Temperature error in specific area using WRFv4.6.1

[Litta]

Hi,
I used the WRF v4.6.1 model with 5 km resolution over the Asian region to generate analysis data. I assimilated surface observations using WRFDA and compared the results with ERA5.
I found large temperature differences in the southern parts of India and Australia. I also compared the results with SYNOP data and observed significant discrepancies around this area. But I got very good score over JAPAN.
This error is due to data assimilation. Please check the attached image and kindly suggest possible solutions.

 

[Litta]

@jakabaesj9 Thank you very much for your suggestion.
I used the CV3 option. I would appreciate your advice on how to reduce this error by adjusting the quality control parameters for those regions.

 

[Litta]

Please check the attached images to see the difference in temperature before and after assimilation. I tried tuning the CV3 option parameters, but there was no change.