Relentless rains continued to plague areas from northern New Jersey into southern New England as lackadaisical Post-Tropical Cyclone Henri pushed across the area. A much more disastrous scene was unfolding west of Nashville in Tennessee: Residents continued to assess damage and search for victims after a severe, localized flash flood early Saturday took at least 21 lives (see below).
After sweeping into southwest Rhode Island as a tropical storm with 60 mph winds at midday Sunday, Henri moved west-northwest into the Hudson Valley of New York while slowing to a crawl, heightening the risk of flooding rains. At 5 p.m. EDT Monday, Henri – newly reclassified from tropical depression to post-tropical cyclone – was centered about 60 miles north-northeast of New York City, moving east at 9 mph. Top sustained winds were down to 25 mph. Henri is predicted to head eastward across southern New England, reaching Cape Cod by Tuesday morning, then accelerate eastward on Tuesday toward Nova Scotia.
There were very few reports of sustained tropical-storm-force winds with Henri, but gusts in the 45- to 70-mph range were common along and just inland from the southern New England coastline.
There remains a moderate risk of excessive rainfall through Tuesday morning near Henri’s track, according to the National Weather Service’s Weather Prediction Center. Widespread amounts of 1-2” and localized 2-4” totals are possible in southern New England.
Henri’s slow westward-then-eastward loop across southern New England over a period of more than 24 hours while still classified as a tropical cyclone is extremely unusual if not unprecedented. Nearly all tropical depressions, tropical storms, or hurricanes in the region move more rapidly north or northeast, hauled toward the Atlantic by strong upper-level winds more typical at that latitude than the exceptionally weak steering currents present on Monday.
Including the predecessor rain bands on Saturday night that inundated parts of New Jersey and the New York City area, the heaviest rainfall amounts in each state from Henri as of 11 a.m. EDT Monday, August 23, included:
Connecticut: 5.00” east-northeast of Manchester;
Delaware: 3.42” 8 miles south of Lewes;
Massachusetts: 2.77” at Feeding Hills;
New Jersey: 8.91” at Cranbury; 8.66” at Jamesburg;
New York: 8.03” at Brooklyn; 7.04” at Central Park;
Pennsylvania: 6.67” 1 miles south of Blakeslee;
Rhode Island: 3.81” 1 mile south of Westerly.
Was this the heaviest rainstorm in New York’s history?
The peak of rainfall in New York City was a two-hour period on Saturday night with 1.94” of rain from 10 to 11 PM EDT and another 1.84” from 11 PM to midnight. Based on an initial automated search of Central Park data, the 1.94” appeared to be the heaviest one-hour total observed there since at least 1943. However, it does not appear to be an all-time record for the site, where records began in 1869. According to station logs tweeted by Brian Brettschneider, Central Park recorded 2.48” of rain in one hour on August 12, 1925, and 2.09” in 30 minutes on July 10, 1905.
Estimates of climatological return periods from the NOAA Atlas-14 product show that Central Park would normally expect a 2”/hour rainfall about once every 10-15 years, perhaps a bit more often in the current human-warmed climate. Substantially more impressive is the two-hour downpour of 3.78” between 10 p.m. and midnight Saturday night, which has an Atlas-14 return period of between 100 and 200 years. (Thanks to Jan Null at Golden Gate Weather Services for the Atlas-14 reference.)
At least 21 killed in Tennessee flash flood
A devastating flash flood early Saturday morning took at least 21 lives in and near the town of Waverly, Tennessee (pop. 4,100), about 65 miles west of Nashville. There were still 40 people missing on Monday morning, and residents and authorities are fearful that many of them have died in the flood.
All but one of the known fatalities occurred in Waverly, a town on Trace Creek, which feeds into Kentucky Lake on the Tennessee River about 10 miles to the west.
As recently as the 1970s, flash floods in the United States sometimes resulted in 100 or more deaths. More recently, with heightened awareness of flash flood risks and improved communication of flash flood warnings, death tolls have tended to be much smaller. It appears the Tennessee disaster is the nation’s deadliest localized flash flood in decades. A flash flood on October 18, 1998, killed 31 people in San Marcos, Texas.
Flash floods are typically driven by rapid water rises in small channels as a result of persistent thunderstorm rains. They are distinct from broader-scale river flooding and from coastal storm-surge flooding during hurricanes. Some larger-scale flooding events, such as the one that killed 27 people across Tennessee on May 1-3, 2010, include both flash and river flooding, as did the deadly inland floods associated with such tropical cyclones as Floyd (1999), Allison (2001), and Harvey (2017).
Saturday’s flooding was well forecast and warned. Many parts of Middle Tennessee had received 2-5” of rain in the days leading up to the Saturday flood. A flash flood watch was issued at 5 p.m. Friday for the affected area, valid from 1 a.m. Saturday through Saturday afternoon, as computer models were already suggesting a band of heavy thunderstorms would develop in the overnight hours. By the predawn hours Saturday, a corridor of intense rainfall was moving south-southeast across the area in a “training echo” fashion. Flash flood warnings were issued for the hardest-hit area beginning at 3:29 a.m., and the warning was upgraded to a flash flood emergency at 7:47 a.m.
