Hydrological Sciences Journal

ISSN: 0262-6667 (Print) 2150-3435 (Online) Journal homepage: http://www.tandfonline.com/loi/thsj20

Modelling hydrological processes in mesoscale lowland river basins with SWAT—capabilities and challenges BRITTA SCHMALZ , FILIPA TAVARES & NICOLA FOHRER To cite this article: BRITTA SCHMALZ , FILIPA TAVARES & NICOLA FOHRER (2008) Modelling hydrological processes in mesoscale lowland river basins with SWAT—capabilities and challenges, Hydrological Sciences Journal, 53:5, 989-1000, DOI: 10.1623/hysj.53.5.989 To link to this article: http://dx.doi.org/10.1623/hysj.53.5.989

Published online: 18 Jan 2010.

Submit your article to this journal

Article views: 927

View related articles

Citing articles: 25 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=thsj20 Download by: [37.44.207.199]

Date: 27 January 2017, At: 13:54

Hydrological Sciences–Journal–des Sciences Hydrologiques, 53(5) October 2008 Special issue: Advances in Ecohydrological Modelling with SWAT

989

Modelling hydrological processes in mesoscale lowland river basins with SWAT—capabilities and challenges BRITTA SCHMALZ, FILIPA TAVARES & NICOLA FOHRER Department of Hydrology and Water Resources Management, Ecology Centre, Kiel University, Olshausenstrasse 40, D-24098 Kiel, Germany [email protected]

Abstract Lowland areas are characterized by specific properties, such as flat topography, low hydraulic gradients, shallow groundwater, and high potential for water retention in peatland and lakes. These characteristics and their dominating hydrological processes have to be assessed and considered for the analysis and modelling of water balances in lowland catchments. The capabilities and challenges of modelling hydrological processes and water balances in mesoscale lowland river basins with the SWAT model are presented. The investigated catchments Stör, Treene and Kielstau are located in northern Germany within a lowland area. Covering areas from 50 to 517 km2, these rural meso-catchments have sandy, loamy and peaty soils. Drainage, in terms of tile drainage and open ditches, has changed the natural water balance. The set-up and modifications of the model as applied in these case studies can be transferred to other similar catchments in lowland areas. The dominating hydrological processes were found to be mainly controlled by groundwater dynamics and storage, drainage, wetlands and ponds. Some groundwater parameters were found to be highly sensitive and they turned out to be the most influential factors for improving simulated water discharge. Key words ecohydrological model; lowland hydrology; SWAT; mesoscale river catchment; groundwater–surface water interactions

Modélisation des processus hydrologiques en basins versants de méso-échelle de plaine avec SWAT–aptitudes et défis Résumé Les zones de plaine sont caractérisées par des propriétés spécifiques telles que une topographie plate, de faibles gradients hydrauliques, des nappes peu profondes et un fort potentiel de rétention d’eau dans des marais et des lacs. Ces caractéristiques et les processus hydrologiques dominants associés doivent être évalués et pris en compte pour l’analyse et la modélisation des bilans hydrologiques des bassins versants de plaine. Les aptitudes et les défis de modélisation, avec le modèle SWAT, des processus et des bilans hydrologiques de bassins versants de plaine de méso-échelle sont présentés. Les bassins versants étudiés de Stör, Treene et Kielstau sont situés dans une plaine du nord de l’Allemagne. Avec des superficies de 50 à 517 km2, ces bassins ruraux de méso-échelle présentent des sols sableux, limoneux et tourbeux. Le drainage enterré et par fossés a modifié le bilan hydrologique naturel. La mise en place et les modifications du modèle pour application à ces études de cas peuvent être transposées à d’autres bassins similaires de plaine. Les processus hydrologiques dominants apparaissent être principalement contrôlés par la dynamique et le stockage de la nappe, le drainage, les zones humides et les retenues. Quelques paramètres hydrogéologiques se sont révélé être très sensibles et donc être les facteurs les plus influents pour améliorer les simulations de débit. Mots clefs modèle éco-hydrologique; hydrologie de plaine; SWAT; bassin versant de méso-échelle; interactions eau de surface–eau souterraine

