Estimating Grassland Parameters from Sentinel-2: A Model Comparison Study

• DOI: 10.1007/s41064-020-00120-1
• Publication Date: 2020-08-05T12:02:54Z
• Authors: M. Schwieder; M. Buddeberg; K. Kowalski; K. Pfoch; J. Bartsch; H. Bach; J. Pickert; P. Hostert
• Keywords: Grassland ecosystem

Abstract Grassland plays an important role in German agriculture. The interplay of ecological processes grasslands secures ecosystem functions and, thus, ultimately contributes to essential services. To sustain, e.g., the provision fodder or filter function soils, agricultural management needs adapt site-specific grassland characteristics. Spatially explicit information derived from remote sensing data has been proven instrumental for achieving this. In this study, we analyze potential Sentinel-2 deriving grassland-relevant parameters. We compare two well-established methods calculate aboveground biomass and leaf area index (LAI), first using a random forest regression second soil–leaf-canopy (SLC) radiative transfer model. Field were recorded on Brandenburg August 2019, used train empirical model validate both models. Results confirm that are suitable mapping spatial distribution LAI quantifying biomass. Uncertainties generally increased with higher values varied average by relative RMSE 11% modeling dry 23% LAI. Similar estimates achieved SLC 30% retrieval, 47% estimation Resulting maps approaches showed comprehensible patterns distributions. Despite variations value ranges maps, very similar. Based results compared comparison validation data, conclude relationship between spectra variables can be quantified map their distributions space. Future research investigate how similar perform across different types, seasons regimes.