Semi-distributed hydrological and water quality models are increasingly used as innovative and scientific-based management tools. However, their application is usually restricted to the gauging stations where they are originally calibrated, limiting their spatial capability. In this study, the semi-distributed hydrological water quality model HYPE (HYdrological Predictions for the Environment) was tested spatially to represent nitrate-N (NO3-N) and total phosphorus (TP) concentrations and loads of the nested and heterogeneous Selke catchment (463 km²) in central Germany. First, an automatic calibration procedure and uncertainty analysis were conducted using the DiffeRential Evolution Adaptive Metropolis (DREAM) tool to simulate discharge, NO3-N and TP concentrations. A multi-site and multi-objective calibration approach was applied using three main gauging stations, covering the most important hydro-meteorological and physiographical characteristics of the whole catchment. Second, the model’s capability was tested to represent further internal stations, which were not initially considered for calibration. Results showed that discharge was well represented by the model at all three main stations during both calibration (1994-1998) and validation (1999-2014) periods with lowest Nash-Sutcliffe Efficiency (NSE) of 0.71 and maximum Percentage BIAS (PBIAS) of 18.0%. The model was able to reproduce the seasonal dynamics of NO3-N and TP concentrations with low predictive uncertainty at the three main stations, reflected by PBIAS values in the ranges from -16.1% to 6.4% and from -20.0% to 11.5% for NO3-N and TP load simulations, respectively. At internal stations, the model could represent reasonably well the seasonal variation of nutrient concentrations with PBIAS values in the ranges from -9.0% to 14.2% for NO3-N and from -25.3% to 34.3% for TP concentration simulations. Overall, results suggested that the spatial validation of a nutrient transport model can be better ensured when a multi-site and multi-objective calibration approach using archetypical gauging stations is implemented. Further, results revealed that the delineation of sub-catchments should put more focus on hydro-meteorological conditions than on land-use features.