Dynamics of global wetlands reflected in the daily time series product Global WaterPack

Kurzfassung

Wetlands are rich ecosystems and essential for human life. They provide wildlife habitats, are crucial to agriculture and fishery, purify and store water, and maintain its availability during dry periods. Key functions and services of wetlands are closely related to the temporal variations of water levels and inundation cycles. Daily time series of optical medium resolution earth observation data with global coverage are now available for more than 15 years, and open the possibility for long-term monitoring of wetland ecosystems. The Global Water Pack is a global and daily 250 m time series product developed at DLR that captures inland waterbody extents, their seasonal dynamics, and their multi-annual alterations. The main input data for the Global Water Pack is daily spectral information of MODIS (Moderate Resolution Spectroradiometer), complemented by auxiliary data including a global DEM and selected thematic MODIS products. Water is first classified on a daily basis using a dynamic thresholding approach. In a subsequent time series analysis, cloud gaps within the resulting daily water masks are reduced. The dataset has been validated based on more than 300 Landsat-based reference classifications revealing an accuracy of more than 90% for pixels of more than 75% sub-pixel water coverage. The Global Water Pack provides crucial information for global wetland ecosystems reflecting the dynamics of open water surfaces. Using examples of major wetlands in Africa, Asia and America, the potential of the Global Water Pack to capture relevant parameters such as inundation frequency and duration, timing of flooding and water retreat, as well as freezing and thawing cycles, is presented. Results are compared with high resolution spatial reference data and in-situ measurements, and the application potential of the Global Water Pack time series for monitoring and assessing wetland ecosystem services and changes are discussed.