Abstract

The 2006 Atlantic hurricane season is summarized and the year’s tropical cyclones are described. A verification of National Hurricane Center official forecasts during 2006 is also presented. Ten cyclones attained tropical storm intensity in 2006. Of these, five became hurricanes and two became “major” hurricanes. Overall activity was near the long-term mean, but below the active levels of recent seasons. For the first time since 2001, no hurricanes made landfall in the United States. Elsewhere in the basin, hurricane-force winds were experienced in Bermuda (from Florence) and in the Azores (from Gordon). Official track forecast errors were smaller in 2006 than during the previous 5-yr period (by roughly 15%–20% out to 72 h), establishing new all-time lows at forecast projections through 72 h. Since 1990, 24–72-h official track forecast errors have been reduced by roughly 50%.

1. Introduction

Overall activity during the 2006 Atlantic season (Fig. 1; Table 1) was near average. There were ten tropical storms, of which five became hurricanes and two became major hurricanes [maximum 1-min winds of greater than 96 kt (1 kt = 0.5144 m s−1), corresponding to category 3 or greater on the Saffir–Simpson Hurricane Scale (Saffir 1973; Simpson 1974)]. These numbers are very close to the 40-yr (1966–2005) averages of 11.1, 6.2, and 2.3, respectively, but represent activity considerably below that of recent seasons. Hurricane activity was relatively compressed, occurring only during a 37-day period between 27 August and 2 October. The last tropical cyclone activity of any kind also occurred on 2 October. Since the beginning of the geostationary satellite era (1966), only the 1983 and 1993 seasons have ended earlier (both on 30 September). In terms of “accumulated cyclone energy” [ACE; the sum of the squares of the maximum wind speed at 6-h intervals for (sub)tropical storms and hurricanes], activity this year was 90% of the long-term (1966–2005) mean, and the lowest observed since the 2002 season.

Fig. 1.

Tracks of tropical storms and hurricanes in the Atlantic basin in 2006.

Fig. 1.

Tracks of tropical storms and hurricanes in the Atlantic basin in 2006.

Table 1.

2006 Atlantic hurricane season statistics.

2006 Atlantic hurricane season statistics.
2006 Atlantic hurricane season statistics.

The near-average activity observed in 2006 appears to have resulted from two opposing factors. Atlantic waters were extremely warm in 2006; indeed, sea surface temperatures (SSTs) in the tropical Atlantic Ocean and Caribbean Sea between 10° and 20°N during the peak of hurricane season were the second highest since 1871 (only 2005 was warmer). While these warm SSTs and other atmospheric factors (e.g., relatively low shear) were conducive to an active hurricane season, the unexpected development of an El Niño event during the late summer appeared to suppress activity. Although the El Niño did not strongly affect the vertical wind shear throughout the basin, it likely caused or reinforced a strong area of anomalous subsidence over the western part of the Atlantic basin (Fig. 2). This sinking led to a warmer, drier, and more stable atmosphere that likely limited tropical cyclogenesis in this area (Bell et al. 2007).

Fig. 2.

Anomaly from the long-term (1968–96) mean of vertical velocity (Pa s−1) for August–October 2006. Areas of anomalous rising motion are shaded. [Data provided by the NOAA/Cooperative Institute for Research in Environmental Sciences Climate Diagnostics Center, Boulder, CO (http://www.cdc.noaa.gov/), based on the National Centers for Environmental Prediction–National Center for Atmospheric Research reanalysis project (Kistler et al. 2001)]

Fig. 2.

Anomaly from the long-term (1968–96) mean of vertical velocity (Pa s−1) for August–October 2006. Areas of anomalous rising motion are shaded. [Data provided by the NOAA/Cooperative Institute for Research in Environmental Sciences Climate Diagnostics Center, Boulder, CO (http://www.cdc.noaa.gov/), based on the National Centers for Environmental Prediction–National Center for Atmospheric Research reanalysis project (Kistler et al. 2001)]

For the first time since the 2001 season, and after twelve hurricanes struck the United States during 2004–2005,1 none did so in 2006. Three tropical storms (Alberto, Beryl, and Ernesto) made landfall in the United States, although only Ernesto had a significant impact, with damage estimated to have been near $500 million. No tropical cyclone–related direct deaths occurred in the United States during 2006. More significant impacts, however, were felt elsewhere in the basin. Alberto affected western Cuba, and Ernesto produced heavy rainfall in portions of Cuba, Haiti, and the Dominican Republic, with five deaths resulting from Ernesto’s rains in Haiti. Florence brought hurricane conditions to Bermuda, and after losing tropical characteristics also brought hurricane-force winds to portions of Newfoundland. Gordon was the first hurricane to affect the Azores since 1991.

The lack of hurricane landfalls on the North American continent in 2006 is attributed to large-scale steering patterns that were markedly different from those present during 2004 and 2005 (Fig. 3). September 2006 was characterized by a strong midlevel trough over the eastern United States, a pattern conducive to steering the four hurricanes that formed during the month out to sea, near or east of Bermuda. This pattern was reminiscent of the dominant steering flow that prevailed during 1995–2003. In contrast, the mean September steering patterns for 2004 and 2005 featured high pressure over the western Atlantic and eastern United States that kept many systems on more westward tracks.

Fig. 3.

Mean 500-mb heights (m) for September (top) 2004–05 and (bottom) 2006. Source of data same as in Fig. 2.

Fig. 3.

