Examining the premise of the "LakeFlow" discharge algorithm in the contiguous United States

Abstract

Lakes and reservoirs are essential storage units of the global drainage system. These stores influence downstream river discharge through detaining and releasing water mass at their interface. LakeFlow is a proposed algorithm to estimate discharge at the lake-river interface for the forthcoming Surface Water and Ocean Topography satellite mission (SWOT). With the direct measurements provided by SWOT, LakeFlow bridges the storage change to an improved parameterization of the inflow and outflow channel properties that are not observable to SWOT but necessary for discharge calculations. The success of LakeFlow relies on a wide existence of the premise, which is that the discharge difference between the primary inflow and outflow river reaches is the first-order control on storage variation. However, additional hydrologic controls and anthropogenic factors, which are not directly observable to SWOT, could also be substantial and therefore undermine the LakeFlow’s premise. For this reason, we have examined the premise conditions for 26 reservoirs in different hydroclimate conditions across the contiguous United States (CONUS), and further constrained the uncertainties by accounting for lateral flow and surface evaporation through a synergy of multi-source remote sensing and modeling datasets. When comparing reservoir storage change with discharge difference between the reservoir’s inflow and outflow river reaches for a two full year, 18 out of 26 reservoirs have a Nash-Sutcliffe efficiency (NSE) greater than 0, exceeding the mean flow benchmark. 12 out of the 26 reservoirs have an NSE greater than 0.5, indicating a satisfactory premise. The majority of these reservoirs with a satisfactory premise are located in less humid regions. By comparing reservoirs that were and were not designed for water supply, our results suggested that human water withdrawal is probably the most significant contributor to a less satisfactory premise. As we expect an overall lesser extent of human water consumption from natural lakes, our results based on reservoirs may serve as a minimum baseline for the actual feasibility of LakeFlow. Simulated lateral flow from the Global Reach-scale A Priori Discharge Estimates for SWOT (GRADES) demonstrates a high effectiveness in improving the premise condition, particularly in humid regions. The inclusion of surface evaporation estimates from the CONUS Reservoir Evaporation Dataset (CRED) dataset appears to be most effective in drier regions. Coupled with both surface evaporation and lateral flow datasets, the LakeFlow premise conditions are improved by a median value of 27.4% for reservoirs in more humid region, and 74.5% in reservoirs that were not built specifically for water supply. These results support the overall feasibility of LakeFlow and motivate continued testing on the accuracy of reservoir inflow/outflow estimations.