Welcome to the new NECN.com! Our Weather New England page lives on, and we hope you enjoy the weather perspective our team offers. Of course, for the latest forecast for your town, be sure to visit the brand new NECN Weather Page at necn.com/weather!
Thursday night's 9 PM NECN newscast aired my report on Monday's Revere tornado - a "rain-wrapped tornado" - a type of tornado with rain on all sides, that can sometimes deliver notoriously little or no warning. In this post, you'll find out exactly what went on "behind-the-scenes," both behind nature's curtain and in the Weather Center at NECN.
The storms began from a piece of atmospheric energy pinwheeling around an upper level storm moving east out of the Great Lakes - as that energy collided with humid, warm air, thunderstorms developed, intensifying in Southern New England Monday morning.
Just after 7 AM, a strong thunderstorm prompted a Severe Thunderstorm Warning in Western Massachusetts, as storms strengthened in an unstable atmosphere with wind shear that favored rotating updrafts - a key factor in not just creating tornadoes, per se, but certainly in creating self-sustaining, long-lived thunderstorms:
By 8:20 AM – some one hour and twelve minutes before the tornado - weak rotation, not strong enough to produce a tornado, was evident within the storms on radar imagery as they approached Boston’s MetroWest region - in fact, in the radar image below, the area showing greatest rotation was on the southern tail of the storm approaching Milford, MA, but storm motion was fast, and rotation was elongated, not conducive of tornado development. I commented to my intern, Lyndon State College meteorology student Aaron Perry, that the storms should have a better chance of capitalizing on the rotational component once their forward speed slowed, likely to happen north of the Massachusetts Turnpike, where the flow aloft was not as fast. In fact, the leading storm nearing Milford would slow so substantially, that the next storm moving northward, barely visible at the bottom of the image, would become the dominant storm, and that is the cell that would produce the tornado in Revere:
For an hour and eighteen minutes, from 8:12 AM to 9:30 AM, in every weather broadcast, I highlighted the storm, pointed out the rotation, and reiterated my concern that the storm was rotating. Though I explained there wasn't strong enough rotation indicated to believe a tornado was imminent, this was a storm to watch carefully. Accordingly, with broad rotation and a strong wind field indicated by Doppler radar, looking at air moving inside the storm, the National Weather Service issued a Severe Thunderstorm Warning during this period, set to expire at 9:30 AM.
At 9:06 AM, during the Severe Thunderstorm Warning, here's the radar analysis - the upper left is "base reflectivity" (what you're used to seeing on TV - a reflection of drop size in the storm), upper right is storm relative velocity (a depiction of which way air is moving, where green and red indicate air moving in opposing directions, ie: rotation), lower left is "correlation coefficient" (assists in determining what is suspended in the sky, and how uniform the sky is - great for quickly deciphering if non-meteorological things are floating...like debris from a tornado), and lower right is a parameter to assist in seeing rotational component, as determined by the computer. Note rotation in the upper right and lower right images...of course, I continued to mention this on-air:
Just minutes later, at 9:13 AM, the southern end of the storm also began to take on some concerning characteristics, though the rotational component was not as strong there on the velocity plot, the storm surely was wrapping up on its southern end, near Westwood, as well:
At 9:30 AM – about an hour and 20 minutes after first tracking the weakly rotating storm - Yours Truly, and undoubtedly the National Weather Service, as well, were in an interesting position. The NWS had carried a Severe Thunderstorm Warning, and I'd been giving live, continuous coverage updates on NECN that the storm was rotating and could still produce a funnel or tornado, and advised our newsroom they should standby for a Tornado Warning for the Boston area because it may be forthcoming...but one critical factor was missing - damage reports. Everyone was seeing heavy rain, many were seeing flash flooding, roads were closing, cars were stuck...but not a single wind damage report had been received and not even a strong wind gust had been reported. Radar showed the rotation weakening in subsequent scans from 9:15 to just prior to 9:30, and Boston's Logan Airport reported a heavy thunderstorm...but a wind from the northeast at 13 mph. So, for the National Weather Service is was a question of how to continue a Severe Thunderstorm Warning for the most populated area in the state when reports showed no strong wind, and of that densely packed populous, not one report of strong wind or damage had been received. We all know that False Alarms degrade public trust and lead to increased complacency when it comes to taking warnings seriously. For me, it went a step farther - I'd just put the most populous area of New England on guard for a rotating storm, and couldn't even find a wind gust over 20 mph. So, at 9:30 the NWS dropped the Severe Thunderstorm Warning, and for the first time in an hour and 18 minutes, I dropped the mention of rotation in the storm and explained on-air what I just explained above - no verification of any damage or strong wind on the ground.
