Critical regression in NOAH-MP snow physics between v4.2 and v4.7 - persistent cold pools

[Ming Chen]

You are right that NCAR maintains a stand-alone NoahMP package. As far as I know, the stand-alone NoahMP performs well when it is forced by real-time observations. However, I am not quite sure whether this package includes the latest updated snow module. I will talk to our expert and he may have more in-depth details regarding this.

 

[tslin2]

OK thank you all, after detailed comparative testing it seems that introducing canopy heat storage effects in v4.4 did the regression in simulation accuracy.

Indeed, going very low with CBIOM parameter effectively reverses these changes. Going 10x lower as suggested in few places, that is from 0.02 to 0.002 is not enough to fully reverse the effect, but 0.001 almost is. I don't know what would the physically correct thing to put there be, but 0.001 seem to work reasonably.

Here is direct comparison, T2, range 273-293 K fixed on all plots.
Left: v4.3
Middle: v4.7.1 with CBIOM = 0.02 (currently default value)
Right: v4.7.1 with CBIOM = 0.001

Thank you all who participated. Should we create a pull request to set a better default value for next model update?

Ivan

View attachment 19345

Dear Ivan,

What is your study region in this figure and vegetation type for cold bias area?

Thanks,
Tzu-Shun