Assessment of soil erosion, sediment yield and basin specific controlling factors using RUSLE-SDR and PLSR approach in Konar river basin, India

Abstract

The study comprehensively assessed the effects of river basin parameters (basin morphology, drainage network, topography, climate, land use land cover (LULC) composition & pattern, and soil properties) on soil erosion (SE) and specific sediment yield (SSY) in the tropical river basin of Konar, India. The revised universal soil loss equation (RUSLE) along with the sediment delivery ratio (SDR) model has been used to quantify the SE and SSY. As the basin parameters are highly codependent with each other, multivariate partial least square regression (PLSR) with variable importance projection (VIP) statistics has been applied to 32 basin parameters to elucidate the linkage with SE and SSY. Multi-spectral LANDSAT-8 image, topographical sheets (1:50,000), gridded (0.25° × 0.25°) daily rainfall data series, laboratory estimates of soil parameters along with other ancillary data have been employed in this regard. Results revealed that wastelands encouraged moderate to very high (5–40.0 t ha−1 yr−1) annual rate of SE compare to other LULC classes. The PLSR studies revealed that the mean slope gradient (S), basin relief (HD), sediment delivery ratio (SDR), patch density (PD), edge density (ED), hypsometric integral (HI), topographic wetness index (TWI) were very strongly influential (b = 0.1–0.16) in SE. Likewise, SDR, mean annual precipitation (P), rainfall-runoff erosivity (R factor) PD, S, and HD were highly (b=0.12–0.22) important variables (VIP = 1.2–2.3) for SSY. Generated spatial information may help in understanding the SE and SSY dynamics in the basin and the knowledge can be used for soil erosion conservation measures in and around the river basins in India.