T E C H N I C A L R E P O R T 0 3 5

2 0 0 6

Ministry of Forests and Range Forest Science Program

Community Structure of Forage Plants Consumed by Black Bears in the Nimpkish Valley, British Columbia

Ministry of Forests and RangeForest Science Program

Community Structure of Forage PlantsConsumed by Black Bears in the Nimpkish Valley, British Columbia

Volker Michelfelder and Kenneth P. Lertzman

The use of trade, firm, or corporation names in this publication is for the informationand convenience of the reader. Such use does not constitute an official endorsement or approval by the Government of British Columbia of any product or service to theexclusion of any others that may also be suitable. Contents of this report are presentedfor discussion purposes only. Funding assistance does not imply endorsement of anystatements or information contained herein by the Government of British Columbia.Uniform Resource Locators (URLs), addresses, and contact information contained inthis document are current at the time of printing unless otherwise noted.

CitationMichelfelder, V., K.P. Lertzman 2006. Community structure of forage plants consumedby black bears in the Nimpkish Valley, British Columbia. B.C. Min. For. Range, Res.Br., Victoria, B.C. Tech. Rep. 035.<http://www.for.gov.bc.ca/hfd/pubs/Docs/Tr/Tr035.htm>

Prepared byVolker MichelfelderB.C. Ministry of Environment4051 18th Avenue Prince George, BC

Kenneth P. LertzmanSchool of Resource and Environmental ManagementSimon Fraser UniversityBurnaby, BC

Prepared forB.C. Ministry of Forests and RangeResearch Branch Victoria, BC

© 2006 Province of British Columbia

Copies of this report may be obtained, depending upon supply, from:Government PublicationsPO Box 9452 Stn Prov GovtVictoria, BC

1-800-663-6105http://www.publications.gov.bc.ca

For more information on Forest Science Program publications, visit our web site at: <http://www.for.gov.bc.ca/scripts/hfd/pubs/hfdcatalog/index.asp>.

Library and Archives Canada Cataloguing in Publication DataMichelfelder, Volker.

Community structure of forage plants consumed by black bears in the Nimpkish Valley, British Columbia

(Technical report ; 035)

"Prepared for B.C. Ministry of Forests and Range, Research Branch."--P.Includes bibliographical references: p. 0-7726-5518-9

1. Forage plants - British Columbia - Nimpkish River Valley. 2. Forests and forestry - Environmental aspects - British Columbia - Nimpkish River Valley 3. Black bear - Feeding and feeds - British Columbia - Nimpkish River Valley. 4. Black bear - Effect of habitat modification on - British Columbia - NimpkishRiver Valley. I. Lertzman, Kenneth P. (Kenneth Peter), 1956- . II. British Colum-bia. Forest Science Program. III. British Columbia. Ministry of Forests and Range.Research Branch. IV. Title. V. Series: Technical report (British Columbia. ForestScience Program) ; 35.

193.3.352 2006 633.209711'2 2006-960047-3

ABSTRACT

Coastal British Columbia is the focus of potential conflict between timberextraction and protection of forest components such as forage plants con-sumed by black bears. To protect foraging habitat and to enhance forageplant abundance, researchers and managers must understand the habitat requirements of these forage plants.

We assessed the community structure of plants consumed by black bearsin the Nimpkish Valley, British Columbia, with respect to environmentalgradients.

Non-metric multidimensional scaling () ordination revealed thatvariation in community structure of forage plants was related to soil nutrientand moisture content, elevation, and tree overstorey dominance. Of thesethree factors, soil nutrient and moisture content had the strongest relationsto forage plant communities. Species richness of forage plants and abun-dance of invasive forage plants generally increased with increasing soilnutrient and moisture content and with decreasing tree overstorey domi-nance. In contrast to invasive forage plants, residual forage plants did notrespond consistently to any of the three factors indicated by the ordination,although abundance of many residual forage plants was depressed where treeoverstorey dominance was high.

To maintain the quality of black bear foraging habitat, forest managersshould: (1) prioritize some forests with nutrient-rich and moist soils for protection; (2) ensure that these forested sites are distributed across the biogeoclimatic variants occupied by black bears; and (3) where harvestingoccurs in these forested sites, avoid harvesting regimes that create large areaswith a dense, structurally homogeneous tree cover.

iii

ACKNOWLEDGEMENTS

Several people contributed substantially to this research project. We wish tothank Alton Harestad for providing considerable advice on various drafts ofthis manuscript. Thanks to Doug Heard, Mark Kepkay, Sarb Mann, SaleemDar, and Jason Smith for their very useful editorial comments. Special thanksto Helen Davis because she provided the data used in the statistical analysisand reviewed this manuscript. And finally, thanks to Tony Hamilton, whoinitiated this project and provided financial support.

This research was funded by a National Science and Engineering ResearchCouncil of Canada Post Graduate Scholarship A award and a grant from theB.C. Ministry of Environment. Data collection for this analysis was fundedby the Habitat Silviculture Protection Account of the joint Federal andProvincial Forest Resource Development Agreement, the Habitat Conserva-tion Trust Fund, Canadian Forest Products Ltd., and Wildlife HabitatCanada.

iv

TABLE OF CONTENTS

v

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Influence of Biotic and Environmental Factors on

Community Structure of Understorey Plants . . . . . . . . . . . . . . . . . 11.2 Life History Strategies of Understorey Plants . . . . . . . . . . . . . . . . . 21.3 Importance of Plant Communities to Management of

Black Bear Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.4 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.1 Data Collection and Experimental Design. . . . . . . . . . . . . . . . . . . . 52.2 Plot Location and Study Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3 Construction of the Data Sets and Outlier Analysis . . . . . . . . . . . . 72.4 Species Diversity Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.5 Multivariate Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.1 Identification of Bear Forage Plants . . . . . . . . . . . . . . . . . . . . . . . . . 123.2 Species Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.3 Gradients Extracted by the Ordination . . . . . . . . . . . . . . . . . . . . . . 143.4 Patterns of Abundance of Forage Plant Groups

in Relation to Environmental Gradients . . . . . . . . . . . . . . . . . . . . . 173.5 Patterns of Forage Plant Abundance Grouped by Season

of Consumption in Relation to Environmental Gradients. . . . . . . 233.6 Patterns of Forage Plant Abundance across Environmental

Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274.1 General Trends in Bear Forage Plant Species Composition

and Abundance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274.2 Relationship between Community Structure of Forage

Plants and Dominance of the Tree Overstorey . . . . . . . . . . . . . . . . 304.3 Relationship between Community Structure of Forage

Plants and Soil Nutrient and Moisture Status . . . . . . . . . . . . . . . . . 334.4 Relationship between Community Structure of Forage

Plants and Elevation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.5 Relationship between Community Structure of Forage

Plants and Environmental Units. . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.6 Management Implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.7 Study Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374.8 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5 Literature Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

1 Comparison of stages within models of forest development . . . . . . . . . 7

2 Environmental matrix variables related to assemblages of forage plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Important autecological characteristics of black bear forage plants occurring in the Nimpkish Valley, B.C. . . . . . . . . . . . . . . . . . . . . . . . . . . 10

vi

1 Location of the study area in the Nimpkish Valley on Vancouver Island, British Columbia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Relationship between Alpha and Beta diversity of forage plants by structural stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3 Ordination of 133 plots in forage species space . . . . . . . . . . . . . . . . . . . . 15

4 Ordination of 133 plots in forage species space with correlation vectors of quantitative and environmental variables . . . . . . . . . . . . . . . 16

5 Ordination of 133 plots in forage species space showing correlation vectors of both species diversity and abundance for groups of forage plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6 Scatter plots of Alpha diversity and the abundance of groups of invasive forage species in relation to ordination axis 2 . . . . . . . . . . . 19

7 Scatter plot of the abundance of graminoids in relation to ordination axis 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

8 Scatter plot of the abundance of Vaccinium forage species in all plots and all residual forage species in relation to ordination axis 3 . . . 21

9 Scatter plot of the proportional summed abundance of all residual forage species in relation to ordination axis 2 . . . . . . . . . . . . . . . . . . . . . 22

10 Scatter plot of the abundance of Dryopteris expansa in relation to ordination axis 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

11 Scatter plot of the abundance of Gaultheria shallon in relation to ordination axis 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

12 Ordination of 133 plots in forage species space showing correlation vectors of forage plants grouped by seasonal importance . . . . . . . . . . . 24

13 Summary of significant results related to the gradient of increasing soil nutrient and moisture content . . . . . . . . . . . . . . . . . . . . 28

14 Summary of significant results related to the gradient of increasing elevation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

15 Summary of significant species abundance results related to the gradient of increasing dominance and contiguity of the tree overstorey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

16 Summary of significant stand structure results related to the gradient of increasing dominance and contiguity of the tree overstorey . . . . . . 29

4 Measures of species diversity for all plants and for bear forage plant groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5 comparison of community structure of forage plants consumed by black bears in the Nimpkish Valley, B.C. . . . . . . . . . . . . . 25

6 Significant maximum indicator values for forage plants consumed by black bears in the Nimpkish Valley, B.C. . . . . . . . 26

7 Multiple comparison of community structure of forage plants between structural stages in the CWHxm2 and vm1 variants . . . 27

8 Management guidelines for choosing forested sites for protection and active management of foraging habitat in the CWH to meet objectives of maintaining the short- and long-term abundance of forage plants, grouped by season of consumption. . . . . . . . . . . . . . . . . . 36

1 INTRODUCTION

A broad array of ecological factors interact at multiple scales to form the vegetation patterns in forests. Forested landscapes, at a scale of up to 20 000ha (Lertzman et al. 1996), contain understorey plant communities that arethe result of complex interactions between deterministic and stochastic fac-tors. These factors are related to plant life history traits (Grime 1977, 2001;Noble and Slatyer 1980; Rabotnov 1986; Grime et al. 1988; Tilman 1988; McCook 1994; Noble and Habiba 1996), site history, disturbance regime, and other biotic and environmental factors (Eis 1981; Haeussler and Coates1986; Halpern 1988; Tilman 1988; Grace 1990; Karakatsoulis and Kimmins1993). How do these diverse factors interact to shape the community struc-ture of understorey plants?

To extract dominant patterns from the complex multi-scale interactionsthat shape plant communities, researchers in British Columbia developedand refined the Biogeoclimatic Ecosystem Classification () system (e.g.,Krajina 1965, 1969; Pojar et al. 1987; Meidinger and Pojar 1991; MacKinnon1992; Green and Klinka 1994). Studies employing the system emphasizethree primary environmental factors that affect the community structure ofunderstorey plants: climate, soil nutrient status, and soil moisture regime.Other studies in North America also found similar relations with one ormore of these three factors (e.g., Arsenault and Bradfield 1995; Swanson et al.1997; Brockway 1998; Ohmann and Spies 1998; Wimberly and Spies 2001;Ohmann and Gregory 2002; Chan et al. 2003). The system is limited,however, by its inability to incorporate important dynamics affecting thecommunity structure of understorey plants, such as temporal changes inspecies diversity after disturbance (i.e., forest succession).

Soil nutrient and moisture availability often limit plant growth and productivity (Kozlowski 1972; Haeussler and Coates 1986). Plant speciescomposition and abundance in forested ecosystems vary along gradients ofsoil nutrient and moisture content (e.g., Gagnon and Bradfield 1987; Klinkaet al. 1989; Karakatsoulis and Kimmins 1993; Halpern and Spies 1995;Hutchinson et al. 1999). The complex interactions among biotic and environmental factors often confound relationships between communitystructure of understorey plants and individual environmental gradients (e.g., Ohmann and Spies 1998). The relative importance of soil nutrient andmoisture content may obscure that of other biotic and environmental factorswith respect to the community structure of understorey plants at the land-scape scale.

In mountainous coastal areas, elevation and continental influence oftendominate the regional climate, and consequently the community structure of understorey plants (Gagnon and Bradfield 1987; Davis 1996). At the land-scape scale of up to 20 000 ha, environmental factors affect the micro- andmeso-climates, which in turn influence the community structure of under-storey plants. Locally, fine-scale topographical features, such as variations in slope and aspect, often further moderate the prevailing regional climate,adding their influence to plant community structure (Haeussler and Coates1986; Klinka et al. 1989; Meidinger and Pojar 1991).

Disturbance and the corollary dynamics of forest succession add spatialand temporal variation to the community structure of understorey plants.This variation is not accounted for in the system at scales relevant to

1.1 Influence of Bioticand Environmental

Factors on CommunityStructure of

Understorey Plants

management of wildlife such as black bears (Ursus americanus). Abundantresources exist for many herbs and shrubs in open, post-disturbance environments. However, after stands reach the canopy closure stage, theonce-abundant resources quickly diminish, often temporarily fostering de-pauperate understoreys with low overall species abundance and richness,especially on productive sites (Alaback 1982a; Oliver and Larson 1990;Halpern and Spies 1995; Franklin et al. 2002). This occurs because intermedi-ate-aged stands tend to have dense, fine-textured canopies, which allow littlelight penetration (Wells 1996). After forest stands reach maturity, variousprocesses, including the mortality of dominant canopy trees, initiate an in-crease in understorey diversity and productivity in canopy gaps (Alaback and Herman 1988; Franklin et al. 2002). Older stands, particularly in coastalBritish Columbia, usually have more canopy gaps than do intermediate seralstages. These gaps allow more light to reach the forest floor, increasing thediversity and abundance of understorey species (Lertzman et al. 1996; Wells1996; Franklin et al. 2002). It is not fully understood how forest successionaldynamics interact with other biotic and abiotic factors to shape the commu-nity structure of understorey plants.

Many ecologists agree that complex successional pathways followed by com-munities of understorey plants can be conceptually simplified by groupingplant species according to similar life history strategies and autecologicalcharacteristics (Connell and Slatyer 1977; Grime 1977, 2001; Noble and Slatyer1980; Alaback 1982b; Tilman 1988; Gitay and Noble 1997; Shugart 1997;Woodward and Kelly 1997; Platt and Connell 2003). One of the broadest classifications of life history strategies applicable to the forests of coastalBritish Columbia categorizes understorey plants relative to their occurrence in undisturbed forests and disturbed areas (Dyrness 1973). Invasive plantspecies (invasives) are either absent in the understorey of undisturbed conif-erous forests (but not necessarily from the seed bed) or restricted to stronglydisturbed microsites, such as sites dominated by windthrow (Schoonmakerand McKee 1988; Halpern 1989). These disturbed microsites often contain adeciduous overstorey. Residual plant species (residuals) are present in undis-turbed coniferous forests prior to disturbance, even if only sparsely (Dyrness1973; Halpern 1989). Invasives and residuals exist along a continuum of di-verse life history strategies, ranging from early-seral invasives, or ruderals,such as the annual Senecio vulgaris (Grime 1977; Rabotnov 1986), to late-seralresiduals, or forest-interior species, such as the perennial Goodyera oblongifo-lia (Halpern et al. 1999).

