To use a catchment-scale model to delineate areas with high and low denitrification capacities in the saturated zone of a catchment, the model must have an accurate spatial description of both general large-scale flow patterns on catchment scale and small-scale flow patterns locally within the catchment. In this study, a coupled groundwater-surface water model based on the MIKE SHE code was developed for the 4.7 km2 Lillebæk catchment in Denmark, where tile drain flow is a major contributor to the stream discharge. The catchment model was calibrated in several steps by incrementally including the observation data into the calibration to see the effect on model performance of including diverse data types, especially tile drain discharge. For the Lillebæk catchment, measurements of hydraulic head, daily stream discharge, and daily tile drain discharge from five small (1–4 ha) drainage areas exist. The results showed that including tile drain data in the calibration of the catchment model improved its general performance for hydraulic heads and stream discharges. However, the model failed to correctly describe the local-scale dynamics of the tile drain discharges, and, furthermore, including the drain data in the calibration did not improve the small-scale spatial dynamics. This is mainly believed to be caused by the model's inadequate simulation of local spatial dynamics in hydraulic heads, which we argue is likely due to the lack of sufficient heterogeneity in the geological model.