Drought-induced damage detection in Iberian Scots pine forests through satellite remote sensing

Drought events are becoming more intense, prolonged, and hotter in the Mediterranean region. Forest biomass and tree competition for limiting resources are also increasing due to reductions in management intensity and rural abandonment, increasing the probability of drought-induced forest decline. Future climate change scenarios predict increased warming and drought intensity, highlighting the importance of developing spatio-temporal monitoring systems to identify drought-induced decline and tree mortality events. Despite satellite remote sensing techniques combined with field data represent the most effective approach for developing forest health monitoring systems, our understanding of which sensors, spectral and spatial resolutions are most suitable for detecting forest decline is still limited. In this study, we assessed the potential of Landsat-8/OLI, Sentinel-2/MSI and PlanetScope/SD sensors to detect drought-induced forest decline in Mediterranean Scots pine forests quantifying the influence of spectral resolution, spatial resolution, level of damage and forest cover. Sentinel-2/MSI showed the highest performance, especially when using spectral indices that include SWIR (NDWIswir2, TCW, MDWI) or red edge bands (NDRE1, MTCI) as indicated by the c. 80% correlation with canopy damage. Optimum spatial resolution for all sensors was 20-30 m; however, forest decline detection was strongly affected by forest cover, with forest cover under 50-60% decreasing dieback detection, probably due to the background signal from soil and understory. Together, our results provide key insights for developing drought-induced forest decline and tree mortality monitoring systems in Mediterranean forests using satellite remote sensing.