Brett Roberts
EF5
I'd be interested to see some type of objective verification, as well as subjective impressions from those who have followed it very closely (likely SPC forecasters). I've followed it on-and-off, and so far, I'm skeptical of its value added. My impression, like yours, has been that it only becomes at all useful inside the 10-14 day range. Yet, by that time, it's often just as easy to look for trends and patterns on the GFS or ECMWF and their ensembles. For example, it appears the CFS began showing the April 17 threat consistently about 10 days out; meanwhile, the operational GFS indicated an impressive western trough with good moisture return from at least 12-14 days in advance. If we start seeing even mildly reliable signals for significant events with 15-30 day lead time, beyond the range of the current global forecast models, then I'll really start paying attention.I continue to keep an eye on CFS forecasts for a hint at what may happen past two weeks. Greg Carbin at SPC has put together a great page for viewing severe storm centered forecast output: http://wxvu.net/spc/cfs_scp/
The model has shown some hint of reliability in terms of getting the large-scale patterns for outbreaks when they're within 10 days or so. Although the forecasts are usually just noise past that, if the latest 10 or so days worth of forecasts verifies, it's going to be a very disappointing May for most. With that said, the CFS probably will not show any signal for local events (like OK last Friday). But that is probably what much of the peak of the season is going to be like - needle-in-a-haystack type days. I only hope that the late arrival of springlike temps and the late departure of snowpack in the north will help keep the jet stream from jumping way north into the far northern US or Canada until later in the season. We shall see, however.
The fact that it looks so dead for May on so many consecutive runs is disconcerting, but being that this is its first year and it's such an abstract statistical tool, I'm not sure we're quite "calibrated" to its output yet. Is it possible it would have looked just as bleak on this date in 2010 or 2011? Who knows. Probably not, but it's easier to stay sane by using that reasoning while we still can!