It happened again last night. What many would consider a rather ordinary summer storm triggered a severe thunderstorm warning from the National Weather Service (NWS) for much of the D.C. area, leaving some meteorologists scratching their heads.
By definition, for a thunderstorm to be severe it must contain damaging winds of 58 mph or greater and/or large hail (at least one inch in diameter).
The storms last night didn’t seem to meet that criteria. Dulles Airport gusted to 44 mph, and the Weatherbug network showed a number of gusts in the 30-40 mph range in the District.
There were no confirmed gusts close to warning levels of around 60 mph.
Gusty? Yes! Severe? No pic.twitter.com/nOin8uNNyj
— TerpWeather (@TerpWeather) August 14, 2013
A concern is that the NWS may be lulling the public towards warning fatigue by issuing warnings for these kinds of storms – which are common in the area during summer. If warnings are issued repeatedly for storms that are merely heavy, but not severe, it’s reasonable to worry the public may start ignoring these warnings. Then, they may not appreciate the relatively rare circumstance when storms are actually producing widespread hazardous conditions and take the needed action.
Data indicate we’ve seen a marked uptick in the number of severe thunderstorm warnings issued over the last 10 years.
From 2003-2007, the NWS issued 1,175 severe thunderstorm warnings for the region, compared to 1,639 warnings from 2008-2012, a 40 percent increase. Granted, 10 years is too short of a time period to definitively conclude the office has become more trigger happy in issuing warnings. It’s possible the weather over the last five years was more severe than the previous five and/or perhaps NWS forecasters have become more skilled at detecting severe storms. But, anecdotally, I can say the NWS seems to be issuing warnings more liberally and the limited data sample support that.
Other meteorologists in the region have also noticed the proliferation in warnings. After the NWS issued a warning for a questionably severe storm in late May, then-WJLA meteorologist Bob Ryan tweeted: “New warning a bit much When those storms came through [with heavy rain] but winds well below criteria gust maybe 30. Every T’storm now severe?”
But officials at the NWS defend their warning practices stressing that while large parts of warned areas may not experience severe storm impacts, thunderstorms often produce highly localized damage, meriting warning issuance.
“When we issue a severe thunderstorm warning, more than 90% of the time, there is significant wind damage in that warning area,” said Chris Strong, the warning coordination meteorologist at the NWS office in Sterling, in an email. “These trees that get knocked down a few at a time have killed numerous people over even just the past few years: in their cars, out in parks, etc. Those events need to be warned on to give people adequate time to get indoors.”
Strong emphasized most storms that come through D.C. are not like the June 2012 derecho leaving behind a wide path of destruction.
“The vast majority of our thunderstorm damage in this area comes from targeted wind gusts that affect communities, or even parts of communities,” Strong said. “They more often than not do not show up on mesonet networks, which are better at picking out larger scale wind events (macrobursts/squall lines).”
Returning to last night’s storms, although I was unable to find any evidence of winds higher than 44 mph, there were three reports of downed trees in the District suggesting the likelihood of locally higher winds perhaps near warning criteria. Also, a Capital Weather Watcher in Reston insisted his neighborhood experienced 60 mph gusts.
— Graham Marsden (@grahambot) August 13, 2013
Last night’s storm represented what’s become a common scenario: a large, densely populated area is warned and, for the overwhelming majority of those impacted by the storm, the storm is not severe. But, in very isolated pockets, severe conditions occur.
So is there anything for NWS to do about the fact many of its warnings will only produce dangerous conditions for a very small fraction of the people in a warned area?
Strong said the answer may be in making the warnings more targeted.
“[W]e will continue to shave down the size of warnings as the science advances,” Strong said. “Hopefully, someday in the future we’ll be able to say the people in Rockville near the intersection of 355 and Tuckerman Lane, there will be a 60 mph wind gust that will likely down several trees on Tuckerman.”
I would argue the NWS should take a hard look at tiering severe thunderstorm warnings. That is, it should issue one kind of less serious warning for a storm that has the potential to produce very localized damaging winds (e.g. pulse storms) and issue a more serious warning for a storm that is likely to produce more widespread damaging winds and/or has a history of producing extensive damaging winds (like a derecho or squall line).
To close, while I am concerned about the issue of the overwarning the public, I accept the NWS’ rationale for its warning policies. But I strongly recommend the NWS work feverishly towards advancing the science to allow more targeted warnings and/or implement a tiered warning system as I just laid out. A priority for NWS has to be effectively communicating risk and not overwarning the public.
(CWG’s Kathryn Prociv contributed to this post, providing the data on NWS warnings from 2003-2012)