Ruderal invasives are predominantly adapted to conditions prevailingafter intense, spatially unpredictable disturbance (Grime 1977, 2001; Rabot-nov 1986; Grime et al. 1988; Bazzaz 1996), including harvesting systemsleaving low (or no) residual trees (Halpern and McKenzie 2001). Studies ofclearcutting in the Oregon Cascades confirm that mature ruderal invasive individuals are most common immediately after disturbance, and least inundisturbed forests (Dyrness 1973; Schoonmaker and McKee 1988; Halpern1989). Ruderals rarely depend upon vegetative propagation; they usually either arrive after disturbance via seeds capable of long-distance dispersal, orpersist as seeds through the disturbance (Grime 1977; Noble and Slatyer 1980;Bell 1991). Storing long-lived seeds in seed banks allows ruderals to rapidlyexploit a site immediately after disturbance (see review by Oakley andFranklin 1998), which usually stimulates germination of dormant seeds in

1.2 Life HistoryStrategies of

Understorey Plants

the soil (Bazzaz 1996). Ruderal invasives are particularly adapted to exploitthe ephemeral period of low competition that immediately follows distur-bance (Grime 1977, 2001; Noble and Slatyer 1980). These adaptations includefast growth, relatively small shoots, and short generations, maximizing seedproduction (Rabotnov 1986; Campbell and Grime 1992). Early-successionalplants such as ruderals can also dramatically increase photosynthesis ratesfollowing intense disturbance in response to increased light (Bazzaz 1996).

Adaptations to intense and frequent disturbance involve physiologicaltrade-offs. Many ruderal early-seral invasives are poor competitors, have low shade tolerance, and rarely invest in vegetative propagation (Connell and Slatyer 1977; Grime 1977; Halpern 1989; Bell 1991; Bazzaz 1996), which facilitates competition in dense vegetation (Grime 2001). Consequently, ruderal invasives flourish immediately after disturbance when competition is low, but then decrease as understorey competition intensifies (Noble andSlatyer 1980) until many understorey species become (temporarily) sup-pressed by overstorey crown closure (Franklin et al. 2002).

Late-seral forest-interior residuals adopt life history strategies in starkcontrast to those of ruderal invasives. Late-seral residuals are especiallyadapted to late-successional environments (R species of Noble and Slatyer1980; R5 species of Halpern 1989; Halpern et al. 1999), characterized by rela-tively low disturbance and limited light (patients of Ramenskii 1935, 1938,cited in Rabotnov 1986; stress tolerators of Grime 1977, 2001). These adapta-tions include slow growth, evergreen habit, greater shade tolerance, andlonger-lived organs than ruderal invaders (Connell and Slatyer 1977; Meier et al. 1995; Bazzaz 1996). Many forest-interior herbs have a higher proportionof below-ground biomass than do early-seral species, indicating more fre-quent below-ground vegetative propagation and nutrient storage capabilities(see citations in Lezberg et al. 1999).

Adaptations to late-seral environments also involve fitness trade-offs.Late-seral, forest-interior residuals disperse propagules slowly, with longergenerations than ruderals. Residual species are also typically unable to sub-stantially increase photosynthesis in response to increased light (Meier et al.1995; Jules 1999; Grime 2001). Forest-interior species such as Chimaphila umbellata require the survival of individuals for vegetative propagation afterdisturbance (Halpern 1989). Studies from the Pacific Northwest show thatmany late-seral forest interior species are not resilient to disturbance, requir-ing many years to recover (Schoonmaker and McKee 1988; Meier et al. 1995;Jules 1997, 1999; Halpern et al. 1999; Grime 2001).

The life history strategies of ruderal invasives and late-seral forest-interiorresiduals represent the extremes of a continuum of understorey life historystrategies. The life history strategies of most shrubs and herbs in coastal tem-perate rainforests of Canada’s west coast fall between these extremes (e.g.,Haeussler and Coates 1986). The low competitive ability and shade toleranceof many annual ruderal invasives often increases with increasing longevity(Grime 1977; Halpern 1989; Bell 1991; McCook 1994). Longer-lived perennialinvasives often proliferate vegetatively as well as by seed, in contrast to mostannual ruderals, explaining why longer-lived invader species are often moreabundant than annual ruderals immediately after stand-replacing distur-bances such as clearcutting (Schoonmaker and McKee 1988; Halpern 1989;Hannerz and Hanell 1993; Grime 2001).

The spectrum of life history strategies includes residuals such as Rubusursinus (a forest-interior subordinate) and Gaultheria shallon (a forest

understorey dominant), which are very tolerant of disturbance (Bunnell1990; Tappeiner and Zasada 1993), often increasing in abundance followingclearcut harvesting (Haeussler and Coates 1986; Halpern 1989; Coates et al.1990). Many other residuals experience an initial decline, followed by an in-crease in abundance (Minore 1984; Haeussler and Coates 1986; Tappeinerand Alaback 1989; Bunnell 1990; Lezberg et al. 1999).

Mature invasives and residuals differ in their abilities to integrate as biological legacies into the post-disturbance environment (Franklin et al.2002). Invasive establishment and maturity are usually restricted to post-disturbance environments. In contrast, mature residuals can remain in thepost-disturbance environment, because they are integral components of theinterior plant community prior to disturbance, and are also relatively re-silient to disturbance (some stress tolerators of Grime 1977; patients ofRabotnov 1986).

Coastal ecosystems are among the most productive sites in British Columbia,making them important areas for timber harvesting. In addition to their eco-nomic value, these forest ecosystems support understorey plants essential asforage for American black bears and grizzly bears (Ursus arctos) (Hamilton1987; Hamilton et al. 1991; Nagy and MacHutchon 1991; Pojar et al. 1991).

Because of the potential for conflict between timber management andwildlife habitat requirements (Bunnell et al. 1999), managers are challengedto increase protection of foraging habitat and actively enhance the abun-dance of forage plants (e.g., Garcia 1985; Noyce and Coy 1990; Deal 2001;Stamp 2003). There is an increasing need to understand habitat requirementsof understorey plant species consumed by wildlife. One question faced bymanagers is “should specific habitats be set aside to maintain the abundanceof forage plants?” If the answer is yes, then we need to determine if and howdifferent habitats are important for different groups of forage plants. Wouldspecific silvicultural prescriptions aid in maintaining or increasing the long-term abundance of forage plants?

There are few studies on the community structure of understorey plantsthat focus on bear forage plants. Although in this study we focus on forageplants important to black bears, other studies on forage plants important togrizzly bears provide additional information, since in coastal British Colum-bia, many plants are consumed by both species (G. MacHutchon, unpublishedreport). Researchers can thus make limited inferences about grizzly bearhabitat based on patterns of forage plants consumed by black bears.

Prior studies evaluated either one plant species at a time (e.g.,MacHutchon et al. 1993; Noyce and Coy 1990) or else assessed the aggregatedabundances of plant species (e.g., Stamp 2003). Although these approachesare appealing because they are simple, they sacrifice ecological insight andstatistical power and are unlikely to resolve complex patterns among plantspecies and communities across environmental gradients. It is difficult toevaluate the combined response of a group of forage species to simultane-ously acting (and interacting) environmental factors (McCune and Grace2002), and the interactions among plant species’ responses. We are not awareof a single study in this region directly quantifying the relative strength andnature of the relationships between community structure of forage plantsconsumed by black bears (or grizzly bears) and various abiotic and bioticfactors.

1.3 Importance ofPlant Communities toManagement of Black

Bear Habitat

Our objective is to examine the species composition and abundance (com-munity structure) of invasive and residual plant species consumed by blackbears in the western variant of the very dry maritime Coastal Western Hem-lock zone (CWHxm2), and the submontane and montane variants of thevery wet maritime subzone (CWHvm1 and CWHvm2, respectively) (Pojar et al. 1987; Meidinger and Pojar 1991). Specifically, we aim to describe the nature and strength of the relationships between the community structure of these forage plants and key environmental gradients and units (e.g., siteseries and structural stages) using a multivariate approach. Results from ouranalyses may be used by managers to better maintain and enhance the abun-dance of forage plants for bears in these ecosystems.

2 METHODS

Data were collected in the Nimpkish Valley, approximately 40 km south ofPort McNeill on northern Vancouver Island, British Columbia, Canada(50°27' latitude, 127°06' longitude) (Figure 1). Davis (1996) assessed thearray of forest stands over 20 000 ha available to black bears for denning, encompassing the home ranges of 21 radio-collared black bears. In 1994,Davis collected all of the data used in this study using random point

2.1 Data Collectionand Experimental

Design

1.4 Objective

Location of the study area in the Nimpkish Valley on Vancouver Island, BritishColumbia. (1999 hillshade coverage from the B.C. Ministry of Forests; 2003biogeoclimatic coverage from the B.C. Ministry of Sustainable ResourceManagement.)

sampling (Marcum and Loftsgaarden 1980). Davis (1996) classified 170 randomly located plots by site series and structural stage and measured vege-tation and stand structural attributes in these plots. To assess the full range of environmental gradients of the forage plants in the analysis, 133 of the 170plots were included in this analysis even though some may have been locatedin parts of the study area where bears do not frequently forage.

Davis (1996) conducted vegetation inventories using 20 × 20-m plots. Abiotic and percentage cover data of plant species in all vegetation layersranging from the moss layer to the dominant tree layer were collected(Luttmerding et al. 1990). To estimate stocking density of live and dead trees,Davis established variable-radius prism cruise plots at the centre of each veg-etation plot (Bull et al. 1990) and measured diameter at breast height of liveand dead trees broader than 17.5 cm. Mean basal area of trees per ha was thenestimated.

Climate is an important determinant of terrestrial ecosystems (Meidingerand Pojar 1991). Within the hierarchical system, ecosystems are classifiedat the biogeoclimatic zone level to reflect variation in regional climate. With-in zones, subzones reflect climatic differences at this scale. Variants withinsubzones represent the influence of finer-scale climatic variation.

We limited our analysis to Davis’ (1996) plots within three biogeoclimaticvariants of the Coastal Western Hemlock (CWH) biogeoclimatic zone. Weexcluded the 19 plots in previously thinned and spaced stands, as well as several outliers (see Section 2.3). Of the remaining 133 plots, 86 were in thewestern very dry maritime variant (CWHxm2), occurring in valley bottomsusually below 400 m elevation; 26 plots were in the very wet submontanevariant (CWHvm1) and 21 plots in the montane variant (CWHvm2). TheCWHvm1 variant occurs above the CWHxm2 variant to about 600 m elevation, and the CWHvm2 variant occurs above the CWHvm1 variant toabout 800 m elevation (Green and Klinka 1994) (Figure 1). The topographyand landforms of the study area are typical of the Northern Island Moun-tains ecosection of the Western Vancouver Island ecoregion (Demarchi1995). The mean annual precipitation over a 30-year period in the study areawas 3244 mm, with only 18% of the annual total occurring in the summer(Hamann and Wang 2005; Wang et al. 2006).

The study area encompasses a range of disturbance histories and foreststructural stages. It is located within Tree Farm Licence #37, which, at thetime of data collection, contained second-growth stands up to 70 years old(Davis 1996). Recent disturbances in the study area included prescribedburns, wildfires, windthrow, juvenile spacing, thinning, and regenerationplanting.

We grouped plots into structural stages based on a seral classification sys-tem frequently used for describing ecosystems in British Columbia, Canada(B.C. Ministry of Environment Lands and Parks and B.C. Ministry of Forests1998). This classification of structural stages is comparable to other classifi-cations (Oliver and Larson 1990; Franklin et al. 2002), but omits somedistinct stages of stand development (Table 1). At the time of sampling, 42 plots were in the shrub/herb structural stage, 14 in the pole/sapling stage, 34 in the young forest stage, 25 in the mature forest stage, and 18 in the oldforest stage.

2.2 Plot Location andStudy Areas

Plots-by-species matrix In collaboration with T. Hamilton (Forest WildlifeBiologist, B.C. Ministry of Environment) and based on a comprehensive com-pilation of plants consumed by black bears in the CWH zone (G. MacHutchon,unpublished report), we developed the list of forage plants included in thisstudy, but did not distinguish preferred forage species because no informa-tion was available. We constructed a 133 × 28 plots-by-species matrix contain-ing percentage cover of plant species consumed by black bears in the CWHzone.

Various data manipulations, including transformation of the raw datainto the Daubermire cover scale (Daubenmire 1959) substantially improvedthe data distribution with respect to multivariate analysis assumptions. Forthe final plots-by-species matrix, percentage cover data were log-transformedafter adding one to eliminate zeroes. Log-transformation reduces the infl-uence of multivariate outliers and the coefficient of variation (Zar 1999;Tabachnik and Fidell 2001), and decreases the influence of the dominantspecies on the data analysis (van der Maarel 1979).

We also excluded five multivariate outlier plots with a compositional dis-similarity greater than 2.3 standard deviations from the average to removetheir bias on multivariate analyses (Tabachnik and Fidell 2001; McCune andGrace 2002). The sample units with multivariate outliers all representedhigh-elevation locations (above 800 m) within the old forest structural stage,leaving only two high-elevation sample units within this structural stage.Thus, the remaining 133 plots do not allow for generalizations with respect tohigh-elevation locations within the old forest structural stage.

Environmental matrix We constructed a matrix of plots-by-environmentalvariables consisting of categorical and quantitative environmental variablesand forest mensuration data. The environmental data included elevation andsoil nutrient and moisture conditions (Table 2). Forest mensuration data in-cluded percentage canopy cover of trees, stocking density, diameter at breastheight, and basal area of live and dead trees (Table 2).

2.3 Construction ofthe Data Sets and

Outlier Analysis

Comparison of stages within models of forest development (adapted from Franklin et al. 2002)

Classification system and development stage

B.C. Ministry of Environment,

Lands and Parks and B.C. Oliver and Franklin et al.

Typical stand agea Ministry of Forests (1998)b Larson (1990) (2002)c

20 yr (3) Shrub/herb Stand initiation Cohort establishmentCanopy closure

30 yr (4) Pole/sapling Stem exclusion(5) Young forest Biomass accumulation,

competitive exclusion80–250 yr (6) Mature forest Understorey reinitiation Maturation 150–250+ yr (7) Old forest Old-growth Vertical diversification

Horizontal diversification

800 yr Pioneer cohort loss

a B.C. Ministry of Environment Lands and Parks and B.C. Ministry of Forests (1998) data are limited to the CWH zone. b Used in this study.c Franklin et al. (2002) data are based on Douglas-fir (Pseudotsuga menziesii)–western hemlock (Tsuga heterophylla) forests

in the Pacific Northwest.

