Communicating Joaquin

Let me vent.

The “cone of uncertainty” has become a “cone of discontent” for me with this storm. 24 hours ago it appeared that there were two possibilities for Joaquin’s track – a bifurcated set of solutions – some toward Bermuda and some toward the Mid Atlantic. As of yesterday both seemed equally probable. What no model really showed was a realistic threat to New York City and New England.

Even so, the National Hurricane Center forecast sort of “split the difference” between the two camps resulting in what I believe is an unrealistic forecast. Their deterministic forecast was based in part from a blend of 2 disparate solutions resulted in this forecast.


This forecast is probably coming from 2 schools of thought. 1) the split the difference forecast results in fewer wild swings advisory to advisory. 2) this forecast is a prudent one if your ultimate goal is to produce a forecast that has the lowest track error. Unfortunately, this is one of the least likely scenarios.

Today, the NHC forecast is equally confounding.


This would give one the indication that southern New England is in the crosshairs of Joaquin. Right? This also gives the end user the sense that a southern New England hit is the most likely scenario. It is our opinion that this is one of the LEAST likely scenarios to occur. How the hell do you communicate that effectively to the viewers? I don’t really know the answer – especially since some local media outlets who skew toward hype are talking about a major overnight trend that’s concerning and troubling for Connecticut.

Back to reality, here are the overnight Euro Ensembles. Where’s most of the clustering?

0_es3 (1)

The operational GFS is trending east in an extreme fashion as well (giant jumps east for each of the last 4 runs) toward the operational European that has insisted on an “out to sea” solution.

The problem is that the National Weather Service and the media has put all of our eggs in the “cone of uncertainty” basket. It works a lot of the times. When it doesn’t work it’s an absolute nightmare. How do you show a 5-day NHC forecast (which people expect to see – you can’t NOT show it) and say it’s really not the most likely scenario.

There aren’t easy answers. This forecast was unusually challenging. The NHC is about as good as it gets with tropical forecasting but communicating risk remains very hard for all of us. This storm is really frustrating.

Tracking Joaquin

There’s no question that Hurricane Joaquin poses a major threat to the United States. The storm continues to strengthen and many of our computer models are in agreement that the east coast is in trouble.


The GFS, GGEM, and UKMet all agree with the 12z tropical models that a Mid Atlantic hurricane strike is becoming increasingly likely by this weekend. However, uncertainty still abounds as the Euro scoots the storm out to sea unlike the model consensus.

The weather pattern over the eastern half of the continent is a classic one for east coast hurricane hits.


A giant/anomalous ridge stretches from Quebec through the Canadian Maritimes and into the North Atlantic. This effectively prevents to storm from curving out to sea. Additionally, models develop a cut-off low over the Appalachians which is typical in a pattern like this with the large ridge acting as an “omega block”. This should force Joaquin into the coast UNLESS it gets too far south.

If Joaquin is able to drop into the Bahamas and begin drifting east it could find a weakness under the ridge and begin scooting toward Bermuda. This is what the European model continues to show. There’s a weakness in the ridge over the central Atlantic – including the remnants of Tropical Storm Ida. There is a fair chance of this occurring.

Another possible (unlikely) outcome is that the ridge is able to shift east enough and the cut-off low forms later or a bit farther north resulting in a track that puts southern New England in the storm’s crosshairs. This is not the most likely scenario.

At this point the best guess is a low impact event here in Connecticut (minor flooding, minor wind issues, minor coastal flooding) with a major/high impact event possible in the Mid Atlantic. This could change, however, so it’s important to stay tuned to future forecast updates!!

A South Coast Special

A doctor at Brigham and Women's Hospital in Boston grabs a few chunks of 2.25

A doctor at Brigham and Women’s Hospital in Boston grabs a few chunks of 2.25″ diameter hail.

It’s hard to believe that after egg sized hail fell in Boston Tuesday afternoon it was not the most impressive severe weather event on Tuesday, August 4th. The morning damaging wind storm produced a swath of damage across Long Island and southern New England that is one of the most impressive I’ve seen in years.

Tuesday morning had all the hallmarks of a daybreak severe weather episode. Strong wind shear, steep lapse rates, and an impressive surge of low level moisture. Our computer models struggled with whether or not storms would develop but I decided to be fairly bullish on the severe weather potential anyway the night prior (as did some other local meteorologists). As it turns out – that was the right call. The morning sounding from Chatham (modified for the pre-storm environment at Warwick, RI) shows an uncapped, highly sheared, and highly unstable environment.


With approximately 3,000 joules of CAPE and more than 60 knots of effective bulk shear when storms were able to develop they became beasts. Storm mode was messy with the clusters resembling giant high reflectivity blob, but on doppler radar velocity data showed several pockets of intense damaging wind.

