Assessing CWSI as an indicator of water stress in oak trees based on high-resolution thermal images

Early detection of decline in forests—even before visual symptoms appear—would allow rapid interventions and would be facilitated by remote sensing indicators derived from high-resolution imagery. Thermal imaging can reveal the difference between canopy and air temperatures (Tc-Ta), which is widely used to assess early stress in both agricultural and natural systems. However, Tc-Ta is highly sensitive to variations in meteorological conditions. The Crop Water Stress Index (CWSI) was formulated to normalise Tc-Ta using theoretical or empirical baselines derived from fully irrigated and water-stressed reference conditions to minimise meteorological effects. While the CWSI is commonly applied to crops, its performance in natural and forested ecosystems remains largely untested. Here, we explored the use of the CWSI to monitor water-stress dynamics in holm oak ( Quercus ilex ) forests in different locations over time. Using an extensive database of aerial images at high spatial and temporal resolution to record canopy temperature continuously at two locations, we developed a standardised baseline specifically for holm oak forests. We also recorded relative humidity, air temperature, and crown temperature and collected field measurements of stem water potential and stomatal conductance. The CWSI was significantly correlated with both stem water potential (R 2 =0.526, p -value<0.05) and stomatal conductance (R 2 =0.517, p -value<0.05), suggesting that it could serve as a reliable indicator of water status in oak. Furthermore, image-based CWSI measurements allowed us to discriminate between different levels of oak decline, highlighting their utility for the dynamic monitoring of water status in forest trees.