Variables related to assemblages of forage plants and environmental factors

Variable name Character Units or symbol description Variable typea

Variables related to terrain, soil, and climatic conditionsELEV elevation metres QSoil nutrient classb soil nutrient class L=low, H=high, ?=unknown CSoil moisture classb soil moisture class L=low, M=medium, H=high CBiogeoclimatic variant biogeoclimatic variant xm2, vm1, vm2 C

Variables related to the tree overstoreyTREET crown cover of all trees percentage cover per plot Q

>2 m tallA1–A3 crown cover of trees percentage cover per plot Q

>10 m tallBASAL basal area of trees and snags m2 per ha Q

with diameter at breast height >17.5 cm

MEAND mean diameter at breast cm Qheight of trees and snags >17.5 cm

DBH3 live trees with an average cm Qdiameter of 57–76 cm

GREEN all live stems stems per ha QN/A structural stage 3=shrub/herb, 4=pole/sapling, C

5=young forest, 6=mature forest, 7=old forest

Variables related to assemblages of forage plantsAlpha richness of forage plants number of species per plot QTOTAL abundance of all forage plants percentage cover per plot QSHADEc abundance of shade-intolerant percentage cover per plot Q

forage speciesMOISTc abundance of very moist to percentage cover per plot Q

very wet soil indicator forage species

Nhighc abundance of nutrient-rich percentage cover per plot Qsoil indicator forage species

% RESID abundance of all residuals percentage cover per plot Qdivided by TOTAL

INVA abundance of all invasive percentage cover per plot Qforage plants

InSh abundance of all invasive percentage cover per plot Qforage shrubs

InFo abundance of all invasive percentage cover per plot Qforage forbs

GRAM abundance of all forage percentage cover per plot Qgraminoids

Rubus abundance of all forage percentage cover per plot QRubus species

VACCT abundance of all forage percentage cover per plot QVaccinium species excluding V. parvifolium

Continued on page 9

To assess the variation in the summed abundance (percent cover) ofgroups of forage species, we constructed a plots-by-traits matrix summariz-ing percentage cover values of forage plants by life history trait, season ofconsumption, functional group membership, and indicator status (Table 2).We tabulated data on the autecological characteristics (Table 3), classifyingspecies as residuals or invaders following the classifications of Dyrness (1973),Dyrness et al. (1974), Halpern (1988, 1989), Schoonmaker and McKee (1988),and Halpern et al. (1992) and autecological accounts (see Table 3). To con-struct the matrix, species were assigned dummy grouping variables by lifehistory trait (1=trait membership, 0=no trait membership) (McCune andGrace 2002). Then, we multiplied this matrix by the plots-by-species matrixcontaining raw species percentage cover values to produce the plots-by-traitsmatrix. In the environmental and plots-by-traits matrix, we log-transformedall variables with a skewness greater than 1.1 (van der Maarel 1979; Zar 1999;Limpert et al. 2001; McCune and Grace 2002). The number and influence ofunivariate outliers in the environmental matrix were substantially reduced asa result, reducing the variation and skewness in the data.

Continued

Variable name Character Units or symbol description Variable typea

Variables related to assemblages of forage plants (continued)VACCPAR abundance of Vaccinium percentage cover per plot Q

parvifoliumGAULSHA abundance of Gaultheria shallon percentage cover per plot QDRYOEXP abundance of Dryopteris expansa percentage cover per plot QATHYFIL abundance of Athyrium percentage cover per plot

filix-femina QTRAU abundance of Trautvetteria percentage cover per plot Q

caroliniensis

Variables related to forage plants grouped by season of consumptiond

Fol S foliage important as spring percentage cover per plot Qforage

Fol Su foliage important as summer percentage cover per plot Qforage

Fol F foliage important as fall percentage cover per plot Qforage

Fr Su fruit important as summer percentage cover per plot Q

forage

a Q=quantitative (continuous), C=categorical.b See Table 6 for details.c See Table 3 for details.d Adapted from MacHutchon, G. 1996. Grizzly bear and black bear foods described for the Coastal Western Hemlock and north-

ern Interior Cedar-Hemlock biogeoclimatic zones of British Columbia. Unpublished report.

Important autecological characteristics of black bear forage plants occurring in the Nimpkish Valley, B.C.a,b

Soil nutrient Soil moisture Wetland

Species regime regimec Elevationd Shade tolerance indicator

Residual speciesTrautvetteria caroliniensis rich F–VM decreasee tolerant–intolerant ?Cornus stolonifera rich VM–W decreasee tolerant–intolerant indicativeAthyrium filix-femina rich VM–W all tolerant facultativeLonicera involucrata rich VM–W low–subalpine tolerant–intolerant facultativeOplopanax horridus rich VM–W low–middle tolerant ?Lysichiton americanum rich W–VW low–middle tolerant–intolerant indicativeRosa gymnocarpa medium VD–MD decreasee tolerant–intolerant ?Dryopteris expansa medium F–VM low–subalpine tolerant facultativeGaultheria shallon poor ? decreasee tolerant facultativeVaccinium parvifolium poor ? decreasee tolerant ?Vaccinium alaskaense poor F–VM low–subalpine tolerant facultativeVaccinium ovalifolium poor F–VM low–subalpine tolerant facultativeVaccinium spp. of minor variable variable variable variable variableimportance

Invasive speciesPopulus balsamifera rich F–VM decreasee intolerant facultativessp. trichocarpa

Epilobium angustifolium rich ? decreasee very intolerant indicative Ribes bracteosum rich VM–W decreasee tolerant–intolerant facultativeRubus spectabilis rich VM–W decreasee tolerant–intolerant facultativeSambucus racemosa rich W–VM decreasee tolerant–intolerant facultativeRubus parviflorus rich ? decreasee tolerant–intolerant facultativeCirsium arvense rich ? low–middle ? ?Lactuca muralis rich F–VM low–middle tolerant–intolerant ?Rubus ursinus medium MD–F decreasee tolerant–intolerant ?Equisetum spp.f medium water-receiving low–alpine tolerant–intolerant facultative

sitesFragaria spp. none–medium none–MD–F decreasee mostly intolerant ?Carex spp. variable variable variable mostly intolerant variablePoaceae (various) variable variable variable variable variableHypochaeris radicata ? ? decreasee tolerant–intolerant ?Prunus emarginata ? moist forests low–middle ? ?

and along

streams

a Adapted from Klinka et. al. (1989), Pojar and MacKinnon (1994), and Meidinger et. al. (2002). Additional autecological information is from Grime 1977, 2001; Bierzychudek 1982; Grime et al. 1988; Klinka et al. 1989; Coates et al. 1990; Haeussler et al. 1990, 1999; Lieffers and Stadt 1994; Pojar and MacKinnon 1994; Willson and Hennon 1997.

b ?=the plant was either not evaluated or its indicator status is unknown or variable.c VD=very dry, MD=moderately dry, D=dry, F=fresh, M=moist, VM=very moist, W=wet.d Occurrence level: increase or decrease indicates trends with increasing elevation.e From Klinka et. al. (1989); all other elevation descriptions from Pojar and MacKinnon (1994). f Data from Equisetum arvense.

We calculated diversity measures for all plants, for all bear forage plants, andfor bear forage plants across various environmental units. The data set didnot differentiate between individual species of the Poaceae family and theCarex, Fragaria, Salix, and Equisetum genera. We calculated Alpha diversity,the average species richness per plot, and Beta diversity, the amount ofspecies compositional change (the total number of observed forage speciesdivided by Alpha diversity) (McCune et al. 2000; Deal 2001). We used para-metric tests of differences in Alpha diversity of bear forage species amongenvironmental units. Where we observed high heteroscedasticity in varianceamong groups, we used non-parametric tests such as the Kruskal-Wallis andMann-Whitney (i.e., Wilcoxon rank sum) (Zar 1999). The sequential Bonfer-roni correction was applied to the Mann-Whitney U-test for multiplecomparisons (Rice 1989; Wells 1996).

To analyze patterns of abundance and species composition (communitystructure) of forage plants consumed by black bears, we performed non-metric multidimensional scaling () ordination (Kruskal 1964; Mather1976) on the transformed plots-by-species matrix. is especially suited for ecological data sets because the method is robust to large numbers of zerovalues and does not assume multivariate normality (Fasham 1977; Clarke1993; Pitkanen 1997; McCune and Mefford 1999; Peterson and McCune 2001). is suitable for data on arbitrary or discontinuous scales, common incommunity ecology (McCune and Grace 2002). Unlike other ordinationmethods, such as canonical correspondence analysis (), is an un-constrained method, in that it does not attempt to maximize correlationamong matrices. Thus, correlations of the forage plants with environmentalfactors are based exclusively on similarities in community structure.

iteratively searches for a global stress minimum. A stable solution isaccepted if the same minimum stress configuration is achieved from severaldifferent starting configurations (Anderson 1971; Kenkel and Orloci 1986;Pitkanen 1997). Stress is estimated from the monotone relationship betweenthe original dissimilarity matrix and the derived distance matrix (Kruskal1964; Kruskal and Wish 1978). Distances in the original dissimilarity matrixare best measured as Sorenson’s (Bray-Curtis) distances because it is mostsuited to the distribution of ecological community data (Beals 1984; McCuneand Mefford 1999). Ordination used Sorenson’s distance between plotssummed over species, a random starting configuration, 100 runs with the realdata, a stability criterion of 0.0007, and 20 iterations (Peterson and McCune2001; McCune and Grace 2002). To ascertain that local minima were avoid-ed, we also ran the ordination once with initial configurations derived fromBray-Curtis ordination (Deal 2001; McCune and Grace 2002).

We chose the 3-dimensional solution as the final solution because it ex-plained the most variability in the community data on the fewest axes (Deal2001), producing both an acceptable stress level and interpretable resultswith respect to plant communities and environmental variables (Clarke 1993;McCune and Grace 2002). We performed a Monte Carlo test with 50 ran-domized runs to ensure that the ordination extracted stronger axes thanexpected by chance (α=0.05). We rigidly rotated the ordination, maximizingthe loading of strongest explanatory variable on axis 1. We evaluated rela-tionships of environmental variables with the ordination axis using thecoefficient of determination (R2), and scatter plots of species abundance inrelation to ordination axis scores (McCune and Grace 2002). Relationships

2.5 MultivariateAnalyses

2.4 Species DiversityMeasures

with R2>0.200 and p<0.001 are considered strong in this study; weaker values are assessed on a case-by-case basis.

To test if plots in various environmental units (such as site series andstructural stages) occupied different regions in species space (i.e., whetherthese units contain differentiated forage plant communities), we used thenon-parametric multi-response permutation procedure () (Mielke1984; Zimmerman et al. 1985) on the rank-transformed Sorensen’s distancematrix (McCune and Grace 2002). is similar to the t-test and one-wayanalysis of variance, but has more relaxed assumptions for the data structure,and is readily applied to multivariate problems. We rank-transformed theSorensen’s distance matrix to enhance the correspondence of the re-sults with the results (McCune and Grace 2002). The -statistic in thisnon-metric is a measure of the chance-corrected within-group agree-ment. If =0, community structure is no different from that expected bychance. When =1, sample units are completely homogeneous within eachgroup. In community ecology, even groups that are significantly differentoften have <0.1; >0.3 is high (McCune et al. 2000; McCune and Grace 2002).

Finally, we used indicator species analysis (Dufrene and Legendre 1997;McCune and Grace 2002) on a plots-by-species matrix containing rawspecies cover values to contrast performance of individual forage speciesacross environmental units. We evaluated the significance of indicator valuesusing a Monte Carlo randomization test with 10 000 runs. While testsfor differences among environmental units, indicator species analysis com-plements results by illustrating those differences (McCune and Grace2002). A species’ indicator value () is its relative abundance × relative fre-quency × 100. Relative abundance is a species’ average abundance in plotswithin an environmental unit divided by its average abundance across allplots. Relative frequency is the proportion of plots in which the species oc-curs within an environmental unit.

For error checking the raw data, we used pro97 (MacKenzie and Klassen1999); for all univariate analyses, we used version 10.0 ( Inc. 1999);and for all multivariate analyses, we used - version 4 (McCune andMefford 1999).

3 RESULTS

Of the 28 bear forage plants found in the study area, 15 are often invasive and13 are often residual plants (Pojar et al. 1984; Klinka et al. 1989; Meidinger etal. 2002; Table 3). Most of these forage plants have indicator value for specificsoil nutrient and/or moisture conditions. Of the 14 indicators of nutrient-richsoils, eight are often invasive. Five forage plants are indicative of nutrient-medium sites, including three that are often invasive. Fourteen forage plantsare indicators of fresh to wet soils, five of which are often invasive (Klinka etal. 1989; Meidinger et al. 2002; Table 3).

The invasive forage plants in the study area range from very shade-intolerant to shade-tolerant to -intolerant, whereas residual forage plants aremostly shade-tolerant and some are shade-tolerant to -intolerant (Klinka etal. 1989; Table 3). Four (out of the 15) invasive forage plants lie at the intoler-ant extreme of the continuum of shade tolerance; five (out of 13) residual

3.1 Identification ofBear Forage Plants

forage plants represent the tolerant extreme. Most other invasive and resid-ual forage plants cluster in the middle of the continuum of shade tolerancebecause they are shade-tolerant to -intolerant.

The 133 plots yielded a total of 136 plant species, of which 28 are consumed by black bears (Table 4). Alpha diversity (average species richness per plot ± ..) of all bear forage plants was 4.8±2.6. Alpha diversity of invasive forageplants was 2.0±2.0 and of residual plants was 2.8±1.3 (Table 4).

Alpha diversity of bear forage plants differed across environmental units(Table 4). Variation in Alpha diversity was strongly affected by nutrient class (Mann-Whitney U-test, p<0.001), moisture class (one-way ,F2,130=17.0, p<0.001), and structural stage (Kruskal-Wallis test, H4,128=41.500,p<0.001), but less to biogeoclimatic variant (Kruskal-Wallis test, H2,130=12.156,p=0.002). Alpha diversity was highest in the high nutrient class (Mann-Whitney U-test, p<0.001) and moisture class (Tukey’s , p<0.001).

3.2 Species Diversity

Measures of species diversity for all plants and for bear forage plant groups

Alpha Standard Coefficient of Beta Number of

Group diversity deviation variation (%) diversity species N

All plant species 15.2 5.5 36 9.0 136 133All shrub species 3.5 1.7 49 8.3 29 133All herb species 6.0 4.1 68 9.8 59 133All bear forage species 4.8 2.6 54 5.8 28 133

Bear forage species per functional groupInvasives 2.0 2.0 100 7.5 15 133Residuals 2.8 1.3 46 4.6 13 133

Bear forage species per BEC variantCWHxm2 4.4 2.5 57 5.7 25 86CWHvm1 6.4 2.6 41 3.3 21 26CWHvm2 4.5 2.0 44 3.1 14 21

Bear forage species per stand structural stageShrub/herb 6.5 2.4 37 3.7 24 42Pole/sapling 5.9 2.7 46 3.6 21 14Young forest 3.5 2.2 63 6.0 21 34Mature forest 2.8 1.6 57 6.4 18 25Old forest 4.4 1.8 41 3.9 17 18

Bear forage species per soil nutrient classa

Low 3.9 2.0 51 5.4 21 75High 6.0 2.8 47 4.3 26 54b

Bear forage species per soil moisture classc

Low 3.6 1.9 53 5.3 19 27Medium 4.2 2.5 60 5.0 21 62

High 6.4 2.3 36 4.1 26 44

a Low=poor to medium, high=rich to very rich (see Green and Klinka 1994). b Four plots could not be classified. c Low=very dry to medium dry, medium=slightly dry to fresh, high=very moist to wet (see Green and Klinka 1994).

