Journal of Advances in Modeling Earth Systems Global and Planetary Change
Comparative Evaluation of Low‐Flow Estimation Performance in Noah‐MP for Improved Hydrological Modeling Across the United States
Abstract
Abstract Accurately simulating low flows is essential for effective water resource management and ecosystem conservation. This study evaluates five runoff and groundwater configurations in the Noah‐MP land surface model—SIMGM, SIMTOP, Schaake96, BATS, and the newly integrated Groundwater for Ungauged Basins (GrUB)—across 543 minimally disturbed reference basins in the contiguous United States. GrUB represents groundwater discharge using a representative‐hillslope formulation with independently derived watershed‐scale parameters and no basin‐specific streamflow calibration. Model performance was assessed using Kling‐Gupta Efficiency (KGE), Nash‐Sutcliffe Efficiency (NSE), Mean Absolute Error (MAE), and bias under general and low‐flow conditions, identified by the 80% exceedance flows () and Tennant method (). Results indicate no configuration performed best across all metrics. During low‐flow periods, Schaake96 consistently produced the lowest MAE and bias, while GrUB generally ranked second and outperformed SIMGM, SIMTOP, and BATS under both low‐flow definitions. Over the full simulation period, GrUB showed comparatively strong overall performance and the lowest median bias. Spatially, the five configurations generally overestimated low flows in many central and eastern basins, whereas localized underestimation occurred in portions of the Midwest and western United States. Overall, Schaake96 showed the strongest low‐flow performance, while GrUB provided relatively small low‐flow errors and balanced overall performance using independently derived parameters without basin‐specific streamflow calibration. These results highlight substantial differences among Noah‐MP runoff and groundwater configurations and their representation of low‐flow processes.