Issues in modifying the WRF restart file

[freekiller]

Dear all,

I encountered a problem when I conducted several sensitive tests with terrain modified in the WRFrestart file. The WRF model (version 4.1.3) was configured in a one-way nesting, triply nested grids with horizontal grid sizes of 9, 3, and 1 km. Firstly, the WRF model, initialized at 0000 UTC 01 July 2021, ran for 12 hours with hourly restart outputs (namely control run). Then, I tried to remove the terrain (i.e., HGT) in the innermost domain (i.e., D03) at the first sixth-hour WRF restart file (i.e., wrfrst_d03_2021-07-01_06:00:00), and restarted the simulation for another 6 hours. However, the simulated results without terrain show close to those of the control run. Besides, the effects can also be viewed from the wind field at 925 hPa (shown below 925hPa.png). I confirmed that the terrain was removed in the wrfrst and consequent wrfout* files (Fig .2).

I can get reasonable results by removing terrain in the wrfinput file (i.e., wrfinput_d01) with the same approach. Therefore, I'm guessing other variables that need to be changed besides modifying the terrain (i.e., HGT) in the WRF restart file.

Any suggestions will be appreciated.

Thanks.

Jerf

 

[kwerner]

Hi Jerf,
Since restart simulations use wrfrst* files, as well as the wrfbdy* file, I would assume the boundary condition file would need to be modified, as well. Otherwise, the conditions at the boundary are still including terrain, which will propogate across the domain.

 

[freekiller]

Hi Jerf,
Since restart simulations use wrfrst* files, as well as the wrfbdy* file, I would assume the boundary condition file would need to be modified, as well. Otherwise, the conditions at the boundary are still including terrain, which will propogate across the domain.

Hi Kwerner,
Thanks for your reply. I will have a try.

Jerf