How to manipulate wrf variables while integrating

Any problems/questions that might not fall into the other topics in this category.

How to manipulate wrf variables while integrating

Postby Souri_atm » Thu Aug 10, 2017 6:23 pm

Dear All,

Suppose we want to assimilate some data every 5 minutes into the model, we may come up with two solutions: i) halting model every 5 minutes and changing wrfrst outputs, and continuing the simulation with the modified initial values and ii) calling the assimilation function inside of WRF while integrating.

I am wondering if anyone knows how to perform the second way. Which file or subroutine is dealing with integration? where am I allowed to call my function?

Thanks a lot!
Souri_atm
 
Posts: 1
Joined: Thu Aug 10, 2017 6:17 pm

Re: How to manipulate wrf variables while integrating

Postby amt1204 » Wed Jun 06, 2018 1:00 pm

To the best of my knowledge, I don't think there is a function within WRF for data assimilation. That is why WRFDA exists (http://www2.mmm.ucar.edu/wrf/users/docs/user_guide_V3.9/users_guide_chap6.htm for user guide and http://www2.mmm.ucar.edu/wrf/users/download/get_sources.html#WRFDA for download). If the options provided in WRFA are not satisfactory (such as you may want to do a more sophisticated EnKF), then perhaps your first strategy should be implemented.

Nonetheless, if you want to edit the integration files, I suggest looking into dyn_em/ (I presume you are looking for ARW and not NMM). In it are all of the files for the dynamical core. I haven't dealt with all (or any, for that matter) of these files. But simply by the names, you can figure out what each does. For example, 'module_first_rk_step_part1.F' is one part of the Runge-Kutta integration algorithm, presumably this is one file you are looking for (another being part2).

Also, I do want to comment on your strategy for data assimilation. You only need to be assimilating every five minutes if you are utilizing either a four dimensional assimilation (such as en extended Kalman Filter or 4D-VAR), nudging, or forming a very short term reanalysis. Even five minutes is a very rapid interval for four dimensional data assimilation- the model needs some time to restore balance.
amt1204
 
Posts: 10
Joined: Mon Dec 22, 2014 9:57 pm

Re: How to manipulate wrf variables while integrating

Postby Phinney » Fri Jul 06, 2018 7:42 am

What time interval would you suggest for four dimensional data assimilation, Amt1204? I guess I thought the intervals should be shorter than necessary.
Phinney
 
Posts: 1
Joined: Fri Jun 29, 2018 5:51 am


Return to Miscellaneous

Who is online

Users browsing this forum: No registered users and 6 guests