Abstract

The mealybug plague (Delotococcus aberiae) is causing serious economic losses to the Spanish agricultural sector, especially to citrus fruits. In this context, our main aim is to leverage the main satellite acquisitions from the EU Copernicus program, that is Sentinel-2 (optical data) and Sentinel-3 (thermal data) observations, to detect the incidence of the mealybug in citrus trees. Particularly, we will determine whether using thermal data the pest can be detected earlier than using only optical data. This study has been developed over the surroundings of Vall d'Uixó (Castellón, Spain) by analyzing, during 2021-2022 season, 50ha of citrus parcels classified depending on the affection level as healthy or unhealthy. Our work is based on integrating Sentinel-2 (optical) and Sentinel-3 (thermal) data. With Sentinel-3 Land Surface Temperature (LST) at 1km spatial resolution, we developed a downscaling method by establishing spatial relationships at 1 km between the Normalized Difference Vegetation Index (NDVI) (derived from Sentinel-2) and LST (from Sentinel-3 SLSTR L2 LST). By performing a linear regression between both parameters at 1 km resolution and assuming that the relationship will hold regardless of resolution, the LST resolution upscale can be performed, considering the NDVI at 10 m as the independent variable. Then, we have studied the relationship between the mealybug affection and the LST. Regarding the optical domain, the NDVI and other optical spectral bands' evolutions (RED, NIR and SWIR) have also been considered. The downscaling methodology has been validated against a 30-meter resolution LST derived from Landsat 9 OLI/TIRS data in 2022 over the studied area, providing a Root Mean Square Error of 2 °C. Preliminary results show significative contrasts between temporal evolutions of parcels with different affection level in NIR, RED and SWIR channels. Furthermore, shows NDVI higher values for healthy parcels, while LST shows lower values with differences up to 3 °C. Preliminary results show that remote sensing data can be successfully leveraged to monitor the mealybug plague in a timely, objective and cost-efficient manner. This work contributes to the development of new monitoring tools for efficient and sustainable action in the fight against natural enemies.