Modeling complex flow in a karst aquifer

wwwelseviercomlocatesedgeo

Sedimentary Geology 18

Modeling complex flow in a karst aquifer

John J Quinn David Tomasko 1 James A Kuiper 2

Environmental Science Division Argonne National Laboratory Argonne IL 60439 USA

Abstract

Carbonate aquifers typically have complex groundwater flow patterns that result from depositional heterogeneities and post-

lithification fracturing and karstification Various sources of information may be used to build a conceptual understanding of flow

in the system including drilling data well tests geophysical surveys tracer tests and spring gaging These data were assembled to

model flow numerically in Germanyrsquos Malm Formation at a site where water disappears from the beds of ephemeral stream

valleys flows through conduit systems and discharges to springs along surface water features Modeling was performed by using a

finite-difference approach with drain networks representing the conduit component of flow laced throughout the porous medium

along paths inferred on the basis of site data This approach represents an improvement over other karst models that attempt to

represent a conduit by a single specialized model node at the spring location or by assigning a computationally problematic

extremely high permeability to a zone By handling the conduit portion of this mixed-flow system with drains a realistic

interpretive flow model was created for this intricate aquifer

D 2005 Elsevier BV All rights reserved

Keywords Mixed-flow karst Malm Formation Groundwater modeling Conduits Hohenfels

1 Introduction

Because carbonate bedrock is potentially affected

not only by fracturing but also by dissolution ground-

water flow in carbonate terrain is typically a combina-

tion of diffuse fracture and conduit flow A regionrsquos

tectonic history may produce multiple sets of fractures

susceptible to solution enlargement and heterogeneities

within the bedrock facies may affect the dissolution

processes Ultimately the karst terrain that develops

may be a mixed-flow system with interacting compo-

nents of diffuse and conduit (or solution-enlarged frac-

0037-0738$ - see front matter D 2005 Elsevier BV All rights reserved

doi101016jsedgeo200511009

Corresponding author Tel +1 630 252 5357 fax +1 630 252

3611

E-mail address quinnjanlgov (JJ Quinn)1 Tel +1 630 252 6684 fax +1 630 252 36112 Tel +1 630 252 6206 fax +1 630 252 3611

ture) groundwater flow (Field 1993 Quinlan and

Ewers 1985)

In porous media settings an understanding of geo-

logic heterogeneity is critical for appropriate conceptual

or numerical modeling (Anderson 1990) and the de-

gree of interconnectedness of permeable zones is key to

understanding the flowfield (Fogg 1986) In karst set-

tings what is vitally important is the recognition and

appropriate modeling of preferential flowpaths such as

conduits Modeling groundwater flow in a karst envi-

ronment can be challenging and often produces results

that are highly uncertain because of the complexity of

flowpaths and lack of site-specific information In the

past many modeling approaches have been used to

simulate flow in a karst environment models using

an equivalent porous medium in which flow is gov-

erned by Darcyrsquos law (Anderson and Woessner 1992)

bblack-boxQ approaches in which functions are devel-

4 (2006) 343ndash351

ig 1 Location of CMTC Hohenfels and regional surface water

atures

JJ Quinn et al Sedimentary Geology 184 (2006) 343ndash351344

oped to reproduce input and output system responses

(recharge and flow at discharge springs) (eg Dreiss

1989ab) models in which the preferred flowpath is

simulated with a very high hydraulic conductivity rel-

ative to the surrounding matrix material (up to eight

orders of magnitude difference) (eg Teutsch 1989

Mace 1995 Eisenlohr et al 1997 Josnin et al 2000)

fracture network simulations in which individual frac-

tures are mapped and then studied (Long et al 1982

Long and Billaux 1987) and open channel equivalents

(Thrailkill et al 1991)

