Detecting juvenile survival effects of habitat actions:: power analysis applied to endangered Snake River spring-summer chinook (Oncorhynchus tshawytscha)

Using 10 years of parr-to-smolt survival rate estimates for passive integrated transponder (PIT) tagged, endangered wild Snake River chinook (Oncorhynchus tshawytscha) (measured in eight spawning streams), we demonstrate that moderate increases in base-case survival may be detectable quickly, as required by recent U.S. National Marine Fisheries Service regulations. The regulations require that effects of tributary habitat actions on juvenile salmonid survival rates be detectable within 5 to 8 years. Simple log-linear regression models were employed where the natural log of survival is a function of parr size, parent stock abundance, tagging location, and year. The analysis uses before-after control-impact (BACI) statistical techniques. The results suggest that multiplicative survival rate changes of 30% should be detectable within 7 years, and increases of 50% within 3 years. Models with higher information-theoretic weights were more powerful than less plausible models. Models using juvenile survival were substantially more powerful than models using spawner-recruit data for the same stocks, but the analysis is no substitute for field studies, which are still very rare.