Presenter: Todd Vandegrift — Precision Water Resources Engineering

Natural and naturalized flows are among the most important inputs to many RiverWare models, especially those that simulate complex basin accounting and water-right administration. These flows define the base water supply a basin would have produced without the impacts of storage/regulation, diversions/consumptive uses/return flows, and imports/exports. Developing appropriate natural flow datasets can be a challenging and sometimes underappreciated modeling problem in its own right, especially when available data is incomplete, inconsistent, or influenced by changing water uses and operations. The CUWCD Natural Flow Model was developed to support the Central Utah Water Conservancy District's CUPSIM and other RiverWare models and hydrologic analyses by calculating naturalized flows for the Provo, Spanish Fork, and Duchesne River basins.

The model employs RiverWare in an intentionally non-traditional manner. Rather than requiring the entire network to solve over the full run period, the model utilizes a RiverWare network to organize upstream data dependencies and uses object methods and dispatching only for specific objects, most notably reservoirs. The model's ruleset consists of a single generalized rule that executes just once, at the end of the final timestep, to calculate naturalized flows at each stream gage or reservoir location and only for periods when all necessary upstream data is available. The network therefore serves as both a calculation framework and a visual description of what is and isn't explicitly accounted for in the resulting naturalized flows.

The calculated naturalized flow records are then filled and extended by a post-processing script outside of RiverWare. The script applies MOVE-type regressions between well-correlated naturalized flow records, taking advantage of the improved relationships that emerge after flows have been naturalized, and steps through gages and missing periods one by one to make the best estimates possible from available data. These methods help build a complete 121-year hydrologic dataset that provides a sound basis for CUWCD's RiverWare modeling, planning studies, and hydrologic analyses.