Simulation of a karst system composed of dendritic

paths (eg Milanovic 1981 White 1988 White and

White 1989) may require a great deal of site-specific

information for the karst channels and flow conditions

(eg elevation slope fill material roughness cross-

sectional area Reynolds number Froude number di-

ameter etc) (Field and Nash 1997 Field 1997) Be-

cause this information is difficult if not impossible to

obtain flow modeling in karst terrain is generally not

performed or simplifying assumptions are used

Interest in modeling various aspects of karst flow

systems is growing as evidenced by several recent

focused conferences conference sessions and collec-

tions of papers (eg Palmer et al 1999 Sasowsky and

Wicks 2000) Many of these papers deal with concep-

tual modeling geochemical modeling karst evolution

or statistical modeling Others are focused on techni-

ques for modeling flow in karst hydrologic systems

The purpose of this paper is to demonstrate an

approach to modeling heterogenous groundwater flow

in a mixed-flow karst setting Various forms of input

data and calibration data are used in the creation of this

model The approach shows promise as a method for

incorporating both the diffuse flow and conduit flow

components of a complicated karst flow system

2 Study area

This study is centered on the 150-km2 Combat

Maneuver Training Center (CMTC) Hohenfels Ger-

many (Fig 1) The region surrounding Hohenfels is

dominated by karst terrain of the Frankische Alb (up-

land) with a network of ephemeral valleys and topo-

graphic relief of roughly 100 m between the dry valleys

and the uplands Deeper valleys which contain peren-

nial rivers and creeks border the northern eastern and

southeastern edges of the facility

The ephemeral valleys contain Tertiary loam and

loess The Upper Jurassic Malm Formation (Fig 2)

dominates both the unsaturated zone and the upper

groundwater flow system The Malm consists mainly

F

fe

of four facies types bedded or massive limestone

bedded dolomite and reef dolomite (Apel 1971) The

exposed and near-surface bedrock at the CMTC is the

Malm Formationrsquos Kimmeridge (Delta) member

(Meyer 1990 Bayerischen Landesamt fur Wasser-

wirtschaft 1990a) The Delta facies are mainly reef

dolomite with minor layered dolomite Carbonate fa-

cies of the Malm Gamma Beta and Alpha members

underlie the Delta member These are in turn underlain

by the Zeta member of the Middle Jurassic Dogger

Formation The Zeta member also known as the Orna-

ten Clay is several meters thick

Bedrock in the Hohenfels region dips gently to the

east (von Freyburg 1969) Fractures exposed in the

Malm at CMTC are mainly oriented west-northwest

and north-northeast (Fuhrmann 1967) This fracture

orientation is consistent with the general pattern of

valley development The CMTC and vicinity have

numerous sinkholes (Fig 3) and the valleys typically

carry water only during a precipitation event (Heigold

et al 1994) The sinkholes sinking streams and spring

outflows are related to the structural history of the site

In order to improve the understanding of the layout of

fracture and conduit systems within a portion of the

CMTC geophysical surveys were performed (Fig 3)

Fig 2 Stratigraphic column for Malm Formation (modified from

Bayerischen Landesamt fur Wasserwirtschaft 1990a)

JJ Quinn et al Sedimentary Geology 184 (2006) 343ndash351 345

Electromagnetic surveying provided mapping of geo-

physical anomalies which are inferred preferential

flowpaths within the carbonate (Heigold et al 1994)

Data for these geophysical lineaments are concentrated

in the north-central portion of the CMTC and data are

limited elsewhere To determine the elevation of the

Ornaten Clay 28 vertical electrical soundings (VES)

were performed across the CMTC (Heigold et al

1994)

Clearly the influence of conduits is a key factor in

groundwater flow in the Hohenfels vicinity and the

groundwater in this karst setting is therefore considered

highly vulnerable to contamination from surficial

sources (Wrobel and Hanke 1987) The conceptual

model of groundwater flow at the CMTC and vicinity

is a mixed-flow karst system with recharge entering the

Malm flowing through the aquifer matrix entering

conduits and traveling rapidly to discharge springs

along the bordering Lauterach River Vils River or

Forellenbach (Forellen Creek) The low permeability

of the Ornaten Clay allows its upper surface to serve

as a lower flow model boundary condition with spatial-

ly variable elevation Only a few dye traces have been

performed to establish connections from several dry

valleys at the site to discharge springs (Fig 4) How-

ever the numerous electromagnetic surveys provide

insight into possible connections from dye release

points other dry valleys or sinkholes to discharge

springs Other types of data are available for the site

but in limited quantities These include drilling logs

water level data and hydraulic conductivity estimates

from pumping tests all from several monitoring wells

at the facilityrsquos two landfills (Fig 4) The pumping tests

yielded values of 00086ndash51 mday for the Malm

Formation (Wolf Blumenthal Ingenieurburo [WBI]

