Decompositions into atmos/ocean vs tracer/velocities

[dhruvbalwada]

We discussed how to frame the different filtering options that we have and how to relate them to each other in today's quick meeting (18 Nov 2022).

Discussion items:

• [ ] Does CESM also look like CM2.6 results?
• [ ] compute Q_a, Q_o, Q_u, Q_T terms (3 of them already computed).
• [ ] what do the mix terms Q_oa, Q_ut look like physically?
• [ ] Is Q*_oa related to coupling?
• [ ] Can we do some sort of variance decomposition or pattern matching between different terms?
• [ ] Is Q_ua most of Q?

[jbusecke]

I have talked to @dhruvbalwada and after confirming that I understand the above have made these figures:

Panel a) shows the full flux for reference. Panel b) shows the diagnosed small scale contribution. Panel c-f show the various decompositions discussed before (see overleaf for details). Finally Panel g-h show the 'cross'/'remainder' terms. b-h) colorbars are scaled to 5% of the colorbar used in a)

Same as above but for sensible heat flux

I also added some text (including definitions for the the Q* terms) in the overleaf. Please take a look at 5.2.1 Observations.

Some things worth discussing:

• Generally the ocean only case looks very similar to the tracer only case, and the atmosphere only looks similar to the velocity only case
• It seems like the velocity/atmosphere dominates the large scale pattern, but the tracer/ocean contributes a significant amount of signal in the WBC regions. These have very high numbers locally and might actually matter a lot for the global average.
• $Q^*_A$ (panel f) looks quite different between latent and sensible heat flux - with dampening occuring off the equator.