Towards The Retrieval of Plant Traits in Grasslands with Hyperspectral Enmap Data

Spatio-temporal information on grasslands traits (e.g., biomass, plant nitrogen concentration) are valuable information for precision agriculture and for the quantification of regional carbon and nitrogen (N) cycles. Remote sensing approaches showed a high potential to provide this kind of information. In this study, we aim to explore the potential of the recently launched hyperspectral satellite EnMAP for the retrieval of plant traits in grasslands. Due to the lack of corresponding field data for EnMAP acquisitions, we first developed different retrieval approaches for above-ground biomass (AGB) and plant N concentration (N%) for a study region in southern Germany based on Sentinel-2 data, i.e., empirical machine learning (ML) and hybrid models. Then, we adapted the hybrid models to EnMAP data (as they do not require field data for model training) and assessed their plausibility.