Cold bias due to fresh snow?

[wallis]

I am running 24hr forecasts initialized from GFS with NoahMP.
Over areas of snow there is a very strong cold bias. I suspect this is because when the model is started it considers this snow to be new snow, even though it has been there for many months.
Is that possible? Is there a way to fix or deal with this?

 

[haiqingsong]

I am running 24hr forecasts initialized from GFS with NoahMP.
Over areas of snow there is a very strong cold bias. I suspect this is because when the model is started it considers this snow to be new snow, even though it has been there for many months.
Is that possible? Is there a way to fix or deal with this?

Which version of WRF do you use？

 

[wallis]

4.4.2. But it is not related to the version.
How is snow age initialized and how should it be initialized?

 

[Ming Chen]

Would you please send me your namelist.input to take a look?
By the way, what variables did you check and find the large cold bias?
Also, please upload your wrfinput file . If the file size is too large, probably you can extract snow-related fields.

 

[Ming Chen]

Can you post a figure that shows snowc, snowh and related cold bias? Thanks.

 

[Ming Chen]

Further looking at this issue, I found that the possible reason for cold bias over snow area would be the overestimated snow cover and albedo in Noah-MP.

Here is how Noah-MP initialize snow properties: noahmp/module_sf_noahmpdrv.F at release-v4.5-WRF · NCAR/noahmp
Basically, it makes sure that the snow temperature, snow depth, SWE and number of snow layers are consistent based on the initial snow depth, SWE, and ground temperature input.

We will investigate the snow cover and snow albedo issues in Noah-MP and hopefully we can fix the cold bias problem soon.

Thank you for reporting this problem.