GSL-south

[AquaAuma]

I think we better document in the data GSL Case properly, and consider each as a survey_unit, what do you think @zoe?

[mpinsky]

@AquaAuma , can you please clarify the issue a bit more for me? It's possible I can answer.
Also tagging @zoekitchel , who got missed in the first comment.

[AquaAuma]

This survey I think is broken down into three sub-surveys by DFO and we should not merge them because they use different designs, which was the main concern

[AquaAuma]

I think it may be the same issue for the northern survey

[mpinsky]

the GSL northern survey has used four vessels with slightly different gear types, documented here: #10 (comment)

[zoekitchel]

Fishglob currently has data from 1970-2019 for GSL-S. From https://waves-vagues.dfo-mpo.gc.ca/Library/115732.pdf, we know:

• the survey was conducted annually with the same vessel and gear until the mid 1980s. In 1985, the RV Lady Hammond replaced the RV EE Prince
• 1970-1983 = random stratified, 1984-1987 = fixed station, 1988-present = random stratified.
• there are three different types of trawl used, with different widths and surface area covered during a standard 30 minute tow (EE Princee = Yankee 36, Lady Hammond = Western IIA, Navicula = 50' Flounder)

The dataset we import from OceanAdapt only includes the Western IIA (1985&1987-2019) and the Yankee 36 (1970-1986). In the cleaning code (get.gsl-s.R), I adjust the area_swept based on the gear type. These two gear types overlap in 1985.

For my taxonomic homogenization work, I only used 1992-2019 because of an additional gear/vessel change in 1992. See https://github.com/AquaAuma/fishglob/issues/72.

[mpinsky]

My notes from talking to Hugues Benoît (DFO) n 2009:

• Daytime sampling 1971-1984, 24 hour sampling 1984-present
• Strata 401, 402, 403 added in 1984
• E. E. Prince 1971-1985 Yankee-36 trawl
• Lady Hammond 1985-1991 Western IIA trawl
• Alfred Needler 1992-2005 Western IIA trawl
• Except 2003 Wilfred Templeman, but almost exactly the same as Alfred Needler
• 2003: missed 402, 425, 436 (Benoit 2006): deep strata, important for deep-water spp. Can fill with a multiplicative model (year*stratum), but assumes no distributional changes
• Teleost 2006-present

He also recommended these for people who want more details:

• Benoit & Swain 2003 standardizing SGSL trawls - CTRFAS.pdf
• Benoit 2006 standardizing the southern Gulf of St. Lawrence - CSASRD.pdf

[AquaAuma]

so based on all this, what do you think is the best to ensure more accurate usage of the dataset in fishglob?

[mpinsky]

Explain what we know and encourage users to consult the original sources to understand other specifics relevant to their analysis.

[zoekitchel]

I agree! I'll add some content to the metadata. I wonder if we should include a similar disclaimer to this on our home page for the repository? (From DATRAS) "*You need to accept the following before you can submit.* *Data policy* The trawl survey database can be accessed depending on your permission level (see data access policy) <https://datras.ices.dk/Data_products/Download/Download_Data_public.aspx#> *'Healthy' warning* Whilst the data have been extensively checked by both the supplying institutes and ICES there are inherent flaws in gathering the information. For example: * In spite of the fact that now all bottom trawl surveys are in the same database and often at least partially overlap in time period or area covered, there are differences between the surveys that may not allow the combination of data from different surveys in one analysis unless this point is specifically addressed. * There is no guarantee that the ground gear types deployed adequately sample all species (e.g. catchability of flatfish in beam trawls is markedly higher than that of GOV). It should be realized that each survey gear catches a gear-dependent subset of the fish community. Also, for the same gear, catchability can be largely modified by rigging, ticklers, or other elements as those described under Gear Exceptions field (for instance bobbins, rockhopper etc.) * A common property of all surveys is that over time species determination skills improve. As a result new taxa will emerge in the dataset that were previously reported as part of another closely-related taxon. * Over time the sampling procedures, gear characteristics, timing of the survey or the area covered may have changed thereby influencing the catches. Expert advice recommendations are given in order to prevent bias in the analyses using DATRAS catch data, caused by such factors. Thus users are urged to treat the data with caution. If the user has any queries on the validity of the data or the conclusions to be drawn they should contact either the ICES Secretariat or the Chairs of the relevant Working Group (IBTS, BITS or BEAM) To guide the users of the trawl data the relevant working groups have provided the users with an expert advise. Please read these before using the data: Expert advice <http://datras.ices.dk/ExpertAdvice.aspx> This is an informative note. For further details on the current data policy, please read carefully the disclaimer included in the download package. I accept the documents and their associated restrictions and agree to acknowledge the data source."

…

On Thu, Sep 7, 2023 at 6:29 AM Malin Pinsky ***@***.***> wrote: Explain what we know and encourage users to consult the original sources to understand other specifics relevant to their analysis. — Reply to this email directly, view it on GitHub <#5 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHQNR7VYTPVHC3UXDJ6X7DLXZHD5FANCNFSM6AAAAAATXRBEN4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- *Zoë Kitchel* PhD Candidate Ecology and Evolution* Rutgers University* she/her/hers *w.* www.pinsky.marine.rutgers.edu *c.* 207.807.4426

[mpinsky]

I agree. A version of this disclaimer would be very appropriate for FISHGLOB as well.

[AquaAuma]

I like that idea! Would any of you volunteer to draft one? Otherwise will do at the end of this month. I'm trying to wrap things up quickly, and I have a lot to do on the dataset itself

[zoekitchel]

I could do it the last week of September, but unfortunately not before then!

[AquaAuma]

that will work, since there were too many tasks on the revision, I've delayed the resubmission deadline again

[AquaAuma]

Just double checked we have this in the metadata docs

So we should be fine? I'll just resolve