Influence of determinant factors towards soil erosion using ordinary least squared regression in GIS domain

• Published in: Applied geomatics 2022-03, Vol.14 (1), p.57-63
• Main Authors: Ahmad, Imran, Shaaikh, Jahier Abbas, Fenta, Assefa, Halefom, Afera, Andualem, Tesfa Gebrie, Teshome, Aseret
• Format: Article
• Language: English
• Subjects: Analysis; Autocorrelation; Calculators; Earth and Environmental Science; Geographic information systems; Geographical information systems; Geographical Information Systems/Cartography; Geography; Geophysics/Geodesy; Hot spots; Inflation (Finance); Measurement Science and Instrumentation; Remote Sensing/Photogrammetry; Soil erosion; Surveying
• ISSN: 1866-9298; 1866-928X
• DOI: 10.1007/s12518-021-00409-9

An attempt has been made in this study to analyze the hotspots of soil erosion in Denbie Woreda, Ethiopia. A total of five determinant factors were multiplied (using raster calculator in GIS domain) and integrated to obtain the soil erosion hotspots of the area. These five determinant factors were used as exploratory variables to understand the spatial behavior of soil loss rates. The ordinary least squared (OLS) regression model diagnostics showed that the R -squared and adjusted R -squared values of the explanatory variables are 0.71 and 0.75 respectively. Variance inflation factor (VIF) values of the OLS range between 1.03 and 1.47 indicating the absence of multicollinearity among explanatory variables. Koenker (BP) statistic is statistically insignificant ( p  > 0.005) indicating that relationships modeled are consistent. The OLS regression model was subject to different tests to confirm its reliability. Spatial autocorrelation tool (Global Moran’s I ) result showed that the residuals exhibit a Gaussian spatial pattern. This study proved that all the determinant factors of Revised Universal Soil Loss Equation played a significant role towards demarcation of soil loss rates.