Are emerging Earth Observations inconsistent in mapping Global Surface Water Extents?

Surface waters play an important role in the hydrologic cycle by providing storage to intercept fluxes of water. Recent advances in Earth observation datasets have helped remotely monitor the dynamic behavior of surface waters and capture their extents with greater spatial detail. However, large variability exists in the way today's Global Surface Water Extent (GSWE) datasets represent these surface waters. We mapped and evaluated the variations in the number and extent of surface waters across five recent GSWE datasets, including European Space Agency Worldcover product (ESA., 2022); HydroLAKES (Messager et al., 2016); Joint Research Centre Global Surface Water Mapping Layers (Pekel et al., 2016); National Elevation Dataset-based (Rajib et al., 2020, Wu et al., 2019); OpenStreetMap Surface Water Layers (Yamazaki et al., 2019) as well as one regional, U.S.-only dataset LAGOS (Cheruvelil et al., 2021). Our findings suggest that the inconsistencies across the selected datasets stem mainly from (i) the definition of surface waters, (ii) data format (vector vs. raster), and (iii) data sources and methods used for waterbody delineation in each dataset. This study will provide novel spatial evidence to understand discrepancies and uncertainties among different GSWE datasets, allowing further research into these differences and providing evidence to choose a suitable dataset for management and policy-related decisions.