MERRA reanalysis fields in specified dynamics runs (FSDW compset)

[duderstadtk@gmail_com]

When using MERRA reanalysis fields as opposed to GEOS5 for our 2000-2001 model runs, we understand that we need to modify the data files in user_nl_cam after creating a case using the FSDW compset. For example, staring a run on October 1, 2000 we need to modify user_nl_cam to include:
met_data_file='2000/MERRA_19x2_20001001.nc'
met_data_path='$DIN_LOC_ROOT/atm/cam/met/MERRA'
met_filenames_list='$DIN_LOC_ROOT/atm/cam/met/MERRA/filenames_list.txt'

Do we also need to modify the following namelist variables…or can we retain these GEOS5 settings for our initial and boundary conditions?
CAM_NML_USE_CASE --> currently sd_waccm_goes5
bnd_topo --> currently USGS-gtopo30_1.9x2.5_phys_geos5_c100929.nc
ncdata --> waccm0_geos5_2x_88L-2005-01-01_c110419.nc

 

[mmills@ucar_edu]

The CAM use case and topography files should remain the same.

If you would like look at an event within a year or two of your start time, you should obtain a new initial conditions (ncdata) file from a previous SD-WACCM run for your start time. If you would like to send me a list of what start dates you want for your runs, I can see if we have ICs available for those times.

I will also look into making available SD-WACCM ICs for January 1 of each year that we have available, on our input data subversion repository.

 

[duderstadtk@gmail_com]

Thanks! This leads in to another question....

We are interested in looking at SPE events in November 2000 and subsequent chemistry and dynamics in the Northern Hemisphere throughout the winter of 2000-2001.

How much of a "spin-up" time should we give the specified dynamics model to allow intermediate chemical species and other physical parameters to provide adequate results?

(The answer to this question will determine the date of the initial conditions.)

 

[mmills@ucar_edu]

I would think you should start no later than the beginning of that year if you are interested in events at the end of the year. The IC files we have are for a chemical mechanism with 122 species, of which the standard 60 species used in WACCM (and SD-WACCM) are a subset. So there will likely be some time period required for the chemistry to adjust in the lower atmosphere, but you will probably not notice for your middle atmospheric studies.

So I would spin up a run from January 1, then save a restart close in time before the events you want to study (say November 1) if you want to tweak the parameters of the event in different runs starting from that point.

 

[mmills@ucar_edu]

We are in the process of adding IC files for SD-WACCM for January 1 of each year from 1989 through 2011 to our subversion input data repository. These will be available under ccsm4_init/wa4_cesm1_1_b02_geos5_2x_sim153f.

 

[duderstadtk@gmail_com]

Thanks for the guidance and path to the ncdata files!

Our current question concerns beginning the spin-up run on January 1, 2000.

The specified dynamics compset FSDW uses default atm/cam/chem/2000-2100_RCP45 surface and aircraft emissions files for CH2O, CO and NOx (namelist variables ext_frc_specifier and srf_emis_specifier). These files begin on January 15, 2000...not January 1, 2000...causing the model run to exit during initialization.

We plan to run the model with MERRA met fields initialized with the ncdata file wa4_cesm1_1_b02_geos5_2x_sim153f.cam2.i.2000-01-01-00000.nc

Therefore, we can either:

1) Begin the spin-up run on 2000-01-15 instead of 2000-01-01 with the default 2000-2100_RCP45 emissions files

- OR -

2) Begin the spin-up run on 2000-01-01 but specify new emissions files within the 1850-2000_emis directory (these files extend to 2009).

Both methods result in completed model runs. Any advice on the best choice...or additional alternatives?

[Recall that our goal is to look at specific SPE events in November of 2000 and their impact through the 2000-2001 winter].

[Also note: the namelist variable srf_emis_type is set to 'SERIAL' in our current runs]

Thanks!

 

[mmills@ucar_edu]

Start on January 15. You will have a comparable spin-up time, and you won't have to deal with changing input files.

 

[snowyang1988@gmail_com]

hello ,I need SD-WACCM ICs for January 1 of 2013, but there is only for 1989-2011 on the input data subversion repository. Where can I get the data updated ?Thank you.Jeff

 

[mmills]

NCAR cannot distribute all of the meteorological fields on our input data subversion repository. You should be able to find what you need on the Earth System Grid at https://www.earthsystemgrid.org by searching for MERRA or GEOS5. Let us know if you have any difficulties finding it.

 

[mmills]

See also, this post, which includes the location on the ESG for MERRA files: https://bb.cgd.ucar.edu/comment/1008077#comment-1008077