Analysis of the Relationship Between Different Drought Indices: Case Study in the Upper Jaguaribe Basin

Abstract

The impacts of droughts vary depending on their intensity, duration, and recurrence, as well as the developmental state and coping capacity of regions and countries. Drought can be categorized into three main types: meteorological, hydrological, and agricultural. Effective mitigation of drought impacts is achievable through an efficient monitoring system. Given that drought is a multifactorial event, the analysis of multiple indices from different classes aids in its monitoring. This study aims to address a research gap by assessing the three types of droughts using indices in semi-arid regions, with the goal of enhancing monitoring and management of these conditions. This work examines the behavior of seven drought indices - SPI, SPEI, EDDI, RDIst, SRI, VHI, and NDVI - calculated for the Upper Jaguaribe Basin using both a moving time window and temporal discretization, for both 3 and 12-month scales. Mutual information value and Pearson correlation were employed as analytical metrics between the indices. Results for meteorological indices SPI, SPEI, and RDIst exhibited significant correlation in both linear and non-linear analyses. EDDI displayed distinct behavior from the other meteorological indices. Notably, EDDI-12 showed higher mutual information than NDVI, suggesting early warning potential for vegetation drought on broader scales. This study analyzes drought propagation through correlation between 3-month scale indices and SPI-3 for February-March-April. Findings demonstrate drought propagation, with vegetation-based indices (VHI and NDVI) and hydrological index (SRI) showing higher correlations in March-April-May, while meteorological indices (RDIst and SPEI) exhibited stronger correlations in February-March-April.

Keywords
drought; drought indexes; Upper Jaguaribe basin