a. Synoptic History
Erin formed from a tropical wave
that crossed from the coast of Africa to the tropical eastern Atlantic Ocean on 22 July 1995.
A large area of disturbed weather and two distinct low-level
circulation centers accompanied the wave. The circulation centers
were oriented from northwest to southeast and moved in tandem
toward the west-northwest over the following five days.
By the 27th, both circulations were generating deep convection
a few hundred miles to the northeast of the Leeward Islands. A day
later, meteorologists at the NHC
Tropical Analysis and Forecast Branch (TAFB, formerly TSAF as in figures) and the
NESDIS
Synoptic
Analysis Branch (SAB) assigned Dvorak technique
T-numbers of 1.5 to the trailing cloud cluster. These numbers increased to T-2.5,
potentially indicative of a tropical cyclone
with 35 knot (tropical storm force) winds by midday on the 30th.
In reality, although the cloud pattern was slowly consolidating and surface pressures were
falling ahead of the system in the Bahamas, development was
retarded by southwesterly vertical wind shear associated with an
upper-level low that was moving southwestward at 10-15 knots across
Florida. Reconnaissance aircraft data from the
U.S. Air Force Reserves (Hurricane Hunters)
on the 28th, 29th, and again during
midday on the 30th indicated that the system did not have a closed
circulation at low levels. Instead it was a very vigorous tropical
wave--winds speeds around 40 knots were reported
from ships in the northern part of the cloud pattern.
Because of the system's potential for development and its
close proximity to the Bahamas and Florida, a special nighttime
reconnaissance mission was requested by the NHC and flown by the
Hurricane Hunters
late on the 30th.
The first
"vortex message"
was transmitted to the NHC shortly after 0100 UTC on the 31st. From
that information it is estimated that the system became
Tropical Storm Erin
at 0000 UTC on the 31st (Fig. 1 [47K GIF],
Table 1).
The upper-level low near Florida affected Erin's movement and
development. Associated steering currents accelerated Erin from 5
to 15 knots and diverted the cyclone
around the northeast side of the low. The temporary and fairly subtle change of heading from
west-northwest to northwest might have been insignificant if Erin
had not been so close to land. Instead, the track of the
center
was deflected to a course that was over or near much of the Bahama
Island chain and then toward a landfall over east-central (rather
than southeast) Florida. As this occurred, enough shearing
persisted to permit only slow strengthening. On the evening of the
31st, Erin became a hurricane while centered near Rum Cay in the
Bahamas. A ragged-looking eye
appeared on satellite pictures on August 1st. Erin made landfall around 0600 UTC on the 2nd near
Vero Beach, Florida as a Category 1 hurricane on the
Saffir/Simpson Hurricane Scale,
with estimated maximum one-minute wind speeds of 75 knots.
Erin's track bent back to west-northwest while the cyclone
crossed the Florida peninsula during the morning and early
afternoon of the 2nd. The cyclone weakened to a tropical storm with
50-knot winds during that period, but remained well-organized.
Upon emerging into the eastern Gulf of Mexico, Erin reintensified
on a track that gradually swung back to northwestward at about 10
knots. Final landfall occurred near Pensacola, Florida during the
late morning of the 3rd. An eye had redeveloped but upper-level
outflow was not particularly impressive on satellite images. Erin
had around 85 knot winds (Category 2) in a small area of its
northeastern eyewall when that part of
the hurricane came ashore near Fort Walton Beach in the western Florida panhandle.
Erin weakened to a tropical storm in southeastern Mississippi
overnight on the 3rd/4th. It was a tropical depression when its
track shifted to the north on the 5th and the east on the 6th. The
depression merged with a frontal system over West Virginia on the 6th.
b. Meteorological Statistics
Erin's intensity was estimated from the data presented in
Figs. 2 and 3 (84K GIF) and Table 2.
Those figures show the curves of Erin's central pressure and maximum one-minute wind speed,
respectively, versus time, along with the observations on which they were based.
The figures contain relevant surface observations and intensity
estimates derived from analyses of satellite images performed by the
TAFB, SAB and the
Air Force Global Weather Central (AFGWC).
The aircraft data came from reconnaissance flights by the
U.S. Air Force Reserve unit
based at Keesler Air Force Base,
Mississippi.