INTRODUCTION Lowland areas are characterized by specific properties, such as flat topography and low hydraulic gradients. There is high potential for water retention in peatland (Kieckbusch et al., 2006) and lakes. Furthermore, there is shallow groundwater to be considered, which results in intensive groundwater–stream water interactions (Sophocleous, 2002; Schmalz et al., 2008). The natural water balance, including the soil water balance, has been changed substantially by human intervention, such as river regulation, pumping stations, and drainage systems such as tile drainage and open ditches. The removal of surplus water causes a drawdown of the groundwater level and a generally reduced flow duration, as well as improved water movement. In the study areas, the original landscape was characterized by far higher groundwater levels and large parts of the riparian areas were swampy. Today, the areas are characterized by many ditches, drainage pipes and by canalization; the river courses have been straightened, deepened and

Open for discussion until 1 April 2009

Copyright © 2008 IAHS Press

990

Britta Schmalz et al.

widened, and small-scale geomorphological depressions have been flattened. The slow flow velocities, resulting from small hydraulic gradients, were intended to be increased by straightening the river courses. Thus the naturally meandering courses of the rivers disappeared from the landscape. To prevent frequent flooding, drainage pipes were installed and ditch systems extended. Therefore, the formerly small-scale water and nutrient cycles became more widespread, resulting in faster and more intensive transport to the river network, and the accelerated discharge, in turn, required further extension of the river profiles. In particular, the river beds in the downstream river courses were substantially deepened, widened and developed to a classical trapezoid discharge profile. By levelling the landscape, natural nutrient sinks in the riparian areas were destroyed. Accumulation zones of nutrients were shifted toward the river. As drainage lowered the soil water level, the water balance of the soil and landscape became more dynamic. The river morphology was changed by a multiplicity of dams, river bed regulations, bed load and sand chambers, and partly canalized passages. Mesoscale hydrological and ecohydrological models could be useful tools to answer questions related to land use and water management. Hydrological and soil parameters of the catchment water balance and water quality can be modelled based on climatic data. These models are suitable to calculate the water and matter fluxes over a wide range of different spatial and temporal scales (Fujita, 1986). Hörmann et al. (2005) gave an overview of the prospects and limitations of ecohydrological models, and discussed the linkage between land use and the water cycle in mesoscale approaches. In general, these models can be used as analytical tools for sustainable river basin management and for estimation of nutrient entry pathways. Ecohydrological models have already been used in lowland catchments: Habeck et al. (2005) modelled nitrogen transport in the Nuthe lowland catchment in northeastern Germany with SWIM (Krysanova et al., 1998), and reproduced the nitrogen flows sufficiently well. Hattermann et al. (2006) incorporated a riparian zone and wetland module into SWIM to consider the effects of riparian zones and wetlands on nutrient transport into surface water of the same lowland catchment. The water balance of wetlands within flood plains of the northeastern German Havel River basin was modelled (Krause & Bronstert, 2007; Krause et al., 2007a,b) with the IWAN model (Krause & Bronstert, 2005); their simulation results proved the close interaction between river and flood plain. To assess lowland catchments, the specific properties of these areas need to be considered. The dominant hydrological characteristics and processes need to be identified and analysed to efficiently simulate the water balance. The objective of this paper is to identify the capabilities and challenges of modelling hydrological processes in mesoscale lowland river basins. The first results from the application of an ecohydrological model are presented. MATERIALS AND METHODS Investigation area The mesoscale investigation sites are parts of a lowland area of Schleswig-Holstein located in northern Germany. Sandy, loamy and peaty soils are characteristic for these catchments. Land use is dominated by arable land and pasture in different proportions. We examined three catchment areas (Fig. 1) which cover different scales. The Stör catchment (468 km2, 25 km river length): The upper part of the Stör catchment to the Willenscharen gauge was considered. The slope of the central model area is small: falling from 90 m and 60 m in the western and eastern parts, respectively, to 2 m a.m.s.l. at the outlet. Only the southwestern region has gradients of more than 3°, but in most of the catchment it is usually smaller than 1° (LVA S-H, 1995). The precipitation is 831 mm/year and the mean annual temperature is 8.3°C. There are mainly sandy soils, few peat soils at the river valleys (fen) and depressions (bog) and gleysols in the eastern parts (Table 1). The largest tributaries of the Stör are the Schwale and the Bünzener Au. The Bünzener Au is the central discharge system of the western part of the study area. About 44% of the area drains into the Bünzener Au (Dobslaff, 2005). The Copyright © 2008 IAHS Press

Modelling lowland hydrological processes in mesoscale river basins

991

Fig. 1 Location of the three investigated lowland catchments.

