Review panel backs October update as best current assessment, asks for fecundity revision and a prompt benchmark

A supplemental review panel convened virtually to evaluate an updated stock assessment for widow rockfish recommended that the October 2025 assessment be treated as the best scientific information currently available for management, but asked the stock assessment authors to rerun the model with a species-specific length–fecundity relationship and requested a near-term benchmark assessment to address outstanding uncertainties.

The panel’s recommendation followed extended discussion about contrasting results from models presented in August and October 2025. Panel members said the October update produced lower estimates of depletion but, counterintuitively, slightly higher estimates of overfishing limits (OFLs) and projected allowable catch in 2027. The panel flagged differences arising from a shift in the assessment’s reproductive metric — from weight-based spawning biomass to spawning output (eggs) — and from a higher internally estimated natural mortality (M) in the October model.

Why it matters: The choice of biological units and parameter values changed the way the model values older, more fecund fish and altered management advice for future catch limits. Panel members and advisers warned that these technical changes could produce substantially different catch advice even when overall biomass and many diagnostics look similar, and they recommended additional runs and documentation to make the pathway from assessment inputs to OFL clear to managers.

Discussion highlights

- Fecundity and depletion: Andre (panel member) noted that “the fecundity change … is not trivial,” arguing that switching to a fecundity (spawning output) basis reduces calculated depletion compared with a weight-based metric because larger, older fish contribute disproportionately to spawning output in the new formulation. The panel emphasized that the change in units — effectively valuing egg production rather than biomass alone — can alter depletion estimates even when the underlying population numbers (the N matrix) are similar across model runs.

- Natural mortality (M) and exploitation rates: The panel recorded that the October model estimated a higher natural mortality (M) than the August model (discussed values in the meeting noted an increase from ~1.22 in the August run to ~1.56 in the October run). At the same time, reported exploitation-rate numbers used for reference points were only slightly different (examples cited included exploitation rates near 0.086 and F50-related exploitation rates near 0.076 and 0.0767 in different tables). Panel members said that, together, the M increase and the fecundity change explain why depletion declined yet OFL and yield advice rose in the October run — a result that is not intuitively obvious and that the panel asked be documented more clearly.

- Data and survey inconsistencies: Several reviewers noted unresolved problems linking fishery-dependent data (catch per unit effort) and fishery-independent surveys. The bottom-trawl survey has recently sampled more older fish in some areas than the midwater trawl fishery age compositions indicate. Panelists and advisers described these differences as a major source of uncertainty, with potential impacts on estimates of natural mortality and the inferred productivity of the stock.

- Spatial and movement uncertainty: Participants raised questions about stock structure and adult movement. The assessment currently treats the coastwide population as a single stock across three states; the panel emphasized that this assumption is untested and that adult movement scales — including differential movement by age — are an important unresolved problem that bears on interpretations of survey and fishery data.

Panel requests and suggested next steps

- Rerun with species-specific fecundity: The panel requested that, if possible, the assessment authors rerun the proposed October base model using the species-specific length–fecundity parameters identified in the cited meta-analysis so the Scientific and Statistical Committee (SSC) can see the direct effect. The panel asked for those results to be made available to the SSC as soon as feasible.

- Decision tables and paired trajectories: To aid the SSC’s decision-making, the panel requested that decision tables and projection outputs include paired trajectories showing both 4+ biomass (or spawning biomass) and spawning output (eggs) so managers can see how depletion and reference points behave in each unit.

- Prioritize a benchmark assessment: Multiple panelists and advisers urged that a benchmark assessment be scheduled and prioritized in the next assessment cycle. The panel said a benchmark would allow a fuller exploration of alternative model formulations, sensitivity analyses (including alternate values for M and steepness), and a more thorough evaluation of the length–fecundity relationship.

Panel conclusion and minority views

The review panel concluded that, of the models in front of them, the October 2025 update best represents the available data and methods but should be accompanied by the requested rerun (species-specific fecundity) and clearer documentation of how the assessment’s biological assumptions affect OFL and ACL advice. Several advisers expressed reservations: some said the October model should be accepted ‘‘as is’’ and left to the SSC and council to consider the rerun if it is available, while others cautioned that the absence of full sensitivity analyses on fecundity and M in this supplemental review leaves residual risk and strengthens the case for an expedited benchmark.

What the panel explicitly did not do

No formal vote to adopt regulations or specific catch limits occurred during this meeting; the panel’s work is advisory to the SSC and the Council. The panel recorded unresolved methodological issues, recommended specific model outputs and documentation for the SSC, and recommended a prioritized benchmark assessment to resolve major uncertainties.

Next steps and timeline

Panel members asked assessment authors and council staff to supply the requested model outputs and rerun results as soon as operationally possible. The panel recommended that the SSC review the October update (and any rerun) at the November meeting, and several panelists urged that, regardless of November timing, a benchmark assessment be scheduled promptly so any important model changes can be addressed in a fuller review cycle.