Monitoring protocol for Sacramento River winter chinook salmon, Oncorhynchus tshawytscha: application of statistical power analysis to recovery of an endangered species


Lindley, Steven T., Michael S. Mohr, and Michael H. Prager
Cover date: 

When monitoring endangered species, natural resource managers require a recovery benchmark and a statistical procedure to test whether the benchmark has been met. We applied statistical power analysis to devise such a procedure for the endangered Sacramento River winter chinook salmon (Oncorhynchus tshawytscha). Winter chinook salmon management currently focuses on population growth rate, and our procedure used a Student’s t-test to evaluate whether the average population growth rate is significantly lower than the management goal of 0.57 per generation. In the test, the null hypothesis was that the growth rate was not lower than the desired rate. In contrast to the usual hypothesis-testing framework, our procedure did not control for the type-I error rate. Instead, it controlled for the statistical power (the complement of the type-II error rate) and uses the resulting type-I error rate, computed from the sample size and other information, for the test. This procedure is conservative for winter chinook salmon in that, if all assumptions are met, it provides the specified level of assurance of detecting dangerously low population growth rates.