1) Synoptic history

A tropical wave moved from Africa to the tropical Atlantic Ocean on 5 August. On 18 August, continuity and soundings from Merida, Mexico, place this weak tropical wave in the vicinity of the Yucatan Peninsula. Also, a weak surface low-formed over the peninsula. This low originated from a thunderstorm complex associated with a westward moving upper-level cyclonic circulation.

Later on 18 August the surface low moved over the Bay of Campeche. Early morning visible satellite imagery showed a low-level cloud circulation center and, a few hours later, a Hurricane Hunter reconnaissance mission confirmed the existence of a closed circulation. With some deep convection and banding present, the system became a tropical depression at 1800 UTC on 18 August over the Bay of Campeche.

The depression did not strengthen right away due to vertical wind shear caused by an upper-level trough over the extreme western Gulf of Mexico. However, the trough moved away and the depression became Tropical Storm Bret late on 19 August while beginning to move slowly northward. Bret steadily strengthened. On the morning of 22 August, it reached 125 kt (a category 4 hurricane on the Saffir–Simpson hurricane scale) just offshore from Brownsville, Texas. Figure 2 is a visible satellite image of Bret on the morning of 22 August. Responding to the presence of a weak midtropospheric ridge over the northwest Gulf of Mexico and to a mid-tropospheric cyclonic circulation over the Rio Grande valley, Bret turned northwestward and its forward speed slowed from near 10 kt to about 5 kt.

Bret's center crossed the Texas coast midway between Brownsville and Corpus Christi and near the middle of Padre Island at 0000 UTC 23 August. By the time of landfall, it had weakened to a category 3 hurricane with 100-kt winds and a pressure of 951 mb. After moving inland, Bret's movement became more westward. Bret continued to weaken while it moved across south Texas and into the high terrain of north-central Mexico, where it dissipated on 25 August.

2) Meteorological statistics

The maximum 1-min surface wind speed of 125 kt early on 22 August is based on a reconnaissance aircraft Global Positioning System (GPS) dropwindsonde vertical wind speed profile (Hock and Franklin 1999). Profiles from Bret's eyewall show that winds reached near 150 kt within 300 m above the surface and were about 125 kt near the surface.

Bret's pressure dropped 35 mb to 944 mb in the 24 h ending at 1200 UTC 22 August and dropped 21 mb during the 6 h ending at 0000 UTC of the same day. This episode of intensification coincides with the hurricane's track over a maximum in the sea surface temperature (SST) field over the west-central Gulf of Mexico. Analyses from Applied Physics Laboratory, The Johns Hopkins University, shows 31°C SST values along Bret's track during the intensification. In addition, values of the upper oceanic heat content (described by Shay et al. 2000) under portions of the hurricane's circulation in the western Gulf of Mexico were high.

Bret was a small hurricane. At its peak, hurricane force winds were confined to a radius of 30–40 n mi from the center in the north semicircle and only 10–20 n mi in the south semicircle. Thus only a small segment of the Texas coast was affected by the core of the hurricane. Kennedy County received most of the hurricane force winds, which were as high as 100 kt over a small portion of the coast of Padre Island. With the center moving inland over a sparsely populated area, few surface reports were available substantiating strong winds. Table 2 lists a selection of surface observations. The highest reported sustained wind is 63 kt at Rincon del San Jose on Padre Island. That instrument failed at 2230 UTC on 22 August just before the center passed nearby. The Port Aransas Coastal Marine Automated Network (C-MAN) station reported maximum sustained winds of 41 kt as the center of the hurricane passed about 60 n mi to the south.

Bret was slow moving and radar estimates suggest maximum storm total rainfall of over 760 mm in Kennedy County. None of the observed rainfall totals in Table 2 come close to that value. Aransas Pass is north of the area of peak rainfall and reported a storm total of 320 mm. The heavy rains accompanying the weakening tropical cyclone caused notable river flooding in the Rio Grande Valley. The Rio Grande River at Laredo and the Aransas River near Skidmore and at Oso Creek crested slightly above flood stage. A 24-h rainfall total of over 350 mm was reported from the Mexican state of Nuevo Leon. The state of Tamaulipas is believed to have received similar amounts.

The Sea, Lake, and Overland Surges from Hurricanes (SLOSH) storm surge model suggested that a narrow region along central and north Padre Island experienced a storm surge of 2.5–3 m. A report from Port Mansfield Pass suggests that 1–1.5 m of water inundated this coastal location. Several cuts were observed in the dunes surrounding Padre Island. The largest of these, near mile marker 50 near the eye's passage, was mistaken as the Mansfield Pass by aircrews inspecting the damage. Other substantial beach erosion was reported near Port Mansfield.

In Aransas County around 2145 UTC 22 August, a tornado reportedly uprooted trees, destroyed a recreational vehicle, a barn, and a shed. Other reports indicate that a tornado touched down in Kingsville around 2245 UTC on 22 August and a tornado was reported in Alice, but the time was not known. Neither led to reported damage.

3) Casualty and damage statistics

Despite Bret's intensity, damage was fairly light. Much of this is due to its landfall over a sparsely populated region in south Texas and the hurricane's small size. The nearest population centers to the south and north of landfall were Brownsville and Corpus Christi. These cities are about 100 n mi apart and were both spared the brunt of the hurricane's core. Brownsville's maximum reported sustained wind was only 29 kt and Corpus Christi's maximum was only 39 kt (see Table 2). There were no deaths.

Property insurance damage claims total $30 million as reported by the Property Claims Services Division of the Insurance Services Office. A standard practice of multiplying by a factor of 2.0 gives an estimated damage total of $60 million. This is a conservative multiplier, based on a sample of past landfall damage estimates.

