Release-recapture models for migration juvenile and adult salmon in the Columbia and Snake Rivers using PIT tag and radiotelemetry data
Release-recapture models for two types of tagging data from migrating salmonids in the Columbia and Snake rivers are presented. The first model uses both juvenile and adult PIT-tag data to analyze the seaward and spawning migrations through the hydrosystem. This branching model accommodates transported smolts, known removals, and multiple adult age classes. Juvenile inriver survival, adult return rates, adult inriver survival, and transportation effects can be estimated from this model, as well as smolt-to-adult return rates (SARs) and the adult age composition. An example of summer Chinook salmon ( Oncorhynchus tshawytscha) released in the Snake River upstream of Lower Granite Dam in 1999 is analyzed. The relative system-wide transportation effect measure was Rˆsys = 1.232 ( SE＾ = 0.036), while the SAR to Lower Granite Dam was estimated to be SAR＾ = 0.0193 ( SE＾ = 0.001). Overall upriver adult survival for the release group was estimated at SA = 0.8175 ( SE＾ = 0.022).The second model analyzes radiotelemetry data from adult salmon to estimate perceived system survival and unaccountable loss during the upriver adult migration. Detections are available at the base and top of dams, and in tributary mouths. A sequence of models is presented, ranging from simple to complex, incorporating memory effects of tributary visits and fallback events. Models are compared using a data set of spring/summer Chinook salmon radio-tagged as adults at Bonneville Dam in 1996. Because these adults came from multiple spawning sites, the estimate of perceived survival from Bonneville to Lower Granite Dam was low, at Ssys = 0.10 ( SE＾ = 0.01). Unaccountable loss, m̂R = 0.28 ( SE＾ = 0.02), was considered the more appropriate performance measure for the adult migration in the case of non-known source fish.A secondary purpose of this dissertation was to compare the analysis of adult data from radiotelemetry and PIT tags. The models in Chapter 3 indicate that for estimating large-scale quantities such as system survival and unaccountable loss, PIT tags may offer comparable information to radio tags. If the projected PIT-tag detection systems in tributary mouths become available and reliable, PIT tags may reasonably replace the more expensive radio tags in estimating large-scale quantities reflecting the adult migration.