Qiang Li, Sam Zipper , Tom Gleeson

Groundwater pumping can reduce streamflow by reducing groundwater discharge and/or inducing streamflow infiltration, which together are referred to as streamflow depletion. Recently, analytical depletion functions (ADFs) have been suggested as rapid and accurate tools for streamflow depletion assessment, but their performance has only been tested in a few hydrogeological settings. To evaluate whether they will be useful tools for other regions with contrasting stream network and hydrogeological characteristics, we compared ADFs to calibrated MODFLOW models in BX Creek and Peace Region with distinct hydrogeological settings (interior plateaus & highlands, and boreal plains, respectively) and spatial scales (165 km2 and 1952 km2, respectively) in British Columbia, Canada. Results showed that ADFs can accurately identify streams most affected by pumping for 100% and 83% of wells in the BX Creek and Peace Region, respectively, and had small prediction errors compared with MODFLOW. Specifically, the mean absolute error of predicted depletion ranged from 2% to 14% of the highest simulated pumping rate over the study period of 30 years, with improved accuracy during the pumping season. We also found contrasting responses of ADF performance to hydrostratigraphic properties such as hydraulic conductivity, aquifer thickness, streambed conductance, and well depth across two domains, indicating that different drivers control ADF accuracy in different hydrogeological settings. Therefore, we conclude that ADFs are useful tools for conjunctive water management, but a good understanding of local hydrogeological conditions is needed to address the potential uncertainty of ADFs for decision-making.