Monitoring spatio-temporal drought dynamics using multiple indices in the dry land of the upper Tekeze Basin, Ethiopia

Abstract

Due to Ethiopia's broad variations in biophysical and climatic variables, an accurate understanding of local-level drought is critical for sustainable drought risk management. This study aims to monitor spatio-temporal drought dynamics over Tekeze basin from 1981 to 2021 using Standard Precipitation Index (SPI), Vegetation Condition Index (VCI), Temperature Condition Index (TCI), and Vegetation health Index(VHI). The analysis relied on data from Moderate Resolution Imaging Spectroradiometer (MODIS) datasets and CHIRPS-v2. Google Earth Engine (GEE) was used to obtain data, process NDVI trends, and calculate drought indices (DI). Man-Kendall trend analysis and Pearson correlation were also employed to examine the trend and association of DI and climate variables. The SPI showed that the basin was affected by moderate, severe, and extreme drought in 1884, 1985, 1987, 1993, 1997, and 2015. TCI and, VCI, analysis indicated that 2002, 2004, 2009, 2015, 2016, and 2017 were severe and extreme droughts. The NDVI showed a decreasing trend throughout most of the basin, except for pocket areas of the managed watershed, area closures, and irrigation sites. Rainfall in July emerged as a critical factor in determining NDVI, LST, TCI, and VCI in July and August. While VCI strongly correlates with precipitation and LST is less correlated. Although all DIs are robust in assessing agricultural droughts, VCI detected more land areas under severe and extreme drought than TCI. The study underscores the importance of larger weights to VCI in correctly classifying drought, particularly in drylands where precipitation is crucial in determining vegetation health.