Multivariate Logistic Regression Model for Soil Erosion Susceptibility Assessment under Static and Dynamic Causative Factors
More details
Hide details
Department of Civil and Environmental Engineering, Universiti Teknologi PETRONAS, 32610 Seri Iskandar, Perak, Malaysia
Department of Water Resources and Environmental Engineering, University of Ilorin, PMB 1515, Ilorin, Kwara State, Nigeria
Centre for Urban Resource Sustainability, Institute of Self-Sustainable Building, Universiti Teknologi PETRONAS, 32610 Seri Iskandar, Perak, Malaysia
Department of Earth Sciences, Quaid-i-Azam University, 45320 Islamabad, Pakistan
Faculty of Industrial Management, Universiti Malaysia Pahang, 26300 Gambang, Pahang, Malaysia
Abdulkadir Taofeeq Sholagberu   

Universiti Teknologi PETRONAS, Department of Civil & Environmental Engineering, Universiti Teknologi PETRONAS, 32610 Seri Iskandar, Malaysia
Submission date: 2018-02-01
Final revision date: 2018-05-27
Acceptance date: 2018-06-03
Online publication date: 2019-05-07
Publication date: 2019-05-28
Pol. J. Environ. Stud. 2019;28(5):3419–3429
Soil erosion is a devastating land degradation process that needs to be spatially analyzed for identification of critical zones for sustainable management. Geospatial prediction through susceptibility analysis assesses the occurrence of soil erosion under a set of causative factors (CFs). Previous studies have considered majorly static CFs for susceptibility analysis, but neglect dynamic CFs. Thus, this study presents an evaluation of erosion susceptibility under the influence of both non-redundant static and dynamic CFs using multivariate logistic regression (MLR), remote sensing and geographic information system. The CFs considered include drainage density, lineament density, length-slope and soil erodibility as static CFs, and land surface temperature, soil moisture index, vegetation index and rainfall erosivity representing the dynamic CFs. These were parameterized to establish geospatial relationships with the occurrence of erosion. The results showed that length-slope had the highest positive impact on the occurrence of erosion, followed by lineament density. During the MLR classification process, predicted accuracies for the eroded and non-eroded locations were 89.1% and 83.6% respectively, with an overall prediction accuracy of 86.6%. The model’s performance was satisfactory, with 81.9% accuracy when validated using the area-under-curve method. The output map of this study will assist decision makers in sustainable watershed management to alleviate soil erosion.