investigation area contains some small ponds and lakes. The larger Einfelder See (Lake) drains into the Stör via the tributary Aalbek. There is significant additional water storage in the Dosenmoor bog (575 ha). The landscape is characterized by many open ditches and drainage pipes. For the Buckener Au catchment, Venohr (2000) estimated that the sub-catchment area drained by open ditches and drainage pipes varies from 32 to 40%. Treene catchment (517 km2, 60 km): The watershed that drains into the river between the source and the catchment outlet at the Treia gauge was considered. The maximum height difference is 76 m. Average precipitation is 872 mm/year and the mean annual temperature 8.2°C. The Bondenau and Kielstau are the source rivers of the Treene. Both flow into Lake Treßsee, where the Treene has its origin. The river morphology of the Treene between Oeversee and Treia has not been changed due to river or flow regulation measures. Indeed, large areas of the catchment are drained by open ditches. Floods in the riparian areas are caused mainly by rain events, snowmelt and the low hydraulic gradient. Larger tributaries to the Treene up to Treia are the Jerrisbek, the Bollingstedter Au and the Jübek. There are three larger lakes: Sankelmarker See (0.56 km2), Südensee (0.64 km2) and Treßsee (0.17 km2) (Dey, 2004). Kielstau catchment (50 km2, 17 km): In a nested approach, we also examined the Kielstau catchment, a sub-catchment of the Treene. The River Kielstau flows through the Lake Winderatter See and has three tributaries from the north: the Moorau, the Levensau and the Hennebach. Various smaller tributaries and water from drainage pipes and open ditches flow into the Kielstau. The drained fraction of agricultural areas in the Kielstau catchment is estimated to be approx. 38% (Fohrer et al., 2007). The Kielstau has been changed markedly from its natural course. The Soltfeld gauge is situated at the catchment outlet. The maximum height difference is about 45 m.

Copyright © 2008 IAHS Press

992

Britta Schmalz et al.

Table 1 Numbers of main structural units: sub-catchments, hydrotopes, and major soil and land-use classes in the three catchments studied. Sub-catchments Hydrotopes Soil types Soil types, % of area (>2%): Haplic podsols Gleyic podsols Hablic cambisols/arenosols Stagnic luvisols Haplic luvisols Stagnic cambisols Stagnic gleysols Haplic gleysols Gleyic anthrosols Sapric histosols Ombric histosols Land-use classes Land use, % of area (>2%): Arable land Pasture Urban Deciduous forest Coniferous forest Fallow/bush land

Stör 22 298 7

Treene 58 681 7

Kielstau 8 117 7

25.0 23.0

8.7 40.8 4.9 24.2 7.3

2.2

12.0 8.0

52.2 16.6 7.2 10.1

12.0 7.0 3.0 5

5.8 2.2 6

48.1 29.5 8.2 2.2 6.9

39.8 37.2 3.3 11.1 1.6 6.3

2.2 8.7 5 55.8 26.1 3.1 8.6 5.6

Database The information required for the model set-up was derived from environmental data of the German authorities: A digital elevation model (50 × 50 m, LVA S-H, 1995), a stream network (LVA S-H 1995, 2003), land use data (250 × 250 m, EEA, 2000; and 25 × 25 m, DLR, 1995) and soil maps (1:100 000, Finnern, 1997; and 1:200 000, BGR, 1999) were used. In addition, time series data from our own measuring campaigns were used, as well as those from German authorities and institutions: meteorological data from Landesamt für Natur und Umwelt des Landes Schleswig-Holstein (LANU-SH), Staatliches Umweltamt in Schleswig (StUA-Schleswig) and Deutscher Wetterdienst (DWD). For model calibration and validation the following discharge data were used: (a) Stör catchment: data from both the German authorities (LANU-SH: 10 gauges, 1990–2002, hourly resolution) and the Ecology Centre, Kiel University (12 gauges, 1992–1994) both calculated to mean daily values. (b) Treene catchment: data (1984–2000) from eight gauges (LANU-SH; hourly resolution, calculated to daily mean values). (c) Kielstau catchment: data from both the German authorities (LANU-SH; 15-min resolution, calculated to daily mean values) and our own measuring campaigns (1986–2005). ECOHYDROLOGICAL MODELLING The SWAT model The river basin model SWAT (Soil and Water Assessment Tool, Arnold et al., 1998) was used in order to assess the water balance in these complex hydrological catchment areas. The SWAT model is a semi-distributed, process-oriented model for simulating water, nutrient and pesticide transport. Simulations are conducted for mesoscale catchments and their hydrological response Copyright © 2008 IAHS Press

