We are running retrospective experiments for the Water in the West projects and we need to get them verified. This is a joint EPAD-AVID project that employs a dual domain (global and nest) to seek improvements for Atmospheric Rivers reaching the west coast of the US. The global domain runs at ~13 km and the nest at ~3 km. The next is located over the N Pacific Ocean and the Western part of the CONUS. In practice, this probably will appear as two two different models from the point of view of the vx system.
The verification we need is the "basic" one. Verification specific to Atmospheric Rivers is being handled separately by colleagues at PSL. I am certain that there are a lot of details to discuss, such as what truth to use. For example, for the global domain we can certainly benefit from comparisons against GFS analyses. And if a global precip dataset is available, like IMERG, that would also be helpful. For the nest, we are interested in vx against radiosondes and sfc obs, in addition to pcp. In summary, a starting point is to setup the same types of vx that GSL normally uses for global and regional models.
Per Matt, the model is a UFS FV3 run using the HAFS workflow.
From Anders:
Here's a location on hera that has both netcdf and grib output for both the global domain and the high-res nest. Output from the global domain is named atmf120.nc and output from the nest is named atm.nest02.f120.nc. The netcdf output is split into atm and sfc files.
As Ligia mentioned, we are hoping to get basic verification up and running as soon as possible (the next month or two would be ideal) since we'll use it for our physics and DA development on this project.
Also:
-the team will be rerunning the model over and over for a few retro cases to test continued improvements... mentioning this since it may be important in terms of figuring out the most agile way to import new runs and get vx results
From LIgia:
We are running retrospective experiments for the Water in the West projects and we need to get them verified. This is a joint EPAD-AVID project that employs a dual domain (global and nest) to seek improvements for Atmospheric Rivers reaching the west coast of the US. The global domain runs at ~13 km and the nest at ~3 km. The next is located over the N Pacific Ocean and the Western part of the CONUS. In practice, this probably will appear as two two different models from the point of view of the vx system.
The verification we need is the "basic" one. Verification specific to Atmospheric Rivers is being handled separately by colleagues at PSL. I am certain that there are a lot of details to discuss, such as what truth to use. For example, for the global domain we can certainly benefit from comparisons against GFS analyses. And if a global precip dataset is available, like IMERG, that would also be helpful. For the nest, we are interested in vx against radiosondes and sfc obs, in addition to pcp. In summary, a starting point is to setup the same types of vx that GSL normally uses for global and regional models.
Per Matt, the model is a UFS FV3 run using the HAFS workflow.
From Anders:
Here's a location on hera that has both netcdf and grib output for both the global domain and the high-res nest. Output from the global domain is named
atmf120.nc
and output from the nest is namedatm.nest02.f120.nc
. The netcdf output is split into atm and sfc files./scratch2/BMC/wrfruc/Samuel.Trahan/westwater/crashing-nested-l127c768-with-output
As Ligia mentioned, we are hoping to get basic verification up and running as soon as possible (the next month or two would be ideal) since we'll use it for our physics and DA development on this project.
Also: -the team will be rerunning the model over and over for a few retro cases to test continued improvements... mentioning this since it may be important in terms of figuring out the most agile way to import new runs and get vx results