Something about WRF

[dsf...@gmail.com]

Hi,

I have some questions about the competetion rules about WRF. Would you mind telling me more specifically.

1.Are the data sets given to us real raw statistics which have to be pre-processed by WPS or Can we get the wrfinput* files and wrfbdy* file?

2.Will the namelist.input be given along with the corresbonding dataset? If so, can we adjust the parameters inside the namelist such as time_step, dynamical parameters and so on?

3.How to Verify the correctness of the results? Will there be a precision requirement?

I am looking forward to your reply!

Regards,
Aden

[John Michalakes]

Hi Aden,

The input datasets will be already pre-processed and ready for ingest into WRF.  These will be either wrfinput or wrfrst files plus the wrfbdy files.  A namelist.input file will be provided for each case.  You're free to modify anything you find in the namelist.input file for a case, but be advised that changing options may make cause problems reaching agreement with the reference solution from which the output from the simulations will be judged.

Regarding verification criteria, that will depend on the case.  But it's likely that verification for at least one of the cases will involve a statistical comparison with a reference output data set using a small program that will be provided for this purpose. The program will output the number of digits agreement between the competitors' output and the reference case.  Other cases may use different criteria and verification mechanism.

Regards,

John

[Jeff Shafer]

John, thank you for your response - it's very helpful!

Jeff