Soil erosion and phosphorus losses under variable land use as simulated by the INCA-P model

We describe the calibration and application of the Integrated Catchment Phosphorus model (INCA-P) to determine the key factors controlling run-off, sediment and phosphorus (P) losses from a small catchment in southeast Norway. The INCA-P model was parameterized for five land use categories: forest, grassland, urban areas and agricultural areas with and without autumn tillage. Monitoring data from the Norwegian Agricultural Environmental Monitoring Program were used for model calibration at the outlet of the catchment (Skuterud). Available measurements and local expert knowledge were used as reference data on sediment and P loss specific to land uses. We successfully applied the INCA-P model to Nordic conditions: the NashSutcliffe (N-S) statistic, calculated for the calibration period, was 0.82, 0.71 and 0.67 for run-off, sediment and P losses, respectively. For the validation period, the sediment and P losses were slightly overestimated. Scatter plots of observed versus modelled values of run-off, sediment and total P losses indicated that in some cases, the INCA-P model did not adequately represent extreme situations. It is likely that the modelled system behaves in a different way under moderate and large rainfall events, and the model does not represent these. We applied the model to five land use change scenarios to evaluate the possibilities for reducing particle-bound P loads at the catchment outlet. Our results indicate that land use change can lead to more significant reduction in particle and P losses than changes in agricultural practices. The most favourable scenario for freshwater ecosystems would be afforestation: changing half of the agricultural areas to forest would reduce sediment and P losses by 44 and 40%, respectively. Changes in agricultural practice could also improve the situation, especially by reducing areas with autumn tillage to a minimum.