Spatiotemporal characterization of agricultural drought in the Sahel region using a composite drought index

• Published in: Journal of arid environments 2022-09, Vol.204, p.104789, Article 104789
• Main Authors: Abdourahamane, Zakari Seybou, Garba, Issa, Gambo Boukary, Aboubakr, Mirzabaev, Alisher
• Format: Article
• Language: English
• Subjects: CHIRPS; Composite drought index; Entropy method; NDVI; Sahel
• ISSN: 0140-1963; 1095-922X
• DOI: 10.1016/j.jaridenv.2022.104789

A composite drought index (CDI) was developed for seasonal drought monitoring at 1 km2 resolution over the West African Sahel (WAS). The CDI was derived from remote sensing data, mainly, the Climate Hazards Group InfraRed Precipitation with Stations (CHIRPS), normalized difference vegetation index (NDVI) and land surface temperature (LST) from the Terra/MODIS satellite. The weights of these input variables were estimated using a combined entropy method and weighted Euclidian distance. The CHIRPS resulted with the highest weight (mean = 0.74) followed by the NDVI (mean = 0.243) and the LST (mean = 0.016). The CDI was found to be well correlated with the standardized precipitation index (SPI) and the standardized precipitation evapotranspiration index (SPEI)—computed from station data. However, the CDI showed a better sensitivity for drought detection following a comparison of drought classes. The suitability of the CDI for agricultural drought monitoring was validated by its good correlation with crop production data, namely maize, millet and sorghum with a Pearson r in the range of 0.29–0.56, 0.40–0.81 and 0.57–0.71, respectively. Finally, a drought database was generated for the WAS, enabling the extraction of drought characteristics at a given location using its geographic coordinates. •A remote sensing based composite drought index (CDI) was developed.•The CDI is suitable for drought characterization at small spatial scale.•The CDI is more sensitive than the SPI and SPEI in detecting drought over the Sahel.•The CDI exhibited a good correlation with maize, millet and sorghum production data.•A drought database was generated for agricultural drought monitoring in the Sahel.