Using ATM_PRESS in combination with M2FLATHER?

[Timh37]

Dear all,

I have a question related to the CPP option ATM_PRESS, which in the case of SOLVE3D imposes the atmospheric pressure onto the 3D pressure gradient equations.

I am running a 3D regional set-up which I'm forcing at the lateral boundaries with an ocean reanalysis that does NOT include the inverse barometer effect in its SSH. Surface forcing is from ERA5 (using bulk formulae), including atmospheric pressure.

So far I've been running with ATM_PRESS off, so no direct effect of atmospheric pressure. However, I'm interested to see the effect of atmospheric pressure. Specifically, I want to know if turning ATM_PRESS=on is equivalent to simply adding the inverse barometer correction as described in for example Stammer & Hutteman (2008), i.e. eta_IB = - p'a_/(rho*g) with p'_a = p_a - global average(p_a), or if there are important non-linear feedbacks with atmospheric pressure - are these included in the code?

When switching on ATM_PRESS, obviously there is a problem at the boundaries when SSH is prescribed that does not include the inverse barometer effect (as discussed here as well: viewtopic.php?f=14&t=1265 ). So, my question is how I can ensure consistency at the boundaries? Is there code to handle that, or should I apply the inverse barometer effect to my SSH boundary conditions? In case of the latter, what reference pressure to use? (e.g. the global_average(p_a)? the average over my specific region? constant or time-varying?).

Thanks in advance.

[m.hadfield]

Non-linear feedbacks are not included in the code.

The PRESS_COMPENSATE option should be used in the case you describe where the model providing the boundary conditions does not allow for the inverse barometer effect.

[Timh37]

Perfect, thanks!