Comparison ARW and NMM

Forum dedicated to older versions of EMS package (WRFEMS v3.2, v3.1 or older). Support is user-to-user based, so please help others if you can.
Post Reply
meteosciez
Posts: 54
Joined: Sat Sep 05, 2009 6:36 am

Comparison ARW and NMM

Post by meteosciez » Mon Sep 28, 2009 12:23 pm

Hi,
I was running until now a 3km domain with WRF-ARW, and I have now defined also a 2.1 km WRF-NMM domain.
While compairing the output of ARW and NMM, I am surprised to observe significant difference with the 2m temperature data, particularly during (or at the end of) the night : NMM temperature seems to be overestimated and are significantly higher than ARW valure ( 3 to 5 degrees differences!) - see http://www.meteo-sciez.com/wxwrf2.php.

I had a look at the forecast data of meteoadriatic, and it seems also that during or at the end of the night, 2m temperature seems to be higher with NMM.

Do you know why ?

Regards
meteo-sciez.fr

meteoadriatic
Posts: 1574
Joined: Wed Aug 19, 2009 10:05 am

Re: comparison ARW and NMM

Post by meteoadriatic » Mon Sep 28, 2009 12:44 pm

Yes, I know about this "issue" with differences in night surface cooling between NMM and ARW. I think this is due to differences in used longwave radiation schemes between them, or something else in model phyisics (PBL scheme maybe). To be sure, you should need to compare ground skin temperatures, so I think if they are lower in ARW, then it is due to longwave radiation scheme, and if they are similar then cause of this difference should be somewhere else. I didn't bother too much with this.

But, I found that observed late-night temperatures in most cases fall somewhere in beetween of NMM and ARW. So, and for other reasons, I combined both outputs in so called WRF-C ("C" as combined) and it seems to do its forecast job better than NMM or ARW alone.

meteosciez
Posts: 54
Joined: Sat Sep 05, 2009 6:36 am

Re: comparison ARW and NMM

Post by meteosciez » Mon Sep 28, 2009 1:10 pm

I have tried a new NMM run with a different PBL scheme : GFS instead of MYJ, and night temperature values are now significantly lower and follow better the topography.
They are now closer to the real temperatures and to the temperatures predicted by WRF ARW.

Image
Image
meteo-sciez.fr

meteoadriatic
Posts: 1574
Joined: Wed Aug 19, 2009 10:05 am

Re: comparison ARW and NMM

Post by meteoadriatic » Mon Sep 28, 2009 2:36 pm

Good to hear that, but be careful, lot of other things also depends on PBL scheme, not only temperatures. So, check carefully that your model still has good overall predicting skill on other variables after that change.

wettersat
Posts: 32
Joined: Fri Aug 21, 2009 9:30 pm

Re: comparison ARW and NMM

Post by wettersat » Mon Sep 28, 2009 2:44 pm

In general what are the differences between ARW and NMM?

Gerald

meteoadriatic
Posts: 1574
Joined: Wed Aug 19, 2009 10:05 am

Re: comparison ARW and NMM

Post by meteoadriatic » Mon Sep 28, 2009 3:09 pm

There are quite heavy differences between these cores. NMM and ARW are dynamic solvers of differential equations, that do their job in much different way. But from user's view of point, most important differences can be summarized like this:

- NMM is created from ETA model in NCEP, ARW from MM5 model in NCAR, both are nonhydrostatic models
- NMM is much faster than ARW
- NMM is primarly used for forecasts, ARW is assumed to be tool for both forecasts and atmospheric research
- ARW has much more complex dynamic and physics settings than very limited choices in NMM
- NMM is not the best choice for multi-level nesting due to some important limitations
- there are endless debates about their predicting skills, NCEP people claim NMM is better, and NCAR stuff of course say the same for ARW.... :)

meteosciez
Posts: 54
Joined: Sat Sep 05, 2009 6:36 am

Re: comparison ARW and NMM

Post by meteosciez » Mon Sep 28, 2009 7:34 pm

meteoadriatic wrote:Good to hear that, but be careful, lot of other things also depends on PBL scheme, not only temperatures. So, check carefully that your model still has good overall predicting skill on other variables after that change.
Yes. I will anyway run both cores for a while to learn their differences.
meteo-sciez.fr

pattim
Posts: 173
Joined: Sun Jun 24, 2012 8:42 pm
Location: Los Angeles, CA, USA

Re: comparison ARW and NMM

Post by pattim » Mon Jul 02, 2012 12:07 am

Wow! Where did these nice plots come from? (what software - part of EMS?)
meteosciez wrote:I have tried a new NMM run with a different PBL scheme : GFS instead of MYJ, and night temperature values are now significantly lower and follow better the topography.
They are now closer to the real temperatures and to the temperatures predicted by WRF ARW.

Image
Image

smartie
Posts: 97
Joined: Sat May 21, 2011 7:34 am

Re: comparison ARW and NMM

Post by smartie » Mon Jul 02, 2012 8:55 am

[quote="pattim"]Wow! Where did these nice plots come from? (what software - part of EMS?)

Patricia,
this is the RIP package- which you now have.
;-)
David

pattim
Posts: 173
Joined: Sun Jun 24, 2012 8:42 pm
Location: Los Angeles, CA, USA

Re: comparison ARW and NMM

Post by pattim » Mon Jul 02, 2012 2:41 pm

smartie wrote:
pattim wrote:Wow! Where did these nice plots come from? (what software - part of EMS?)

Patricia,
this is the RIP package- which you now have.
;-)
David
Thanks David! I can't wait to try it out. <<goosebumps!>>

Post Reply