Simulation of nutrient reduction scenarios with three different models. Does the selection of model affect the recommended set of measures?

Abstract This article compares and discusses the results of applying three different models in order to quantify the phosphorus loads in Kapos catchment, Hungary. The models range from the very simple TEOTIL operating on statistical data and export coefficients, via the semi‐complex INCA‐P to the highly complex and process‐based SWAT. All models have been calibrated with use of basically the same input and calibration data. The models have, to the extent they are capable, been used further to simulate a set of theoretical pollution reduction scenarios. The calibration results showed that all the models are basically able to simulate historical loads of TotP on a catchment level, except maybe those years with particularly high or low loads (INCA‐P in 1999 and TEOTIL in 2002). The scenario analysis showed that the models are to a varying degree capable of simulating the simplified scenarios, where TEOTIL seems to a limited extent suitable for such analysis. This study demonstrated that the selection of model can influence the output of the scenario analysis significantly, and the selection of tool to be applied should therefore be made with great care. The most obvious example is the comparison of the simulation of the reduced runoff from area sources (scenario E) with the simulation of reduced discharges from point sources (scenario F). Based on the simulation results, TEOTIL considers reduction in nutrient runoff from areas to be the most efficient measures, while INCA‐P and SWAT results would suggest focusing on the reduction of point‐source discharges.