Estimation of drought-induced forest decline of Scots pine in the Navarrese Pyrenees (Spain)

Identifying and predicting areas affected by drought-induced forest decline using remote sensing is key for monitoring forest-responses to drought, as well as for forest conservation and management. This is particularly important in the Mediterranean Basin, a climate change hotspot, where drought-induced forest decline has become increasingly frequent in recent decades. In this study, we focused on Pinus sylvestris forests in the Navarrese Pyrenees to quantify drought-induced forest decline, examined its relationship with climatic and hydrological variables, and assessed how the temporal evolution of greenness and wetness indices, namely Tasselled Cap Greenness (TCG) and Wetness (TCW) correlates with tree canopy damage levels. We parameterised a canopy damage model using 20 m grids from Sentinel-2 imagery and a Support Vector Machine algorithm. The model was calibrated with data from a site experiencing severe drought-induced forest decline and a healthy site. Then, we extended the prediction of tree canopy damage across the pine distribution in the same provenance region, with similar ecological conditions. The model achieved a strong fit of R2 = 76%, with an average uncertainty of CI95% = 14.6%, although it increased at higher canopy damage. The highest tree canopy damage was observed in areas of low elevations and precipitation, coupled high temperatures and evapotranspiration rates. Canopy damage showed a negative correlation with TCW and with TCG. However, during summer, areas with high tree canopy damage had the highest greenness, which could be due to background signals from the understory, surpassing the TCG levels of healthier forests.