NWSFO Boston Storm Series Report
#99-2
(To animate satellite imagery, hold down right mouse button, and select play)
Tropical Storm Floyd
Event Date: September 16, 1999
This page was last updated on September 23, 1999
Damage Reports and Fatalities/Injuries
Products and Service Assessment
After forming in the central Atlantic Ocean on September 8, 1999, Tropical Storm Floyd entered southern Connecticut during the evening of September 16. Floyd quickly departed southern New England during the early morning hours of September 17th. With its passage through NWSFO Boston's County Warning Area (CWA), Floyd predominantly impacted southern New England with heavy rain which led to small river and stream flooding, minimal mainstem river flooding, and wind damage associated with heavy squalls.
Tropical Storm Watches and Warnings, along with Hurricane Warnings were issued by the Tropical Prediction Center (TPC) for the NWSFO Boston CWA. NWSFO Boston issued Hurricane Local Statements, High Wind Watches and Warnings, Wind Advisories, Flood Potential Statements, Flood Watches, Flood Warnings, Public Information Statements, and Special Weather Statements. These products generally proved to be accurate and timely.
Floyd's huge impact to the entire east coast of the United States and its associated media coverage caused enhanced local media interest and emergency manager's desire for information. NWSFO Boston initiated two new methods to assist in communication with these two important user groups: "Meet-me" conference calls for the emergency management community, and scheduled press briefings held in the conference room of NWSFO Boston facility for local media. Both of these practices were very beneficial in providing these vital external user groups, hence the general public, with a consistent message on the threat of Floyd to southern New England.
Floyd formed in the central Atlantic on September 8, 1999. Floyd moved at a west-northwest track for about a week, strengthening during this timeframe. Floyd peaked as a strong category 4 hurricane on the Saffir-Simpson Scale during the evening hours of September 13th as it neared the Bahamas. As Hurricane Floyd skirted the Bahamas, he turned to the northwest, paralleling the Florida and Georgia Coast. The close approach to the southeast United States coast prompted mass evacuations in Florida, Georgia, and South Carolina. Problems with evacuations occured throughout the southeast, with South Carolina's Governor issuing an apology to his constituents with regard to mishandling of evacuations.
Hurricane Floyd actually had little sensible impact to Florida, Georgia, and South Carolina. Floyd finally made landfall near Cape Fear, North Carolina during the predawn hours of September 16. In the wake of Floyd, massive flooding occured over much of eastern North Carolina over the next several days.
Floyd continued at hurricane strength as it left North Carolina. A University of Oklahoma Doppler-on-Wheels measured sustained winds of 81 mph with a gust of 123 mph over eastern North Carolina during the early morning hours of September 16. This measurement, along with a North Carolina emergency management report sustained wind of 98 mph with a gust to 130 mph, prompted TPC to upgrade the Tropical Storm Warning for southern New England to a Hurricane Warning during the midnight shift on Septemeber 16th. Floyd weakened during the late morning of September 16th, and this caused a flip-flop back to a Tropical Storm Warning for southern New England in the September 16 afternoon advisory out of TPC. The complete path of Floyd is shown in Figure 1.
Figure 1. Tropical Storm Floyd's track and intensity from September 8-17, 1999.
As shown in Figure 2, the Eta model 36-hour forecast of mean sea level pressure valid at 1200 UTC 17 Septemeber was fairly accurate in its position with an approximate 50 mile positional error. Interestingly, for Hurricane Edouard in 1996, the Eta model also showed skill at forecasting the tropical cyclone's position at 36-hours.
In contrast, TPC's position at 36-hour position valid at 1200 UTC 17 September was approximately over Newport, Rhode Island, approximately 120 miles too far south.
Typically, forecasting tropical cyclones beyond 36-hours is difficult. Floyd proved to be no different. Forecasts at the 36 to 72 hour range from TPC were of little benefit. Additionally, there was considerable conflict between the TPC forecast track and the Hydrologic Prediction Center (HPC) tracks at these extended ranges. This has a considerable impact on briefing emergency managers and on developing NWSFO Boston public and marine products.
