From old vertical levels (WRF 3.6) to hybrid vertical levels (WRF 4.1.4)

Topics related to the dynamics/numerics of the WRF model
Post Reply
Posts: 8
Joined: Mon May 04, 2020 1:26 pm

From old vertical levels (WRF 3.6) to hybrid vertical levels (WRF 4.1.4)

Post by davidsuculum » Mon May 04, 2020 1:47 pm

Hi all.

I would like to convert a list of vertical levels that I had set for an old version of WRF (WRF 3.6) to the new hybrid vertical coordinates in WRF 4.1.4. Please note that the vertical levels I had been using in WRF 3.6 were ones I defined, not those automatically generated. (For instance, I put more vertical levels in the lower layers of the atmosphere).

Perhaps, (if this is a more accessible option than the previous one), other approach could be to specify a list of vertical levels in terms of the pressure and obtain the correspondent vertical hybrid coordinate.

Thank you very much,

Posts: 105
Joined: Mon Apr 23, 2018 9:03 pm

Re: From old vertical levels (WRF 3.6) to hybrid vertical levels (WRF 4.1.4)

Post by davegill » Mon May 04, 2020 5:58 pm

The eta locations in the terrain following (TF) vertical coordinate are defined by a normalized reference pressure, with an identically isobaric surface at the model lid. The eta locations for the hybrid vertical (HV) coordinate are defined to adjust towards an isobaric surface at an identified eta value (etac in the namelist).

The term "convert" is not entirely possible, you cannot cleanly take TF data and make it into HV data. You could do vertical interpolation at each (i,j), but that is likely to introduce yet more noise in your initialization step and along the boundaries. Also, any effort to make old data work with later versions of the real program and WRF model is really going off on your own, and there is no support available for your changes that would be required in the metadata, additional HV fields, and boundary files.

A better idea would be as you suggest - just use the same eta levels from your previous TF simulations and rerun the new version of the real program. You can easily use the existing monotonic list of eta values from the v3.6 code and put those levels in the namelist for the v4 (and after) WRF-system pre-processor (real) and the model (WRF). In the &domains namelist record, just include your selected eta values when you re-run the real program (here is an example of the syntax; of course, fill in your own levels):

Code: Select all

 eta_levels                          = 1.000, 0.990, 0.978, 0.964, 0.946,
                                       0.922, 0.894, 0.860, 0.817, 0.766,
                                       0.707, 0.644, 0.576, 0.507, 0.444,
                                       0.380, 0.324, 0.273, 0.228, 0.188,
                                       0.152, 0.121, 0.093, 0.069, 0.048,
                                       0.029, 0.014, 0.000,
The original metgrid data from v3.6 probably would work for the v4 and later real program (which would then feed the new HV coordinate into the WRF model). Note that the newer (for example, v4.2) metgrid output would have different default static field values compared to your original v3.6 metgrid files, so any sort of sensitivity testing for the HV vs TF results would have additional confounding factors.
Dave Gill

Post Reply

Return to “Dynamics and Numerics”