Geospatial modelling of rainfall erosivity in the humid tropics using remotely sensed data

Intense rainfall events characterise the humid tropical small island developing states (HuT-SIDS). Therefore, appropriate computation of rainfall erosivity is essential in predicting soil loss in the face of climate change. The rainfall erosivity index is calculated using kinetic energy and rainfall intensity of individual storm events requiring data that are often unavailable in HuT-SIDS. The objectives of our study were: (i) to investigate event-based rainfall erosivity using two erosive rainfall indices; and (ii) to create iso-erodent lines using Global Precipitation Measurement Integrated Multi-satellite Retrievals (GPM IMERG). We developed a Storm Event Tool (SE-TOOL) in the ArcGIS environment using Python scripts to automate the processing of sub-hourly precipitation data. The methodological approach integrated Python Scripts, Model Builder, geostatistical analytics and GPM IMERG data to calculate the erosivity indices, offering a novel approach to soil loss modelling. The results showed that the mean EI30 index, the product of total storm kinetic energy (E) and its maximum continuous 30-min intensity (I30), was 5080 MJ-mm-/ha-h. The Kinetic Energy greater than 25 mm (KE >25) index, recorded a mean erosivity of 1928 MJ/ha. The difference between the GPM IMERG and the Automated Weather Station was not significant (t (30) = 2.07; p < 0.05, 0.67). EI30 was determined to be the more appropriate erosivity index for predicting soil loss in the HuT-SIDS. The integration of higher temporal and spatial resolution data and including all storm events above the 0 mm threshold, increased the model accuracy, providing baseline data for soil erosion risk assessment.