Extreme climatic events, such as heatwaves and prolonged droughts, significantly contribute to forest degradation by increasing tree mortality and dieback. This phenomenon results in the loss of tree organs, thereby reducing ecosystem resilience. Although climate-driven dieback has been extensively studied in deciduous trees, relatively few studies have monitored this phenomenon in evergreen forests.

This study aims to investigate dieback affecting Mediterranean evergreen forests, using the Gargano promontory (Apulia, Italy) as a case study, where significant dieback was recorded in the summer of 2024. Specifically, we focus on forests dominated by Quercus ilex and Pinus halepensis, aiming to identify key environmental variables (geomorphological, hydrological, and climatic) that explain differential dieback patterns, and to assess the ecosystem's capacity for recovery through remote sensing techniques.

We analysed spectral indices representing eco-physiological attributes, such as leaf chlorophyll content, productivity, and canopy biomass, during stress (April–August) and potential recovery (September–March) periods in dieback-affected areas. Our analysis utilised monthly Landsat 8 imagery to calculate a suite of spectral indices serving as proxies for these eco-physiological parameters. Using linear regression models, we identified dieback-affected areas by detecting significant negative trends in these indices between the stress and recovery periods. Subsequently, we applied a Random Forest model to correlate observed dieback trends with environmental predictors.

Our preliminary findings indicate two key outcomes: firstly, dieback intensity varies significantly across the landscape; secondly, there is a strong relationship between geomorphological variables and the patterns, severity, and recovery potential of these forests.

This research underscores the potential of satellite imagery as a powerful tool for monitoring eco-physiological stress in Mediterranean evergreen forests, highlighting their vulnerability to climate-driven dieback events.