Multi-scale drought-vegetation relationships in the Jebel Marra Highlands, Sudan: evidence from Landsat NDVI time series, SPEI, and BFAST analysis

Résumé fourni par la source

Dryland ecosystems are highly sensitive to climatic variability, particularly drought, which influences vegetation productivity, ecosystem function, and land degradation processes. Understanding vegetation–drought relationships is therefore important for assessing ecosystem dynamics in semi-arid environments. This study investigated long-term vegetation dynamics and drought conditions in the Jebel Marra Highlands of western Sudan using Landsat-derived Normalized Difference Vegetation Index (NDVI) time series and multi-scale Standardized Precipitation–Evapotranspiration Index (SPEI) data for the period 2000–2024. Monthly NDVI time series were derived from Landsat imagery, while SPEI was calculated from Climate Hazards Group InfraRed Precipitation with Station Data (CHIRPS) precipitation and ERA5-Land potential evapotranspiration datasets for 3-, 6-, and 12-month accumulation periods. Temporal trends were evaluated using the Mann–Kendall test, vegetation–drought relationships were examined using correlation and lagged-correlation analyses, and vegetation regime shifts were identified using the Breaks For Additive Seasonal and Trend (BFAST) method. The NDVI time series exhibited a significant positive long-term trend ( τ = 0.24, p < 0.001), indicating an overall increase in vegetation greenness during the study period despite substantial interannual variability. Contemporaneous correlation analysis identified a significant positive relationship between NDVI and SPEI-3 ( r = 0.442, p < 0.001), whereas relationships with SPEI-6 ( r = −0.074, p = 0.191) and SPEI-12 ( r = 0.044, p = 0.443) were not significant. Lagged-correlation analysis showed varying temporal associations among drought timescales, with maximum correlations occurring at a lag of 2 months for SPEI-3 ( r = 0.833) and 3 months for SPEI-6 ( r = 0.868). BFAST analysis identified consistent vegetation breakpoints around 2006–2007, 2012, and 2018 across multiple model parameterizations in the tropics. Seasonal analyses showed lower NDVI values during drought periods, particularly during the peak growing season (July–September). Overall, the results demonstrate that vegetation variability in the Jebel Marra Highlands was more strongly associated with short-term moisture conditions than with concurrent medium- or long-term droughts. The identified structural changes in the NDVI time series indicate periods of substantial vegetation reorganization during the study period. The integration of Landsat observations, locally derived SPEI datasets, and breakpoint analysis provides a useful framework for monitoring vegetation dynamics and hydroclimatic variability in semi-arid mountain environments.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.