Prisma Images for the Estimation of Functional Traits and Biodiversity Indices in Mid-Latitude Forests

In this contribution, we exploit hyperspectral data cubes collected by the PRISMA satellite to develop and test machine learning regression algorithms (MLRAS) for mapping forest traits and estimating spatial patterns of plant beta diversity in mixed forest ecosystems. The MLRAs were trained using PRISMA data and ground data collected in correspondence of the satellite overpasses. Based on their ecological importance, leaf chlorophyll content, leaf water content, leaf mass area, leaf nitrogen content and leaf area index were investigated. Finally, emerging methods aimed at the estimation of forest biodiversity based on spectral diversity were applied to hyperspectral images.Results showed that MLRAs performed well in leaf and canopy traits retrieval with high predictive capacity for all the models tested. Moreover, the spatio-temporal patterns of functional traits generally agreed with the spectral diversity patterns.