy interaction with precipitation) have not been tested. It is also not known how well the model is able to simulate watershed hydrology in more gently sloped terrain, although, contrary to, for example, DHSVM, PREVAH does allow for groundwater contributions, which in theory should lead to better performance.
Prms (mms)A1.14
The Precipitation-Runoff Modeling System (PRMS) is a “modular-design, deterministic, distributed-parameter modeling system developed by the USGS23 to evaluate the impacts of various combinations of precipitation, climate, and land use on streamflow, sediment yields, and general watershed hydrology” (Leavesley et al. 2005). To better allow for model updates and adding process components, the architec-ture and modular structure of PRMS were redesigned and formed the basis for the Modular Modeling System (MMS) (Leavesley et al. 2005).
Model DescriptionA1.14.1
Watersheds are divided into HRUs based on characteristics including slope, aspect, elevation, vegetation type, soil type, land use, and precipitation distribution (Leavesley et al. 2005). The model incorporates a single vegetation layer, two conceptual soil zones (recharge zone and a lower zone), subsurface and groundwater reservoirs, and operates at a daily time step (Tables 4, 6, and 8).
Three empirical procedures (Table 4) are available to compute potential evapotranspiration: using pan-evaporation data, potential evapotranspiration as a function of daily mean air temperature and possible hours of sunshine, and the analytical Jensen-Haise technique (Jensen et al. 1969). Potential evapotranspiration is first satisfied from interception storage, retention storage on impervious surfaces, and evaporation/sublimation from snow surfaces, and the remaining potential evapotranspiration demand is then applied to the soil-zone storage (Leavesley et al. 2005). Snowpack water balance is
23 http://water.usgs.gov/software/PRMS/
83
computed daily and the energy balance is computed twice daily (for 12-hour periods) using a hybrid (analytical) energy balance method (simplifies latent and sensible heat terms as a function of air temper-ature and is computed only on days with precipitation) (Leavesley et al. 2005). Overall, PRMS ranks as a mixed model for simulating SVAT processes (Table 6).
Infiltration, percolation, and runoff are simulated using empirical methods (Table 4). Infiltration to the HRU soil zone first fills the soil-zone reservoir to its maximum storage capacity, and subsequent infiltration is routed to the subsurface and groundwater reservoirs (Leavesley et al. 2005). The subsur-face reservoir’s relatively rapid flow may occur in the vadose and groundwater zones during rainfall and snowmelt events (Leavesley et al. 2005). The groundwater reservoir simulates the slower component of flow from the groundwater zone. PRMS does not have a specific allowance for road hydrology (Tables 4 and 6).
There is no explicit routing of channel flow in PRMS (daily mode). However, channel reservoir components can be used to simulate the storage and routing response of channel reservoirs using a linear-storage routing procedure or a modified-Puls routing procedure (Leavesley et al. 2005). Reservoir inflows are computed as the sum of the streamflow contributions from all HRUs and the parts of sub-surface and ground-water reservoirs above the channel reservoir, and may include the outflow of up to three upstream channel reservoirs (Leavesley et al. 2005).
Meteorological inputs are daily precipitation, maximum and minimum air temperature, and solar radiation. The modular design of MMS provides a flexible framework for continued model-system enhancement and hydrologic-modeling research and development. Overall, the model ranks as being of relatively high complexity (Table 5) due to its large number of input parameters (although less than HSPF) and limited efforts to make the model user friendly. A GIS interface has been developed for MMS to facilitate model development, application, and analysis (Leavesley et al., 2005).
Model ApplicationsA1.14.2
The following PRMS (MMS) published model applications were reviewed:
• “Estimatesofground-waterrecharge,baseflow,andstreamreachgainsandlossesintheWil-lamette River Basin, Oregon” (Lee and Risley 2002). PRMS was used to estimate the water budget of 216 sub-watersheds in the Williamette River and to compute long-term average recharge and baseflow over a 24-year period. The watershed has a temperate climate and is approximately 30 000 km2.
• “GSFLOW–CoupledGround-WaterandSurface-WaterFlowModelbasedontheintegrationofthe Precipitation-Runoff Modelling System (PRMS) and the Modular Ground-Water Flow Model (MODFLOW-2005)” (Markstrom et al. 2008). Coupled with MODFLOW, PRMS was used to simulate runoff and groundwater recharge at Sagehen Creek in California, a USGS Hydrologic Benchmark Network Basin located on the east slope of the northern Sierra Nevada. The watershed drains an area of approximately 27 km2. The simulation period was 24 years, and was composed of 128 HRUs in PRMS and 90-m grid cells (in MODFLOW).
• AsimulationmodelofBoulder’salpinewatersupply(PaytonandBrendecke1985).Themodelwasapplied to the Boulder watershed, a 28-km2 watershed located along the continental divide 24-km west of Boulder, Colorado. The watershed ranges in elevation from 2900 to 4100 m. The watershed itself is predominantly alpine with steep rocky slopes, small glaciers, and tundra vegetation. The lower third of the watershed is covered with subalpine fir and Englemann spruce.
84
In the Williamette River study, simulated annual flows compared well with observed annual flows from 47 selected sub-watersheds of varying size, both unregulated and regulated, for the 1973–1996 water-year period.
In the Colorado study, a drawback of PRMS was that the snowpack is assumed to be uniformly dis-tributed over an entire HRU and melt proceeds at a uniform over the entire HRU. Thus, at the end of the melt season, the snow cover on an HRU goes from 100% coverage to none in a single day, producing a rather drastic (and unrealistic) cessation of contribution to runoff.
The results from these studies indicate that PRMS has the capability to model hydrological processes typically associated with the Pacific Northwest (e.g., steep watersheds and snow processes), and can also be linked to a groundwater flow model for simulating baseflow to stream networks, springs, and lower- elevation valley systems.
rhessysA1.15
RHESSys24, the Regional Hydro-Ecological Simulation System, is a GIS-based, modelling framework that integrates water and biogeochemical cycling and transport over spatially variable terrain at small (first-order streams) to medium (fourth- and fifth-order streams) scales (Tague and Band 2004). RHESSys combines both physically based process models and a method for partitioning and parameterizing the landscape (Tague and Band 2004).
Model DescriptionA1.15.1
RHESSys uses a hierarchical approach to partition a landscape into hydrologically distinct units, includ-ing patches (essentially HRUs, areas with similar soil and land use characteristics), hillslopes (portions of a watershed that drain to a single point or stream reach), climate zones and watersheds (to organize stream routing). Multiple vegetation strata can be included in the model, while soil water storage is con-ceptually divided into an unsaturated and a saturated zone (Tague and Band 2004). The model operates at a daily time step (Tables 4, 6, and 8).
Evaporative fluxes from each canopy layer include water intercepted by the canopy, sublimation of intercepted snow, and transpiration (Tague and Band 2004), calculated using physically based equations (Table 4). Snowmelt is computed using a quasi-energy budget approach that takes into account radia-tion and a combination of controls due to sensible and latent heat fluxes, and advective fluxes (from rain-on-snow) (Coughlan and Running 1997). Overall, RHESSys ranks as an analytical model for simu-lating SVAT processes (Table 6).
At each time step, net throughfall from canopy layers and snowmelt are added to surface detention storage and infiltrated into the soil using Phillip’s infiltration equation (Tague and Band 2004). Pon-ded water that is not infiltrated within the daily time step becomes detention storage, which, beyond a surface detention storage capacity parameter, will become overland flow. Drainage from the unsaturated zone to the saturated zone is limited by field capacity of the unsaturated zone. Two approaches are used to calculate runoff, including a topograhically defined wetness index (TOPMODEL; Beven and Kirkby 1979) and an explicit analytical routing model (Table 4), adapted from DHSVM (Wigmosta et al. 1994).
Open channel routing requires explicitly identifying the location of stream channels. RHESSys allows for the simulation of road hydrology, including interception of surface and shallow subsurface runoff, precipitation interception, and routing of road flow using explicit information on road network geom-etry and inter-connectivity, and the location of stream crossing structures (Tables 4 and 6).
24 http://fiesta.bren.ucsb.edu/~rhessys/
85
Additionally, RHESSys includes the following model components (Tables 4 and 6) (Tague and Band 2004):
• SpatialinterpolationofclimatevariablesusingtheMTN-Climmodel(Runningetal.1987).Themodel uses topography (elevation, slope, and aspect) and climate station information to spatially distribute input climate variables over complex terrain.
• Eco-physiologicalmodel(BIOME-BGC;RunningandHunt1993)toestimatecarbon,water,andpotential nitrogen fluxes from different canopy cover types. The models include the allocation of carbon and nitrogen to the various tissues (leaves, roots, and stems) that make up overall biomass to represent plant growth.
• Annualplantmortalityratemaybesimulatedasafixedpercentageofcurrentbiomass(TagueandBand 2004). Future versions of RHESSys may incorporate a variable mortality rate as a function of environmental stressors that increase susceptibility to disease, blow down, etc. In the current ver-sion, episodic changes in vegetation such as forest harvesting or fire can be implemented through a dynamic redefinition of the stratum-level, carbon- and nitrogen-state variables (Tague and Band 2004).
• Temporaleventcontrol(TEC)allowstheusertodefinethetimingandnatureofdisturbanceevents (e.g., forest harvesting, fire, or road construction) that occur during the simulation period.
Minimum meteorological inputs for RHESSys are precipitation and daily minimum and maximum temperature. Additional inputs, depending on model complexity, may include duration of rainfall, shortwave and longwave radiation, relative humidity, and wind speed. If radiation and relative humidity (vapour pressure deficit) are not provided as input, they are calculated internally in the model. Parameterization and data management within RHESSys are complex due to the multiple levels of spatial partitioning (topographic, land-cover, and soil map layers) and the associated parameter sets. RHESSys incorporates a number of interface programs, which format input data as required by the model, including GIS-based, terrain-partitioning programs and RHESSys-specific programs that derive landscape representation from GIS images and establish connectivity between spatial units. These various programs can be run in stand-alone mode or as part of an integrated RHESSys interface, RAINMENT (Band et al. 2000).
Model ApplicationsA1.15.2
The following RHESSys published model applications were reviewed:
• “Evaluatingexplicitandimplicitroutingforwatershedhydro-ecologicalmodelsofforesthydrol-ogy at the small catchment scale” (Tague and Band 2001a). This study compared two implicit routing routines: (1) the TOPMODEL approach and (2) explicit routing, based on the DHSVM method. These two approaches were applied to Watershed 2, a 60-ha watershed in the H.J. Andrews Experimental forest. This study indicated the advantage of the explicit routing approach to model lateral soil water distribution. The TOPMODEL approach was more sensitive to random error in input data.
• “Simulatingtheimpactofroadconstructionandforestharvestingonhydrologicresponse”(Tagueand Band 2001b). RHESSys was used to explore the implications of road-cut depth and road-drainage patterns on seasonal hydrologic responses, including runoff production, soil moisture and ecological processes, such as evapotranspiration, for a small watershed in the western Oregon Cascades, US. Results showed the potential for an ecologically significant change in soil moisture in the area downslope from the road. These changes are mediated by the drainage patterns associated with roads, specifically whether road culverts serve to concentrate or to diffuse flow. The modelled effects on seasonal watershed outflow were less significant, but did show clear temporal patterns
86
associated with climate pattern, hillslope drainage organization, and road construction. Com-parison between modelled- and observed-outflow response suggested that the model does not yet capture all of the processes involved in assessing the effects of forest road construction.
• “ApplicationoftheRHESSysModeltoaCaliforniasemiaridshrublandwatershed”(Tagueetal.2004). RHESSys was evaluated by comparing model predictions of monthly and annual stream-flow to stream guage data and by comparing RHESSys behaviour to MIKE-SHE. The models were applied to Jameson, a 34-km2 sub-watershed of the Gibraltar watershed, located in the Santa Ynez Mountains, north of the City of Santa Barbara, California. The study period used for model evaluation was the 1974–1980 water years. This study illustrated a trade-off between using addi-tional parameters and storage terms and limiting the need for extensive calibration. The additional parameters in MIKE-SHE, and inclusion of groundwater storage zone, allowed this model to more accurately capture observed streamflow behaviour. The results from the California study appear to have led to inclusion of a second, deeper groundwater drainage system in the RHESSys model (Table 4 and 6; Tague et al. 2004).
• “Impactsoftimberharvestingontheflowregimeofacoastalstream;fromtheheadwaterstotheentire watershed” (Krezek et al. 2007). This model was applied to the Carnation Creek experimen-tal watershed, located on the west coast of Vancouver Island, British Columbia. Streams from the headwaters to the entire watershed were chosen for the analysis. Review of the Carnation Creek study indicates that RHESSys may not have been able to simulate rainfall runoff mechanisms correctly in this coastal watershed. RHESSys tended to simulate substantial overland flow during storm events, whereas, in reality, subsurface preferential flow through root networks is the domi-nant runoff mechanism contributing to peak flow generation (Beckers and Alila 2004). This lack of preferential flow representation is a limitation of most hydrologic models. Preferential flow can be incorporated into models with a predominant groundwater emphasis such as InHM and Hydro-GeoSphere.
• “Scale-dependenceofnaturalvariabilityofflowregimesinaforestedlandscape”(Sanfordetal.2007). A distributed hydrologic model was used to characterize the natural flow regime of water-sheds from first to fifth order within tributaries of the Batchawana River in the Algoma Highlands of central Ontario. A 30-year simulated flow record was used to calculate natural variability in the flow regime. Flow variability under wetter conditions was similar across all watersheds, regardless of scale. Conversely, flow variability under drier conditions was scale-dependent, with smaller wa-tersheds (< 600 ha) showing a large range in variability and larger watersheds (> 600 ha) showing a smaller range in variability, converging toward a constant with increasing area. Suggestions for model improvements were also made, mainly concerning the formulations of decomposition and respiration rates in biogeochemical models.
• “WaterandcarbonfluxesofEuropeanecosystems:AnevaluationoftheecohydrologicalmodelRHESSys” (Zierl et al. 2007). This study investigated whether the regional hydro-ecological simula-tion system RHESSys is a suitable tool for long-term global change impact studies under selected climatic conditions of Europe. A validation of RHESSys was performed using daily, monthly, and yearly data on (1) streamflow and snow cover in five alpine watersheds and (2) water and carbon fluxes at 15 EUROFLUX sites. The analyses confirmed that RHESSys is a suitable tool for studying global change impacts on mountain hydrology. Regarding the simulation of the carbon cycle, this investigation detected both data and model limitations.
• “Spatialpatternsofsimulatedtranspirationresponsetoclimatevariabilityinasnowdominatedmountain ecosystem” (Christensen et al. 2008). RHESSys was used to assess elevation differences in sensitivity of transpiration rates to the spatiotemporal variability of climate variables across the Upper Merced River watershed, Yosemite Valley, California, USA. Low elevations (1200–1800 m) showed little inter-annual variation in transpiration due to topographically controlled high soil
87
moistures along the river corridor. Annual conifer stand transpiration at intermediate elevations (1800–2150 m) responded more strongly to precipitation variability. Transpiration at the high-est elevations (2600–4000 m) showed strong sensitivity to air temperature. Overall, model results suggest elevation differences in vegetation water use and sensitivity to climate were significant and will likely play a key role in controlling responses and vulnerability of Sierra Nevada ecosystems to climate change.
• RHESSyshasalsobeenappliedtothelarge42000km2 Big Thompson watershed in Colorado with alpine rock, tundra, and forested headwater watersheds (Landrum et al. 2002). However, details of this study are lacking in the conference proceedings.
Jointly, these publications indicate that RHESSys is a useful model for addressing complex resource management issues in the Pacific Northwest, including eco-hydrology25 and climate change questions, but that model improvements are perhaps needed in the areas of simulating the effects of forest roads (Tague and Band 2001b), biochemical processes (Sanford et al. 2007), and geochemical cycling (Zierl et al. 2007).
ssArrA1.16
The Streamflow Synthesis and Reservoir Regulation (SSARR26) model was developed by the US Army Corps of Engineers, North Pacific Region to “provide mathematical hydrologic simulations for systems analysis as required for the planning, design, and operation of water control works” (US-Army Corps of Engineers 1991). The SSARR model has been further developed for operational river forecasting and river management activities (US-Army Corps of Engineers 1991).
Model DescriptionA1.16.1
The model is semi-distributed and works with elevation bands for snowmelt calculations. It includes a single soil layer and a single vegetation layer (i.e., overstorey and understorey are not distinguished). The model can operate at time intervals ranging from 0.1 to 24 hours (Tables 4, 6, and 8).
The watershed components of the model are empirical. SSARR computes snowmelt based on a tem-perature-index approach with a correction for rainfall (US-Army Corps of Engineers 1991). Index-based methods are used to simulate precipitation interception, evapotranspiration (Thornthwaite formula for PET combined with soil moisture index to limit actual ET), soil moisture, runoff (empirically derived relationship with soil moisture) and baseflow (deep groundwater reservoir based on infiltration index) (Table 4) (US-Army Corps of Engineers 1991).
Flow is routed through the river system and reservoir component of the model routes from upstream to downstream points through channels, lake storage, and reservoirs under free-flow or controlled-flow conditions (US-Army Corps of Engineers 1991).
The input/output is highly complex (Table 5) and calibration is cumbersome. As of 1991, there is no user support available, and there will be no further program updates or modifications to the model. Overall, the model has limited functionality in a forest management context (Table 6), is typically ap-plied in medium to large watersheds (Table 7) for river forecasting and reservoir management (see below), and is therefore not recommended for forest management applications.
25 Interdisciplinary area linking hydrology with ecological processes (e.g., plant physiology) involved in the water cycle. Eco-hydrology research investigates the effects of hydrological processes on the distribution, structure, and function of ecosystems, and the effects of biotic processes on elements of the water cycle (Baird and Wilby 1999).26 www.nwd-wc.usace.army.mil/report/ssarr.htm
88
Model ApplicationsA1.16.2
The following SSARR published model applications were reviewed:
• “ClimatechangeimpactsonfloodingintheElbowRiverwatershed”(ValeoandXiang2005).The1200 km2 Elbow River watershed, located in the foothills of the Rocky Mountains, was simulated using the SSARR model. Peak flow rates, volumes, and times to peak were documented along with the full hydrographs for both short-term (single events) and long-term (six month) simulation periods. (Valeo and Xiang 2005)
• “Techniquesforpredictionofrunofffromglacierizedareas”(Power1985).TheSSARRmodelwasapplied by the US Corps of Engineers in various watersheds of the Columbia River Basin to simu-late hydrographs from ablation of mountain snowpacks.
• “AnapproachtospatiallydistributedsnowmodellingoftheSacramentoandSanJoaquinbasins,California” (Daly et al. 2000). This study implemented a simple approach to modelling processes of snow hydrology with explicit spatial distribution using SSARR (Daly et al. 2000).
For each of the latter two studies, processes related to snow accumulation, melting, and presence of glaciers (on the Columbia River projects) required modification to SSARR. For parts of the Columbia River, the SSARR model can produce long-term water supply forecasts with accuracy but requires con-tinually adjusting the model to match observed hydro-meteorological conditions. For the upland snow accumulations in Sacramento and San Joaquin watersheds, snow water equivalent was not represented well and requires further work (Daly et al. 2000).
swAtA1.17
The Soil Water Assessment Tool (SWAT) is a watershed-scale model developed to predict the impact of management on water, sediment, and agricultural chemical yields (Gassman et al. 2007).
Model DescriptionA1.17.1
In SWAT, watersheds are divided into multiple sub-watersheds, which are then subdivided into hydro-logic response units (HRUs) (Gassman et al. 2007). The HRUs are areas of homogeneous land use and soil characteristics, and represent percentages of the sub-watershed area that are not considred in a spatially explicitly manner in the model simulation (Gassman et al. 2007). The model considered two soil layers (root zone and unsaturated zone), together with conceptual shallow and deeper aquifer stores, a single vegetation layer, and operates on a daily time step (Tables 4, 6, and 8).
