Insights from Dayflow: A Spatiotemporally Continuous Historical Streamflow Reanalysis Dataset for the Conterminous United States
Abstract
Accurate historical streamflow data can support a broad spectrum of water resource applications. While gauge observations provide the most reliable estimates, they are often limited in space and time resulting in a significant data gap in ungauged basins. Process-based hydrologic models can be data/computation-intensive, and statistics-based models can be region/stream-specific, posing challenges for historical streamflow reconstructions. In this study, we implement a scalable modeling framework that integrates simulated runoff from the Variable Infiltration Capacity (VIC) model and observed streamflow from selected US Geological Survey (USGS) gauge stations using the Routing Application for Parallel computatIon of Discharge (RAPID) routing model for efficient data assimilation. The outcome of the modeling effort is called Dayflow (https://hydrosource.ornl.gov/dataset/dayflow-V1) - a 36-year (1980-2015) reconstructed daily streamflow dataset for ~2.7 million NHDPlusV2 stream reaches across the CONUS. A comprehensive evaluation of Dayflow shows that Kling-Gupta Efficiency (KGE) > 0.5 at 56% and bias within ± 20% at 63% of the 7,526 USGS gauge locations. Hierarchical streamflow assimilation shows an overall improvement, notably in the western semiarid-to-arid regions. Comparisons to other streamflow reanalysis datasets at the national scale from National Water Model and the global scale from GRADES (https://www.reachhydro.org/home/records/grades) show improved KGE and reduced bias. Investigations of relationships among streamflow prediction errors and key hydrologic, hydroclimatic, and geomorphologic basin characteristics reveal some strong region-specific patterns, which may help advance future modeling efforts. Insights from Dayflow might enable a better understanding of hydrologic conditions in a changing environment in locations not represented by current streamflow monitoring networks.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.H43B..07G