1992) Because the report does not include stratigraphic

logs or well construction diagrams for the wells it is

unclear which portions of the Malm were tested In

addition a few spring-flow measurements are available

from springs along the Lauterach River

In regions of moderate rainfall the formation of

fracture porosity through solution occurs rapidly in

the tens of meters immediately below the bedrock

surface where the groundwater flow system is open

and accessible to rapid recharge Throughout the

CMTC most recharge occurs in the uplands on the

flanks of the dry valleys where meteoric waters have

direct access to the Malm Formation Precipitation at

the Hohenfels measuring station has an average annual

value of 648 mm Accounting for evapotranspiration

and runoff the average annual groundwater recharge in

the area is about 200 mm (WBI 1992)

3 Methodology

31 Approach

Prior applications of finite-difference or finite-ele-

ment techniques in modeling flow in karst systems have

made use of specialized model cells or nodes in an

attempt to replicate some aspect of the flow system

Examples include assigning a drain (eg Yobbi 1989)

or a general head boundary (eg Dufresne and Drake

1999) to each model cell that represents an outlet

spring However the flow removed from the system

by the drain or general head boundary is limited to

seepage from the adjacent upgradient model cells the

contributions from upgradient conduits are ignored In

Fig 3 Sinkholes and geophysical lineaments in a portion of CMTC Hohenfels


the current study connected pathways of drain cells

simulate the conduit portion of mixed-flow karst aqui-

fers and are conceptually more accurate and realistic

For this study preferred flowpaths in a karst envi-

ronment are simulated by using the drain feature of the

finite-difference code MODFLOW (Harbaugh et al

2000) The upgradient end of the preferred flowpath

coincides with the location of a known surficial feature

(eg sinkhole lineament fracture) especially those

features that were the location of a dye release and

terminates downgradient at a surficial discharge point

Intermediate points are assigned on the basis of an

inferred flowpath determined from losing stream seg-

ments ephemeral stream beds fracture lineaments

other surficial characteristics and the results of geo-

physical surveys combined with the results of dye

traces The total discharge for the modeled conduit is

the sum of the discharges of each drain along the

branching drain network

This approach is similar to using a discrete singular

fracture set model (Teutsch and Sauter 1991) without

incorporating detailed information on the fractures

Rather the modeling addresses key hydrogeologic fea-

tures on a scale of less than 100 m to several kilometers

This scale is most important when considering flow and

transport (Thrailkill 1986) By using this method the

numerical instability associated with modeling an ex-

treme permeability contrast between a preferential

flowpath and the adjacent aquifer materials is avoided

and the influence of the conduit on the surrounding

aquifer can be more realistically demonstrated

This same approach has also been used on a Mis-

souri site (Quinn and Tomasko 2000) Although the

Missouri site lacked the benefit of geophysical linea-

ment data it had the advantage of a high density of

drilling data and water level data numerous tracer tests

and continuous spring gaging

32 Model construction

The modeling domain includes the entire CMTC and

off-base areas to the west and south Extending the

domain to regional boundaries takes advantage of hy-

drologic boundaries and allows for potential future

refinement of the model with more data Specified

head boundary conditions were applied to the perennial

creeks and rivers according to their typical stages No-

flow boundaries were applied to watershed drainage

divides which were assumed to approximate ground-

water divides The finite-difference modeling grid has

Fig 4 Selected springs tracer release points and landfill locations


cells at a uniform resolution of 50 m Because of the

scarcity of information on lithology and on the degree

of karstification with depth the model was constructed

with one vertical layer

More than 2100 drain cells were concentrated in the

portion of the modeling domain that has abundant

geophysical lineament results and several tracer tests

The drain feature in MODFLOW was originally devel-

oped to simulate agricultural drainage tiles that remove

water from an aquifer at a rate proportional to the

difference in water level (head) between the aquifer

and some fixed drain elevation as long as the head in

the aquifer is above that elevation (Harbaugh et al

2000) If the head in the aquifer falls below that of

the drain no additional water removal occurs For the

computations presented in this study drain elevations at

the discharge points of the preferred flowpaths were

assumed to be equal to the elevations of associated

