WRR Publication

A new publication in the journal Water Resources Research (Tiwari and Mascaro, 2026) provides a statistical characterization of high flow volumes across the United States for an assessment of managed aquifer recharge projects. The approach provide a consistent foundation for screening and siting locations where floodwaters are used to recharge aquifers. Congratulations to the authors associated with the Center for Hydrologic Innovations!

• Abstract: Managed aquifer recharge using floodwater (Flood-MAR) is a promising strategy to mitigate groundwater depletion. Its effective implementation depends on reliable information on the availability and frequency of high flows. However, outside of California, the United States lacks a nationwide assessment of high-flow volumes (HFV) suitable for Flood-MAR planning, particularly at ungaged locations. Here, we present a continental-scale statistical characterization of HFV frequency at the outlets of 2,115 subbasins across the conterminous United States (CONUS). HFV is defined as the volume exceeding the 80th, 90th, and 95th quantiles of daily streamflow, thresholds commonly adopted to satisfy environmental flow requirements. We estimate daily streamflow time series, Q(t), using four methods for prediction in ungaged basins (PUB), extract independent HFV events, and model their frequency using the generalized extreme value distribution. Leave-one-out cross-validation at 701 stream gages shows that the accuracy of estimated HFV depends on the ability of the PUB methods to capture over-threshold flows and the serial correlation structure of Q(t). HFV quantiles for Flood-MAR-relevant return periods are estimated as weighted averages of the PUB methods, along with their associated uncertainties. Results indicate that HFV availability is highest and uncertainty is lowest in wetter regions, including northern California, the Pacific Northwest, and portions of the Lower Mississippi Basin, reflecting greater potential for Flood-MAR. In contrast, arid and semiarid basins across the central and southwestern United States exhibit lower HFV availability and higher uncertainty. These results provide a consistent, CONUS-wide foundation for first-level Flood-MAR siting and screening.