Distributed, physically based river basin models are receiving increasing importance in integrated water resources management (IWRM) in Germany and in Europe, especially after the release of the new European Water Framework Directive (WFD). Applications in mesoscale catchments require an appropriate approach to represent the spatial distribution of related catchment properties such as land use, soil physics and topography by utilizing techniques of remote sensing and GIS analyses. The challenge is to delineate scale independent homogeneous modelling entities which, on the one hand may represent the dynamics of the dominant hydrological processes and, on the other hand can be derived from spatially distributed physiographical catchment properties. This scaling problem is tackled in this regional modelling study by applying the concept of hydrological response units (HRUs). In a nested catchment approach three different modelling conceptualisations are used to describe the runoff processes: (i) the topographic stream-segment-based HRU delineation proposed by Leavesley et al. [Precipitation-Runoff-Modelling-System, User's Manual, Water Resource Investigations Report 83-4238, US Geological Survey, 1983]; (ii) the process based physiographic HRU-concept introduced by Flugel [Hydrol. Process. 9 (1995) 423] and (iii) an advanced HRU-concept adapted from (ii), which included the topographic topology of HRU-areas and the river network developed by Staudenraush [Eco Regio 8 (2000) 121]. The influence of different boundary conditions associated with changing the landuse classes, the temporal data resolution and the landuse scenarios were investigated. The mesoscale catchment of the river Ilm (A similar to 895 km(2)) in Thuringia, Germany, and the Precipitation-Runoff-Modelling-System (PRMS) were selected for this study. Simulations show that the physiographic based concept is a reliable method for modelling basin dynamics in catchments up to 200 km(2) whereas in larger catchments, where lateral processes dominate, the other concepts have advantages. (C) 2003 Elsevier Ltd. All rights reserved.