springs or levels in surficial receiving waters At the

upstream end of the flowpaths the elevations were

estimated from drilling logs potentiometric maps of

the shallow groundwater systems and bedrock maps

To produce smooth transitions between cells elevations

of drains in model cells located along the inferred

conduit were estimated by linear interpolation from

the upgradient end through any intermediate nodes to

the downgradient end

In addition to drain elevations the drain conduc-

tance must also be specified This lumped parameter

incorporates information on characteristics of the drain

and its immediate surroundings as well as the head loss

between the drain and the aquifer (Harbaugh et al

2000) For simplicity a high conductance value was

selected to eliminate the need for drain-specific data

that would be difficult to obtain A value of 100 mday

per meter of conduit length was converted to drain

conductance for each drain cell This high value of

conductance promotes the removal of water from the

simulated conduits Although uncertainty is associated

with the conductance assigned to the drains use of this

high value produces a reasonable effect on potentio-

metric surfaces

able 1

ater level calibration data

alibration target Approximate average

water level (m)aModel-predicted

value (m)

ld Landfill

Monitoring Wells

416 414

ew Landfill

Monitoring Wells

420ndash441 422

a Krause (1997) and WBI (1992)


MODFLOW model construction was facilitated

using the Groundwater Modeling System (GMS) (Brig-

ham Young University [BYU] 2005)

The VES survey of the site provided localized top-

of-Dogger elevations consistent with regional mapping

(von Freyburg 1969) This surface served as the mod-

elrsquos bottom surface ranging in elevation from about

460 m above sea level in the northwest to about 300 m

in the southeast

Elevation data were obtained from Shuttle Radar

Topography Mission (SRTM) data (NASA 2000)

These data are provided in geographic coordinates at

a cell size of 36 s (approximately 76 m) As is typical

with SRTM data some small gaps of missing data were

present in the files SRTMFill v10 software (3D Na-

ture 2003) was used to fill the gaps The data were

projected to Universal Transverse Mercator (UTM)

Zone 32 WGS84 datum

An animation illustrating the model including

ground surface elevation conduit orientation and the

top of the Ornaten Clay is available on the Elsevier

website as an Electronic Supplement to this paper

Calibration is a procedure in which flow model

parameters are adjusted so that output better represents

flow and head measurements Adjusting drain eleva-

tions is one means of achieving model calibration By

changing the drain elevations and the length of inferred

conduits the match to target heads at site monitoring

wells can be improved

4 Results and discussion

The Hohenfels site has a limited monitoring well

networkndashsix wells located at the landfills near the cen-

ter of the modeling domainndashthat can be used for cali-

bration purposes A suitable calibration of the model

was achieved by first adjusting drain elevations to

ensure active flow in all inferred conduits and then

by adjusting the Malmrsquos hydraulic conductivity to

match the heads at the well locations The calibrated

hydraulic conductivity value was 15 mday which is

consistent with available pumping test data This value

was assigned uniformly in the model because of a lack

of areally distributed information across the study area

Results of the limited calibration data are shown in

Table 1

The simulated potentiometric contours are consis-

tent with overall regional flow directions (Andres and

Wirth 1985) but they also display the localized flow

directions resulting from the influence of the conduits

laced through the porous medium Fig 5 depicts the

drain pathways and resulting equipotentials in the cen-

T

W

C

O

N

ter of the study area as well as the flow vectors of the

calibrated model Localized changes in the direction of

groundwater flow illustrate the interaction of ground-

water with the adjacent surface water bodies and sug-

gest that the boundary conditions for the model are

defensible That is at the Forellenbach and the Vils and

Lauterach rivers water-level contours bend upstream

showing expected groundwater discharge Along the

modeled conduits the contours behave in the manner

expected for a mixed-flow karst environment dis-

charge from a diffuse medium and flow into open

conduits Groundwater converges on the preferential

flowpath and lines of equal potential point upstream

The calibrated modelrsquos water budget indicates that in

the portion of the modeling area containing drains

approximately 95 of the water that enters the system

as recharge exits as conduit flow to springs and only 5

is diffuse discharge to the northern boundary the Lau-

terach River One-time spring gaging measurements are

available for springs at Schmidmuhlen and Papiermuhle

(Bayerischen Landesamt fur Wasserwirtschaft 1990b)