993

Modelling lowland hydrological processes in mesoscale river basins

units (HRU; hydrotopes). The HRUs are defined by unique land-use and soil combinations within each sub-basin. The advantages of SWAT include the possibility to perform spatially differentiated analyses, to investigate seasonal dynamics, land-use changes, and different management options. For this study, we worked with AVSWAT2000 (Neitsch et al., 2002a,b) with the ArcView3.x interface (Di Luzio et al., 2002) and used daily time steps. Parameterization of the model The ecohydrological modelling with SWAT was conducted by dividing the three study catchments into sub-catchments and hydrotopes. Different soil types and land-use classes were assigned (Table 1). Sensitivity tests were conducted to assess the most sensitive parameters for the model set-up in these particular lowland catchments. For the calibration of the Kielstau catchment, this was performed by changing each parameter ten times within its allowable range. One parameter at a time was adjusted, while the others were kept unchanged. Table 2 shows the variables considered, their definition and their allowable range in SWAT. It was not the intention of this study to apply autocalibration to achieve optimum fit between measured and modelled curves. The examination of lowland areas is a representation of a different landscape type with specific processes that are not considered a priori in the model set-up. The Table 2 Variables used for sensitivity analyses, their definition and allowable range (SWAT). Parameter group Groundwater

Variable name GW_DELAY GWQMN ALPHA_BF REVAPMN

Drainage

Infiltration Soil

Pond

GW_REVAP RCHRG_DP DDRAIN TDRAIN GDRAIN CN2 SOL_Z SOL_AWC SOL_K PND_FR PND_ESA PND_PSA PND_EVOL PND_PVOL

Wetland

WET_FR WET_MXSA WET_NSA WET_MXVOL WET_NVOL

Definition Delay time for aquifer recharge (d) Threshold water level in shallow aquifer for baseflow (mm H2O) Baseflow recession constant Threshold water level in shallow aquifer for revap or percolation to deep aquifer (mm H2O) Revap coefficient Aquifer percolation coefficient Depth to subsurface drain (mm) Time to drain soil to field capacity (h) Drain tile lag time (h) SCS runoff curve number for soil moisture condition II Depth from soil surface to bottom of layer (mm) available water capacity Saturated hydraulic conductivity (mm/h) Fraction of the sub-basin area draining into the pond Surface area of the pond when filled to the emergency spillway (ha) Surface area of the pond when filled to the principal spillway (ha) Volume of water held in the pond when filled to the emergency spillway (104 m3 H2O) Volume of water held in the pond when filled to the principal spillway (104 m3 H2O) Fraction of the sub-basin area draining into the wetland Surface area of the wetland when filled to the maximum water level (ha) Surface area of the wetland when filled to the normal water level (ha) Volume of water held in the wetland when filled to the maximum water level (m3 H2O) Volume of water held in the wetland when filled to the normal water level (m3 H2O)

Allowable range 0–500 0–5000 0–1 0-500 0.02–0.2 0–1 0–2000 0–72 0–100 35–98 0–3500 0–1 0–2000 0–1 0–200 0–1000 0–200 0–100 0–1 0–3000 0–3000 0–300 0–300

Copyright © 2008 IAHS Press

994

Britta Schmalz et al.

