search

pfmc-assessments / canary_2023

Other
5 stars 1 forks source link

ABC buffer in timeseries is different than that in the forecast? #78

Closed brianlangseth-NOAA closed 1 year ago

brianlangseth-NOAA commented 1 year ago

@iantaylor-NOAA I notice that in the timeseries output of the report file, there is a column called ABC_buffer. However, it is different than the buffer applied within the forecast file. 

When using get_buffer(c(2023:2034), sigma = 0.5, pstar = 0.45) I get the following, which is what I entered in the forecast file  #_Year Fraction    2023    1.000    2024    1.000    2025    0.935    2026    0.930    2027    0.926    2028    0.922    2029    0.917    2030    0.913    2031    0.909    2032    0.904    2033    0.900    2034    0.896

But when I do model$timeseries[,c("Yr","ABC_buffer")] I get  2023   0.923552 2024   0.921031 2025   0.916297 2026   0.914683 2027   0.911792 2028   0.907943 2029   0.903955 2030   0.901249 2031   0.901293 2032   0.904792 2033   0.911342 2034   0.919891

I would think these align. Do you know why they dont?

brianlangseth-NOAA commented 1 year ago

Thanks to @chantelwetzel-noaa and @iantaylor-NOAA for discussing. Pulling model$timeseries gives forCatch over OFL and thus is not the actual ABC buffer that should be used; that is what is taken from get_buffer() and what should be in the forecast file.

Note that the output from SSexecutivesummary in g_Projections_ES.csv has a column called ABC, which if the stock is < 40% is actually the ACL.