Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 10:30 AM ET, 05/30/2008

Severe Weather Risk More than Slight

By Dan Stillman

The D.C. area is under a "slight risk" for severe thunderstorms tomorrow. That's not as insignificant as it may sound. Graphic courtesy the Storm Prediction Center.

The D.C. area is under a "slight risk" for severe thunderstorms tomorrow, and I don't like it one bit.

It's not that I'm averse to heavy downpours, loud claps of thunder, gusty winds, large hail and the potential for tornadoes. In fact, as a meteorologist, I think all these are pretty exciting. My beef is with the the word "slight," and the misleading message it may send to the public at large.

Keep reading for more on my problem with "slight," and see our full forecast for the latest on tomorrow's potential for severe storms and what kind of weather will follow into next week.

The "slight risk" designation comes from the National Weather Service's Storm Prediction Center (SPC) , which routinely issues forecasts for severe weather across the country, and indicates a 15-30% chance that a severe thunderstorm will happen within roughly 25 miles of any given point inside the area circled in the graphic above, according to SPC. A severe thunderstorm is defined as "a thunderstorm producing hail that is at least penny size, 0.75 inches in diameter or larger, and/or wind gusts to 58 mph or greater, and/or a tornado."

Technically speaking, I'd say "slight" is a fair translation of a 15-30 percent chance. The truth of the matter, however, is that the potential for severe weather tomorrow in the D.C. area and throughout the mid-Atlantic is real and significant, and often is on days the region is classified under the "slight risk" category. Yet, my hunch is that many people interpret "slight" along the lines of its dictionary definition -- "of little importance ... trivial" -- which is a dangerous assumption considering the damage to life and property a severe thunderstorm or tornado can cause.

In gambling terms, if Big Brown was given only a slight chance of winning the Triple Crown in next Saturday's Belmont Stakes, I'd probably bet against him. (For the record, oddsmakers are giving the horse a much better chance than that.)

Furthermore, while I don't have any hard stats on this, my recollection is that most of the severe thunderstorms and tornadoes that occur in the mid-Atlantic occur on "slight risk" days. This is mainly because, unlike the central United States and Southeast, the mid-Atlantic rarely sees the atmospheric conditions needed to be categorized under a "moderate" or "high" risk.

A skilled weather forecaster and communicator, if he or she even reports the risk assigned by SPC, will put that information into context by providing a more in-depth explanation of the potential for severe storms (see our latest forecast, for example). Problem is, people don't always get their weather information straight from the source. Often, it's filtered by a radio DJ or broadcast in the form of an Internet headline, both of which may mention a "slight risk" with little or no elaboration.

What "slight risk" fails to communicate, at least in my mind, is that the risk of severe storms is quite real compared to a day on which there's a chance of merely garden-variety thunderstorms. Tomorrow, for example, several ingredients that would support the formation of severe storms -- including an unstable atmosphere, changing wind speed and direction with height, and the jet stream swinging over our region -- are expected to be in place.

Even so, it's entirely possible that few or no severe storms will materialize. But that doesn't change the fact that the atmosphere will be much more ripe for the development of severe weather than it normally is -- a situation not well-portrayed by the term "slight risk," in my opinion.

Considering I started this complaining, I should probably step up to the plate and suggest an alternative label to "slight." Instead, I'll let you take the first crack using the comments link below.

By Dan Stillman  | May 30, 2008; 10:30 AM ET
Categories:  Thunderstorms  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Stormy Saturday? Otherwise Nice
Next: CommuteCast: Setting the Stage for Saturday


Speaking of severe weather, anybody have any insight into this event in central New York two Saturday's ago? A tornado formed well separated from the actual thunderstorm... pretty interesting.

Posted by: CM | May 30, 2008 11:00 AM | Report abuse

What we have here is a simple case of disagreement between you and the SPC regarding the threat tomorrow. You feel the threat to be "significant", the SPC does not.

