The Deluge Has Ended

After nearly 4.5″ of rain in parts of southeastern Connecticut the heavy rain has come to an end. Here’s a look at some of the rain totals across the state – many of which are quite impressive!

  • Westbrook – 4.71″
  • Norwich – 4.46″
  • North Haven – 3.49″
  • Bridgeport (ASOS) – 3.15″
  • North Grosvenordale – 3.12″
  • Hampton – 3.08″
  • West Hartford – 2.22″
  • New Hartford – 2.06″


The rain was focused along a warm front that managed to sneak into southern Connecticut. The winds were out of the south and east in Groton, Westerly, and Westhampton Beach while Providence, Willimantic, and New Haven had northerly winds. An impressive theta-e gradient as well with mid 50s for temperatures and dew points in Groton and low 40s elsewhere in Connecticut.

The Yantic River managed to reach moderate flood stage with 4.5″ of rain through the basin while other rivers in northeastern Connecticut reached minor flooding benchmarks.



Not a surprise to see minor flooding given the setup with a strong low level jet on top of a stalled warm front near the south coast. A nice surge of high precipitable water air moved in overnight all in a region of strong quasi-geostrophic ascent and even a bit of elevated instability. Classic “sultan of sandbag” signal for southeastern Connecticut, Rhode Island, and southeastern Massachusetts.

New London County Hailstorm

Hail (and a screw driver) from Mark in Preston

Hail (and a screw driver) from Mark in Preston

When you think of severe weather in Connecticut you probably don’t think about New London County. The southeastern corner of the state is a pit for snow and 99 times out of 100 is where thunderstorms go to die. With the exception of hurricanes – weather weenies in New London County are used to let down after let down after let down.

Today, however, brought a severe storm bonanza to New London County. 3 separate storms brought severe hail to New London County including a report (and picture to go along with it!) of significant hail (hen egg size) in Ledyard.

2" diameter hail in Ledyard

2″ diameter hail in Ledyard from Josh

The severe weather threat today was a bit nebulous but still well forecast. A cold front had bisected the state by midday with a narrow corridor of marginal/moderate instability juxtaposed with fairly strong deep layer shear. Here’s the 18z RAP sounding for Groton around the time of the largest hail reports.

RAP BUFKIT analysis at 17z

RAP BUFKIT analysis at 17z

Dew points over 70F were underneath steep mid level lapse rates – on the order of 7.2 C/KM between 500 and 700mb. Impressive! While MLCAPE values were only around 1000 j/kg a fair amount of that CAPE was located in the hail growth zone (-10 to -30C) and 0-6km shear values were near 40 knots.

It shouldn’t come as a surprise that many of these storms exhibited mid level rotation and the mesocyclones were able to sustain some pretty impressive hail stones. I’ll also throw this out there that the models today were consistently underdone with progged CAPE values today. For example, the 12z OKX sounding had 2500 j/kg of CAPE which was far higher than model forecasts or even 12z model analyses. It’s conceivable that the 1000 j/kg of MLCAPE analyzed at 18z on the RAP may have also been underdone compared to reality.

Handfulls of quarter size hail in Groton

Handfulls of quarter size hail in Groton

The storm that dropped the largest hail pulsed up shortly before 2 p.m. over Montville and the Thames River just a (hail)stone’s throw from Mohegan Sun. The 1754 UTC volume scan from OKX shows an impressive hail core with 65 dbz echoes up to the -20c isotherm (over 20,000 ft AGL) and 72 dbz around 10,000 ft AGL.

1754 volume scan KOKX 88d

1754 volume scan KOKX 88d

While meteorological echoes (part anvil being blown northeast by 70 knot southwesterlies near the Equilibirum level and part other junk) masked the hail spike on the OKX radar but the radar out of Taunton showed an impressive (spatially and vertically) three body scatter spike.

KBOX 88d correlation coeffecient - note blue shading southwest of storm showing hail spike

KBOX 88d correlation coeffecient – note blue shading southwest of storm showing hail spike

Shortly after the storm pulsed over Montville the core of the storm dropped and so did golf ball and hen egg size hail in the far northwest corner of Ledyard near the Thames River and along and west of Avery Hill Road just south of Route 2A. Dual Polarization products showed the hail core descending with a clear signal of hail in Ledyard by 1758 UTC (at 4500 ft AGL).


1758 UTC KOKX volume scan. Clockwise from top left – 0.5 degree base reflectivity, correlation coefficient, specific differential phase, differential reflectivity

Within an area of high reflectivity (Z between 50 and 60 dbz) you can see a noticeably depressed area of correlation coefficient – in some cases near 0.90. There are also areas of ZDR near zero or even subzero which shows that hail is dominating the signal (hail tumbles as it falls so appears spherical to dual pol radar which leads to a differential reflectivity value near 0).

