Downscaling CFS files

[janac4]

Hello everyone,
I am familiar with WRF model using GFS data as input forcing files, but now I want to analyse CFS seasonal operational forecasts. Since the resolution is very coarse (100km) I want to perform a downscaling using WRF.
Is there a way to do it not dynamically? I'm thinking about running a 1-hour simulation WRF run so the first wrfout is the result of the downscaling of the first CFS input file, but I'm not sure this is right.
Also, I found that with all_ic_times=true I can obtain wrfinput files at each interval time. How is this data created? Could it be interpreted as a downscaling method of the input files? Is there any physics involved?

 

[Ming Chen]

By running WRF driven by CFS data, you can obtain dynamically downscaled data.

However, if you simply interpolate CFS data to WRF grids (e.g., by producing wrfinput at all times), the results are still from CFS and I won't think such data are 'downscaled' data.

 

[janac4]

I understand, thanks.
Is there any guide to run WRF with CFS data? I found the CFS repository but there are multiple files filtered by variable or layer type (pressure levels, flux variables, ocean, etc) and also files from CFSR which I assume I don't need if I want to run a forecast simulation (data in the future).
Also, these seasonal forecasts are large in time (up to 9 months), so I would appreciate any guidance on interval time-step for example as I have only run 3-4 forecasted days runs.

 

[janac4]

I understand, thanks.
Is there any guide to run WRF with CFS data? I found the CFS repository but there are multiple files filtered by variable or layer type (pressure levels, flux variables, ocean, etc) and also files from CFSR which I assume I don't need if I want to run a forecast simulation (data in the future).
Also, these seasonal forecasts are large in time (up to 9 months), so I would appreciate any guidance on interval time-step for example as I have only run 3-4 forecasted days runs.

I found that PGB CFS files should be used, but it is also important to have SST data in seasonal simulations. Some people use flux CFS files coupled with PGB but others say that SST variable is in OCN files.
I don't really know what to use. Also, OCN files have many variables and different resolutions (either 0.5º or 1º), is it possible to run WRF with PGB files (1º) with OCN (0.5º)? Will this work?

 

[Ming Chen]

if CFS data provide all variables required for running WRF (U, V, T, Q, surface and soil data ), then it should be fine to run WRF driven by CFS .
I am not sure whether CFS data include SST, if so, you can use SST from CFS. If not, you may download other SST data and use it for WRF run. her his a document that demonstrates how to use different SST:

https://www2.mmm.ucar.edu/wrf/users/tutorial/presentation_pdfs/201707/wps_advanced.pdf