The quality of water in rivers depends on numerous hydrological and anthropogenic factors. The Mūša catchment, belonging to the Lielupė river basin district (RBD) in the northern part of Lithuania, was taken for water quality investigation. In this catchment, 63% of the territory is under arable land.

The conceptual FYRIS model was chosen to identify the impact of the sources of pollution with total nitrogen (N) in the Mūša river. The modelling encompasses the 1997–2006 period.

After calibration, the model efficiency coefficient was E = 0.46, i. e. fairly good, and the correlation coefficient was r = 0.69.

While modelling the variation of nitrogen concentrations during the study period (1997–2006), the results did not correspond to the monitored concentrations equally well in all subcatchments. One of the main reasons for the disagreement between the model and monitoring data can be a vast dispersal of the monitoring data. The model represents the monitoring results in the Mūša, Daugyvenė and Tatula rivers quite well. There is a strong linear correlation between the model and the monitoring results. The determination coefficient of the regression equation R² varies from 0.48 to 0.56.

Analysis of the total nitrogen load to the Mūša catchment from different pollution sources has shown that about 87% of it comes from arable land, 10% enters from waste water treatment plants (WWTP), households and urban territories, and only 3% of all nitrogen within the catchment comes from wooded territories and pastures.

Keywords: water pollution, nitrogen, load, catchment, model