**3. Results**

The data collected are presented by thematic maps. From the atmospheric and climatic data collected on the field, the CWSI for each waypoint has been evaluated and associated with its geographical coordinates in GeoTIFF format. This way, the georeferenced CWSI map can be layered on the crop area through a Geographic Information Software (QGIS).

The map from satellite data is plotted to show the different values of the CWSI among the different regions in the field. Figure 2a shows that the CWSI is quite homogeneous and ranges from 0.4 to 0.9 over the whole field. The distribution of CWSI values based on the ground-collected data is more irregular and uneven. Areas A and B show values of CWSI ranging mainly from 0.5 to 1, with few squares having an index value below 0.5, while areas C and D show values ranging from 0 to 0.2.

As pointed out in Table 1, the values of CWSI can be grouped into four classes that differentiate between 0, "no water stress", up to 1, "critical water stress". The map in Figure 2b highlights that grouping. To further inspect the importance of each factor to the CWSI value, we have performed a Global Variance Based Sensitivity Analysis [5], computing the main and total effects of the contribution of each variable to the variance of the output. The results are displayed in Figure 3.

**Figure 2.** (**a**) CWSI map evaluated through atmospheric data by Landsat 8 and Copernicus database (14 July 2018); 30 × 30 m resolution. (**b**) CWSI map based on ground-collected data (4 June 2022); 10 × 10 m resolution.

**Figure 3.** Main and total effects, and 0.95 confidence intervals for variables: Atmospheric Temperature (Ta), Canopy Temperature (Tc), UV Radiation (UV), Atmospheric Pressure (p), Relative Humidity (Ur), Wind Speed (u2) and Ground Height (h).