One thing that’s somewhat interesting is that the southeastern part of the storm has very high KDP values – nearly 4 deg/km while the northwest part of the storm was much lower. While the hail signal was present throughout this region the spike in KDP over the center of Ledyard may indicate a lot of water coated sub-severe hail while the lower KDP over the Thames River and northwest Ledyard in the hail core was where the larger/non-water coated hail was falling. Indeed, this matches up with the reports we received of hail near (or even over) 2″ on the Thames River in far northwest Ledyard with dime to quarter size hail in the center of town.

Golf ball size hail in Ledyard melted to this size about 10 minutes after it fell. Thanks Tom for the picture!

Golf ball size hail in Ledyard melted to this size about 10 minutes after it fell. Thanks Tom for the picture!

I went back and searched through SPC’s storm database and then double checked some of the events in NCDC’s storm data. It appears that the 2″ hail from Ledyard is the largest hail event reported in New London County since 2″ hail was reported in Old Lyme during the 1995 super hailstorm. Incidentally, the largest hail report I can find in New London County was from Groton in May 1969 where baseball size hail was reported. 2.75″ hail (baseball size) only shows up 4 times in the Connecticut SPC storm event database since 1955 and 3 of the 4 were from 1995 and all have been along or east of the Connecticut River.

Hurricane Carol – The Monster of 1954

Hurricane Carol in Groton Long Point (photo credit

Hurricane Carol in Groton Long Point (photo credit

Hurricane Carol was one hell of a storm. The category 3 hurricane made landfall near Groton and brought devastating flooding and wind damage to parts of southeastern Connecticut and Rhode Island (from South County to Providence). Carol is a bit forgotten – the impact was less than 1938 and it was overshadowed in the state by the remnants of Connie and Diane in 1955 that paralyzed the state with a biblical flood.

The reanalysis by Chris Landsea and the work by Jarvinen shows a landfall of Carol near Groton. This is about 20 miles east of the “Best Track” landfall location in Old Saybrook. The reanalyzed landfall location makes sense given the damage documented in Connecticut.

Hurricane Carol reanalyzed track and wind swath (Landsea et al.)

Hurricane Carol reanalyzed track and wind swath (Landsea et al.)

Carol made landfall on the morning of August 31, 1954. The storm brought category 2 force winds (sustained near/over 100 mph) in Groton and Stonington and a vicious storm surge. In New London the storm tide reached 9.6 ft MLLW, only exceeded by the 1938 storm which was 10.6 ft MLLW. The surge in New London at the tide gauge was 6.5 feet though a study by the Army Corps of high water marks after Carol revealed surges of 8-10 feet were common throughout southeastern Connecticut. In Stamford – many miles from the landfall location – the storm tide reached 10.3 ft NGVD which was only exceeded by Sandy and the 1938 hurricane.

Hurricane Carol in Connecticut (NOAA)

Hurricane Carol in Connecticut (NOAA)

Hurricane Carol is widely viewed to be the “most tropical” of hurricanes to hit Connecticut. What’s meant by that is that while the storm was likely going extratropical transition – Carol was the most purely tropical of the storms to strike Connecticut. Here’s one example of that – take a look at this picture taken from the old Griswold Hotel in Groton near the mouth of the Thames River.

Flooding during the eye of Carol (Courtesy: Town of Groton)

Flooding during the eye of Carol (Courtesy: Town of Groton)

The storm had a classic “eye” and the precipitation distribution from the storm – both east and west of the storm track – was relatively symmetrical. That’s very unusual for a landfalling New England hurricane!

In Bridgeport the monthly climate report indicates 1.62″ of rain fell on the day with a peak sustained wind out of the NNE at 40 mph with a gust to 60 mph. The report read as follows “Hurricane Carol did extensive damage to the shorelines of Connecticut. Storm center passed 50 miles east of station. Lowest pressure noted at 0920 EST, with NNE winds of 40 MPH and gusts to 60 MPH noted at 0928 EST. Airport was inundated with a maximum of 2 feet of water.” The lowest pressure reading in Bridgeport was 28.87″.

In Hartford at Brainard 2.36″ of rain fell. No wind data was recorded. In New Haven at Tweed 2.63″ of rain fell with a sustained wind of 38 mph out of the northeast. The report read “Heavy storm on the 31st. Gusts to 65 MPH 9-10 A.M. Caused heavy water manage along shore. Lowest pressure 28.77″ 9:10 A.M.”

At Windsor Locks 1.95″ of rain fell with a suatained wind of 56 mph out of the northwest! A 64 mph wind gust was recorded in the monthly report.

Cooperative observer reports for August 1954 include some interesting highlights.

