Uncertainty in the expected responses of aquatic ecosystems can be a barrier to environmental flow implementation. Robust assessments of flows when they occur can reduce this uncertainty and inform future flow implementation. True control river segments are usually not available, and many assessments rely on before-after designs that can lead to erroneous inference if replication is low, important environmental drivers cannot be controlled, or experimental flows are not prescribed randomly. Model-based approaches provide an alternative to design-based approaches to analysis. Specifically, models that consider both controlled and uncontrolled factors can be used to forecast responses under both the observed conditions and the expected conditions in the absence of flows (i.e., a counterfactual).
Here we assess the impacts of “cool mix” flows released from Lake Powell through Glen Canyon Dam into the Grand Canyon segment of the Colorado River during 2024 using counterfactual forecasting analyses. Cool mix flows involved releasing a portion of the daily flow through river outlet works located 100 feet deeper in the reservoir than the hydropower generating tubes from which water is typically released. Cool mix was designed to minimize spawning and somatic growth of Smallmouth Bass (Micropterus dolomieu: SMB). SMB were first observed reproducing below Glen Canyon Dam in 2022 coincident with the lowest reservoir elevations and warmest release temperatures in five decades. Cool mix was expected to increase phosphorous in addition to cooling water temperatures. Cool mix was expected to increase the somatic growth of Rainbow Trout (Oncorhynchus mykiss) as increasing phosphorous should increase tailwater ecosystem productivity and cooling temperatures favored this cold-water species. Cool mix was also expected to decrease the growth of Humpback Chub (Gila cypha) as they prefer warmer temperatures. Observed water quality, SMB catch, and somatic growth for all species was consistent with forecasts under a variety of assumptions, however uncertainty in flow effects varied.