1) Synoptic history

Dennis originated from a tropical wave that crossed the coast of Africa on 17 August. The system moved westward with little significant weather until 21 August, when its shower activity increased a few hundred miles northeast of the Leeward Islands. A low-level circulation developed over the next 2 days and the convection became better organized. A reconnaissance flight on 23 August failed to find a surface circulation, but aircraft data indicated a circulation was present aloft at 850 mb. Later, surface observations showed a closed circulation, and it is estimated that Tropical Depression 5 formed early on 24 August about 190 n mi east of Turks Island. Reconnaissance data and ship reports indicated further intensification, and the depression became Tropical Storm Dennis later that day, and a hurricane early on 26 August.

Dennis initially moved at 9–12 kt, but slowed to an erratic 3 kt on 25 August when steering currents weakened due to the passage of a midlatitude trough to the north. Once Dennis reached hurricane strength, it began a more steady northwestward motion near the eastern Bahamas. This motion continued into 28 August.

Westerly shear persisted, preventing additional significant strengthening until late on 27 August. After the shear decreased, Dennis reached a peak intensity of 90 kt on 28 August and maintained that speed until early on 30 August. Even at peak intensity, Dennis was not a tightly wound hurricane. The eye was 30–40 n mi wide and, on several center fixes, the Hurricane Hunter aircraft did not report an eye. The radius of maximum winds was, at times, as large as 70–85 n mi.

A second midlatitude trough caused Dennis to turn gradually northward on 28–29 August, and accelerate east-northeastward on 30–31 August. This turn kept the center about 60 mi south of the North Carolina coast. The east-northeast motion continued until the trough passed Dennis on 31 August. Then steering currents collapsed and the cyclone slowed to an erratic drift about 110 n mi east of Cape Hatteras, North Carolina. This erratic drift lasted into 2 September.

During this time, Dennis was affected by the cold front associated with the midlatitude trough. A combination of vertical shear and entrainment of cool dry air into the circulation decreased the convection and caused weakening. Dennis weakened to a tropical storm on 1 September, and on 1 and 2 September may have, in fact, been as much subtropical or extratropical as tropical. Despite the lack of convection, surface observations indicate maximum sustained winds were near 45 kt on 2 September, some of which were due to the pressure gradient between Dennis and a strong surface ridge north of the front. Winds in excess of 34 kt or greater were observed as far north as New Jersey.

A large high pressure cell over the eastern United States forced Dennis southward late on 2 September. This motion over warmer water probably helped initiate a deep convective burst on the next day. Then, Dennis turned northwestward toward the North Carolina coast as the ridge moved east into the Atlantic. This motion continued on 4 September along with reintensification. Dennis was just below hurricane strength when it made landfall at the Cape Lookout National Seashore just east of Harkers Island, North Carolina, on 4 September. Dennis continued inland and weakened to a depression over central North Carolina. Even in dissipation, Dennis continued to move erratically along a zigzag northward course. Dennis became extratropical on 7 September and was absorbed into a larger extratropical low on 9 September.

2) Meteorological statistics

Dennis's path brought it near the eastern Bahamas on 27–28 August. The only official report of tropical storm force winds in the Bahamas was from a C-MAN station at Settlement Point, Grand Bahama, which reported 34-kt sustained winds with gusts to 46 kt on 29 August. A selection of surface observations is listed in Table 3. Also, unofficial reports from amateur radio operators indicated sustained winds of up to 55 kt with gusts to 65 kt in the Abaco Islands.

Aircraft reconnaissance data suggest that Dennis maintained 90-kt surface wind speeds for 36 h from 28 to 30 August. Just before landfall on 4 September, an aircraft measured 71-kt wind at 850 mb and a 984-mb pressure. Vertical wind profiles from several hundred GPS dropwindsondes (Franklin et al. 2000) indicate that the average ratio between surface and 700-mb flight-level winds is about 0.90 and the ratio between surface and 850-mb winds is about 0.80. Thus, Dennis is estimated to be a 60-kt tropical storm at landfall.

Dennis tracked parallel to the Florida and Georgia coasts, with tropical storm force winds near the coasts. A 41-kt gust was reported from the St. Augustine, Florida, C-MAN station. On 29 August, the core of Dennis passed just east of National Oceanic and Atmospheric Administration (NOAA) buoy 41010, which reported a 57-kt sustained wind with a gust to 72 kt and a minimum pressure of 980.2 mb.

Dennis's first pass near the mid-Atlantic coast on 30 August caused sustained tropical storm force winds with gusts to hurricane force in coastal North Carolina and gusts to tropical storm force in coastal South Carolina from Charleston northward. The maximum reported sustained winds were 53 kt with gusts to 77 kt at Oregon Inlet. Gusts to 96 kt at Wrightsville Beach and 85 kt at Hatteras Village suggest that sustained hurricane force winds may have occurred along the coasts of New Hanover and Dare Counties. Sustained hurricane force winds of 81 kt with gusts to 97 kt were measured at the Frying Pan Shoals C-MAN station (145-ft elevation) with a minimum pressure of 977.2 mb.

The landfall of Dennis on 4 September produced tropical storm force winds over portions of eastern North Carolina and coastal southeastern Virginia. The large circulation of Dennis also affected shipping over the western Atlantic. On 30 August, the ship Zim U.S.A. reported a sustained wind speed of 65 kt while located about 120 n mi east-southeast of the center of the then 90-kt tropical cyclone.

Few detailed observations of storm surge are available from areas affected by Dennis. Storm tides up to 1.5 m above normal were reported along much of the North Carolina coast on both 30 August and 4 September. Areas along the Neuse River reported tides 3 m above normal tide level on 30 August, while areas along the Pamlico River reported similar heights on 4 September. Portions of the South Carolina and southeastern Virginia coasts experienced tides about 1 m above normal. Amateur radio reports from the Bahamas indicate similar tide heights as the eye passed over the Abacos. Because Dennis meandered off the North Carolina coast for several days, the above normal tides were prolonged. This led to extensive beach erosion along portions of the North Carolina and southeastern Virginia coasts.

