iHAMOCC: change N-fixation limitation from surface mixed layer to euphotic zone? about blom HOT 9 CLOSED

I also think it's best to go with option (b). Scientifically, it'd be more logical when comparing different simulations or when merging code versions.

from blom.

With the new coordinates the upper 2 layers of the model are quite thin, i.e. the volume for N-fixation is small und thus PP will be reduced somewhat. So in this regard, the change belongs to the new vertical coordinates.

from blom.

If PR #154 is accepted, the N-fixation should take place over the entire mixed layer depth (defined by the kOBL index). This will hopefully solve the issue with the 2 top layers being very thin, at least for the first implementation. It has the advantage that the code change should be "minimal", in the sense that the vertical coordinate change will not coincide with another conceptual code change.

from blom.

I thought the idea is to limit N2 fixation to the euphotic zone, and not the whole MLD. According to this, N2-fixation requires lights: https://www.sciencedirect.com/science/article/pii/S1385110104000930, though probably during deep MLD, temperature limitation come into effect.

from blom.

I agree that we should have N2 fixation in the euphotic zone, but in the first instance I would like to have a version that is mostly backwards compatible with CMIP6 simulations, so that we can evaluate "pure" coordinate changes. Perhaps this will only be a short-lived intermediate step, but I think it is safer to do it one step at a time.

from blom.

Maybe it's worth thinking about another 2D field similar to kmle in #154 to indicate the depth of the euphotic zone? - so making it possible to differentiate between MLD and euphotic zone depth - then one has the possibility to switch between the two cases during run time for the cyanobacteria by initializing the field either via i) the current MLD or ii) a constant value (or iii) potentially even a value based on light intensity).

from blom.

The depth of the euphotic zone is currently set to a fixed value dp_ez = 100.0, but the k-index kwrbioz corresponding to this depth is already a 2D field. It should be fairly straight forward to extend the set_vgrid routine to change the calculation for kwrbioz. I suppose it could have a significant effect in ocprod if we decide to change the definition of this field.

from blom.

Hi Tomas, I wouldn't touch ocprod here (since, as you pointed out, setting kwrbioz differently would change a lot in ocprod) and was more thinking about an analogue of kwrbioz for the cyanobacteria only (e.g. kwrcyaz or similar) - which you then can use as described earlier.

from blom.

As a starting point, we could just use kwrbioz as it is, also for the nitrogen fixation? If the process needs enough light(?), then this would be more justifiable than using mld. But since this would break backwards-compatibility we should probably introduce this at a well defined point...

from blom.