Data-driven assessment of climate change and vegetative cover dynamics in traditional oases

Abiod Valley, Aurès region, Algeria. This work focuses on the relation between climatic forcing and vegetation cover dynamics in traditional oases. This study pursues two main objectives: (1) estimate the vegetative surface cover of traditional oases from satellite images, (2) quantify the impact of climatic variables on vegetation dynamics and assess future scenarios. We propose a methodology that leverages satellite imagery and derived indices (NDVI, NDMI) to quantify vegetation cover and water stress events at the oasis spatial scale. We then assess the feedback between climate and vegetation cover at monthly and yearly scale through multivariate analyses based on vector autoregression (VAR) and vector error correction (VEC) models. Our findings reveal an appreciable decrease in vegetation cover over the last decade across three considered traditional oases in the study region. The monthly scale analysis suggests a lagged effect of climatic variables, especially cumulative precipitation, on vegetation water stress. The long term VEC prediction of climatic variables aligns with GDDP-CMIP6 climate projections, forecasting an increase in average temperature and potential evapo-transpiration. A significant decline in vegetative surface cover is predicted by 2050 from the analysis of yearly data, highlighting the critical need for water management interventions to safeguard oasis ecosystem and prevent desertification. • Dynamic method for estimation of oasis area and water stress from remote sensing. • Autoregressive data-driven model coupling oasis vegetation with climatic variables. • Oases system decline in the Aures region (Algeria) is quantified. • Future sharp decline of vegetation is predicted calling for actions.