computing diagnostic variables

[Marco Miani]

Hi.

I am merging a series of wrfout files, in this specific case only variables P and T:

# Load available files

ds_list = []

for f in wrf_files:

try:

ds_list.append(xr.open_dataset(f)[["P", "T"]].isel(bottom_top=0))

except FileNotFoundError:

print(f"Missing file: {f}, skipping.")

# Merge while keeping missing time steps as NaN

ds_merged = xr.concat(ds_list, dim="Time", fill_value=np.nan)

After ds_merged is completed, saved to netcdf on disk, and reopened with wrf.getvar I am failing to get T2 (air temp at 2 meters).

Variables T and P alone (as shown above in the snippet) are not sufficient to compute (internally) the diagnostic variable. My questions, if you know:

1. what are the needed variables to compute T2, so to include them in the merging?

2. Is there any better, more efficient, faster, less CPU-intesive way do merge/concat a large list of files? I am doing this because I only need a handful of variables, and only at surface level. wrf get var does not seem to offer this flexibility, xarray does. So I am doing the heavy lifting in xarray, saving to disk and reopening in wrf.

Thanks for the advice and happy coding
Marco

[Will H]

My guess is it is using a combination of these two commands

https://github.com/NCAR/wrf-python/blob/main/src/wrf/g_dewpoint.py  - search 2m

https://github.com/NCAR/wrf-python/blob/main/src/wrf/g_temp.py - and this main calculation?

But i'm not sure.  

I also see that T2 is a variable with a header that can be called with wrf.getvar

<class 'netCDF4._netCDF4.Variable'>
float32 T2(Time, south_north, west_east)
    FieldType: 104
    MemoryOrder: XY
    description: TEMP at 2 M
    units: K
    stagger:
    coordinates: XLONG XLAT XTIME
unlimited dimensions: Time
current shape = (1, 405, 435)
filling on, default _FillValue of 9.969209968386869e+36 used