Figure 3 shows the official TPC forecast track from Advisory #28, issued at 1500 UTC Tuesday, September 14th. It shows the position of Floyd at 0000Z/17, or about the time of maximum impact to southern New England, about 300 miles too slow to the southwest. The TPC track also forecast Floyd into and past the stalled cold front (not shown), which was draped across central New York state and Pennsylvania.
Twenty-four hours later, TPC Advisory #32, released at 1500 UTC Wednesday, September 15, is shown as Figure 4. While certainly better then Advisory #28, this forecast still was too slow in bringing tropical storm force winds into the NWSFO Boston CWA. A Tropical Storm Warning was posted for much of southern New England coast with the afternoon forecast package on Wednesday the 15th.
During the midnight shift of Thursday, September 16th, TPC upgraded the Tropical Storm Warning to a Hurricane Warning based on a University of Oklahoma Doppler-on-Wheels which measured sustained winds of 81 mph with a gust of 123 mph over eastern North Carolina. Another observation, this one from a North Carolina emergency management agency, reported a sustained wind of 98 mph with a gust to 130 mph in New Bern, North Carolina. These reports prompted the TPC forecaster to reduce the initial intensity of Floyd down to 90 kts, but rather having a commensurate decrease in the intensity for the 24-hour forecast, he increased it from 60 to 65 kts to hurricane strength. As it turned out, the last reported hurricane-force sustained wind occured at 14Z at Duck Pier C-Man, North Carolina.
The highest sustained wind from Floyd in the NWSFO Boston CWA was recorded at the Buzzards Bay Coastal Marine Network (CMAN), with sustained 47 knots from 180 deg at approx 0300 UTC September 17th. A peak gust of 57 knots was recorded at the Buzzards Bay CMAN between 0400 and 0500 UTC September 17th.
The highest peak gust in the NWSFO Boston CWA was recorded at the New Bedford Hurricane Barrier, with a 64 knot gust at 0600 UTC Sep 17. Elsewhere, a gust of 63 knots was recorded at Brewster at approx 0545 UTC September 17, and a gust to 62 knots was recorded at Hyannis.
The lowest pressure in the NWSFO Boston CWA was recorderd at Hartford ASOS (Brainard Field - HFD), with 985.4 mb at 0253 and 0259 UTC September 17. Just outside of the CWA, the Meriden, CT ASOS recorded 984.5 mb at 0156 UTC September 17.
The western portion of NWSFO Boston's CWA, which was to the left of Floyd's track, is where the greatest amount of rain fell. The swath of maximum storm total rainfall values within the CWA occured on the east slopes of the Berkshires, with 7 to 9 inches routinely observed. The rainfall was most likely enhanced by upslope on the strong southeast fetch along with being on the west side of Floyd's theta-e axis which advected moist tropical air over a seasonably cool air mass. The peak stom total rainfall within the CWA occured at Bristol, CT, with 10.80 inches. Overall, the rainfall was characterized by a steady moderate-to-heavy rain along and west of Floyd's track, with no torrential rainfalls reported. Figure 5 is an animation of mosaiced reflectivity from the National Climatic Data Center. The maximum one hour rainfall that was recorded by ASOS was at Blue Hill Observatory (MQE), with 0.84 inches in an hour ending at 2100 UTC September 16 and at Westfield, MA (BAF) ending at 0000 UTC September 17.
The rainfall to the east of Floyd's track was showery in nature, with only 1 to 4 inches reported. It appeared from both radar and spotter observations that the showery convective squalls on the east side of Floyd were the source of the damaging wind gusts. It appears that the rain-soaked ground may have played a major role in the tree damage because of the low magnitude of wind speeds that were observed were not commensurate with the scattered damage.
The maximum storm surge occured at the Fox Point Hurricane Barrier, with a 4.2 foot surge recorded between 0215 and 0345 UTC on September 17. Other storm surge values on September 17th include the Providence National Ocean Survey (NOS) site with 3.7 feet recorded at 0206 UTC, Newport NOS 2.6 feet at 0136 UTC, New Beford Hurricane Barrier 2.5 feet at 0200 UTC, Boston NOS 1.6 feet at 2100 UTC, and Nantucket NOS 1.3 feet at 0312 UTC.
