This exercise provides the opportunity to go through the entire set-up and run process for WPS and WRF, but instead of us dictating how you should set up your case, you can be creative and play around with domains.

Note For your actual research you may need high resolution domains, possibly with a large number of grid points, but due to computation limitations, for this exercise, please use resolution and sizes similar to those you've been practicing with so far. This exercise is only meant to cover the steps and considerations you need when making your own case – not the actual size of the problem.

Recall that you need to follow the 5 processes depicted below. You will run the WPS processes in the wps directory, and you'll run real and wrf in the wrf directory.

Download Input Data

If you already have data, or a data site you are familiar with, feel free to use that. Otherwise, we have several data sets available for your use.

• If interested in obtaining real-time GFS forecast data, you can read how to do so here. These are only available for a short period of time, so it's best to choose a time period in the past few days.


• Feel free to use any of the datasets provided to you for the tutorial exercises. Just keep in mind that these data are specific and may not give the results you are interested in for your particular domain.

  /glade/campaign/mmm/wmr/wrf_tutorial/input_data/colorado_blizzard/fnl_20210 : Colorado Blizzard (Nov 2019) Global Data
  
  /glade/campaign/mmm/wmr/wrf_tutorial/input_data/severe_outbreak/fnl_20230 : March-April 2023 Severe Weather Outbreak Global Data
  
  /glade/campaign/mmm/wmr/wrf_tutorial/input_data/matthew/gfs.0p25.2016100 : Hurrican Matthew (Oct. 2016) Global Data
  
  /glade/campaign/mmm/wmr/wrf_tutorial/input_data/climate_cfsv2/cdas1.2016100 : Oct 2016 Global Data - keep in mind you'll need to use 2 different Vtables (see the Climate Run Example for info on the Vtables).
  
  



• You can also visit this page to see other options for free input data. This site provides several data types available from NSF NCAR's RDA team. All types are available in /glade/campaign/collections/rda/data/ds* (where "ds*" is the data type number).

Set-up and Run Geogrid

• Edit namelist.wps to configure the simulation domain. Use namelist options such as 'e_we/e_sn/i_parent_start/j_parent_start/dx/dy' to determine the size and placements of your grid. Determine whether you want any nested grids, and if so, make the appropriate namelist changes for multiple domains.

  Here are some questions you might want to consider:
  
  

  1. Do I need to use nested domains to reach my target resolution?
  
  
  2. Are there any potential problems with the locations of my boundaries?
  
  
  3. How long will it take a typical air parcel to pass from one side of my domain to the other (consider the importance of domain size)?
  
  
  4. Based on other exercises that I've run, roughly how computationally expensive do I expect the domain setup to be, and how long do I expect this simulation to take?

  
  
• Use util/plotgrids_new.ncl or the 'ncview' option to view your domain. Play around with the namelist until you have the domain where you want it. If you would like to set up a nested run, set the coarse domain first, and once you are satisfied with the coarse domain, use the above-mentioned tools to outline the nest domain (i.e., to find out the starting and ending endices of the fine domain). Then you can compute the actual number of grid points in the nest domain.


• Now run geogrid.

Set-up and Run Ungrib

• You will need to change the namelist dates to match your case. Also make sure interval_seconds is set correctly to reflect the number of seconds between data intervals for the data you are using.
  
  
• Link the GRIB data into the current directory.


• Link–in the correct Vtable.


• And finally, run ungrib to generate intermediate files.

Set-up and Run Metgrid

Simply run metgrid to create the met_em.d0* files for the real.exe program.

Set-up and Run WRF

1. Edit namelist.input to reflect your case domain and date information. Use what you have learned through previous exercises to determine what variables need to be modified.


2. Make sure to link–in the appropriate files and then run real and wrf.

View the rsl.out* files to ensure the run was successful. If you wish, view some of your data with the available graphics options.