eDNA biodiversity from space: predicting soil bacteria and fungi alpha diversity in forests using DESIS satellite remote sensing

Environmental DNA (eDNA) allows thousands of microbiological taxonomic units to be identified from a small amount of environmental matrix, e.g. soil, a leaf or water. However, the spatial and temporal diversity of the microbiological world over contiguous and extensive areas remains unknown, primarily due to the typically large distances between <i>in situ</i> field samples which are usually collected infrequently or even one-off. Here we describe a fundamentally different approach to alpha biodiversity modelling, by coupling remote sensing image spectroscopy from the DESIS satellite with eDNA profiling using Gaussian Process Regression. Alpha diversity is mapped at a fine resolution of 30 m, and expressed using metrics including functional richness, Shannon index as well as phylogenetic diversity. Up to 50% variance in bacterial alpha diversity, and 40% of fungal alpha diversity variance, are explained by our modelling. Further, we demonstrated how ecological relations are signalled by the microbiological communities, in three European temperate forests. Our findings demonstrated for the first time using the combination of eDNA and image spectroscopy, that key ecological variables relate to alpha diversity, thereby accelerating our understanding of the spatial variation in bacterial and fungal communities across large and complex forest landscapes.