Protecting Spring-Run Chinook Salmon: The Use of a Two-Dimensional Water Temperature Model to Evaluate Alternative Hydroelectric

Waterpower XVI

The relationship between stream flow and water temperature in the waters affected by Pacific Gas and Electric Company’s (PG&E) DeSabla-Centerville Hydroelectric Project, FERC No. 803, (Project) was an important relicensing issue. The Project diverts cool water from the West Branch of the Feather River (WBFR) into Butte Creek, a stream that supports the largest population of spring-run Chinook salmon in California. Annually, PG&E and resource agencies develop a coordinated plan to maximize the cool water benefits in Butte Creek through changes in Project operations (e.g., timing and magnitude of releases from Project reservoirs). The development of a predictive stream temperature model, focusing on summer months (June through September) when high water temperatures can be a limiting factor, improved our ability to manage stream temperatures. The Army Corp of Engineers CE-QUAL-W2.v.3.2 (W2) was used for modeling the portion of the study area that is operationally adjusted to control temperatures in spring-run Chinook salmon summer holding habitat. W2 is a two-dimensional, laterally averaged, hydrodynamic, water temperature and water quality model. It was successfully used in this high gradient stream system by producing a hydro-dynamically equivalent model of the plunge-pool topographies typical of most of our study area. The model accommodated multiple waterbodies representing reservoirs and streams, multiple inflows and outflows, time-varying boundary conditions, and layer/segment addition and subtraction. W2 is a finite difference equation model that can compute water temperatures at sub-minute time intervals; as such it effectively modeled daily variations in water temperatures (i.e., daily minimums, means and maximums). We calibrated W2 with two years (2004, 2005) of summer water temperature data from stream, canal and reservoir locations, travel time (from dye studies) at four stream locations, and stream wetted width versus flow relationships from 22 instream flow transects. Error statistics indicated model calibration was successful in producing good to excellent performance across all metrics. Simulations were conducted using 2005 hydrology (above normal) and meteorology (hot) and 2001 hydrology (dry) combined with the hot 2005 meteorology. To investigate the effects of alternate operational scenarios on summer-time Butte Creek water temperatures a total of 32 simulations were conducted. Effectiveness of an alternative to manage water temperatures was determined by comparing a base case (reflecting current operations) with the alternative. Simulation results were then used by relicensing participants to develop instream flow recommendations. 


Ed Cheslak, Chris J. Berger, Robert Annear, Scott Wells