Updates to CE-QUAL-W2 models for select U.S. Army Corps of Engineers reservoirs in the Willamette Valley Project and an inter-reservoir reach of the Middle Fork Willamette River, northwestern Oregon

Abstract

Mechanistic models capable of simulating hydrodynamics and water temperature in rivers and reservoirs are valuable tools for investigating thermal conditions and their relation to dam operations and streamflow in river basins where upstream water storage and management decisions have an important influence on river reaches with threatened fish populations. In particular, models allow managers to investigate how new, untried operations or hypothetical structures might influence streamflow and temperature conditions downstream. CE-QUAL-W2 is a two-dimensional (laterally averaged) hydrodynamic water-quality model that has previously been used to investigate the downstream effects of dam operations and other anthropogenic influences on stream temperature in the Willamette River Basin in northwestern Oregon, a region with two populations of fish species designated as threatened under the Endangered Species Act. By linking CE-QUAL-W2 river models to models of upstream, large Willamette Valley Project dams and reservoirs, these models can be used to investigate how dam operations at individual dams can influence streamflow and thermal conditions in downstream river reaches as an integrated system. Integrated model simulations that include the large dams and reservoirs linked to downstream river reaches can help managers develop a better understanding of tradeoffs associated with potential retrofits or operational changes across the multipurpose dams in the Willamette Valley Project, the effect of dam management on downstream tributaries and the Willamette River, and the resulting potential effect on threatened fish populations and habitat conditions. River models capable of simulating river corridors downstream from U.S. Army Corps of Engineers dams were previously updated and integrated to simulate conditions that occurred from March through October of 2011 (a cool and wet year), 2015 (a hot and dry year), and 2016 (a moderately hot and dry year) using CE-QUAL-W2 version 4.2. These river models encompass the following: Coast Fork Willamette and Middle Fork Willamette Rivers, the Row River, and Fall Creek downstream from Cottage Grove, Dexter, Dorena, and Fall Creek Dams, respectively; South Fork McKenzie River downstream from Cougar Dam; McKenzie River downstream from its confluence with the South Fork McKenzie River; South Santiam River downstream from Foster Dam; North Santiam River downstream from Big Cliff Dam; and Willamette River from its start at the confluence of the Middle Fork Willamette and Coast Fork Willamette Rivers to Willamette Falls (river mile 26.0; near West Linn, Oregon). This report documents model modifications, boundary condition data sources or estimation methods, and goodness-of-fit statistics for six CE-QUAL-W2 reservoir models and one river model upstream from the existing river models. These models simulate (1) Hills Creek Lake; (2) Lookout Point Lake and Dexter Reservoir on the Middle Fork Willamette River; (3) the Middle Fork Willamette River reach between Hills Creek Dam upstream and Lookout Point Lake downstream; (4) Cougar Reservoir on the South Fork McKenzie River; (5) Green Peter Lake on the Middle Santiam River and Foster Lake on the South Santiam River; and (6) Detroit Lake and (7) Big Cliff Reservoir on the North Santiam River. These CE-QUAL-W2 models were built by a variety of researchers to simulate a range of conditions in past years; this report documents their upgrade to U.S. Geological Survey (USGS) edition 7 of version 4.2 of CE-QUAL-W2 and updates each model to simulate conditions from January through December of 2011, 2015, and 2016. Also included in this report is an explanation of modifications to the CE-QUAL-W2 source code that constitute USGS edition 7 of CE-QUAL-W2 version 4.2. Each of the models described in this report can be run in isolation or linked to downstream models as a “system model” to simulate conditions in tributaries and (or) in the Willamette Valley Project as a whole. As part of the model updates described in this report, some model parameters were adjusted to improve stability or decrease model error, and boundary conditions including meteorological, hydrologic, and temperature inputs were developed and updated for model years 2011, 2015, and 2016, as necessary. In some cases, the data sources used to drive previous model versions were no longer available, which required the development and checking of new data sources or estimation techniques. Goodness-of-fit statistics for outflow from the dams and in simulated river reaches generally show a good model fit, with the models simulating subdaily water temperatures at most comparable locations with a mean absolute error of generally less than 1 degree Celsius (°C) and a reasonably low bias. Model simulation of the thermal vertical profiles in each reservoir also produced an overall mean absolute error of generally less than 1 °C for all 3 years, with the exception of the Hills Creek Lake Model and the Cougar Reservoir Model in years when the reservoirs did not fill (2015 and 2016). Both of these models have known calibration issues and tend to be sensitive to the choice of certain structural parameters in the model. Overall, the calibration process was focused on obtaining model settings that led to realistic water temperature predictions in all 3 years (2011, 2015, and 2016) without over-calibrating specifically to any single year. A complete investigation of model error for these reservoir submodels was beyond the scope of this investigation but could be undertaken in the future if better model performance for these two reservoirs is desired.