Skip Ribbon Commands Skip to main content
Sign In

Centre for Development and Environment (CDE)



Analyse der Bodenerosion mit der AVErosion-Extension für ArcView

Chisholm, Michael

Publisher: CDE, University of Bern

The aim of this thesis is to describe the application and evaluation of the AVErosion extension for ArcView 3.x. AVErosion is a GIS-Tool for assessing soil erosion that is still largely unused in Switzerland. The tool was developed at the University of Tübingen in Germany based on erosion prediction analyses by SCHÄUBLE (1999, 2005). AVErosion is available for downloading from the Internet free of charge ( AVErosion is based upon the Universal Soil Loss Equation (USLE) by WISCHMEIER & SMITH (1978). As an empirical model, USLE quantifies soil erosion based on the product of the erosionrelevant factors of precipitation (R), soil type (K), crop system (C), topography (L & S) and soil erosion management practices (P). AVErosion offers two versions of the model: the traditional USLE formula as well as the modified USLE (MUSLE87) by HENSEL & BORK (1988). The modification concerns the calculation of topographic factors. Whereas USLE only predicts soil erosion on a single slope, MUSLE87 is designed to make predictions for catchments. Both models require the same input data, which are the factors R, K, and C, a digital elevation model (DEM) for the determination of the factors L, S and P, as well as a layer defining the area or plots to be calculated. The output is a map indicating the long-time average risk of soil loss in tonnes per hectare and year [t*ha-1*a-1]. For each calculation AVErosion produces two outputs in the form of raster grids: one map with the average soil loss in grid resolution and another map showing the average risk of soil erosion for each agricultural plot. In this thesis the tool was applied to three research areas in Switzerland: Estavayer in the Canton of Fribourg, and Oberaargau and Frienisberg in the Canton of Berne. The DEM is required in order to calculate the topography-relevant factors L, S and P. Three DEMs were available in three different resolutions: the DHM100 (BFS) used by FRIEDLI (2006) with a resolution of 100x100m, the DHM25 (swisstopo) with a resolution of 25x25m and finally the DTM-AV (swisstopo) with an ultra-high resolution of 2x2m. Aside from the difference in quality, the DEM used also dictates the resolution of the final result. The R, K, and C factors used in this thesis have been taken from FRIEDLI (2006), who created a soil erosion map for Switzerland in a hectare resolution. Despite the fact that the factors from FRIEDLI (2006) are only available in a resolution of 100x100 metres, it makes sense to use them in all calculations made in this thesis since the erosion predictions made here can be directly compared to the previous calculation by FRIEDLI (2006). Moreover, by retaining these factors, we can evaluate the influence of the DEM on the soil erosion model. Implementing USLE/MUSLE87 in different resolutions shows the major influence that topographic factors have on the result of the calculations. For the lower resolutions of 100x100m and 25x25m, both USLE and MUSLE87 predict a similar level of soil erosion. Conversely, the disparities at the 2x2m level are very clear. At the highest resolution, MUSLE87 predicts approximately twice the amount of soil erosion as USLE. USLE predicts greater soil erosion on the 25m scale than on the 2m scale. However, the higher the resolution, the higher the rate of soil erosion predicted by MUSLE87, with around twice the amount of soil erosion predicted on the 2m-scale than on the 100m-scale. Modelled using the DTM-AV, MUSLE87 depicts areas of high soil erosion risk with great accuracy when compared with photographs taken on the spot. This, however, applies only to the relative risk. At the time of publication there was insufficient field survey data available to verify the absolute erosion rates predicted by AVErosion. Compared to USLE calculations by FRIEDLI (2006), AVErosion predicts more severe soil erosion. For the Estavayer research area AVErosion predicts an average soil loss of 3.6 t*ha-1*a-1, compared to FRIEDLI’s (2006) predicted 2 t*ha-1*a-1. This is largely due to the calculation of slope length. Whereas FRIEDLI (2006) used a standard slope length of 100 meters, AVErosion obtains slope lengths from the DEM. This results in extremely long slopes, especially at low resolutions such as the DHM100 (BFS). Predictions of average soil loss per plot exist for the Frienisberg area, where PRASUHN & GRÜNIG (2001) have long been conducting active research. A comparison of their results with a high-resolution DTM-AV plot calculation using MUSLE87 shows a better concurrence of calculation results. The median for soil loss predicted by PRASUHN & GRÜNIG (2001) amounted to 2.45 t*ha-1*a-1, while AVErosion modelling calculated a median of 2.96 t*ha-1*a-1. The conclusion is that AVErosion is a well-structured and practical tool for efficiently predicting long-term soil loss rates. The different resolutions of DEM show the major impact that topography has on erosion. Moreover, the high-resolution DTM-AV has demonstrated its potential for application in future geographical studies. The erosion maps drawn up as part of this thesis can help to identify and analyse plots at risk from soil erosion. As mentioned, the absolute risk of soil erosion was not determined in this thesis. So far not taken into consideration is the actual land use factor C. In a next step the land use factors could be taken into consideration. Mapping projects determining C factors (among others) are in progress and their results and implementation into USLE/MUSLE87-Modelling will help allow comparing modelled results with the erosion mapping.
University of Bern | Centre for Development and Environment (CDE) | Hallerstrasse 10 | CH-3012 Bern | Tel: +41 31 631 88 22
 © University of Bern | Imprint