Dear Vijayan Gv,

There are many ways to change the initial conditions while running WRF. However, I would recommend you to modify the wrfinput files rather than to modify the model input data. First, you may need the following Fortran code (read_wrf_nc.f) to read and modify the variables in your wrfinput files. Then, you can study the effects of changing initial conditions on simulation by comparing the wrfout files before and after the modification.

Good luck!

You can modify the WRFinput_**.nc file generated after running real.exe using any NETCDF editing packages in python, MATLAB, etc. Just change the values of the required arrays in input files to whatever you want and then run wrf.exe to generated WRF outputs

Thankyou for your replies, Zuhang Wu and Tejasvi Chauhan .

I have a further query though. A peer of mine suggested that I should try nudging as the model might not work properly if I were to change the values of some parameter at specific locations alone (for eg, if I were to change the winds at just one pressure level or the temperature over a few gridpoints). When is it necessary that I do nudging and when is it sufficient to directly adjust the fields in the wrfinput files?

Hi, Vijayan Gv

When you run the WRF model, it will takes a few hours of "spin-up" time to ensure the stability of model integration. Thus, I'm not worried about the problem caused by changing a few values in the initial field. Of course you can use nudging to moderate the spin-up if necessary.