Soil erosion risk assessment in the Umzintlava catchment (T32E), Eastern Cape, South Africa, using RUSLE and random forest algorithm

ABSTRACT

The Revised Universal Soil Loss Equation (RUSLE), based on remotely sensed data, is an important tool for assessing erosion prone areas and serves as a guide towards soil conservation efforts. Besides being a crucial data source from which RUSLE parameters can be derived, remotely sensed data can also be used independently to delineate erosion features. This study aims to assess soil erosion risk in the Umzintlava catchment using two independent methods, i.e. RUSLE and Random Forest (RF), and explore the relationship between soil loss and erosion factors as represented by different RUSLE parameters. To achieve this, rainfall, soil, digital elevation, and satellite data were used. The results indicate that a considerable portion (>90%) of the catchment area is of ‘very low’ to ‘low’ erosion risk, while the remainder suffers ‘moderate’ to ‘extremely high’ erosion risk. Among erosion factors, the LS-factor (slope length and steepness) showed strong correlation with soil erosion (p < 0.001; R² = 0.954). This suggests that areas with steep slopes are the most vulnerable to hillslope erosion, whereas gully erosion is prominent in areas with gentle to nearly flat slopes. The integration of RUSLE-derived soil loss and RF-derived erosion features successfully delineated the spatial patterns of soil erosion across the Umzintlava catchment, providing useful information on erosion risk at least costs. This information is instrumental in targeted management interventions to combat soil erosion within the catchment.

KEYWORDS:

• Soil erosion risk
• Umzintlava catchment
• Revised Universal Soil Loss Equation (RUSLE)
• erosion factors
• random forest

Acknowledgments

This paper is part of a research project carried out with financial support from the National Research Foundation (NRF) under grant [number 107329]. The authors are thankful to the South African Weather Services (SAWS) for providing climatic data, the United States Geological Survey (USGS) for availing DEM data at no cost, and the South African National Space Agency (SANSA) for providing free SPOT images.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Research Foundation [107329].