Exploring the potential of remote sensing application to assess locust damage: A case study of the first and second wave locust invasion in Laikipia and Samburu counties, Kenya.

Abstract

The 2020-2021 desert locust invasion in Kenya caused massive agricultural damage, threatening food security in Arid and Semi-Arid Lands. Control and assessment of the damage are complicated to begin with and were further hindered by the current travel restrictions in place due to the Covid-19 pandemic. To contribute to ground-based data, remote sensing analysis offers an opportunity to assess the damage situation from a different viewpoint. It offers the advantage of a greater spatial scale analysis covering areas that would be inaccessible with ground-based observation. This thesis addresses the feasibility of using remote sensing for desert locust damage assessment studies in two Kenyan counties: Laikipia and Samburu. The computation of the impact by desert locusts was executed through the development of a Normalized Difference Vegetation Index (NDVI) change detection model using Sentinel-2 Surface Reflectance (SR) to compute the before and after invasion NDVI difference image. The NDVI change was computed to show areas that were damaged after the desert locust invasions in 2020 and 2021. The results of the model showed that Samburu county was adversely affected by both desert locust invasions correlating with ground-based household assessments. Laikipia recorded the highest cropland Land Use and Land Cover (LULC) damage after both the first and second invasion exacerbating food insecurity in the county. For further research, the developed model could be enhanced by ground-based data in the future.