  • Baltic, CT recorded 4.10″ of rain “HURRICANE” was listed in the important wx conditions space.
  • Brooklyn, CT recorded 4.00″ of rain
  • The observer in Colchester wrote a great blurb about Carol – though I can only decipher about half of it! colchester
  • The Danbury coop observer reported wind gusts near 60 mph
  • In Derby the coop observer reported the following “High winds on Aug. 31. Hurricane “Carol” considerable damage to trees. An additional amount of rain – 1.05″ fell from 8:30 a.m. to 12:30 p.m. on Aug 31″ That 1.05″ is in addition to the 1.56″ reported at 8:30 a.m. in Derby.
  • Even in Falls Village the observer mentioned the wind on the 31st “wind on the 31st caused considerable damage to untilities.”
  • In Groton the observer wrote “Aug 31. Winds of hurricane force from 9:00 A.M. to 11:30 A.M. did much damage to buildings, trees, and boats in this area. Power and telephone lines severely damaged causing loss of service for several hours.”
  • In Mansfield at the dam the coop observer from the Army Corps mentioned on the 31st “Note: temperature readings may be inaccurate because the box blew over during height of storm”
  • Here’s what the Middletown observer wrote: middletown
  • Here’s the observer’s remarks in New London at Fort Trumbull. newlondon
  • In Norfolk, observer Norman Smith summarized Carol this way, “The feature of the month that was most notable was the passage on the 31st of an Atlantic hurricane. The storm here brought 2.59″ of beneficial rain with shifting gale winds from the E NE and NW. There was some damage to power and telephone lines and roads were blocked by uprooted trees and branches.”
  • In Putnam 4.25″ of rain fell and the observer noted the hurricane passed east of Putnam with considerable damage to crops and trees with some property damage.
  • In Storrs 3.35″ of rain fell. The observer wrote, “On the 31st of August Hurricane “Carol” hit eastern Connecticut hard with 60 mile wind – higher gusts – and nearly 2″ of rain between 8 and 11 a.m.
  • 4.36″ of rain fell in Westbrook – here’s the observation remarks.westbrook

The observations show a core of very heavy rain – 4″-6″ of rain near the center of the storm with less rain on the western periphery. Again this symmetry is unusual and shows that the eyewall was likely quite in tact and the storm was still quite “tropical” at the time of landfall.

Carol produced $50,000,000 in damage in 1954 dollars in Connecticut and killed 65 across New England.  The storm is the only category 2 to strike Connecticut in modern times (1938 was a category 3 in Connecticut) and remains the strongest storm to strike the state in the last 59 years.

Friday Severe Weather and Tornado Threat

I’ve been watching Friday’s severe weather potential with a bit of interest since Sunday. It hasn’t been clear whether ingredients would come together for severe weather with different computer models offering up very different outcomes. This afternoon, both the 18z NAM and GFS show the potential for a rather sizable severe weather event including tornadoes.


The synoptic setup is pretty straightforward. A deep closed-low near the southern tip of Hudson Bay continues spinning with a number of shortwaves advancing east across the region. At the same time, mid level wind fields strengthen during the day tomorrow approaching 40 knots out of the west at 500mb while a weak wave of low pressure forms along a front to the west strengthening the southerly flow in the boundary layer.


Now the GFS (the NAM has been showing this for a while) develops a strong low level jet over the region Friday. This creates a highly sheared environment in the 0-3km layer. Take a look at these hodographs from the 18z GFS and NAM for 18z tomorrow valid at 18z Friday.


The long and curved hodographs are very impressive. In addition the models are developing some surface based instability tomorrow. The extent of the destabilization is very unclear, however.

18z GFS sounding for Hartford valid 18z Friday

18z GFS sounding for Hartford valid 18z Friday

So what will happen? This setup does seem like it could morph into a classic low CAPE/high shear severe weather day if the current depiction by the GFS and NAM comes to pass. Hail is quite unlikely but damaging winds and tornadoes are possible. The threat will depend on the amount of destabilization which is unclear and dependent on cloud cover and convection earlier in the day. That said, lifted condensation levels (LCLs) are quite low so any storm that begins rotating can produce a tornado.

While the NAM is likely overdone at 18z Friday – many of the parameters, especially the 70 j/kg of 0-3km CAPE and storm relative helicity >300 m2/s2, is sufficient for tornadoes and even significant tornadoes per Davies, 2006.

WVIT Severe Threat

We’ll have to watch tomorrow’s setup closely – I’ll keep you advised. If some of the model solutions turn out to be correct tomorrow may be a very busy last day of work before I leave for vacation!

Another EF-1 Tornado in Connecticut


When Nick Stanczyc posted this picture on Instagram someone forwarded it to me within moments. I couldn’t believe it! The classic looking tornado was just about a half mile away from the intersection of Rt 30 and Rt 31 in Mansfield near Storrs.

