Meridional wind noise/oscillation

[fgibbon]

Hi everyone,

I've been working with CESM2.1.5 and we've found something in the meridional wind field of various simulations that we are concerned about.

At high latitudes, there appears to be noise in the V component of wind causing a saw-tooth pattern in the zonal-mean V and discontinuities in the spatial pattern. This is present in aquaplanet simulations I have run myself, as well as in CMIP6 data from NCAR (available on the UK's CEDA archive) including AMIP, historical, piControl experiments.

Here is a link to a Github repository which shows this artefact: https://github.com/fgibbon/cesm_v_wind

We are wondering if this is a known issue, and if it is something we ought to be concerned about affecting our analyses.

Many thanks,
Freya

 

[pel]

Hi Freya,


Thank you for your detailed analysis!

The potential issue you describe was discussed in this paper:

https://journals.ametsoc.org/view/journals/mwre/139/9/2011mwr3607.1.xml

In idealized tests, the authors identified small-scale meridional waves that persist and are not fully eliminated by either second- or fourth-order divergence damping operators. I was not aware that these would also appear in full-physics simulations with complex parameterizations.

A related factor is that the Fourier filters (applied in addition to the divergence damping) effectively remove linear and nonlinear computational instabilities, but they do so only in the zonal direction. No equivalent filtering is applied in the meridional direction. This asymmetry could explain why the noise manifests in the meridional direction rather than the zonal one.

To help mitigate grid-scale noise in divergence fields (including away from the polar regions), we implemented fourth-order divergence damping to replace the second-order divergence damping. This is described in the following paper:

https://journals.sagepub.com/doi/10.1177/1094342011410088

But as you show this is not enough to remove the noise caused by the polar filter (even makes it worse!).

Peter

 

[pel]

It might be possible to alleviate the noise using the recommendation in Jared's paper:

"Employing a damping coefficient that neglects the latitudinal variation of the grid cell area will likely damp these meridional waves more effectively, but the polar singularity will then be more apparent."

One might need to significantly reduce time-step though ...