Submitted to: International Conference on Aeolian Research
Publication Type: Proceedings
Publication Acceptance Date: January 6, 2002
Publication Date: January 7, 2002
Citation: Funk, R., Skidmore, E.L., Hagen, L.J. Comparison of wind erosion measurements in Germany with simulated soil losses by WEPS. Proceedings of ICAR5/GCTE-SEN Joint Conference. 2002. v. 02-2 ix. p. 235-238.
Interpretive Summary: The first comparison between measured and simulated soil losses by WEPS in Germany shows satisfying results. This includes the total soil loss for an event, the spatial variations on the field and the temporal changes in transport capacity. The estimation of all missing parameters was handled very carefully with respect to all available information, to reduce the uncertainty and to minimize subjectivity.
The Wind Erosion Prediction System (WEPS) s a process-based model that simulates daily weather and field conditions along with hourly wind speeds and erosion. Its physical basis should allow model application to regions outside the United States for which it was originally developed. The objective of this study was to compare results from wind erosion measurements with simulated soil losses as a first example of WEPS use in Germany. Another purpose was to introduce methods and techniques for quantifying wind erosion in preparation for model comparison studies to be undertaken within the Global Change and Terestrial Ecosystems Soil Erosion Network. The field data were obtained from a small (2.5 ha) square erosion plot, where all erosion events during a three-year experiment were recorded. Two sediment sampler designs were used to measure the horizontal soil transport with high temporal and spatial resolution. Comparisons between measured and simulated erosion were based on signle erosion events and outputs from the rosion submodel of WEPS. Because some input parameters had to be estimated, multiple runs of the model were performed with varying roughness parameters to set the simulated threshold wind speed equal to that measured during the first erosion event after a tillage operation. This initial data set was used as the basis for the gradually changing roughness, crust cover and the fraction of erodible material on crust for all following erosion events depending on rainfall or erosion. Thus, the accuracy of the simulation depends much more on the relation between the erosion events than on a good agreement for one single event. the results showed excellent agreement between measurements andsimulation with R square greater than 0.9, which is mainly attributed to the good agreement with the four largest erosion events where transport exceeded 100 kg per meter width. Excluding these events, the R square was reduced to about 0.6 for all other erosion events. Spatial and temporal variability of the soil transport were also reasonably simulated.