Assessment of evapotranspiration simulations in the Malše basin

Abstract

The application of the distributed hydrological model brings the benefits of assessment of the spatially distributed quantities which are hard to measure in the field over a larger area, e.g. evapotranspiration. The Malše River basin has been chosen for the evaluation of evapotranspiration simulation by the distributed hydrological model, SWIM (Soil and Water Integrated Model). The primary interest in this analysis was to assess the ability of the hydrological model to simulate the actual evapotranspiration on larger scales and to evaluate its dependence on the landscape characteristics such as the vegetation cover, soil type, and average precipitation amount during the simulation. Annual actual evapotranspiration in each hydrotope was evaluated in the simulation period of 1985–1998. Because of the lack of the data observed (evapotranspiration), the model was calibrated on the discharge time series. The credibility was quantified using Nash Sutcliffe efficiency which was more than 0.7. The main trends of the simulated actual evapotranspiration were evaluated and assessed as satisfactory. The differences in the soil types did not seem significant for the evapotranspiration variation, the monthly average values among soil types differing by ± 10% except histosol. On the other hand the differences in the land-use categories strongly influenced the amount of evapotranspiration (–30; +50%). It appears that the model SWIM overestimates the actual evapotranspiration in the spring and, on the other hand, underestimates that in the autumn according to the comparison with the only data available in the entire Climate Atlas of the Czech Republic.