Question about how iron_flux is converted into dissolved Fe tendencies (which module handles it?)

[yujiezhang]

Hi everyone,


I am currently working on modifying iron inputs in CESM2’s ocean biogeochemistry (POP2 + MARBL), and I would like to understand exactly how iron_flux is converted into tracer tendencies for dissolved iron (Fe) inside the model(e.g., mmol Fe m⁻³ s⁻¹).

 

[mlevy]

The surface fluxes for all tracers (which includes those returned from MARBL, which includes iron) are used by the ocean GCM in the vertical diffusion and mixing processes.

For POP2, I believe all the fluxes are converted to tendencies in the vdiff() subroutine, specifically by the multiplication of dzr(k) (1/layer thickness, which has units of m⁻¹). Note that it's easy to get lost in the variable names, but STF refers to surface tracer fluxes and VTF = STF in the top layer of ocean cells (there is a merge() statement setting VTF earlier in the routine)

         VDTK(:,:,n) = merge((VTF(:,:,n,bid) - VTFB)*dzr(k), &
                             c0, k <= KMT(:,:,bid))

 

[yujiezhang]

Thank you very much for your clear explanation. It's really helpful to understand how surface fluxes are processed in POP2 through the vertical diffusion subroutine.

 

[yujiezhang]

Hi mlevy,

I am encountering an issue with my CESM2.1.3 simulation. I used the SSP585BPRP restart files from 2020 and added an extra iron flux in the ecosys_forcing module. The fertilization region is between 30N-35N in latitude and 210-215 in longitude. I then compared this experiment with a control run.

The problem is that the sea surface iron concentration in other regions responded too rapidly, changing globally within just a few months. I am unsure what is causing this. To check if it was due to air-sea feedback, I ran a test using the G1850 configuration, and the results were normal.

If I want to run future scenario experiments, what suggestions could you give me?

 

[mlevy]

As we talked about in your other thread, there isn't a good out-of-the-box way run a future simulation with a data atmosphere... we just don't have the datasets to use for forcing those runs. I was talking with a colleague about your results, and he was surprised at the lack of noise the G compset plots -- perhaps it's just a combination of the contour levels you chose for plotting and the size / strength of your iron patch, but the basic mechanism driving global changes is the change in chlorophyll affecting heat uptake. In the fully-coupled run, this feedbacks to the atmosphere (changes in ocean heat uptake result in different heating patterns in the atmosphere), so it's not surprising that there are global differences in the first month. However, even in the G case (which does not feed back to the atmosphere), changes in chlorophyll still lead to differences in ocean heat uptake and that will change circulation patterns. I wonder if you'll see global implications of iron fertilization in your G case if introduce more contour levels around 0.