Two minutes later, hidden in a blanket of torrential rain, the funnel descended and four minutes of terror in winds up to 120 mph ensued.
Within minutes, a debris signature was detected on radar. I didn't grab a still image of that radar image, as it was time to prep and get on-air, but our weather data provider, WSI, did send out an image of the correlation coefficient product I'd mentioned above - note the circled area, indicating where radar picked up on something non-uniform in the sky, nearly always debris.
I pointed to the debris signature with Aaron and fellow meteorologist Joe Joyce, then rushed to the weather wall to give the update. But there was one problem: Debris in the air would mean a tornado had been on the ground for at least a few minutes prior to that debris showing up on radar. Everything lined up on radar, but how would it be possible for the following to transpire:
- A tornado developed, touched down and caused damage minutes earlier...which would be somewhere a few minutes after 9:30, but...
- No reports of damage were coming in to NECN...
- No reports of damage were coming into the National Weather Service, else they'd have issued a Tornado Warning...
- No reports of a funnel were received to either outlet
- Presumably a few to several minutes had passed since the apparent tornado touched down in a densely populated City just five miles north of Downtown Boston, but 6 minutes later, nobody saw it, and no damage was reported.
What were the chances, I wondered, that a tornado could have touched down in such a populated area, when everyone has a cell phone and social media runs rampant to the point that a door not closing properly on a public transit train downtown makes the news, but not one report had come in? It was a moment of true conflict between technological signs that were irrefutable, but logic that didn't compute. Within moments, the Tornado Warning was issued - damage reports were, indeed, starting - the tornado had already touched down. Coverage continued live, on-air, with street-by-street forecasts of the rotation...but it was too late. The tornado had already come…and gone...eight minutes before the warning was issued, and two minutes after I'd stopped cautioning a tornado may develop. Cars had been overturned, and citizens in impacted parts of town were contending with fallen live power lines amidst a gauntlet of downed trees, overturned cars and flooded roadways.
In a humid, tropical airmass, clouds hang low to the ground – what may otherwise have been a funnel observed developing in a cloud overhead, instead becomes a funnel developing on top of you, embedded entirely in the cloud, hidden by rain, touching down almost immediately. In New England, it is the setup that affords the least advance warning for developing tornadoes, and a setup that would lead to a quick touchdown that surprised residents, meaning no observation of a funnel could be made, let alone a report of one. When the storm came and went in minutes, residents would be left wondering what hit them, and the localized nature of damage to the 3/8 mile wide path would further delay realization of exactly what hit, which explains the slight delay in damage reports - naturally, it takes a few moments to realize what hit you when you couldn't see it coming.
Thankfully, there were no serious injuries, but from a forecasting perspective, this storm reminds meteorologists and viewers, alike, of inherent uncertainty in nature, and of the importance that storm information be expressed with a range of possible outcomes made clear, and those possible outcomes taken seriously. Just like in Westfield and Springfield, MA, back in 2011, for an hour and 20 minutes leading up to touchdown, I warned that rotation could produce a tornado. The difference: in 2011, we did not waver with that warning, and then received a report of a touchdown. In 2014, with two minutes left to the touchdown (literally), the storm was fumbled. Lesson learned - respect the unpredictability of nature at all times, not just 118 of 120 minutes. Thankfully, no life was lost in the teaching.