Agricultural management practices are among the major drivers of agricultural nitrogen (N) loss. Legislation and management incentives for measures to mitigate N loss should eventually be carried out at the individual farm level. Consequently, an appropriate scale to simulate N loss from a scientific perspective should be at the farm scale. A data set of more than 4000 agricultural fields with combinations of climate, soils and agricultural management which overall describes the variations found in the Baltic Sea drainage basin was constructed. The soil–vegetation–atmosphere model Daisy (Hansen et al. 2012) was used to simulate N loss from the root zone of all agricultural fields in the data set. From the data set of Daisy simulations, we identified the most important drivers for N loss by multiple regression statistics and developed a statistical N loss model. By applying this model to a basin-wide data set on climate, soils and agricultural management at a 10 × 10 km scale, we were able to calculate root-zone N losses from the entire Baltic Sea drainage basin and identify N loss hot spots in a consistent way and at a level of detail not hitherto seen for this area. Further, the root-zone N loss model was coupled to estimates of nitrogen retention in catchments separated into retention in groundwater and retention in surface waters allowing calculation of the coastal N loading.