These measurements are compared in Table 2 to model-

calculated fluxes which are automatically calculated

through MODFLOW runs within GMS Although mod-

eled conduit discharges do not match the measured

values the conditions at the time of each measurement

are unknown Spring discharge in karst terrains can have

highly variable flow depending on recent weather con-

ditions Therefore the available measurements are not

regarded as exact calibration targets but rather they are

tabulated to show that the method can provide results

approximating the values of spot measurements (Table

2) The calibration of the steady-state model could be

improved if spring gaging data were available at more

locations and ideally as continuous recordings

The resulting conduit discharge and diffuse dis-

charge along the Lauterach boundary are somewhat

inconsistent with an isotope tracer study elsewhere in

the Malm Formation which showed conduit discharge

to be 25ndash70 of total groundwater flux (Weise et al

2001) The Hohenfels proportions however compare

favorably with other carbonate aquifers (Worthington

1999)

Fig 5 Equipotential contours drain cell locations and flow vectors in the north-central portion of the CMTC site Location shown in Fig 3


The method is most accurate for the central portion

of the study area because of the surrounding drain

systems It is a method that is scale-dependent proper

use of this finite-difference approach depends on

matching a conceptual model that is appropriate to a

particular scale of study

At other study areas application of this technique

would involve the selection of both appropriate grid

spacing and also a vertical resolution achieved by

dividing the model into multiple layers The layering

would depend on available information on lithologic

and hydrogeologic heterogeneities including the depth

and degree of weathering in the carbonate

5 Conclusions

This paper presents a method of numerically mod-

eling the heterogeneities of flow in a karst environment

Table 2

Spring discharge calibration data

Calibration target One-time discharge

measurement (Ls)aModel-predicted

value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

a Bayerischen Landesamt fur Wasserwirtschaft (1990b)

