[ECmean]: 3d ocean variable regridding errors

[jhardenberg]

machine

Levante

catalog

climatedt-phase1

version

main

version

main

What happened?

When using AQUA-core main the ECmean diagnostic reports the following :

climate_metrics-ecmean.log:2026-02-18 00:36:40 :: Regridder :: ERROR    -> Regridder for vertical coordinate depth not found.
climate_metrics-ecmean.log:2026-02-18 00:36:40 :: Regridder :: ERROR    -> Cannot regrid variables ['so', 'thetao']
climate_metrics-ecmean.log:2026-02-18 00:36:59 :: Regridder :: ERROR    -> Regridder for vertical coordinate depth not found.
climate_metrics-ecmean.log:2026-02-18 00:36:59 :: Regridder :: ERROR    -> Cannot regrid variables ['so', 'thetao']
climate_metrics-ecmean.log:2026-02-18 00:37:59 :: Regridder :: ERROR    -> Regridder for vertical coordinate depth not found.
climate_metrics-ecmean.log:2026-02-18 00:37:59 :: Regridder :: ERROR    -> Cannot regrid variables ['so', 'thetao']
climate_metrics-ecmean.log:2026-02-18 00:38:18 :: Regridder :: ERROR    -> Regridder for vertical coordinate depth not found.
climate_metrics-ecmean.log:2026-02-18 00:38:18 :: Regridder :: ERROR    -> Cannot regrid variables ['so', 'thetao']

These are annoying because the log contains 'ERROR' (which tends to upset (understandably) operators). On the other hand, these seem harmless, because our current bias and PI tables do not use 3D ocean variables....

Code to reproduce the issue

Error messages

No response

Are you interested in making a pull request?

None

[mcadau]

I've noticed this pseudo-error too while running aqua analysis and I was quite puzzled, indeed.
So, the idea would be to let ECmean ignore the 3D Ocean variables, so that the regridder is not triggered by those variables?

Anyway, since I took a deeper look on ECmean code while facing #168, if you want I can investigate this issue more in detail. What do you think?