INNOVATIVE INDICES FOR WATER STRESS DETECTION FROM HYPERSPECTRAL INFRARED OBERVATIONS

Monitoring surface and vegetation conditions is crucial for analyzing the impact of climate change on natural resources. Detecting vegetation stress using remote-sensing data is essential for understanding these changes and taking action against extreme events like land and forest dryness caused by summer heatwaves in the Mediterranean region. Commonly used satellite indices for this purpose include the Normalized Difference Vegetation Index (NDVI), followed by the Leaf Area Index (LAI), Fraction of Absorbed Photosynthetically Active Radiation (FAPAR), Surface Soil Moisture (SSM), and physical parameters such as near-surface air temperature (Ta), obtained from both remote sensing data and on-site measurements. However, it is a well-known fact that NDVI cannot distinguish between barren soil and distressed vegetation and, while surface temperature and air temperature have an influence on soil moisture, correlation among these three quantities is not straightforward to evaluate. This thesis aims at demonstrating the effectiveness of two newly-developed thermodynamical indices, the Emissivity Conrast Index (ECI) and theWater Deficit Index (WDI), in assessing vegetation stress and woodland degradation, specifically in southern Italy from 2014 to 2022. ECI is based on infrared surface emissivity, closely related to land cover, while WDI directly measures surface water loss. These indices have been calculated using physical parameters derived from observations acquired by the Infrared Atmospheric Sounding Interferometer (IASI), then upscaled and remapped on a regular grid using an Optimal Interpolation (OI) scheme. A comparison with other traditional indices is presented, further validating the applied methodology. Additionally, it is shown how the synergy between ECI and WDI can be exploited to identify the criminal origin of a fire event, specifically the Mount Vesuvius arsons of summer 2017, uncoupling the fire outbreak from the heatwave that affected the Mediterranean area during that same period. Finally, the Weather Reasearch and Forecasting (WRF) model is used with two different sets of global forecasts as input -Global Forecast System (GFS) and European Centre for Medium-RangeWeather Forecasts (ECMWF)-to calculate WDI, surface and dew-point temperature (Ts and Td) maps for July 2017 over southern Italy. The results are compared against the WDI maps obtained from IASI retrievals, showing a high level of similarity, as well as limited yet interesting differences, mainly related to the specific implementations of the global Numerical Weather Prediction (NWP) models from which the input forecasts were derived.