by assigning sequences of adjacent model cells with

drains to simulate conduits Although sparse site data

limit the accuracy of the current model in the Hohenfels

study area it serves as an example of a straightforward

approach to modeling the intricate groundwater flow in

a karst terrain This model is considered interpretive

with a focus on demonstrating a method rather than

providing a refined calibrated solution

With improved coverage of site data (eg tracer

tests drilling data well tests water level data geophys-

ical surveys spring surveys spring flow gaging) an

interpretive model such as this could evolve into a more

effective tool for testing conceptual models identifying

data gaps assessing water resources or comparing

remediation scenarios

Acknowledgements

The comments and suggestions of Steve Worthing-

ton Timothy Eaton and an anonymous reviewer pro-

vided valuable improvements to this paper This work

was supported in part by the US Department of De-

fense US Army under interagency agreement

through US Department of Energy contract W-31-

109-Eng-38 and in part by the US Department of

Energy Office of Environmental Restoration and

Waste Management under contract W-31-109-Eng-38


Appendix A Supplementary data

Supplementary data associated with this article can

be found in the online version at 101016jsedgeo

200511009

References

Anderson MP 1990 Aquifer heterogeneitymdasha geological perspec-

tive In Bachu S (Ed) Proceedings of the Fifth Canadian

American Conference on Hydrogeology National Water Well

Association Dublin Ohio pp 3ndash22

Anderson MP Woessner WW 1992 Applied Groundwater Mod-

eling Simulation of Flow and Advective Transport Academic

Press Inc New York

Andres G Wirth H 1985 Grundwassergleichenkarte von Bayern

(1 500000) Bayerisches Landesamt fur Wasserwirtschaft

Munchen

Apel R 1971 Hydrogeologische Untersuchungen im Malmkarst der

Sudlichen und Mittleren Frankenalb Geologica Bavarica 64

268ndash355

Bayerischen Landesamt fur Wasserwirtschaft 1990a Geologische

Karte von Bayern (1 25000) 6736 Velburg Bayerischen Land-

esamt fur Wasserwirtschaft Munchen

Bayerischen Landesamt fur Wasserwirtschaft 1990b Verzeichnis der

Quellen in Bayern Bayerischen Landesamt fur Wasserwirtschaft

Munchen

Brigham Young University 2005 GMS Groundwater Modeling Sys-

tem Version 50 Engineering Computing Graphics Laboratory

Brigham Young University Provo Utah

Dreiss SJ 1989a Regional scale transport in a karst aquifer 1

Component separation of spring flow hydrographs Water

Resources Research 25 (1) 117ndash125

Dreiss SJ 1989b Regional scale transport in a karst aquifer 2

Linear systems and time moment analysis Water Resources Re-

search 25 (1) 126ndash134

Dufresne DP Drake CW 1999 Regional groundwater flow model

construction and wellfield site selection in a karst area Lake City

Florida Engineering Geology 52 129ndash139

Eisenlohr L Bouzelboudjen M Kiraly L Rossier Y 1997

Numerical versus statistical modeling of natural response of

a karst hydrogeological system Journal of Hydrology 202

244ndash262

Field MS 1993 Karst hydrology and chemical contamination

Journal of Environmental Systems 22 (1) 1ndash26

Field MS 1997 Risk assessment methodology for karst aquifers

(2) solute-transport modeling Environmental Monitoring and As-

sessment 47 23ndash37

Field MS Nash SG 1997 Risk assessment methodology for karst

aquifers (1) estimating karst conduit-flow parameters Environ-

mental Monitoring and Assessment 47 1ndash21

Fogg GE 1986 Groundwater flow and sand body interconnected-

ness in a thick multiple-aquifer system Water Resources Re-


Fuhrmann D 1967 Stratigraphische und Tektonische Untersuchun-

gen auf den Kartenblattern Kastl und Velburg (1 25000) Dipl-

Arb University of Erlangen Germany

Harbaugh AW Banta ER Hill MC McDonald MG 2000

MODFLOW-2000 the US Geological Survey Modular Ground-

Water Model - User Guide to Modularization Concepts and the

Ground-Water Flow Process Open-File Report 00-92US Geo-

logical Survey Reston Virginia

Heigold PD Thompson MD Borden HM 1994 Geophysical

Exploration in the Lautertal at the Combat Maneuver Training

Center Hohenfels Germany Argonne National Laboratory

Argonne Illinois ANLESDTM-82

Josnin J-Y Pistre S Drogue C 2000 Modelisation drsquoun systeme

karstique complexe (basin de St Chaptes Gard France) un outil

de synthese des donnees geologiques et hydrogeologiques Cana-

dian Journal of Earth Sciences 37 1425ndash1445

Krause H 1997 Results of groundwater monitoring in the vicinity

of the new sanitary landfill in October 1997 Prepared by Dr

Reitzler and Heidrich Gmbh Nurnberg Germany October 13

Long JCS Billaux DM 1987 From field data to fracture network

modeling an example incorporating spatial structure Water


Long JCS Remer JS Wilson CR Witherspoon PA 1982

Porous media equivalents for networks of discontinuous fractures

Water Resources Research 18 (3) 645ndash658

Mace RE 1995 Geostatistical description of hydraulic properties in

karst aquifers a case study in the Edwards Aquifer Proceedings

International Symposium on Groundwater Management Ameri-

can Society of Civil Engineers New York pp 193ndash198

Meyer RFK 1990 Erlauterungen zur Geologischen Karte Blatt

6736 Velburg (1 25000) Bayerisches Geologisches Landesamt

Munchen

Milanovic PT 1981 Karst Hydrogeology Water Resources Pub-

lications Littleton Colorado

NASA 2000 Shuttle Radar Topography Mission Data Available at

ftpedcsgs9crusgsgovpubdatasrtm accessed April 2004

Palmer AN Palmer MV Sasowsky ID (Eds) 1999 Karst

Modeling Proceedings Karst Waters Institute Special Publica-

tion vol 5 Karst Waters Institute Charles Town WV

Quinlan JF Ewers RO 1985 Ground water flow in limestone

terraces strategy rationale and procedure for reliable efficient

monitoring of ground water quality in karst areas Proceedings

Fifth National Symposium on Aquifer Restoration and Ground

Water Monitoring National Water Well Association Dublin

Ohio pp 197ndash234

Quinn J Tomasko D 2000 A numerical approach to simulating

mixed flow in karst aquifers In Sasowsky I Wicks C (Eds)

Groundwater Flow and Contaminant Transport in Carbonate

Aquifers AA Balkema Rotterdam Holland pp 147ndash156

Sasowsky I Wicks C (Eds) 2000 Groundwater Flow and Con-

taminant Transport in Carbonate Aquifers AA Balkema Rot-

terdam Holland

Teutsch G 1989 Groundwater models in karstified terrains

two practical examples from the Swabian Alb (S Germany)

