Location: Southwest Watershed Research Center
Project Number: 2022-13610-012-36-R
Project Type: Reimbursable Cooperative Agreement
Start Date: Jul 1, 2018
End Date: Jun 30, 2021
1) Integrate high resolution radar rainfall data with process-based erosion models and apply the models to study erosion processes in instrumented watersheds in U.S. (Walnut Gulch Experimental Watershed) and Israel (Ramot Menashe Watersheds). 2) Assess erosion risk, i.e., erosion rates for different recurrence intervals, using a stochastic framework that links high resolution rainfall weather generators with the erosion models. This will allow to examine potential solutions for high erosion risk areas in the studied watersheds. 3) Characterize the conditions inducing high erosion rates in terms of storm properties and soil surface conditions. 4) Examine the utility of satellite-based rainfall data of relatively high resolution for erosion estimation and risk assessment.
Two erosion models will be used in this research: the Rangeland Hydrology and Erosion Model for rangelands in the USDA-ARS Walnut Gulch Experimental Watershed in the US and the Dynamic Water Erosion Prediction Project (DWEPP) model for croplands in Ramot Menashe Watersheds (RMW) in Israel. Both are process-based models incorporated into the KINEROS2 watershed and erosion model and serve as the hillslope infiltration, runoff, and erosion engine for overland flow model elements. Radar-rainfall data in Israel is available from two C-Band instruments located in the Tel Aviv area about 50–70 km from the RMW area. In the US, National Weather Service NEXRAD radar-rainfall estimates will be obtained from the Tucson, Arizona location. These radar-rainfall data sources will be used to drive the KINEROS2 model to obtain erosion estimates that will be compared to observed erosion estimates from the WGEW and the RMW watershed respectively. A High space-time Resolution rainfall Weather Generator (HiReS-WG) was developed for the Israeli study region which will also be applied to the WGEW. A Spatially Explicit Rainfall Generator developed in the US will also be employed. Weather generator rainfall outputs will be used as rainfall input to the RHEM (for WGEW) and DWEPP (for RMW) models. The model runs and the set of model parameters include both present state of the watersheds and hypothetical states representing relevant scenarios (such as changes in land cover or conservation practices). Spatially distributed erosion and sediment from all hillslope and channel modeling elements in the studied watersheds will be computed for the ensemble data. These outputs will be used for two types of analyses: 1) For each hillslope and channel modeling element we will analyze the series of storm erosion rates and apply extreme value analysis methods to derive erosion risk maps, i.e., erosion rates for different recurrence intervals. The multiple ensembles will allow computing uncertainty range around these estimates. 2) For areas in the studied watersheds where high erosion risk is estimated we will use the model to suggest realistic changes that can be made (e.g., changes in land cover or soil conservation practice) to reduce that risk. 3) From the simulated erosion rates for all ensemble data we will examine the events where highest erosion rates were obtained, and compare it to the other cases, to better understand what characterizes these extreme cases. In this analysis, both storm properties (such as rain intensity, areal coverage, storm speed, total amount) and surface properties (e.g., beginning vs. end of growing season) will be considered. Whether or not satellite-based rainfall products can be used for erosion modeling is still not known, and in this proposal we intend to investigate this issue. We focus on those satellite-products that have the highest resolution. This includes the CMORPH 30-min, 8-km data set that already was found suitable for extreme rainfall analysis and the 30-min, 10-km IMERG product from the newly available GPM satellite. The analysis will be performed at the event and watershed scales.