Historical Streamflow Reanalysis at the National Scale using Hierarchical Routing and Data Assimilation

Abstract:

Historical streamflow records are generally limited to relatively few observational records or require data-intensive models for individual streams. Process-based hydrologic models may be generalized and applied to areas with similar hydrologic characteristics; however, model performance can be diminished in areas with significant flow regulation (i.e. dams, diversions, etc.). To address the practical limitations of existing data and modeling approaches, we present a national dataset using spatially distributed runoff information, a hierarchical Routing Application for Parallel computation of Discharge (RAPID) river routing scheme, and data assimilation with around 3000 US Geological Survey (USGS) National Weather Information System (NWIS) stations to produce complete 1981–2015 historical records for the 2.7 million Conterminous US reaches in the medium resolution National Hydrography Dataset Plus V2 (NHDPlusV2). Results of this national-scale modeling effort are compared to other streamflow estimation methods (e.g. Vogel and EROM), with particular interest in improving the representation of streamflow downstream of points of regulation. We further combine information from the NHDPlusV2 and USGS Gages II datasets, and infrastructure datasets like the National Inventory of Dams to identify areas and characteristics where there are the greatest improvements over other methods, and where the greatest deficiencies remain. This dataset is especially relevant for large-scale applications in the water-energy nexus and may help users reduce the amount of effort required to process historic observations or develop assessment-specific models.