Proceedings Solving Ground Water Problems with Models In-

ternational Ground Water Modeling Center Indianapolis Indiana

pp 929ndash953

Teutsch G Sauter M 1991 Groundwater modeling in karst ter-

rains scale effects data acquisition and field validation Proceed-

ings Third Conference on Hydrogeology Ecology Monitoring

and Management of Ground Water in Karst Terrain National

Water Well Association Dublin Ohio pp 17ndash35

Thrailkill J 1986 Models and methods for shallow conduit-flow

carbonate aquifers Proceedings Environmental Problems in Karst

Terranes and Their Solutions Conference National Water Well


Thrailkill J Sullivan SB Gouzie DR 1991 Flow para-

meters in a shallow conduit-flow carbonate aquifer Inner


Bluegrass Karst Region Kentucky USA Journal of Hydrology

129 87ndash108

von Freyburg B 1969 Tektonische Karte der Frankischen Alb und

ihrer Umgebung Erlanger Geologische Abhandlungen vol 77

Erlanger Geologische Abhandlungen Erlangen Germany

Weise SM Rau I Seiler K-P 2001 Long-term storage behaviour

of karstic aquifer deduced by multi-trace investigations EGS

XXVI General Assembly European Geophysical Society Nice

France

White WB 1988 Geomorphology and Hydrology of Karst Terrains

Oxford University Press New York

White WB White EL 1989 Karst Hydrology Concepts from the

Mammoth Cave Area Van Nostrand Reinhold New York

Wolf Blumenthal Ingenieurburo (WBI) 1992 Study concerning the

existing landfill Hohenfels Phase I Hochstrasse 8b 8540 Red-

nitzhembach Germany prepared for CMTC Hohenfels June 12

Worthington SRH 1999 A comprehensive strategy for understand-

ing flow in carbonate aquifers In Palmer AN Palmer MV

Sasowsky ID (Eds) Karst Modeling Proceedings Karst Waters

Institute Special Publication vol 5 Karst Waters Institute

Charles Town WV pp 30ndash37

Wrobel J-P Hanke K 1987 Karten der Gefahrdung der Grund-

wasser in Bayern durch Nitrat GLA Fachberichte vol 3 Bayer-

isches Geologisches Landesamt Munchen pp 3ndash25

Yobbi D 1989 Simulation of steady-state ground water and spring

flow in the upper Floridian aquifer of coastal Citrus and Hernando

Counties Florida Water-Resources Investigations Report 88-

4036 US Geological Survey Tallahassee FL

3D Nature 2003 SRTMFill v10 Software Available at http

www3dnaturecomsrtmfillhtml accessed April 2004

ig 1 Location of CMTC Hohenfels and regional surface water

atures


oped to reproduce input and output system responses

(recharge and flow at discharge springs) (eg Dreiss

1989ab) models in which the preferred flowpath is

simulated with a very high hydraulic conductivity rel-

ative to the surrounding matrix material (up to eight

orders of magnitude difference) (eg Teutsch 1989

Mace 1995 Eisenlohr et al 1997 Josnin et al 2000)

fracture network simulations in which individual frac-

tures are mapped and then studied (Long et al 1982

Long and Billaux 1987) and open channel equivalents

(Thrailkill et al 1991)

