Tornado Wrap-Up, Why Forecasts Went Wrong
A little bit of soul-searching today... A number of blog readers this morning are asking: What went wrong with the forecast? Well, it's time to cut the hype and enter the no-rotation zone, to borrow from O'reilly (and by "no rotation zone" I mean we didn't get much rotation and it wasn't in the right zone, bah-dum-shi! I'll be here all week).
Seriously though, we're really jumping the gun even looking at this issue - realize that this storm is not over. We still will see hail, wind and possibly even tornado reports today as the system moves eastward. But since the blog readers and Forum users are chomping at the bit, I'll take a look at these questions now and revise tomorrow if necessary (note that I am not discussing the morality of "hoping" for severe storms here - obviously in the end, the less severe storms, the better for the public!). Here is some video of the damage from the tornadoes, large hail and flooding that did occur (check out the tower down on the SUV about halfway through, and the huge waves on Lake Superior near the end!). Sam Champion from ABCNews reports:
http://wwwa.accuweather.com/video/video_inline_launchV3.asp?video=http://qt.accuweather.com/flashvideo/vblog_ferrell/1-VPID04112008gmatornado&category=&thumb=http://qt.accuweather.com/flashvideo/vblog_ferrell/1-VPID04112008gmatornado.jpg&hd=false
NOTE: VIDEO MAY HAVE ADS; VIDEOS NOT AVAILABLE ON WIRELESS DEVICES
Here are the stats so far as far as Spotter Reports from this outbreak:
And here are the complaints at hand:
The outbreak wasn't big as predicted. Actually, it might have been. Many meteorologists were throwing around the "hundreds" number as far as severe spotter reports, and if you look at this entire outbreak, which started April 7th, you will see that we have had 467 reports, and over 500 warnings were issued. This is not something to sniff at, in fact it's the second largest outbreak this year (unless we get another 119 reports, then it will beat Super Tuesday, though those reports occurred over only a two day period).
There weren't as many tornadoes as predicted. This is true. The number that I heard being bandied about earlier this week, here at AccuWeather and elsewhere, was "100." Up until yesterday, we had a total of 29 tornado reports from this storm. That's a far cry from Super Tuesday, which had 134, and from the 1974 outbreak, which was another storm meteorologists used for comparison (there were 148 tornadoes then; if that happened this week, the number of storm reports would have likely been much higher, perhaps twice that, because of duplicates).
Thursday's tornadoes were predicted in the wrong area This is trickier to measure. Below is a map showing The SPC's forecasts, the storm reports, and a comparison to Super Tuesday.
Technically a Moderate Risk area should have "6-19 tornadoes" so yes, they missed that, because the Moderate Risk area in fact had only a couple. Note that had they kept their second forecast (from the day before) instead of shaving off most of Missouri, they would have caught more tornadoes. The good thing, however, is that all tornadoes were contained within the Slight Risk, so people in that area should have been aware of the risk of tornadoes if the media did their job (although some will argue that Slight Risks occur so often, it's like crying wolf). The bad news (forecast-wise) is that Arkansas, which was always "ground zero" for the Moderate Risk, reported no tornadoes.
It's easy of course, to pick on the SPC forecasts since I don't work there, but they are the "official" forecast that the media looks to when warning the public. I'll let other forecasters here defend their own forecasts on their own blogs; I spoke with Henry briefly this morning and he thought the jet stream was to blame - it came in too fast and caused tornadoes in different areas than originally predicted, plus unexpected cloud cover held down instability. Here's a particularly troubling map issued on our Breaking Weather News Page (PREMIUM | PRO Thursday morning:
The text was better: "dangerous storms can be expected from the Gulf Coast to the western Great Lakes." Clearly, we would have been better off outlining the SPC areas for the map, because we missed the Iowa tornadoes entirely (or they were hidden in a "Slight" risk that we didn't show on the map). But hey, forecasting isn't an exact science and I'm sure our meteorologists did the best they could with the information available.
My personal opinion in regards to the over-hyping of this outbreak (which I participated in) is that the meteorologists (everywhere) were paying too much attention to matching the overall weather pattern to 1974, and not spending enough time looking at the many wonderful Severe Weather Indices [JessePedia] from the Forecast Models [JessePedia]. As I noted in my blog on Tuesday and Thursday, these never looked very impressive, or when they did, they were in different geographical locations. Why didn't I promote that fact more? Honestly, since I'm so out-of-the-loop these days in regards to what our forecasters are thinking, I assumed there was something obvious I was missing, and I didn't want to be the guy who said "this outbreak is going to be lame" and we get hundreds of tornado reports.
P.S.: I'd love to see a comparison or some thoughts on the 4-KM WRF Model Simulated Radar vs. the actual radar for this storm. I've uploaded some movies here* if anyone has time to go through them. Note in the actual movie, it includes the early-week as well as the late-week storm, but the WRF loops are only for the late-week storm.
*Quicktime required for Actual Radar Loop. "Save Target As" required for IE.