The first such pocket moved across the North Shore of Long Island producing widespread wind damage. When it left the Island a second area of strong winds clipped Groton and Stonington while the core of the wind moved ashore in South County, Rhode Island.

okx11 fatality occurred in Mystic from a falling tree from the northern storm while the southern storm produced an incredible 71 mph wind gust on Great Gull Island (near The Race), a 61 mph wind gust on the Stonington Borough breakwater and widespread wind damage in far southeastern Stonington (Pawcatuck and Greenhaven) as well as Westerly and Charlestown. In Charlestown, an 83 mph wind gust was measured and falling trees injured several campers at Burlingame State Park.

Trees down on Greenhaven Road in Stonington

Trees down on Greenhaven Road in Stonington

KGON 041018Z AUTO 28027G43KT 1 1/2SM R05/4000VP6000FT +TSRA BR BKN014 OVC033 23/21 A2984 RMK AO2 PK WND 24043/1015 LTG DSNT ALQDS RAB10 TSE01B07 P0012 T02280211

KWST 041034Z AUTO 22016G48KT 1/2SM TSRA FG FEW004 BKN012 OVC018 21/19 A2987 RMK AO2 PK WND 23048/1028 VIS 1/4V1 3/4 LTG DSNT ALQDS RAB21 TSB20 P0030 T02110189

Meanwhile, a second cluster of thunderstorms was bringing an exceptional wind storm to portions of the Providence metro area. In terms of damaging straight-line winds the radar signature over Warwick/Cranston was one of the most impressive you’ll see in New England.

box1The first sign of trouble was over western Rhode Island in the town of Coventry. In about 5 minutes the core of the storm exploded – with 65 dBz over at 30,000 feet by 6:12 a.m. Within the next 20 minutes what starts as a relatively benign looking velocity signature in the low levels of the atmosphere turns into a monster with a large core of destructive winds (at least 5 miles wide with >50 knots).


A closeup of the storm over Warwick and Cranston shows just how intense the storm was with a few pixels of radial velocity approaching 80 knots.


As luck would have it this passed very close to TF Green airport where a wind gust of 67 mph was measured as the storm moved through.

KPVD 041034Z AUTO 30027G52KT +TSRA 17/16 A2985 RMK AO2 PK WND 29058/1028 LTG DSNT ALQDS RAB14B29 TSE01B11 P0057 T01720161 RVRNO $

What’s interesting is that this wind occurred about 2 miles south of where radar was picking up the strongest velocity. In fact, the velocity sampled over KPVD was 56 knots (at 1300 ft) which is almost exactly the peak gust measured at the airport was. It stands to reason that areas on the Warwick/Cranston line saw wind gusts up to 90 mph as the storm roared in and the atmosphere was able to efficiently mix to the surface.

Courtesy: WX1BOX Warwick/Cranston RI - N1EGS-John Buco

Courtesy: WX1BOX Warwick/Cranston RI – N1EGS-John Buco

A few takeaways from this event that really stand out.

One, The the Storm Prediction Center frequently doesn’t outlook or issue watches for these “south coast specials”. They’re characterized bya relatively narrow zone of high theta-e air advects north from the ocean resulting in a small area of substantial CAPE near the coast. What’s frustrating is that even though there’s only a small geographic area at risk – the population for that area is HUGE – including Long Island, southern New England, and even the New York City area. A severe thunderstorm watch would have been helpful to increase situational awareness of forecasters (both TV and NWS) around the region.

Two, even though reflectivity imagery didn’t have a classic signature for a widespread destructive wind event – velocity data from both OKX and BOX was spot on in identifying the areas hardest hit. Even a giant blob of reflectivity (as opposed to a well-defined squall line) can produce significant severe winds.

Three, this event shows that true severe thunderstorm events can be just as damaging – if not more than a tornado event. In fact, the 125,000 power outages from this thunderstorm event in Rhode Island was worse than Hurricane Sandy. This event was one of the few around here to verify Severe Thunderstorm Warnings with multiple 50 knot measured gusts. This storm was the real deal. It’s important to differentiate between the run of the mill severe storms that take down a couple diseased tree limbs from the ones that have the potential to produce widespread and life-threatening weather. This is one of our biggest challenges here in southern New England during convection season.

A Nino Roaster?

The Pacific Ocean is on fire. Well, not literally on fire, but it’s pretty damn toasty out there. Chris Farley would be proud of this El Nino. The large mass of unusually warm water in the equatorial Pacific continues to grow and become even warmer – possibly rivaling some of the strongest El Ninos in recent memory.


Computer model forecasts continue to show a strengthening of this El Nino – and most experts I’ve spoken to believe this El Nino will be strong come winter.


The question is – what will a strong El Nino mean for New England? The three most powerful winter El Ninos since 1950 have produced below average snow in Connecticut. In fact, the mean and median snowfall for the top 10 El Nino events is below normal for both Hartford and Bridgeport.