Dennis affected the mid-Atlantic states twice within a week, and other weather systems affected the region during the same period. The maximum reported rainfall was 486 mm at Ocracoke, North Carolina, with 150–250 mm reported elsewhere over portions of eastern North Carolina. Amounts of 75–150 mm occurred elsewhere over eastern North Carolina, extreme eastern South Carolina, and over portions of southeastern Virginia. Rainfalls were generally 75 mm or less elsewhere over eastern South Carolina and less than 25 mm in Florida and Georgia.

Dennis significantly contributed to saturating the ground of the mid-Atlantic coastal states and this set the stage for the severe flooding caused by Hurricane Floyd 2 weeks later. Rainfall data from the Bahamas indicates a maximum total of 102 mm at Cat Island. Heavier amounts likely occurred on Eleuthera and in the Abaco group near the eye of Dennis.

One tornado was reported with Dennis on 4 September. This F2 tornado in Hampton, Virginia, caused 15 injuries, 6 of them serious.

3) Casualty and damage statistics

Four deaths reported in Florida were related to high surf spawned by the hurricane. No deaths are known due to winds, rains, storm tides, or tornadoes associated with Dennis.

The Property Claims Services Division of the Insurance Services Office reports insured losses due to Dennis totaled $60 million in North Carolina and Virginia. As in Bret, a two-to-one ratio was applied to the insured losses, giving $120 million total property damage. Press reports indicate that agricultural losses in North Carolina and Virginia were $37 million, making a total estimated damage from Dennis of $157 million.

1) Synoptic history

Floyd can be traced to a tropical wave that emerged from western Africa on 2 September. The wave proceeded westward across the eastern tropical Atlantic at about 15 kt, with little change in structure. On 6 September, a curved band of deep convection began to develop and Tropical Depression 8 formed on 7 September about 1000 mi east of the Lesser Antilles.

A deep-layer ridge prevailed to the north of the cyclone such that the associated steering current was west-northwestward at 12–15 kt for 2 days. The cloud pattern became sufficiently well organized for the system to become Tropical Storm Floyd early on 8 September and about 750 n mi east of the Leeward Islands. Floyd slowly strengthened and became a hurricane on 10 September about 200 n mi east-northeast of the northern Leeward Islands.

As Floyd was nearing hurricane status, a midtropospheric trough near 60°–65°W caused a slowing of the forward speed, and a turn toward the northwest. This motion continued until 11 September, keeping the hurricane well to the northeast of the Leeward Islands. On 11 September, Floyd's upper-level outflow was disrupted over the southern semicircle by the above-mentioned trough and by an anticyclone over the eastern Caribbean Sea. Consequently, after strengthening to 95 kt early on 11 September, the hurricane weakened to 85 kt by 12 September. Early on 12 September, rising mid- to upper-tropospheric heights to the north of Floyd forced the hurricane to turn toward the west. Maximum sustained winds increased to 135 kt and the central pressure fell about 40 mb by early on 13 September. Figure 4 is a visible satellite image of Floyd about 8 h prior to the maximum wind speed reaching 135 kt.

A possible contributor to the significant strengthening of Floyd was the presence of enhanced upper oceanic heat content along its track. Analyses from the Physical Oceanography Division, Atlantic Oceanographic and Meteorological Laboratory (AOML), NOAA, showed relatively high values of heat content just to the east of the Bahamas a day or two before Floyd passed through the area.

Floyd moved toward the central Bahamas until late on 13 September when it turned west-northwestward. The eye passed just 20–30 n mi northeast and north of San Salvador and Cat Islands on the night of 13 September. Floyd's eyewall passed over central and northern Eleuthera on the morning of 14 September, and after turning toward the northwest, Floyd struck Abaco Island on the afternoon of 14 September. By the time the hurricane hit Abaco, it had weakened somewhat from its peak, but Floyd was still a borderline category 3/4 hurricane.

A mid- to upper-tropospheric trough over the eastern United States eroded the subtropical ridge over the extreme western Atlantic and Floyd continued to turn gradually to the right. The center of the hurricane paralleled the central Florida coast, passing about 95 n mi east of Cape Canaveral on 15 September. By later that afternoon, Floyd was east of the Florida–Georgia border and headed northward toward the Carolinas.

Although there was a fluctuation in intensity, caused by an eyewall replacement event discussed in the next section, overall, the intensity of Floyd diminished from 13 to 15 September. Two large-scale factors probably contributed to the weakening: the entrainment of drier air at low levels from the northwest, and increasing south-southwesterly vertical shear.

After turning toward the north-northeast with its forward speed increasing to near 15 kt, Hurricane Floyd made landfall near Cape Fear, North Carolina, early on 16 September as a category 2 hurricane, with estimated maximum winds near 90 kt. Floyd was losing its eyewall structure as it made landfall. Continuing to accelerate north-northeastward, Floyd's center passed over extreme eastern North Carolina on the morning of 16 September and over the greater Norfolk, Virginia, area a little later. Floyd then weakened to a tropical storm and moved swiftly along the coasts of the Delmarva Peninsula and New Jersey later on. By the time it reached Long Island on 17 September, its forward speed had increased to near 29 kt. The system decelerated as it moved into New England.

By late on 16 September and early on 17 September, Floyd merged with a frontal zone along the Atlantic seaboard. Floyd took the form of a frontal low and became extratropical by the time it reached the coast of Maine on the morning of 17 September. The cyclone turned toward the northeast and then east-northeast, moving over the coast of New Brunswick late on 17 September, Prince Edward Island early on 18 September, and Newfoundland late on 18 September and early on 19 September. Floyd's extratropical remnant merged with a large extratropical low over the North Atlantic and was no longer a distinct entity by later on 19 September.

