Error "TSK" is unrealistic" when running real.exe of WRF

[cancan_6]

My simulation period is 2008-12-15 to 2010-01-01. The WPS parts were shown successful, but when I was running the real.exe, the model stopped at 2009-05-29_12:00:00 with the error "TSK" is unrealistic, problems for seaice post"; and only wrfinput_ d01 obtained, while wrfinput_d02 and wrfinput_d03 didn't occur. I tried to run the month May of 2009 only, the same error occurred (model also stopped at 2009-05-29_12:00:00).

The error part in rsl.error file and namelist.input were attached for reference. Thanks a lot in advance for any suggestions to solve this problem.

 

[kwerner]

Hi,
I have a couple of thoughts that could potentially help.

1) I assume you do have surface input data, correct?
2) Because you are running a long simulation, you should be using a higher-resolution SST input, along with the sst_update=1 option. We recommend this to anyone simulating more than 5 days.
3) You should turn off cu_physics for domain 3 because at 3km resolution, the model doesn't need any cumulus parameterization.

 

[cancan_6]

Thanks a lot for your sugestions. With these changes (turn off cu_physics of domain3 and set the sst_update parameters respectively), I still got the same error. I also tried to set smooth_cg_topo=.true. as mountainous area is included in my study region, but same error occurred.

FNL data and GEOG data are used for my simulation.

 

[cancan_6]

When I used other FNL file to replace the FNL file of 2009-05-29_12:00:00, the real.exe could run successfully and all wrfinput files obtained. In this situation, I may try to use other reanalysis data. I may figure out this problem later, but I still wonder why this happened. Thanks.