Table 2 lists a selection of surface observations. The
highest wind at the surface was a gust to 128 knots reported in
association with a tornado at Providenciales in the Turks and Caicos Islands.
Several reports of hurricane force winds (WMO-standard 10-
minute average) were received from the Bahamas, including 68 and 70
knots during the passage of the northeast part of the eyewall over
Cat Island at 0200 UTC and 0400 UTC, respectively, on August 1st.
These 10-min winds are about 80 percent of the 86 knot
maximum 10-second 850 mb flight-level winds encountered by the reconnaissance
aircraft. Several amateur radio reports included gusts to around
90 knots in the Bahamas. The ship Tampa was in the northeastern
eyewall at 1200 UTC on the 1st when it reported 70 knot winds.
The basis for the 75-knot wind speed estimate along the
Florida east coast was a one-minute wind speed of 74.6 knots
recorded by a Florida Institute of Technology
anemometer which made one observation per hour at Sebastian Inlet. This wind
appears to coincide with the passage of one of Erin's strongest convective
cells at that time (0500 UTC), which was located in the
northwestern eyewall. While somewhat higher winds could have been
expected to occur offshore in the (normally stronger) northeastern
eyewall, Doppler radar data for that area suggests that the peak
winds (inbound toward Melbourne) at the lowest tilt angle were only
slightly stronger, around 85 knots. The maximum 850 mb
flight-level wind speed then was around 85 knots.
A wind speed of 85 knots is estimated at
1330 UTC on 3 August near Fort Walton Beach. This took place in a small area within
Erin's strongest sector, the northeastern eyewall, as it swept
across the shoreline. That estimate is based largely on
NWS Mobile
office Doppler wind data which showed inbound wind speeds exceeding
100 knots in a few volume samples centered at about 9,800 feet
above the coast from 1320 to 1400 UTC. The peak 850 mb flight-level wind speed
leading up to this time was 92 knots in the
northeastern eyewall near 1200 UTC, but subsequent excursions into
that part of the hurricane were precluded by the hurricane's close proximity to land.
Doppler velocities decreased by about 15 knots over the
following two hours and 75 knots is the estimated
maximum surface wind speed when the center of the eye came ashore around 1600 UTC.
Hence, the coastal region immediately west of Fort Walton Beach,
including Pensacola, experienced Category 1 conditions, though
gusts to near 100 knots likely occurred. The FAA
system of six anemometers at Pensacola Regional Airport (PNS) registered a
maximum 30-second wind speed of about 60 knots.
The highest wind speed measured at an official reporting station in the Florida
panhandle was an 88-knot gust at the Pensacola Naval Air Station
(NPA). Amateur radio operators relayed unofficial observations of
gusts near 95 knots to the NHC.
The hurricane's lowest pressure of 973 mb was reported by the
Hurricane Hunters
near 1330 UTC and again near 1600 UTC on the 3rd. The latter measurement
placed the center of Erin near the coast and in the southern part of the
eye as seen on surface radar.
The Melbourne National Weather Service Office
estimated that Erin generated a 2 to 4 foot storm tide
during the Florida east coast landfall. Storm tides averaged 1 to 2 feet along the west-
central Florida peninsula. According to the Melbourne office, up
to about 12 inches of rain fell southwest through northwest of
their site. Several small, brief tornadoes occurred over east-
central Florida well after Erin made landfall. One tornado caused
minor damage in Titusville. Another occurred near Lake Lizzie,
killing two horses. A couple of weak tornadoes were also reported
over northeast Florida and in the panhandle near Hurlburt Air Force Base.
Storm tides were estimated at 6 to 7 feet just west of Navarre
Beach and 3 to 4 feet along Pensacola Beach. Up to about 5 inches
of rain was reported from the panhandle.
c. Casualty and Damage Statistics
There were no deaths reported in the Bahamas or in Florida.
A total of six deaths occurred in the Atlantic and Gulf of Mexico
waters off Florida. All drowned. The 234-foot gambling and cruise
ship Club Royale sank in the Atlantic 90 miles east of Cape
Canaveral and three crew members are presumed dead. A 15-year old
surfer drowned in a rip current off Palm Beach County. A man and
daughter in an inflatable boat were swept from the Cape San Blas
area into the Gulf of Mexico where they presumably drowned.
