In pacific-hake/hake-assessment: Stock Assessment for Pacific Hake

Request 2 -- "Sensitivity to Age-1 index plots"

Show fit to survey index for sensitivity removing the age-1 index (plots from the assessment presentation on slides 16-20), and any other plots the JTC finds to be helpful.

• Created plots from slides 16, 18, 19, and 20 of the assessment presentation using the model which does not include the age-1 relative index

• Changes in the survey index fit include a reduction in the 2019, 2021, and 2023 medians when compared to the base model

• Very small change in the survey age composition fit for age-3 fish in 2021. Other than that there are no discernible differences in the age composition fits

Fit to the acoustic survey (age-1 index removed)

model <- sens_models[[2]][[2]]

plot_survey_fit_mcmc(model,
                     type = "acoustic",
                     n_posts = 1000,
                     y_lim = c(0, 6),
                     glow = TRUE,
                     glow_color = "black",
                     glow_offset = 0.5,
                     leg_ymax = 4.7,
                     leg_sep = 0.7,
                     leg_font_size = 10 / .pt)

Fit to the acoustic survey (base model) for comparison

plot_survey_fit_mcmc(base_model,
                     type = "acoustic",
                     n_posts = 1000,
                     y_lim = c(0, 6),
                     glow = TRUE,
                     glow_color = "black",
                     glow_offset = 0.5,
                     leg_ymax = 4.7,
                     leg_sep = 0.7,
                     leg_font_size = 10 / .pt)

Fit to acoustic survey age composition data

:::::: {.columns} ::: {.column width="40%"} Essentially the same as last year's assessment Overestimated: * \textcolor{blue}{1999} cohort in most surveys * \textcolor{yellow}{2010} cohort in 2017 survey * Underestimated: * some young cohorts in 2003--2011 * \textcolor{black}{2014} cohort in 2017 survey * \textcolor{yellow}{2021} cohort in 2023 survey :::

::: {.column width="60%"}

plot_age_comp_fit(model,
                  n_col = 4,
                  type = "survey",
                  x_breaks = seq(2, 15, 2),
                  label_font_size = 4,
                  label_loc = c(12, 0.7))

::: ::::::

Fit to fishery age composition data

:::::: {.columns} ::: {.column width="40%"} Essentially the same as last year's assessment Large \textcolor{Purple}{1999}, \textcolor{Pink}{2010}, \textcolor{Purple}{2014}, and \textcolor{Orange}{2016} cohorts fit particularly well Some over- and under-fitting in \textcolor{Pink}{1980} and \textcolor{Purple}{1984} cohorts The \textcolor{Green}{2020} cohort is fit better than last year * Overall, fishery data in r assess_yr - 1 is fit well and implies age-3 and -4 fish made up a large contributon of the catch :::

::: {.column width="60%"}

plot_age_comp_fit(model,
                  label_font_size = 3,
                  type = "fishery",
                  n_col = 4)

::: ::::::

Pearson residual for fit to the age data

:::::: {.columns} ::: {.column width="50%"} \center Fishery \center

plot_pearson_bubbles(model,
                     type = "fishery",
                     leg_pos = "top")

:::

::: {.column width="50%"} \center Survey \center

plot_pearson_bubbles(model,
                     type = "survey",
                     alpha = 0.7)

::: :::::: Dark bubbles: observed > expected

White bubbles: observed < expected

Pearson residual for fit to the age data (base model) for comparison

:::::: {.columns} ::: {.column width="50%"} \center Fishery \center

plot_pearson_bubbles(base_model,
                     type = "fishery",
                     leg_pos = "top")

:::

::: {.column width="50%"} \center Survey \center

plot_pearson_bubbles(base_model,
                     type = "survey",
                     alpha = 0.7)

::: :::::: Dark bubbles: observed > expected

White bubbles: observed < expected

pacific-hake/hake-assessment documentation built on Feb. 17, 2025, 1:58 p.m.