set-up of the model without autocalibration helps to identify model deficiencies for lowland processes, which would otherwise be masked by parameter tuning. RESULTS AND DISCUSSION Model modifications The first modelling runs did not represent the streamflow correctly. They showed underestimation of measured high winter discharge peaks, early response discharges in the late summer peaks, and overestimation of baseflow, mainly during the summer. Thus, the modelled discharge dynamics had reacted too slowly. Therefore, the model parameterizations were adapted in consideration of the conditions in these lowland catchments, mainly referring to the groundwater storage and dynamics and the drainage. For the Stör catchment the following modifications were introduced in the model (Dobslaff, 2005): (a) The drained area comprises between 2 and 62% of the different sub-catchments. Thus, drainage was considered for fen substrates, gleysols and podsols at a depth of 80 cm. The calibration was carried out based on the hydrotope data for sub-basins. (b) The water table depth varied between 0.1 and 9.0 m and the GW_DELAY was found to be from 0 to 2.6 d. (c) Baseflow recession constants ALPHA_BF were used for the Sarlhusen (0.0155), Padenstedt (0.0118) and Willenscharen (0.0073) gauges. These factors were calculated from discharge data applying a baseflow separation program developed by Arnold & Allen (1999). For the Treene catchment (Dey, 2004; Dey et al., 2004): (a) Drainage was implemented for 14% of the catchment area. (b) An increase in the recession constant ALPHA_BF to 0.0286 (at Treia gauge) and 0.0309 (at Mühlenbrück gauge) was derived by analysis of data from summer 1994. The values for the other sub-basins were set higher to present a realistic dynamic. (c) A percolation time of two days for the hilly areas of the catchment was assumed. For the Kielstau catchment the following parameters were modified (Tavares, 2006): (a) Groundwater retention time (GW_DELAY) of 50 d. (b) Explicit modelling of ponds and wetlands. (c) Drainage (DDRAIN as 600 mm and TDRAIN as 24 h). Groundwater parameters High sensitivity of some groundwater parameters was demonstrated. Due to the lack of information concerning the groundwater processes, these parameters were calibrated by semi-intuitive trial-and-error processes. The initial attributed parameters of the Kielstau model set-up were based primarily on the calibrated parameters of Dey (2004) for the uppermost sub-basin of the Treene catchment. Even though this sub-basin corresponds to the Kielstau catchment, it failed to simulate the groundwater regime in the Kielstau area accurately. The groundwater delay time parameter (GW_DELAY) describes the delay time for the water that moves past the lowest depth of the soil profile by percolation or bypass flow before becoming shallow aquifer recharge. It depends on the depth of the water table and the hydraulic properties of the geological formation in the vadose and groundwater zones. Figure 2 shows the effect of different values of groundwater delay time on streamflow in the last calibration year (1999) in the Kielstau catchment. The initial value of zero was changed in order to obtain a better fit of the modelled results to the observed data and was allowed to vary between 0 and 200 d. A reduction of the parameter value to 50 d allowed a better fit for the periods of low flow. Dobslaff (2005) found that GW_DELAY proved to be of low sensitivity for the Stör catchment. This factor could change the discharge considerably, with a realistic time delay of 100 d, as well as 30 d in the lateral regions, only in the upper river section of the Dosenbek, in a bog area. Copyright © 2008 IAHS Press

Modelling lowland hydrological processes in mesoscale river basins

995

Fig. 2 Comparison of parameterization of groundwater delay (GW_DELAY) for the year 1999, Kielstau catchment.

The threshold depths for baseflow (GWQMN) and re-evaporation (REVAPMN), as well as the groundwater “revap” coefficient (GW_REVAP) affect the amount of groundwater flow. In addition, these parameters control the upwelling of groundwater into the unsaturated soil zone. For the Kielstau catchment, these parameters were changed in the model set-up. The following groundwater calibration produced a change of small dimension which, nevertheless, allowed the improvement of the quality of the Kielstau simulation. In the Stör catchment these parameters affect only the end of periods with small or no precipitation, in which the height of the groundwater level decreased. The volume of the baseflow cannot be reduced uniformly in the summer months. These groundwater parameters were able to further lower the baseflow (Dobslaff, 2005). The parameter ALPHA_BF is sensitive to discharge. As Dobslaff (2005) shows, the dynamics of the draining behaviour of the groundwater storage are increased by larger values for this coefficient. However, a larger ALPHA_BF value leads to small baseflow during the summer months, while the discharges in winter are overestimated. Selecting a larger ALPHA_BF coefficient could reduce the base discharge during the summer months by faster drainage behaviour of the groundwater, but it did not succeed in lowering the discharge dynamics during the winter months by other model parameters (Dobslaff, 2005). Ponds and wetlands Some parts of the three investigation areas are occupied by wetlands and water impoundments, which are of high relevance for the hydrological regime. In particular, the calibration of pond and wetland parameters allowed a better fitting of the surface runoff simulation for the Kielstau catchment with 8.7% fen. This was particularly relevant for the periods of low flow. The observed peaks, which were already underestimated in the model simulation without ponds and wetlands, became further underestimated. Nevertheless, in an overall perspective, the inclusion of ponds and wetlands provides an improved calibration, as well as a more realistic representation of the catchment (Fig. 3). Soil, drainage and surface runoff The delivery of water from the soil profile to the groundwater and to the river proved to be too slow in the model parameterizations studied, as compared to observed data. Many lowland areas were drained in order to speed up the transport of water from precipitation to the river discharge, and this should be interpreted in the model. The discharge dynamics after precipitation events could be reproduced well by the implementation of agricultural drainage in the model. Figure 4 demonstrates the influence of drainage on the discharge dynamics in the Stör catchment (Dobslaff, 2005). Also in the Kielstau catchment, discharge dynamics were Copyright © 2008 IAHS Press

996

Britta Schmalz et al.