All Bahamas islands from Mayaguana to Grand Bahama suffered
damage characterized by the Bahamas Department of Meteorology as
mostly minor. Some structural damage, sunken boats, crop loss and
flooding was reported. Losses known to date for Abaco, Grand
Bahama, Mayaguana, and Exuma total $400,000.
The American Insurance Services Group estimated $375 million
as the loss to insured property in the United States caused by Erin
($350 million in Florida, $20 million in Alabama, and $5 million in
Mississippi). Because the total loss is usually estimated by the
NHC to be up to about double the insured loss, the total U.S. loss
is tentatively estimated at $700 million.
Wind damage occurred over east-central and northeast Florida.
Thousands of homes and businesses suffered damage in Brevard
county. Less significant damage occurred in other counties in the
region. Freshwater flooding from rainfall occurred in the
Melbourne and Palm Bay areas and westward in some spots to the
Florida gulf coast. Beach erosion occurred along the central
Florida east coast, with damage mainly to boardwalks, beach
accessways and the dune system. Light to moderate beach erosion
was also reported northward to the Georgia border. Minor erosion
occurred along the west-central Florida coast.
The most significant structural damage for the final landfall
occurred on Pensacola Beach, Navarre Beach, around Mary Esther and
in northeast Pensacola. More than 2,000 homes were damaged there
and crop losses were reported. Some beach erosion was reported
west of Navarre Beach. Farther inland, about 100 homes were
damaged in Alabama. Widespread tree, power line and crop damage
extended inland.
d. Forecast and Warning Critique
Prior to Erin becoming a tropical cyclone and the NHC
initiating advisories, forecasts and warnings for the precursor
tropical wave/gale system were issued in High Seas Forecasts of the
NWS Tropical Prediction Center. During that period, NHC Tropical
Weather Outlooks indicated that the wave could soon become a
tropical depression
or tropical storm. Nevertheless, no lead time
(in the traditional sense of tropical storm or hurricane watches or
warnings) was available to the Turks and Caicos Islands and the
southeastern Bahamas. For the future, it might be worth
considering whether and how to issue such watches/warnings (and,
perhaps, track forecasts) for systems near land that could rapidly
become a threatening tropical cyclone.
Compared to the most recent 10-year averages, the NHC track
forecast errors for Erin were of about normal magnitude at 12 and
24 hours and much smaller than normal at 48 and 72 hours
(Table 3).
The first few NHC forecasts and the corresponding numerical
guidance did not accurately incorporate the effects of the upper-
level low near Florida on the path of Erin, and generally showed
the cyclone making landfall over southeast Florida. Intensity
forecasts were generally quite good, although in the first two
forecasts not enough strengthening was shown because the
deleterious effects of strong vertical wind shear were incorrectly forecast to persist.
Some people in the Pensacola area indicated that they did not have
sufficient notice of Erin's approach (see Table 4).
Table 5
shows lead times for that area of about 37, 25, and 21 hours for
the tropical storm watch,
tropical storm warning, and
hurricane warning, respectively.
Although there was no hurricane watch,
the other lead times are close to normal and, based on past experience,
should have been sufficient to accomplish the necessary tasks to
protect life and property. In fact, there were no lives lost in that area.
Rather than lack of lead time, it appears that the hurricane
warning was not taken seriously. Comments suggest one reason was
that the NHC forecasts did not show the cyclone center moving
directly over Pensacola. This is a critical misuse of NHC's
forecasts. The users of NHC advisory information are encouraged to
be familiar with potential track errors (Table 3)
and to understand that warning areas are designated with those uncertainties in mind.
In addition, cyclones which move along a course roughly parallel to
the coast pose an additional problem because even a slight sideways
jog of the hurricane or nonuniformity of the coastline can result
in landfall. In the limiting case, the center of the eye can
remain just offshore, but the entire coast could experience the
eyewall and its destructive hurricane conditions. In the future,
to ameliorate this kind of situation, the NHC will further their
efforts to deemphasize the precise forecast track in favor of the
threat implied by a watch or warning.