The only report of significant beach erosion was from Norton Point Beach in West Tisbury on Marthas Vineyard, where a half-mile of barrier beach was eroded 30 to 50 feet.
Flooding was the most serious consequence from Tropical Storm Floyd, with numerous reports of small river and stream flooding, predominantly in western Massachusetts, coicident with the heaviest rains. Most notable were the Quinnipiac River in Southington, CT and the Westfield River in Westfield, MA. Several roads in Hampden County were washed out. Flooded roadways forced boat rescues in Bristol, Connecticut. An emergency manager stated that most every brook in Hampden County was out of its banks. The Fall River Dam in Bernardston, MA suffered a catastrophic failure and will likely not be rebuilt. Some damage was sustained at the Chicopee Falls Dam and the Bray Pond Dam in western Massachusetts. A bridge in Granby, MA was determined to be at risk for collapse, isolating about a half dozen homes. A few mainstem rivers out west (Connecticut, Farmington, and Westfield Rivers), and the Piscataquog River in the east, went into minor flood on Friday, September 17th, but receded below their minor flood stage in several days.
Minor to isolated moderate coastal flooding was reported along the south coast. A shore road was washed out at Matunack, a beach community in South Kingston, RI. A NWS meteorologist surveyed the Sakonnet Point area of Rhode Island. He noted that an area of blacktop parking lot 300 x 60 feet was washed out and approximately 300 boulders roughly the size of wheel barrows were washed up onto the parking area. Also, the low coast road along Horseneck Beach in Westport, MA had 2 inches of sand and some small rock rubble covering about a mile long stretch. There have been several other reports of wash overs along the south coast, without any significant damage.
The most significant impact was extensive small stream and some river flooding in Western Massachusetts, Hartford county Connecticut, and to a lesser extent in Northern Rhode Island. Wind damage was scattered throughout the CWA, with downed trees and powerlines reported between 2000 UTC 16 September and 0200 UTC 17 September. For the most part, only minor coastal flooding was reported along the south coast of Rhode Island and Massachusetts, in part due to a relatively low astronomical high tide.
There are three known fatalities that were indirectly attributable to the storm. The news media reported a fatality due to a single car crash in Monson, MA, and a Rhode Island hospital (in Providence) patient died after his breathing apparatus failed due to a loss of power. The hospital's emergency generator apparently failed to kick in after commercial power was lost. The news media also reported a drowning in Connecticut from a person who attempted to kayak the swollen Farmington River the day after the storm. One injury due to a falling limb was reported in shirley, MA but has not been confirmed. An indirect injury occurred in Whately, MA, when a damage assessor fell into a washed-out area. No other injuries have been reported thus far. Dollar damages are not yet available.
NWSFO Boston started three major external coordination initiatives during Floyd:
The feedback from the user community has been very positive. In essence, this office has been praised for providing a timely and accurate assessment of the threat from Floyd. The state emergency managers (input was received from Connecticut, Massachusetts, New Hampshire, and Rhode Island) indicated that the Hurricane Local Statements and other forecasts/statements from this office were valuable. A similar message has been received from other federal agencies such as FEMA Region I and the New England Division of the Corps of Engineers.
There was particular praise regarding the value of the emergency management coordination conference calls provided leading up to and into the event. We received a few logistical suggestions regarding the coordination calls, which we will implement. For the first time that any of us can remember, our emergency management coordination calls not only attracted a variety of partnership organizations via phone but also in person presence of the Director of FEMA Region I, the Director of the Massachusetts Emergency Management Agency (MEMA), the Director of MEMA Area II, top ranking State Police officials, and the Charles River Dam Manager. Other key emergency management decision makers on the conference calls included the American Red Cross, Steamship Authority, and U.S. Coast Guard. FEMA Region I complimented us on the coordination associated with Floyd and noted that they held follow up conference calls with federally recognized Indian tribes, including the Narragansetts, the Passamaquoddy Indians (Maine), and the Mashautucket - Pequots (Connecticut). The Rhode Island Emergency Management Agency shared with us that the Governor of Rhode Island personally dialed into at least a couple of the coordination calls and found our briefings very useful.
