Soil Erosion under Spatially Heterogeneous Field Conditions

Abstract

Soil erosion is a significant problem for most of the agroecosystems worldwide, as it is one of the primary causes of soil degradation caused by a loss of the topsoil layer. Most of the existing soil erosion models provide insight into the soil erosion drivers. However, the distributed and small-scale nature of erosion processes makes it difficult to quantify the severity of the erosion problem at the field scale under spatially heterogeneous soil and terrain conditions. A thorough analysis for quantitatively assessing the spatial and temporal variability of soil erosion patterns under highly heterogeneous field conditions is of great importance. This thesis presents a systematic analysis of (i) the limitations and applicability of existing modeling approaches (ii) the within-field variability of a selected field site in East Germany where the effects and interactions between soil characteristics, topography, rainfall intensity, and soil cover on soil erosion and surface runoff, carbon and nitrogen losses have been observed and (iii) the insights into the uncertainties involved in the simulation of sub-field scale runoff and soil erosion processes due to the model structure and parameter estimation. With respect to the first objective, 51 Soil erosion models based on their representation of the soil erosion process and their suitability for assessing soil erosion for more complex field designs were investigated. The results showed that a particular shortcoming of most of the existing field scale models is their one-dimensional nature. Moreover, only a few models are suitable for dynamic soil erosion assessments at the field-scale with heterogeneous soil and slope conditions. To date, there are no field-scale dynamic models available considering complex agricultural systems for the dynamic simulation of soil erosion. For the second objective, on a selected field with about 6 ha, which covers several landscape elements, intensive measurements were conducted to systematically analyze the influence of heterogeneous field conditions and rainfall intensities along with their complex interactions on sediment losses and surface runoff. Results revealed that (i) Rainfall intensity with 40.6% showed the highest contribution to sediment yield followed by Slope steepness (23.8%), Vegetation cover (17.74%), Silt and organic matter content (14.77%), and Depth to loamy layer (3.17%), indicating a strong rainfall intensity-erosion relationship. Based on the results, it was suggested that the highest sediment yield should be obtained with the factor levels combination 16-18 % Silt and organic matter, 0% vegetation cover, 3-5 % slope steepness and a rainfall intensity between 3.4-4 mm/min. Finally, the accuracy of soil erosion models with different model structures in the same field with heterogeneous soil and terrain conditions was tested. For this purpose, two soil erosion modelling approaches (Freebairn, Rose) were coupled with the Lintul5 and SlimWat models within the SIMPLACE framework. The results indicated that the simulation of water erosion with the both the Freebairn and the Rose approach were influenced by the performance of the runoff and crop growth sub-models. Sensitivity analysis revealed that runoff and slope angle were most important components in both Freebairn and Rose models in predicting sediment yield followed by the parameter “soil erodibility” in Freebairn and the parameter “efficiency of entrainment” in the Rose approach respectively. The Freebairn model had slightly higher accuracy (RMSE=0.69 t ha-1 d-1) of sediment yield predictions than the Rose model (RMSE=0.83 t ha-1 d-1). This thesis highlights the need of an improved understanding of complex interactions among multiple factors influencing the soil erosion process. The coupled model developed within this thesis marks a promising approach that should be tested for a wider range of crops, soils, and climate conditions to be applied for modeling studies at larger spatial scales.