Apparently, a second problem was that Erin was "only" a
tropical storm when the hurricane warning was issued on the
afternoon of August 2. Residents were "surprised" to find out that
Erin had become a hurricane (as was forecast) when they awoke on
the morning of the 3rd. Hence, the public did not give enough
attention to the intensity forecast--but paid too close attention
to details of the track forecast. In both instances, the hurricane
warning should have been the overriding consideration driving public response.
Acknowledgments
Some information in this report was provided by the Forecast Office
of the Bahamas Department of Meteorology and by NWS offices in the watch and warning areas.
Table 4
Hurricane Erin watch and warning summary
Date/Time (UTC) | Action | Location |
31/0330 |
Tropical Storm Warning issued |
Central and Southeast Bahamas |
Tropical Storm Watch issued | Northwest Bahamas |
Tropical Storm Watch issued |
Florida east coast Sebastian Inlet southward through Florida
Keys including Dry Tortugas and Florida Bay and Florida west coast Venice southward |
31/0900 | Tropical Storm Warning issued |
Northwest Bahamas |
31/1500 | Hurricane Warning issued |
Sebastian Inlet southward through Florida Keys including
Dry Tortugas and Florida Bay and all of the Bahamas |
Hurricane Watch issued |
Florida west coast from Venice southward to Everglades City and for Lake Okeechobee |
01/0200 | Hurricane Warning discontinued |
Southeastern Bahamas |
01/0300 |
Tropical Storm Warning issued |
North of Sebastian Inlet to New Smyrna Beach |
Hurricane Watch issued |
Florida west coast Bayport southward and for Lake Okeechobee |
01/0900 | Hurricane Warning issued |
Lake Okeechobee |
01/1500 | Hurricane Warning issued |
Sebastian Inlet to New Symrna Beach |
Tropical Storm Warning issued |
New Smyrna Beach to St Augustine |
Hurricane Warning discontinued | Central Bahamas |
Hurricane Watch issued | North of Bayport to Cedar Key |
01/1800 | Hurricane Warning discontinued |
Florida Keys from Key Largo southward |
01/2100 |
Tropical Storm Warning issued |
Florida west coast from Fort Myers to Suwanee River |
Tropical Storm Watch issued |
Suwanee River to Apalachicola |
Hurricane Watch discontinued | Fort Myers southward |
02/0000 | Hurricane Warning discontinued |
Florida east coast south of Hallandale |
02/0100 | Hurricane Warning discontinued |
New Providence and Andros Islands |
02/0300 |
Tropical Storm Warning issued |
Suwanee River to Apalachicola |
Tropical Storm Watch issued | Apalachicola to Pensacola |
02/0700 | Hurricane Warning discontinued |
Florida east coast southward from Deerfield Beach |
02/0900 |
Hurricane Warning discontinued |
Remaining areas |
Tropical Storm Warning issued |
Florida east coast from Fernandina Beach southward to Jupiter Inlet including
Lake Okeechobee |
Tropical Storm Warning issued |
Florida Gulf Coast Longboat Key to Apalachicola |
02/1000 | Hurricane Warning discontinued |
Remainder of Bahamas |
02/1100 | Tropical Storm Warning discontinued |
Lake Okeechobee |
02/1500 |
Tropical Storm Warning issued |
Apalachicola to Pensacola |
Tropical Storm Watch issued |
Pensacola to mouth of Pearl River |
02/1900 |
Hurricane Warning issued |
Suwannee River to mouth of Pearl River |
Hurricane Watch issued |
South of mouth of Pearl River to mouth of Mississippi River including city of New Orleans |
02/2300 | Hurricane Warning issued |
Mouth of Pearl River to mouth of Mississippi River including city of New Orleans |
03/0100 | Tropical Storm Warning discontinued |
Long Boat Key to mouth of Suwannee River |
03/0300 |
Hurricane Warning issued |
Mouth of Mississippi River to Grand Isle Louisiana |
Hurricane Watch issued |
Grand Isle to Morgan City Louisiana |
03/0500 | Hurricane Warning issued |
Grand Isle to Morgan City |
03/0900 | Hurricane Warning discontinued |
East of Apalachicola |
03/1900 | Hurricane Warning discontinued |
West of mouth of Pearl River and for New Orleans |
03/2100 | Hurricane Warning discontinued |
Remaining areas |