[Wrf-users] Problem with WRF-ARW BMJ scheme after V2.0.3.1?

[Jim Dudhia]

Yes, it is a known change in BMJ, that they reduced its activity in  
2.1. For 2.2 they are restoring a lot of its
old activity. I found out how to modify the 2.1 code to do this in a  
quick way, but you need the 2.2 version
for the official NCEP change,
Jimy

On Jul 27, 2006, at 1:42 PM, Gary Lackmann wrote:

> Hello,
>
> With WRF-ARW V2.0.3.1 we had noted consistent behavior of the BMJ  
> CP scheme relative to other models (e.g. MM5, workstation Eta) when  
> we compared between cases and simulations.  It is a fairly active  
> scheme.  When we began running WRF V 2.1 and beyond, we noted a  
> very much reduced level of BMJ activity.  This was gleaned not just  
> from the "convective" precipitation totals, but also from the  
> heating and moisture tendencies. There is about an order of  
> magnitude less parameterized precipitation in the newer versions,  
> using BMJ.
>
>
> In contrast, with the KF scheme, the totals are consistent with  
> what we saw in the earlier versions of the model, and the totals  
> are much greater than those for the BMJ for the cases we've studies.
>
>
> Recently, working with graduate student Megan Gentry, we ran the  
> Hurricane Ivan case using several different permutations of WRF.  I  
> am attaching a plot of the convective precipitation (CTOT) and SLP  
> for Ivan using V2.0.3.1 as compared to V2.1.2 (which is very  
> similar to all other 2.1 or later versions we've tried).  Clearly,  
> there is a huge difference.  This issue is not only seen with this  
> case- we've run convective cases, other hurricanes, mid-latitude  
> cyclones, etc. and the story is the same.
>
>
> For most of the WRF model runs that we've been doing recently, we  
> have used the KF CP scheme due to this perceived issue.  But we  
> wonder if others have noticed this, are aware of it, have fixed it,  
> etc. Perhaps this issue is well-known and will be addressed in the  
> next release?  If so, we would like to know before we invest any  
> effort in trying to fix it ourselves.  We've investigated so far as  
> to determine that the issues probably lies outside of changes to  
> module_cu_bmj.F.
> Thanks for any info you would like to share about this.
>
> Best Wishes,
>
> Gary Lackmann
> <ctot_comp24.gif>
> _______________________________________________
> Wrf-users mailing list
> Wrf-users at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/wrf-users