Several options for calculating evapotranspiration can be employed, with the most rigorous one being physically based (Table 4). Evapotranspiration values estimated externally to SWAT can also be input for a simulation run. Estimating snowmelt is based a temperature-index approach within elevation bands, with no apparent correction for rain-on-snow events (Neitsch et al. 2005). Overall, SWAT ranks as a mixed model for simulating SVAT processes and the effects of forest management (Table 6).
Soil moisture redistribution, runoff, and groundwater flow are calculated analytically, and include processes such as preferential (bypass) flow and perched water tables (Neitsch et al. 2005). Groundwater recharge is partitioned between shallow and deep aquifers (Neitsch et al. 2005). The shallow aquifer can support return flow to stream systems and evapotranspiration from plants with sufficiently deep roots (Neitsch et al. 2005). Water that recharges the deep aquifer is assumed lost from the system.
Surface runoff in SWAT is simulated using the SCS Curve Number or Green-Ampt methods (Neitsch et al. 2005). Streamflow routing is analytical and road hydrology is not incorporated in the model (Tables 4 and 6).
89
The SWAT model includes a plant growth component (Table 6), which computes plant growth under optimal conditions (adequate water and nutrient supply and a favourable climate) by simulating the development of leaf area and the conversion of light interception to biomass (Neitsch et al. 2005). Differences in growth between plant species are defined by parameters contained in a plant growth database (Neitsch et al. 2005). The model also contains components that can represent lakes, reservoirs, and groundwater (Tables 4 and Table 7), and simulate nutrient cycling, sediment yield, and fate of pesticides (water quality).
Climatic variables required by SWAT consist of daily precipitation, maximum/minimum air tem-perature, solar radiation, wind speed, and relative humidity. These variables can be either input from observed data or generated in the model. SWAT ranks as being of high complexity (Table 5), which is a function of its wide range of simulation capabilities. GIS-based interfaces are used to facilitiate the set up of SWAT, and provide a straightforward means of translating digital data (land use, topography, and soils) into model inputs. Additional SWAT support tools are continually being developed (Gassman et al. 2007). The model is freeware.
Conceptual diagrams in Neitsch et al. (2005) suggest that SWAT is best applied in gradual terrain with a substantial groundwater component to the overall watershed budget. The model can be applied in rain or snow settings, but the degree-day snowmelt method is not adapted for rain-on-snow (mixed) condi-tions. The model has been applied in small to large watersheds (Table 7).
Model ApplicationsA1.17.2
Gassman et al. (2007) reviewed historical SWAT development as well as a total of 115 model applica-tions in the US, Europe, and worldwide. These studies illustrate the versatility of SWAT in simulating the hydrology of diverse watersheds. Generally, this review illustrated the ability of SWAT to replicate hydro-logic processes at a variety of spatial scales on an annual or monthly basis. However, model performance has been inadequate in some studies, especially when comparisons of predicted output were made with time series of measured daily flow. These shortcomings in simulating watershed hydrology at relatively short time scales (days) is attributed to the simplified representation of runoff processes with the SCS curve number method. One drawback of the HRU approach is that explicit spatial representation of riparian buffer zones, wetlands, and other features is not possible.
The following SWAT published model applications were reviewed in more detail:
• “HydrologiccalibrationandvalidationofSWATinasnowdominatedRockyMountainwater-shed, Montana, U.S.A.” (Ahl et al. 2008). This study employed 54 HRUs to simulate the 1993–2002 streamflow of Tenderfoot Creek, a high-elevation watershed with approximately 85% coniferous forest cover. More than 70% of annual precipitation in the 2251-ha watershed falls as snow (Ahl et al. 2008). Seasonally, SWAT performed well during the spring and early summer snowmelt runoff period, but was a poor predictor of late summer and winter baseflow. Model sensitivity to the sur-face runoff lag parameter reflected the influence of frozen soils on runoff processes.
• Hydrologiccomparisonbetweenaforestedandawetland/lakedominatedwatershedusingSWATin Ontonagon River basin of northern Michigan (Wu and Johnston 2008). Model calibration and validation was satisfactory. Streamflow simulation discrepancies occurred during summer and fall months, and dry years. Snowmelt parameters were not transferable between adjacent watersheds. The wetland/lake-dominated watershed exhibited substantially different temporal flow trends than did the adjacent forested wetland, with lower peak flows in April and higher baseflows during sum-mer months.
90
Overall, results indicated that SWAT can provide reasonable predictions of annual, monthly, and daily streamflow from forested watersheds. Further model refinement is needed to improve representa-tion of snow cover and melt and to improve performance during the baseflow period (summer and fall months). These findings are echoed by Krysanova and Arnold (2008) as part of a special issue on ad-vances in ecohydrological modelling with SWAT. This special issue also suggests that many studies have demonstrated the robustness of SWAT in simulating sediment and nutrient concentrations and loads.
The SWAT agricultural model is being adapted to the boreal forest environment under the lead of re-searchers from Lakehead University (Ontario), as part of the Forest Watershed and Riparian Disturbance (FORWARD) project27, with research watersheds in the Swan Hills on the western Canadian Boreal Plain (AB) and in the Legacy Forest on the Central Canadian Boreal Shield (ON). SWAT-related FORWARD projects include:
• “Plantgrowthsimulationforlandscape-scalehydrologicalmodeling”(Kiniryetal.2008).Thisproject incorporated a process-oriented plant model for simulating crops, perennial grasses, and woody species into the daily time-step hydrological transport model, SWAT.
• “Incorporatingwaterquantityandqualitymodellingintoforestmanagement”(Li,X.etal.2008b).Millar Western Forest Products Ltd. (MWFP) has the right to harvest trees, grow trees, manage the forest, and plan activities that assure forest productivity and industry profitability without jeopar-dizing the quality of the environment. Thus, as part of obtaining provincial government approval, the company has to submit a Detailed Forest Management Plan that includes a comprehensive assessment of the environmental implications of forestry operations and the mitigation of im-pacts. A hybrid modelling tool is proposed that relies on inexpensive remote-sensing data, with few ground-truthing requirements, to model streamflow, suspended solids, and nutrients in streams on the Boreal Plain. Incorporating modelling tools into the MWFP planning process provided MWFP with additional strategies to operate in an environmentally sensitive manner.
The latter study further strengthens the notion that SWAT may be a useful tool for physical and chemical water quality assessments.
ubc-uf Peak flow modelA1.18
Development of a hydrologic process model to predict peak flow changes for mountain pine beetle-affected areas in BC is currently ongoing as a collaborative effort (under the lead of Dr. M. Weiler) between the University of British Columbia (UBC) and the University of Freiburg (UF), Germany.The model description below is largely based on preliminary documentation that was kindly provided by the model developers.
Model DescriptionA1.18.1
The model is designed to simulate ungauged watersheds. However, the model could also be calibrated against streamflow data as well as SWE data. The model is fully distributed, but does not explicitly include soil or vegetation layers. There will be two versions of the model available. The first version estimates peak flow changes due to land cover changes on a mesoscale using spatial information at 25-m resolution. The second version will model peak flow changes due to land cover changes on a macroscale by upscaling information (elevation and dominant runoff processes) from 25- to 400-m grid cells. The models will operate at a daily time step (Tables 4, 6, and 8).
27 http://forward.lakeheadu.ca/index.htm
91
The model consists of three major modules: climate input, runoff generation, and land cover modi-fication. The climate input module determines the mean annual snowmelt as well as the maximum rainfall, based on long-term climatic averages. It also takes into account the reduction of rainfall due interception processes by forest vegetation. This information is then used to determine the timing and magnitude of peak flows for every third-order watershed. The runoff module delineates Dominant Run-off Processes (DRP) such as channel interception, Horton overland flow, saturation overland flow, and shallow surface flow. In the land cover modification module, total runoff is calculated based on rainfall and (or) snowmelt inputs that account for vegetation modification resulting from man-made and nat-ural disturbances, and are based on the contribution from each runoff generation process. The following paragraphs give a more detailed description of each module.
Evapotranspiration is calculated using factor reductions for forest, grassland, and clearcut. Canopy rainfall interception is included, while snow interception is under development. Snowmelt is calculated using an empirical degree-day method (Kuusisto 1980; Rango and Martinec 1995). To characterize these processes, the model relies on experimental data regarding forest canopy interactions with precipitation processes (Berris and Harr 1987; Cheng 1989; Winkler 2001; Maloney et al. 2002). Overall, the model ranks as empirical with respect to the simulation of SVAT processes (Table 6).
Runoff-generation processes considered are Horton infiltration excess overland flow, saturation excess overland flow, shallow subsurface flow, and channel interception (Table 4). For the delineation of each DRP, the following hydrological indices/parameters have been created: overland flow distance, vertical distance to channel network, average gradient to the stream, and topographic wetness index. The latter index takes into account the upslope drainage area and the average gradient to the stream. In compari-son to the wetness index from TOPMODEL (Beven and Kirkby 1979), which uses the upslope area and the local slope, the local slope is replaced by the average gradient to the stream for this wetness index to account for more realistic conditions for subsurface flow. Road hydrology and runoff in burned areas can be represented in a simplified fashion as Hortonian overland flow areas and as information on con-nectivity to the stream (Table 6).
The model extracts all grid cells within each third-order watershed (See section 2.3.3 for definition of stream order) in the Fraser River watershed and calculates the average melt-time series for each water-shed. To apply the model to larger watersheds, routing effects have to be explicitly included in the model. The derivation of a simple and robust flow-routine module is one of the current key activities.
In the current model application, the climatic database includes the mean monthly climatic precipi-tation and temperature data derived from the PRISM methodology for a 400-m grid spacing for the province of BC (Spittlehouse 2006). For modelling, the monthly values have been interpolated to daily temperature and precipitation for each grid cell. The model appears to be of medium complexity (Table 5). A database of BC containing the information required to run the model at a 400 m grid is provided (DEM, climate, forest cover, pine cover (%), DRP, and MPB). However, some data processing and GIS analysis may be required.
Model development has focused on assessing peak flow changes, but there may be additional output capabilities (Table 2.20, Appendix 2). The main advantages of the UBC-UF peak flow model are the relative simplicity of its philosophy and the fact that ready map databases will be provided to generate model inputs. The main drawbacks of the model are that it is still under development and has, to date, undergone limited testing(see below). The UBC-UF peak flow model has been developed specifically for use in BC, using provincial databases. The model should be applicable to third-order streams (typically small watersheds) and larger watersheds across BC (Table 7).
92
A model that would presumably be similar in philosophy, but focused on low flow prediction, is under development as a collaborative UBC-UF effort. For the purposes of this review, model names are considered tentative.
Model ApplicationsA1.18.2
The 25-m resolution model has been tested on selected watersheds in the Kootenay and Prince George regions of BC (Weiler et al. In prep.). However, specific model applications cannot be reviewed at this point and additional model testing is recommended.
Overall, the model takes an innovative approach towards assessing peak flow changes due to forest disturbances across BC; it attempts to bridge the gap between the high data requirements of physically based models and the limited data that is typically available in operational management situations by making optimal use of existing databases and combining these with simplified representations of dom-inant watershed processes. The main drawback of the model is that it is still under development and has seen limited testing to date, i.e., its capabilities in forest management need to be ground-truthed. The UBC-UF peak flow model has been developed specifically for use in BC (as per reliance on provincial databases) and would require modification for application in Alberta.
ubcwmA1.19
The UBC Watershed Model (UBCWM) was originally developed for daily streamflow forecasting on the Fraser River system in British Columbia (Quick et al. 1995). The UBCWM is being used operationally for forecasting daily streamflows in sub-watersheds of the Fraser River system that are subject to snow-melt floods from the mountain snowpacks of the Coast, Columbia, and Rocky Mountains in British Columbia.
Model DescriptionA1.19.1
The model was originally designed for forecasting runoff from mountain watersheds and, for this reason, the model is divided into area-elevation bands (Quick et al. 1995). The model incorporates a conceptual soil moisture storage reservoir and a single canopy layer (Tables 4, 6, and 8).
Potential evapotranspiration is calculated in an empirical fashion using maximum daily temperatures. Snowpack accumulation is estimated based on temperature and elevation. Snow melt is modelled using an analytical approach (Table 4) that includes energy balance components expressed in terms of com-monly available meteorological data variables (precipitation, maximum and minimum temperatures) based on research by the US-Army Corps of Engineers (Assaf 2007). Overall, UBCWM ranks as a mixed model for simulating SVAT processes.
Infiltration and runoff are simulated using empirical equations (Table 4). The partitioning of snow-melt and rain between runoff components (very slow, slow, medium, and fast) is controlled by the soil moisture model (Quick et al. 1995). Water allocated to each component is subject to a routing proced-ure based on linear reservoirs (Quick et al. 1995). The UBWCM can be used with a companion routing model (UBC Flow Model) to examine sub-watershed flows (M. Schnorbus, pers. comm., May 2009). This capability can be used to model, in a simple way, the response of large heterogeneous watersheds as an amalgamation of sub-watersheds connected by a routing network. The quick- and medium-runoff components use a set of reservoirs, while the slower components are represented by a single reservoir (Quick et al. 1995). Each component is summed to produce runoff for each elevation band, and for the watershed at each time step. The model has no provision for simulation of road hydrology (Table 6).
93
The model is able to account for glacial melt, lake storages, and lake routing (Tables 4 and 7). The UB-CWM is currently used as an operational flow forecasting tool by the BC River Forecast Centre28 and BC Hydro. As such, it has seen wide use across BC and there is also consulting experience with the model. Micovic and Quick (1999) used the UBCWM and data from a number of watersheds in BC to develop regional parameter sets to predict streamflow in ungauged watersheds in BC. The study demonstrated that, as long as the precipitation inputs are sufficient and that the impermeable fraction of the watershed can be estimated, the standard parameter set could be used to obtain reasonable model results for water-sheds that were not used to generate the parameter sets (Micovic and Quick 1999).
The model operates from meteorological inputs of daily maximum and minimum temperatures and precipitation. Overall, the UBCWM ranks as a model of medium complexity (Table 5). Each elevation band is defined according to the following characteristics: band area, mean elevation, forested fraction, density of the forest canopy, impermeable fraction, and the orientation index. The model is freeware. However, it appears that with the retirement of the model developer (Dr. M. Quick, UBC) the model is no longer being actively maintained.
To date, the model has been successfully applied to small to medium mountainous watersheds in rain, snow, and mixed regimes (Table 7). The model has seen some research applications in a forest manage-ment context (see below), which have indicated the following advantages and disadvantages (Dr. Y. Alila, pers. comm., Jan. 2009):
• Modestoveralldatarequirements.Themodelcantypicallybesetupandcalibratedtostreamflowdata in about a 2-week period.
• Simplifiedforestcoverrepresentation(specificationofcrownclosure/forestdensity),andtheuseofelevation bands limit flexibility in simulating forest management scenarios.
Model ApplicationsA1.19.2
The following UBCWM applications were reviewed:
• “Acomponentbasedwaterchemistrysimulatorforsmallsubalpinewatersheds”(HudsonandQuick 1997). This paper presents the development and testing of a component-based water chem-istry simulator and is based on a comprehensive study of water chemistry processes conducted in small subalpine watersheds in the Southern Interior of British Columbia. The UBCWM was used as the basis for the model. Empirical relationships between chemical concentrations in forest-floor runoff, groundwater and interflow, and associated hydrological parameters and outflow rates were developed to model streamflow chemistry. Model results were in close agreement with observed streamflow chemistry.
• Effectsofforestharvestingandregenerationonpeakstreamflowinacoastalwatershed(Hudson2000). The UBCWM was used at Russell Creek, a fourth-order tributary sub-watershed of the Tsitika River on northeastern Vancouver Island that drains an area of 30.8 km2. The model used rainfall intensity to predict peak flows and demonstrated a decline in peak flows concurrent with a decline in equivalent clearcut area.
• “AssessingsnowpackrecoveryofwatershedsintheVancouverForestRegion”(Hudson2000).TheUBCWM was used to examine the effects of regenerating forest canopy conditions on snow accu-mulation and melt. The study was conducted at upper Gray Creek on the Sunshine Coast, at about 1000-m elevation in the Mountain Hemlock biogeoclimatic zone, and consisted of an analysis of snow course data collected between January 1993 and May 1997 in stands at different stages of regeneration (Hudson 2000).
28 www.env.gov.bc.ca/rfc/
94
• “Hydrologicrecoveryofthepeakflowregimeintwosnowdominatedmeso-scalewatershedsinthe Southern Interior of British Columbia: An investigation using a conceptual hydrologic model” (Luo and Alila 2006). UBCWM was used at two interior snow-dominated watersheds, Bellvue Creek (73 km2) and Whiteman Creek (112 km2) in the Okanagan Basin, to quantify the effects of total clearcut of an entire watershed and investigate the nature of the relationship between these effects and hydrologic recovery over time. The long-term simulation of each watershed was based on available meteorological data (Bellvue Creek, 33 years; Whiteman Creek, 76 years).
• “ApplicationofGISinthedeterminationofprobablemaximumflood”(Micovic2006).TheUBCWatershed Model (UBCWM) was used to derive the PMF estimates for several BC Hydro dams in southwestern BC. The Strathcona Dam watershed (1193 km2) and the Ladore Dam watershed (245 km2) within the Campbell River System were used for this simulation
• “HydrologicresponsetoscenariosofclimatechangeinsubwatershedsoftheOkanaganbasin,British Columbia” (Merritt et al. 2006). A model of the 8000-km2 Okanagan Basin in the Southern Interior region of BC was constructed to explore scenarios of climate change on the volume and timing of runoff from tributaries, and the availability of water resources for the various stakehold-ers in the watershed.
• “Delineatingthelimitsonpeakflowandwateryieldresponsestoclearcutsalvagelogginginlargewatersheds” (Alila and Luo 2007). In this study, the UBC Watershed Model was used to predict the effects of forest harvesting on streamflow characteristics. Because of the lumped nature and the av-eraging of processes over time and space associated with such a conceptual model, the investigation in this study was restricted to simulating the effects of clearcut logging over the entire watershed, excluding the effects of forest roads, spatially distributed logging, and tree regrowth.
Based on the above literature, the UBC Watershed Model is useful for forest management applica-tions, including examining hydrologic recovery in harvested stands and developing a link between equivalent clearcut area (ECA) and peak flows. In addition, the model has proven to be successful at modelling snow accumulation and melt (Hudson 2000).
Environment Canada has used climate change scenario information in conjunction with the UBC Watershed Model to produced preliminary projections of the potential changes in snowpack and runoff in the Capilano Watershed (Hutchinson et al. 1999).
In the 1970s, the model was also tested and adopted by the Prairie Provinces Water Board for moun-tain snowmelt forecasting in the Saskatchewan River system headwaters (Quick and Pipes 1977).
vicA1.20
The Variable Infiltration Capacity (VIC29) macroscale hydrologic model solves full water and energy bal-ance equations (Liang et al. 1994). VIC is a research model and, in its various forms, has been applied to many watersheds including the Fraser River, Columbia River, the Ohio River, the Arkansas-Red Rivers, and the Upper Mississippi Rivers. The application of the VIC model to climate change studies in cold mountainous terrain has been reviewed by Werner and Bennett (2009).
Model DescriptionA1.20.1
The VIC model represents surface and subsurface hydrologic processes on a spatially distributed (grid cell) basis. In typical applications, grid cells have ranged in resolution from 1/8th to 2 degrees per side, although model applications are underway at a higher resolution of 1/16th degree (see below). With this 1/16th-degree grid size (e.g., about 7 km in the longitudinal direction by 3.5-km north–south in
29 http://www.hydro.washington.edu/Lettenmaier/Models/VIC/
95
southern BC), VIC can only be applied to watersheds on the order of 500 km2 (based on a minimum of 15 grid cells) or larger. The minimum planning scale in BC (single grid cell) is about 30 km2. Sub-grid scale variation in vegetation and soil characteristics can be approximated by partitioning grid cell areas to different vegetation and soil classes using a statistical representation. The model includes three soil layers and a single vegetation layer and operates at a sub-daily time step (Tables 4, 6, and 8). Topographic variability is represented in the VIC model through the use of elevation bands; up to five elevation bands can be applied in VIC to represent the changes in elevation across the grid cell and to determine sub-grid scale variations in precipitation.
