FV3 CAPE performance example - Home - EMC Model Evaluation Group

Hi Jack,

Interesting example and thanks very much for sharing it.  What seemed to be the case is that, in general, the FV3's CAPE was mostly higher in the warm sector than the GFS.  It'll be interesting to see if that is a systematic difference or not.

Mark

Hi Jack,

Thanks very much for pointing out this example. I'll be covering the 1 May severe weather event during tomorrow's MEG meeting, though I won't be delving too much into the medium-range guidance. This is certainly an interesting case comparison between the GFS and FV3.

Although the CAPE values appear too high in the FV3 at F216, I'm quite impressed with how spot on the FV3 was with the location of the key surface features (00Z 2 May surface analysis shown for reference).

Note how the FV3 had a much better handle than the GFS with the lows in southeast Canada and southeast Colorado, the associated surface trough/frontal boundary, and the dryline at F216. The following three images are GFS-FV3 comparisons for SLP, 2-m temperature, and 2-m dewpoint valid at 00Z 2 May.

The CAPE values in the warm sector are definitely still something to keep an eye on, but the overall synoptic "success" at F216 is encouraging to see. A quick glance at F168 and F120 suggests the FV3 had a better handle on these features at those leads as well (though it was clearly too fast with the cold front at F168).

When we "last met," discussing the FV3's long-lead-time CAPE performance, I had included a couple of 384-hr forecasts from both the GFS and FV3, valid for 00Z May 25.  Both were showing rather monstrous-looking tropical systems in the GOM.  

Now, fast forward 192-hrs, and it's comforting to see that, especially the FV3, both models have backed off that forecast, and are generally much more benign.

That said, however, now both the GFS and FV3 show a potentially-impacting tropical latitudes system in their current 240-hr forecasts, valid 48 hours later at 00Z May 27th--Saturday night of Memorial Day weekend.  And to see where each model has their low center, I've switched off TT, to use the comparison image generated by EMC.

The FV3 has a 981mb system on the tip of the Yucatan, while the GFS has a 985mb system over FL's space coast. 

While, as usual, neither model is likely to be correct at such forecast projections, it is noteworthy that the ECMWF also depicts a 995mb system in the vicinity, just west of Tampa.

In summary, LOTS of bouncing around at long-lead forecast projections, especially with trying to resolve a potential system at tropical latitudes in May, but, of course, always good to keep an eye on.  Now on to make a new post of what I really wanted to share.... 

Jack,

(17 May 2018-550PM EDT) - Thanks for providing sample from these ultra-extended forecast periods (384 hr!). We noticed the 240hr 00Z/17 May suite of guidance had "lows" in various locations of the eastern Gulf of Mexico.

It'll be interesting how it turns out. And the 00Z/06Z 17 May FV3 run had some interesting solutions in the extended as well...

Steve Z-SOO/WFO Sterling, VA (LWX)