2) Meteorological statistics

The peak intensity of Floyd, 135 kt, is based upon an estimate of 90% of the highest aircraft reconnaissance flight-level (700 mb) winds of 149 kt on 13 September. The minimum dropsonde-measured central pressure was 921 mb on the same day.

Floyd is estimated to have been a 90-kt hurricane at landfall in North Carolina, based on a 10-m anemometer measurement of sustained winds of 83 kt on 16 September near Topsail Beach, North Carolina. There were also unofficial reports of peak wind gusts to 120 kt (at eight stories elevation) at Wrightsville Beach and 104 kt at the Wilmington Emergency Operations Center.

Table 4 lists a selection of surface observations from land stations and data buoys. Floyd's eye passed over NOAA data buoy 41010, located about 105 n mi east-northeast of Cape Canaveral on 15 September. That buoy reported maximum 8-min averaged winds of 72 kt at an anemometer height of 5 m. At least two factors would imply a higher value for the 1-min, 10-m wind speed: 1) extrapolating upward from 5- to 10-m elevation, and 2) the presence of waves above 15 m in height. The best track intensity of Floyd when it passed over the buoy was 100 kt, as indicated by dropsonde and aircraft flight level wind data. The center of the hurricane passed about 25 n mi west of the Frying Pan Shoals C-MAN station located about 30 n mi southeast of Cape Fear on 16 September. This station reported winds sustained at 86 kt for a 20-min period at an anemometer height of 44 m.

On 13 September, just after Floyd reached its maximum strength, a sequence of microwave satellite images (not shown), as well as aircraft radar, gave evidence of a concentric eyewall replacement cycle. Willoughby and Black (1996) describe Hurricane Andrew's concentric eyewalls. First, during the deepening phase, there was a dominant inner eyewall with an eye diameter of 20–25 n mi. Later on, after peak intensity was reached, there was an indication of concentric eyewalls. Figure 5a shows a radar image from a NOAA WP-3D research aircraft late on 13 September that shows the inner and outer eyewalls. Then, there was an apparent eyewall replacement, as suggested by measured vertical wind profiles and radar imagery from a NOAA WP-3D research aircraft as well as by the microwave images. The inner eyewall was dissipating while Floyd was centered near Eleuthera. This corresponded to a weakening of the hurricane to near 105 kt. The outer convective ring became the new eyewall by the time Floyd was centered over Abaco with a 50 n mi eye. This is shown in Fig. 5b, an aircraft radar image at 2028 UTC on 14 September. Afterward, the new eye failed to contract significantly, while Floyd restrengthened just slightly as it reached Abaco. After the disintegration of the inner eyewall, the large-scale environment, as noted previously, became less favorable. Consequently, after leaving the Bahamas, Floyd never regained its former intensity.

Heavy rainfall preceded Floyd over the mid-Atlantic states due to a preexisting frontal zone and the associated overrunning. Hence, even though the tropical cyclone moved fairly quickly, precipitation amounts were very large. Myrtle Beach, South Carolina, recorded 408 mm. At Wilmington, North Carolina, the storm total of 484 mm included a 24-h record of 383 mm. Newport News, Virginia, also reported over 400 mm and amounts over 300 mm were reported in Maryland, Delaware, and New Jersey. Rainfall totals in the 200–250-mm range extended across southern New York and into New England. Storm surge values up to 3.1 m were reported along the North Carolina coast.

A number of tornadoes were sighted in eastern North Carolina. There was a confirmed tornado in Bertie County and another in Perquimans County. The latter tornado destroyed two houses and damaged three or four others. At least 10 tornadoes were reported by spotters in the Newport/Morehead City County Warning area, and these apparently caused some structural damage. Four tornadoes or funnel clouds were seen in the Wilmington area, but no damage was apparent.

3) Casualty and damage statistics

There were 57 deaths directly attributable to Floyd, 56 in the United States and 1 on Grand Bahama Island. The death toll by state is as follows: North Carolina, 35; Pennsylvania, 6; New Jersey, 6; Virginia, 3; Delaware, 2; New York, 2; Connecticut, 1; and Vermont, 1. Most of these deaths were due to drowning in freshwater flooding. Floyd was the deadliest hurricane in the United States since Agnes of 1972.

In the United States, the Property Claims Services Division of the Insurance Services Office reports that insured losses due to Floyd totaled $1.325 billion. In comparison to most hurricane landfalls, Floyd caused an inordinately large amount of freshwater flood damage, which probably alters the two to one damage ratio. The damage estimate is $4.5 billion.

4) Warnings

It is notable that hurricane warnings were issued for the U.S. east coast from south Florida northward to Plymouth, Massachusetts. The last time such an event occurred was during Hurricane Donna of 1960. According to information provided to the Federal Emergency Management Agency, over two million people were evacuated for Floyd in the United States. This is reported to be the largest evacuation in U.S. history.

1) Synoptic history

A tropical wave moved from west Africa to the Atlantic on 10 September, accompanied by convective banding and some evidence of a low-level cloud circulation. The developing tropical cyclone's track was south of the Atlantic subtropical ridge and toward the west-northwest from 10–16 September. The best track starts on 11 September, south of the Cape Verde Islands, when the system became a tropical depression. It reached tropical storm strength on 12 September and became a hurricane on 13 September. Gert continued to strengthen to 130 kt by 16 September. Figure 6 is a visible satellite image of Gert near the time of 130-kt wind speed.

There was a weakness in the ridge ahead of Gert and along Floyd's track. While intensifying, Gert responded to this weakness and slowly turned north and then north-northeastward during 16–21 September. Its center came within about 300 n mi to the northeast of the Leeward Islands during this turn. With some fluctuations in intensity, winds remained near 115 kt through 19 September, after which weakening commenced. Gert weakened to 60 kt by 23 September, having moved to near southeastern Newfoundland. It then became extratropical and merged with another extratropical low pressure system.

