A new multi-source dataset describing surface water regimes across North American wetlandscapes

Surface water plays a critical role in wetland ecosystem functioning, but can be difficult to track over time due to challenges brought by vegetation canopy cover, temporal data gaps, or other limitations. Here, we describe the development of a new data product called "Surface Water Regimes (SWaRe)", which aims to provide a synoptic characterization of seasonal hydrologic profiles across North American wetlandscapes. SWaRe is derived from time series imagery from optical Landsat-8 and Sentinel-2 as well as synthetic aperture radar (SAR) Sentinel-1 sensors, and is being produced using a suite of automated surface water detection algorithms. Sub-pixel estimates of surface water extent from the optical data sets allow for the monitoring of hydrologic dynamics within very small or heterogeneous wetlands. On the other hand, automated detection of surface water from SAR images offers the potential for a consistent record of inundation, independent of cloud-cover or time of satellite overpass. With these datasets, we generated consistent, gap-free composites of surface water extent at temporal intervals approaching daily resolution. We show that by leveraging the spectral and temporal resolutions of the optical and SAR data streams, respectively, SWaRe is capable of capturing both seasonal inundation dynamics as well as episodic flooding in dynamic and heterogenous wetlandscapes. SWaRe is expected to support ongoing initiatives to update national-scale wetland inventories as well as research in wetland hydrology and biogeochemistry.