On the 24th anniversary of one of the state’s most violent tornado outbreaks we got hit again – but by a much weaker storm. The first sign of trouble came around 1:30 when a rather unimpressive thunderstorm began to show signs of rotation over Watertown, Thomaston, and Plymouth.


For a period of time while near Route 8 there was a Delta-V of 40 knots on adjacent gates in the storm’s radial velocity (i.e. 20 knots inbound and 20 knots outbound) about 4,000 ft above the ground. In addition, the reflectivity signature indicated a small hook echo or appendage associated with the storm. Unlike the July 1st storm (which was a low topped supercell) the rotation was seen through a large portion of the storm from about 4,000 ft at the lowest radar sample to 22,000 ft.

The 17z RAP analysis sounding shows an environment that was capable of producing rotating storms. The environment was quite unstable with surface based CAPE values exceeding 2000 j/kg. In addition there was fairly impressive directional and speed shear through the lowest 3km of the atmosphere. Because of this there was fairly sharp curvature of the hodograph between 1 and 2km above the ground. It’s no surprise that a somewhat discrete cell ahead of the main cluster was able to take on supercellular characteristics.


While the radar presentation for the storm in Watertown was arguably the most impressive of the day the storm did not produce damage and did not produce a tornado. The likely reason why is that there was little 0-1km shear – the bulk of the shear was above that level. The height of the lifted condensation level or LCL (think of this as the cloud base) was around 800 meters. While low, this was not as low as we saw during the July 1st tornadoes or the subsequent storms in northeastern Connecticut (where LCLs were near 300 meters!)

The overall synoptic environment was characterized by a deep trough in the Great Lakes with a shortwave rotating underneath it through New York and New England. QG forcing for ascent was greatest ahead of this with fairly widespread shower and thunderstorm coverage.


Shortly after 5 p.m. – 2 supercells had developed in Connecticut. One produced some damage near Tolland and another in Coventry and Mansfield. Here’s the 21z sounding from the RAP analysis at the Windham Airport. Most of the shear is located in the 0-1km layer with a helicity of 111 m2/s2. There is about 1100 j/kg of sb CAPE and LCL heights are VERY low – about 300 meters!


The storm in Tolland shortly after 4 was totally unimpressive on radar. There’s some signs of a weak low level mesocyclone but the storm itself was about as exciting on radar as a snow shower. Even so, it was apparently able to produce some damage in parts of Tolland just west of the Green as it crossed I-84. This damage was ruled to be from straight line winds and not a tornado with trees knocked down in a southwest to northeast fashion along the storm’s path.

The most likely “cause” for the wind damage in Tolland wasn’t a classic microburst but rather a damaging rear flank downdraft that produced a corridor of >50 knot southwesterly winds along the storm’s path. This would make sense given the fact many people saw a funnel cloud though there was no conclusive evidence of a touchdown.


As the storm moved northeast the mesocylone remained and even strengthened a bit out toward Willington and Union. It’s unclear to me what happened shortly after 5 p.m. The first mesocylone over Tolland weakened as it crossed into Stafford and Willington while a second, and stronger, mesocyclone develop on the southeast flank over Willington and then moving into Ashford and Union. This turned into the more powerful mesocyclone and peaked with a low level delta-V of about 40 knots and a tornado warning from the folks in Taunton. Whether this was simply the mesocyclone “cycling” or if this was a splitting supercell it’s hard to say based on the radar data I’m looking at now. Anyone have any ideas?


The most impressive storm of the late afternoon developed shortly after 5 p.m. to the southwest of the Willington/Union storm. Rotation was evident as the storm left Glastonbury and it strengthened in Hebron and Andover. The storm produced an EF-1 tornado from Andover to Mansfield with the worst damage concentrated in an area of Coventry near Coventry Lake.


Given the super low LCLs and enough turning in the lowest kilometer of the atmosphere it’s not a surprise that this storm produced.

Given the advances in doppler radar (super resolution!) and the proliferation of smart phones with cameras and social media it’s not a surprise we’re hearing a lot more about these tornadoes than we have in the past. Tornadoes are nothing new here in Connecticut and even though it seems we’ve been getting an unusual amount of late this is really the way it has always been. The state is vulnerable to tornadoes and while most of our tornadoes are weak (EF0 or EF1) we have a long history of significant and violent tornadoes. Being able to hear about damage within moments and get pictures of the tornado shortly after touchdown thanks to everyone with an iPhone or Droid allows us to cover tornadoes and severe weather much more effectively.

While it’s been a busy year for tornadoes in Connecticut – to be honest – we’ve just been sort of unlucky. Marginal setups have been able to produce spinners. Hopefully yesterday’s was the last tornado of the year!