Alpha diversity was higher in the youngest structural stage than in the in-termediate stages (Table 4, Figure 2). Alpha diversity in the shrub/herb stagewas not significantly different from that in the pole/sapling stage (Mann-Whitney U-test, p=0.413). In contrast, Alpha diversity in the shrub/herb stagewas significantly different from that in the young forest and mature foreststages (Mann-Whitney U-test, p=<0.001). Alpha diversity in the shrub/herbstage was not significantly different from the old forest stage (Mann-WhitneyU-test, p=0.002).

Beta diversity, or the amount of species compositional change, also dif-fered across environmental units (Table 4, Figure 2). Beta diversity of all bearforage species combined was 5.9 (n=133), higher than the value of individualenvironmental units, except the young forest and mature forest structuralstages (Table 4), which were the most heterogeneous.

8

7

6

5

4

3

2

1

0

Div

ersi

ty

Shrub/ Pole/ Young Mature Old herb sapling forest forest forest

Alpha diversity

Beta diversity

Relationship between Alpha and Beta diversity of forage plants by structuralstage. Error bars are 95% confidence intervals.

The ordination accounted for 84.8% of the variation in the data set, with35.2%, 27.6%, and 22.0% of the variation loaded on axes 1, 2, and 3, respec-tively. The first axis expressed a gradient of increasing soil nutrient andmoisture content. Plots in nutrient-rich and moist sites mostly concentratedon the right side of axis 1, whereas nutrient-poor and dry sites concentratedon the left (Figure 3a, 3b). Our interpretation was strengthened by the posi-tive correlation of the summed abundance (percentage cover) of bear foragespecies that are indicators of nutrient-rich and moist sites with axis 1 (R2=0.594and 0.658, respectively, both p<0.001, Figure 4a, 4b). The weaker second axiswas strongly related to plot elevation (R2=0.676, p<0.001; Figure 4a).

The third axis was weakest and was strongly related to increasing domi-nance and contiguity (connectedness, adjacency) of the tree overstorey,shown by the increase of various measures of overstorey abundance such astotal tree cover and basal area per ha of stems with a diameter greater than17.5 cm (all R2>0.200, p<0.001) (Figure 4b, 4c). This interpretation was supported by the negative correlation of the summed abundance of shade intolerant forage species with the gradient of overstorey dominance(R2=0.223, p<0.001).

3.3 GradientsExtracted by the

Ordination

Axi

s 2

Axis 1

LH?

Axi

s 2

Axis 1

LMH

Axi

s 2

Axis 1

xm2vm1vm2

Ordination (NMS) of 133 plots in forage species space. Axis1 is related to soil nutrient and moisture content; axis 2 isrelated to elevation. See Table 2 for classes and variantdescriptions.

(a)

(b)

(c)

Axi

s 2

Axis 1

Elev

Moist *

Nhigh*

(a)

Axi

s 3

Axis 1

A1–A3Basaldbh 3*MEAND

Shade*

TREET

Green

Moist *

Nhigh*

(b)

Axi

s 3

Axis 2

Shade*Nhigh*

Basal

MEANDdbh 3*

Elev

TREETA1–A3

(c) 34567

73

Ordination (NMS) of 133 plots in forage species space with correlation vectors ofmostly environmental variables (see Table 2 for details; * indicates log-transformation).Axis 1 is related to soil nutrient and moisture content; axis 2 is related to elevation; axis3 is related to overstorey dominance and contiguity. Vectors represent the strength(R2>0.200, p<0.001) and direction of correlations between variables and axis scores(vectors scaled to 150% for visual clarity). Arrows along axis 3 represent structuralstages (see Table 1).

The summed abundance of all bear forage plants increased with decreasingdominance and contiguity of the tree overstorey (R2>0.200, p<0.001), mostlyindependently of ordination axes 2 and 3 (Figure 5). Species richness increasedwith decreasing dominance and contiguity of the tree overstorey, and increas-ing soil nutrient and moisture content (both R2>0.200, p<0.001), mostlyindependently of the gradient related to elevation. Although species richnessdid not strongly correlate with the gradient related to elevation, species rich-ness followed a fairly uniform pattern along this gradient, being elevated atmedium elevations (Figure 6).

Similarly to the result for all forage species, invasive forage plants re-sponded strongly to the gradient related to the tree overstorey. The summedabundance of all invasive plants and invasive shrubs (particularly Rubusspp.) increased with decreasing dominance and contiguity of the tree over-storey, and with increasing soil nutrient and moisture content, mostlyindependently of the gradient related to elevation (both R2>0.200, p<0.001;Figure 5). The summed abundance of invasive forbs increased with decreas-ing dominance and contiguity of the tree overstorey, mostly independentlyof axes 2 and 3 (R2>0.200, p<0.001). The summed abundance of allgraminoids increased with increasing soil nutrient and moisture content,mostly independently of the other two ordination gradients (R2>0.200,p<0.001). Graminoid combined abundance was also lower where the treeoverstorey was very dominant (Figure 7). Summed abundance of all invasiveplants, invasive shrub abundance, and graminoid abundance were all lowestat high elevations (Figure 6).

Unlike invasive forage plants, the summed abundance of residual forageplants did not correlate strongly with the ordination (Figure 5). Summedabundance of residuals was depressed in plots with a more contiguous anddominant tree overstorey (Figure 8). To investigate further patterns of abun-dance for residual plants, we overlaid the proportional summed abundanceof all residual plants (relative to total forage plant abundance) on the ord-ination. The proportional summed abundance of residual forage plantsincreased with increasing overstorey gradient, and with decreasing soil nutri-ent and moisture content, mostly independently of increasing elevation(R2>0.200, p<0.001; Figure 5a, 5b). The number of plots with low propor-tional abundance of residuals decreased roughly with increasing elevation(Figure 9).

Some residual forage plants also correlated with the ordination at thespecies level or in sub-assemblages (assemblages of subsets of residualspecies). All Vaccinium forage species abundance responded strongly to in-creasing elevation. V. parvifolium decreased with increasing elevation, mostlyindependently of the other two ordination gradients (R2>0.200, p<0.001;Figure 5a, 5c). In contrast, the summed abundance of all other Vacciniumspecies (excluding V. parvifolium) increased with increasing elevation, most-ly independently of the other two ordination gradients (R2>0.200, p<0.001;Figure 5a 5c). The summed abundance of all Vaccinium species (excluding V. parvifolium) followed an approximately unimodal distribution along thetree overstorey gradient (Figure 8), being lowest at the extremes.

Other residual forage plant abundance increased primarily with increasingsoil nutrient and moisture content. The abundance of ferns and Trautvetteriacaroliniensis followed this pattern, mostly independently of the other two ordination gradients (R2>0.200, p<0.001; Figure 5a). Two species showingunique responses were the fern Dryopteris expansa, which was less abundant

3.4 Patterns ofAbundance of Forage

Plant Groups inRelation to

EnvironmentalGradients

Axi

s 2

Axis 1

(a)

VACCT*

DRYOEXP*ATHYFIL* TRAU*

Alpha

GAULSHA*

%RESID

VACCPAR*

GRAM* Rubus* InSh*INVA*

Axi

s 3

Axis 1

(b)

%RESID

InFo*Total

ATHYFIL*TRAU*

Alpha

GAULSHA*GRAM*

Rubus*InSh*INVA*

Axi

s 3

Axis 2

(c)

%RESID

InFo*

Total

VACCT*

Alpha

VACCPAR*

Rubus*InSh*

INVA*

Ordination (NMS) of 133 plots in forage species space showingcorrelation vectors of both species diversity and abundance forgroups of forage plants. For axes, see Figure 4; for variables, seeTable 2; for a list of species, see Table 3.

12

8

4

0

Alpha diversity(species richness)

2.0

1.0

0.0

Invasive abundance

2.0

1.0

0.0

Invasive shrub abundance(including Rubus species)

2.0

1.5

1.0

0.5

0.0

Graminoid abundance

Ordination axis 2

Scatter plots of Alpha diversity and the (log) abundance of groups of invasiveforage species in relation to ordination axis 2 (increasing elevation). The lineindicates two standard deviations of the running mean.

(a)

(b)

(c)

(d)

at lower elevation (Figure 10), and the shrub Gaultheria shallon, whose abundance increased with decreasing soil nutrients and moisture, mostly independently of the other two ordination gradients (R2>0.200, p<0.001;Figure 5). G. shallon decreased where the tree overstorey was very dominant(Figure 11). None of the other residual forage plants, either as individuals orin sub-assemblages, correlated strongly with the ordination.

2.0

1.5

1.0

0.5

0.0

Graminoid abundance

Ordination axis 3

Scatter plot of the (log) abundance of graminoids in relation to ordination axis 3(increasing dominance and contiguity of the tree overstorey). The line indicates twostandard deviations of the (log) abundance.

2.0

1.5

1.0

0.5

0.0

Vaccinium parvifolium abundance

1.2

0.8

0.4

0.0

Vaccinium species abundance(except for

V. parvifolium)

2.0

1.0

0.0

All residual species abundance

Ordination axis 3

Scatter plot of the (log) abundance of Vaccinium forage species and all residualforage species in relation to ordination axis 3 (increasing dominance and contiguityof the tree overstorey). The line indicates two standard deviations of the runningmean.

80

40

0

All residuals (proportionalsummed abundance)

Ordination axis 2

1.2

0.8

0.4

0.0

Dryopteris expansa

Ordination axis 2

Scatter plot of the proportional summed abundance of all residual forage species inrelation to ordination axis 2 (increasing elevation).

Scatter plot of the (log) abundance of Dryopteris expansa in relation to ordinationaxis 2 (increasing elevation). The line indicates two standard deviations of therunning mean.

The summed abundance of forage plants, grouped by season of consump-tion, responded strongly to the three main environmental gradients (wherestrong is R2>0.200 and p<0.001; Figure 12). The abundance of importantspring foliage forage plants increased strongly with increasing soil nutrientand moisture content, increasing elevation, and decreasing overstorey domi-nance and contiguity. The abundance of forage plants whose foliage isimportant fall forage increased strongly with both increasing soil nutrientand moisture content and increasing elevation. The abundance of both for-age plants whose foliage is important summer forage and whose fruits areimportant summer forage increased strongly with decreasing overstoreydominance and contiguity. Overlaying summed abundances of forage specieson the ordination did not reveal any strong patterns of abundance for anyother forage groups.

3.5 Patterns of ForagePlant Abundance

Grouped by Season ofConsumption in

Relation toEnvironmental

Gradients

Gaultheria shallon abundance

Ordination axis 3

2.0

1.5

1.0

0.5

0

Scatter plot of the (log) abundance of Gaultheria shallon in relation to ordinationaxis 3 (increasing dominance and contiguity of the tree overstorey). The lineindicates two standard deviations of the running mean.

Axi

s 2

Axis 1

(a)

Fol F

Fol Sp

Axi

s 3

Axis 1

(b)

Fol F

Fol SpFol Su

Fr Su

Axi

s 3

Axis 2

(c)

Fol F

Fol Sp

Fol SuFr Su

Ordination (NMS) of 133 plots in forage species space showing correlation vectors offorage plants grouped by seasonal importance. For descriptions of axes, see Figure 4.For variables, see Table 2. For species lists, see Table 3.

We used the multi-response permutation procedure () to test for dif-ferences in community structure (i.e., species composition and abundance)between environmental units, such as site series. Indicator species analysisanalyzed how various environmental units differed (McCune and Grace2002).

The ordination arranged plots roughly by soil nutrient and moistureregimes (Figure 3a, 3b), supporting the differences found in Alpha diversity(see Section 3.2). Of all environmental units studied, plots grouped by site series produced the highest effect size or -statistic, indicating thatspecies composition and abundance of bear forage plants differed the mostacross site series (Table 5). This supports the ecological criteria for distin-guishing site series (Meidinger and Pojar 1991).

Indicator species analysis revealed that many bear forage plants weresignificantly associated with plots in nutrient-rich and/or moist sites (Table6). Only Gaultheria shallon was significantly associated with both nutrient-poor and dry sites (Table 6). Vaccinium species of minor importance as bearforage were also significantly associated with dry sites (Table 6).

Along the altitudinal gradient (axis 2) of the ordination, plots roughlyclustered together by variant (Figure 3c). Although Alpha diversity betweenvariants did not differ (Kruskal-Wallis test, H2,130=12.156, p=0.002), supported community structure distinctions among variants with respect tobear forage species (=0.208, p<0.001). The performance (i.e., relative abun-dance × frequency) of Gaultheria shallon, Vaccinium parvifolium, and Rubusursinus was best within the CWHxm2 (Table 6). Other Rubus species, Epilo-bium angustifolium, Trautvetteria caroliniensis, and Oplopanax horridusperformed best within the CWHvm1 (Table 3). Vaccinium species (excludingV. parvifolium) performed best in the higher-elevation CWHvm2 variant(Table 6).

3.6 Patterns of ForagePlant Abundance

across EnvironmentalUnits

MRPP comparison of community structure of forage plants consumed by black bears in the Nimpkish Valley, B.C.

Number Chance corrected

Grouping variablea of groups within-group agreement (A) P

Variant 3 0.221 <0.001Black bear habitat typeb () 7 0.402 <0.001Site seriesb (per variant) 17 0.511 <0.001Soil moisture class (high nutrient 3 0.100 <0.001

class only)Soil nutrient class (high moisture 2 0.117 <0.001

class only)Structural stage (xm2) 5 0.136 <0.001Structural stage (vm1) 4 0.267 <0.001

Structural stage (vm2) 5 0.015 0.363

a See Tables 1, 2, and 4 for details on soil nutrient and moisture classes and structural stages. b Black bear habitat types (s) are site series combined across and within variants based on similarities in moisture and

nutrient regimes (Davis 1996).

