Spatial and temporal dynamics of nutrients in two agricultural catchments in southeast Sweden

Abstract: Many lakes and coast areas in Sweden and all over the globe suffer from eutrophication, caused by an overload of phosphorus (P) and nitrogen (N). In Sweden these emissions from point sources, e.g. waste water treatment plants, have been reduced the last de cades and focus is now on reducing the diffuse emissions from agriculture and waste water from rural households. Since diffuse sources cover large areas and it is not possible to monitor everywhere, thus water quality modeling, of discharge and nutrient co ncentrations in rivers and water bodies is a common method to estimate the nutrient transport. In this thesis the HYPE - model was setup in two agricultural catchments, the Skenaån catchment and E23, in Östergötland County in southeast Sweden, during 2006 - 2 011 and 2009 - 2011 respectively. Several sub - catchments were parameterized, based on their landuse, soil type, distribution of household sewage treatment systems etc. The aims were to investigate if the model could show similar spatial distribution of sourc es of nutrients and the same temporal dynamic during high runoff events exemplified by snowmelt periods, compared to the observed concentrations. Much of the nutrient losses occur during these events and during a snowmelt - period in March 2010 intensive wat er sampling was taken, in order to compare it with model - simulations. Moreover, the aims were to study if calibration and increased effort in including more local information, and if a time resolution of hours instead of days, could improve the model simul ation. This was done by the use of local set ups of the HYPE - model in the two catchments. The observed results showed higher stream nutrient concentrations in areas with higher portion of arable land which spatial sources the model could catch overall. Ho wever, the model was less successful simulating temporal dynamics of P and especially the seasonality failed, where nitrate dynamics was better simulated. The discharge was in general simulated successful and was largely improved by including local measure ments of precipitation, instead of using a national gridded database. The temporal dynamics of modeled transport of nitrate was successful and indicated no needs of further local information or calibration compared to the national model setup S - HYPE, but w as less successful for total - phosphorus, phosphate and particulate phosphorus. Local calibration improved the simulation of P - transport where including more local information, like local precipitation, did not. Neither did a higher time resolution improve the temporal simulation of the mentioned P - fractions. The phosphorus modeling, both simulation of pathways and timing is very difficult, since the knowledge is still poor concerning the complex mechanisms governing the losses of phosphorus, which is confir med in this study. The processes are even more difficult in these small catchments, having no lakes, since in other catchments where lakes exists it is easier to get a decent fit between observed and modeled data. This is due to retention parameters in lak es could more easily be changed to improve the simulation, however these does not explain the reality.