The average person does not know or care about the SPC (probably more than 50% have never heard of the SPC). When a local forecaster believes the threat to be significant, he or she can easily communicate that mindset to the public by simply including a "headsup" in their forecast: "a good chance of thunderstorms tomorrow afternoon, some of these storms may be severe". That tells the public much more than any designation by the SPC, and in reality is in general agreement with the SPC "slight" designation.

Posted by: Augusta Jim | May 30, 2008 11:09 AM | Report abuse

Dan, good post. I wonder if the fact that the internet has made all these products readily available might be part of the "translation problem." I might be wrong, but I would guess the 1 day, 2 day, and beyond outlooks put out by SPC are for meteorologist consumption more than the general public.

I agree that the use of "slight" might be confusing to some degree, but for those who follow forecasts from the SPC the risk gradient makes some sense. Generally, when a moderate risk goes up, we know there is a pretty good likelihood of a larger outbreak. And when a high risk goes up (usually only 3-5 times a year and often in the plains like yesterday), we know a very bad outbreak, often with damaging EF3-EF5 tornadoes or a major derecho, is on the way.

Sure, "slight" does not completely paint the picture, but since they forecast for the whole country and often outline large regions I don't really disagree with their terminology. Take the slight risk graphic in this post -- it's quite likely that not everyone in that green circle will actually see severe weather (as defined). Even here, where the slight risk for tomorrow is higher in probability than other places (30%), many locations will not reach the "official" criteria.

Maybe there should be a better way to designate levels of concern, but I think this is more of a guidance package for local NWSFOs in the leadup to an event, and the SPC later hones in on the greatest areas of near-term concern with a watch.

Posted by: Ian, Capital Weather Gang | May 30, 2008 11:44 AM | Report abuse

It really doesn't pass the logic test when you say:

"Technically speaking, I'd say "slight" is a fair translation of a 15-30 percent chance. The truth of the matter, however, is that the potential for severe weather tomorrow in the D.C. area and throughout the mid-Atlantic is real and significant, and often is on days the region is classified under the "slight risk" category."

...what, then, is "real and significant"? Is that more than 15-30%? And even if a storms occurs inside the area shown on the map -- which stretches from Oklahoma through Arkansas (tornado alley, anyone?) all the way to the Eastern Seaboard -- what percentage of tornadoes that happen in that area are in or near the greater DC metro area?

I think the problem isn't in the use of the word slight, but with the way the word "severe" is buried within the forecast. Hey, 58 MPH winds are dangerous and worth taking special care to avoid, that's no joke. But a thorough reading of the forecast tells me that the morning will likely be nice and that weather doesn't become a huge factor until the afternoon or later. The upshot: Don't cancel plans for the AM, but when you see big clouds heading our way know that they might be packin' a wallop!

Posted by: i hate walks | May 30, 2008 11:45 AM | Report abuse

Augusta Jim -- I think Ian hits the nail on the head when he mentions the Internet making products that are meant primarily for meteorologists more available to the mainstream ... I briefly alluded to this with my "Internet headline" example in the post.

Thanks to the Internet, the forecast of "slight risk" -- which may be accurate from a technical standpoint but misleading from the perspective of effectively warning the public -- can get trickled down to the average person without further elaboration.

Posted by: Dan, Capital Weather Gang | May 30, 2008 11:57 AM | Report abuse

i hate walks -- It's admittedly a nuanced argument. The issue I have is that, while mathematically it's not off-base to call a 15-30% chance of severe storms "slight," I think it sends the wrong message to the public (and often it does reach the average person, even if it was intended for a more technical audience -- see previous comment). To me, a 15-30% chance of a severe thunderstorm or tornado presents a significant threat. Whereas I have no problem using the word slight to describe a 15-30% chance of a rain or snow shower, which poses little or no danger.

Posted by: Dan, Capital Weather Gang | May 30, 2008 12:27 PM | Report abuse

SPC added hatching to the outlook with the 1730z update. Interesting to note that DC and Baltimore are right on the fringe of this. Forecast will likely be refined as we get closer and closer to tomorrow.