2) Meteorological statistics

Dvorak satellite wind speed estimates place the time of Gert's peak intensity of 130 kt near 0000 UTC on 16 September. Nearly 24 h of reconnaissance data, starting about 0800 UTC on 16 September, indicated that the wind speed weakened to 120 kt and then reached 125 kt early on 17 September.

The center of Gert passed about 115 n mi east of Bermuda on 21 September. The maximum 10-min wind speed reported from the airport was 39 kt with a peak gust to 64 kt. A gust to 76 kt was reported from an exposed coastal location. There were three instances of 1-min mean winds speeds between 66 and 70 kt from a harbor location during the morning of 21 September. A rainfall total of 13 mm was reported from Bermuda. Gale conditions were also experienced on the Avalon Peninsula on southeast Newfoundland.

3) Casualties and damages

Bermuda experience some coastal erosion along its east and south sides.

Two persons drowned on 20 September, when a large and unexpected wave swept them out to sea. They had been standing at the water's edge at Schoodic Point in Acadia National Park, Maine. The local marine patrol described the wave as a “rogue wave.” This event is believed to be related to large swells generated by Gert, even though the hurricane was located more than 1000 n mi south-southeast of Maine at the time. There were also news reports of 8-m-high waves sweeping over the coast of the southeast tip of Newfoundland. Three persons were swept into the water while trying to secure their boat. All were rescued.

1) Synoptic history

The tropical wave that produced Harvey moved off the west coast of Africa late on 4 September. From 4 to 14 September, the wave's trek across the tropical Atlantic into the eastern Caribbean Sea was uneventful, due, in part, to the disruptive effects of the upper-level outflow from Hurricane Floyd. By 16 September, convective activity began to increase over the western Caribbean Sea and a broad area of low pressure formed. Upper-air observations from the northwest Caribbean showed an associated midlevel circulation near the Cayman Islands. The broad area of low pressure drifted northwestward into the Gulf of Mexico over the next few days. Early on 19 September, satellite imagery showed increasing deep convection near and east of the circulation center and a tropical depression formed about 350 n mi west-southwest of St. Petersburg, Florida.

Upper-level outflow around the depression improved. By early on 20 September, surface winds were near 40 kt and the system became Tropical Storm Harvey, about 300 n mi west-southwest of St. Petersburg. Over the next 24 h, Harvey's central pressure dropped a modest 7 mb and the storm is estimated to have reached a peak intensity of 50 kt on 20 and 21 September. Satellite images showed that the system did not become better organized during this period because of westerly vertical wind shear. Data from a NOAA Gulfstream IV high-altitude aircraft on the afternoon of 20 September showed 25–35-kt westerly upper-level winds over the cyclone. This contributed to the displacement of Harvey's center to the northwest edge of the deep convection and restricted the outflow over the western semicircle. Satellite imagery also suggested that dry air was being entrained into the circulation from the northwest.

After turning from a north to a northeast heading on the afternoon of 19 September, Harvey moved slowly eastward on 20 September. By early on 21 September, the system turned southeastward and the forward speed increased to 9 kt in response to midlevel northwesterly flow. Harvey's track bent back to the east by midmorning of 21 September and the cyclone accelerated in advance of a midlatitude trough approaching from the west. Harvey made landfall near Everglades City, Florida, around 1700 UTC on 21 September with maximum sustained winds of 50 kt and a minimum central pressure of 999 mb. It crossed Florida and moved offshore over the Atlantic near Fort Lauderdale. Later that afternoon, Harvey was absorbed by a developing extratropical cyclone centered just off the coast of South Carolina. This extratropical cyclone then moved over portions of the Canadian Maritime Provinces, causing heavy rainfall. Several weather stations along the Fundy coast of New Brunswick received over 200 mm, and up to 302 mm fell in northern Nova Scotia, making for the worst flooding in 30 years. Between 50 and 100 mm of rain fell in Labrador.

2) Meteorological statistics

Harvey's maximum sustained winds of 50 kt are based on a reconnaissance aircraft wind speed of 58 kt at a flight level of 457 m. The ship Liberty Sun reported 47 kt on 20 September. The Molasses Reef C-MAN recorded a sustained wind of 47 kt, with a gust to 59 kt, and Fowey Rocks Light reported 45-kt sustained winds with a gust to 51 kt.

The highest official sustained surface wind observed over land was 32 kt at the Key West Airport on 21 September. The Turkey Point Nuclear Power Plant recorded a 10-min, 46-kt wind, at a 9-m elevation, while the highest gust, 48 kt, was recorded at Tenraw in the Everglades (25.6°N, 81.9°W). The lowest pressure observed in south Florida was 999.4 mb at the Fort Lauderdale Airport, also on 21 September.

The highest storm-total rainfall recorded in Harvey was 255 mm at the Naples Conservancy in Collier County. Naples Lakewood measured about 250 mm of rain as did Naples/Collier County Emergency Management at their operations center. Substantial street flooding was reported in portions of Collier and Lee Counties. Rainfall totals of 140 and 171 mm were observed at Immokalee and Everglades City, respectively. Storm total rainfall across Miami-Dade, Broward, and Palm Beach Counties ranged from 19 mm (West Palm Beach Airport) to 72 mm (Coral Springs).

The maximum recorded storm surge was 0.7 m at Fort Myers with estimates from 0.5 to 1.0 m elsewhere in Charlotte County. Storm surge up to 0.6 m occurred elsewhere in southwest Florida and in the Florida Keys. Tidal flooding was reported in Everglades City including the county airport where a portion of the runway was flooded, resulting in the closure of the airport. Minor coastal flooding was reported along the south-facing portions of the Florida Keys and the west-facing shores of Florida Bay. Sections of Highway A1A in the Keys were closed due to the flooding. The combined effect of wave action and the storm surge resulted in minor beach erosion in Sarasota County, along the south-facing shores of the Keys, and in the backcountry of Everglades National Park.

