Case studies

[ChristineStawitz-NOAA]

A place to ask questions about FIMS case studies

[ChristineStawitz-NOAA]

Should we do this here or in the Google space?

[cmlegault]

I just uploaded asap2fims_v2.R here. It now runs with the latest and greatest version of FIMS.

[msupernaw]

Interesting results! NAA looks pretty good, however FIMS seems to estimate lower SSB values. Did FIMS meet the convergence criteria?

…

On Thu, Nov 30, 2023 at 2:39 PM Chris Legault ***@***.***> wrote: I just uploaded asap2fims_v2.R here <https://drive.google.com/drive/folders/1WD8eilQORw1rq1aEQr7LFWYlbzSmWwqP?usp=drive_link>. It now runs with the latest and greatest version of FIMS. — Reply to this email directly, view it on GitHub <#530 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABFUSEBJHMJZTQ6HUB5MVPTYHDOFTAVCNFSM6AAAAAA76SKKDSVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TOMRRGY4TE> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

-- Matthew Supernaw *Scientific Software Developer* *National Oceanic and Atmospheric Administration* *Office Of Science and Technology* *NOAA Fisheries | *U.S. Department of Commerce Phone 248 - 396 - 7797

[cmlegault]

I used the additional Newton step trick from Cole to get really good convergence in FIMS. The SSB values are due to an undocumented feature in the FIMS SSB calculations that applies a 0.5 multiplier (see Issue #521).

[Cole-Monnahan-NOAA]

I'm getting an error when running this line

fishing_fleet$log_obs_error$value <- log(sqrt(log(as.numeric(mean(rdat$control.parms$catch.tot.cv[,1], na.rm=TRUE)^2) + 1)))

Error: Not compatible with requested type: [type=list; target=double].
In addition: Warning message:
In fishing_fleet$log_obs_error$value <- -2 : Coercing LHS to a list

Should I go back to a previous commit?

[iantaylor-NOAA]

If you're using current main branch, then I think the log_obs_error needs to be a vector based on this change: https://github.com/NOAA-FIMS/FIMS/pull/532/files.

[Cole-Monnahan-NOAA]

I tried that and got the same error so I went back to the commit referenced in the code and now it seems to run.

[ChristineStawitz-NOAA]

Thanks @iantaylor-NOAA you are correct. In addition to changing the input, the names of these two values need to be changed wherever they are called to reflect that it's a vector and not a scalar quantity anymore:

fishing_fleet$log_obs_error <- rep(log(0.01^2 + 1))), nyears)
fishing_fleet$estimate_obs_error <- FALSE

See here for full syntax: https://noaa-fims.github.io/FIMS/articles/fims-demo.html

[Cole-Monnahan-NOAA]

I got pretty far on bridging my case study but am finding some strange behavior with the initial NAA and plus group calculations. Perhaps this could be a seaside chat topic?

[ChristineStawitz-NOAA]

Thanks! Can you make the seaside chat tomorrow to discuss?

[Cole-Monnahan-NOAA]

Yes I hope to be there unless I'm still feeling sick. I have some really informal stuff prepared that we can walk through.