This repository evaluates ABC (Allowable Biological Catch) and TAC (Total Allowable Catch) projections for Bering Sea and Aleutian Islands (BSAI) groundfish fisheries. The analysis compares two-year model projections against final values and a naive rollover baseline to assess forecast reliability for fisheries management.
doc/- Main analysis document (Quarto), bibliography, and rendered outputsR/- Exploratory R scripts for data analysis and visualizationdata/- Source data (BSAI_OFL_ABC_TAC.csv with OFL/ABC/TAC records 1986-2025)
The main analysis (doc/evaluation-abc-tac-projections.qmd) uses Quarto with R and covers:
- Interannual variability (CV) of ABC/TAC across species
- Two-year vs final value comparisons (percent and absolute changes)
- Rollover baseline comparison
- Log-linear TAC~ABC regressions with cross-species effects
- Dynamic Structural Equation Model (DSEM) for TAC prediction
Core: tidyverse, here, scales, GGally
Visualization: ggthemes, ggdag, patchwork, viridis, wesanderson
Modeling: dsem
Tables: gt, knitr, flextable, YesSiR
quarto render doc/evaluation-abc-tac-projections.qmdOutputs both HTML (with embedded resources) and PDF formats.
The CSV contains columns: AssmentYr, ProjYear, lag, Area, Species, ABC, OFL, TAC, OY, Order
lag=1: Final values used that yearlag=2: Two-year projection made the prior yearOY=1: Records included in Optimum Yield calculations- Seven main species account for ~90% of total ABC: Pollock, Yellowfin sole, Pacific cod, Atka mackerel, Northern rock sole, Flathead sole, Pacific ocean perch
- Percent differences scaled by two-year ABC:
(value_lag1 - value_lag2) / ABC_lag2 - Absolute percent error for model vs rollover comparisons
- Coefficient of variation for interannual variability