Reducing uncertainty in sub-stock composition to better understand harvest-population diversity trade-offs in Canadian origin Yukon Chinook

The Project Investigators evaluated the value of population composition information for management decision based on inferred trade-offs between harvest and biological risk in Canadian-origin Yukon River Chinook. Using historical scale samples and genetic stock identification, they decomposed aggregate Chinook salmon returns into individual population compositions from 1985 – 2019. They then used closed loop simulations to compare the biological and fishery performance of harvest policies based on a range of information about population level composition over time. The first, least informative scenario (DS.1), reconstructed population level dynamics over time based on a single recent year of population composition data (2019); the second, moderately informative, scenario (DS.2) was based on population compositions averaged across all year with data (1985 – 2019); the last most informative scenario was based on all years of population composition data coupled with a state-space run-reconstruction that accounted for incomplete observations over time. For each scenario a harvest policy was identified based on the inferred trade-offs between aggregate harvest and risk to individual populations and then the policy was projected forward in time. They found that while harvest performance was similar across all three scenarios, improved population composition information resulted in management decisions that reduced the risk of overfishing weak populations by up to 60% and increased the likelihood of meeting individual population spawner goals by 23-48% thereby increasing equity in access to Chinook across the system. Collectively these analyses illustrated the potential decision making value of improved information on population composition and strengthen the foundation of information upon which future work can understand the benefits of, and risk to, population diversity in the system.


The project budget funds were spent at a lower amount than originally planned for due to not being able to find all of the historical scale samples the Project Investigators were originally hoping to process at the Department of Fisheries and Oceans (DFO) Scale Lab. This led to a corresponding decrease in the amount of administrative and operations funds being spent as well. There was an administrative error in the original project funds entered into the PCSRF database. The original amount should have been $87,994.50. Even at this budgeted amount, the project funds were under budget in project disbursements for the DFO Molecular Genetics Lab - processing the scale samples for genetic stock identification came in lower than was budgeted for.

The principle investigators propose to strategically increase the annual number of scales analyzed from 1982-2005. Over 8,000 scales have been analyzed to date to determine annual sub-stock composition of the run. Since 2006 an average of 500 scales have been analyzed annually, however, our current 2019 AYK-SSI Proposal Submission2understanding of sub-stock composition from 1982 to 2005 is based on only ~ 150 scales from each year. Increasing the annual number of scales to 250 is predicted to eliminate biases and reduce uncertainty in sub-stock composition estimates by up to 15%. More precise, and less biased, sub-stock composition estimates will translate into reductions in uncertainty that will benefit the sub-stock run-reconstructions and stock-recruitment analyses that are the basis of the management simulation modeling that is currently underway. In addition, they will allow for a unique opportunity to conduct a formal Value of Information analysis on the consequences of improved understanding of population diversity on the performance of alternative mixed-stock harvest strategies for salmon in large river basins.