Posted by: Kenny L. | May 30, 2008 1:43 PM | Report abuse

I believe the comments are missing an important point. Namely, though I haven't checked the exact numbers yet, climatologically for any given summer day in the region the probability for thunderstorms is ~ 15-30%. The "slight risk" is that some of these may be severe. If it were possible (as in some situations) to narrow down the region or specify where within it the odds were higher, that information would be provided by SPC.

I agree this is not the best choice of words This product is intended for meteorologists (as "guidance", not actual forecasts), but nevertheless media often cite or show this graphic, and that's all (e.g., the Today Show this morning). In these circumstances there should be at least a statement to pay close attention to your local sources of weather information (hopefully the Weather Gang being the first choice). The appropriate word as mentioned previously is "heads-Up" ==> stay tuned.

Posted by: Steve Tracton, Capital Weather Gang | May 30, 2008 1:51 PM | Report abuse

The La Plata tornado (4.28.02) happened on a moderate risk day, although the 1630z day one update took the DC/MD area out of the moderate risk (morning convection had left behind significant cloudiness). The moderate risk was re-established at the time of the 2000z update.

The College Park tornado (9.24.01) happened on a slight risk day, although the wording of SPC's convective discussions sharpened as the day went along, and the 2000z update mentioned the specific possibility of tornadoes in the DC region.

Regarding the use of the word "slight," I think varying population densities complicates the issue. In other words, if the SPC tags western Nebraska with a "slight risk" in a day two outlook (meaning a 15% to 30% chance of severe weather happening within a 25-mile radius of any given point), that is necessarily a "slighter" risk in the aggregate than if that given point is, say, the Ellipse, where the 25-mile surrounding zone has a lot more people/structures that could be affected by a severe weather event. Now perhaps SPC has some way of calibrating their outlooks to take these variations in density into account. Also, aren't the SPC probability values keyed to a specific area's severe weather climatology, with the percentages deviations from climatological norms? So a "10 percent risk" of tornadoes in that point in the Nebraska panhandle would represent a greater aggregate risk for tornadoes than would a "10 percent risk" centered on the Ellipse?

Posted by: catatonia | May 30, 2008 2:14 PM | Report abuse

The standard formula for estimating risk is:
(Probability of occurrences) x (Consequences of Occurrence) = Risk

The problem here is the confabulation of "risk" with "probability."

The 15-30% chance represents the SPC's best estimate of the PROBABILITY that severe weather will occur.

As Catatonia notes above, the CONSEQUENCES of a severe storm are much greater in a densely populated area.

If forecasters wish to be precise, they should distinguish probability of occurrence from consequences.

The effect of telling the public that there is a slight change of severe weather is lot different than the effect of telling the public that there is a 15-30% chance of a tornado touching down within 20 miles of anyone's house.

Posted by: Sasquatch | May 30, 2008 2:39 PM | Report abuse

Dan, are you saying that anytime there's a chance of bad weather (which would be about 250 times a year) that the forecasters should go in all-out panic mode? Isn't there already too much hype whenever a storm (or a simple shower) looms somewhere on the horizon?
There's something in here about the boy who cried wolf or the expression "you doth protest too much."

Posted by: GuySharp | May 30, 2008 2:42 PM | Report abuse

GuySharp -- I'm with you that there is often too much hype for weather that is not worthy of the hype. In no way am I saying that forecasters should go into "panic mode" anytime there's a chance of severe weather, nor do I think the post in any way whatsoever implies that.

Posted by: Dan, Capital Weather Gang | May 30, 2008 3:01 PM | Report abuse

The probability of a thunderstorm on a given day in DC is:
Jan 0.6%
Feb 0.7%
Mar 3.9%
Apr 9.0%
May 14.8%
Jun 19.7%
Jul 21.0%
Aug 16.8%
Sep 8.0%
Oct 3.5%
Nov 2.0%
Dec 0.0%

Posted by: Steve, Capital Weather Gang | May 30, 2008 3:02 PM | Report abuse

THANK YOU THANK YOU THANK YOU for finally addressing this issue that has bugged me for a while. They always say "slight risk" and everyone shrugs it off as nothing, and when their car is being blown away or pelted with golf-ball sized hail, they yell at the meteorologists and say "You said SLIGHT risk, tell that to my car!!!"