VIC is a physically based model with respect to simulating SVAT processes (Table 6) that include evaporation from the soil layers, evapotranspiration from vegetation, canopy interception evaporation, and snowmelt.
Infiltration and runoff mechanisms in the VIC model are empirical (Table 4) due to the macroscale nature of the model. Infiltration is simulated using the variable infiltration capacity curve while surface runoff, interflow, and baseflow (ARNO method) are incorporated for simulating total runoff from the grid cell (Liang et al. 1994). Quickflow occurs when precipitation exceeds the variable infiltration capac-ity and then runs off. VIC does not account for interflow. The model runs one grid cell at a time over a desired time period to produce time series of runoff, baseflow, evaporation, and other physical variables for each grid cell. Quickflow and baseflow are routed to produce streamflow at points of interest in the watershed. There is no horizontal transfer of water from grid cell to grid cell and therefore a routing model must be used as a post-processing tool to calculate streamflow as a composite of both surface runoff and baseflow. The model has no specific provision for including road hydrology (Table 6).
Recent work has focused on improving the VIC model for use in northern regions, including representing frozen soils; spatial variability in snow and frost, lakes and wetlands; and blowing snow (Cherkauer and Lettenmaier 1999; Cherkauer et al. 2003; Bowling et al. 2004). A regional (deep) groundwater component for VIC is apparently under development (Werner and Bennett 2009). A glacial representation is lacking at present (Tables 4 and 7), however parameterization of glacier melt contribu-tions to streamflow may be added to the VIC model.
Required meteorological inputs include temperature and precipitation at a minimum while solar radiation is needed for energy balance calculations (e.g., snowmelt) and can be estimated within the model. Overall, the model is of high complexity (Table 5), requires substantial data collection, pre-pro-cessing, and GIS analysis. VIC is freeware. No technical support is available, unless specific arrangements have been made.
VIC is a macroscale model and is therefore not useful for typical site-specific of small watershed for-est management applications; however it does have potential for regional-scale planning and associated policy development. However, its inclusion in this synthesis is warranted as a possible tool for assessing effects of climate change and large-scale disturbances (e.g., mountain pine beetle infestation and salvage harvesting).
Model ApplicationsA1.20.2
This section is based on the review of VIC applications for climate change applications in cold mountainous watersheds by Werner and Bennett (2009). VIC has been run at various scales, ranging from 2 to 1/8th degrees in the Columbia and Colorado River Basins (Hamlet and Lettenmaier 1999, 2000; Mattheussen et al. 2000; Nijssen et al. 1997, 2001; Christensen et al. 2004). For the Columbia River Basin, Nijssen et al. (1997) compared data and simulation results for nine streamflow gauges over two separate 12-year periods. Mean annual runoff was generally well predicted, while seasonal flow volumes were
96
more variable. This discrepancy was partially caused by the resolution of the grid cells (1 degree), which did not accurately account for the variability in elevation.
For a study focused on the entire western US, an analysis of snow processes and VIC results showed broad general agreement between observed trends in snow water equivalent (SWE) for three different times: (1) SWE measured on April 1, (2) actual measured spatial patterns of snow accumulation and melt for the period from 1950–1997, and (3) model simulations over the same period (Hamlet and Lettenmaier 2005). There were differences in at-site measurements and simulations of SWE, primarily due to differences in measurements versus model scales and spatial variability of precipitation relative to grid-based precipitation estimates. When the VIC snow model is run with detailed site data, the model closely represents daily snow accumulation measurements.
In general, VIC is suitable for simulating watershed hydrology for large river basins in the PNW. How-ever, there are a number of barriers to successful implementation, including the particular challenge of accurately representing spatial distribution of meteorological variables in complex terrain. This chal-lenge is common to all model applications in mountainous settings, and not VIC specific (i.e., similar issues are noted for WATFLOOD; Section A1.23.2).
Currently, the VIC model is being applied in an MPB context for the Fraser River, and in a climate change context in the Columbia River, the Peace River, and the Campbell River by the Pacific Climate Impacts Consortium with the support of the University of Washington and the MOE River Forecast Centre.
wasim-ethA1.21
The Wasserhaushalts-Simulations-Modell (WaSiM-ETH30) is a fully distributed, physically based model for estimating climate change impacts for subalpine and alpine regions. Compared to PREVAH, WaSiM-ETH is a physically based runoff model and has more flexibility in separating surface runoff from interflow, allowing, in general, a better reproduction of flood events (Gurtz et al. 1999). The application of WaSiM-ETH to climate change studies in cold mountainous terrain has been reviewed by Werner and Bennett (2009).
Model DescriptionA1.21.1
The spatial and temporal discretization of the model can accomodate grid cell sizes of virtually any dimensions (from centimetres to several kilometres), and time steps between 1 minute to several days (Werner and Bennett 2009). The model includes a layered soil model and layered vegetation (Tables 4, 6, and 8).
The original model version incorporated a set of interpolators for meteorological data; a set of evap-oration algorithms (with the most rigorous one being physically based, Table 4); several methods for simulating snow accumulation and melt (including the empirical degree-day method and a modified (analytical) energy balance method); an interception module; and topography-driven correction of radiation, temperature, and precipitation (Schulla and Jasper 2000). Overall, the model ranks as an ana-lytical model for simulating SVAT processes (Table 6).
In the original model, runoff generation was based on the TOPMODEL approach (Beven and Kirkby 1979). Improvements in recent model versions include the following:
• LayeredsoilmodelbasedontheRichards’equation(i.e.,aphysicallybasedapproach;Table4)• Multi-layergroundwatermodel• Layeredvegetation
30 www.wasim.ch/en/index.html
97
• Irrigationmodule• Dynamicplantphenology• Glaciermodel(accumulationofsnow;meltofsnow,ice,andfirn)• Interfaceforon-line-couplingwithothermodels(climatemodels,groundwatermodels.etc.)• Lakemodel(includinginteractionwithgroundwater)• Surfacedischargerouting
The river (channel) routing accounts for reservoirs and lakes as well as branches and artificial abstrac-tions using time-dependent dynamic abstraction rules (Tables 4 and 7) (Werner and Bennett 2009).
WaSIM-ETH requires significant meteorological-driving data (air temperature, precipitation, wind speed, global radiation, vapour pressure or relative air humidity, and relative sunshine duration) to run in a fully physically based configuration. WaSiM-ETH can also be coupled to regional climate models. Overall, this model ranks as being of high complexity (Table 5). Graphical- and command line-driven tools are available for pre- and post-processing of driving data and outputts. According to the WaSIM-ETH website, as of 2009, there are an estimated 50 institutions currently applying the model (universities, governmental agencies, and consulting engineers). WaSiM-ETH is freeware.
The main drawbacks of the model would appear to be its relative complexity and associated high meteorological input requirements, and that model experience, to date, only resides in Europe. However, the developer (Dr. J. Schulla) will start work for the Pacific Climate Impacts Consortium in July 2009, bringing expertise with the WaSim-ETH model to Western Canada.
Model ApplicationsA1.21.2
The following WaSiM-ETH published model applications were reviewed:
• “Thehydrologicalroleofsnowandglaciersinalpineriverbasinsandtheirdistributedmodel-ling” (Verbunt et al. 2003). The WaSiM-ETH model was applied to three Swiss high-alpine river watersheds with different portions of glacierized areas. Continuous discharge simulations were performed at a spatial resolution of 100 m and at a temporal resolution of 1 hour for the period from 1981–2000 and compared with hourly discharge observations measured at the watershed outlets (Verbunt et al. 2003).
• “Differentialimpactsofclimatechangeonthehydrologyoftwoalpineriverbasins”(Jasperetal.2004). For the Thur watershed (1700 km2) and the Ticino watershed (1515 km2), possible future changes in the natural water budget relative to the 1981–2000 (Thur) and 1991–2000 (Ticino) baselines were investigated by forcing WaSiM-ETH with a set of climate scenarios for monthly mean temperature and precipitation.
• “Distributedhydrologicalmodellingofaheavilyglaciatedalpineriverbasin”(Kloketal.2001).Inthis study, a glacier model component was successfully integrated into WaSiM-ETH and PREVAH to simulate the discharge of a heavily glaciated watershed. The model was tested on a high-alpine sub-watershed of the Rhone River (central Switzerland) of which 48% is glaciated. Continuous dis-charge simulations were performed for the period 1990–1996 and compared with hourly discharge observations (Klok et al. 2001).
• “HydrologicsimulationsintheRhinebasindrivenbyaregionalclimatemodel.”(Kleinnetal.2005). This article describes a model chain for studying streamflow responses to climate variations and anthropogenic climate change. The model chain was developed for the Rhine River upstream of Cologne, a 145,000-km2 river system in Central Europe, north of the Alps. It encompassed a regional climate model (RCM) at grid spacings of 56 and 14 km, and the distributed WaSiM-ETH model with a grid spacing of 1 km. The model chain was found to reproduce observed month-to-month variations of winter precipitation and streamflow. This result provides confidence that
98
the model chain is able to represent key processes related to streamflow variations in response to climate variations and climate change.
WaSiM-ETH, including the glacier model component, accurately simulated the seasonal variations in discharge, the diurnal discharge fluctuations, and the time at which the melt period starts and termin-ates in the Klok et al. (2001) study. A drawback of the model is that, in autumn, the discharge volumes and the runoff peaks are often overestimated. The overestimation could be due to the passageways in the glacier that are narrowing or to the method that is used to calculate the melt rate.
The model has also been applied for modelling the impacts of land use and drainage density on the water balance of a lowland–floodplain landscape in northeast Germany (Krause et al. 2007) and for investigating the effects of land use change on hydropower (Verbunt et al. 2005).
Overall these studies suggest that WaSiM-ETH can be successfully applied in small to large watersheds and in both complex mountainous settings (including watersheds with glacial melt contributions) as well as in more gradual terrain (Table 7). However, forest hydrology specific aspects of the model (e.g., canopy interaction with precipitation) have not been tested.
water balance model by QuAlhymoA1.22
“The Water Balance Model (powered by QUALHYMO) is a public domain, on-line decision support and scenario modeling tool for promoting rainwater management and stream health protection through implementation of “green” development practices”31. The model is continually being modified to meet the needs of users. As such, it has become a widely used and accepted tool for innovative stormwater management.
The original version of the Water Balance Model allowed users to evaluate the effectiveness of site planning that incorporates source controls on stormwater management (such as absorbent landscap-ing, infiltration facilities, green roofs, and rainwater harvesting) and achieving development targets for rainwater capture and runoff control under different land uses, soil, and climate conditions.
The Water Balance Model powered by QUALHYMO allows users to simulate four situations (site surface alteration, site controls on baseflow discharge, detention pond storage, and stream erosion) that integrate the site with the watershed and the stream. These situations improve the definition and ap-plication of source controls for rainwater runoff volume and rate reduction to match pre-development baseline values and post-development conditions.
As a valuable stormwater management tool, the Water Balance Model, powered by QUALHYMO, is widely used across Canada (links provided under the BC website referenced above will direct the in-terested reader to other provinces). However, the model has no apparent use in a forest management context. As such, the model is not presented in Tables 4 to 8.
wAtflood, clAss, and meshA1.23
The distributed hydrologic model, WATFLOOD32, developed at the University of Waterloo, has been applied to a study floods, climate change impacts, and environmental impact studies (Kouwen 2008). The application of WATFLOOD to climate change studies in cold mountainous terrain has been reviewed by Werner and Bennett (2009).
31 http://bc.waterbalance.ca/32 www.watflood.ca/
99
Model DescriptionA1.23.1
WATFLOOD uses Grouped Response Units (GRUs), in which process parameters are tied to land cover, and land cover mixes can vary from watershed element to watershed element (Werner and Bennett 2009). The GRU approach helps to maintain computational efficiency. WATFLOOD is insensitive to grid size providing there are sufficient grids to adequately represent the drainage system and the vari-ability in the meteorological data (Kouwen 2008). It has been used with grid sizes from 1 to 45 km2 and for watershed areas from 15 to 1 700 000 km2 (Kouwen 2008). WATFLOOD includes a four-zone model representing the land surface, the upper subsurface zone (saturated and varying depth), the unsaturated zone, and a saturated lower zone. By chaining up to 100 annual events at an hourly time step, continuous simulation can be approximated (Tables 4, 6, and 8).
Evapotranspiration is calculated using empirical methods while snowmelt can be calculated using either a temperature-index approach (Table 4) or a hybrid (analytical) radiation-temperature approach, although this latter option does not appear be available to model users (Kouwen 2008). Overall WAT-FLOOD ranks as a mixed model for simulating SVAT processes (Table 6). WATFLOOD also includes a crude glacier melt model (Tables 4 and 7).
Infiltration is calculated using a Green-Ampt infiltration model (analytical), while stormflow is gener-ated using a Hortonian runoff model. River network routing options include storage routing, a coupled lower zone-wetland-stream routing model, a coupled lower zone-prairie coulee-stream routing model, lake routing, and reservoir operating rules (diversions) (Kouwen 2008). The groundwater component of the model is only conceptual (Tables 4 and 7).
At a minimum, the model requires temperature and precipitation as meteorological inputs. Overall, WATFLOOD ranks as being of relatively high complexity (Table 5). WATFLOOD has a GUI pro-grammed in MS Visual Basic for Windows and can be linked to the Green Kenue GUI that is also used for HBV-EC (Section A1.6). WATFLOOD emphasizes making optimal use of spatially referenced (GIS) information such as remotely sensed data, radar-rainfall data, LANDSAT, or SPOT land use and/or land cover data.
Known limitations of WATFLOOD include its simplified snowmelt calculations. More rigorous routines have not been incorporated as “this would significantly complicate the model and require con-siderably more detailed information about the spatial variations of terrain, aspect, vegetation cover, and meteorologic conditions” (Kouwen 2008; pages 2–26). As such, the model is best applied in gradual ter-rain (small to large watersheds) under either rain or snow conditions (Table 7) but not in mixed settings as the temperature-index approach does not handle rain-on-snow events.
Some of the above-mentioned limitations of WATFLOOD can be overcome with the WATCLASS model, which combines the strength of WATFLOOD in lateral water routing with those of the Can-adian Land Surface Scheme (CLASS) in simulating vertical water and energy fluxes (Soulis et al. 2000). CLASS is a model of similar philosophy and application scale as VIC (Section A1.20), and originally developed for use with the Canadian Global Climate Model (GCM) to provide feedback on parameters such as surface albedo and radiative and turbulent energy fluxes (Verseghy 1991; Verseghy et al. 1993). In WATCLASS, the vertical water flux calculations of WATFLOOD are replaced by the more physically based methods of CLASS, while the lateral routing routines of WATFLOOD are retained (Soulis et al. 2000). In 2008, a new model known as “Standalone MESH” was developed from WATCLASS, together with a GUI called ParaMESH33. MESH output can also be analyzed with the Green Kenue GUI.
33 http://halfront.wxe.sk.ec.gc.ca/html/documents/store/1_0_sa_MESH.html
100
Both WATCLASS and MESH are being developed and maintained by Environment Canada. Use of WATCLASS has not been widespread, but this may change with development of MESH and ParaMESH. The WATFLOOD model was reviewed in this synthesis.
Model ApplicationsA1.23.2
This section is largely based on the review of WATFLOOD applications for climate change applications in cold mountainous watersheds by Werner and Bennett (2009). WATFLOOD was used for a ground-water separation study in Boreal wetland terrain to compare model-derived estimates of groundwater contributions to streamflow to data for five watersheds along the lower Liard River near Fort Simpson, Northwest Territories. The study area (approx 5000 km2) was characterized by meandering streams, extensive peat lands (bogs and fens), and discontinuous permafrost and was composed of roughly 25% deciduous, 20% coniferous, 35% transitional, and 20% wetland. The model was run on an hourly time step for the spring and summer periods of 1997–1999. Overall, the groundwater proportioning in WAT-FLOOD was reasonably representative of groundwater volumes determined from isotope separation (Stadnyk et al. 2004).
WATFLOOD has also been applied in two mountainous, snowmelt-dominated watersheds in BC: the Columbia and Peace rivers. For the Columbia River study, WATFLOOD was applied at a 1-km grid cell size (Bingeman et al. 2006). Model outputs agreed reasonably well with the long-term data record (91 years) for the Columbia River at Nicholson. Errors during the summer period appeared to occur when rainfall events supplemented peak streamflow caused by melting snow and ice. Flood frequency curves for Mica Dam were similar for measured and modelled flows, except for an over-estimation of the high-est flows. Model estimates of snow volume were made for 15 snow course locations in the Columbia River basin in BC, and differences between measured and modelled values ranged from -27% to 240%. The model tended to underestimate high snow water equivalent (SWE) and overestimated the low SWE, while mean SWE was modelled accurately. The differences were attributed to: land classification errors, elevation differences, temperature, and precipitation modelling errors and other microclimate effects inherent to the specific location of the snow survey station. These issues are similar to those noted for VIC (Section A1.20.2) and illustrate the difficulties that must be overcome in large-scale model applica-tions (Werner and Bennett 2009).
For the Peace River watershed, simulations using a 45 x 45-km grid were run for 24 years and com-pared to 14 streamflow observation stations (Toth et al. 2006). Errors in the simulations were greatest near the mountainous headwaters. Errors decreased for higher-order streams, as the quality of climatic input data increased. Errors during the snowmelt period highlight the difficulties of mesoscale model-ling of snow interception/sublimation, translocations of blowing snow between watersheds, the effect of slope aspect on melt timing, and issues with parameterizing models for snowmelt infiltration into seasonally frozen soils. WATFLOOD does not specifically address any of these factors, which may be important in large cold-regions watersheds (Werner and Bennett 2009).
wePPA1.24
The USDA’s Water Erosion Prediction Project (WEPP34) model simulates erosion based on stochastic weather generation, infiltration theory, hydrology, soil physics, plant science, hydraulics, and erosion mechanics (Flanagan and Livingston 1995). WEPP was developed by the USDA National Soil Erosion Research Lab (NSERL).
34 www.ars.usda.gov/Research/docs.htm?docid=10621
101
Model DescriptionA1.24.1
WEPP is a semi-distribted model that simulates the conditions that impact soil erosion for hillslopes or small watersheds (which are subdivided into hillslopes). The model includes multiple soil horizons, a single canopy layer, and a daily time step for predicting evapotranspiration and percolation (Tables 4, 6, and 8).
The water balance component of WEPP is based on algorithms developed for the SWRRB (Simulator for Water Resources in Rural Basins) model, with modifications to improve simulating rainfall intercep-tion, soil drainage, and soil evaporation (Flanagan and Livingston 1995). Evapotranspiration equations in the model rank as analytical (Table 4). Snowmelt calculations are empirical but utilize temperature, radiation, vapour transfer, and precipitation through various coefficients. Overall, WEPP ranks as a mixed model for simulating SVAT processes (Table 6).
Infiltration, drainage and subsurface runoff are simulated in an empirical fashion (Table 4), using storage routing techniques to predict flow through each soil layer in the root zone, based on water inputs (rainfall, snowmelt), soil moisture storage and evapotranspiration (Flanagan and Livingston 1995). Initial conditions for infiltration (Green-Ampt model) are defined by soil moisture content in the upper layer. Drainage below the root zone is considered lost to a lower layer and is removed laterally.
