Infill tow-level covariates

[slarge]

@sgaichas and I were discussing some of the environmental data collected on the surveys and how they can be used in VAST and other models. As it stands, SST coverage is about 90% and BT coverage is probably a bit lower. Returning a data.frame() with a tow-level OISST and/or GLORYs estimated values would be a huge advantage for models that don't play well with missing covariates (e.g., GAM, VAST, etc).

@kimberly-bastille, @andybeet, and I (link , link, and link) have independently developed some code to aggregate OISST, so it probably wouldn't be a big lift. It might also be a useful action to run in the background.

Probably low priority for the time being, but it could be a useful addition.

[sgaichas]

I'll be working out how to merge with the survey diet data (including NEAMAP) in the next couple of weeks. Watch this space for updates.

[andybeet]

@slucey I think it would be a good idea to process and serve the environmental data from surveys within this package. I think we could process it independently then have an option to join it with the catch.

[kimberly-bastille]

SST - @atyrell3 and I worked up (ecopull) some gh actions that pull and calculate SST data. Depending on the exact needs of this project it could be pretty easy to port over and edit.

BT - Bottom Temp conversations for the SOE will hopefully be underway this summer/fall (2022) and I think could be useful here as well. Figuring out a workflow that works for both products (+ more) should be included in the conversation.

[slucey]

Surface and Bottom temp from the survey tows is already included in the get_survey_data pull. We could easily make a stand alone function that only pulls that data then link it up to @kimberly-bastille and @atyrell3 package or other packages to fill in missing data.