CAM4 potential temperature (TH) field systematically corrupted

[lehner@climate_unibe_ch]

Hi there

We've been running the fully coupled CESM1.0.1 (compset B_1850-2000_BGC-BPRP) with CAM4 in the 0.9x1.25_gx1v6 and write out 3-dimensional potential temperature (TH) in the atmosphere history files. Unfortunately, TH seems to be corrupted in the lower sigma levels as can be seen in the 1-month output of the lowest level attached (TH.png). The problem vanishes once you go to levels higher than the highest topography. We tested other compsets, write-out frequencies, and of course checked all the other variables. TH seems to be the only variables corrupted and this is independent of the compset.

It is currently unclear to us whether this is a bug in the inter-/extrapolation scheme of CAM4 or not. Clearly, this output cannot be used to interpolate to pressure levels, not can one transfer it back to normal temperature.

Any shared experience or advice on this issue would be greatly appreciated.

With many thanks and best wishes,
Flavio

 

[santos]

I strongly suspect that, when this field was put in, the purpose was to use it in combination with the zonally averaged fields. As a result, I doubt that it is actually computed properly in locations where zonal averaging is not being done, i.e. in latitudes and pressure levels where a significant number of points are not present due to orography.

 

[santos]

Yes, it appears that TH is simply an intermediate step in the calculation of the zonally averaged fields that are output by the circulation diagnostics. So it's really for use in combination with those fields (and/or for debugging those diagnostics), and not calculated properly in locations where those fields are not valid.If you want a full 3D potential temperature field, you should probably use T and the pressure coordinates to create it yourself in post-processing, rather than using TH.

 

[lehner@climate_unibe_ch]

Thanks, Sean, for the quick answers!Indeed, we had for some streams accidentally output TH instead of T (so we lack T in this case), which is why we were interested in whether it's possible to go to T from TH. Unfortunately, this means we lost the T 3D field.