Precipitation is routed to Hortonian overland flow when rainfall intensity exceeds soil infiltration capacity. Surface runoff is characterized by depression storage, which is directly related to soil surface micro-topography. Hillslope runoff occurs once storage is satisfied and is simulated using kinematic wave equations. At the watershed scale, the model supports linking hillslope profiles to channels and im-poundments (e.g., farm ponds, terraces, culverts, filter fences, and check dams). The WEPP model does not explicitly include hydrodynamic channel-network flood-flow routing (Conroy et al. 2006). WEPP calculates the peak runoff rate at the channel (sub-watershed) or watershed outlet using either a modi-fied version of the Rational equation (similar to that used in the EPIC model) or the method used in the Chemicals, Runoff, and Erosion from Agricultural Management Systems (CREAMS) model (Flanagan and Livingston1995).
The WEPP model contains sediment erosion and yield-prediction capabilities for hillslopes and roads and a crop-growth model (Tables 6 and 8). The model calculates the average sediment yield for each hillslope, and simulates channel soil detachment, sediment transport and deposition, and may include impoundments to remove sediment from the channel flow (Flanagan and Livingston 1995). The US Forest Service has developed several interfaces to WEPP for road erosion and fuel management predic-tions35. The X-DRAIN tool has been developed based on long-term WEPP simulations for a range of US climates, soil textures, topographic conditions, and road network parameters, such as spacing of cross-drains, road gradient, length of the buffer between roads and streams, and steepness of the buffer on sediment yield (Elliot et al. 1998). X-DRAIN is a user-friendly computer program used to determine optimum cross-drain spacing for existing and planned roads, and to provide recommendations for constructing and decommissioning roads, and estimating sediment yield for a given road or road system (Elliot et al.1998).
35 http://forest.moscowfsl.wsu.edu/fswepp/
102
Driving data required for WEPP include a minimum of precipitation, temperature, and solar radia-tion. Additional data on wind and dew point temperature or relative humidity can be used to support analytical evapotranspiration calculations. Overall, WEPP ranks as being of high complexity (Table 5), which reflects its wide-ranging functionality. A web interface is available to run WEPP simulations on servers at the USDA NSERL. A prototype GIS version is also available for download. The WEPP model and associated utilities are freeware.
Because the WEPP model does not explicitly include hydrodynamic channel network, flood flow routing, its functionality is limited to small watersheds (Conroy et al. 2006, Table 7). Furthermore, as indicated by the model application review (below), WEPP may have limited overall functionality in simulating the hydrology of forested watersheds.
Model ApplicationsA1.24.2
The following WEPP published model applications were reviewed:
• “EvaluationofrunoffpredictionfromWEPP-basederosionmodelsforharvestedandburnedforested watersheds” (Covert et al. 2005). This study evaluated runoff predictions generated by GeoWEPP (Geo-spatial interface to the Water Erosion Prediction Project) and WEPP. Three small (2 to 9 ha) watersheds in the mountains of the interior Pacific Northwest were monitored for sev-eral years following timber harvest and prescribed fires. Observed climate variables, percent ground cover, soil erodibility values, and GIS-derived slope data were used to drive the models. Predictions of total yearly runoff generated by the GeoWEPP and WEPP models were compared to total yearly runoff measured at each watershed.
• “Modelingerosionfrominslopinglow-volumeroadswithWEPPWatershedModel”(Tysdaletal.1999). In this study, sensitivity analysis and validation was carried out to determine the ability of the WEPP hydrologic model to predict erosion from in-sloping forest roads in the Oregon Coast Range. Various road lengths with contributing areas ranging from 81 to 889 m2 were used.
• “Acoupleduplanderosionandinstreamhydrodynamic-sedimenttransportmodelforevaluatingsediment transport in forested watersheds” (Conroy et al. 2006). WEPP’s hillslope erosion model was linked to CCHE1D channel model (sediment transport 1D networks) to model flow and sedi-ment load at the 473 ha North Fork Caspar Creek Experimental Watershed in coastal, northern California, which was harvested between 1989 and 1991.
In the seasonal runoff predictions in the Covert study, the modified WEPP model was most accurate for the spring months (higher runoff) but was a poor predictor for other seasons when the measured runoff rates were low. The GeoWEPP model successfully incorporated digital-elevation data, but the WEPP version used to process the data did not adequately represent the hydrological processes of for-ests. The lateral flow modifications that were added to the WEPP model improved predictions of runoff in forests, thus suggesting that further refinement of these calculations may improve the accuracy of WEPP-based models when applied to forest environments (Covert et al. 2005).
One notable limitation of WEPP arose in the Conroy et al. (2006) study. The model stores only daily summary information, even though it generates sub-daily hillslope runoff and sediment delivery infor-mation. This limitation precludes the model from being fully integrated with other hydrologic models that require sub-daily time series data, and resulted in misrepresenting days with more than one peak flow event. However, results show that linking the upland erosion model WEPP to a channel sediment transport model, CCHE1D, increased the overall accuracy of surface runoff/erosion estimates associated with implementing effective erosion control measures within ungauged watersheds.
103
The last study showed that, when used correctly, the WEPP hydrologic model can be useful in predicting runoff and sediment yields for in-sloped, low-volume roads. WEPP can account for such variations as topography, soil properties, management practices, and climate, all of which cause differ-ences in low-volume road erosion (Tysdal et al. 1999).
Thus, overall, these studies suggest that WEPP may have limited functionality in simulating forested watershed hydrology, and that its functionality mainly lies in predicting sediment erosion and yield from forest roads and hillslopes.
wrenss (winwrnshyd and ecA-Ab)A1.25
The Water Resources Evaluation of Non-Point Silvicultural Sources (WRENSS36) handbook provides a method to describe and evaluate changes to the water resource resulting from non-point silvicultural activities (US-Enviromental Protection Agency 1980). Two electronic procedures have been developed from the original WRENSS procedure: WinWrnsHyd (a Microsoft Access™ database implementation of the hydrology section of the handbook (Troendle and Leaf 1980; Swanson 2005), and Equivalent Clear-cut Area-Alberta (ECA-AB), which estimates changes in average annual streamflows (Silins 2002; Silins 2003). The description of ECA-AB was kindly provided by Dr. U. Silins, University of Alberta, and is reproduced with permission.
Model DescriptionA1.25.1
The WRENSS handbook and its companion electronic program, WinWrnsHyd, is a lumped parameter black-box hydrologic model, with no explicit soil or vegetation representation. The models provide for changes in average annual streamflows (yield) under different forest management regimes, and allow for forest regrowth using growth and yield curves (Tables 4, 6, and 8). The procedure does not account for runoff or channel routing.
The main use of WinWrnsHyd is in simulating annual water yields (Swanson 2005). Time series of water yield can be simulated with differing initial regrowth conditions to assist in analyzing past, present, and future harvesting patterns (Table 8) (Swanson 2005). Changes in water yield as well as the details of the effects of climatic and silvicultural variables on snow accumulation, disposition, and evapotranspiration can also be examined (Swanson 2005). “The rain and snow-dominated procedures were derived from simulated results produced by two models: PROSPER, a plant-soil-water model was used to simulate the rain-dominated procedure of the WRENSS handbook, and WATBAL, a snowmelt water balance model was used to simulate the snow dominated procedures” of the WRENSS hand-book (Swanson 2005, p. 108). Equations underlying the evapotranspiration calculations appear to be analytical, accounting for leaf area index, while snowmelt calculations are empirical. Changes in evapo-transpiration are achieved through empirical water use modifiers. The parameters that affect the results of the snow-dominated procedure are clearcut size, basal area, and tree height regrowth (aerodynamic roughness) (Swanson 2005). Overall, WRENSS and WinWrnsHyd rank as mixed procedures for simu-lating the effects of forest harvesting (Tables 4 and 6). The time-series version of WinWrnsHyd requires regrowth equations for the tree species found in the applicable WRENSS-defined hydrologic/climatic region (Swanson 2005).
The hydrologic procedures have been modified by the Canadian Forest Service as user-selectable op-tions, including (Swanson 2005; p. 108):
• aprovisiontocorrectwinterprecipitationfortheeffectofwindonsnowcatchingaugesfittedwith the commonly-used shields;
36 www.epa.gov/warsss/rrisc/handbook.htm
104
• aprovisiontoestimatesublimationfromthewintersnowpack;• aprovisiontocorrectprecipitationforelevationdifferencebetweenawatershedandthegauge
location; and • thecalibrationofthemodeltoawatershedwithknownorestimatedannualyield.Thecalibration
option can be used in lieu of, or in addition to, the snow catch and elevation options.
ECA-AB is a hydrologic procedure that, similar to WRENSS, estimates changes in average annual streamflows. The revised streamflow generation procedures include incorporating many components of the WRENSS procedures for simulating annual water yield. However, unlike WRENSS, ECA-Alberta does not explicitly simulate evapotranspiration (ET), but requires user-supplied information on long-term precipitation and streamflow (in the watershed or regional averages) to estimate changes in evapotranspirationand streamflow resulting from forest disturbance. Changes in streamflows are based on the area harvested in a watershed, rate of forest regrowth, and water balance calculations of generated runoff (determined from long-term monthly precipitation and annual streamflows). ECA-AB provides a relatively simple framework for evaluating the hydrologic effects of forest practices with modest input-data requirements. However, the accuracy of the model output depends primarily on providing it with accurate information on the hydrologic recovery of forest stands after disturbance, and on the availabil-ity of representative regional streamflow and precipitation data.
WinWrnsHyd and ECA-AB rank as low complexity procedures (Table 5) that can be used to rapidly assess changes in annual yield due to different forest management practices. The main drawback of these models is that their use does not extend beyond annual yield calculations. An algorithm to simulate peak flow changes due to forest harvesting has been developed in WinWrnsHyd (Swanson 2005) based on empirical relationships with water yield, but use of this model feature is not recommended given the lack of accounting for runoff processes in the model.
Model ApplicationsA1.25.2
The following model applications were reviewed:
• “HydrologicalimplicationsofsalvageharvestinglodgepolepineinHydraulicCreekwatershed”.(Golding 1986). The WRENSS model was applied to simulate the effects of different harvesting systems on streamflow. The author stressed the limitations of using the model, which calculates relative yield changes, rather than absolute streamflows.
• “OpportunisticuseofhydrologicdatatoassessCFS-WRENSSforforestryinterpretationsinSouthCentral B.C.” (Gluns and Eremko 1988). The authors evaluated the potential use of WRENSS in south-central BC. When the model was applied to data from an existing paired watershed study, it closely predicted the change in average annual water yield (mm) following clearcutting 27% of the watershed area. However, the CFS-WRENSS model poorly predicted actual water yields, a use for which it was not designed. It appeared that the snow evaporation component of the model is not applicable. The authors concluded that the model has applicability in south-central BC, but only to assess relative impacts of different harvesting regimes. Additional work is needed before absolute effects on streamflow can be modelled.
• “UsingtheWRNSHYDProceduretoestimatelong-termcumulativeeffectsofaspenclearcuttingon water yield in Alberta” (Swanson 1994). Regrowth of aspen clearcuts was estimated and input into WRENSS to obtain an estimate of the cumulative effects of harvest through time. The net effect in WRENSS of the combined effects of clearcutting on snow accumulation, snow transport, loss in transport, and evapotranspiration is zero for any single clearcut in aspen forests of this region 25 years after harvest. The cumulative effects of periodic harvests within a 546-km2 water-shed were estimated to determine possible harvesting regimes to maintain the annual water yield
105
increase below 15% at the first significant user, a criterion established by Alberta Environment to minimize flooding impacts on downstream residents and water users.
• “HydrologicrecoveryofaspenclearcutsinnorthwesternAlberta”(SwansonandRothwell2001).Several deciduous forest sites in northwestern Alberta ranging from 20 to 40 ha in size were modelled over 60 years to determine the change in generated runoff from original harvest. The ob-jective of this study was to better define the period during which aspen clearcuts could significantly affect water yield and the potential for flooding. The aspen regrowth equations were programmed to reach maximum basal area at age 60 for this project. This allowed forest managers to successfully relate cumulative areal and temporal harvests to water yield change.
• “HydrologiceffectsofmountainpinebeetleinfestationsinwesternAlberta”(RothwellandSwan-son 2007). WRENSS was used to simulate the hydrologic effects of mountain pine beetle tree mortality on annual water yield and peak flows in the east-slope watersheds of western Alberta. Simulations were done at Forest Management Unit 2 (FMU 8), a 2200-km2 area dominated by lodgepole pine stands.
Overall, these studies show the WRENSS procedure as useful for evaluating the relative effects of existing and future harvests to minimize any detrimental effects on water users and other resources. Care may be needed when assessing absolute effects on streamflow, at least in portions of BC (Golding 1986; Gluns and Eremko1988). The model is currently relatively widely used in Alberta.
wrmmA1.26
The Water Resources Management Model (WRMM) was developed by Alberta Environment as a plan-ning tool for surface water resources utilization in river basins (Ilich et al. 2000). The model may be applied to relatively simple river systems or it can represent extensive, complex river basins with a net-work of reservoirs, hydro-plants, instream channels, diversion canals, irrigation, and other consumptive uses.
The model enables easy and repeated analysis of the river basin’s response to differing combinations of water supplies, demands, and water management structures (Ilich et al. 2000). The model simulates planning alternatives within a river basin by allocating the water supply to present or projected uses. Use of the model has evolved in two forms:
• BasinPlanning,whichuseshistoricalsupplyanddemanddatatoprojectfutureconditions,allow-ing the assessment of long-term water use alternatives; and
• OperationalPlanning,whichevaluatestheshort-termfuture(e.g.nextfewdaysorweeks)conse-quences of different operational strategies.
The WRMM has been extensively used by Alberta Environment for water-use planning in the South Saskatchewan River Basin.
The WRMM does not simulate watershed processes (runoff, etc.) in that water supply from head-waters or local runoff (i.e., natural inflows to the river system) must be specified by the user from measurements or from simulated forecast flows from another model. As such, the model has no func-tionality in forest management applications and is not included in Tables 4 to 8.
wuAmA1.27
The Water Use Analysis Model (WUAM) is designed by the Economics and Conservation Branch, Ecosystem Sciences and Evaluation Directorate of Environment Canada, primarily to provide projec-tions of multi-sectoral water uses in a drainage basin context (Kassem 1992; Kassem et al. 1994). WUAM depicts a river basin as a dendritic network of nodes (representing tributaries or sub-basins) and links
106
(representing the flow path between nodes), with water use projections and water balance calcula-tions carried out at the node level using monthly time intervals (Kassem 1992; Kassem et al.1994). The model is also able to consider water diversions and inter-jurisdictional water apportionment, analyze the impacts of water price on water use, model reservoir operations, account for water use priorities, and analyze water rationing and usage cutbacks when available water supplies are approached or exceeded (Kassem 1992; Kassem et al. 1994).
As with WRMM, water supplies are simulated based on specified natural streamflow time-series data at selected points within the drainage basin, i.e., the model does not simulate watershed processes (run-off etc.) such that water supply from headwaters, local runoff or in-diversion must be specified by the user from measurements or simulated forecast flows. As such, the model has no functionality in forest management applications and is not included in Tables 4 to 8.
107
Model name Agricultural Catchments Research Unit (ACRU)
Development group Department of Agricultural Engineering of the University of Natal in Pietermaritzburg, South Africa. A version of the model adapted for Alberta conditions was developed at the University of Lethbridge (Stefan Kienzle)
Model information URL www.beeh.unp.ac.za/acru/
Manual Yes
Tutorial No
Model support Dr. Stefan Kienzle is willing to provide training for a fee; original training, without the snow modelling, can be provided in South Africa (Pietermaritzburg).
Model cost A few 100 US$
Computing requirements Equipment PC
Software GIS
Source code available Contact developer
Model type Semi-distributed
Model scales Input time step Daily
Output time step Daily
Grid size Sub-basins
Application scale Small to medium watersheds
Planning scale Sub-basins
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes (University of Lethbridge model version only)
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Fraction of rainfall
Snow accumulation Threshold temperature
Snowmelt Temperature index, accounting for precipitation
Snow interception Fraction of snowfall
Evapotranspiration Physically based
Infiltration Empirical
Overland flow Empirical
Subsurface hillslope runoff Empirical
Groundwater flow Empirical
Roads No
Streamflow routing Yes
Frozen soil No
Lakes/wetlands Yes
table A2.1 Agricultural Catchments Research Unit (ACRU)
APPendix 2 model review tAbles
108
Model name Agricultural Catchments Research Unit (ACRU)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes (not in UNP version of model)
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Sediment erosion, nutrient fluxes; irrigation, reservoir operations
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, land classification, soil depth, and texture classes
Meteorology data Daily precipitation and minimum and maximum temperature
Overstorey vegetation parameters Crop type
Understorey vegetation parameters None
Soil parameters (deep layers) Soil texture, hydraulic conductivity
Soil parameters (root zone layers) Soil texture, hydraulic conductivity, root depths
Routing Slope, length, roughness, shape
Data processing requirements High; GIS required
Key model assumptions or limitations
Predominantly an agricultural model but has been applied in forested environments
Constraints Data requirements High
Level of expertise Professional to academic
Level of effort High
Model adaptability Forest regrowth capability No (model does have a crop yield component for economic analysis)
Complex stand characteristics No
Forest mortality or fire risk No
Alteration of land cover details Yes,but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.1 continued
109
Model name BROOK90
Development group C. Anthony Federer
Model information URL http://home.roadrunner.com/~stfederer/brook/compassb.htm
Manual Yes, available upon request
Tutorial No
Model support [email protected]
Model cost Free (If model is used regularly, $30 for individual licence and $50 for site licence)
Computing requirements Equipment PC, Apple, or Unix workstation with a minimum of 512Mb RAM
Software WinZip, TextPad, FileEx
Source code available Yes (Fortran, C, QuickBasic, VisualBasic)
Model type Lumped model
Model scales Input time step Daily
Output time step Detailed output (spatial variables) at user specified times
Grid size N/A
Application scale Small watersheds
Planning scale Watershed
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No allowance for ROS events
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Calculated with a canopy resistance of zero and aerodynamic resistances based on canopy height, coupled with a canopy capacity and an average storm duration
Snow accumulation Yes
Snowmelt Based on a degree-day factor; accounts for snowpack temperature and liquid water content
Snow interception Identical to rain interception with no allowance for melt of intercepted snow
Evapotranspiration Energy-based calculations for overstorey, understorey and soil using Shuttleworth and Wallace
Infiltration Darcy’s Law accounting for soil moisture level
Overland flow Saturation excess overland flow
Subsurface hillslope runoff The downslope flux from each layer depends on the slope and slope length parameters as well as on the hydraulic conductivity of the layer.
