Overwinter survival and growth of juvenile coho salmon, Oncorhynchus kisutch, in Freshwater Creek, California

Coho salmon Oncorhynchus kisutch are an anadromous species that typically spend their first winter in freshwater habitats before outmigrating to sea the following spring as smolts. Coho salmon in California are a state and federally listed species, requiring efforts to monitor populations and estimate key demographic rates. Estimates of juvenile survival through the winter and the total abundance of smolts are an important component of this population monitoring. These estimates are typically based on smolt trapping during spring smolt outmigration and do not account for an unknown proportion of the population that emigrates from natal stream habitat at other times of the year. I used a mark-recapture study to estimate (1) apparent overwinter survival, or the proportion of fall-marked individuals leaving during typical spring smolt outmigrant sampling, and (2) the probability of early emigration, or the proportion of fall-marked individuals leaving prior to the initiation of outmigrant sampling. Sampling was conducted in a small coastal stream in northern California over two winters from 2010-2012. "Apparent" overwinter survival estimates for six different reaches ranged from 13.2- 49.3% in 2010-2011 and 10.5- 19.5% in 2011-2012; however, I estimated that 2-25% of juvenile coho salmon emigrated early from each reach in the first year and 6-27% in the second year. The largest observed emigration rates were in the main stem reaches in both years. Further analysis revealed that survival and emigration behavior depend on characteristics of the individual and its location and habitat at the time of marking. The best-supported models in the survival analysis included individual covariates for watershed area (an index of flow, stream width, and gradient) of fall sampling locations and fork length in the fall, suggesting that both size and location within the watershed are important determinants of survival. The best supported models in the early emigration analysis included group covariates based on the reach in which the individuals were first encountered in the fall. Watershed area and fork length were also important for predicting early emigration in the 2011-2012 winter period.