Along the gradient of increasing dominance and contiguity of the treeoverstorey (axis 3 of the ordination), plots separated somewhat by structuralstage. Plots in the youngest stage (shrub/herb) concentrated in the lowerportion of axis 3 (Figure 4c), while plots in the young forest and mature for-est stages concentrated more in the upper portion, and plots in the old foreststage more in the middle of this axis. Plots in various structural stages alsodiffered in Alpha diversity (Kruskal-Wallis test, H4,128=41.500, p<0.001;Table 2). Young and mature forest plots extended over much of axis 3, de-picting the high species compositional heterogeneity (Table 2), also reflectedin the overlap between structural stages (see arrows in Figure 4c).

results supported ordination results, showing some differences incommunities of bear forage plants among structural stages. In the CWHxm2and vm1 variants, structural stages differed in species composition and abun-

Significant maximum indicator values (IVmax) for forage plants consumed by black bears in the Nimpkish Valley, B.C.a

IVmax per grouping variableb

Frequency of Structural stagee

occurrence Soil nutrient Soil moisture (xm2 and

Species (% of all plots) classc classd Variant vm1 only)

ResidualsAthyrium filix-femina 36 67.2 (high) 71.0 (high)Dryopteris expansa 32 43.4 (high) 27.6 (high) 25.3b (4)Trautvetteria caroliniensis 13 29.3 (high) 30.0 (high) 19.8b (vm1) 20.2b (7)Oplopanax horridus 5 11.1 (high) 13.6 (high) 23.1 (vm1)Cornus stolonifera 8 8.9b (high) 13.0 (high)Gaultheria shallon 55 72.8 (low) 58.3 (low) 46.4 (xm2)Vaccinium ovalifolium 29 23.5b (high) 70.9 (vm2) 24.1b (7)Lysichiton americanum 5 13.6 (high)Vaccinium spp. 2 7.0b (low) 9.5b (vm2)Vaccinium alaskaense 16 50.9 (vm2)Vaccinium parvifolium 77 41.1b (xm2) Rosa gymnocarpa 5 13.0b (7)

InvasivesRubus spectabilis 52 78.2 (high) 74.0 (high) 44.3 (vm1)Sambucus racemosa 17 33.1 (high) 32.4 (high) 19.9b (vm1)Rubus parviflorus 21 25.8 (high) 25.2 (vm1) 28.7 (4)Poaceae (various) 14 22.4 (high) 20.2 (high)Carex spp. 7 11.1 (high) 20.5 (high)Rubus ursinus 20 29.4 (xm2)Epilobium angustifolium 29 26.7b (vm1) 77.6 (3)Lactuca muralis 20 24.6 (3)Hypochaeris radicata 6 20.8b (3)

Prunus emarginata 5 12.5b (4)

a Species with very low occurrence cannot be statistically significant. b Significant at α=0.05. All other values are significant at α=0.01.c Low=very poor to medium, high=rich to very rich soil nutrient regime (Green and Klinka 1994). d Low=very dry to medium dry, medium=slightly dry to fresh, high=very moist to wet soil moisture regime (Green and Klinka

1994).e 3=shrub/herb, 4=pole/sapling, 5=young forest, 6=mature forest, 7=old forest structural stage (B.C. Ministry of Environment,

Lands and Parks and B.C. Ministry of Forests 1998).

dance (=0.136, 0.267 respectively, both p<0.001; Table 5), but not in theCWHvm2 (=0.015, p=0.363).

Multiple comparison confirmed further differences between plotsby structural stage in the CWHxm2 and vm1 (Table 7). The composition andabundance of bear forage plants did not differ significantly between plots in the pole/sapling and shrub/herb structural stages (=0.028, p=0.066). Differences became stronger in the older structural stages compared to theshrub/herb structural stage (young forest =0.112, p<0.001; mature forest=0.141, p<0.001; Table 7). Compared to the mature forest structural stage,the species composition and abundance of bear forage plants in the old foreststructural stage was more similar, but still significantly distinct from theshrub/herb structural stage. Because the sampling design did not permitblocking for the effects of soil nutrients and moisture content and elevation,these results must be interpreted cautiously.

None of the forage species was significantly associated with the young andmature forest structural stages (Table 6) according to indicator species analy-sis in the CWHxm2 and CWHvm1. Except for the fern Dryopteris expansa, allforage species significantly associated with the shrub/herb or pole/saplingstructural stages were invasive species. All forage species significantly associ-ated with the old forest stage were residuals (Table 6).

Multiple MRPP comparison of community structure of forage plants betweenstructural stages in the CWHxm2 and vm1 variants

Structural stages Chance corrected

compared (n) within-group agreement (A) P

3 (33) vs. 4 (10) 0.028 0.0663 (33) vs. 5 (34) 0.112 <0.0013 (33) vs. 6 (22) 0.141 <0.001

3 (33) vs. 7 (13) 0.128 <0.001

4 DISCUSSION

Most of the variation in community structure of forest plants in the studyarea (84.8%) was related to three environmental gradients (Figures 2, 3). The gradient of soil nutrient and moisture content explained most (35.2%) of the variation in community structure of forage plants (see Figure 13 for asummary). Altitude explained 27.6%, of the variation, and the gradient of increasing dominance and contiguity of the tree overstorey explained 22.0%of the variation in community structure of forage plants (see Figures 14, 15,and 16 for summaries). Species richness of forage plants and the abundanceof invasive forage plants increased with increasing soil nutrient and moisturecontent and with decreasing dominance and contiguity of the tree canopy(Figures 5, 13, 15). In contrast, the abundance of residual forage plants did notrespond consistently to any of the three environmental gradients (Figure 5),but decreased where the tree overstorey was very dominant (Figures 8, 11).Invasive forage plants generally performed best early in forest succession,whereas residual forage plants performed best in later stages (Table 6, Figure16). No forage plants performed well during intermediate successional stages

4.1 General Trends inBear Forage Plant

Species Compositionand Abundance

Low High

Indicator species response

Abundance of invasive species

Alpha diversity

Soil nutrient class

Soil moisture class

13 of 28 speciesperformed best

No speciesperformed best

Gaultheria shallonperformed best

Low

Low

High

High

HighLow

Low Medium

High

Ordination axis 1: soil moisture and nutrient content

Summary of significant results related to the gradient of increasing soil nutrient andmoisture content (ordination axis 1). For details, see Tables 4, 5, and 6 and Figure 5.

Summary of significant results related to the gradient of increasing elevation (ordinationaxis 2). For details, see Tables 4, 5, and 6 and Figures 5, 6, 9, and 10.

Low High

Indicator species response

Abundance ofVaccinium parvifolium

Abundance of Vaccinium spp.excluding V. parvifolium

Biogeoclimatic variant (CWH)

Plot elevation

Ordination axis 2: elevation

Most Vaccinium speciesperformed best

All significant indicators exceptVaccinium species performed best

HighLow

HighLow

vm2xm2

LowHigh

vm1

Low High

Proportional abundance ofresidual species

Alpha diversity

Abundance of invasive species

Abundance of all foragespecies

Dominance and contiguity oftree overstorey

Ordination axis 3: dominance and contiguity of the tree overstorey

HighLow

LowHigh

LowHigh

LowHigh

HighLow

Summary of significant species abundance results related to the gradient of increasingdominance and contiguity of the tree overstorey (ordination axis 3). For details, see Tables4, 6, and 6 and Figures 4, 5, 7, 8, and 10.

Summary of significant stand structure results related to the gradient of increasingdominance and contiguity of the tree overstorey (ordination axis 3). Lines indicate thedistribution of structural stages. For details, see Tables 4, 5, 6, and 7 and Figure 4.

Low High

Indicator species performance

Forest structural stage

Dominance and contiguity oftree overstorey

Ordination axis 3: dominance and contiguity of the tree overstorey

HighLow

Young forest – Mature forest: No species performed best

Old forest: Residualspecies performed best

Shrub/herb – Pole/sapling: Invasive speciesperformed best except one residual

Shrub/herb - Pole/sapling

Young forest - Mature forest

Old forest

(i.e., young forest and mature forest), corroborated by lower average speciesrichness (Alpha diversity) during these stages (Table 3).

Our results confirm the sparse information currently available on patternsof abundance of groups of forage plants important to black bears and grizzlybears. Researchers in British Columbia and elsewhere identified sites withhigh precipitation (Koehler and Pierce 2003), sites rich in soil nutrients, wet-lands and other sites high in soil moisture (Stamp 2003), and sites lacking adense tree cover (Zager 1980; Martin 1983; Bratkovich 1985; MacHutchon etal. 1993; Hammer 1996) as habitats with abundant bear forage plants. Our ordination results indicate that nutrient-rich and moist sites, and sites withopen canopies produce abundant understorey plants important to blackbears. We did not directly measure precipitation, but the high affinity ofmany forage plants for soil moisture (Tables 4, 6) support a positive abun-dance and performance response of these forage plants to an increase inavailable water.

Overstorey dominance as a surrogate for understorey conditions The pre-vailing disturbance regime substantially affects the structure of the treeoverstorey in coastal British Columbia (Lertzman et al. 1996, 2002; Spies1997; Gavin et al. 2003a). Its coastal temperate rainforests are noted for thenear-absence of recent fires and the dominance of late-seral forests in un-managed landscapes (Gavin et al. 2003b), where the primary disturbanceregime creates small gaps (Lertzman et al. 1996). The results of this studysupport other findings indicating changes in the structure of the overstoreyimpact the development of understorey plants (e.g., Klinka et al. 1985, 1996;Halpern 1989; Huffman et al. 1994; Hanley and Brady 1997; Franklin et al.2002). Although the gradient related to the tree overstorey was the weakest ofthe three main factors affecting community structure of forage plants in thestudy area, it was consistently negatively correlated with the total abundanceand richness of forage plants, particularly invasive forage plants (Figure 5).

The interception of light by the overstorey canopy has a significant impacton the development of certain understorey plants (Alaback 1982b; Haeussleret al. 1990; Klinka et al. 1996; Hanley and Brady 1997; McKenzie et al. 2000;Roburn 2003). Low levels of light restrict understorey growth, whereas plantgrowth, berry production, and sexual reproduction increase at higher lightlevels (Stewart 1988; Tappeiner and Alaback 1989; Bunnell 1990; Haeussler etal. 1990; Huffman et al. 1994; Klinka et al. 1996).

Studies of forest succession suggest that the degree of overstorey domi-nance is not just a surrogate for the level of understorey light. For example,in some mature forests in western Washington, various measures related tothe degree of overstorey dominance were surrogates of understorey dynam-ics directly contingent upon time (McKenzie et al. 2000). During early standdevelopment, the degree of overstorey dominance is a surrogate for the degrees of understorey light (Drever and Lertzman 2003) and ground distur-bance (Vanha-Majamaa and Jalonen 2001), which in turn impact understoreydevelopment (Halpern and McKenzie 2001; Roberts and Zhu 2002). Never-theless, it is not clear if overstorey-understorey relationships observedfollowing partial cutting apply to unmanaged forests.

Different responses of invasives and residuals to increasing dominance of the tree overstorey Few studies have explicitly examined the differentialresponses of invasives and residuals to increasing dominance of the tree

4.2 Relationshipbetween Community

Structure of ForagePlants and Dominanceof the Tree Overstorey

canopy. In our study, invasive forage plants responded positively to increas-ing canopy openness and decreasing dominance of the overstorey. These findings are consistent with research on understorey plant succession aftercomplete removal of the tree canopy in mature and old-growth westernhemlock and Douglas-fir forests in western Washington and Oregon (Dyr-ness 1973; Halpern 1988, 1989; Schoonmaker and McKee 1988; Halpern andSpies 1995). Invasives established and increased in abundance within severalgrowing seasons after clearcutting. Short-lived invasive herbs, then invasiveshrubs, and finally residual shrubs dominated soon after disturbance(Schoonmaker and McKee 1988). Invasive herbs and shrubs are more adapt-ed than residuals to early-seral conditions (Grime 2001), and will likelybenefit most from disturbances that open up the tree canopy if competitionfor understorey resources is reduced as a result (Noble and Slatyer 1980;Grime et al. 1988; Halpern 1989).

Our results on the relationship of invasives to the tree overstorey are comparable to findings from studies on succession following partial cutting(Halpern et al. 1999). In Sweden, clearcuts had a higher increase in invasivesthan did shelterwood treatments (reduction from 650 trees/ha to 200 and 140trees/ha) in mature spruce stands (Hannerz and Hanell 1993; see Section 3.5).However, it is not clear how directly findings from partial cutting studies canbe applied to unmanaged forests.

Except for Vaccinium species, residuals in our study did not respond neg-atively to increasing canopy openness or contiguity (Figures 5, 8), and wereless abundant where the tree overstorey was very dominant (Figures 8, 11).This indicates that many of the residuals in the study area fall in the middleof the continuum of life history strategies between early-seral invasives andlate-seral residuals. No residual forage plants observed in the Nimpkish Valley depend on forest interior and/or late-successional conditions. Auteco-logical descriptions show that many of these residuals benefit from increasedlight associated with decreased dominance of the tree overstorey and someare relatively resistant to various disturbances (Alaback 1982a; Minore 1984;Klinka et al. 1989; Coates et al. 1990; Pojar and MacKinnon 1994). Afterclearcutting in mature and old-growth forests in western Washington andOregon, the abundance of many residuals recovered to pre-disturbance lev-els prior to canopy closure, reflecting their relatively high resilience, althoughintensive disturbance delayed the recovery of some species (Schoonmakerand McKee 1988; Halpern and Spies 1995).

Vaccinium forage species in the Nimpkish Valley seem relatively sensitiveto disturbance. Their abundance was depressed at very low levels of canopydominance and contiguity compared to all residual forage species, presum-ably where ground disturbance is high (Figure 8). Although most residualsexperience a transient decline immediately after disturbance, we were notable to detect this effect for most residuals because our data did not includeplots in the two earliest (sparse/bryoid and herb) structural stages sinceshrubs were never completely removed after clearcutting.

We expected the abundance of most residual forage plants (as a propor-tion of the total abundance of forage plants) to increase with increasing treeoverstorey dominance and contiguity because invasive plants do not grow as well as residuals in undisturbed forests (Dyrness 1973; Halpern 1989). Increasing tree overstorey dominance and contiguity should present a steep-er competitive exclusion gradient for invasive rather than residual forageplants. Ordination results and indicator species analysis confirmed this (Figure 5, Table 6).

Relationship between community structure of forage plants and foreststructural stages Patterns of change in forest structure can be associatedwith recognizable stages of forest development (e.g., Franklin 1982; Spies andFranklin 1988; Oliver and Larson 1990; Wells 1996; Franklin et al. 2002). Wefound some differences in the community structure of forage plants acrossstructural stages and communities. The shrub/herb stage had higher speciesrichness than most other structural stages (Table 3) and differed distinctlyfrom most other structural stages, particularly from the intermediate-agedstages in the CWHxm2 and CWHvm1 variants (Table 7, Figure 4c). No for-age plants had an affinity for intermediate structural stages (i.e., young andmature forest).

