Paul Bates (Univeristy of Bristol, United Kingdom); Paper Team The SWOT Global Hydrology Assessment (28 authors, from around the world)

Basic global river knowledge is modelled, not measured, and models calibrated on <1% of global rivers reify into freshwater knowledge past, present, and future. We use the Surface Water Ocean Topography (SWOT) satellite to assess our understanding of the global river system represented by six simulations of global rivers [including both ML and physics-based]. In rivers with enough sufficient high-quality SWOT data, representing 33% of global flow (8,200 km3/yr) and 19% of global rivers, SWOT suggests that up to 5,400km3/yr of modelled flows may be in error. Attribution of extreme errors suggests models struggle with dams, Arctic Freshets, and wetland systems. A SWOT proxy map of likely global model skill resembles a map of a human hydrologic activity, suggesting adding to suggestions that current basic hydrology knowledge does not reflect humanity’s role in the water cycle. Finally, we prove SWOT measurements can correct model errors as anticipated, and ultimately SWOT’s primary data suggest ‘naturalized’ models ignoring human action can should now be abandoned. As SWOT continuously improves, future models, and therefore future basic global knowledge, will improve.