Use of Statistical Bootstrapping to Calculate Confidence Intervals for the Fall Chinook Salmon Run Reconstruction to Lower Granite Dam

Abstract

AbstractMeeting escapement and recovery goals for Pacific salmon Oncorhynchus spp. is greatly facilitated by accurate and precise estimates of run size and stock composition. Although this is particularly important for endangered Pacific salmon runs like the Snake River fall Chinook salmon O. tshawytscha evolutionarily significant unit, few studies provide estimates with confidence intervals (CIs) and standard errors (SEs). We developed a run reconstruction method for fall Chinook salmon returning to Lower Granite Dam on the Snake River. We showed that composition estimates of fish returning to large river systems can be obtained through subsampling. The methods reported here yield CIs around those estimates. We used compound bootstrap methods to derive CIs and SEs on return numbers and composition. In general, if the number of fish returning for a group was 300 or more, we were 90% confident that our return estimate represented the true number within 10%. Estimates for smaller groups were generally imprecise. The analytical technique described here provides fishery managers with information necessary for assessing escapement and recovery goals and addressing conservation and harvest allocations of adult Snake River fall Chinook salmon.