Error when running init_atmosphere_model

[kawilson22]

Hello, I have the most recent version of MPAS and am trying to run a simulation using GFS data from 2011. I am able to obtain the static file with no errors, but when I try to then create the init file, I am met with the following error:

ERROR: extrap_type == 2 not implemented for target_z >= zf(1,nz)
ERROR: *****************************************************************
ERROR: Error in interpolation of t(k,iCell) for k= 55, iCell= 3
CRITICAL ERROR: *****************************************************************

I have attached the log and namelist files for reference. Thank you.

 

[Ming Chen]

Did you use GFS model level data or pressure level data? Where did you download the data?

Is this a global or regional MPAS run?

Please set the option

config_extrap_airtemp = 'lapse-rate'

Then try again. Let me know whether it fixes the problem.

If the case still failed, please send me your streams.init_atmosphere to take a look. Mesh information will also be helpful. Thanks.

 

[kawilson22]

I believe I am using GFS model data from Index of /data/global-forecast-system/access/historical/analysis/201104. I am doing a global run with the x1.10242 mesh.

I changed config_extrap_airtemp as you mentioned and ran into the same error, but it now occurs in the log.init_atmosphere.0016.err file and my out file is different (attached). I have also attached my streams file.

 

[Ming Chen]

Hi,

Would you please download the latest version of MPAS-v8.3.1, and get GFS data from CISL RDA ( NCAR RDA Dataset d084001), then try again?

I understand that GFS data are only available after 2015-01-01 in CISL RDA, while your study period is for 2011. Please change your model time accordingly so that you can use CISL RDA GFS data. This test will help us determine what could be the possible reason behind your issue.

I am suspicious that the trouble you have is actually a data issue, not that something is wrong in the model. Please keep me updated about the results.

Thanks.

 

[kawilson22]

Hello,

I downloaded the latest version of MPAS and ran a simulation using GFS data from the link from 2023. Now all steps of the init_atmosphere_model work fine, so I think it is an issue with GFS. Is there another dataset like ERA or MERRA-2 that would be better for a 2011 case, or are there model parameters that I need to change in order to account for the GFS differences? Thank you!

 

[Ming Chen]

I am sorry to answer late but this is because I am on a business trip.

I would suggest that you try the ERA5 data, which can be easily processed for initializing MPAS. Please download the source codes at :

GitHub - NCAR/era5_to_int: A simple Python script for converting ERA5 model-level netCDF files to the WPS intermediate format

A simple Python script for converting ERA5 model-level netCDF files to the WPS intermediate format - NCAR/era5_to_int

github.com

Then follow the instruction to run the script and produce intermediate file for MPAS initialization.

Let me know if you ahve any issues.

 

[kawilson22]

Thank you, I reran MPAS with the ERA5 input data and it worked. Is it correct that I only need the intermediate ERA5 file for the start time of the model initialization? And that I need to download all data files from both the pressure level analysis and surface analysis to use era5_to_int?

 

[Ming Chen]

Thank you, I reran MPAS with the ERA5 input data and it worked. Is it correct that I only need the intermediate ERA5 file for the start time of the model initialization?

You are right that only ERA5 at the initial time is required if you run global MPAS simulation.

For regional simulation, all the ERA5 files during the integration period are required.

And that I need to download all data files from both the pressure level analysis and surface analysis to use era5_to_int?

This is correct. Surface analysis is needed because we need initial values of surface/soil information for LSM module.