Adding MARBL to the library to support simultaneous simulations

[lgloege]

I am curious if there is work being done to add MARBL to OceanBioME? I think this would be a great addition to the mCDR community since it allows running counterfactual and intervention experiments simultaneously with the same physics. Folks in the ROMS community were noticing spurious numerical noise when differencing two independent simulations and MARBL helped them get around this. I will gladly contribute if work on this front is already underway. If not, would be great to discuss how to get the ball rolling. I am happy to lead this effort,

[johnryantaylor]

Hi @lgloege, As far as I know noone is working to integrate MARBL into OceanBioME. It would be wonderful if this is something that are able to contribute! I agree that it would be particularly useful for comparisons with ROMS/MARBL for mCDR. You might also look at the PISCES model (another IPCC-class biogeochemical model) which is implemented in OceanBioME. @jagoosw implemented PISCES in a modular way with the idea that it would be possible to combine different biogeochemical models together (e.g. take phytoplankton from one and nutrients from another).

[jagoosw]

Hi @lgloege

This sounds like a great idea! I also don't think anyone is working on this but if you'd be happy to lead I can help. A good place to start might be this docs page if you've not seen it already, and the PISCES implementation might be a good guide on how we'd like to structure more complicated models.

When I wrote the PISCES implementation I tried to modularise it a bit, but its not there yet and some work will need to be done to change out parts of the model (e.g. the phytoplankton), so if you have any ideas on that it would be great to discuss.

Would it be helpful if I made a branch with an outline of the code to get this going?

[lgloege]

Hi @jagoosw

Yes, it would be very helpful if you could create a branch with an outline of the code. Thank you. I come from a python background and recently got into julia via Oceananigans/OceanBioME. Having a project like this will force me to hone my julia skills. Looking forward to starting this!

@vtamsitt may also be interested in this effort

[jagoosw]

Do you know where I could find more explict documentation with the formulas or even a list of all of the tracers?

I'm struggling to work out what is explicitly modelled, it says 15 living Diaz/Dia/Nano * (C, P, Si, Fe, CaCO₃) + Zoo, but I guess one of the phytoplankton doesn't track one of the tracers to make it 15? And then 17 non living but I can't find a list of what they are (DIC, Alk, O2, NO3, NH4, Fe, Si, PO4, DOC, DOP, DOFe?, ...?).

[lgloege]

There is the MARBL read-the-docs page. I was unable to find a list of tracers. Once we outline the code to the best of our knowledge we could see if Jesse Vance would double check it? He knows the model well since he has been using it with ROMS recently.

[jagoosw]

@lgloege

I've made a branch here. It might be helpful for you to refer to the PISCES implementation to work out how todo it.

I have no idea how best to modularise/abstract but we can discuss that as you go along.

Also, I noticed in the docs that MARBL uses centimeters, we've tried to stick to SI units in our models with the exception of concentration which we usually do in mmol / m^3, and we still have PAR in einstein/s but might try and depreciate that.

[lgloege]

Thanks, i'll take a look at the branch and PISCES implementation this week. I have some ideas how to modularize the code, but want to think on it some more before I say anything.