Using dynamic state variable models to predict steelhead life history in four California watersheds

Background/Question/Methods Steelhead/rainbow trout (Oncorhynchus mykiss) are a falcutatively anadromous, iterporarous species with a highly variable life history. This highly plastic life history presents management challenges, particularly with regards to conserving life history diversity and predicting the demographic impacts of altered life histories. I used a state-dependent life history model to predict the effects of changes in the juvenile growth environment on steelhead life histories, and applied this model to data collected in four California watersheds, two large and heavily modified rivers in the Central Valley and two small and less impacted coastal creeks. I characterized optimal state-dependent strategies as a function of fish size and growth rate during decision windows preceding the actual smolting and spawning events, allowing the identification of threshold sizes and growth rates at which fish are predicted to adopt alternate life history pathways. Using these thresholds, I predicted the strategies taken by individual fish growing along varying trajectories under current conditions, and predicted short term responses to these old rules and long term responses to newly evolved rules in altered environments.

Results/Conclusions For the coastal creeks, the model correctly predicted that anadromy would predominate, with most fish smolting at older ages. Residency was correctly predicted to be very rare except above waterfalls. Predictions for the Central Valley rivers had to be integrated over large uncertainty in survival rates. Still, the model was robust in its predictions for the American River, predicting that anadromous fish smolting at young ages would predominate for almost all plausible parameter estimates. Predictions for the Mokelumne River were much more sensitive to survival estimates, but included correct predictions of a mix of residency and anadromy for plausible inputs. Over the short-term, I predict that plausible changes in growth rates are likely to change the balance of smolt ages but not the residency/anadromy balance on the coastal rivers, while they might more readily change the residency/anadromy balance on the Mokelumne while having little effect on life histories in the American River, where young smolts are still predicted to predominate. Over the long-term, I predict that maintaining passage for emigrating and returning fish is most important for the preservation of anadromy. One important message from this modeling is that the interactions between growth rate and marine and freshwater survival are complex, such that life histories should not be expected to consistently respond to changes in any one factor in a predictable way.