The use of Sentinel-2 satellite data for spatially variable rate fertilisation of durum wheat crop in the centre of Sicily

Remote Sensing (RS) using Unmanned Aerial Vehicles (UAVs) and satellites, specifically the Copernicus Sentinel-2 satellite, is a valuable technique in Precision Agriculture (PA). It employs the Normalised Difference Vegetation Index (NDVI) as an indicator of vegetation vigour to generate maps for spatially variable rate fertilisation. This study aims at assessing the potential of RS from the Copernicus Sentinel-2 satellite in sensing the spatial variability of vegetation vigour and plant leaf water content in durum wheat fields of inland Sicily. The objective is to integrate this information into a Decision Support System (DSS) for generating fertilisation maps divided into Management Zones (MZs). For this case study, a continuous durum wheat field of 4.23 hectares ca., located in the centre of Sicily, was chosen. Google Earth Pro was used to identify and delineate the study area, and the resulting polygon was saved as a .kml file. This vector map was incorporated into the AgroInsider online platform. The investigation spanned the growing seasons 2021-2022 (cv. Arcangelo) and 2022-2023 (a mix of cv. Arcangelo, Simeto and Garigliano). NDVI and Normalised Difference Water Index (NDWI) values were collected, and NDVI images were extracted from the AgroInsider platform during the key phenological growth stages, i.e. seeding, sprouting, raising and harvest. By analysing the NDVI graph, the optimal fertilisation time (06/02/2023) and the sprouting onset (23/03/2022 and 13/03/2023 for the two growing seasons, respectively) were identified. The raising onset (12/04/2022 and 07/04/2023 for the two growing seasons, respectively) was determined by comparing NDVI and NDWI graphs. Subsequently, fertilisation maps divided into two MZs were extracted from the AgroInsider platform for 01/02/2023, immediately before the optimal fertilisation date. The RS data from the Sentinel-2 satellite proved highly effective in discerning the spatial variability of vegetation vigour and plant leaf water content in the marginal areas of inland Sicily. This underscores the utility of RS as an integral component of a DSS, capable of efficiently generating fertilisation maps divided into MZs.