A gradient analysis in the CWH zone in the Fraser Valley and on southernVancouver Island also reported significant differences in the species compo-sition of understorey plants between young structural stages without domi-nant tree cover and all other structural stages (Klinka et al. 1985). Invasivessuch as Epilobium angustifolium had a higher affinity, and consequently ahigher diagnostic value, for the two youngest structural stages. None of theresidual vascular plants had a higher affinity for the two youngest structuralstages when compared to all other structural stages according to classificationsystem proposed by Mueller-Dombois and Ellenberg (1974). Our results alsoshow that, except for one residual species, invasive forage plants performedbest in the structural stages with the least dominant tree overstorey (i.e.,shrub/herb and pole/sapling) (Table 6). Indicator species analysis also sup-ported the suggestion of Klinka et al. (1985) that understorey species abun-dance was generally suppressed during intermediate stages of forestsuccession.

There are several explanations for why plots in different structural stagesoften had overlapping forage plant community structure. First, life historytheory proposes multiple successional pathways for sites at the same struc-tural stage (Noble and Slatyer 1980; Klinka et al. 1985; Hamilton 1988).Deterministic and stochastic factors related to timing and intensity ofground disturbance, availability of propagules, and other biotic and environ-mental factors affect the successional pathway at a given site (Noble andSlatyer 1980; Pickett et al. 1987; Schoonmaker and McKee 1988). Althoughplots in our study often originated from different disturbances, we were not able to account for differences related to disturbance history that likelycaused different successional trajectories in communities of forage plants atthe same structural stage.

Second, the model of structural stages used in our analysis does not reflectcertain distinct stages of forest development. For our analysis we adopted amodel of structural stages developed by the B.C. Ministry of Environment,Lands and Parks and the B.C. Ministry of Forests (1998) adapted from Oliverand Larson (1990) (Table 1). Neither model explicitly accounts for canopyclosure, which, except for the stand-initiating disturbance, is the most dramatic event affecting the rate and degree of change in stand condition,causing significant shifts in environmental factors, composition, and functionof forest ecosystems (Franklin et al. 2002). These significant shifts cause majorenvironmental changes in the understorey (Franklin et al. 2002). Severalother distinct developmental stages are also not accounted for in the BritishColumbia model (Table 1). Therefore, it is not surprising that structuralstages did often not clearly separate along the axis of increasing dominanceof the tree overstorey, reflected by the overlapping arrows in Figure 4c.

Finally, the legacy effect (Franklin et al. 1997, 2002) likely blurs differencesin the community structure of forage plants expected along the continuumrelated to the tree overstorey. The total abundance of forage plants is inverse-ly related to increasing dominance and contiguity of the tree overstorey(Figure 5). Trends found along individual axes in community structure offorage plants support fundamental differences at the extremes of the contin-uum realted to the tree overstorey. Young and intermediate-aged structuralstages represent extremes in dominance of the tree overstorey. From an un-derstorey plant’s perspective, young structural stages have the least, andintermediate-aged structural stages the most, dominant canopy. Simply con-sidering the continuum of overstorey structure, we would thus expect thecommunity structure of forage plants in intermediate-aged structural stagesto be the most differentiated from the youngest structural stages. However,arranged along a temporal axis, young and intermediate-aged structuralstages are overlapping and adjacent (Table 1). We can thus expect a legacy ef-fect related to understorey development during the young structural stage topersist in the intermediate structural stage (Franklin et al. 2002). Althoughmost invasives do not survive canopy closure, some may endure on disturbedmicrosites (Dyrness 1973). The intermediate position of some residual forageplants along the continuum of life history strategies allows them to establishbefore canopy closure, surviving this phase of forest development, even ifsuppressed (e.g., Bunnell 1990; Halpern et al. 1999). Researchers in Douglas-fir–dominated forests in Oregon and Washington also observed that,although understorey species abundance changed significantly across struc-tural stages, many occurred in all stages (Halpern and Spies 1995), likelybecause the legacy effect blurs differences that would otherwise be apparent.

In the study area, soil nutrient and moisture content are the most importantdeterminants of forage plants community structure. This gradient accountedfor the most variation. More forage species, especially invasive species, oc-curred on richer and moister sites, and most species performed best (i.e., hada significant and maximum indicator value) on these sites (Figure 5, Table 6),supporting the literature.

Although our results do not clearly reveal whether the forage plants in the study area responded more to changes in soil nutrients or soil moisture,most forage plants respond positively to both (Taylor and MacBryde 1977;Haeussler and Coates 1986; Klinka et al. 1989; Haeussler et al. 1990; Mei-dinger et al. 2002). Of the 20 plants indicative of fresh to wet soils and ofnutrient-medium or -rich soils, only five (primarily Vaccinium spp.) are notknown indicators of both increased soil nutrients and moisture (Table 4).Most forage plants that do not have a documented soil nutrient or moistureindicator value often perform better on richer and moister sites (Taylor andMacBryde 1977; Coates et al. 1990; Pojar and MacKinnon 1994).

Compared to the group of invasive forage plants, the residual forageplants did not respond as homogeneously to any of the main environmentalgradients (Figure 5), supporting previous studies. Accordingly, residual for-age plants include species with affinities for nutrient-rich and moist sites aswell as nutrient-poor and/or drier sites (Table 4).

Differences in life history strategy between invasives and residuals also explain why invasives responded more consistently to the gradient of in-creasing soil nutrient and moisture content than did residuals. Unlikeinvasives, residual forage plants are not restricted to forests with an open

4.3 Relationshipbetween Community

Structure of ForagePlants and Soil

Nutrient and MoistureStatus

canopy (Dyrness et al. 1974; Halpern 1989). For residuals on sites with a dom-inant tree overstorey, light is probably more limiting than soil nutrientsand/or moisture, limiting their response to soil resources.

Elevation accounted for the second highest proportion of variation in com-munity structure of forage plants. Elevation causes complex environmentalgradients that co-vary with factors such as increasing precipitation and de-creasing temperature with increasing elevation. Environmental factors suchas soil chemistry also change with elevation (Ohmann and Spies 1998).

Forage plants differed in their responses to the altitudinal gradient. Dry-opteris expansa was less abundant at lower elevations (Figure 10). Vacciniumspecies responded strongly to elevation, and performed best in the higher- elevation CWHvm2 variant, except for V. parvifolium (Table 6, Figure 5). Incontrast, all other significant indicator forage species performed best in theCWHxm2 or CWHvm1 variants (Table 6). Many invasives were less abun-dant at higher elevations (Figures 6, 9), substantiating autecological studies(Klinka et al. 1989; Pojar and MacKinnon 1994; Table 1).

Researchers and managers often divide the landscape into smaller environ-mental units, such as site series and structural stages, to distinguish sites byabundance and species composition. It is important to know which criteriaare best suited to partition distinct environmental units. Utilizing the threeenvironmental gradients observed in this study should provide environmen-tal units with more distinct plant communities than using other criteria.Although the sampling intensity in our study was too low to partition thelandscape across these three environmental gradients, our results showedthat site series reflected the variance associated with the two strongest ordi-nation gradients to produce the most distinct groups of forage plants (Table5), integrating the effects of changes in soil nutrient and moisture contentand elevation.

Davis (1996) combined site series into black bear habitat types (s) to reduce the number of categories. These s were site series combinedacross and within variants based on similarities in moisture and nutrientregimes (see Davis 1996). Our analysis revealed that s produced less dis-tinct groups of forage plants than did site series, probably in part because notall s distinguish between variants, particularly between differences inelevation, the second strongest ordination gradient (Table 5).

Both site series and various structural stages in the CWHxm2 andCWHvm1 variants contained distinct units of forage plants (Table 5), and together reflect each ordination gradient. At least in lower-elevation variants,further differentiating site series into structural stages would capture evenmore distinct bear forage plant community units, but there was insufficientstatistical power to test this hypothesis.

Our results indicate that active management and/or protection of suitableareas could maintain or increase the abundance of forage plants in land-scapes used by black bears in the CWH zone. The choice of tree harvestingregime can increase the abundance of forage plants at the stand level. If soildisturbance is minimized and there is sufficient adjacent forest containingbear forage plants, harvesting regimes of moderate to high intensity up toclearcutting would likely increase the short-term abundance of invasive for-age plants up until crown closure (Arnott and Beese 1997; Beese and Bryant

4.6 ManagementImplications

4.5 Relationshipbetween Community

Structure of ForagePlants and

Environmental Units

4.4 Relationshipbetween Community

Structure of ForagePlants and Elevation

1999; Mitchell and Beese 2002). There would be an initial decline immediate-ly after harvest, but then many residual forage plants would increase to orsurpass their pre-disturbance levels of abundance (consult Section 1.2 for adefinition of residual and invasive forage plants). Since none of the residualforage plants is a forest-interior or late-successional species, benefits fromtimber harvesting will be lost with canopy closure. Thus, to the extent thatblack bear habitat is a management priority, tree canopy closure should bedelayed as long as possible, and vegetation management treatments facilitat-ing establishment and management of the next tree harvest should beavoided.

After harvesting, residuals will likely recover faster in larger openings andretention systems if forest patches are retained to provide a source of propag-ules for residual recolonization, creating habitat islands (Franklin et al. 1997).Retained forest patches would serve as “stepping-stones” for residual forageplants recolonizing the harvested area.

To improve foraging habitat quality at the landscape level over the longterm, forest managers should avoid harvesting regimes that create large areaswith a dense, structurally homogeneous tree cover that typically develops 15–45 years after clearcutting. In the absence of other management actions facilitating the next harvest rotation, in the short term clearcutting or com-parable harvesting systems with short rotations would convert forest standsto a highly productive bear forage early structural stage. However, over themedium term, this increases the proportion of intermediate-aged foreststands with a homogeneous, contiguous overstorey (Deal 2001; Franklin etal. 2002) with depauperate understoreys (Alaback 1982b; Wells 1996). Undernormal industrial forest management practice, these stands would dominatethe landscape (Lertzman et al. 1997). Alternative harvesting regimes likelyimprove the long-term quality of foraging habitat by promoting a more het-erogeneous overstorey with more variable understorey light levels (Franklinet al. 1997; Drever and Lertzman 2003). Alternative harvesting regimes mayfeature longer rotations with variable retention of various stand componentsas biological legacies.

Most importantly, to improve the quality of spring and fall foliage forag-ing habitat in the long term, forest managers should avoid or minimizeharvesting in nutrient-rich and moist areas (Table 8). Where this is unavoid-able, managers should avoid harvesting regimes that produce large areas witha dense, structurally homogeneous tree cover, which can be achieved by re-taining some areas in unharvested patches. When it is desirable to maintainthe abundance of certain residual forage plants such as Vaccinium speciesand Gaultheria shallon, managers should also set aside and actively manageareas with lower soil nutrient and moisture status (Table 4, Figure 5).

To the extent that maintaining black bear populations is a managementpriority, forest managers should reserve and actively manage foraging habitatthroughout the range of elevations and variants in the CWH zone. Althoughlower-elevation sites are important for most forage plants (Tables 3, 6), high-er-elevation areas, particularly in the CWHvm2 variant, provide importanthabitat for many fall foliage bear forage plants (Figure 12). Retaining foraginghabitat throughout the landscape is particularly important for female blackbears with cubs because these bears are vulnerable to predation while travel-ling to foraging sites.

To maintain sustainable levels of habitat abundance and quality across thelandscape, forest managers should vary the intensity of protection and man-

agement activities over time and space. For example, where bear forage is notlimiting, strategies to increase its abundance can be relaxed. Where bear for-age is limiting, management prescriptions aimed at preserving existingfeeding sites and increasing forage abundance should be emphasized.

For areas with similar ecology and climate as the study area, we estab-lished a set of guidelines aimed to maintain or increase the short-term (priorto crown closure) and long-term (after crown closure) seasonal availabilityof forage plants (Table 8). In general, forest managers can best maintain andincrease the seasonal availability of forage by focusing management and pro-tection efforts based on three criteria: soil nutrient and moisture regimes,elevation, and tree canopy cover, for which harvesting regime combined withsuccession could be a surrogate.

Because the management guidelines described in Table 8 draw from find-ings that apply to a specific set of conditions, these prescriptions apply only

Management guidelines for choosing forested sites for protection and active management of foraging habitat inthe CWH to meet objectives of maintaining the short- and long-term abundance of forage plants, grouped byseason of consumption

Management objectives (choose one or more)

Increase short- and

Protection of long-term abundance

Forage groupa foraging habitat of forageb Timber harvesting

Foliage spring • Prioritize protecting • Remove overstorey using • Remove overstorey using Foliage fall sites with higher low to medium intensities low to medium levels

soil nutrient and of variable retention. of variable retention.moisture content. • Delay canopy closure as • Delay canopy closure as

• Prioritize protecting long as possible and avoid long as possible and avoidhigher-elevation sites. vegetation management vegetation management

treatments. treatments.• Choose sites with higher

soil nutrient and moisture content for harvesting.

• Choose sites at higher elevation for harvesting.

Foliage summer • Protect sites within all • Remove overstorey using • Remove overstorey usingFruit summer nutrient and moisture low to medium intensities low to medium levels of

regimes. of variable retention. variable retention.• Protect sites at lower • Delay canopy closure as • Delay canopy closure as

elevations (xm2, vm1 long as possible and avoid long as possible and avoidvariants). vegetation management vegetation management

treatments. treatments.• Sites with all soil nutrient

and moisture regimes at all elevations are suitable for harvesting.

Fruit spring • Heterogeneous species ecologies or low abundance of forage plants in study area, manage for

Fruit fall individual species following Table 2.

a Foliage forage groups include graminoids and Equisetum species. Lysichiton americanum is the only member of the forage group with roots or corms important and should thus be managed individually.

b Only applicable when forage abundance is very low due to a dense tree overstorey.

to areas similar to the study area. Similar approaches could be devised forother regions if local knowledge of the ecosystem is used to assess sites basedon community structure of forage plants.

The study area boundaries encompassed the home ranges of black bears inthe Nimpkish Valley (Davis 1996). Thus, inferences beyond the home rangesof these bears must be made cautiously. To extend these conclusions to thesethree variants throughout the CWH zone requires more geographicallyextensive data collection.

We were unable to include plots in the sparse/bryoid and herb structuralstages and within other non-forested areas in our analysis, because theseplots did not occur in the study area when the field work was conducted.Thus, the patterns observed in this study do not reflect the variation repre-sented in such potentially very important foraging areas.

Our study did not account for the productivity of the forage plants, whichdepends on light levels. Shrubs such as Gaultheria shallon may be abundantunder moderately closed-canopy forests but only produce negligibleamounts of forage for black bears. Inter-seasonal nutritional differences between forage plants were also not studied, which would likely show sub-stantial variation within the study area.

When managing the availability of bear forage plants, managers have littleguidance for choosing an adequate spatial scale. For example, it is often notclear if it is more appropriate to manage for a forage plant group at very fine(site series modifier) or much coarser () resolution.

Indicator species analysis can be used to find the environmental unit mostcongruent with the ecological amplitude of a group of species. The most ap-propriate resolution for a given forage group can be determined by plottingthe sum of indicator values of a given group of forage plants for successivelevels of hierarchical management units (e.g., variant, , site series, siteseries modifier). The best resolution would occur where the summed indica-tor values for the forage group are maximized (Dufrene and Legendre 1997;McCune and Grace 2002). Of course, ecological and management considera-tions other than the ecological amplitude of a species may entail a differentscale.