There were two tornadoes. An F0 touched down briefly in Collier County near Paradise Point taking the roof off one house, while the other touched down in Dade County with no reported damage.

The Canadian Hurricane Centre reported that some areas of southeastern New Brunswick and northern Nova Scotia received over 250 mm of rain over a 30-h period from Harvey's extratropical remnants. Oxford, Nova Scotia, reported 302 mm and 50–100 mm fell in Labrador.

3) Casualty and damage statistics

There are no known casualties. Property damage estimates supplied by the Property Claims Services Division of the Insurance Services Office, indicate that Harvey caused about $7.5 million in insured losses in southern Florida. The total estimated damage from Harvey is $15 million.

1) Synoptic history

A broad area of low pressure accompanied by disorganized clouds and thunderstorms prevailed over the southwestern Caribbean Sea from 8 to 10 October. This system did not show signs of tropical cyclone development until a tropical wave reached the western Caribbean Sea on 11 October. On 12 October a U.S. Air Force Reserve reconnaissance plane found an incipient low-level circulation and a broad low pressure area of 1006 mb just to the northeast of the coast of Honduras. However, the circulation was too ill-defined to be classified as a tropical depression. Satellite imagery during the night showed that the thunderstorms increased and banding features and upper-level outflow became quite distinct. Surface and upper-air data from Grand Cayman, and satellite intensity estimates, indicate that a tropical depression formed in the northwestern Caribbean Sea early on 13 October and reached tropical storm status 6 h later. Irene moved on a general northward track, then slowed before curving to the north-northeast just to the southwest of the Isle of Youth, Cuba, where it made its first landfall on 14 October. Radars from Cuba and Key West showed the center of Irene moving on a north-northeast track over western Cuba. The center of the tropical cyclone then crossed the Havana and Ciudad Havana provinces late on 14 October. Irene reached hurricane status over the Florida Straits and the calm of the center moved over Key West near midday on 15 October. Most of the hurricane force winds were confined to the east of Irene's center over the lower to middle Florida Keys. Irene made its fourth landfall near Cape Sable, Florida, and then moved across southeast Florida bringing tropical storm conditions and 250–500 mm of rain. During the period while Irene was crossing Florida, sustained hurricane force winds appeared to be limited to squalls offshore of the east coast of Florida, as reported by reconnaissance aircraft and indicated by surface observations and Doppler radar.

Irene moved back over water in northern Palm Beach County near Jupiter a little after 0000 UTC on 16 October. It retained hurricane strength and moved on a general northward track paralleling the Florida east coast heading for the Carolinas. An upper-level trough, sweeping eastward across the eastern United States, forced Irene on a fast northeastward track. The core of Irene missed the mainland Carolinas but produced very heavy rains inland. It then brushed North Carolina's Outer Banks before moving out to sea.

During a 12-h period beginning on the evening of 18 October, Irene's central pressure dropped from 978 to 958 mb and the winds increased from 70 to 95 kt. This may be attributable to a combination of a trough interaction and the tropical cyclone moving over the warm waters of the Gulf Stream. Thereafter, Irene continued to accelerate and, on 19 October, was absorbed by a much larger extratropical low near Newfoundland. The resultant system became an intense extratropical storm over the North Atlantic.

2) Meteorological statistics

Irene was monitored by three Cuban weather radars located at Havana, Isle of Youth, and Pinar del Rio. Operationally, Irene was upgraded to hurricane status just before landfall over the Isle of Youth. However, numerous observations from Cuba and a postanalysis of satellite imagery indicate that Irene was most likely a tropical storm while crossing Cuba. Peak wind gusts reported from Cuba were 68 kt at the Havana forecast office.

A selection of surface observations is given in Table 5. The highest sustained wind observed was 69 kt at Big Pine Key in the Florida Keys. Observations from the South Florida Water Management District indicate that gusts to hurricane force were experienced near Lake Okeechobee. Based on the Miami Weather Surveillance Radar-1988 Doppler signatures, these gusts were likely produced by small-scale mesocyclone-induced downbursts. Four weak tornadoes occurred in Broward and Palm Beach counties.

There was serious urban flooding in southeast Florida with a maximum rainfall amount reported in south Florida of 443 mm in South Miami. Coastal sections of North Carolina and Virginia also recorded amounts in the 200–300-mm range, with lesser amounts in Georgia and South Carolina.

The rapid intensification of Irene on 18 October off the North Carolina coast was documented by a reconnaissance plane investigating the hurricane during that period. Their report indicated a very small closed eyewall of about 3 n mi in diameter and 114 kt winds at a flight level of 850 mb.

3) Casualty and damage statistics

There were eight deaths indirectly caused by Irene. Five people were electrocuted (four in Broward County and one in Miami-Dade County) when they walked into puddles that concealed downed power lines. There were three drownings when vehicles were driven into canals obscured by flooding (one in Palm Beach County and two in Broward County). In addition, there were three persons injured by tornadoes in Broward County.

Irene caused considerable damage due to flooding in south Florida. In some residential areas, flooding lasted for a week, displacing several hundred people and isolating thousands more. The total losses (agricultural and property) were estimated near $600 million mostly in Dade, Broward, and Palm Beach Counties. Additional losses to near $200 million occurred in the rest of Florida. An estimated 700 000 customers lost electricity.

1) Synoptic history

Jose originated from a tropical wave that moved off the west coast of Africa on 8 October. The wave moved slowly westward across the tropical Atlantic for several days. By 15 October, when the system was located about midway between Africa and the Lesser Antilles, shower activity became better organized. The disturbance became a tropical depression on 17 October, about 550 n mi east of the southern Windward Islands. There was well-defined upper-tropospheric outflow over the depression. Moving west-northwestward, the depression strengthened into Tropical Storm Jose on 18 October.

