4 Discussion
4.1 Intended use (New SEDS/NuSEDS data entry and review)
The updated guidance is intended to support consistent estimate-type assignment during data entry and review by:
- Making method and evidence expectations explicit (rather than implicit in a table-only lookup).
- Recording conservative qualifier codes that explain why a candidate Type cannot be supported.
- Improving interpretability for downstream users by separating enumeration (field) methods from estimation (analysis) methods.
Estimate Types are not a binary “usable/unusable” label. Lower Types may still be informative for specific purposes (e.g., presence/absence, qualitative local context, or within-program indices), but are not interchangeable with census-like or fully qualified absolute-abundance estimates.
4.2 Implications for downstream assessments
Estimate Types are commonly used as screening filters in downstream workflows (including WSP rapid status assessments) where method properties and evidence quality affect interpretation (C. A. Holt et al. 2009; Fisheries and Oceans Canada 2005). When Type assignment is inconsistent, threshold-based decisions can become sensitive to classification noise, particularly near cutoffs (e.g., around the Type 4 boundary). More broadly, measurement error in spawner abundance can distort stock–recruit inference and downstream decisions, reinforcing the need to record uncertainty and detectability-related metadata where available (Walters and Ludwig 1981).
By producing explicit qualifier codes (e.g., missing method identification, timing/coverage gaps, insufficient documentation), the updated guidance makes the reasons for conservative classification transparent and enables targeted improvements in metadata capture.
4.3 Recommended metadata additions (schema) and implementation status
A central finding of this update is that the current Type table alone cannot communicate the information required for consistent interpretation of real-world escapement estimates. Several recurring qualifiers are better represented as explicit metadata elements rather than as new public estimate Types.
This report therefore recommends (and the companion implementation supports) additional metadata elements for New SEDS/NuSEDS uploads, including:
- Run-window coverage evidence (missed start/end of run).
- Device/site uptime and unobserved intervals.
- Breach/bypass context and severity.
- Infilling/interpolation method and justification.
- Timing relative to run timing (and visibility/conditions where relevant).
- Combined-method components and an integration method summary.
- Calibration source and diagnostics for calibrated time series.
- Revision history (what changed and when).
- Quantitative uncertainty when available (e.g., CV or SE).
These elements are summarized in Methods (Table 2.2). In the companion toolkit implementation, these elements are reflected in the returned qualifier codes so that Type assignment is transparent to data users (NuSEDS Escapement Estimates Toolkit Working Group 2026).
4.4 Practical implementation monitoring (lightweight)
A full empirical validation program is outside current resourcing. However, New SEDS/NuSEDS stakeholders can still track practical indicators of data completeness and interpretability that are directly tied to the guidance.
Suggested monitoring metrics include:
- Proportion of estimates with METHOD_UNKNOWN.
- Proportion of estimates downgraded for DOC (insufficient documentation evidence).
- Frequency of specific qualifiers (e.g., RUN_COVERAGE, UPTIME, BREACH_BYPASS, TIMING, VISIBILITY) by program or time period.
- Distribution of Types (1–6) over time within programs, to detect abrupt changes driven by metadata practices rather than biology.
These metrics support targeted improvements (e.g., documenting breach severity more consistently) without requiring a resource-intensive validation study.
4.5 Limitations
This guidance summarizes interpretability given recorded methods and evidence. It does not correct biased inputs, replace program expertise, or guarantee biological accuracy. The most common practical limitation is incomplete metadata: when critical qualifiers are not recorded, the guidance applies conservative classification (often Type 5) to avoid overstating interpretability.
4.6 Governance and change control
Because estimate Types are used as screening tools and can influence downstream inference, any changes to method families, thresholds, or qualifier rules should be versioned and reviewed. Changes should be accompanied by an updated worked example set and regression/path tests in the companion implementation so that behavior changes are deliberate and transparent.
4.7 Next steps
- Finalize a minimal, public-facing metadata schema for New SEDS/NuSEDS that supports the guidance (including quantitative uncertainty fields where available).
- Maintain a small set of worked examples (anonymized as needed) to demonstrate expected behavior on representative cases.
- Use implementation monitoring metrics (above) to prioritize metadata capture improvements that most strongly affect interpretability.