Browsing by Author "Ruohonen, Viljami"

Global climate change is predicted to have a major impact on northern temperate and arctic zones climates. Rainfall and temperatures are likely to increase in the future decades. Increasing erosion is one of the major threats linked to these phenomena. In Finland specifically, winter months are expected to face the most radical change in climate. This master's thesis aims to answer the following questions: how much does erosion increase in changing climate, what are the factors that either increase or control erosion rates, and how well does the physically based SWAT- model behave in the settings of Southern Finland. Two gauging stations inside the study area were used to calibrate and validate the physical SWAT- model (Härkälänjoki and the lower-part of Lepsämänjoki). Total of 16 parameters controlling runoff were used. NS & R2 as well as P- & R-factors were chosen to quantify and measure model behavior. In the calibration period, the R2 & NS values for Härkälänjoki were 0,29 and -0,05, and for Lepsämänjoki 0,40 and 0,34. In the validation period, the R2 & NS values for Härkälänjoki were 0,30 and -0,57, and for Lepsämänjoki 0,54 and 0,21. In the calibration period, the P & R-values were 0,33 and 0,24 for Härkälänjoki and 0,32 and 0,29 for Lepsämänjoki. In the validation period, the P & R values were 0,42 and 0,57 for Härkälänjoki and 0,50 and 0,63 for Lepsämänjoki. The results of R2 and NS and the other criteria indicate that Lepsämänjoki responded better to model calibration and validation, being satisfactory. Values for Härkälänjoki were generally unsatisfactory. The RCP8.5 scenario used in this thesis increases erosion by approx. 94% compared to the validation period. However, spatial variability is considerable. Average annual erosion in the climate change scenario was 1,22 t/ha. Winter months contribute most to the total annual erosion. Areas covered by evergreen forests and located near the edges of the study area experience least erosion, while areas near the downstream of the river experience the most erosion. SWAT model behaved reasonably well, and worked well with high-quality GIS data. However, the model still needs some further adjustment and development in order to have more consistent and user-friendly interface.