Jose turned northward when a mid- to upper-tropospheric trough produced a weakness in the Atlantic subtropical ridge. Jose became a hurricane late on 19 October about 100 n mi east of the Leeward Islands. As it neared these islands, Jose reached its peak intensity of 85 kt at 1200 UTC 20 October. Turning back to a west-northwest heading, Hurricane Jose struck the northern Leeward Islands, passing over Antigua around midday on 20 October. The eye then moved near St. Barthelemy and St. Martin early on 21 October.

Southwesterly vertical shear caused weakening, as Jose moved over the northern Leeward Islands. Jose weakened to a tropical storm by the time it reached Tortola in the British Virgin Islands, on the morning of 21 October. A little later on 21 October, the cyclone turned back toward the northwest and the center passed about 40 n mi northeast of the eastern tip of Puerto Rico. With a large mid- to upper-tropospheric trough positioned over the western North Atlantic, Jose turned northward, then north-northeastward on 22 October. The storm continued north-northeastward at a faster forward speed on 23 October. Its structure was still disrupted by southwesterly shear. Early on 24 October, however, microwave satellite data indicated that the deep convection again was over the low-level center and Jose regained hurricane strength. The hurricane passed about 250 n mi east of Bermuda around midday on 24 October, and the forward speed increased markedly. The rejuvenation of the tropical cyclone was short lived and Jose weakened back to a tropical storm by 25 October. It continued to accelerate into the North Atlantic, losing tropical characteristics later that day. It then merged with a larger midlatitude low and associated front.

2) Meteorological statistics

The estimate of a peak intensity of 85 kt is based on a wind of that speed measured at 10 m by a reconnaissance aircraft GPS dropwindsonde. The highest measured wind speed from a surface reporting station was 70 kt with a gust to 89 kt at Antigua. Sustained winds of 65 kt were measured at St. Maarten. The highest sustained wind reported from the Virgin Islands was 52 kt with a gust to 60 kt at St. John. Higher wind speeds likely occurred over portions of the British Virgin Islands. In Puerto Rico, winds were mostly below tropical storm force. There was an unofficial measurement of sustained winds of 30–39 kt with a gust to 48 kt from Costa Azul Beach in Luquillo.

Very heavy rain fell well after the passage of the center over the northern Leeward Islands, in association with bands of thunderstorms extending east and southeast of the center. Rainfall totals were as high as 350 mm over portions of the islands.

3) Casualty and damage statistics

Two deaths were caused by Jose, one in Antigua and one in St. Maarten. Damage in Antigua was characterized as “minor.” In St. Maarten, the heavy rains caused extensive flooding and mud slides that damaged roads and homes, especially in low-lying areas. United States (Puerto Rico and the U.S. Virgin Islands) damage totals were less than $5 million.

1) Synoptic history

Satellite imagery shows that the remnants of a cold front moved slowly southward across the western Caribbean Sea beginning on 22 October. A broad area of low pressure gradually formed over much of the Caribbean during this time and cloudiness and thunderstorms became concentrated over the southwestern Caribbean Sea on 26 October. On the next day, low-level cloud lines began to show a closed circulation just north of Panama. On 28 October, a reconnaissance aircraft reported a well-defined low-level closed circulation about 150 n mi east of Bluefields, Nicaragua, indicating that a tropical depression had formed.

Katrina was a tropical storm for about 6 h late on 30 October, while making landfall on the coast of Nicaragua just south of Puerto Cabezas. For the rest of its 4 days, Katrina was a tropical depression that moved on a mostly northwestward track across Nicaragua and Honduras, back over the waters of the northwest Caribbean, and then across northern Belize and the Yucatan Peninsula. The depression dissipated on 1 November just north of the Yucatan Peninsula when it was absorbed by a cold front.

2) Meteorological statistics

The system was monitored by reconnaissance aircraft on 28 and 29 October while located in the southwestern Caribbean Sea. The basis for naming Katrina a tropical storm was a 43-kt flight-level wind observation on 29 October. Satellite-based rainfall estimates suggest that 250–375 mm of rain may have occurred over portions of Nicaragua and Honduras, with lesser amounts for the Yucatan Peninsula. San Andres, Colombia, an island about 100 n mi east of Nicaragua, reported 91 mm of rain in 6 h on 28 October.

3) Casualties and damages

It is possible that the rainfall described above caused some flash flooding over mountainous terrain over Central America. No reports of damages or casualties were received.

1) Synoptic history

A broad area of low pressure was first identified in the southwest Caribbean Sea early on 8 November. Although thunderstorm activity remained poorly organized, locally heavy rains and strong gusty winds occurred over the northwestern Caribbean Sea and adjacent Central America and Mexico for several days. Early on 13 November, satellite imagery showed that the system was gradually becoming better organized. Based on 30-kt surface winds and a 1003-mb central pressure from aircraft reconnaissance that afternoon, a tropical depression formed about 150 n mi south of the Cayman Islands. By mid morning of 14 November, the overall organization of the depression was improving and the depression became a tropical storm on the morning of 14 November.

Later that day, a reconnaissance mission found maximum flight-level winds of 84 kt and a central pressure of 988 mb. Lenny became a hurricane about 150 mi southwest of Kingston, Jamaica. Satellite imagery showed a banding-type eye, 15–20 n mi in diameter.

Lenny's extended west-to-east motion across the Caribbean is unprecedented. For its first 48 h, Lenny moved on an east to east-southeastward course. This motion was induced by the flow around the southern portion of a deep-layer trough located over the western Atlantic. Several short-wave troughs helped to amplify the mean western Atlantic trough through the period, increasing the westerly steering flow. From 1200 UTC 15 November to 1800 UTC 16 November, Lenny's mean forward speed was 14 kt toward the east. The unusual nature of this steering pattern is illustrated by an anomaly chart of November mean 500-mb heights. This chart (not shown) shows a very large negative anomaly (minimum height of −60 m) centered a few hundred nautical miles north of Puerto Rico.