Key emergency managers had been briefed on the potential for an impact from Floyd as early as the end of the previous week (when there appeared to be a risk of an impact as early as Tuesday or Wednesday). The issuance of Heavy Surf Advisories Tuesday evening (9:15 pm) with a Special Weather Statement provided an early outlook to the public on the potential impact of Floyd.
At 5 am Thursday, the eastern half of our interior county warning area had been placed under a High Wind Warning and most western zones under a High Wind Watch. This provided about 14 hours lead time (winds reaching high wind criteria entered our county warning area around 7 pm Thursday evening). The High Wind Watch for the western zones was upgraded to a warning at 11 am (providing about 8 hours lead time). Our wind forecasts were a little high (e.g. peak gusts 80 - 90 mph southeast coast versus observed 65 - 75 mph) but seemed to fairly depict the threat. Furthermore, it’s reasonable to expect stronger gusts in squalls had Floyd tracked about another 50 or 60 miles farther east with more of the core circulation over the still relatively warm water east of the Delmarva region. It’s questionable whether we actually reached measured high wind criteria in many of our Connecticut Valley zones but did receive quite a number of reports of downed trees and limbs. The rain soaked leaves and ground may have been factors.
A Flood Potential Outlook (ESF) was issued 5 am Wednesday. This provided 36 to 48 hours notice. A Flood Watch (FFA) was issued 5 pm Wednesday. This provided 24 to 31 hours lead time (to the first small stream and river flooding). The Flood Warnings for small streams generally had a lead time of 2 to 5 hours, with as little as 30 minutes (Hampden County) to as much as 5 hours (Providence County). The Quinnipiac River at Southington, CT may have experienced the most severe flooding of any "gaged" smaller rivers in our county warning area. The Flood Warning that included (and specifically mentioned the Quinnipiac) was issued at 3:17 pm. This river reached flood stage at 5:40 pm. The River Flood Warnings generally had a lead time of 7 to 13 hours, with as little as 1 to 2 hours (UNVC3, GOFN3) and as much as 13 hours (HFDC3) and 23 hours (MDDC3).
Implement again emergency management coordination teleconferences and selected media briefings in future tropical cyclone threats. The Emergency management coordination teleconferences should be considered for major extratropical cyclone threats as well (especially when a combination of blizzard conditions and severe coastal flooding could occur).
Written media guidelines would help in the control of the media presence and communication of necessary ground rules.
Need to ensure that more frequent reports from USGS river gages are available during major flood threats (i.e., when gaged streams/rivers approach critical stages).
Need to make sure that one or two fresh people are available to proactively compose statements (Severe Weather Statements, Flood Statements, and/or Short Term Forecasts as appropriate) dealing with the short term threat on the leading edge of the most significant portion of the storm (be it wind, flood, or storm surge).
The cooperative nature of all staff who participated in this event (scheduled duty meteorologists/HMTs, augmented Hurricane Team members, Electronic Technicians) should be praised and considered as a role model. One staff member noted a close analogy with the teamwork displayed during the May 31, 1998 severe weather outbreak (which occurred coincidently with our Open House).
Staffing was adequate for this event, but there was little slack. The same level of staffing may be tenuous in a Category 3 or 4 direct hit. There is a clear risk of burnout with the more extreme events especially when one considers the intensity of pre-storm preparation, the pressures of post-storm assessment, and the potentially prolonged nature of riverine flooding.
The TDL real-time storm surge program on the internet (courtesy of Will Shafer) was a valuable tool for us. It helped us with the timing and magnitude of tropical storm force winds and enhanced our confidence in the storm surge projections (including which areas to retain the greatest emphasis on).
On a scale beyond this office, we believe the following points should receive attention:
greater consideration to HPC track projections north of 35 degrees north latitude in developing the official tropical cyclone track,
need to re-emphasize the importance of earlier coordination with NHC, and
a reaffirmation of tropical cyclone watch/warning philosophy.