Remote sensing across scales and platforms: monitoring Castelporziano nature reserve forest insect outbreaks

Abstract Stone pines ( Pinus pinea ) are a dominant species in Mediterranean coastal forests, playing a crucial ecological and landscape role. However, these ecosystems are experiencing increasing pressure from rising temperatures, prolonged droughts, and other stressors, which weaken trees and make them more susceptible to natural disturbances, such as insect infestations, leading to significant forest decline. This study presents a multi-platform, temporal, and spatial resolution remote sensing approach to detect and monitor insect outbreaks. The study area is Castelporziano Presidential Estate (Rome, Italy), a pine-dominated periurban forest of high ecological value and severely affected by Tomicus destruens (Coleoptera: Curculionidae) and Toumeyella parvicornis (Hemiptera: Coccidae). Very high-resolution Pléiades and DigitalGlobe images were used to map forest disturbances accurately and to assess the accuracy of the yearly forest disturbance maps obtained through Sentinel-2 data. Sentinel-2 data enabled the production of yearly disturbance maps in near real time (NRT), improving the Three Indices Three Dimensions (3I3D) unsupervised forest disturbance algorithm. Lastly, PlanetScope data were used to calculate the Mean Time Lag (MTL) between disturbance occurrence and detection in Sentinel-2 NRT images, revealing an average detection delay of 8.5 days. Additionally, the unsupervised model’s performance was evaluated, yielding an overall accuracy of 85%. The accuracy of the Sentinel-2 forest disturbance map was assessed against the map derived from Pléiades/DigitalGlobe, achieving an overall accuracy of 92%. These findings confirm that integrating multi-source remote sensing data enhances the timeliness and reliability of insect outbreak monitoring, offering valuable tools for managing Mediterranean forests threatened by climate change and biotic agents. Graphical Abstract