Groundwater flow Empirical reservoir
Roads No
Streamflow routing No
Frozen soil No
Lakes/wetlands No
table A2.2 BROOK90
110
Model name BROOK90
Model outputs Full hydrograph For first-order watersheds only
Annual yield For first-order watersheds only
Peak flow For first-order watersheds only
Low flow For first-order watersheds only
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Preferential flow
Key required inputs (may need to refer to documentation for comprehensive list)
Map files No
Meteorology data Daily temperature and maximum/minimum temperature required (daily solar radiation, vapor pressure, and wind speed are desirable)
Constants Canopy parameter, Stefan Boltzman constant, solar constant, latent heat of sublimation of snow, latent heat of fusion of water, vonKarmen constant, psychromoter constant
Overstorey vegetation parameters Albedo, days of initial leaf out and major leaf fall, LAI, stem area index
Understorey vegetation parameters N/A
Soil parameters (deep layers) Thickness of unsaturated zone; number of soil parameters is flexible and depends on runoff mechanism(s) simulated
Soil parameters (root zone layers) Soil water release curve, thickness of root zone; number of soil parameters is flexible and depends on runoff mechanism(s) simulated)
Channel routing None
Data processing requirements Medium
Key model assumptions or limitations
No channel routing
Constraints Data requirements Medium
Level of expertise Medium
Level of effort Basic GUI makes for medium effort
Model adaptability Forest regrowth capability No but input can be manipulated to reflect input from growth models
Complex stand characteristics No; mixed stands are represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.2 continued
111
Model name Cold Regions Hydrologic Model (CRHM)
Development group University of Saskatchewan
Model information URL www.usask.ca/hydrology/crhm.htm
Manual Yes; 30-page overview
Tutorial No
Model Support Limited; see website for contact info
Model Cost None (academic distribution and agreement)
Computing requirements Equipment PC
Software GIS (ArcView); C++ for modification; Excel for post-processing
Source Code Available Yes
Model type Semi-distributed
Model scales Input Time Step Sub-daily
Output Time Step Sub-daily
Grid Size Metres; model is object-oriented, based on HRUs that have variable spatial units and connectivity for a region
Application Scale Small to medium watersheds
Planning scale Landscape segments such as hillslopes (HRUs)
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Yes
Snow accumulation Accumulation and blowing snow transport model
Snowmelt Various choices, including: energy balance model, fractal melt/depletion model, simplified melt models, or radiation and temperature index method
Snow Interception Interception and sublimation
Evapotranspiration Penman-Monteith, Granger and Pomeroy, Shuttleworth and Wallace
Infiltration Various choices, including: Green and Ampt approach, and methods by Gray, Granger for frozen soils
Overland flow Soil moisture balance for modelling drainage and runoff
Subsurface hillslope runoff Soil moisture balance for modelling drainage and runoff
Groundwater flow Yes
Roads No
Streamflow routing No; streamflow response modelled by lag and route method
Frozen soil Yes
Lakes/wetlands No
table A2.3 Cold Regions Hydrologic Model (CRHM)
112
Model name Cold Regions Hydrologic Model (CRHM)
Model outputs Full hydrograph Simplified channel routing
Annual yield Simplified channel routing
Peak flow Simplified channel routing
Low flow Simplified channel routing
SWE Yes
Evapotranspiration Yes
Water Balance Yes
Soil Moisture Yes
Infiltration Yes
Water table Yes
Overland flow No
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed Runoff Yes
Other Effects of frozen soil on water movement
Key required inputs (may need to refer to documentation for comprehensive list)
Map Files Spatial data (basin area, DEM) imported/generated with internal GIS tools
Meteorology data Variable, depending on specific module selections within CRHM (precipitation, temperature, radiation, humidity, wind speed, and direction)
Overstorey vegetation parameters
Vegetation height, albedo, fetch distance
Understorey vegetation parameters
N/A
Soil parameters (deep layers) Lag time, storage constant
Soil parameters (root zone layers)
Ground slope, thickness, bulk density, porosity, heat capacity, soil moisture conditions, hydraulic conductivity
Channel Routing Lag time, storage constant
Data processing Requirements High; HRUs developed using GIS tool within CRHM based on user understanding of spatial data, terrain, and relevant hydrological processes
Key model assumptions or limitations
Simplified streamflow routing
Constraints Data requirements Moderate; highly dependent on complexity of processes selected in various modules
Level of expertise High; selection and use of each module would require knowledge of each specific model/process
Level of effort High
Model adaptability Forest-regrowth capability No, but input can be manipulated to reflect input from growth models
Complex stand characteristics No; mixed stands are represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.3 continued
113
Model name Distributed Hydrology Soil Vegetation Model (DHSVM)
Development group University of Washington
Model information URL www.hydro.washington.edu/Lettenmaier/Models/DHSVM/index.shtml
Manual Limited (web pages)
Tutorial Yes
Model support Limited support available by special arrangement ([email protected]; DHSVM user support list server)
Model cost None (GIS software will need to be purchased)
Computing requirements Equipment PC, Apple, or Unix workstation with a minimum of 512Mb RAM
Software GIS (ArcInfo); Gnu C-compiler (freeware)
Source code available Yes
Model type Fully distributed
Model scales Input time step Sub-daily
Output time step Same as input; detailed output (spatial variables) at user specified times
Grid size 10 to several 100 m
Application scale Small to medium watersheds up to 10 000 km2
Planning scale Flexible
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Overstorey (trees) and understorey
Rainfall interception Maximum canopy rainfall interception threshold; mass and energy balance for evaporation of intercepted rainfall
Snow accumulation Rain/snow temperature thresholds
Snowmelt Physical; snow mass and energy balance
Snow interception Rate and maximum canopy snow interception threshold; mass and energy balance for melt and evaporation of intercepted snow
Evapotranspiration Energy-based calculations for overstorey, understorey, and soil
Infiltration Darcy’s Law accounting for soil moisture level
Overland flow Empirical with flow velocity dependent on grid size and time step
Subsurface hillslope runoff Pixel-by-pixel routing based on topographic gradient, water mass balance, and Darcy’s Law
Groundwater flow DHSVM does not have an explicit representation of groundwater flow
Roads Can be incorporated but not required
Streamflow routing Can be internal (DEM calculated using GIS), forced (stream network file), or unit hydrograph
Frozen soil No
Lakes/wetlands No
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
table A2.4 Distributed Hydrology Soil Vegetation Model (DHSVM)
114
Model name Distributed Hydrology Soil Vegetation Model (DHSVM)
Model outputs, cont. Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater No
Road flow Yes
Watershed runoff Yes
Other Mass wasting, sediment erosion
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, vegetation type, soil type, and depth
Meteorology data Temperature, precipitation, humidity, wind speed, shortwave and longwave radiation, temperature lapse rate and precipitation lapse rate for one or more stations at model input time step
Constants Reference height for meteorology data, rain LAI multiplier, snow LAI multiplier, rain/snow temperature thresholds, snow roughness, snow water capacity, ground roughness
Overstorey vegetation parameters
Fractional coverage, trunk space, height, LAI, albedo, aerodynamic attenuation, radiation attenuation, clumping factor, maximum snow interception capacity, maximum release drip ratio, snow interception efficiency, stomatal resistance (min/max), moisture threshold, vapor pressure deficit, RPC, number of root zones, root zone depths, root fractions in each zone
Understorey vegetation parameters
LAI, albedo, root fractions in each zone
Soil parameters (deep layers)
Lateral conductivity and exponential decrease, porosity
Soil parameters (root zone layers)
Surface albedo, vertical conductivity, porosity, maximum infiltration, pore size distribution, bubbling pressure, field capacity, wilting point, bulk density, thermal conductivity, thermal capacity
Routing Stream network file, stream map file, stream class file, road network file, road map file, road class file, unit hydrograph, and travel time files (last two files replace need for all preceding files)
Data processing Requirements
Most data requires GIS processing including maps (topography, soil type and depth, vegetation) and streamflow routing and roads; climate data typically also needs to be processed to address data gaps and quality
Key model assumptions or limitations
Steeply sloped terrain with thin soil veneer
Constraints Data requirements DHSVM data requirements are high compared to typical data availability; most physical parameters can be based on literature values
Level of expertise Senior professional to academic level; model user needs to be knowlegable in physical hydrology, computer modelling, and GIS
Level of effort High; considerable GIS and data processing required
Model adaptability Forest-regrowth capability No, but input can be manipulated to reflect input from growth models
Complex stand characteristics
No; mixed stands are represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios; model can handle map input from meteorological models (MM5)
table A2.4 continued
115
Model name Forest Hydrology Model (ForHyM)
Development group University of New Brunswick, Department of Forestry
Model information URL N/A
Manual No
Tutorial No
Model support Unlikely
Model cost Free
Computing requirements Equipment PC or Apple workstation
Software Not known
Source code available No
Model type Lumped
Model scales Input time step Daily or weekly
Output time step Same as input
Grid size N/A
Application scale Small watersheds
Planning scale Watershed
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation Overstorey (trees)
Rainfall interception Yes; based on theoretical relationship
Snow accumulation Rain/snow temperature thresholds
Snowmelt Accounts for snowpack temperature, ET and liquid water content
Snow interception Yes; based on theoretical relationship
Evapotranspiration Potential ET equation based on relationships with air temperature
Infiltration Yes
Overland flow No
Subsurface hillslope runoff No
Groundwater flow No; model does not address seepage losses to groundwater
Roads No
Streamflow routing No
Frozen soil No
Lakes/wetlands No
Model outputs Full hydrograph For first-order watersheds only
Annual yield For first-order watersheds only
Peak flow For first-order watersheds only
Low flow For first-order watersheds only
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
table A2.5 Forest Hydrology Model (ForHyM)
116
Model name Forest Hydrology Model (ForHyM)
Model outputs, cont. Infiltration Yes
Water table No
Overland flow No
Subsurface hillslope runoff Yes
Groundwater No
Road flow No
Watershed runoff Yes
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files No
Meteorology data Monthly values for: mean air temperature, precipitation, and mean snow fraction of precipitation
Constants LAI, field capcity, and permanent wilting percentage for deciduous versus coniferous trees
Overstorey vegetation parameters
LAI, stem area index, proportions of coniferous and deciduous trees
Understorey vegetation parameters
N/A
Soil parameters (deep layers) N/A
Soil parameters (root zone layers)
Field capacity, thickness and permanent wilting percentage for each of forest floor, soil and subsoil, clay fraction of soil and subsoil, texture of soil and subsoil
Channel Routing None
Data processing Requirements Minimal
Key model assumptions or limitations
No channel routing
Constraints Data requirements Low to medium
Level of expertise Professional to academic
Level of effort Medium
Model adaptability Forest regrowth capability No, but input can be manipulated to reflect input from growth models
Complex stand characteristics No; mixed stands are represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.5 continued
117
Model name Forest Water Dynamics (ForWaDy)
Development group B. Seely and J. P. Kimmins.
Model information URL www.forestry.ubc.ca/ecomodels/moddev/forwady/forwady.htm
Manual No
Tutorial No
Model Support Contact developer
Model Cost Free
Computing requirements Equipment PC or Apple workstation
Software Contact developer
Source code available Contact developer
Model type Lumped
Model scales Input time step 0.25 days
Output time step Same as input
Grid size N/A
Application scale Small watersheds
Planning scale Watershed
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Overstorey (trees) and understorey
Rainfall interception Yes; based on theoretical relationship
Snow accumulation Rain/snow temperature thresholds
Snowmelt Accounts for snowpack temperature, ET, radiation melt and liquid water content
Snow Interception Yes; based on theoretical relationship
Evapotranspiration Energy budget approach for canopy trees, understorey plants and forest floor
Infiltration Yes
Overland flow No
Subsurface hillslope runoff No
Groundwater flow No; current model does not address seepage losses to groundwater
Roads No
Streamflow routing No
Frozen soil No
Lakes/wetlands No
Model outputs Full hydrograph For first-order watersheds only
Annual yield For first-order watersheds only
Peak flow For first-order watersheds only
Low flow For first-order watersheds only
SWE Yes
Evapotranspiration Yes
Water Balance Yes
table A2.6 Forest Water Dynamics (ForWaDy)
118
Model name Forest Water Dynamics (ForWaDy)
Model outputs, cont. Soil Moisture Yes
Infiltration Yes
Water table No
Overland flow No
Subsurface hillslope runoff Yes
Groundwater No
Road flow No
Watershed Runoff Yes
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map Files No
Meteorology data Mean, maximum and minimum air temperature, solar radiation, total precipitation, snow fraction
Constants Albedo for the canopy and understorey
Overstorey vegetation parameters
Percent cover by conifers and hardwoods
Understorey vegetation parameters
Seasonal conifer and hardwood LAI, seasonal understorey percent cover, rooting depths for trees, rooting depths for understorey, root occupancy in each layer, canopy resistance
Soil parameters (deep layers)
Soil texture class of each soil layer, coarse fragment content of layers
Soil parameters (root zone layers)
LF layer mass (kg/ha), humus depth and bulk density, depth of soil layers (rooting zone), soil texture class of each soil layer, coarse fragment content of layers
Routing None
Data processing Requirements
Minimal
Key model assumptions or limitations
No channel routing
Constraints Data requirements Medium
Level of expertise Professional to academic
Level of effort Medium
Model adaptability Forest regrowth capability No, but model can be coupled to FORECAST and FORCEE
Complex stand characteristics
No; mixed stands are represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.6 continued
119
Model NameHydrologiska Byråns Vattenbalansavdelning–Environment Canada (HBV-EC)
Development group Environment Canada Pacific and Yukon Region (D. Hutchinson), UBC Dept of Geography (D. Moore)
Model information URL http://chc.nrc-cnrc.gc.ca/Numerical/Downloads/Green_Kenue_e.html
Manual Yes; described under Green Kenue website
Tutorial Yes; available at ftp://kenueftp.chc.nrc.ca/GreenKenue/
Model support Contact developers
Model cost None; licence agreement with National Research Council of Canada
Computing requirements Equipment PC
Software GIS (ArcView); Green Kenue GUI for HBV-EC
Source code available Contact developers
Model type Semi-distributed
Model scales Input time step Daily
Output time step Daily
Grid size Model area is divided into elevation bands and uniform grid to define forested, open, lake, and glacial areas
Application scale Small to medium watersheds
Planning scale Landscape segments such as hillslopes (HRUs)
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial Yes
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation No explicit accounting for forest cover (as “layer”)
Rainfall interception Fraction of rainfall
Snow accumulation Threshold temperature
Snowmelt Temperature index, accounting for retention capacity of snow and partial thaw
Snow interception Fraction of snowfall
Evapotranspiration ET is scaled relative to potential ET and soil moisture storage
Infiltration Percolation modelled based on soil moisture storage for given elevation band and user specified field capacity
Overland flow Generated when soil moisture storage is exceeded
Subsurface hillslope runoff
Cumulative runoff is summed for forested and open areas, and added to “fast” and “slow” linear reservoirs; runoff from glacier areas is routed seperately, based on SWE of glaciated area
Groundwater flow Cumulative runoff is summed for forested and open areas, and added to “fast” and “slow” linear reservoirs; runoff from glacier areas is routed seperately, based on SWE of glaciated area
Roads No
Streamflow routing Simplified channel routing; triangular weighting function
Frozen soil No
Lakes/wetlands No
table A2.7 Hydrologiska Bryåns Vattenbalansavdelning–Environment Canada (HBV-EC)
120
Model NameHydrologiska Byråns Vattenbalansavdelning–Environment Canada (HBV-EC)
Model outputs Full hydrograph Simplified channel routing
Annual yield Simplified channel routing
Peak flow Simplified channel routing
Low flow Simplified channel routing
SWE Yes
Evapotranspiration Yes
Water Balance Yes
Soil Moisture Yes
Infiltration Yes
Water table No
Overland flow Yes
Subsurface hillslope runoff
Yes
Groundwater Yes
Road flow No
Watershed Runoff Yes
Other glacial melt, lakes
Key required inputs (may need to refer to documentation for comprehensive list)
Map Files DEM, land classification (forest, open, lake, glacier)
Meteorology data Daily precipitation, daily mean temperature, daily potential ET, correction factors for elevation and gauge errors
Overstorey vegetation parameters
Canopy factors for sunlight blocked, interception fraction, ratio of melt compared to open areas
Understorey vegetation parameters
None
Soil parameters (deep layers)
Empiriral reservoir parameters
Soil parameters (root zone layers)
Field capacity, lower limit for ET
Routing Linear reservoir parameters
Data processing Requirements
Green Kenue GUI offers windows-driven method (drop down menus) for model setup
Key model assumptions or limitations
Rudimentary representation of forest effects on watershed hydrology; simplified channel routing
Constraints Data requirements Medium
Level of expertise Professional to academic
Level of effort Medium
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk
No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.7 continued
121
Model NameHydrologic Engineering Center’s Hydrologic Modelling System (HEC–HMS)
Development group US Army Corps of Engineers
Model information URL www.hec.usace.army.mil/software/hec-hms/
Manual Yes
Tutorial Yes; example applications and training workshops available
Model support No
Model cost None
Computing requirements Equipment PC, Solaris UNIX, Linux
Software GIS (ArcView or companion product: HEC-GeoHMS)
Source code available No
Model type Semi-distributed
Model scales Input time step Sub-daily or greater
Output time step Sub-daily or greater
Grid size N/A
Application scale Small to large watersheds
Planning scale Sub-basins
Model calibration method Internal automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Not specified
Snow accumulation Temperature threshold dependent
Snowmelt Temperature-index method, distributed based on elevation bands
Snow interception Not specified
Evapotranspiration Monthly average PET by Priestly-Taylor
Infiltration Various methods including: constant or exponential rate, SCS curve number, Green-Ampt, Smith-Parlange
Overland flow Various methods including: unit hydrograph of Clark, Snyder or SCS, user-specified unit hydrographs, kinematic wave
Subsurface hillslope runoff Flow from sub-basins can be accounted for by various methods: constant input, linear reservoir method, non-linear Boussinesq method
Groundwater flow Yes, several reservoirs
Roads No
Streamflow routing Various methods: simple time lag, Muskingum routing, modified Puls method, Muskingum-Cunge method for complex channel geometry
Frozen soil No
Lakes/wetlands Yes
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
table A2.8 Hydrologic Engineering Center's Hydrologic Modelling System (HEC-HMS)
122
Model NameHydrologic Engineering Center’s Hydrologic Modelling System (HEC–HMS)
Model outputs, cont. SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table No
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other lakes/wetlands
Key required inputs (may need to refer to documentation for comprehensive list)
Map files Not specified
Meteorology data Meteorological data are input to a meteorological model that may be applied in a gridded or Theissen polygon approach to the model area; data may originate from specific gauges stations or be generated statistically based on required storm frequency and exceedance probability (US National Weather Service data)
Overstorey vegetation parameters
Crop coefficient for distribution of vegetative cover within model area
Understorey vegetation parameters
N/A
Soil parameters (deep layers)
Linear reservoir parameters for baseflow store; depends on complexity of modelling
Soil parameters (root zone layers)
Linear reservoir parameters for interflow store; depends on complexity of modelling
Routing Depends on complexity of modelling
Data processing requirements
Basins are defined and comprised of hydrologic elements (sub-basin, reach, reservoir, junction, diversion, source, sink)
Key model assumptions or limitations
Empirical ET and snowmelt equations; SCS curve number runoff methods predominantly a model for river networks
Constraints Data requirements High
Level of expertise Professional to academic
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.8 continued
123
Model Name Hydraulic Evaluation of Landfill Performance (HELP)
Development group US Environmental Protection Agency, US Army Corps of Engineers, University of Hamburg Institute of Soil Science
Model information URL http://el.erdc.usace.army.mil/elmodels/helpinfo.html www.geowiss.uni-hamburg.de/i-boden/mitarb/kberger_e.htm
Manual Yes
Tutorial Yes
Model support Limited for publically available code; training and support available from Schlumberger Water Services for Visual HELP version
Model cost Free code publically available from http://el.erdc.usace.army.mil/products.cfm?Topic=model&Type=landfill; Visual HELP with GUI available from www.swstechnology.com/software_product.php?ID=11
Computing requirements Equipment PC
Software GIS and Python for distributed recharge modelling
Source code available No
Model type Semi-distributed when combined with GIS
Model scales Input time step Daily
Output time step Daily; summaries for longer timeframes
Grid size None (1D columns)
Application scale Soil columns and combinations/groups of columns for larger areas of any size
Planning scale Soil columns and combinations/groups of columns for larger areas of any size
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation 1-layer of vegetation, growth and decay model relative to maximum LAI
Rainfall interception Interception calculated by empirical relationship for above ground biomass
Snow accumulation Snow accumulation based on temperature threshold
