Document Type
Thesis
Abstract
The Yukon River is the longest river in Alaska, stretching 3,700 km from British Columbia, Canada to the Bering Sea. Chinook salmon (Oncorhynchus tshawytscha) harvest in the U.S. portion of the river is managed by the Alaska Department of Fish & Game (ADF&G), which requires robust inseason abundance predictions for returning Canadian-origin Chinook salmon to regulate fisheries and comply with the Yukon River Salmon Agreement. Chinook salmon returning to the Yukon River are an important resource for residents of the region; however, in recent years little-to-no harvest opportunity has been available due to low run sizes which are complicated by a high degree of uncertainty in annual run size predictions. Currently, the ADF&G treats the preseason run size forecast separately from inseason run size projections based on information collected at the Pilot Station Sonar (PSS) and Eagle Sonar projects. As a result, the ADF&G inseason projection methods may not fully account for uncertainty in the data or the increase in precision and accuracy of inseason information relative to preseason forecasts as the season progresses. To address lack of integration between preseason and inseason projections, I implemented a Bayesian updating approach where preseason forecasts for Canadian-origin Chinook salmon are updated with inseason sonar passage information to project the total end of season abundance. I explored differences in inseason projection accuracy among multiple methods for relating PSS and Eagle Sonar daily passage to the season total Canadian-origin Chinook salmon abundance, as well as genetic stock identification data collected at PSS. Next, I developed a run timing model based on environmental covariates, to address uncertainty in abundance projections associated with interannual variation in run timing. I found that treating the preseason forecast as a prior and updating it daily with inseason passage observations resulted in more accurate projections of the season total Canadian-origin Chinook salmon run size and that run timing information did not improve run size projection accuracy.
Publication Date
8-17-2024
Recommended Citation
Lambert, Aaron Wallace, "Development of a Bayesian framework for Canadian-origin Yukon River Chinook salmon inseason abundance projection and the exploration of run timing information on run size projection accuracy" (2024). Fisheries . 409.
https://ualaska.researchcommons.org/uaf_fisheries_facpubs/409
Handle
http://hdl.handle.net/11122/15515