Fig. 3 Comparison of measured and modelled stream discharge with and without pond and wetland simulation, for the year 1997, Kielstau catchment.

Fig. 4 Stör catchment: influence of drainage on discharge dynamics at the “Grauel oben” gauge during March and December 1995 (* ALPHA_BF = 0.003; after Dobslaff, 2005).

increased, and the underestimation of the high discharge peaks was decreased after drainage implementation. Further attempts to improve the model performance were made as follows: Surface runoff is extremely sensitive to parameter CN2 (Singh et al., 2004) which is the SCS runoff curve number (CN2 in the HRU management file). For the Kielstau catchment the decrease in the CN2 values resulted in decreasing runoff and increasing infiltration, baseflow and recharge. Problems occur when attempting, via CN2, to increase surface runoff in the high peak periods, since runoff is modelled to grow over the whole time period. Copyright © 2008 IAHS Press

Modelling lowland hydrological processes in mesoscale river basins

997

The overestimation of surface runoff throughout the simulation led to changes of the parameter SOL_AWC, the available water capacity. This parameter revealed a relatively low sensitivity in the Kielstau model set-up, in contrast to what was found by White & Chaubey (2005). Soil thickness turned out to be a sensitive parameter for the too slow-reacting discharge dynamics. For many soil units, the thickness within the riparian areas was significantly reduced from below by a high groundwater level. The decrease in soil thickness generates more dynamic discharge behaviour after precipitation events (Dobslaff, 2005). Besides, in the sub-catchment areas with drained organic soils, the saturated hydraulic conductivity of the soil (SOL_K) has influence on the dynamics of the discharge in the summer months (Dobslaff, 2005). Model performance The modelling of the water balance with SWAT of the Stör, Treene and Kielstau catchment areas showed a good correlation between measured and modelled discharges (Table 3; Fig. 5). The discharge dynamics were reproduced in a better way. Figure 6 shows, as an example, the stepwise calibration of the SWAT model for the Kielstau The most remarkable improvement occurred after changing GW_DELAY. Table 3 Model performance at the catchment outlets. Stör calibration * Stör validation * Treene calibration † Treene validation † Kielstau calibration Kielstau validation * Dobslaff, 2005. † Dey, 2004.

Period 01.01.1992–31.12.1996 01.01.1997–31.12.2002 1994–1995 1997–2000 01.01.1990 –31.12.1999 01.01.2000– 01.05.2005

Nash-Sutcliffe index 0.76 0.70 0.89 0.86 0.71 0.63

Correlation coefficient 0.87 0.84 0.95 0.94 0.82 0.75

These model parameterizations demonstrate that catchments in lowland areas only show a good agreement between measured and modelled water discharge if the specific lowland conditions are considered and properly implemented in the model. The lowland river catchments have a high fraction of peatlands in the river valleys which have a high potential of water retention. The lakes, small ponds and landscape depressions act in the same way. In order to use these areas agriculturally, most of them are drained by pipes or ditches which have a strong influence on the water balance. Water from precipitation events in the drained catchments has a faster and more direct response than the flow through the real soil profile in unmodified lowland areas. Furthermore, because of the shallow groundwater, there is high interaction between groundwater and surface water. This results in a high groundwater dynamic and in an intensive exchange of water between these water bodies. This has to be taken into consideration in the ecohydrological model set-ups. CONCLUSIONS The Stör catchment, the Treene watershed and its sub-catchment Kielstau in northern Germany served as example regions for lowland areas. However, the results of this study are transferable to other lowland catchment areas. The dominating hydrological processes were derived from measurements and ecohydrological modelling. The influence of groundwater dynamics and storage, drainage networks, and wetlands and ponds was very pronounced. The groundwater parameters turned out to be the most influential factors for lowering the baseflow and improving the simulation of water discharge. With this process knowledge, the water balance was Copyright © 2008 IAHS Press

998

Britta Schmalz et al.

(a)

(b)

(c)

Fig. 5 Measured and modelled discharge at the catchment outlets: (a) Stör catchment (Fohrer et al. 2006); (b) Treene catchment (Dey 2004); and (c) Kielstau catchment, during calibration (different time scales).

