Wind components on different grids

Atmosphere AUS2200
1

Hi everyone,

The output wind component variables from AUS2200 appear to be on different grids (staggered grids?) with:

“U COMPNT OF WIND AFTER TIMESTEP” fld_s00i002
on a lat, lon grid

and
“V COMPNT OF WIND AFTER TIMESTEP” fld_s00i003,
on a lat_0, lon_0 grid.

I was wondering what are the appropriate options for working with these different grids. I think in a previous meeting @pjs548 mentioned that wind component variables on the same grid were available. From discussing with @cbengel I understand that by default AUS2200/RNS do not output these wind component variables that are on the same grid. @andrewb1 I think is also interested in this.

If it is possible to output the wind component variables on the same grid in future AUS2200 runs could I get some help/tips on how to set up AUS2200 to ouput these variables?

        I've currently been looking in:

app/um
and doing a rose edit and navigating to:
STASH Requests and Profiles - STASH Request

where it seems you can select or deselect different variables.
Am I on the right track here? Is this the place where you can select what variables to output from AUS2200?

Thanks
Chris

2

Yes, U will be offset by half a grid step in longitude and V by half a step in latitude.

You’re on the right track with the output. Try maybe s15i002 ‘U WIND ON MODEL LEVELS B GRID’. You’ll need to select options for dom (vertical levels), tim (time processing) and use (which file to output to)

1 Like
3

Hi @cchambers, is regridding the staggered winds to a common grid an option for the analysis you want to do, or do you require native model output?

I usually regrid everything to the thermodynamic grid in postprocessing / analysis, so that all variables are collocated for the purpose of analysis.

2 Likes
4

Hi @bethanwhite , yes that is an option but I was not sure how best to do it. Yesterday @andrewb1 also put up a python script on GitHub which today has been added to the Cookbook that also does this:
https://atmosphere-cookbook.readthedocs.io/en/latest/Recipes/AUS2200_destagger_winds.html
Do you use a similar method to this? I’m yet to check it fully but will try out soon on AUS2200 data.

5

Hi @cchambers, was trying to find this code snippet for you yesterday but was having login issues with NCI (from the Met Office). I’ve been doing a straight regrid in iris as follows, hopefully the logic explains itself - this is to get the uwind onto the thermodynamic grid. (The same logic can be applied to the vwind).

u_10m_tgrid = u_10m.regrid(temp_1pt5m, iris.analysis.Linear())

Docs:
https://scitools-iris.readthedocs.io/en/latest/userguide/interpolation_and_regridding.html

I guess this might not be useful if you’re not working in iris, but I am sure other packages have similar regridding functions.

1 Like