The Daily Erosion Project - Adding Channels and Forests

The Daily Erosion Project (DEP) began about 20 years ago as a daily sheet and rill erosion estimator for the state of Iowa, USA by using weather data from US National Weather Service NEXRAD radar and automated weather stations as well as USDA National Resource Inventory (NRI) survey data accessed at the township level and Soil Survey Geographic (SSURGO) soil properties data. In 2014 we revised our framework to use remote sensing methods to gather all inputs at the field level. These efforts have led to the development of the Agricultural Conservation Planning Framework, https://acpf4watersheds.org, a spatial database and geoprocessing toolbox that allows one to generate candidate locations for site specific conservation practices. The addition of automated Digital Elevation Model (DEM) enforcement techniques allowed development of a new sampling strategy to estimate sheet and rill erosion on hundreds of thousands of agricultural flowpaths across the Midwestern United States; https://dailyerosion.org. This is done by using the US 12-digit Hydrologic Unit Code (HUC) basin numbering system in conjunction with an even finer system of delineation to define approximately 100 ha DEP subcatchments. One agricultural flowpath is randomly selected and modeled in each subcatchment and the mean of the HUC12 is the estimated erosion. However, DEP estimates still do not include ephemeral gully or channel erosion processes and only occur on agricultural lands, thus DEP still lacks simulation of many erosion processes. Our objective in this project is to develop methods to add sheet and rill erosion estimates on forested lands and channel erosion estimates on agricultural lands to DEP. Considered alone, the addition of sheet and rill erosion forest estimates is somewhat straightforward as many inputs remain the same. However, the basic land management unit used by the DEP processing pipeline is the agricultural field, so expanding to include forested areas required developing new algorithms. To achieve this, we started with the LANDFIRE canopy cover database and street map datasets from the US Forest Service (USFS) and Open Streetmaps, respectively. LANDFIRE specifies the amount of canopy cover on forested lands throughout the United States, enabling selection of appropriate forest management for each 30m pixel from USFS files developed for each 10% canopy cover increment. Additionally, forest erosion is highly sensitive to channelization due to road networks, so hillslope erosion estimates are truncated at roads to keep the modeled areas constrained to places where sheet and rill erosion processes dominate. Last, modeled hillslopes are considered either agricultural or forested, but not both (agricultural hillslopes truncate at forest and forested hillslopes truncate at agricultural fields) to maintain estimates for each part separately. Adding channel erosion estimates to DEP is also necessary to adequately estimate overall erosion but requires alteration of current DEP practices to use the watershed version of WEPP (rather than hillslope). In furtherance of this goal, new algorithms are being developed to define and parameterize basins inside DEP subcatchments, develop flow networks for the basins, parameterize flow network channel geometry from QL2 LiDAR and to map land cover and management in channels as most datasets do not resolve vegetated waterways. Basin and flow network delineation can be accomplished by setting critical grid order thresholds similar to the current DEP process for hillslope termination. Model results for forest erosion are consistent with expectations but extensive field validation presents difficulties. Model results can be combined with the grid order threshold process to identify areas of potential channel erosion and the relative likelihood that erosion will occur in a given channel. The addition of these estimates to DEP helps further soil degradation process modeling and will enable future analyses into erosion forms and dynamics across different temporal and geological scales.