Fig. 6 Statistical development of model efficiency and correlation during the calibration steps (Kielstau): A: base model; B: GW_DELAY = 200 d; C: GW_DELAY = 50 d; GWQMN = 1250 mm; D: GW_REVAP = 0.2; E: REVAPMN = 0.01 mm; F: ponds & wetlands; G: drainage; H: CN2 adjustments. Copyright © 2008 IAHS Press

Modelling lowland hydrological processes in mesoscale river basins

999

successfully simulated with the ecohydrological mesoscale model SWAT. During the calibration, Nash-Sutcliffe index values of 0.76, 0.89, 0.71 were obtained for the Stör, Treene and Kielstau catchments, respectively with correlation values of 0.87, 0.95 and 0.82. REFERENCES Arnold, J. G. & Allen, P. M. (1999) Automated methods for estimating baseflow and ground water recharge from streamflow records. J. Am. Water Resour. Assoc. 35(2), 411–424. Baseflow separation programme: http://www.brc. tamus.edu/swat/soft/baseflow.html. Arnold, J. G., Srinivasan, R., Muttiah, R. S. & Williams, J. R. (1998) Large area hydrologic modelling and assessment, Part I.: Model development. J. Am. Water Resour. Assoc. 34(l), 73–89. BGR (1999) Bodenübersichtskarte im Maßstab 1:200 000 (BÜK 200). Verbreitung der Bodengesellschaften, Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover, Germany. Dey, T. (2004) Räumlich differenzierte Einzugsgebietsmodellierung für den tidefreien Bereich der Treene. Diploma Thesis, Ecology Centre, Kiel University, Germany. http://www.hydrology.uni-kiel.de/lehre/abschlussarbeiten/da_tdey.pdf. Dey, T., Horn, A., Hörmann, G. & Fohrer, F. (2004) Räumlich differenzierte Einzugsgebietsmodellierung am Beispiel des tidefreien Bereichs der Treene. In: Neue Methodische Ansätze zur Modellierung der Wasser- und Stoffumsätze in Großen Einzugsgebieten (7. Workshop zur großskaligen Modellierung in der Hydrologie, München, November 2003) (ed. by R. Ludwig, D. Reichert & W. Mauser), 111–122. Kassel University Press, Kassel, Germany. Di Luzio, M., Srinivasan R., Arnold J. G. & Neitsch, S. L. (2002) Arcview Interface for SWAT2000—User’s Guide. Report TR-193, Texas Water Resources Institute, College Station, Texas. DLR (1995) Landsat TM5-Szene aus dem Jahr 1995, Auflösung 25 × 25 m. Deutsches Zentrum für Luft- und Raumfahrt Köln. Dobslaff, N. (2005) GIS-basierte Modellierung von Wasserhaushalt und Abflussbildung am Beispiel des Einzugsgebietes der oberen Stör. Diploma Thesis, Ecology Centre, Kiel University, Germany. http://www.hydrology.uni-kiel.de/lehre/abschlussarbeiten/da_dobslaff.pdf. EEA (European Environment Agency) (2000) Corine Landcover 250 m-Raster. European Environment Agency, Copenhagen, Denmark. Finnern, J. (1997) Böden und Leitbodengesellschaften des Störeinzugsgebietes in Schleswig-Holstein – Vergesellschaftung und Stoffaustragsprognose (K, Ca, Mg) mittels GIS. Schriftenreihe des Instituts für Pflanzenernährung und Bodenkunde der Universität Kiel, Nr. 37, Kiel, Germany. Fohrer, N., Dey, T., Dobslaff, N. & Hörmann, G. (2006) Räumlich differenzierte hydrologische Modellierung norddeutscher Tieflandeinzugsgebiete am Beispiel der Treene und der Stör. Ecosys. Suppl. 46, 107–120. Fohrer, N., Schmalz, B., Tavares, F. & Golon, J. (2007) Ansätze zur Integration von landwirtschaftlichen Dränagen in die Modellierung des Landschaftswasserhaushalts von mesoskaligen Tieflandeinzugsgebieten. Hydrologie & Wasserbewirtschaftung 51(4), 164–169. Fujita, T. T. (1986) Mesoscale classifications: their history and their application to forecasting. In: Mesoscale Meteorology and Forecasting (ed. by P. S. Ray), Ch. 2, 18–35. American Meteorological Society, Boston, Massachusetts, USA. Habeck, A., Krysanova, V. & Hattermann, F. (2005) Integrated analysis of water quality in a mesoscale lowland basin. Adv. Geosci. 