Accessing Vertical Coordinates for CONUS404-PGW Zarr

[ballight42]

Hi,

I am currently working with the hourly conus404-pgw datasets and had a question about vertical coordinates for 3D variables like U and V. I am trying to interpolate these variables to specific pressure levels. But I noticed that the bottom_top dimension in these variables is missing. If it was intentionally omitted, is there a recommended way to interpolate these variables onto the vertical levels for the PGW scenario? Many thanks in advance for your time and I will appreciate any instructions:)

The code I used to access data:

import intake
import fsspec
endpoint_url = "https://usgs.osn.mghpcc.org/"
zarr_url = "s3://hytest/conus404-pgw/conus-pgw_hourly.zarr"
fs = fsspec.filesystem("s3", anon=True, client_kwargs={'endpoint_url': endpoint_url})
mapper = fs.get_mapper(zarr_url)
ds = xarray.open_zarr(mapper)
ds['U']

The output coordinates for variable U:

xarray.DataArray'U'time: 376945 y: 1015 x_stag: 1368)

which I expected to be

xarray.DataArray'U'time: 376945 bottom_top:50 y: 1015 x_stag: 1368)

[pnorton-usgs]

Hello Yan,

The raw WRF model output files do indeed have the U- and V-components of wind on multiple atmospheric levels. The zarr dataset that we provide on our OSN POD is created from a subset of WRF model output variables. As such, the U and V variables in the zarr dataset only include the lowest model level of those wind components. To get the original multi-level versions of those variables you would need to download the model output 3D variables from the NCAR RDA (@amsnyder - do you have a link for RDA access to the CONUS404-PGW 3D files?). The metadata for U and V are a bit imprecise and I will update the description in the zarr datasets to denote that only the lowest model level is included.

[ballight42]

Hi Parker,

Thanks for the clarification! That really helps, I'll use the full WRF output for the pressure level data as suggested. I'm also looking for the most efficient way to download these files. Is there an RDA link for this similar to the one for CONUS404? I appreciate any suggestions to speed up data transfer.

[lulinxue]

Hi Parker,

Thanks for the clarification! That really helps, I'll use the full WRF output for the pressure level data as suggested. I'm also looking for the most efficient way to download these files. Is there an RDA link for this similar to the one for CONUS404? I appreciate any suggestions to speed up data transfer.

Hi Yan, this is Lulin Xue from NSF NCAR. I am the lead of CONUS404 datasets and related projects. The CONUS404 PGW dataset is not officially available on GDEX (replacing RDA). I wonder if you have an account on NCAR HPC (Derecho and/or Casper). If this is the case, you can directly access the CONUS404 PGW data on either system. I can provide more details through emails if that works for you. Please reach out to me at xuel@ucar.edu or conus404@ucar.edu for any CONUS404 related questions.

Thanks!
Lulin