Evaluating Long-Term One-Way Atmosphere-Hydrology Simulations and the Impacts of Two-Way Coupling in Four Mountain Watersheds

Document Type

Article

Publication Date

5-2022

Abstract

Joint hydrologic-atmospheric model frameworks offer novel insights into the terrestrial hydrologic cycle and the potential for improved predictive capabilities for stream discharge and other hydrologic fluxes. In this study, we examine both one- and two-way coupled integrations of the Weather Research and Forecasting (WRF v3.8.1) atmospheric and WRF-Hydro (v5.0) hydrologic models for four 1000–2000 km2 snow-dominated mountain watersheds (1500–2100 m mean elevation) in Idaho's Rocky Mountains. In watersheds where anthropogenic withdrawals are minimal (3 of 4 watersheds), we simulate stream discharge with high confidence (KGE > 0.63) for a 20 year period in the uncoupled scenarios, and find that WRF winter precipitation accumulations have less than 15% average error for all but two of the fourteen comparison NRCS Snotel sites. However, annual streamflow biases are highly correlated (r2 > 0.8 in some cases) with the annual errors in WRF cold-season precipitation, suggesting that process representation of winter orographic precipitation limits hydrologic predictability. In the second part of the study, we evaluate the potential for ‘two-way’ model coupling to influence hydrologic predictability by examining a 2 month case-study period with active spring season convective precipitation. We quantify the impacts of resolving hillslope-scale soil water redistribution on the ABL, and find that while resolving overland and saturated subsurface soil moisture flow influences soil moisture distributions and surface energy fluxes, the impact on precipitation is non-systematic, as precipitation is generally atmospherically controlled during the study period. Consequently, future efforts should focus on improving winter orographic process representation, as streamflow is highly sensitive to errors in these processes.

Link to Full Text

Share

COinS