Spatiotemporal Mapping of Drought Impacts Across Continents: A Cluster-Based Approach

Drought events pose significant challenges to ecosystems and human societies, necessitating precise methodologies for their identification and analysis. This study introduces a clustering technique to establish a robust framework for identifying drought objects. The identification process incorporates spatial proximity metrics, Haversine distance calculations, and periodic boundary handling to detect coherent drought-affected regions. Drought objects are further refined by applying a land-sea mask to exclude oceanic areas and merging small-scale clusters to maintain relevance. The study highlights the value of tracking drought objects over time and the critical insights this provides into the spatio-temporal dynamics of droughts.

The methodology enables a dynamic understanding of drought patterns, producing outputs such as high-resolution cluster maps with spatial characteristics, including the severity and area of each cluster. These characteristics are developed using drought events reported in the Geocoded Disasters (GDIS) dataset and are linked to the impact data, such as the number of people affected and economic damage caused by the events. These findings are vital for disaster risk reduction, climate impact studies, and policy-making. By integrating spatial analysis with the clustering, this study provides a comprehensive and reproducible approach to linking the geographical extent and intensity of drought events to their impacts.