Estimating missing age distribution data for Chinook salmon for use in run reconstruction and estimation of productivity

Chinook salmon from a single cohort return over multiple years to spawn.  Thus age composition for spawning returns is important in conducting run reconstruction which is needed to determine changing productivity of this ESA listed species from the Pacific Northwest.  Age data has not been collected annually over the time span to be analyzed so a method is needed to fill in the missing data.  Neither an average annual nor cohort age distribution is satisfactory since abundance varies greatly from year to year making neither of these options viable.  A method has been developed to weight an average age distribution for a given cohort by the relative abundance of returns for the year the age class returned.  This method has been used successfully for Puget Sound Chinook returns in our analyses of productivity and rebuilding harvest limits.  It has also been tested on salmon data sets with a complete set of age data, randomly missing our data and estimating it with our method.  The accuracy of the estimated age and productivity based on various properties of abundance and age distributions of the population is discussed.