Modeling riverine nutrient transport to the Baltic Sea: a large-scale approach.

We developed for the first time a catchment model simulating simultaneously the nutrient land-sea fluxes from all 105 major watersheds within the Baltic Sea drainage area. A consistent modeling approach to all these major watersheds, i.e., a consistent handling of water fluxes (hydrological simulations) and loading functions (emission data), will facilitate a comparison of riverine nutrient transport between Baltic Sea subbasins that differ substantially. Hot spots of riverine emissions, such as from the rivers Vistula, Oder, and Daugava or from the Danish coast, can be easily demonstrated and the comparison between these hot spots, and the relatively unperturbed rivers in the northern catchments show decisionmakers where remedial actions are most effective to improve the environmental state of the Baltic Sea, and, secondly, what percentage reduction of riverine nutrient loads is possible. The relative difference between measured and simulated fluxes during the validation period was generally small. The cumulative deviation (i.e., relative bias) [Sigma(Simulated - Measured)/Sigma Measured x 100 (%)] from monitored water and nutrient fluxes amounted to +8.2% for runoff, to -2.4% for dissolved inorganic nitrogen, to +5.1% for total nitrogen, to +13% for dissolved inorganic phosphorus and to +19% for total phosphorus. Moreover, the model suggests that point sources for total phosphorus compiled by existing pollution load compilations are underestimated because of inconsistencies in calculating effluent loads from municipalities.