“…When model performance is poor in a particular hydrological aspect, model modifications are based on hydrological expert knowledge that, whilst subjective, is the only currently available way to adjust the model. The Kennet catchment in southern England is chosen as a complex case study that represents a number of the modelling challenges; however, the methodology and the results are of interest beyond this study due to the similarities across the hydrological modules of different third-generation LSMs, and also the broad importance of chalk aquifers and deep-groundwater systems (Brouyere et al, 2004;Downing et al, 1993;Kloppmann et al, 1998;Pinault et al, 2005;Dahan et al, 1998Dahan et al, , 1999Nativ and Nissim, 1992;Nativ et al, 1995).…”