search

r4atlantis / atlantisom

Atlantis operating model. Generates data sets from Atlantis scenarios.
https://r4atlantis.github.io/atlantisom
8 stars 6 forks source link

catch.nc and catch.txt do not add up #21

Open alketh opened 8 years ago

alketh commented 8 years ago

@kellijohnson Can you upload the catch.txt files?

kellijohnson-NOAA commented 8 years ago

@alketh the files are now uploaded with e6c59d851f4f9d5b21423e017123a551598ca794

kellijohnson-NOAA commented 8 years ago

@alketh I think there is an error in the calc_biomass_age function. @hodgsone and I were chatting and she remembers that not all of the catch needs to be converted using the redfield ration, thinking that vertebrates are already in wet weight, or something along those lines. Also I checked the atlantis code and it seems like maybe we also need to look at volume of the layer but I am not sure.

sgaichas commented 5 years ago

I think we have a couple of problems with CATCH.nc. First is a bug in the output for models using codebases prior to late 2015 (I'm working with California Current right now and it's vintage is mid-2015). From Beth Fulton's note in the wiki:

Note that in CATCH.NC the numbers at age had issues in old versions. For versions pre-dating December 2015 you need to divide the numbers at age by (86400 * number_water_column_layers_in_the_box). New versions post December 2015 are now in numbers without needing further adjustment.

I just wrote code to correct the catch numbers at age for California Current and will try to add that to run_truth

The second problem is the CATCH.nc output file does not include a layer, so the aggregation of structn and resn over layers (using medians) in lines 58-67 in calc_biomass_age may be resulting in this mismatch. If catch happened in a particular layer, it is aggregated in the output CATCH.nc file, so we don't know which layer has the appropriate weight info. The calc_biomass_age function works on the truth$nums output because no aggregation of structn and resn is necessary, all have a layer field.

One workaround is to assume the total catch in the catch.txt output file is correct and use that directly as catch biomass for atlantisom. We will have only annual total catch and no way to break it into seasons, but maybe that is ok.

gavinfay commented 5 years ago

Unsure if this is helpful but,

Re: your inline comment at 61-62 of calc_biomass_age, In the harvest.prm file, there is (or was) a parameter named: Effort_vdistrib which describes the proportion of effort by water column layer for each fleet.