AccuWeather rips Weather Service for dismissing tornado threat in Oklahoma Wednesday

Randy Jennings

May 18, 2013
Article posted to Washington Post's Capital Weather Gang Blog today at lunch time:

It refers to this marketing material AccuWeather Enterprise Solutions released critical of NWS and promoting their services:,_Oklahoma_Tornado.pdf
My 2 cents:
I agree with David Titley in the article, “on any given day, there’s always one forecast that is best”. I don’t think the criticism of NWS Norman is fair, nor do I think AccuWeather should pick one case and try to use it for marking purposes. While I’m sure AccuWeather has a talented team of meteorologists, I would never depend on their service. I have worked for a client of theirs. I remember a case a few years back where I was in our company EOC and there was a tornado on the ground a couple of miles from our location and we where in the path and in the NWS warning polygon. I first heard about the warning over ham radio on the Skywarn net (the Skywarn volunteer at the NWS WFO told us as it was being typed up). Seconds later in came over NOAA all hazards radio. A minute or so later local TV was covering it, and a few minutes later phones started ringing with the local government’s CodeRed call. Just before the warning was cancelled, we finally got an email from AccuWeather saying that all was safe – we could ignore the NWS warning, as their meteorologists had determined it not to be a threat to us. In a time of emergency, I don’t think email is a reliable or timely form of communication. I also don’t think one should second guess a NWS warning, especially when you have several thousand employees to move to safety.
  • Like
Reactions: cbhopper

Jeff Duda

Resident meteorological expert
Staff member
Oct 7, 2008
Broomfield, CO
Haha, nice plug for Radarscope, but I use PYKL3 (sorry!).

I agree with the sentiment in your blog, Steve. The main article in the OP also said it well. This was a case of cherry picking, and I suspect there are issued with false alarms that were conveniently omitted from the AccuWeather write-up. Unfortunately, that write up will probably sway some potential customers into buying into AccuWeather, but in my opinion what was said in that article makes them look misleading and dishonest. It's also an insult to the public sector and even academic sectors for them to say what they did.
Apr 17, 2006
Owasso Oklahoma
Visit site

NWS WFO Tulsa was real close to nailing the Sand Springs tornado even with their morning threat map. Their threat percentage went up later in the day I do not have that map. Their warning was issued quickly as the tornado was out by Lake Keystone.
Aug 16, 2009
Amarillo, TX
We will be discussing this on The Ghost Train podcast this Sunday. We'll have a live Q&A during the show so viewers can interact with us. This topic will most definitely come up. I urge you to watch this Sunday @ 7:30 Central.
Apr 5, 2010
Omaha, Nebraska
I've always been more of a Wunderground than Accuweather guy, but they do make a good point. In my opinion, some offices seem more likely to drag their feet on putting out TOR warnings than others. They are apt to over-warn on high threat days, and lag behind on low-threat days. It feels like once per year I'm out chasing and get a funnel or touchdown a couple minutes before the office warns here in Nebraska, typically on a low-risk day. If the point of a warning system is to give 15 minutes, then why wait until you have visual confirmation? By then it's too late.

I understand the risk of over-warning (boy who cried wolf), but if you've got classic velocity signatures on the radar (which is the issue in the specific Moore, OK case), regardless of other circumstances, why not warn the cell? That's what should be the discussion here. The signature was obvious, why didn't they call it and error on the side of safety?


Mar 1, 2004
Lansing, MI
Rocye - are you serious? You are talking about the Moore OK cell? It was not a classic signature by any means. If that cell got a tornado warning with 15 minutes lead time, we'd be issuing a warning on every "mean-looking" cell.
Apr 14, 2011
Alexandria, LA
On Marcus' show last night, Skip Talbot did a great job dissecting the AccuWeather release point-by-point. The short version is, if AccuWeather tornado-warned the OKC/Moore area for the specific reasons they allude to in the PDF, then it's purely coincidence that the 3/25/15 tornado happened to touch down in their warned area; because the images used in the release show that their "superior forecasting model" allegedly predicted a mesocyclonic supercell tornado, which the Moore tornado was decidedly not. The PDF is written with the expectation that the person reading it will be unaware of the difference.