We parameterized and applied a deterministic salmon production model to infer the degree to which river flows and temperatures may limit freshwater production potential of the Klamath River in California. Specific parameter requirements, data sources, and significant assumptions are discussed in detail. Model simulations covered a wide variety of historical hydrologic and meteorologic conditions for 40+ years of environmental data.
The model was calibrated only qualitatively, appearing to perform well in predicted outmigrant timing, but overestimating growth. Egg-to-outmigrant survival was near that reported for other rivers north of the Klamath River.
Predicted production potential appeared to be determined by multiple causes involving both regularly occurring habitat-related constraints and irregularly occurring exposure to high water temperatures. Simulated production was greatest in years of intermediate water availability and was constrained in both dry and wet years, but for different reasons. Reducing mortality associated with limitations to juvenile habitat, if possible, would be expected to have the highest payoff in increasing production. Water temperature was important in determining predicted production in some years but overall was not predicted to be as important as physical microhabitat. No single mortality cause acted as a true “bottleneck” on production.
Model uncertainty is addressed through a sensitivity analysis. Predicted habitat area may be a large source of model uncertainty and sensitivity, but collectively, model parameters associated with timing of events (for example spawning, fry emergence, and emigration) or related triggers control much of the model sensitivity.
Though model uncertainty remains, one can begin to explore potential alternatives to reduce production limitations. Specific recommendations are made regarding future study and reducing uncertainty.