Snowmelt Snow melt calculated by temperature index method allowing for re-freezing and rain-on-snow
Snow interception Not specified
Evapotranspiration Modified Penman equation (energy available for ET)
Infiltration Initial infiltration calculated by surface water balance (accounting for rainfall, snowmelt, runoff, evaporation), vertical drainage in the soil column is calculated per layer based on Darcy flow by unit gradient
Overland flow Runoff calculated by SCS curve number but not routed
Subsurface hillslope runoff No
Groundwater flow No
Roads No
Streamflow routing No
Frozen soil Yes
Lakes/wetlands No
table A2.9 Hydraulic Evaluation of Landfill Performance (HELP)
124
Model Name Hydraulic Evaluation of Landfill Performance (HELP)
Model outputs Full hydrograph No
Annual yield No
Peak flow No
Low flow No
SWE Yes
Evapotranspiration Yes
Water Balance Yes
Soil Moisture Yes
Soil Moisture Yes
Infiltration Yes
Water table No
Overland flow No
Subsurface hillslope runoff No
Groundwater No
Road flow No
Watershed Runoff Yes
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map Files DEM (for approximate slope aspect only), soil type, vegetation
Meteorology data Growing season, average annual windspeed, average quaterly humidity, latitude, precipitation, temperature, solar radiation
Overstorey vegetation parameters
LAI, evaporative zone depth
Understorey vegetation parameters
Lumped with overstorey vegetation (i.e., one vegetation layer)
Soil parameters (saturated zone)
Water table depth (base of model)
Soil parameters (unsaturated zone)
Porosity, wilting point, field capacity, hydraulic conductivity, SCS curve number, slope, layers and thicknesses
Channel Routing None
Data processing Requirements
Climate data can be generated from internal weather generator; soil layer data can be selected from database
Key model assumptions or limitations
No channel routing; SCS curve number runoff methods; developed as landfill model
Constraints Data requirements Medium
Level of expertise Medium
Level of effort Medium
Model adaptability Forest regrowth capability Yes
Complex stand characteristics
No; mixed stands are represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.9 continued
125
Model name Hydrologic Simulation Program–Fortran (HSPF)
Development group USGS/US-EPA
Model information URL http://water.usgs.gov/software/HSPF/
Manual Yes (WinHSPF user manual; Window’s help file)
Tutorial Yes (part of WinHSPF user manual)
Model support AQUA TERRA Consultants; USGS HSPF user list server
Model cost None (WinHSPF is distributed as part of US-EPA BASINS analysis system): www.epa.gov/waterscience/BASINS/
Computing requirements Equipment Unix or DOS and 128 Mb of RAM
Software ArcView Version 3.1 or better and Spatial Analist
Gui Yes (WinHSPF, BASINS)
Source code available No
Model type Semi-distributed
Model scales Input time step 1 minute to 1 day
Output time step Same as input
Grid size land segments with similar hydrologic characteristics
Application scale Small to large watersheds
Planning scale Grouped land segment to watershed scale
Model calibration method Yes; HSPFexpert can be used for model calibration
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Single canopy
Rainfall interception interception of rain or water yielded by melting snowpack
Snow accumulation Rain/snow temperature threshold
Snowmelt Snow accumulation and melt using an degree-day or energy balance approach
Snow interception No; only snowmelt interception considered (see above)
Evapotranspiration Emperical from PET and water available in storage reservoirs (canopy interception, upper soil zone, lower soil zone, baseflow, active groundwater)
Infiltration Infiltration capacity based
Overland flow Yes
Subsurface hillslope runoff Upper zone, lower zone, and baseflow reservoir
Groundwater flow Not explicitly simulated; deep percolation is considered “lost” to system
Roads No
Streamflow routing Inputs from pervious and impervious land segments are routed downstream through user-defined reaches
Frozen soil No
Lakes/wetlands Yes
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
table A2.10 Hydrologic Simulation Program–Fortran (HSPF)
126
Model name Hydrologic Simulation Program–Fortran (HSPF)
Model outputs, cont. Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Water quality
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM
Pervious or impervious land segments with simular hydrologic characteristics
Meteorology data Precipitation and temperature as minimum; temperature lapse rate is optional if defaults are not good; solar radiation, dew point, wind velocity, and cloud cover are needed for energy-based snowmelt calculations
Constants KMELT and TBASE (parameters for degree-day snowmelt method), maximum water content of snowpack, snow albedo, ground heat snowmelt rate
Vegetation parameters Canopy interception capacity (monthly or constant), PET, lower zone ET fraction, base flow ET fraction, Manning’s roughness co-efficient for surface runoff, shaded fraction of land segment (snowmelt calculations), fraction of land covered by forest
Upper zone, lower zone and baseflow parameters
Upper zone storage, lower zone storage nominal, porosity, infiltration index, infiltration exponent, infiltration ratio, length-of-surface flow path, slope-of-surface flow path, surface retention storage, upper zone storage nominal, interflow index, interflow recession constant, variable groundwater recession, active groundwater recession constant, fraction of groundwater to deep aquifer or inactive storage, and active groundwater ET fraction
Channel routing Length, width, depth and hydraulic properties of each reach, hydraulic routing weighting factor
Data processing requirements
Limited data processing requirements except for climate data; data processing is facilitated through GUI
Key model assumptions or limitations
The model is mainly intended for water quality simulations (urban and agricultural settings); simplified forest cover representation as evidenced by Carnation Creek work (Eugene Hetherington)
Constraints Data requirements High
Level of expertise Professional to academic
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.10 continued
127
Model name HydroGeoSphere
Development group University of Waterloo and Laval University
Model information URL www.science.uwaterloo.ca/~mclaren/
Manual Yes
Tutorial No; example files provided
Model support Limited; contact Rob McLaren ([email protected])
Model cost None (academic distribution and agreement)
Computing requirements Equipment PC, Unix
Software GIS, GridBuilder mesh generator (by Rob McLaren) for watershed scale models, TECPLOT for visualisation of output
Source code available Yes
Model type Fully-distributed
Model scales Input time step Variable and adaptive during simulation
Output time step Same as input; detailed output for pre-specified variables
Grid size Centimetres to hundreds of metres
Application scale Soil columns, research plots, watersheds to 1000 km2
Planning scale Flexible
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival No
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation No
Rainfall interception Maximum threshold for interception
Snow accumulation No; off-line calculation of snow accumulation is required
Snowmelt No; snowmelt may be simulated as specified flux rate
Snow interception No
Evapotranspiration Kristensen and Jensen (1975) method (AET estimated from PET, LAI, root zone parameters)
Infiltration 3D variably saturated subsurface hillslope runoff by Richard’s equation (finite element or finite difference methods)
Overland flow 2D diffusion-wave equation
Subsurface hillslope runoff 3D variably saturated subsurface hillslope runoff by Richard’s equation (finite element or finite difference methods)
Groundwater flow 3D variably saturated subsurface hillslope runoff by Richard’s equation (finite element or finite difference methods)
Roads Can be incorporated based on land surface/subsurface properties
Streamflow routing Internal computation based on model topography (DEM) and hydraulic properties of land surface (i.e., specified for known stream routes)
Frozen soil No
Lakes/wetlands As boundary conditions or part of simulation
table A2.11 HydroGeoSphere
128
Model name HydroGeoSphere
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE No
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow Yes
Watershed runoff Yes
Other Water quality
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, geologic layering, vegetative cover
Meteorology data Climate data pre-processed to generate rainfall and PET data for use in model
Overstorey vegetation parameters
LAI, field capacity, wilting point, oxic and anoxic limits, interception value, root density function, root depth, three fitting parameters
Understorey vegetation parameters
Lumped with overstorey vegetation (i.e., one vegetation layer)
Soil parameters (saturated zone)
Hydraulic conductivity, specific storage, porosity, anisotropy ratio
Soil parameters (unsaturated zone)
Soil moisture curves (vanGenuchten functions, Brooks-Corey functions, or tabulated relationships for capillary pressure-moisture content and relative permeability)
Channel routing Manning’s N, microtopography height, minimum mobile water depth
Data processing requirements
Climate data and PET must be calculated off-line; soil layering and geologic features must be described for entire model region and have corresponding XYZ data files
Key model assumptions or limitations
No accounting for snow processes
Constraints Data requirements High; full-scale catchment models can be developed from conceptualized knowledge of soil and geologic framework
Level of expertise High; model user needs to be knowlegable of both hydrology, hydrogeology and numerical modelling
Level of effort High; considerable effort is needed to pre-process climate and landscape data for inclusion in robust groundwater flow model
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.11 continued
129
Model name Integrated Hydrology Model (InHM)
Development group University of Waterloo (initial development), Stanford University (1998 to present)
Model information URL http://inhm.org/
Manual No; some instruction at http://68.183.67.101/InHM/docs/GettingStarted.htm
Tutorial No; example files provided at http://68.183.67.101/InHM/docs/examples/abdul.coupled.field/field.index.htm
Model support Limited; contact Joel VanderKwaak ([email protected])
Model cost None (licence agreement on website)
Computing requirements Equipment PC
Software TECPLOT; for visualisation of output
Source code available Yes
Model type Fully distributed
Model scales Input time step Variable and adaptive during simulation
Output time step Same as input; detailed output for pre-specified variables
Grid size Centimetres to hundreds of metres
Application scale Soil columns and small research watersheds (few square km)
Planning scale Flexible
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival No
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation No
Rainfall interception No
Snow accumulation No; off-line calculation of snow accumulation is required
Snowmelt No; snowmelt may be simulated as specified flux rate
Snow interception No
Evapotranspiration No; ET rate may be defined as negative flux (scaled based on computed moisture condition)
Infiltration 3D variably saturated Subsurface hillslope runoff
Overland flow 2D diffusion-wave equation
Subsurface hillslope runoff 3D variably saturated subsurface hillslope runoff (finite element method)
Groundwater flow 3D variably saturated subsurface hillslope runoff (finite element method)
Roads Can be incorporated based on land surface/subsurface properties
Streamflow routing Internal computation based on model topography (DEM) and hydraulic properties of land surface (i.e., specified for known stream routes)
Frozen soil No
Lakes/wetlands As boundary conditions or part of simulation
table A2.12 Integrated Hydrology Model (InHM)
130
Model name Integrated Hydrology Model (InHM)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE No
Evapotranspiration No
Water balance No
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow Yes
Watershed runoff Yes
Other Water quality; sediment erosion implemented at the University of Stanford
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, geologic layering, vegetative cover
Meteorology data Off-line processing to generate specified rainfall and ET rates
Overstorey vegetation parameters
None
Understorey vegetation parameters
None
Soil parameters (saturated zone)
Hydraulic conductivity, specific storage, porosity, anisotropy ratio
Soil parameters (unsaturated zone)
Soil moisture curves (vanGenuchten functions, Brooks-Corey functions, or tabulated relationships for capillary pressure-moisture content and relative permeability)
Routing Manning’s N, microtopography height, minimum mobile water depth
Data processing requirements
Climate data and interaction with land cover (vegetation) must be calculated off-line and input as specified flux across top of model domain; soil layering and geologic features must be described for entire model region and have corresponding XYZ data files
Key model assumptions or limitations
Assumptions about potential ET–actual ET relationships; no accounting for snow processes
Constraints Data requirements High; full-scale catchment models can be developed from conceptualized knowledge of soil and geologic framework
Level of expertise High; model user needs to be knowlegable of both hydrology, hydrogeology and numerical modelling
Level of effort High; considerable effort is needed to pre-process climate and landscape data for inclusion in robust groundwater flow model
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.12 continued
131
Model name MIKE-SHE
Development group DHI Water & Environment (former Danish Hydraulic Institute)
Model information URL www.dhigroup.com/Software/WaterResources/MIKESHE.aspx
Manual Yes
Tutorial Yes
Model support Yes; international support from [email protected]; training courses
Model cost $12 000
Computing requirements Equipment PC
Software GIS (ArcView)
Source code available No
Model type Fully-distributed
Model scales Input time step Variable; time-series pre-processor included for data management and input
Output time step User-specified
Grid size Centimetres to hundreds of metres; uniform grid size
Application scale Soil columns, research plots, watersheds
Planning scale Flexible
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation No
Rainfall interception Maximum threshold for interception
Snow accumulation Threshold dependent (user specified)
Snowmelt Degree-day method
Snow interception Same as rain
Evapotranspiration Kristensen and Jensen (1975) method (AET estimated from PET, LAI, root zone parameters)
Infiltration 1D by Richards equation, simplified two-layer root zone model, or gravity model; vertical unsaturated flow calculated for each model grid cell or larger areas of common soil properties (grouped)
Overland flow 2D diffusion-wave equation
Subsurface hillslope runoff 3D saturated flow (finite difference method) for flow occuring below the water table; fast and slow linear reservoirs as alternative approach
Groundwater flow 3D saturated flow (finite difference method)or linear reservoir approach
Roads Can be incorporated based on land surface/subsurface properties
Streamflow routing Forced by stream network file generated by MIKE-11 module
Frozen soil No
Lakes/wetlands Yes
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
table A2.13 MIKE-SHE
132
Model name MIKE-SHE
Mode outputs, cont. Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Water quality
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, soil type, geologic layering, vegetative cover
Meteorology data Temperature, precipitation, PET
Constants rain/snow temperature thresholds, ground roughness (Manning’s N, microtopography height)
Overstorey vegetation parameters
LAI, field capacity, wilting point, interception value, root zone depth, three fitting parameters
Understorey vegetation parameters
Lumped with overstorey vegetation (i.e., one vegetation layer)
Soil parameters (saturated zone)
Hydraulic conductivity, specific storage, porosity, anisotropy ratio
Soil parameters (unsaturated zone)
Built-in soil database with pedo-tranfer functions; user may specify soil moisture curves (vanGenuchten functions, Brooks-Corey functions, or tabulated relationships for capillary pressure-moisture content and relative permeability)
Routing Explicit connection to river/stream reaches when using linear reservoir approach; runoff routed to rivers by 2D overland flow; river flow calculated by MIKE-11 in parallel to MIKE-SHE simulation (by Muskingum routing to the Higher Order Dynamic Wave formulation of the Saint-Venant equations)
Data processing requirements
Most data requires GIS processing including maps (DEM, soil types, vegetation, hydrogeology) and stream network (in GIS and MIKE-11); time-series and map editors are included with MIKE-SHE to help convert raw data to MIKE-SHE standardized format
Key model assumptions or limitations
Simplified representation of snowmelt
Constraints Data requirements High; full-scale catchment models can be developed from conceptualized knowledge of soil and geologic framework
Level of expertise High; model user needs to be knowlegable of hydrology, hydrogeology, and GIS to be efficient
Level of effort High; GUI assists with model setup, but user must be able to choose between different methods/approaches for each module (overland flow, 1D unsaturated flow, 3D saturated flow)
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.13 continued
133
Model name MODHMS
Development group HydroGeoLogic Software Systems
Model information URL www.hglsoftware.com/Modhms.cfm
Manual Yes
Tutorial Yes
Model support Yes; support from [email protected]; training courses
Model cost Available for research purposes or by hiring HGL under consulting contract
Computing requirements Equipment PC
Software GIS (ArcView)
Knowledge Average GIS skills for importing spatial data; model has GUI
Source code available No
Model type Fully distributed
Model scales Input time step Variable and adaptive during simulation
Output time step Same as input; detailed output for pre-specified variables
Grid size Centimetres to hundreds of metres
Application scale Soil columns, research plots, watersheds
Planning scale Flexible
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival No
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation No
Rainfall interception Canopy interception threshold
Snow accumulation No; off-line calculation of snow accumulation is required
Snowmelt No; snowmelt may be simulated as specified flux rate
Snow interception No
Evapotranspiration Kristensen and Jensen (1975) method (AET estimated from PET, LAI, root zone parameters)
Infiltration 3D variably saturated Subsurface hillslope runoff by Richards’ equation (finite difference method)
Overland flow 2D diffusion-wave equation
Subsurface hillslope runoff 3D variably saturated Subsurface hillslope runoff by Richards’ equation (finite difference method)
Groundwater flow 3D variably saturated Subsurface hillslope runoff by Richards’ equation (finite difference method)
Roads Can be incorporated based on land surface/subsurface properties
Streamflow routing Forced by stream network linked to top of model domain; channel flow is modelled as 1D diffusive-wave equation (model accepts different channel shapes and control structures)
Frozen soil No
Lakes/wetlands Yes
table A2.14 MODHMS
134
Model name MODHMS
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Water quality
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, soil type, geologic layering, vegetative cover
Meteorology data Climate data pre-processed to generate rainfall and PET data for use in model
Overstorey vegetation parameters
LAI, field capacity, wilting point, oxic and anoxic limits, interception value, root density function, root depth, 3 fitting parameters
Understorey vegetation parameters
Lumped with overstorey vegetation (i.e., one vegetation layer)
Soil parameters (saturated zone)
Hydraulic conductivity, specific storage, porosity, anisotropy ratio
Soil parameters (unsaturated zone)
Soil moisture curves (vanGenuchten functions, Brooks-Corey functions, or tabulated relationships for capillary pressure-moisture content and relative permeability)
Routing Manning’s N, microtopography height, minimum mobile water depth
Data processing requirements
Climate data and PET must be calculated off-line; soil layering and geologic features must be described for entire model region and have corresponding XYZ data files
Key model assumptions or limitations
No representation of snowmelt
Constraints Data requirements High; full-scale catchment models can be developed from conceptualized knowledge of soil and geologic framework
Level of expertise High; model user needs to be knowlegable of hydrology, hydrogeology, and GIS to be efficient
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.14 continued
135
Model namePrecipitation-Runoff-Evapotranspiration-Hydrotope model (PREVAH)
Development group Eidgenössische Technische Hochschule (ETH); SDC Swiss Flood Forecasting Assistance Project
Model information URL http://hydrant.unibe.ch/PREVAH/index.htm
Manual Yes
Tutorial Yes
Model support Limited; from developers (Dr. Massimiliano Zappa)
Model cost None
Computing requirements Equipment PC
Software GIS (ArcInfo)
GUI WinPREVAH
Source code available Yes
Model type Semi-distributed
Model scales Input time step Daily
Output time step Daily
Grid size Landscape units (HRUs)
Application scale Small to medium watersheds
Planning scale HRUs such as hillslopes
Model calibration method Calibration runs in WinPREVAH
Hydrologic regimes simulated Glacial Yes
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Fraction of rainfall
Snow accumulation Threshold temperature
Snowmelt combination of a temperature index and an energy balance approach
Snow interception Fraction of snowfall
Evapotranspiration Physically based
Infiltration Empirical
Overland flow Empirical
Subsurface hillslope runoff Empirical
Groundwater flow Empirical (two stores)
Roads No
Streamflow routing Yes
Frozen soil No
Lakes/wetlands No
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
table A2.15 Precipitation-Runoff-Evapotranspiration-Hydrotope model (PREVAH)
136
Model namePrecipitation-Runoff-Evapotranspiration-Hydrotope model (PREVAH)
Model outputs, cont. Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Glacial melt; surface water control structures (dams, etcc)
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, land classification, soil depth, and soil classes
Meteorology data Precipitation, air temperature, air humidity, global radiation, relative sunshine duration, and wind speed
Constants Many constants depending on process representation
Overstorey vegetation parameters
Minimal stomatal resistances; root depth, LAI, vegetation density
Understorey vegetation parameters
None
Soil parameters (deep layers)
Hydraulic conductivity, field capacity
Soil parameters (root zone layers)
Hydraulic conductivity, field capacity, root depths
Routing Slope, length, roughness, shape
Data processing requirements
High; GIS required
Key model assumptions or limitations
Applied primarily in steeply sloped terrain
Constraints Data requirements High
Level of expertise Professional to academic
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.15 continued
137
Model namePrecipitation Runoff Modelling System/Modular modelling system (PRMS/MMS)
Development group USGS
Model information URL wwwbrr.cr.usgs.gov/projects/SW_MoWS/software/oui_and_mms_s/prms.shtml
Manual Yes
Tutorial Yes
Model support [email protected]
Model cost Free
Computing requirements Equipment 550 Kilobytes (Kb) of available Random Access Memory (RAM), math coprocessor, about 2Mb of hard disk space, MSDOS 6.0 or greater.
