Towards improved USLE-based soil erosion modelling in India: A review of prevalent pitfalls and implementation of exemplar methods

Abstract

One of the most common approaches to modelling soil erosion worldwide has been the implementation of the original Universal Soil Loss Equation (USLE) and its revised version, the RUSLE. However, despite its widespread use, often there are discrepancies in the methods used to compute it and in the values elicited for the five individual factors that comprise this function. Such pitfalls subsequently skew the final results obtained and often many studies also fail to adequately examine the accuracy of the enumerated soil loss amounts. We examine these aspects with respect to the raft of USLE-based studies undertaken in India over the last few decades, reviewing a total of 100 investigations in this regard. Results reveal that almost all studies had either over- or underestimated at least one of the five factors, thereby possibly misrepresenting the actual soil loss occurring from their examined areas. Even more worryingly, most studies had failed to document their methods succinctly or in sufficient detail to ascertain their efficacies or provide viable templates for replication elsewhere. Our results also show a marked spatiality in the pursuance of such studies, with these being mostly undertaken in the eastern part of the country, even though the proportionate land affected by soil erosion is considerably less in this region. Thus regions where the USLE would be most pertinent for implementation towards land management have seen a lower number of applications. We hope that by avoiding the missteps highlighted in this paper and following the subsequently detailed exemplar methods of conducting such an investigation along with the relevant model accuracy and uncertainty checks, the USLE can be best utilised in these regions and in the rest of the country for soil erosion mitigation. Though focused on India, the methods outlined can also be used to conduct the most accurate possible USLE-based soil erosion modelling elsewhere.