Forest Road/Stream Crossings Erosion Vulnerability Assessment using Geospatial Hydrologic Models under Extreme Precipitation Events

Climate-change induced extreme precipitation events have major consequences on high-gradient forest road/stream crossings and drainage structures (culverts/bridges) resulting in siltation and scouring. The goal of this study is to develop geospatial technology- based hydrologic models incorporating design rainfall intensities to identify erosion hazards and vulnerability risks to these structures. The study has been completed in three watersheds at USDA-Forest Service (FS) long-term experimental forests in: i) a high gradient Coweeta Hydrology Laboratory, NC, ii) a low-gradient Santee Experimental Forest, SC, and iii) a medium gradient Alum Creek Experimental Forest, AR. Two automated hydrologic geospatial models were developed in this study in ArcGIS ModelBuilder platform: i) Streambank Erosion Spatial Vulnerability Assessment (SBEVA) model and ii) Automated Revised Universal Soil Loss Equation (RUSLE) model for potential erosion and stream bank vulnerability estimation. The SBEVA model was developed with landuse classified using OBIA on 1m NAIP imageries, 3m DEM, and 10m gSSURGO raster based soil characteristics, and design flood discharges calculated using the latest NOAA-PFDS (Precipitation Frequency Data Server) based 100-yr interval 24-hr duration rainfall obtained using partial duration series storm data. Vulnerability weighted-probability scale developed through Delphi method determined stream crossings vulnerability to scouring. RUSLE model was developed to estimate pixel based erosion amount using newly computed R-factor from the latest NOAA-PFDS I₃₀ (100-yr, 30-minutes rain intensity) data. K, L, and S-factors were determined from gSSURGO. C and P-factor rasters were created with NAIP-classified data and latest field information. Zonal Statistics tool quantified total eroded soil exported from individual ArcHydro created watersheds to each crossing and determined a scale of clogging vulnerability. The combined vulnerability estimates from both the SBEVA and RUSLE hydrology models established the most vulnerable locations and the structures, which were groundtruthed. As the RUSLE2 model does not assess streambank erosion, SBEVA model inclusion help develop a better decision support tool for the USDA-FS and other agencies for safeguarding these forest structures.