Tillage effects on soil erosion on agricultural lands in semi-arid Turkey

The thesis evaluates the effects of tillage on soil erosion on agricultural lands in semi-arid Turkey. Data collection has been undertaken in two-winter seasons, 1997-8 and 1998-9. Data collection was carried out under natural rainfall conditions. Land left bare and untilled during the winter season, on which the soil surface was compacted, has a high runoff potential. Tilled surfaces contribute less runoff because of high depressional storage induced by the surface roughness following tillage. After two years of tests, runoff values show that across-slope tillage is able to decrease the surface runoff to around 60% of that on down-slope tilled land and around 20% of that on untilled land. The results of this study show that overall, runoff is largely controlled by treatment. Soil loss results mirror those from runoff from the three treatments, but the differences are not as great. Tillage is more effective in reducing runoff than it is in reducing soil loss. Furthermore, across-slope tillage when compared to down-slope tillage reduces runoff to a greater extent that it does soil loss. Measurements of splash detachment (the principal source of sediment detachment in interrill areas) shows that rates of detachment do not vary among the three treatments to the same extent as does soil loss.;During the 1998-9 field season rills formed on the most sand-rich soil of the six study sites (site 2). Except for rill cross section size on the across-slope tilled plot, rill properties do not conform well to predictions of previous models. Further work is needed to improve the prediction of rill properties.;The final part of this thesis is concerned with the application of an erosion model to predict the erosion from the experimental plots and to compare the predictions of this model with the field data. Runoff and soil loss were tested using the EUROSEM model and from the results of these tests an evaluation of important parameters in the model was made. The results show that the EUROSEM model performs poorly in predicting both runoff and soil loss from the plots.