Quantifying the Uncertainty of a Juvenile Chinook Salmon Race Identification Method for a Mixed‐Race Stock

Expected daily FL ranges (length at date) of juvenile Chinook Salmon Oncorhynchus tshawytscha have been used throughout California's Central Valley to identify federally listed winter‐run and spring‐run juveniles in a mixed four‐race stock. Accurate race identification is critical both to species recovery and to management of the water supply for 25 million people and a multibillion‐dollar agricultural industry. We used genetic race assignment of 11,609 juveniles sampled over 6 years to characterize the accuracy of the length‐at‐date approach, specifically by testing two of its central assumptions: (1) juvenile FL distributions do not overlap between races on a daily basis; and (2) the growth rates that are used to project FL at date are accurate. We found that 49% of FLs for genetically identified juveniles occurred outside the expected length‐at‐date ranges for their respective races, and we observed a high degree of overlap in FL ranges among the four races. In addition, empirical growth rates were well below those from which length‐at‐date criteria were derived. Given the high degree of FL overlap between races, we conclude that modification of the length‐at‐date method will not substantially reduce identification error. Thus, we recommend that genetic assignment be used at least as a supplemental approach to improve Central Valley Chinook Salmon race identification, research, and management. Received January 7, 2014; accepted July 22, 2014