The record 1997-1998 El Nino was accompanied by a mighty dud of a winter. Only 8.9″ of snow fell in Bridgeport (average is 29.0″) and in the Hartford area only 28.7″ of snow fell* (average is 48.4″). The 1982-1983 El Nino which peaked over the winter was only somewhat below normal in terms of snowfall, 46.4″ in Hartford and 23.0″ in Bridgeport. Rounding out the top 3 the winter of 1991-1992 was another dud with 23.6″ of snow in Hartford and 16.5″ in Bridgeport. Awful.

The top 10 El Nino and top 10 La Nina events are all relatively blah when it comes to snowfall.

Top 10 La Nina Events (since 1950)

  • Hartford mean snowfall 46.4″ / median snowfall 41.9″
  • Bridgeport mean snowfall 26.3″ / median snowfall 22.6″

Top 10 El Nino Events (since 1950)

  • Hartford mean snowfall 41.9″ / median snowfall 40.8″
  • Bridgeport mean snowfall 26.0″ / median snowfall 21.5″

Also of interest is that the top 10 El Nino events in Hartford have an average winter temperature above normal. The top 10 events average a winter temperature of 29.5 degrees (median of 29.7 degrees) compared to the 30-year average of 29.1 degrees. Additionally the top 3 El Nino events (’82, ’91, ’97) were all well above average with a mean of 31.5 degrees – more than 2 degrees above normal.


El Nino is only one part of the puzzle. While odds favor a warmer than normal and less-snowy than normal winter in Connecticut meteorologists will likely have their hands full. A strong El Nino will essentially put the subtropical jet stream on steroids. Plenty of moisture and energy will be around for “fun” storms – likely many messy storms with snow, mix, rain, etc. Even a “below normal” winter can produce a big one – remember the Megalopolitan snowstorm of 1983?


*Notes: To compensate for missing snowfall data in the official records I’ve used our weather observer in Collinsville for snow totals and adjusted them by a factor of 0.86 (I got this from comparing the mean snowfall at BDL and Collinsville for available years). This isn’t perfect but it is the best we’ve got. Also, to compute the top 10 El Nino/La Nina I averaged the MEI over December, January and February to come up with a winter ENSO strength. All stats are for ENSO events 1950-present.

Giant Hail in Maine

Courtesy: Bart Haley / Berwick, ME

Courtesy: Bart Haley / Berwick, ME

Getting baseball size chunks of ice to drop out of the sky in New England takes a special kind of storm. That kind of storm blew up over portions of southern New Hampshire and Maine on Sunday.


Not surprisingly, it took a fairly unique environment to produce this giant hail. The most important feature was a remnant plume of elevated mixed layer air that resulted in steep mid level lapse rates across New England. What we mean by this is an unusually steep drop in temperature between about 10,000 and 25,000 feet above our heads. We can quantify this by measuring the temperature difference over certain levels.


Between 700mb and 500mb the lapse rate on the evening Gray, ME weather balloon launch was 7.2 C/km. The Chatham sounding (away from thunderstorms and probably more representative of the pre-storm environment) had a very impressive 7.6 C/km lapse rate. Research by Banacos and Ekster has shown that many significant severe weather events in the northeast occur when an elevated mixed layer is present.

The EML is quite common in the plains – not so common here. With that in place the stage was set for a pretty impressive severe weather event. CAPE on the order of 3,000 j/kg and large CAPE in the hail growth zone (the part of the atmosphere between -10c and -30c) and enough effective wind shear (30 to 35 knots in southern Maine) was enough to give us some hail beasts.


This cross section of the storm from the Gray, ME radar shows the exceptional hail core. 70+ dbz values were up to 27,000 ft AGL and a giant three body scatter spike was present down radial.

Prior to dropping baseballs – radar gave a clear indication that this storm was about to drop some big ice.


On the GYX dual pol products you can see the high reflectivity in the upper left panel (huge TBSS too), while on the upper right you can see the low CC in that hail core. CC values ranged from 0.75 to 0.9 in the lowest part of the hail core along with ZDR (bottom left) well below zero. This is a very solid signature for significant hail – and in this case baseball size hail. Below – you can see the storm at peak intensity at about 20,000 feet above the ground with a giant hail spike accompanied by an 80 dbz pixel!!

80 dbz and giant TBSS at 21,000ft AGL

80 dbz and giant TBSS at 21,000ft AGL

Prior to reaching peak strength the storm exhibited very impressive wet hail growth between -20 and -30c. You can see the depressed CC (indicating mixed phase hydrometeors) way above the freezing level – a clear signature for big hail growth. Also present (not shown) was a large ZDR column indicating a tremendous amount of liquid water being lofted by the storm’s updraft.


GYX 5.1 degree tilt – sampling storm approx 27kft AGL

So that’s the story of the giant Maine hail. Never underestimate a sounding with fat CAPE!

Courtesy: Bart Haley / Berwick, ME

Courtesy: Bart Haley / Berwick, ME