A model-based approach to assessing the potential response of chinook salmon to habitat improvements

The recovery strategy for Snake River salmon relies on improvements to freshwater habitat; however, the benefit resulting from such actions is unknown. To address this, I modeled survival as a function of habitat variables, and predicted survival under alternative future habitat states. Predictions suggest that the potential for improving survival through restoration is high for some populations and low for others, while the potential for reducing survival by further impairing habitat is great for all. Additionally, I addressed a basic question relevant to the current recovery strategy: what constitutes good habitat? Based on spawning site selection patterns, I created models relating site use to habitat for two populations. Patterns were similar for both populations; however, models had high stream specificity with regards to depth, causing models to transfer poorly between populations. These results provide information on habitat selection with implications for others using a similar approach to assess habitat suitability.