Active adaptive management for native fish conservation in the Grand Canyon: Implementation and evaluation

My first objective was to evaluate the efficacy of a large scale non-native fish removal effort to benefit endemic fishes of the Colorado River within Grand Canyon. During 2003--2006, over 23,000 non-native fish, primarily rainbow trout Oncorhynchus mykiss, were removed from a 9.4 mile reach of the Colorado River. These removals resulted in a rapid shift in fish community composition from one dominated by cold water salmonids (>90%), to one dominated by native fishes and the non-native fathead minnow Pimephales promelas (>90%). Concurrent with the mechanical removal, data collected within a control reach of the river suggested a systemic decline in rainbow trout unrelated to the fish removal effort. Thus, the efficacy of the mechanical removal was aided by an external systemic decline, particularly in 2005--2006.My second objective was to improve current knowledge of humpback chub Gila cypha growth to aid in length-based age determination, and to provide a tool to evaluate temperature-dependent changes in growth rate. I estimated a temperature-dependent growth function for humpback chub by predicting more than 14,000 growth increments from a mark-recapture database. Results suggest that humpback chub growth is strongly dependent on temperature and that previous growth curves based on paired age-length data tend to over-estimate the age of small fish and under-estimate the age of large fish.My third objective was to update humpback chub stock assessment procedures following guidance from an external review panel. These recommendations were primarily to develop model selection procedures and to evaluate the effect of error in length-based age determination. I used both Pearson residual analysis and Akaike Information Criterion to evaluate candidate models---leading to the conclusion that the most general assessment model was required to adequately model patterns in capture probability. I used the temperature-dependent growth relationship to estimate probabilistic relationships between age and length. These age-length relationships were then used in Monte Carlo simulations to capture the effect of ageing error on subsequent estimates of recruitment and adult abundance. The results indicate that the adult humpback chub population has likely increased between 20--25% since 2001.My fourth objective was to evaluate whether there was any evidence of effect from past adaptive management actions or uncontrollable factors on Grand Canyon fish populations, and to make recommendations for further adaptive management program development. These results are largely inconclusive except that the combined policy of mechanical removal and increased water temperatures is temporally correlated with increased native fish abundance in the mainstem Colorado River near the confluence of the Little Colorado River, a reach deemed critical habitat for humpback chub. I recommend that the adaptive management program invest additional effort in developing more explicit and measurable resource goals, particularly for focal Colorado River resources. I further recommend that additional investment in monitoring of juvenile native fish survival and growth in the mainstem is needed to adequately evaluate future adaptive management experiments. Finally, additional predictive capability is needed to both formalize a priori hypotheses about juvenile native fish survival and recruitment, and to screen future policy options.