C4: Understanding the spatial and temporal dynamics of vegetation in wetscapes 2.0 through multi-sensor Earth observation time series data
Project leader: Prof. Dr. Sebastian van der Linden (University of Greifswald)
A better understanding of functional processes in and around rewetted fens requires information on vegetation cover and its dynamics at high spatial and temporal resolution and at a very high level of thematic detail, including, e.g., cover fractions of (peatland) vegetation types or species, vitality or biomass. Data from current and imminent Earth observation (EO) satellites provide this information on vegetation cover with the required level of detail, accuracy and spatial resolution, and with an areal/temporal coverage that cannot be achieved with drone-based or airborne imagery. C4 aims to produce spatially continuous maps that (i) achieve a higher level of thematic detail than regularly available maps (e.g. by introducing additional peatland vegetation classes), (ii) quantify vegetation cover annually (species coverage) and intra-annually (fractions of green and non-photosynthetic vegetation throughout time), and (iii) help to extrapolate biomass estimates. This will be used to analyse the effects of rewetting on vegetation cover and to discribe or disaggregate identified processes throughout space and/or time. Furthermore, the input for process-based modelling shall be refined. We will use a Sentinel-1/Sentinel-2 (Sen1/Sen2) and Landsat 8/9 (L8/L9) data cube (2017 onwards, for all sites) to (i) derive time series and spectral-temporal metrics, (ii) create machine learning-based maps of annual (peat-)land cover classes and percent cover fractions and (iii) quantify the intra-annual dynamics of green and dead vegetation components (e.g. for reed). This will allow to describe abundance and vitality of vegetation on the WETSCAPES2.0 research sites and beyond at high spatial and temporal resolution, to be used in B1-6, C1-3, C5, S1-4. By integrating hyperspectral spaceborne data from EnMAP and PRISMA for the Core Sites and the L-LExps, we will achieve an even higher level of thematic detail and accuracy, and derive spatially and temporally resolved biomass estimates that will provide new opportunities for several other projects of the CRC, e.g. A1, B1, B3, B4, B6, C5, S3-4. Results from C4 will function as baseline products for the historical analyses in S3. Our work is based on reference data from very high-resolution satellite images, airborne and drone campaigns and high-frequency field spectroscopy as well as field data especially from A1, B1, B2, B4, B6, C2, C3.