During adaptive management, researchers and managers should establishlong term monitoring plots to replicate and complement these findings. Theforest development trajectory after disturbance has long-term effects on un-derstorey plant communities. Long-term monitoring is needed to clarify therelationship between forest seral dynamics and community structure of for-age plants. In particular, the long-term effects of conventional and alternativeharvesting regimes on forest seral dynamics need further examination.

The full range of forage plant communities within the CWH zone in theNimpkish Valley and elsewhere, including wetlands, non-forested areas, andvery young structural stages, should be sampled to determine the applicabili-ty of our findings beyond the study area and across the CWH zone. Samplingshould follow a stratified and random design for optimal analysis.

Researchers and managers working in the CWH zone should integrate in-formation on differences in the forage value of forage plants. Once a rankingfor forage plant value is established, it can be easily integrated into futureanalyses and management recommendations by weighting abundance values.

4.8 Future Work

4.7 Study Limitations

5 LITERATURE CITED

Alaback. 1982a. Forest community structural change during secondary succession in southeast Alaska. In Proc. symp. On forest succession and stand development research in the Northwest. J.E. Means (editor).For. Res. Lab., Oregon State Univ., Corvallis, Oreg., pp. 70–79.

———. 1982b. Dynamics of understorey biomass in Sitka spruce-western hemlock forests of southeast Alaska. Ecology 63:1932–1948.

Alaback, P.B. and F.R. Herman. 1988. Long-term response of understorey vegetation to stand density in Picea–Tsuga forests. Can. J. For. Res.18:1522–1530.

Anderson, A.J.B. 1971. Ordination methods in ecology. J. Ecol. 59:713–726.

Arnott, J.T. and W.J. Beese. 1997. Alternatives to clearcutting in BC coastalmontane forests. For. Chron. 73:670–678.

Arsenault, A. and G.E. Bradfield. 1995. Structural-compositional variation inthree age-classes of temperate rainforests in southern coastal BritishColumbia. Can. J. Bot. 73:54–64.

Bazzaz, F.A. 1996. Plants in changing environments: linking physiological,population, and community ecology. Cambridge Univ. Press, Cambridge, U.K.

B.C. Ministry of Environment, Lands and Parks and B.C. Ministry of Forests.1998. Field manual for describing terrestrial ecosystems. Resources Inventory Branch, B.C. Min. Environ., Lands and Parks, and B.C. Min.For., Res. Br., Victoria, B.C. Land Manage. Handb. 25.

Beals, E. 1984. Bray-Curtis ordination: an effective strategy for analysis ofmultivariate ecological data. Adv. Ecol. Res. 14:1–55.

Beese, W.J. and A.A. Bryant. 1999. Effect of alternative silvicultural systemson vegetation and bird communities in coastal montane forests ofBritish Columbia, Canada. For. Ecol Manage. 115:231–242.

Bell, F.W. 1991. Critical silvics of conifer crop species and selected competitive vegetation in northwestern Ontario. Ont. Min. Nat. Resour., Northwestern Ont. For. Technol. Dev. Unit, Tech. Rep. 19, Thunder Bay, Ont., Rep. 3310.

Bierzychudek, P. 1982. Life histories and demography of shade-tolerant temperate forest herbs: a review. New Phytol. 90:757–776.

Bratkovich, A.A. 1985. Grizzly bear habitat components associated with pastlogging practices on the Libby ranger district, Kootenai National forest.In Grizzly bear habitat symp. G.P. Contreras and K.E. Evans (editors).U.S. Dep. Agric. For. Serv., Ogden, Utah. Gen. Tech. Rep. -207, pp.180–184.

Brockway, D.G. 1998. Forest plant diversity at local and landscape scales inthe Cascade Mountains of southwestern Washington. For. Ecol. Manage.109:323–341.

Bull, E.L., R.S. Holthausen, and D.B. Marx. 1990. How to determine snagdensity. Western J. Appl. For. 5:56–58.

Bunnell, F.L. 1990. Reproduction of salal (Gaultheria shallon) under the forest canopy. Can. J. For. Res. 20:91–100.

Bunnell, F.L., L.L. Kremsater, and E. Wind. 1999. Managing to sustain vertebrate richness in forests of the Pacific Northwest: relationshipswithin stands. Environ. Rev. 7:97–146.

Campbell, B.D. and J.P. Grime. 1992. An experimental test of plant strategytheory. Ecology 73:15–29.

Chan, S.S., S.R. Radosevich, and A.T. Grotta. 2003. Effects of contrastinglight and soil moisture availability on the growth and biomass allocation of Douglas-fir and red alder. Can. J. For. Res. 33:106–117.

Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Austral. J. Ecol. 18:117–143.

Coates, D., S. Haeussler, and J. Mather. 1990. A guide to the response ofcommon plants in British Columbia to management treatments. For. Can. and B.C. Min. For., Victoria, B.C., Handb. 008.

Connell, J.H. and R.O. Slatyer. 1977. Mechanisms of succession in naturalcommunities and their role in community stability and organization.Amer. Nat. 1 1 1:1119–1144.

Daubenmire, R.F. 1959. A canopy coverage method of vegetation science.Northwest Sci. 33:43–64.

Davis, H. 1996. Characteristics and selection of winter dens by black bears incoastal British Columbia. MSc thesis. Simon Fraser Univ., Burnaby,B.C.

Deal, R.L. 2001. The effects of partial cutting on forest plant communities ofwestern hemlock-Sitka spruce stands in southeast Alaska. Can. J. For.Res. 31:2067–2079.

Demarchi, D.A. 1995. Ecoregions of British Columbia, 4th ed. 1:2 000 000map. B.C. Min. Environ., Lands and Parks, Victoria, B.C.

Drever, C.R. and K.P. Lertzman. 2003. Effects of a wide gradient of retainedtree structure on understorey light in coastal Douglas-fir forests. Can. J.For. Res. 33:137–146.

Dufrene, M. and P. Legendre. 1997. Species assemblages and indicatorspecies: the need for a flexible asymmetrical approach. Ecol. Monogr.67:345–366.

Dyrness, C.T. 1973. Early stages of plant succession following logging andslash burning in the western Cascades of Oregon. Ecology 54:57–58.

Dyrness, C.T., J.F. Franklin, and W.H. Moir. 1974. A preliminaryclassification of forest communities in the central portion of the western Cascades of Oregon. U.S. Int. Biol. Prog., Univ. Wash., Seattle, Wash., Conif. For. Biometr. Bull. 4.

Eis, S. 1981. Effect of vegetative competition on regeneration of white spruce(Picea glauca). Can. J. For. Res. 11:1–8.

Fasham, M.J.R. 1977. A comparison of nonmetric multidimensional scaling,principal components and reciprocal averaging for the ordination ofsimulated coenoclines and coenoplanes. Ecology 58:551–561.

Franklin, J.F. 1982. Old-growth forests in the Pacific Northwest: an ecologicalview. In Old-growth forests: a balanced perspective. Univ. Oreg., Bur.Gov. Res. Serv., Eugene, Oreg., pp. 5–27.

Franklin, J.F., D.R. Berg, D.A. Thornburgh, and J.C. Tappeiner. 1997. Alternative silvicultural approaches to timber harvesting: variable retention systems. In Creating a forestry for the 21st century. K.A.Kohm and J.F. Franklin (editors). Island Press, Washington, D.C., pp. 11–30.

Franklin, J.F., T.A. Spies, R. Van Pelt, A.B. Carey, D.A. Thornburgh, D.R.Berg, D.B. Lindenmayer, M.E. Harmon, W.S. Keeton, D.C. Shaw, K.Bible, and J.Q. Chen. 2002. Disturbances and structural developmentof natural forest ecosystems with silvicultural implications, using Douglas-fir forests as an example. For. Ecol. Manage. 155:399–423.

Gagnon, D. and G.E. Bradfield. 1987. Gradient analysis of west central Vancouver Island forests. Can. J. Bot. 65:822–833.

Garcia, E.R. 1985. Grizzly bear direct habitat improvement on the KootenaiNational Forest. In Grizzly bear habitat symp. G.P. Contreras and K.E.Evans (editors). U.S. Dep. Agric. For. Serv., Ogden, Utah. Gen. Tech.Rep. -207, pp. 185–189.

Gavin, D.G., L.B. Brubaker, and K.P. Lertzman. 2003a. An 1800-year recordof the spatial and temporal distribution of fire from the west coast ofVancouver Island, Canada. Can. J. For. Res. 33:573–586.

———. 2003b. Holocene fire history of a coastal temperate rain forest basedon soil charcoal radiocarbon dates. Ecology 84:186–201.

Gitay, H. and I.R. Noble. 1997. What are functional types and how should we seek them. In Plant functional types their relevance to ecosystemproperties and global change. T.M. Smith, H.H. Shugart, and F.I.Woodward (editors). Cambridge Univ. Press, Cambridge, U.K., pp.3–19.

Grace, J.B. 1990. On the relationships between plant trades and competitiveability. In Perspectives on plant competition. D. Tilman (editor). Academic Press, San Diego, Calif., pp. 51–65.

Green, R.N. and K. Klinka. 1994. A field guide to site identification and interpretation for the Vancouver Forest Region. B.C. Min. For., Victoria, B.C. Land Manage. Handb. 28.

Grime, J.P. 1977. Evidence for the existence of three primary strategies inplants and its relevance to ecological and evolutionary theory. Amer.Nat. 111:1169–1194.

———. 2001. Plant strategies, vegetation processes, and ecosystem properties. 2nd ed. Wiley, New York, N.Y.

Grime, J.P., J.G. Hodgson, and R. Hunt. 1988. Comparative plant ecology.Unwin, London, U.K.

Haeussler, S., L. Bedford, J.O. Boateng, and A. MacKinnon. 1999. Plant community responses to mechanical site preparation in northern interior British Columbia. Can. J. For. Res. 29:1084–1100.

Haeussler, S. and D. Coates. 1986. Autecological characteristics of selectedspecies that compete with conifers in British Columbia: a literature review. B.C. Min. For., Victoria, B.C. Land Manage. Rep. 33.

Haeussler, S., D. Coates, and J. Mather. 1990. Autecology of common plantsin British Columbia: a literature review. For. Can. and B.C. Min. For.,Victoria, B.C. Rep. 158.

Halpern, C.B. 1988. Early successional pathways and the resistance and resilience of forest communities. Ecology 69:1703–1715.

———. 1989. Early successional patterns of forest species-interactions of life-history traits and disturbance. Ecology 70:704–720.

Halpern, C.B., J.A. Antos, K. Cromack, and A.M. Olson. 1992. Species interactions and plant diversity during secondary succession. Northwest Environ. J. 8:203–205.

Halpern, C.B., S.A. Evans, C.R. Nelson, D. McKenzie, D.A. Liguori, D.E.Hibbs, and M.G. Halaj. 1999. Response of forest vegetation to varyinglevels and patterns of green-tree retention: an overview of a long-termexperiment. Northwest Sci. 73:27–44.

Halpern, C.B. and D. McKenzie. 2001. Disturbance and post-harvest groundconditions in a structural retention experiment. For. Ecol. Manage.154:215–225.

Halpern, C.B. and T.A. Spies. 1995. Plant species diversity in natural andmanaged forests of the Pacific Northwest. Ecol. Appl. 5:913–934.

Hamann, A. and T. Wang, 2005. Models of climate normals for genecologyand climate change studies in BC. Agricultural and Forest Meteorology128: 211–221

Hamilton, A.N. 1987. Classification of coastal grizzly bear habitat for forestryinterpretations and the role of food in habitat use by coastal grizzlybears. MSc thesis. Univ. British Columbia, Vancouver, B.C.

Hamilton, A.N., C.A. Bryden, and J.C. Clement. 1991. Impacts of glyphosateapplication on grizzly bear forage production in the Coastal WesternHemlock zone. Wildl. Br., B.C. Min. Environ., Victoria, B.C. Rep. 165

Hamilton, E. 1988. A system for the classification of seral ecosystems withinthe biogeoclimatic ecosystem classification system: first approximation.Res. Br., B.C. Min. For. and Lands, Victoria, B.C., Lands Res. Rep.87004–.

Hammer, D. 1996. Buffaloberry (Shepherdia canadensis (L.) Nutt.) fruit production in fire successional bear feeding sites. J. Range Manage.49:520–529.

Hanley, T.A. and W.W. Brady. 1997. Understorey species composition andproduction on old-growth western hemlock-Sitka spruce forests ofsoutheast Alaska. Can. J. Bot. 75:574–580.

Hannerz, M. and B. Hanell. 1993. Changes in the vascular plant vegetationafter different cutting regimes on a productive peatland site in centralSweden. Scand. J. For. Res. 8:193–203.

Huffman, D.W., J.C. Tappeiner, and J.C. Zasada. 1994. Regeneration of salal(Gaultheria shallon) in the central coast range forests of Oregon. Can. J. Bot. 72:39–51.

Hutchinson, T.F., R.E.J. Boerner, L.R. Iverson, S. Sutherland, and E.K.Sutherland. 1999. Landscape patterns of understorey composition andrichness across a moisture and nitrogen mineralization gradient inOhio () Quercus forests. Plant Ecol. 144:177–189.

Jules, E.S. 1997. Consequences of forest fragmentation for the understoreyplant, Trillium ovatum (Liliaceae). In Conservation and management of native plants and fungi. M.V. Wilson (editor). Native Plant Soc.Oreg., Corvallis, Oreg., pp. 201–206.

———. 1999. Mechanisms of reduced trillium recruitment along edges ofold-growth forest fragments. Cons. Biol. 13:784–793.

Karakatsoulis, J. and J.P. Kimmins. 1993. Vegetation development and Douglas-fir growth following clearcut harvesting in the CWHdm subzone: 10-year results from the Blaney chronosequence. B.C. Min.For., Victoria, B.C. Land Manage. Rep. 81.

Kenkel, N.C. and L. Orloci. 1986. Applying metric and nonmetric multi-dimensional scaling to ecological studies: some new results. Ecology67:919–928.

Klinka, K., H.Y.H. Chen, Q.L. Wang, and L. de Montigny. 1996. Forestcanopies and their influence on understorey vegetation in early-seralstands on west Vancouver Island. Northwest Sci. 70:193–200.

Klinka, K., V.J. Krajina, A.Ceska, and A.M. Scagel. 1989. Indicator plants ofcoastal British Columbia. Univ. British Columbia Press, Vancouver,B.C.

Klinka, K., A.M. Scagel, and P.J. Courtin. 1985. Vegetation relationshipsamong some seral ecosystems in southwestern British Columbia. Can.J. For. Res. 15:561–569.