5, 13–17. Hattermann, F. F., Krysanova, V., Habeck, A. & Bronstert, A. (2006) Integrating wetlands and riparian zones in river basin modelling. Ecol. Modell. 199 (4), 379–392. Hörmann, G., Horn, A. & Fohrer, N. (2005) The evaluation of land-use options in mesoscale catchments—prospects and limitations of eco-hydrological models. Ecol. Modell. 187, 3–14. Kieckbusch, J., Schrautzer, J. & Trepel, M. (2006) Spatial heterogeneity of water pathways in degenerated riverine peatlands. Basic and Appl. Ecol. 7, 388–397. Krause, S. & Bronstert, A. (2005) An advanced approach for catchment delineation and water balance modelling within wetlands and floodplains. Adv. Geosci. 5, 1–5. Krause, S. & Bronstert, A. (2007) The impact of groundwater–surface water interactions on the water balance of a mesoscale lowland river catchment in northeastern Germany. Hydrol. Processes 21, 169–184. Krause, S., Bronstert, A. & Zehe, E. (2007a) Groundwater–surface water interactions in a north German lowland floodplain— implications for the river discharge dynamics and riparian water balance. J. Hydrol. 347, 404–417. Krause, S., Jacobs, J. & Bronstert, A. (2007b) Modelling the impacts of land-use and drainage density on the water balance of a lowland floodplain landscape in northeast Germany. Ecol. Modell. 200(3–4), 475–492. Krysanova, V., Müller-Wohlfeil, D.-I. & Becker, A. (1998) Development and test of a spatially distributed hydrological/water quality model for mesoscale watersheds. Ecol. Modell. 106 (1–2), 261–289. LVA S-H (1995) Digitales Geländemodell DHM50 für Schleswig-Holstein. Gitterweite 50 × 50 m. Landesvermessungsamt Schleswig-Holstein, Kiel, Germany. LVA S-H (2003) Basisgewässernetz des Landes Schleswig-Holstein. ATKIS-Objektart: Gewässerlinie. Landesvermessungsamt Schleswig-Holstein, Kiel, Germany. Neitsch, S. L., Arnold, J. G., Kiniry, J. R., Srinivasan, R.& Williams, J. R. (2002a) Soil and Water Assessment Tool—User’s Manual. Version 2000. Blackland Research & Extension Center, Temple, Texas, USA. Neitsch, S. L., Arnold, J. G., Kiniry, J. R., Srinivasan, R., Williams, J. R. & King, K. W. (2002b) Soil and Water Assessment Tool—Theoretical Documentation. Version 2000. Blackland Research & Extension Center, Temple, Texas, USA. Schmalz, B., Springer, P. & Fohrer, N. (2008) Interactions between near-surface groundwater and surface water in a drained riparian wetland. In: Groundwater–Surface Water Interactions: Process Understanding, Conceptualization and Modelling (Proc.Perugia Symp., July 2007) (ed. by C. Abesser, T. Wagener & G. Nuetzmann), 21–29. IAHS Publ. 321. IAHS Press, Wallingford, UK. Singh, J., Knapp, H. V. & Demissie, M. (2004) Hydrologic modeling of the Iroquois River watershed using HSPF and SWAT. Illinois State Water Survey Contract Report. Illinois Department of Natural Resources, USA.

Copyright © 2008 IAHS Press

1000

Britta Schmalz et al.

Sophocleous, M. (2002) Interactions between groundwater and surface water: the state of the science. Hydrogeol. J. 10, 52–67. Tavares, F. (2006) Continuous, spatially distributed, stream flow and quality assessment of a lowland catchment in Northern Germany. MSc Thesis, Ecology Centre, Kiel University, Germany. http://www.hydrology.uni-kiel.de/lehre/abschlussarbeiten/msc_tavares.pdf. Venohr, M. (2000) Einträge und Abbau von Nährstoffen in Fließgewässern der oberen Stör. Diploma Thesis, Ecology Centre, Kiel University, Kiel, Germany. White, K. L. & Chaubey, L. (2005) Sensitivity analysis, calibration, and validations for a multisite and multivariable SWAT model. J. Am. Water Resour. Assoc. 41(5), 1077–1089.

Received 19 April 2007; accepted 13 August 2008

Copyright © 2008 IAHS Press