decision table info

[iantaylor-NOAA]

Links from decision table discussion in today's class:

• 9 models for 2019 widow decision table: https://drive.google.com/drive/folders/1CJ9wlCuUs9q7AxK1tRxEbAeYtX8QyvNr
• table_decision() function in the Yellowtail Rockfish repo: https://github.com/pfmc-assessments/yellowtail_2025/blob/main/Rscripts/table_decision.R
  • applied in exec summary at https://github.com/pfmc-assessments/yellowtail_2025/blob/main/report/01_executive_summary.qmd#L381-L395
  • applied in main tables at https://github.com/pfmc-assessments/yellowtail_2025/blob/main/report/08_tables.qmd#L467-L481
  • models will get read in this section, but haven't been developed yet, so only placeholder: https://github.com/pfmc-assessments/yellowtail_2025/blob/main/report/002_load_tables.qmd#L122-L147
• r4ss::SS_ForeCatch() function to create a table of fixed catches (e.g. from harvest control rule) from base model to input into forecast file for low and high state of nature

[kinmic]

Hi @iantaylor-NOAA

I'm a little unsure of how the forecast catch is calculated for the 0.25 and 0.45 catch streams. Can you explain what
'ACLp* =0.45 sigma =0.50' actually is and how it corresponds to the forecast catches defined by GMT for the first two forecast years (are these the ACL? or the ACT? - Sorry I get a little confused with the WC reference points)

[okenk]

That's why the buffer from PEPtools::getbuffer() for the first two years 1, i.e., there is no buffer. We just use the GMT catches as they are.

> PEPtools::get_buffer(2025:2036, sigma = 0.5, pstar = 0.45)
The sigma value of 0.5 adjusts yearly by the rate of 0.075.
   year buffer
1  2025  1.000
2  2026  1.000
3  2027  0.935
4  2028  0.930
5  2029  0.926
6  2030  0.922
7  2031  0.917
8  2032  0.913
9  2033  0.909
10 2034  0.904
11 2035  0.900
12 2036  0.896

[kinmic]

Hi Kiva,

Ah yes ok that helps. I wasnt aware of the getbuffer function. So I add the output of that getbuffer() into a model as hcr, then use SS_ForeCatch() to generate the fleet wise catch amounts?

[okenk]

Yes, see this code for how to implement it using r4ss:
https://github.com/pfmc-assessments/yellowtail_2025/blob/main/Rscripts/bridging_runs.R#L495-L505

[okenk]

Sorry, there are also two separate types of forecast files:

1. base model, use the code I linked to above, where you hard code in 2025 and 2026, and then use the buffer from the get_buffer() function
2. decision tables, where you have a control file for a non-base model, and then you hard code in catches in the forecast for 2025-2036, using the ACL from the base model.

[kinmic]

Oh awesome, this maks perfect sense now. Thanks Kiva!