Simulation of a karst system composed of dendritic

paths (eg Milanovic 1981 White 1988 White and

White 1989) may require a great deal of site-specific

information for the karst channels and flow conditions

(eg elevation slope fill material roughness cross-

sectional area Reynolds number Froude number di-

ameter etc) (Field and Nash 1997 Field 1997) Be-

cause this information is difficult if not impossible to

obtain flow modeling in karst terrain is generally not

performed or simplifying assumptions are used

Interest in modeling various aspects of karst flow

systems is growing as evidenced by several recent

focused conferences conference sessions and collec-

tions of papers (eg Palmer et al 1999 Sasowsky and

Wicks 2000) Many of these papers deal with concep-

tual modeling geochemical modeling karst evolution

or statistical modeling Others are focused on techni-

ques for modeling flow in karst hydrologic systems

The purpose of this paper is to demonstrate an

approach to modeling heterogenous groundwater flow

in a mixed-flow karst setting Various forms of input

data and calibration data are used in the creation of this

model The approach shows promise as a method for

incorporating both the diffuse flow and conduit flow

components of a complicated karst flow system

2 Study area

This study is centered on the 150-km2 Combat

Maneuver Training Center (CMTC) Hohenfels Ger-

many (Fig 1) The region surrounding Hohenfels is

dominated by karst terrain of the Frankische Alb (up-

land) with a network of ephemeral valleys and topo-

graphic relief of roughly 100 m between the dry valleys

and the uplands Deeper valleys which contain peren-

nial rivers and creeks border the northern eastern and

southeastern edges of the facility

The ephemeral valleys contain Tertiary loam and

loess The Upper Jurassic Malm Formation (Fig 2)

dominates both the unsaturated zone and the upper

groundwater flow system The Malm consists mainly

F

fe

of four facies types bedded or massive limestone

bedded dolomite and reef dolomite (Apel 1971) The

exposed and near-surface bedrock at the CMTC is the

Malm Formationrsquos Kimmeridge (Delta) member

(Meyer 1990 Bayerischen Landesamt fur Wasser-

wirtschaft 1990a) The Delta facies are mainly reef

dolomite with minor layered dolomite Carbonate fa-

cies of the Malm Gamma Beta and Alpha members

underlie the Delta member These are in turn underlain

by the Zeta member of the Middle Jurassic Dogger

Formation The Zeta member also known as the Orna-

ten Clay is several meters thick

Bedrock in the Hohenfels region dips gently to the

east (von Freyburg 1969) Fractures exposed in the

Malm at CMTC are mainly oriented west-northwest

and north-northeast (Fuhrmann 1967) This fracture

orientation is consistent with the general pattern of

valley development The CMTC and vicinity have

numerous sinkholes (Fig 3) and the valleys typically

carry water only during a precipitation event (Heigold

et al 1994) The sinkholes sinking streams and spring

outflows are related to the structural history of the site

In order to improve the understanding of the layout of

fracture and conduit systems within a portion of the

CMTC geophysical surveys were performed (Fig 3)












1994)












































3 Methodology

31 Approach









































































































metric surfaces

able 1




value (m)

ld Landfill

Monitoring Wells

416 414

ew Landfill

Monitoring Wells

420ndash441 422











in the southeast









Zone 32 WGS84 datum


























Table 1







T

W

C

O

N











































1999)
















5 Conclusions



Table 2




value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

















Acknowledgements














200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France

































1994)












































3 Methodology

31 Approach









































































































metric surfaces

able 1




value (m)

ld Landfill

Monitoring Wells

416 414

ew Landfill

Monitoring Wells

420ndash441 422











in the southeast









Zone 32 WGS84 datum


























Table 1







T

W

C

O

N











































1999)
















5 Conclusions



Table 2




value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

















Acknowledgements














200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France



















































































































metric surfaces

able 1




value (m)

ld Landfill

Monitoring Wells

416 414

ew Landfill

Monitoring Wells

420ndash441 422











in the southeast









Zone 32 WGS84 datum


























Table 1







T

W

C

O

N











































1999)
















5 Conclusions



Table 2




value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

















Acknowledgements














200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France

































































metric surfaces

able 1




value (m)

ld Landfill

Monitoring Wells

416 414

ew Landfill

Monitoring Wells

420ndash441 422











in the southeast









Zone 32 WGS84 datum


























Table 1







T

W

C

O

N











































1999)
















5 Conclusions



Table 2




value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

















Acknowledgements














200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France






















able 1




value (m)

ld Landfill

Monitoring Wells

416 414

ew Landfill

Monitoring Wells

420ndash441 422











in the southeast









Zone 32 WGS84 datum


























Table 1







T

W

C

O

N











































1999)
















5 Conclusions



Table 2




value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

















Acknowledgements














200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France





































5 Conclusions



Table 2




value (Ls)

Schmidmuhlen 73 155

Papiermuhle 205 35

















Acknowledgements














200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France


























200511009

References







Press Inc New York



Munchen



268ndash355






Munchen
















244ndash262











































Munchen













Ohio pp 197ndash234







terdam Holland





pp 929ndash953














129 87ndash108







France
























129 87ndash108







France






















Documents

Modeling complex flow in a karst aquifer