During a 24-h period beginning midday on 16 November, the central pressure dropped 34 mb. Lenny's maximum sustained surface winds of 135 kt and minimum central pressure (933 mb) occurred at 1800 UTC 17 November when the hurricane was centered about 20 n mi south of St. Croix in the U.S. Virgin Islands. However, the 135-kt winds were confined to the southeast quadrant of the hurricane over water. Figure 7 is a visible satellite image of Lenny just before reaching 135-kt wind speed.

Lenny moved into a col between two midlevel ridges late on 17 November and its forward motion slowed. The hurricane drifted east-northeastward before turning southeastward early on 19 November. Despite what appeared in satellite imagery as a favorable upper-level outflow pattern, Lenny weakened, perhaps due to oceanic upwelling. Lenny was gradually weakening when its center passed very slowly over St. Maarten during the afternoon of 18 November, over Anguilla later that evening, and over St. Barthelemy early on 19 November. Lenny weakened to a tropical storm on 19 November just south of St. Barthelemy. It made its final landfall in Antigua late on 19 November.

Lenny continued to move southeastward and the forward motion increased to near 8 kt. The motion turned again toward the northeast as Lenny weakened to a depression on 21 November. The depression turned back to the east early on 21 November and dissipated on the morning of 23 November about 600 n mi east of the Leeward Islands.

2) Meteorological statistics

The estimated peak intensity of Lenny, 135 kt, is based on 90% of the 149-kt flight-level (700 mb) wind speed reported at 1702 UTC 17 November. A GPS dropwindsonde measured 183 kt at 891 mb, a record dropwindsonde wind speed in a hurricane. The lowest pressure recorded by the aircraft was 934 mb at 1929 UTC 17 November. However, the minimum central pressure of 933 mb was assigned based on the lower 700-mb height from the 1702 UTC fix.

Table 6 lists a selection of surface observations from land stations. The highest official sustained surface wind speed observed over land was 73 kt at the St. Maarten Meteorological Office in Philipsburg on 19 November. However, the meteorological antenna fell down during the strongest winds. Hamilton Airport on St. Croix recorded sustained 1-min winds of 60 kt, with a gust to 80 kt. The lowest official pressure observed in the northern Leeward Islands was 972.1 mb at the St. Maarten Meteorological Office.

The maximum sustained surface wind speed reported via amateur radio operators was 81 kt on St. John in the U.S. Virgin Islands on 17 November. During the peak of the hurricane, Saba recorded a gust of 145 kt (792-m elevation) before the anemometer blew away. The highest measured gust of 97 kt on St. Croix coincides with Lenny's closest approach. The minimum central pressure at St. Croix, 980 mb, was also recorded at this time.

Lenny's slow drift across the northern Leeward Islands resulted in very large rainfall amounts over a 36-h period. For many locations, heavy rain over several days was the primary impact of Lenny. Portions of French St. Martin, St. Barthelemy, and Guadeloupe received record rains. The highest rainfall total in Dutch St. Maarten was 700 mm, recorded at the meteorological office in Philipsburg. Even higher amounts are likely to have occurred in French St. Martin. V. C. Bird International Airport in Antigua reported 465 mm.

Rainfall totals up to 200 mm occurred across the U.S. and British Virgin Islands and resulted in widespread flooding. The largest official rainfall total in Puerto Rico was 239 mm on Magueyes Island, Lajas.

Fredericksted in St. Croix was inundated by an estimated 4–6-m storm surge. The maximum reported storm tide was 0.9 m at the NOAA National Ocean Service gauge in Lime Tree Bay on St. Croix.

Lenny's approach from the west produced unprecedented wave and storm surge impact on westward-facing coasts and harbors. The meteorological service in Dutch St. Maarten reported that southern and western coastal areas were significantly impacted by wave action. The Météo-France meteorological station in Gustavia on St. Barthelemy estimated waves up to 4.9 m in the harbor on 17 November. A platform near La Desirade, just east of Grande-Terre, Guadeloupe, recorded a significant wave height of 3.0 m at 2300 UTC 20 November and estimates range up to 4 m in the harbor.

Lenny generated large waves and swells that propagated across much of the southern and eastern Caribbean, affecting the Guajira Peninsula of Colombia, Aruba, Bonaire, Curacao, and much of the remainder of the Leeward and Windward Islands.

3) Casualty and damage statistics

Lenny was responsible for 17 direct deaths: 3 in Dutch St. Maarten, 2 in Colombia, 5 in Guadeloupe, 1 in Martinique, and 6 offshore. Two of the deaths in St. Maarten were caused by flying debris while the other was the result of a collapsed roadway. The remainder of the onshore fatalities, based on media reports, are presumed to be due to freshwater flooding. Two of the offshore deaths occurred when the sailing yacht Vdar was lost somewhere in the southern Caribbean Sea.

Lenny's heavy rain, and wave and storm surge produced considerable damage on many of the islands in the northeast Caribbean. St. Croix sustained moderate damage. Many boats were washed ashore along the north coast of the island. Several boats sank In Christiansted Harbor and structures in the south and east portions of the island suffered roof damage. Lenny also impacted the agricultural areas of the island. Both the Dutch and French portions of St. Maarten/Martin were severely impacted with many buildings damaged and boats damaged or lost. In St. Lucia, at least 70 homes were reported damaged. In Saba, the airport tower and several other buildings were severely damaged. Guadeloupe sustained a large amount of damage along the west coast due to the wave action, and inland due to heavy rains. In Grenada, 10 homes were destroyed and 21 small boats were lost. There were also reports of damage in St. Vincent and the Grenadines, and Montserrat.

Insured losses of $165 million have been reported from Puerto Rico and the U.S. Virgin Islands. Using a factor of 2.0, the total U.S. damages from Lenny is estimated at $330 million.