Software Contact developer
Source code available Yes (FORTRAN)
Model type Semi-distributed
Model scales Input time step Daily
Output time step Daily
Grid size N/A
Application scale Small to large watersheds
Planning scale Landscape units (GRUs) such as hillslopes
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation 1 layer
Rainfall interception Yes
Snow accumulation Temperature thresholds
Snowmelt Energy balance
Snow interception Yes, but can only vaporize
Evapotranspiration Several methods available
Infiltration Empirical
Overland flow Empirical
Subsurface hillslope runoff Linear reservoir
Groundwater flow Linear reservoir
Roads No
Streamflow routing Simplified methods used in continuous simulation option
Frozen soil No
Lakes/wetlands No
table A2.16 Precipitation Runoff Modelling System/Modular modelling system (PRMS/MMS)
138
Model namePrecipitation Runoff Modelling System/Modular modelling system (PRMS/MMS)
Model outputs Full hydrograph Simplified channel routing
Annual yield Simplified channel routing
Peak flow Simplified channel routing
Low flow Simplified channel routing
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Sediment erosion
Key required inputs (may need to refer to documentation for comprehensive list)
Map files Not specified
Meteorology data Daily precipitation, maximum and minimum air temperature, solar radiation, longwave radiation, lapse rate (for mountainous watersheds), pan evaporation
Overstorey vegetation parameters
Seasonal cover density, maximum interception storage depth on vegetation, winter cover density for the predominant vegetation above the snowpack
Understorey vegetation parameters
N/A
Soil parameters (deep layers)
Hydraulic conductivity of the transmission zone
Soil parameters (root zone layers)
Effective value of the product of capillary drive and moisture deficit at field capacity and wilting point
Routing Depending on method
Data processing requirements
Numerical specification of terrain and drainage network requires GIS data processing
Key model assumptions or limitations
Snow accumulation and melt; problems noted with abrups cessation snowmelt runoff from HRUs; simplified channel routing
Constraints Data requirements High
Level of expertise High
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.16 continued
139
Model name Regional Hydro-Ecologic Simulation System (RHESSys)
Development group Christina Tague, University of California, Santa Barbara
Model information URL http://fiesta.bren.ucsb.edu/~rhessys/
Manual Yes
Tutorial Yes
Model support [email protected]
Model cost Free
Computing requirements Equipment PC
Software GIS
Source code available Yes
Model type Semi-distributed
Model scales Input time step Daily
Output time step Daily, monthly, and yearly detailed output
Grid size Depends on DEM resolution and scale of application
Application scale Small to medium watersheds
Planning scale Patch or hillslope (used for distributing climate forcings)
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Multiple layers
Rainfall interception Yes
Snow accumulation Temperature thresholds
Snowmelt Adaptation of degree-day method to include radiation and ROS
Snow interception Yes
Evapotranspiration Penman-Montheith
Infiltration Darcy’s Law accounting for soil moisture level
Overland flow Empirical with flow velocity dependent on grid size and time step
Subsurface hillslope runoff Pixel-by-pixel routing based on topographic gradient; water mass balance and Darcy’s Law
Groundwater flow Empiral store
Roads Can be incorporated but not required
Streamflow routing Can be internal (DEM-calculated using GIS), forced (stream network file) or unit hydrograph
Frozen soil No
Lakes/wetlands No
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
table A2.17 Regional Hydro-Ecologic Simulation System (RHESSys)
140
Model name Regional Hydro-Ecologic Simulation System (RHESSys)
Model outputs, cont. Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow Yes
Watershed runoff Yes
Other Nutrient fluxes
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, hillslope, soil types, LAI, slope, aspect, impervious, land use, vegetation
Meteorology data Daily precipitation, daily maximum and minimum temperatures; can also download data from NCDC
Overstorey vegetation parameters
Fractional coverage, trunk space, height, LAI, albedo, aerodynamic attenuation, radiation attenuation, clumping factor, maximum snow interception capacity, maximum release drip ratio, snow interception efficiency, stomatal resistance (min/max), moisture threshold, vapor pressure deficit, RPC, number of root zones, root zone depths, root fractions in each zone
Understorey vegetation parameters
LAI, albedo, root fractions in each zone
Soil parameters (deep layers)
Lateral conductivity and exponential decrease, porosity
Soil parameters (root zone layers)
Surface albedo, vertical conductivity, porosity, maximum infiltration, pore size distribution, bubbling pressure, field capacity, wilting point, bulk density, thermal conductivity, thermal capacity
Routing Stream network file, stream map file, stream class file, road network file, road map file, road class file, unit hydrograph, and travel time files (last two files replace need for all preceding files)
Data processing requirements
Most data requires GIS processing including maps (topography, soil type and depth, vegetation) and streamflow routing and roads; climate data typically also needs to be processed to address data gaps and quality
Key model assumptions or limitations
None
Constraints Data requirements Data requirements are high compared to typical data availability; most physical parameters can be based on literature values
Level of expertise Senior professional to academic level; model user needs to be knowlegable in physical hydrology, computer modelling, and GIS
Level of effort High; considerable GIS and data processing required
Model adaptability Forest regrowth capability Yes
Complex stand characteristics
No; mixed stands are represented through average parameters
Forest mortality or fire risk Both under development; can currently be forced as input
Alteration of land cover details
Yes; temporal control file
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.17 continued
141
Model name Streamflow Synthesis and Reservoir Regulation (SSARR)
Development group US Army Corps of Engineers, North Pacific Region
Model information URL www.nwd-wc.usace.army.mil/report/ssarr.htm
Manual Yes
Tutorial No
Model support No
Model cost None
Computing requirements Equipment IBM-compatible personal computers
Software None
Source code available No model development as of 1991
Model type Semi-distributed
Model scales Input time step Sub-daily
Output time step Sub-daily
Grid size N/A
Application scale Medium to large watersheds
Planning scale Elevation bands
Model calibration method Manual method or external automated procudure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Yes
Snow accumulation Temperature-threshold dependent
Snowmelt Temperature-index method; distributed based on elevation bands with correction for precipitation
Snow interception Yes
Evapotranspiration Empirical; Thornthwaite formula
Infiltration Empirical; soil moisture index
Overland flow No
Subsurface hillslope runoff Empirical (direct runoff reservoir)
Groundwater flow Empirical (baseflow reservoir)
Roads No
Streamflow routing Yes, including reservoir operations and water use
Frozen soil No
Lakes/wetlands lakes
table A2.18 Streamflow Synthesis and Reservoir Regulation (SSARR)
142
Model name Streamflow Synthesis and Reservoir Regulation (SSARR)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table No
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Lakes, reservoir operation, streamflow forecasting
Key required inputs (may need to refer to documentation for comprehensive list)
Map files Not specified
Meteorology data Temperature and precipitation
Overstorey vegetation parameters
PET values, ET indices, interception capacity
Understorey vegetation parameters
N/A
Soil parameters (deep layers)
Infiltration index (baseflow reservoir), time of storage index
Soil parameters (root zone layers)
Soil moisture ET/runoff indices, surface-subsurface separation
Routing Depends on complexity of modelling
Data processing requirements
High, depending on reservoir operations complexity
Key model assumptions or limitations
Predominantly a flow forecasting and reservoir operation model; empirical representation of watershed processes
Constraints Data requirements High
Level of expertise Professional to academic
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.18 continued
143
Model name Soil and Water Assessment Tool (SWAT)
Development group Jeff Arnold (USDA Agriculture Research Service)
Model information URL www.brc.tamus.edu/swat/index.html
Manual Yes
Tutorial Yes, online help and paid workshops
Model support Nancy Sammons ([email protected]), Jeff Arnold ([email protected])
Model cost Free
Computing requirements Equipment PC or Unix workstation
Software Contact developer
Source code available Yes (FORTRAN)
Model type Semi-distributed
Model scales Input time step Daily, continuous for 1 to100 years
Output time step Detailed output (spatial variables) at user specified times
Grid size tens of metres
Application scale Small to large watersheds
Planning scale Hydrologic Response Unit (HRU) landscape units
Model calibration method
Internal automated procedure
Hydrologic regimes simulated
Glacial No
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Yes
Snow accumulation Rain/snow temperature thresholds
Snowmelt Yes; controlled by air and snowpack temperature, the melting rate and the areal coverage of snow
Snow interception No
Evapotranspiration Several methods
Infiltration Simplified as precipitation minus runoff
Overland flow SCS curve number is generated
Subsurface hillslope runoff Reservoir
Groundwater flow Reservoir
Roads No
Streamflow routing Flow is routed through the channel using a variable storage coefficient method developed by Williams (1969) or the Muskingum routing method
Frozen soil No
Lakes/wetlands Lakes
table A2.19 Soil and Water Assessment Tool (SWAT)
144
Model name Soil and Water Assessment Tool (SWAT)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Sediment erosion, nutrient fluxes, water quality
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, landuse/land cover
Meteorology data Daily precipitation, maximum/minimum air temperature, solar raditation, wind speed and relative humidity; can be from observed data records or generated during simulation
Overstorey vegetation parameters
LAI, canopy height, root depth, stomatal conductances, nitrogen uptake parameters
Understorey vegetation parameters
N/A
Soil parameters (deep layers)
Deep aquifer percolation fraction, specific yield, groundwater delay time, recharge delay time, baseflow recession constant
Soil parameters (root zone layers)
Soil hydrologic group, root depth, water capacity, hydraulic conductivity, percent sand/silt/clay, texture
Routing Users are required to define the width and depth of the channel when filled to the top of the bank as well as the channel length, side slope, and Manning’s N
Data processing requirements
Most data requires GIS processing including maps (topography, soil type and depth, vegetation) and streamflow routing; climate data typically also needs to be processed
Key model assumptions or limitations
Empirical snowmelt and ET, SCS Curve Number runoff method
Constraints Data requirements High
Level of expertise High
Level of effort High
Model adaptability Forest regrowth capability Yes (crop regrowth)
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.19 continued
145
Model name UBC–UF Peak Flow Model
Development group University of British Columbia, University of Freiburg
Model information URL None yet
Manual Yes
Tutorial No
Model support No
Model cost None (IDL software will need to be purchased)
Computing requirements Equipment PC workstation with a minimum of 512Mb RAM
Software SAGA, R, IDL
Model type Fully-distributed
Model scales Input time step Daily
Output time step Same as input
Grid size 10s to 100s of metres (currently working with 25 and 400 m spatial resolution)
Application scale meso- to macroscale
Planning scale Impact scale due to land cover changes or forest management can be as small as 1 ha
Model calibration method The model is designed to simulate ungauged basins; however, the model could also be calibrated; current development to interface the model with parameter estimation and uncertainty software
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation Pine coverage, general land cover types (no distinction in overstorey and understorey)
Rainfall interception Canopy rainfall interception factor
Snow accumulation Rain/snow temperature thresholds
Snowmelt Rain/snow temperature thresholds, degree-day model
Snow interception Currently under development
Evapotranspiration Factor reduction for forest, grassland, and clearcut
Infiltration Only indirectly included by application of runoff factors
Overland flow Mapping of areas on which dominantly Hortonian overland flow occurs using land cover characteristics (roads, burned areas) and connectivity to the stream
Saturation overland flow Mapping of areas on which dominantly saturation overland flow occurs using vertical distance to open water and topographic index; validated in several watersheds in BC
Subsurface hillslope runoff Mapping of areas on which dominantly subsurface runoff occurs using overland flow distance and gradient to the stream
Groundwater flow No representation of groundwater flow
Roads Can be incorporated into the delineation of the dominant runoff processes (Hortonian Overland Flow)
Streamflow routing Currently under development
Frozen soil No
Lakes/wetlands No
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow No; another model being developed for this
SWE Yes
Evapotranspiration Yes
Water balance Yes
table A2.20 UBC–UF Peak Flow Model
146
Model name UBC–UF Peak Flow Model
Model outputs, cont. Soil moisture No
Infiltration No
Water table No
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater No
Road flow As Hortonian overland flow areas
Watershed runoff Yes
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, vegetation types, disturbances (MPB, clearcut), hydrology (streams, river network), roads
Meteorology data Temperature, precipitation
Constants Rain/snow temperature thresholds, degree-day factor, interception factors, runoff factors, snowmelt factors, flow velocity parameters
Vegetation parameters Pine coverage, general land cover types (no distinction in overstorey and understorey)
Understorey vegetation parameters
No
Soil parameters (deep layers)
No
Soil parameters (root zone layers)
No
Routing Stream network file, travel time files
Data processing requirements
Most data requires GIS processing (precipitation, climate, topography, land cover, streamflow routing); climate data typically also needs to be processed to address data gaps and quality; a database of BC containing data (DEM, climate, forest cover, pine cover (%), dominant runoff process (DRP), MPB (infestation at 400 m will be provided)
Key model assumptions or limitations
For use in BC and focused on peak flows (model has additional claimed output capabilities as per above list)
Constraints Data requirements Data requirements are medium compared to typical data availability (about 15 parameters); database at 400 m spatial resolution will be available
Level of expertise Professional to academic level; model user needs to be knowlegable in physical hydrology and GIS
Level of effort Medium; knowledge of GIS and basic hydrological knowledge is required
Model adaptability Forest regrowth capability No; dynamic vegetation module is included; changes in vegetation cover can be reflected by manipulating the input data
Complex stand characteristics
No; within a single pixel mixed stands would be represented through average parameters (interception factor)
Forest mortality or fire risk Yes; under current development to interface the model with parameter estimation and uncertainty software
Alteration of land cover details
Yes; model inputs can be adapted as data becomes available
Future climate data Yes
table A2.20 continued
147
Model name University of British Columbia Watershed Model (UBCWM)
Development group University of British Columbia
Model information URL None
Manual Yes; Quick (1995)
Tutorial Not known
Model support Limited
Model cost None
Computing requirements Equipment PC
Software None
Source code available Yes
Model type Semi-distributed
Model scales Input time step Daily
Output time step Hourly or daily
Grid size Model is divided into elevation bands
Application scale Small to large watersheds
Planning scale Elevation bands
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial Yes
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Canopy treated as single layer for each elevation band
Rainfall interception Yes
Snow accumulation Elevation and temperature dependent
Snowmelt Energy balance method based on daily temperature range
Snow interception Yes
Evapotranspiration Empirical
Infiltration Yes
Overland flow Linear reservoir method, dependent on soil moisture, treated as “fast” flow to watershed outlet
Subsurface hillslope runoff Linear reservoir method, dependent on soil moisture, treated as “interflow” to watershed outlet
Groundwater flow Linear reservoir method, dependent on soil moisture, treated as either “slow” or “very slow” flow to watershed outlet
Roads No
Streamflow routing Yes
Frozen soil No
Lakes/wetlands Lakes
table A2.21 University of British Columbia Watershed Model (UBCWM)
148
Model name University of British Columbia Watershed Model (UBCWM)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table No
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Glacial melt
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, land classification (forest, open, lake, glacier)
Meteorology data Daily precipitation, daily mean temperature, daily PET, correction factors for elevation and gauge errors
Overstorey vegetation parameters
Crown closure
Understorey vegetation parameters
None
Soil parameters (deep layers)
Empiriral reservoir parameters
Soil parameters (root zone layers)
Field capacity, lower limit for ET
Routing Linear reservoir parameters
Data processing requirements
Medium
Key model assumptions or limitations
Use of elevation bands and simplified forest cover representation
Constraints Data requirements Medium
Level of expertise High
Level of effort Medium
Model adaptability Forest regrowth capability UBCWM has been linked with TIPSY, a growth and yield program developed by MoF that provides electronic access to the managed stand yield tables generated by TASS (Tree and Stand Simulator) and SYLVER (Silvicultural treatments on Yield, Lumber Value, and Economic Return)
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.21 continued
149
Model name Wasserhaushalts-Simulations-Modell (WaSiM-ETH)
Development group Eidgenössische Technische Hochschule (ETH)
Model information URL www.wasim.ch/en/index.html
Manual Yes (in English)
Tutorial Yes
Model support Limited; from developers
Model cost None
Computing requirements Equipment PC
Software GIS (ArcInfo)
Source code available Yes
Model type Fully distributed
Model scales Input time step Sub-daily
Output time step Sub-daily
Grid size From cm to several km
Application scale Small to large watersheds
Planning scale Flexible
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial Yes
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Layered vegetation
Rainfall interception Fraction of rainfall
Snow accumulation Threshold temperature
Snowmelt Combination of a temperature index and an energy balance approach
Snow interception Fraction of snowfall
Evapotranspiration Physically based
Infiltration TOPMODEL or 3D richards
Overland flow Yes
Subsurface hillslope runoff TOPMODEL or 3D richards
Groundwater flow Reservoir or 3D Richards
Roads No
Streamflow routing Yes
Frozen soil No
Lakes/wetlands Lakes
table A2.22 Wasserhaushalts-Simulations-Modell (WaSiM-ETH)
150
Model name Wasserhaushalts-Simulations-Modell (WaSiM-ETH)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other Lakes, glacial melt, surface water control structures (dams, etc), water quality
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, land classification, soil depth, and soil classes
Meteorology data Precipitation, air temperature, air humidity, global radiation, relative sunshine duration, and wind speed
Constants Many constants depending on process representation
Overstorey vegetation parameters
Minimal stomata resistances, root depth, LAI, vegetation density
Understorey vegetation parameters
None
Soil parameters (deep layers)
Method dependent
Soil parameters (root zone layers)
Method dependent
Routing Slope, length, roughness, shape
Data processing requirements
High; GIS required
Key model assumptions or limitations
None
Constraints Data requirements High
Level of expertise Professional to academic
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.22 continued
151
Model name Water Balance Model for BC (based on QUALHYMO model)
Development group The British Columbia Inter-Governmental Partnership (BCIGP)
Model information URL http://bc.waterbalance.ca/
Manual Yes; set of articles on website
Tutorial No
Model support Yes; [email protected]
Model cost Free 30-day trial account (registration required); full version costs $1000 annually
Computing requirements Equipment PC, Apple, or Unix workstation with a minimum of 512Mb RAM
Software Flash Player
Knowledge Limited - very user friendly
Source code available Not known
Model type Black box (lumped)
Model scales Input time step Sub-daily
Output time step Same
Grid size N/A
Application scale Small sites to small/medium watersheds
Planning scale Urban sites
Model calibration method Not applicable
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Not known
Hydrologic processes modelled
Vegetation Not known
Rainfall interception Not known
Snow accumulation Not known
Snowmelt Yes
Snow interception Not known
Evapotranspiration Yes
Infiltration Yes
Overland flow SCS runoff curves
Subsurface hillslope runoff Yes
Groundwater flow Conceptual store
Roads As impervious areas
Streamflow routing No
Frozen soil No
Lakes/wetlands No
table A2.23 Water Balance Model for BC (based on QUALHYMO model)
152
Model name Water Balance Model for BC (based on QUALHYMO model)
Model outputs Full hydrograph No
Annual yield No
Peak flow No
Low flow No
SWE No
Evapotranspiration No
Water balance Model computes site water balances
Soil moisture No
Infiltration Yes
Water table No
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Site runoff
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files None
Meteorology data User selects climate data (parameters unknown) from a list of weather stations
Constants Retardance roughness, rational coefficient
Overstorey vegetation parameters
None
Understorey vegetation parameters
None
Soil parameters (deep layers)
None
Soil parameters (root zone layers)
Native soil type, coverage area and depth, percent composition of clay and sand, (model then assigns hydraulic conductivity, maximum water content, field capacity, and wilting point), land use, and surface conditions (% impervious)
Routing None
Data processing requirements
Minimal, if any, data processing requirements
Key model assumptions or limitations
