As Utah pursues an integrated Great Salt Lake (GSL) Basin planning strategy, estimates of environmental water demands are needed not only at gaged sites but across the many ungaged and non-reference stream reaches where flow targets may be set. This report provides this information by estimating reference-condition functional flows throughout the GSL Basin stream network. Functional flows are discrete seasonal components of the annual hydrograph (winter low flow, high flow ascension, peak, and recession, and summer low flow) with documented ecological, geomorphic, and/or biogeochemical importance. Lane et al. (2026) proposed a functional flows framework for GSL Basin’s rivers and a ‘calculator’ for quantifying the annual functional flow metrics at reference-condition streamgages, incorporating input from stakeholders representing a diverse array of federal, state, and local organizations and agencies. This report applies the functional flows conceptual model and calculator to generate natural functional flow predictions across the basin. Collectively, the functional flows framework, calculator, natural functional flow predictions, and associated tools are intended to help maximize the ecosystem benefits of water delivered as part of Great Salt Lake recovery efforts.

In this study, we applied a machine-learning approach to predict the interannual ranges of 18 functional flow metrics across the GSL basin stream network. Random forest models were trained on 2,540 gage-years of streamflows from 180 minimally disturbed reference gages across Utah, the Bear River Basin, and the Middle Rockies ecoregion, using publicly available geospatial and climate predictors. Model performance was evaluated with leave-one-out cross-validation, and a regionalized quantile-mapping correction was applied to address the range-compression bias inherent to ensemble tree models. Seventeen of the 18 flow metrics achieved good or better performance, with the models relying on physically meaningful predictors— groundwater recharge for magnitude metrics, air temperature for timing, and baseflow index for rates of change—lending confidence to the approach. The resulting functional flow metric predictions span ~2,500 stream segments (a 4,561-km network) and are delivered in a publicly accessible geodatabase.

These natural functional flow predictions represent the seasonal and interannual streamflow ranges that would have supported native river ecosystems in the absence of major impairments. The predictions provide the best estimate of natural flows for GSL Basin streams available at this time. They also serve as consistent, broadly applicable environmental flow targets that can help characterize and quantify how much water is needed for water quality and environment – one of many elements being considered as part of the GSL Basin Integrated Plan. However, specific water management decisions remain with local resource managers and should incorporate site-specific refinement of flow targets where data/resources permit. More broadly, the ecological management goals identified by stakeholders through the functional flows framework process can help with setting and modeling water resource system performance measures (WRe and USBR 2024).