Application of the SWAT Model for a Tile-Drained Lowland Catchment in North-Eastern Germany on Subbasin Scale

Tile drainage is a widespread practice in agriculturally dominated lowlands with naturally high groundwater tables. A realistic estimation of the stream flow composition including tile drainage is an essential precondition for identifying major flow sources of nutrients. In this study, the Soil Water Assessment Tool (SWAT) was applied to the partially tile-drained Warnow catchment in north-eastern Germany to evaluate the effect of tile drainage systems on stream flow composition on a subbasin scale. In addition, model performance was tested after excluding tile drainages from the calibrated model setup. A sensitivity analysis revealed the highest sensitivities for parameters concerning evapotranspiration, soil characteristics, and groundwater flow, with a large variability in sensitivity ranks among the subbasins. Nash-Suttcliffe-Efficiencies (NSE) varied strongly among the subbasins for the tile-drained model setup ranging from 0.22 to 0.81 for the calibration and from −0.81 to 0.66 for the validation period. The percentage of tile flow varied between 0.3 and 31.9 %, and reflected statistically significantly (p > 0.05) the spatial extent of tile-drained areas within the subbasins. Excluding tile drainages from the model setup led to a strong decrease in model quality and to a changed stream flow constitution dominated by groundwater. The results of our study indicate that the SWAT model realistically represented the actual fractions of tile flow on discharge on the subbasin scale within the Warnow catchment. Therefore, we conclude that the incorporation of tile drainage systems is essential to calculate flow components accurately. Copyright Springer Science+Business Media Dordrecht 2013