Urban stormwater runoff model
Constraints Data requirements Fairly minimal, but also limits accuracy of results
Level of expertise Low; user friendly interface, and guidance for each input; no expert knowledge of hydrology or previous modelling experience is necessary
Level of effort Very low; initial runs can be completed in minutes
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes; model inputs can be adapted to reflect urban landscape
Future climate data No
table A2.23 continued
153
Model name Variable Infiltration Capacity (VIC) model
Development group University of Washington
Model information URL www.hydro.washington.edu/Lettenmaier/Models/VIC/
Manual Limited (web pages)
Tutorial No
Model support Limited support available by special arrangement ([email protected]; VIC user support list server)
Model cost None (GIS software will need to be purchased)
Computing requirements Equipment UNIX, freebsd, Linux, and DOS operating systems
Software GIS (ArcInfo), Gnu C-compiler (freeware)
Source code available Yes
Model type Macroscale model with statistical representation of subgrid variability (snow elevation bands; fractional coverage of soil and vegetation types)
Model scales Input time step Sub-daily
Output time step Same as input; detailed output (spatial variables) at user specified times
Grid size 1/8 to 2 degrees (~10 to 200 km)
Application scale Medium to large watersheds to global scale
Planning scale 25 km2 or greater
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Overstorey (trees) and understorey
Rainfall interception canopy interception of rainfall, evaporation of intercepted water, and transpiration
Snow accumulation Rain/snow temperature thresholds
Snowmelt Snow accumulation and melt using an energy balance approach for a two-layer model
Snow interception Canopy interception of snowfall, accumulation of intercepted snow and rain, and loss of accumulated snow due to melting and blowing
Evapotranspiration Energy-based calculations for vegetation and soil
Infiltration Conceptual (variable infiltration curves)
Overland flow Not simulated
Subsurface hillslope runoff The upper soil layer is designed to represent dynamic response to rainfall/snowmelt events (surface runoff), while the lower layer is used to simulate seasonal soil moisture behaviour (ARNO baseflow model)
Groundwater flow (aquifers) Not simulated
Roads Not simulated
Streamflow routing Runoff and baseflow are routed to edge of grid cell and subsequently routed through stream network to outlet
Frozen soil Yes
Lakes/wetlands Yes
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
table A2.24 Variable Infiltration Capacity (VIC) model
154
Model name Variable Infiltration Capacity (VIC) model
Model outputs, cont. Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater No
Road flow No
Watershed runoff Yes
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, land cover, soil type
Meteorology data At a minimum temperature and precipitation at model input time step; additional meteorological data (atmospheric density and pressure, shortwave radiation, vapor pressure, wind speed) can also be specified as model input; alternatively they are calculated internally in model
Constants Height for wind speed data, height for humidity data, rain/snow temperature thresholds, minimum observable wind speed; fraction of grid cell that receives precipitation (sub-grid precipitation variability), snowpack roughness
Vegetation parameters Number of vegetation types in grid cell, fraction of grid cell covered by vegetation type, number of root zones, root zone thicknesses, root fraction in each zone, LAI (one per month), flag for overstorey present, architectural resistance, minimum stomatal resistance, shortwave albedo, roughness length, displacement height, shortwave radiation ET threshold, radiation attenuation factor, wind attenuation, trunk ratio
Soil parameters Variable infiltation curve parameter, maximum velocity of baseflow, fraction of maximum velocity where non-linear baseflow begins, fraction of maximum soil moisture where non-linear baseflow occurs, baseflow exponent, saturated hydraulic conductivity, exponent for variation of conductivity with soil moisture, soil moisture diffusion parameter, initial moisture content, average soil temperature, soil thermal damping depth, bulk density, field capacity, wilting point, residual moisture content, soil roughness
Routing Flow direction file, flow velocity file, flow diffusion file; grid cell contributing fraction
Total number of different inputs
47
Data processing requirements Most data requires GIS processing including maps (topography, soil type and depth, vegetation) and streamflow routing and roads; climate data typically also needs to be processed to address data gaps and quality
Key model assumptions or limitations
Macroscale model not applicable to typical forest management situations
Constraints Data requirements VIC data requirements are high compared to typical data availability; most physical parameters can be based on literature values
Level of expertise Senior professional to academic level; model user needs to be knowlegable in physical hydrology and GIS
Level of effort High; considerable GIS and meteorological data processing required
Model adaptability Forest regrowth capability No; however, VIC input can be manipulated to reflect input from forest growth models
Complex stand characteristics No; mixed stands would be represented through average parameters
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.24 continued
155
Model name WATFLOOD
Development group University of Waterloo
Model information URL www.civil.uwaterloo.ca/Watflood/intro/intro.htm
Manual Yes; www.civil.uwaterloo.ca/Watflood/Manual/manualstart.htm
Tutorial No; example data available
Model support Limited
Model cost $2 000
Computing requirements Equipment PC
Software Excel or Grapher for output; additional support programs from WATFLOOD developers for data preparation and output preparation
Source code available No
Model type Semi-distributed
Model scales Input time step Hours
Output time step Hours
Grid size Kilometre scale; grid has many GRUs of common properties, which are summed for more efficient computation
Application scale 20 to 2 million km2
Planning scale GRUs such as hillslopes
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Maximum threshold for interception; scaled as a function of PET
Snow accumulation Depth depends on percent of snow-covered area, variable for different land classes
Snowmelt Temperature-index and radiation-temperature algorithms
Snow interception No
Evapotranspiration PET by Priestly-Taylor or Hargreaves methods, AET estimated from PET reduction
Infiltration Philip formula (based on Green and Ampt approach)
Overland flow Infiltration excess, depression storage, and Manning equation for channel flow
Subsurface hillslope runoff Interflow (upper zone storage) by storage-discharge relation
Groundwater flow No; drainage to lower zone is calculated, but groundwater flow is not simulated explicitly, and baseflow is calculated by power function
Roads No
Streamflow routing Storage routing technique for simplified channel cross-sections
Frozen soil No
Lakes/wetlands Yes
table A2.25 WATFLOOD
156
Model name WATFLOOD
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table No
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater Yes
Road flow No
Watershed runoff Yes
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM
Meteorology data Temperature, rain/snow gauge data (or radar data), radiation data
Overstorey vegetation parameters
Forest vegetation coefficient for PET-AET reduction (tall versus short vegetation), interception fractions
Understorey vegetation parameters
Lumped with overstorey vegetation (i.e., one vegetation layer)
Soil parameters (saturated zone)
Empirical parameters
Soil parameters (unsaturated zone)
Soil moisture and temperature coefficients for PET-AET reduction, depth and resistance of interflow layer
Routing Channel roughness, bankfull versus drainage table
Data processing requirements
Model can be run with Green-Kenue GUI
Key model assumptions or limitations
Simplified snowmelt calculations; more rigorous routines have not been incorporated as “this would significantly complicate the model and require considerably more detailed information about the spatial variations of terrain, aspect, vegetation cover and meteorologic conditions”
Constraints Data requirements High
Level of expertise High
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.25 continued
157
Model name Water Erosion Protection Project (WEPP)
Development group USDA Agricultural Research Service (ARS), Soil Conservation Service (SCS)
Model information URL www.ars.usda.gov/Research/docs.htm?docid=10621
Manual Yes
Tutorial Yes
Model support [email protected]
Model cost Free
Computing requirements Equipment IBM/compatible personal computers running under MS-DOS 5.0+ operating system environments
Software Cligen, ver 4.3, GIS (ArcInfo)
Source code available Contact developer
Model type Semi-distributed
Model scales Input time step Daily
Output time step Same
Grid size Hillslope
Application scale Small or large watersheds
Planning scale Hillslope
Model calibration method Manual or external automated procedure
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed Yes
Hydrologic processes modelled
Vegetation Single layer
Rainfall interception Yes
Snow accumulation Estimates snow on the ground on hourly or daily basis
Snowmelt Energy budget incorporating air temperature, solar radiation, vapour transfer, and precipitation
Snow interception Same as rainfall (no distinction made)
Evapotranspiration Yes; simplified (Ritchie eqn)
Infiltration Green and Ampt equation
Overland flow Broad sheet flow is assumed for overland flow using two sets of regression equations, one for peak runoff rate and one for runoff duration
Subsurface hillslope runoff Algorithm based on DRAINMOD, effect of water table fluctuations on runoff and erosion is simulated
Groundwater flow No
Roads Stand-alone representation of roads for erosion calculations only
Streamflow routing Simplified methods
Frozen soil No
Lakes/wetlands No
table A2.26 Water Erosion Protection Project (WEPP)
158
Model name Water Erosion Protection Project (WEPP)
Model outputs Full hydrograph Simplified channel routing
Annual yield Simplified channel routing
Peak flow Simplified channel routing
Low flow Simplified channel routing
SWE Yes
Evapotranspiration Yes
Water balance Yes
Soil moisture Yes
Infiltration Yes
Water table Yes
Overland flow Yes
Subsurface hillslope runoff Yes
Groundwater No
Road flow No
Watershed runoff Yes
Other Sediment erosion
Key required inputs (may need to refer to documentation for comprehensive list)
Map files DEM, vegetation types, soil types
Meteorology data Daily data for: precipitation amount, duration, time to peak rainfall, peak rainfall, temperature (maximum, minimum, and dew point), temperatures, solar radiation, average wind speed and direction
Overstorey vegetation parameters
Type of vegetation (crop or range), plant growth parameters, tillage sequences and effects on soil surface and residue, dates of harvesting or grazing; if necessary, description of irrigation, weed control, burning and contouring
Understorey vegetation parameters
Lumped with overstorey
Soil parameters (deep layers)
(For up to 10 layers) thickness, initial bulk density, initial hydraulic conductivity, field capacity, wilting point, cation exchange capacity, and contents of: sand, clay, organic matter, and rock fragments
Soil parameters (root zone layers)
Surface Albedo, initial saturation, interrill and rill erodibility, and critical shear
Routing Yes
Data processing requirements
High
Key model assumptions or limitations
Predominantly set up as hillslope model with limited watershed-scale utility; empirical ET, simplified channel routing
Constraints Data requirements High
Level of expertise High
Level of effort High
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
Yes, but no time varying input; snapshots only
Future climate data Yes; input can be adjusted for alternative scenarios
table A2.26 continued
159
Model nameWater Resources Evaluation of Non-Point Silvicultural Source (WRENSS)
Development group US-EPA developed original WRENSS procedure; Robert Swanson(Canadian Forestry Service) developed WinWrnsHyd
Model information URL www.epa.gov/warsss/rrisc/handbook.htm
Manual Yes, WRENSS handbook
Tutorial No
Model support Limited
Model cost Free
Computing requirements Equipment PC workstation (Microsoft Access is not available for Macs)
Software Microsoft Access 2000 or later version
Source code available No
Model type Lumped
Model scales Input time step Monthly
Output time step Annual
Space discretization Lumped, empirical/black box
Application scale Small to medium watersheds
Planning scale Watershed
Model calibration method Not required but can be calibrated to streamflow
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No; separated into rain-dominated procedure (PROSPER) and snow-dominated procedure (WATBAL)
Hydrologic processes modelled
Vegetation Specified as cover type
Rainfall interception Unclear
Snow accumulation Yes
Snowmelt Based on the WATBAL hydrologic model
Snow interception Yes, interception coefficient
Evapotranspiration Conceptual based on precipitation (WATBAL) or PET (PROSPER)
Infiltration No
Overland flow No
Subsurface hillslope runoff No
Groundwater flow No
Roads No
Streamflow routing No
table A2.27 Water Resources Evaluation of Non-Point Silvicultural Source (WRENSS)
160
Model nameWater Resources Evaluation of Non-Point Silvicultural Source (WRENSS)
Model outputs Full hydrograph No
Annual yield Yes
Peak flow In WinWrnsHyd (untested)
Low flow No
Swe No
Evapotranspiration Yes
Water balance Yes
Soil moisture No
Infiltration No
Water table No
Overland flow No
Subsurface hillslope runoff No
Groundwater No
Road flow No
Watershed runoff No
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files None
Meteorology data From database
Constants Evapotranspiration modifier coefficient; elevation, aspect, latitude, forest cover type, silvicultural application, rooting depth, wind speed
Overstory vegetation parameters LAI
Understory vegetation parameters No understorey/overstorey discrimination
Soil parameters (deep layers) none
Soil parameters (root zone layers) none
Routing none
Data processing requirements All of the parameters needed to operate WRENSS are available from sets of regionalized curves for all forested regions in the United States and Canada.
Key model assumptions or limitations
WATBAL was intended to be a site-specific model, but geographic regional coefficients and modifiers “will yield reasonable results which are applicable in the respective regions”. Claimed but untested capability to assess peak flow changes.
Constraints Data requirements Low (about 10 separate inputs)
Level of expertise Low
Level of effort Low
Model adaptability Forest regrowth capability By specifying growth curves
Complex stand characteristics No
Fire risk or forest mortality No
Alteration of land cover details Yes; model allows for time varying forest cover
Future climate data No
table A2.27 continued
161
Model name Equivalent Cut Area–Alberta (ECA-AB)
Development group US-EPA developed original WRENSS procedure; ECA-AB by Dr. Uldis Silins, University of Alberta
Model information URL www.ales.ualberta.ca/rr/silins.cfm
Manual No
Tutorial No
Model support Limited
Model cost None
Computing requirements Equipment PC
Software Excel
Source code available No
Model type Lumped
Model scales Input time step Monthly
Output time step Annual
Space discretization Lumped, empirical/black box
Application scale Small to medium watersheds
Planning scale Watershed
Model calibration method Not required but can be calibrated to streamflow
Hydrologic regimes simulated Glacial No
Nival Yes
Pluvial Yes
Mixed No; separated into rain-dominated procedure (PROSPER) and snow-dominated procedure (WATBAL)
Hydrologic processes modelled
Vegetation Specified as cover type
Rainfall interception Unclear
Snow accumulation Yes
Snowmelt Based on the WATBAL hydrologic model
Snow interception Yes, interception coefficient
Evapotranspiration Conceptual based on precipitation (WATBAL) or PET (PROSPER)
Infiltration No
Overland flow No
Subsurface hillslope runoff No
Groundwater flow No
Roads No
Streamflow routing No
Model outputs Full hydrograph No
Annual yield Yes
Peak flow No
Low flow No
SWE No
Evapotranspiration No
table A2.28 Equivalent Cut Area–Alberta (ECA-AB)
162
Model name Equivalent Cut Area–Alberta (ECA-AB)
Model outputs, cont. Water balance No
Soil moisture No
Infiltration No
Water table No
Overland flow No
Subsurface hillslope runoff No
Groundwater No
Road flow No
Watershed runoff No
Other None
Key required inputs (may need to refer to documentation for comprehensive list)
Map files None
Meteorology data Average annual precipitation and streamflow
Constants None
Overstorey vegetation parameters
Tree species (pine, white spruce, black spruce, deciduous), area of harvest and total watershed, year of harvest, quality of site (poor, fair, good)
Understorey vegetation parameters
No understorey/overstorey discrimination
Soil parameters (deep layers)
None
Soil parameters (root zone layers)
None
Routing None
Data processing requirements
Model requires user supplied information on long-term precipitation and streamflow (in the watershed or regional averages) to estimate changes in ET and streamflow resulting from forest disturbance
Key model assumptions or limitations
Estimate of cumulative effect of various harvest events and empirically derived hydrologic recovery
Constraints Data requirements Low (about 10 separate input types)
Level of expertise Low
Level of effort Low
Model adaptability Forest regrowth capability By specifying growth curves
Complex stand characteristics
No
Fire risk or forest mortality No
Alteration of land cover details
Yes; model allows for time varying forest cover
Future climate data No
table A2.28 continued
163
Model name Water Resources Management Model (WRMM)
Development group Alberta Environment (Tom Tang)
Model information URL www3.gov.ab.ca/env/water/regions/ssrb/wrmmoutput/WRMM/index.asp
Manual Yes
Tutorial No
Model support Contact developer ([email protected])
Model cost None
Computing requirements Equipment Microcomputers
Software None
Source code available Unlikely; proprietory software
Model type River Nodes
Model scales Input time step Days
Output time step Days
Grid size N/A
Application scale Dentritic river networks
Planning scale N/A
Model calibration method N/A
Hydrologic regimes simulated Glacial N/A; model uses natural measured or computed streamflows as input
Nival N/A; model uses natural measured or computed streamflows as input
Pluvial N/A; model uses natural measured or computed streamflows as input
Mixed N/A; model uses natural measured or computed streamflows as input
Hydrologic processes modelled
Vegetation None
Rainfall interception No
Snow accumulation No
Snowmelt No
Snow interception No
Evapotranspiration No
Infiltration No
Overland flow No
Subsurface hillslope runoff No
Groundwater flow No
Roads No
Streamflow routing Yes, including reservoir operations, power generation, and water use
Frozen soil No
Lakes/wetlands No
table A2.29 Water Resources Management Model (WRMM)
164
Model name Water Resources Management Model (WRMM)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE No
Evapotranspiration No
Water balance No
Soil moisture No
Infiltration No
Water table No
Overland flow No
Subsurface hillslope runoff No
Groundwater No
Road flow No
Watershed runoff No
Other Reservoir operation, power generation, water use
Key required inputs (may need to refer to documentation for comprehensive list)
Map files No
Meteorology data No
Overstorey vegetation parameters
No
Understorey vegetation parameters
No
Soil parameters (deep layers)
No
Soil parameters (root zone layers)
No
Routing Natural flows
Data processing requirements
Depending on reservoir operations complexity
Key model assumptions or limitations
Not a watershed hydrology model
Constraints Data requirements Depending on reservoir operations complexity
Level of expertise Medium to high
Level of effort Depending on reservoir operations complexity
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
No
Future climate data Yes; using streamflow forecasts from hydrologic models
table A2.29 continued
165
Model name Water Use Analysis Model (WUAM)
Development group Economics and Conservation Branch, Ecosystem Sciences and Evaluation Directorate, Environment Canada
Model information URL None
Manual 100-page demonstration
Tutorial 100-page demonstration
Model support Contact developer (Atef Kassem, Environment Canada, Ottawa, [email protected])
Model cost None
Computing requirements Equipment microcomputers
Software None
Source code available Contact developer (Atef Kassem, Environment Canada, Ottawa, [email protected] )
Model type River Nodes
Model scales Input time step Monthly
Output time step Monthly
Grid size N/A
Application scale Dentritic river networks
Planning scale N/A
Model calibration method N/A
Hydrologic regimes simulated Glacial N/A; model uses natural measured or computed streamflows as input
Nival N/A; model uses natural measured or computed streamflows as input
Pluvial N/A; model uses natural measured or computed streamflows as input
Mixed N/A; model uses natural measured or computed streamflows as input
Hydrologic processes modelled
Vegetation None
Rainfall interception No
Snow accumulation No
Snowmelt No
Snow interception No
Evapotranspiration No
Infiltration No
Overland flow No
Subsurface hillslope runoff No
Groundwater flow No
Roads No
Streamflow routing Yes, including reservoir operations and water use
Frozen soil No
Lakes/wetlands No
table A2.30 Water Use Analysis Model (WUAM)
166
Model name Water Use Analysis Model (WUAM)
Model outputs Full hydrograph Yes
Annual yield Yes
Peak flow Yes
Low flow Yes
SWE No
Evapotranspiration No
Water balance No
Soil moisture No
Infiltration No
Water table No
Overland flow No
Subsurface hillslope runoff No
Groundwater No
Road flow No
Watershed runoff No
Other Reservoir operation, water use
Key required inputs (may need to refer to documentation for comprehensive list)
Map files No
Meteorology data No
Overstorey vegetation parameters
No
Understorey vegetation parameters
No
Soil parameters (deep layers)
No
Soil parameters (root zone layers)
No
Routing Natural flows
Data processing requirements
Depending on reservoir operations complexity
Key model assumptions or limitations
Not a watershed hydrology model
Constraints Data requirements Depending on reservoir operations complexity
Level of expertise Medium to high
Level of effort Depending on reservoir operations complexity
Model adaptability Forest regrowth capability No
Complex stand characteristics
No
Forest mortality or fire risk No
Alteration of land cover details
No
Future climate data Yes; using streamflow forecasts from hydrologic models
table A2.30 continued
Science, In
novation
, and
Sustainab
ility: Investin
g in
British
Colum
bia’s K
now
ledg
e-based
Natural R
esource Sector
Cover and text
2 0 0 9
F O R R E X S E R I E S 2 5
Cover and text
ISSN 1495-964X
TM
Review of Hydrologic Models for Forest Management and Climate Change Applications in British Columbia and Alberta
Forrex Forum for Research and Extension in Natural Resources