*end rant*

But yeah, I am really glad that this was addressed. I do think that the SPC saying "slight risk" does underplay the significance a bit. But GuySharp also has a point...if meteorologists say that the threat is "significant" too many times, they'll become numb to it, just like many people I know are numb to and/or don't care about the Emergency Alert System when it comes on. That's the one thing that sucks about weather. When you're predicting it, you're always walking that razor thin line between wolf-crier or down-player...both of which are bad, bad titles to have slapped on your forehead if you're a reputable meteorologist.

Posted by: weatherdudeVA (Lake Ridge) | May 30, 2008 3:09 PM | Report abuse

I gotta side with Dan on this one. We have a Heat Index that takes many factors into account for how the temperature "feels" on a given day. Similarly, a severe weather index of some sort could be useful here... taking into account the probability of the severe weather, population density (catatonia), consequences (sasquatch), and other factors. Nobody worries about how a UV index is calculated, they just need to know what it is that day, and they can then take the proper skin precautions.

While mathematically the same, a 15-30% chance of stubbing my toe should not have the same index rating as a 15-30% chance of being torn to pieces by a pack of rabid dogs. The chance is similar, but the potential harm if the event occurs is quite different.

If the point it to warn people, it should be taken into account that, unfortunately, using the word "slight" mitigates the concern the average person has towards the weather report. You can blame the receiver of the report if you wish for this, but people have busy and complicated lives. If the chance of severe weather is a 2 on a 10 point scale but the damage done would be a 7 or 8 if it were to occur, you cannot allow people to walk away from your report thinking the danger level for that day is "slight."

Posted by: StormCellerJane | May 30, 2008 4:58 PM | Report abuse

I disagree with the initial post. The SPC issues slight, moderate, or high risks based upon the probability that a certian severe weather event will occur within 25 miles of a given point. Proabailities are based upon the amount of severe weather reports that the forecaster would expect from a certian situation i.e A high risk would assume that many tornadoes would be reported with at least a few of them being strong or long lived...this "slight" risk that you are all complaining about is noting more than a portrayal of the uncertainty of the forecast, or the forecaster deciding that the severe weather will lack in coverage or intensity. I understand that there are other people using the outlooks such as emeregency managers etc. and that these people may not understand the probabilities behing the word slight, but it is just that, A SLIGHT risk. Severe weather events such as tomorrow would constitute a high end slight risk...It would be foolish for the SPC to have too many categories for their categorical outlook. People would be getting confused as to what high, higher, highest, low, medium...You get the point...

Posted by: Anonymous | May 30, 2008 11:01 PM | Report abuse

For me, the prognostic discussion is much more important than the charts or probabilities. A slight risk means many things, and in many cases, it simply indicates that the ingredients for a severe weather outbreak simply haven't come together to the point that a higher probability or moderate risk is warranted. Most slight risks aren't upgraded until the day or even the afternoon of the expected event because it takes that long for the models (and conditions) to meet the profile for a severe weather outbreak. Most of the time the probabilities don't come into focus until several hours before the severe weather outbreak begins. So I study the prognostic discussion and for me, it has more importance than the risk label. I realize the public as a whole is too ready to react to adjectives but we must all realize that simple adjectives do not begin to characterize the complex set of parameters that contribute to a severe weather outbreak.

As far as DC is concerned tomorrow, I wouldn't be surprised if the hatched area weren't drawn westward in later outlooks, and I would be equally surprised if DC didn't wind up in a moderate risk zone on Saturday. I'm still seeing derecho as the main weather event.

Posted by: Steve Wasko | May 30, 2008 11:54 PM | Report abuse

I think the slight, moderate and high work well for probability. I think we are needing a way to detail the possible impact from any severe weather. Maybe an impact severity scale similar to the EF scale.

Posted by: DW | May 31, 2008 10:42 AM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company