Modelling the Impact from Combined Changes on Riverine Nutrients Across Sweden

Swedish authorities currently explore how the next generation of environmental objectives for 2050 should be formulated taking into account climate change, changes in land use and deposition of pollutants. To provide scientific support, we have modelled a suite of combined scenarios to predict the non-linear dynamics of riverine nutrients, using a national multi-basin model system for Sweden, called ‘S-HYPE’. The model is process-based and the set-up covers more than 450 000 km² and produces daily values of nutrient concentration and water discharge in 37 000 catchments.

Net nutrient fluxes from large catchments in modified landscapes are influenced by land use characteristics but also by landscape composition, e.g. occurrence of lakes on the main stream network or connectivity of nutrient source areas to the stream network. This is accounted for in the model concept. Forestry management changes were investigated for potential conflicts with national environmental protection goals.

The scenarios included three forestry management scenarios, three down-scaled climate change projections based on Hadley and Echam GCMs, and one nitrogen deposition scenario. The aim was to evaluate basin-wide impacts on TN and TP fluxes as well as feedbacks between the respective scenarios. Combined scenarios were thus modelled.

Results indicate a low impact of changes in forestry management and atmospheric deposition for both concentrations and loads; whereas changes in water balance dynamics through changes in climate forcing can lead to regionally strong changes in nutrient concentrations and fluxes. The impact assessment through a chain of multiple scenarios and the various numerical models involved, accumulate a considerable uncertainty. While a formal assessment of overall uncertainties is difficult for the chosen approach, we discuss the modeled impacts in the light of expectable uncertainties resulting from scenario assumption choices and limitations in the model framework.