Pacific Salmon Foundation: Salmon Watersheds Program

A pair of Chinook salmon on their spawning redd on the Tatchun River in central Yukon. Photo by Peter Mather.
Our Projects >

Exploring How Incomplete Information Affects Biological Status Assessments

Canada’s Wild Salmon Policy (WSP) identifies standardized monitoring of the biological status of Conservation Units (CUs) as a critical element of implementing strategies aimed at conserving wild salmon. This information on biological status is essential for designing, implementing, and evaluating management strategies and recovery plans for threatened salmon CUs. Over the past decade, the Pacific Salmon Foundation (PSF) has been working to advance the implementation of the WSP by taking a data-driven approach to assessing the biological status of Pacific salmon CUs throughout British Columbia following the approaches outlined in the WSP. These assessments of biological status, along with other information on salmon CUs, are made freely available to the public through an online data visualization platform called the Pacific Salmon Explorer.

The estimates of biological status shown on the Pacific Salmon Explorer often involve “run reconstructions” that infill missing spawner data, expand observed spawner abundance to account for unmonitored streams, assign catch to individual populations, and quantify age-at-return. The influence of these assumptions on the accuracy of status assessments is unknown, but of increasing concern as monitoring efforts decline, making these assumptions all the more necessary for deriving estimates of biological status. This project evaluated how common assumptions surrounding the expansion of spawner abundance, assignment of catch to CUs, and age-at-return may bias the kinds of assessments of biological status undertaken by the PSF and visualized in the Pacific Salmon Explorer. 

A basic run reconstruction involves expanding spawner abundances from observed streams to the entire CU (steps 1–3, blue), assigning catch to CUs (step 4, green), and assigning return to brood years (step 5, tan). The black arrows indicate the data required to assess status via percentile and spawner-recruitment benchmarks.

Outcomes

We found that common assumptions in run reconstructions had relatively little impact on status outcomes. Specifically, the probability of misclassifying biological status of CUs was not affected by: (1) reduced monitoring coverage (which affects the magnitude of expansion for spawner abundance), or (2) declines in the carrying capacity of some spawning populations (which was expected to affect the accuracy of expansions). Underestimating spawner abundance did tend to result in more misclassifications when using spawner-recruitment benchmarks, but the effect was relatively small. Overestimating catch (if salmon returning to other areas are incorrectly counted towards the catch for the CU being assessed) tended to result in more misclassifications than underestimating catch under the spawner-recruitment benchmarks; overestimating catch by 50% led to a 10% increase in misclassifications, which is potentially significant given the uncertainty in assigning catch to CUs.

This project has shown that biological status assessments are robust to the common assumptions associated with run reconstructions, even in the face of declining monitoring coverage that has been observed in salmon-bearing regions throughout BC. However, misclassifications increased when catch was overestimated, highlighting the need to improve catch allocation to CUs (for example, through genetic stock identification).

The collection of such data has been challenging, in part because the goals of the WSP – to preserve the incredible diversity within salmon species – are not reflected by the coarse scale at which salmon fisheries are managed (Management Areas often span multiple CUs). Further research is required in order to understand how inherent biases in metrics and benchmarks depend on underlying status, and recommend criteria under which different benchmarks are reliable. However, despite the many unknowns, this project has demonstrated that efforts to assess biological status using imperfect and incomplete data are worthwhile.

Related Links