Koehler, G.M. and D.J. Pierce. 2003. Black bear home-range sizes in Wash-ington: climatic, vegetative, and social influences. J. Mammal. 84:81–91.

Kozlowski, T.T. 1972. Physiology of water stress. In Wildland shrubs—their biology and utilization: an international symposium. J.R. Goodin (editor). U.S. Dep. Agric. Ogden, Utah. Gen. Tech. Rep., -1 pp. 229–244.

Krajina, V.J. 1965. Biogeoclimatic zones and classification of British Columbia. Ecology of Western North America 1:1–17.

———. 1969. Ecology of forest trees in British Columbia. Ecology of Western North America 2:1–146.

Kramer, P.J. and T.T. Kozlowski. 1979. Physiology of woody plants. Academic Press, New York, N.Y.

Kruskal, J.B. 1964. Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika 29:1–27.

Kruskal, J.B. and M. Wish. 1978. Multidimensional scaling. Sage Publica-tions, Beverly Hills, Calif.

Lavender, D.P., C.M. Parish, C.M. Johnson, G. Montgomery, A. Vyse, R.A. Willis, and D. Winston (editors). 1990. Regenerating British Columbia’s forests. Univ. British Columbia Press, Vancouver, B.C.

Lertzman, K., D. Gavin, D. Hallett, L. Brubaker, D. Lepofsky, and R. Mathewes. 2002. Long-term fire regime estimated from soil charcoal in coastal temperate rainforests. Cons. Ecol. 6:5.

Lertzman, K., T. Spies, and F. Swanson. 1997. From ecosystem dynamics toecosystem management. In The rainforests of home: profile of a NorthAmerican bioregion. E. Wolf (editor). Island Press, Washington, D.C.,pp. 361–382.

Lertzman, K.P., G.D. Sutherland, A. Inselberg, and S.C. Saunders. 1996.Canopy gaps and the landscape mosaic in a coastal temperate rain forest. Ecology 77:1254–1274.

Lezberg, A.L., J.A. Antos, and C.B. Halpern. 1999. Belowground traits ofherbaceous species in young coniferous forests of the Olympic Peninsula, Washington. Can. J. Bot. 77:936–943.

Lieffers, V.J. and K.J. Stadt. 1994. Growth of understorey Picea glauca, Calamagrostis canadensis, and Epilobium angustifolium in relation to overstorey light transmission. Can. J. For. Res. 24:1193–1198.

Limpert, E., W.A. Stahel, and M. Abbt. 2001. Log-normal distributions across the sciences: keys and clues. BioScience 51:341–352.

Luttmerding, H.A., D.A. Demarchi, D.V. Lea, D.V. Meidinger, and T. Vold.1990. Describing ecosystems in the field. 2nd ed. B.C. Min. Environ.,Victoria, B.C. Manual 11.

MacHutchon, G., S. Himmer, and C. Bryden. 1993. Khutzeymateen Valleygrizzly bear study: final report. B.C. Min. Environ., Lands and Parks,Wildl. Br., Victoria, B.C. Habitat Res. Rep. 31.

MacKenzie, W. and R. Klassen. 1999. VPro97 Version 2 for Access 8. B.C.Ministry of Forests, Victoria, B.C.

MacKinnon, A. 1992. Use of the biogeoclimatic ecosystem classification system in British Columbia. For. Chron. 68:100–120.

———. 2003. West coast, temperate, old-growth forests. For. Chron.79:475–784.

Marcum, C.L. and D.O. Loftsgaarden. 1980. A nonmapping technique forstudying habitat preferences. J. Wildl. Manage. 44:963–968.

Martin, P. 1983. Factors influencing globe huckleberry fruit production inNorthwestern Montana. Intl. Conf. Bear Res. Manage. 5:159–165.

Mather, P.M. 1976. Computational methods of multivariate analysis in physical geography. John Wiley and Sons, London, U.K.

McCook, L.J. 1994. Understanding ecological community succession. Causalmodels and theories, a review. Vegetatio 110:115–147.

McCune, B. and J.B. Grace. 2002. Analysis of ecological communities. MjMSoftware Design, Gleneden Beach, Oreg.

McCune, B., J. Hutchinson, and S. Berryman. 2002. Concentration of rareepiphytic lichens along large streams in a mountainous watershed inOregon, . The Bryologist 105:439–450.

McCune, B. and M.J. Mefford. 1999. -. Multivariate analysis of ecological data, version 4. MjM Software Design, Gleneden Beach,Oreg.

McCune, B., R. Rosentreter, J.M. Ponzetti, and D.C. Shaw. 2000. Epiphytehabitats in an old conifer forest in western Washington, . The Bryologist 103:417–427.

McKenzie, D., C. B. Halpern, and C. R. Nelson. 2000. Overstorey influenceson herb and shrub communities in mature forests of western Washington, . Can. J. For. Res. 30:1655–1666.

Meidinger, D., T. Lee, G.W. Douglas, G. Britton, W. MacKenzie, and H.Qian. 2002. British Columbia plant species codes and selected attributes. Version 4 database. B.C. Min. For., Res. Br., Victoria, B.C.

Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. B.C. Min.For., Res. Br., Victoria, B.C., Special Rep. Ser. 6.

Meier, A.J., S.P. Bratton, and D.C. Duffy. 1995. Possible ecological mechanisms for loss of vernal-herb diversity in logged eastern deciduous forests. Ecol. Appl. 5:935–946.

Mielke, P.W. Jr. 1984. Meteorological applications of permutation techniquesbased on distance functions. In Handbook of statistics, Volume 4. P.R.Krishnaiah and P.K. Sen (editors). Elsevier Science Publ., The Hague,Netherlands, pp. 813–830.

Minore, D. 1984. Vaccinium membranaceum berry production seven yearsafter treatment to reduce overstorey tree canopies. Northwest Sci.58:208–212.

Mitchell, S.J. and W.J. Beese. 2002. The retention system: reconciling variableretention with the principles of silvicultural systems. For. Chron.78:397–403.

Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology. Wiley, New York, N.Y.

Nagy, J.A. and A.G. MacHutchon. 1991. Annual progress report—year 1(1989/90), annual working plan—year 2 (1990/91) Khutzeymateen Val-ley grizzly bear study. B.C. Min. Environ., Victoria, B.C. Wildl. Work.Rep. -48.

Noble, I.R. and G. Habiba. 1996. A functional classification for predicting the dynamics of landscapes. J. Veg. Sci. 7:329–336.

Noble, I.R. and R.O. Slatyer. 1980. The use of vital attributes to predict successional changes in plant communities subject to recurrent disturbances. Vegetatio 43:5–21.

Noyce, K.V. and P.L. Coy. 1990. Abundance and productivity of bear foodspecies in different forest types of northcentral Minnesota. Intl. Conf.Bear Res. Manage. 8:169–181.

Oakley, B.B. and J.F. Franklin. 1998. Bitter cherry (Prunus emarginata) distribution, successional dynamics, and implications for the role ofthe seed bank. Can. J. Bot. 76:1725–1732.

Ohmann, J.L. and M.J. Gregory. 2002. Predictive mapping of forest composition and structure with direct gradient analysis and nearestneighbor imputation in coastal Oregon, . Can. J. For. Res.32:725–741.

Ohmann, J.L. and T.A. Spies. 1998. Regional gradient analysis and spatial pattern of woody plant communities of Oregon forests. Ecol. Monogr.68:151–182.

Oliver, C.D. 1981. Forest development in North America following major disturbances. For. Ecol. Manage. 3:153–168.

Oliver, C.D. and B.C. Larson. 1990. Forest stand dynamics. McGraw-Hill,New York, N.Y.

Peterson, E.B. and B. McCune. 2001. Diversity and succession of epiphyticmacrolichen communities in low-elevation managed conifer forests in western Oregon. J. Veg. Sci. 12:511–524.

Pickett, S.T.A., S.L. Collins, and J.J. Armesto. 1987. Models, mechanisms and pathways of succession. Bot. Rev. 53:335–371.

Pitkanen, S. 1997. Correlation between stand structure and ground vegetation: an analytical approach. Plant Ecol. 131:109–126.

Platt, W.J. and J.H. Connell. 2003. Natural disturbances and directional replacement of species. Ecol. Monogr. 73:507–522.

Pojar, J., K. Klinka, and D. Demarchi. 1991. Coastal Western Hemlock zone.In Ecosystems of British Columbia. D. Meidinger and J. Pojar (editors). B.C. Min. For., Res. Br.,Victoria, B.C., pp. 95–112

Pojar, J., K. Klinka, and D.V. Meidinger. 1987. Biogeoclimatic ecosystem classification in British Columbia. For. Ecol. Manage. 22:119–154.

Pojar, J. and A. MacKinnon. 1994. Plants of coastal British Columbia. LonePine Publishing, Edmonton, Alta.

Pojar, J., R. Trowbridge, and D. Coates. 1984. Ecosystem classification andinterpretation of the Sub-Boreal Spruce zone, Prince Rupert Forest Region, British Columbia. B.C. Min. For., Victoria, B.C. Land Manage.Rep. 17.

Rabotnov, T.A. 1986. Types of plant strategies. Soviet J. Ecol. 16:123–130.

Ramenskii, L.G. 1935. Principal apparatus, major concepts, and terms in theproduction typology of lands, geobotany, and ecology. Soviet Botany4:25–42.

———. 1938. Introduction to comprehensive soil-geobotanical study oflands. Sel’khozgiz, Moscow, Russia. (In Russian.)

Rice, W. 1989. Analyzing tables of statistical tests. Evolution 43:223–225.

Roberts, M.R. and L.X. Zhu. 2002. Early response of the herbaceous layer toharvesting in a mixed coniferous-deciduous forest in New Brunswick,Canada. For. Ecol. Manage. 155:17–31.

Roburn, A.E. 2003. Light transmission and understorey vegetation in two old-growth riparian stands: a study in spatial pattern. MRM thesis.Simon Fraser Univ., Burnaby, B.C.

Rochele, J.A. 1980. Mature forests, literfall and patterns of forage quality asfactors in the nutrition of black-tailed deer on Northern Vancouver Island. PhD dissertation. Univ. British Columbia, Vancouver, B.C.

Schoonmaker, P. and A. McKee. 1988. Species composition and diversityduring secondary succession of coniferous forests in western Cascademountains of Oregon. For. Sci. 34:960–979.

Shugart, H.H. 1997. Plant and ecosystem functional types. In Plant functionaltypes: their relevance to ecosystem properties and global change. T.M.Smith, H.H. Shugart, and F.I. Woodward (editors). Cambridge Univ.Press, Cambridge, U.K., pp. 20–43.

Spies, T.A. 1991. Plant species diversity and occurrence in young, mature, andold-growth Douglas-fir stands in western Oregon and Washington. InWildlife and vegetation of unmanaged Douglas-fir forests. M.H. Huff(editor). U.S. Dep. Agric. For. Serv., Portland, Oreg. Gen. Tech. Rep.-285, pp. 91–109.

———. 1997. Forest stand structure, composition, and function. In Creatinga forestry for the 21st century: the science of ecosystem management.K.A. Kohm and J.F. Franklin (editors). Island Press, Washington, D.C.,pp. 11–30.

Spies, T.A. and J.F. Franklin. 1988. Old growth and forest dynamics in theDouglas-fir region of western Oregon and Washington. Natural AreasJ. 8:190–201.

Inc. 1999. Base 10.0 user’s guide. Inc., Chicago, Ill.

Stamp, J.P. 2003. Abundance of forage plants and seasonal habitat value forgrizzly bears in the North Cascades, British Columbia. MRM thesis.Simon Fraser Univ., Burnaby, B.C.

Stewart, G.H. 1988. The influence of canopy cover on understorey development in forests of the western Cascade Range, Oregon. Vegetatio 76:79–88.

Swanson, F.J., J.A. Jones, and G.E. Grant. 1997. The physical environment asa basis for managing ecosystems. In Creating a forestry for the 21st century: the science of ecosystem management. K.A. Kohm and J.F.Franklin (editors). Island Press, Washington, D.C., pp. 229–237.

Tabachnik, B.G. and L.S. Fidell. 2001. Using multivariate statistics. 3rd ed. Allyn and Bacon, Boston, Mass.

Tappeiner, J.C. and P.B. Alaback. 1989. Early establishment and vegetativegrowth of understorey species in the western hemlock-Sitka spruceforests in southeast Alaska. Can. J. Bot. 67:318–326.

Tappeiner, J.C. and J.C. Zasada. 1993. Establishment of salmonberry, salal,vine maple, and bigleaf maple seedlings in the coastal forests of Oregon. Can. J. For. Res. 23:1775–1780.

Taylor, R.L. and B. MacBryde. 1977. Vascular plants of British Columbia: a descriptive resource inventory. Univ. British Columbia BotanicalGarden, Vancouver, B.C. Tech. Bull. 4.

Tilman, D. 1988. Plant strategies and the dynamics and structure of plantcommunities. Princeton Univ. Press, Princeton, N.J.

Van der Maarel, E. 1979. Transformation of cover-abundance values in phytosociology and its effects on community similarity. Vegetatio39:97–114.

Vanha-Majamaa, I. and J. Jalonen. 2001. Green tree retention in Fennoscan-dian forestry. Scand. J. For. Res. Suppl. 3:79–90.

Wang, T., A. Hamann, D.L. Spittlehouse, and S.N. Aitken. 2006. Develop-ment of scale-free climate data for western Canada for use in resourcemanagement. Intl. J. Climatology 26(3):383–397.

Wells, R W. 1996. Developmental trends of stand structure and tree mortalityin coastal western hemlock forests. MRM thesis. Simon Fraser Univ.,Burnaby, B.C.

Wells, R.W., K.P. Lertzman, and S.C. Saunders. 1998. Old-growth definitionsfor the forests of British Columbia, Canada. Natural Areas J.18:279–292.

Willson, M.F. and P.E. Hennon. 1997. The natural history of western skunkcabbage (Lysichiton americanum) in southeast Alaska. Can. J. Bot.75:1022–1025.

Wimberly, M.C. and T.A. Spies. 2001. Influences of environment and disturbance on forest patterns in coastal Oregon watersheds. Ecology82:1443–1459.

Woodward, F.I. and C.K. Kelly. 1997. Plant functional types: towards adefinition by environmental constraints. In Plant functional types:their relevance to ecosystem properties and global change. T.M. Smith,H.H. Shugart, and F.I. Woodward (editors). Cambridge Univ. Press,Cambridge, U.K., pp. 47–65.

Zager, P. 1980. The influence of logging and wildfire on grizzly bear habitatin northwestern Montana. PhD dissertation. Univ. Montana, Missoula, Mont.

Zar, J.H. 1999. Biostatistical analysis. Prentice Hall, Upper Saddle River, N.J.

Zimmerman, G.M., H. Goetz, and P.W. Mielke. 1985. Use of an improvedstatistical method for group comparisons to study effects of prairie fire.Ecology 66:601–611.