Human impact, particularly land cover changes (e.g., agriculture, construction) increase erosion and sediment loading into streams. Benthic species are negatively affected by silt deposition that coats and embeds stream substrate. Given that riparian buffers are effective sediment filters, riparian restoration is increasingly implemented by conservation groups to protect stream habitats. Limited funding and a multitude of impaired streams warrant the need for cost-effective prioritization of potential restoration actions. We created a decision-support framework for conservation agencies and aquatic resource managers to prioritize riparian restoration efforts. Our framework integrates GIS data and field surveys into a statistical model to predict instream silt from estimates of upland soil loss and riparian filtration capacity. We focus specifically on prioritizing sites in upper sections of the Roanoke and Nottoway river basins (Virginia, US) based on observed records of Roanoke logperch (Percina rex), an imperiled sediment-sensitive species. Our statistical approach examines soil characteristics, land cover, precipitation, topography, and annual soil loss estimates from the empirically derived Revised Universal Soil Loss Equation, combined with land cover-based riparian filtration capacity as potential stream habitat predictors. We found riparian filtration capacity to be a significant predictor of silt cover, while precipitation was a significant predictor of embeddedness. Spatial scale was also a factor, in that spatial variance in silt cover and embeddedness was more accurately predicted at smaller spatial extents. Ultimately, our model can be used as a prioritization tool for mitigating high siltation areas, or for protecting low soil erosion areas.
Keywords: Decision support systems; RUSLE; Riparian filtration; Sediment model; Soil erosion.