Seasonal lags and legacy effects on vegetation cover and carbon uptake in Kuwait from 2014 to 2024: A MODIS GPP–ET–NDVI analysis

Abstract

Abstract Remote sensing provides a consistent framework for quantifying vegetation dynamics in hyper-arid environments where field observations are sparse. This study integrated MODIS-derived gross primary productivity (GPP), evapotranspiration (ET), and normalized difference vegetation index (NDVI) with GPM IMERG precipitation to evaluate seasonal variability, carbon–water coupling, rainfall-response lags, and cumulative hydroclimatic controls across Kuwait during 2014–2024. Vegetation activity peaked in spring, when mean GPP reached 2.25 ± 0.41 g C m⁻² d⁻¹ and NDVI 0.26 ± 0.05, compared with 0.92 ± 0.21 g C m⁻² d⁻¹ and 0.12 ± 0.03 in summer. ET was strongly associated with GPP (R² = 0.680, p < 0.001) and NDVI (R² = 0.533, p < 0.001), whereas same-month precipitation explained substantially less variation in GPP (R² = 0.189, p = 0.003) and NDVI (R² = 0.081, p = 0.061). Cross-correlation analysis showed that both vegetation indicators responded strongly when rainfall preceded vegetation activity by one month (GPP: r = 0.711; NDVI: r = 0.678; both p < 0.001), with the NDVI response persisting longer than GPP. At the cumulative interannual scale, ET remained significantly associated with GPP (R² = 0.524, p = 0.012) and NDVI (R² = 0.630, p = 0.004), whereas cumulative precipitation was not significant. These findings demonstrate that vegetation dynamics in Kuwait are better characterized by ecosystem water use and antecedent rainfall than by contemporaneous or cumulative precipitation alone, highlighting the importance of rainfall timing and short-term hydroclimatic legacy effects in hyper-arid ecosystems.