Rainfall analyses based on a blend of radar and rain-gauge data show a corridor of 10-20” amounts over the past week (see Figure 2), with most of that falling during the Saturday storm. The highest rainfall amounts measured by CoCoRaHS and NWS cooperative observers in the area included:
17.26” about 9 miles north of Centerville (CoCoRaHS);
17.20” at McEwen (NWS COOP);
13.76” about 6 miles west-southwest of Dickson (CoCoRaHS).
The 17.20” at McEwen is being investigated as a potential 24-hour rainfall record for the state of Tennessee. The CoCoRaHS reading of 17.26” would actually top the McEwen amount, presuming it fell within a 24-hour period. However, verifying this CoCoRaHS report would be a challenge, given the observer’s report from Saturday: “Unfortunately, after 5,000 plus reports this will be the last from this station. The flood took the deck and rain gauge down river, then wrecked the interior, and finally collapsed the brick back side of the house, leaving open framework.”
Three systems to watch in the Atlantic
After a brief quiet spell, the Atlantic may turn busy quickly starting later this week. At 2 p.m. Monday, the National Hurricane Center, NHC, highlighted three areas for potential development.
A disturbance in the eastern Atlantic south of the Cape Verde Islands, classified as Invest 98L, may be the first of the three to develop. On Monday afternoon, 98L featured a large but disorganized cluster of strong showers and thunderstorms (convection) in a moist atmosphere (mid-level relative humidity of 75%), with some spin already evident. The system will be passing over warm waters with sea surface temperatures (SSTs) of 27-28 degrees Celsius (81-84°F) for the next several days. Moderate to strong wind shear at times, together with infusions of dry air from the Sahara Air Layer just to its north, will be inhibiting factors for 98L, but the shear is likely to decrease by midweek.
NHC gives 98L a 10 percent chance of developing into at least a tropical depression by Wednesday and a 40 percent chance by Saturday.
Another disturbance, 97L, was located a few hundred miles west of the Cape Verde Islands in the central Atlantic, marked by a band of only moderately strong convection. Much like 98L, 97L will be passing over warm 27-28°C waters by Tuesday within a relatively moist atmosphere (midlevel RH of 55-60%), but it will be hindered by even stronger wind shear (15-25 knots through Wednesday).
NHC gives 97L a 10 percent chance of developing into at least a tropical depression by Wednesday and a 40 percent chance by Saturday. If it develops, 97L will likely end up recurving out to sea long before getting to the western Atlantic.
Perhaps most troublesome by the end of this week could be the broad area of disturbed weather located over the southeast Caribbean on Monday. Although unimpressive for now, this disorganized complex of convection is expected to gradually organize as it moves west toward the western Caribbean. NHC gives this system a near-zero chance of development through Thursday but a 40% chance between Thursday and Saturday.
The storm formerly known as Linda sweeps into Hawaii
Although it is now classified as a remnant low by the Central Pacific Hurricane Center (CPHC), what was once category 4 Hurricane Linda was making its presence felt in Hawaii on Monday.
As it moved westward, ex-Linda continued to boast a well-organized low-level circulation on Monday, evident in satellite loops, with pockets of 30-knot wind evident on ASCAT scatterometer imagery. Phase space diagrams from Florida State University showed on Monday that Linda still had a symmetric, warm-core circulation. As recently as Sunday night, satellite imagery had shown a cluster of strong thunderstorms holding their own on the north side of ex-Linda (see Figure 6 above). However, all deep convection was gone from Linda by Monday morning. Had the convection not vanished, there would have been a reasonable case for redesignating the storm as Tropical Depression Linda. (In the area served by NHC, a tropical cyclone can officially persist for as long as 18 hours after the disappearance of convection before being considered a remnant low.)
Ex-Linda’s center will pass over or just south of Oahu on Monday afternoon, August 23. A flash flood watch was in effect for the entire state of Hawaii. Linda’s remnants were predicted by the National Weather Service office in Honolulu to bring widespread 2-4” rains with isolated totals of up to 10”. The islands will also face winds of up to 30 mph, with some gusts as high as 50 mph, especially on the northern side of ex-Linda’s circulation. “The smaller islands from Molokai and Lanai to Kauai appear to be in direct line to experience strong and gusty winds, along with heavy rainfall and thunderstorms that could potentially produce flash flooding,” the NWS/Honolulu office wrote in its forecast discussion at 4 a.m. HST Monday.
Overall, ex-Linda could bring as much wind and rain to parts of Hawaii as Henri brought to New York and southern New England.
Linda had developed into a tropical storm on August 10 in the Eastern Pacific and was classified as a remnant low by CPHC on August 20. It’s easy to see how Linda’s circulation has held together so well into Hawaii: Linda survived longer than expected after it evolved into a powerful annular hurricane – the type dominated by a large eye and a single robust ring of convection that gives it a truck-tire appearance. Annular storms are typically among the most resilient and the most resistant to weakening of all tropical cyclones.
Website visitors can comment on “Eye on the Storm” posts. Please read our Comments Policy prior to posting. Comments are generally open for 30 days from date posted. Sign up to receive email announcements of new postings here. Twitter: @DrJeffMasters and @bhensonweather
Comments