Mean 500-mb heights (m) for September (top) 2004–05 and (bottom) 2006. Source of data same as in Fig. 2.

2. Storm and hurricane summaries

The individual cyclone summaries that follow are based on National Hurricane Center (NHC) poststorm meteorological analyses of a wide variety of (often contradictory) data described below. These analyses result in the creation of a “best-track” database for each storm, consisting of 6-hourly representative estimates of the cyclone’s center position, maximum sustained (1-min average) surface (10 m) wind, minimum sea level pressure, and maximum extent of 34-, 50-, and 64-kt winds in each of four quadrants around the center. The tracks and basic statistics for the season’s storms and hurricanes are given in Fig. 1 and Table 1, respectively (also see online at http://www.nhc.noaa.gov/pastall.shtml).2 The life cycle of each cyclone (as indicated by the dates given in Table 1) is defined to include the tropical or subtropical depression stage, but does not include remnant low or extratropical stages.

For storms east of 55°W, or those not threatening land, the primary (and sometimes sole) source of information is geostationary weather satellite imagery, interpreted using the Dvorak (1984) or Hebert–Poteat (Hebert and Poteat 1975) techniques. Ships and buoys occasionally provide important in situ observations on these cyclones. For systems posing a threat to land, in situ observations are also generally available from aircraft reconnaissance flights conducted by the 53rd Weather Reconnaissance Squadron (“Hurricane Hunters”) of the Air Force Reserve Command (AFRC), and by the National Oceanic and Atmospheric Administration (NOAA) Aircraft Operations Center (AOC). During reconnaissance flights, minimum sea level pressures are either measured by dropsondes released at the circulation center or extrapolated hydrostatically from flight level. Low-level winds in the eyewall or maximum wind band are often measured directly using global positioning system (GPS) dropwindsondes (Hock and Franklin 1999), but more frequently are estimated from flight-level winds using empirical relationships derived from a 3-yr sample of GPS dropwindsonde data (Franklin et al. 2003). During NOAA reconnaissance missions, surface winds can be estimated remotely using the Stepped-Frequency Microwave Radiometer (SFMR) instrument (Uhlhorn and Black 2003). For storms close to land, weather radars, buoys, and conventional land-based surface and upper-air observations supplement the satellite and reconnaissance data. In key forecast situations, the kinematic and thermodynamic structure of the storm environment is obtained from dropsondes released during operational “synoptic surveillance” flights of NOAA’s Gulfstream IV jet aircraft (Aberson and Franklin 1999).

Several satellite-based technologies play an important role in the analysis of tropical weather systems. Foremost of these is multichannel passive microwave imagery [e.g., from the Tropical Rainfall Measuring Mission (TRMM) satellite], which over the past decade has provided radarlike depictions of systems’ convective structure (Hawkins et al. 2001) and is of great help in assessing system location and organization. The SeaWinds scatterometer onboard the Quick Scatterometer (QuikSCAT) satellite (Tsai et al. 2000) provides surface winds over large oceanic swaths. While the QuikSCAT generally does not have the horizontal resolution to determine a hurricane’s maximum winds, it can often be used to estimate the intensity of weaker systems and to determine the extent of tropical storm–force winds. In addition, it can be helpful in resolving whether an incipient tropical cyclone has acquired a closed surface circulation. Finally, information on the thermal structure of cyclone cores is provided by the Advanced Microwave Sounder Unit (AMSU; Velden and Brueske 1999). Intensity estimates derived from such data in some cases can be superior to Dvorak classifications (Herndon and Velden 2004).

A number of organizations have developed Web sites that have proven to be extremely helpful for tropical cyclone forecasting and postanalysis. These sites include the Naval Research Laboratory (NRL) Monterey Marine Meteorology Division Tropical Cyclone Page (http://www.nrlmry.navy.mil/tc_pages/tc_home.html), with its comprehensive suite of microwave products, the cyclone phase diagnostics page of the Florida State University (http://moe.met.fsu.edu/cyclonephase/), which is frequently consulted to help categorize systems as tropical, subtropical, or nontropical, and the tropical cyclone page of the University of Wisconsin—Madison Cooperative Institute for Meteorological Satellite Studies (http://cimss.ssec.wisc.edu/tropic/tropic.html), which contains a variety of useful satellite-based synoptic analyses.

In the cyclone summaries below, U.S. property damage estimates have been generally estimated by doubling the insured losses reported by the Property Claim Services of the Insurance Services Office. The reader is cautioned, however, that great uncertainty exists in determining the cost of the damage caused by tropical cyclones when they make landfall. Descriptions of the type and scope of damage are taken from a variety of sources, including local government officials, media reports, and local National Weather Service (NWS) Weather Forecast Offices (WFOs) in the affected areas. Tornado counts are based on reports provided by the WFOs and/or the Storm Prediction Center. Hard copies of these various reports are archived at the NHC.

Although specific dates and times in these summaries are given in coordinated universal time (UTC), local time is implied whenever general expressions such as “afternoon,” “midday,” etc. are used.

a. Tropical Storm Alberto, 10–14 June

Alberto formed from an area of disturbed weather that persisted for several days over Central America and the northwestern Caribbean Sea. Convection in the area increased on 8 June with the arrival of a westward-moving tropical wave, and pressures began to fall between the Yucatan Peninsula and Cuba where the convection was most concentrated. By 0600 UTC 10 June, the disturbance had acquired a surface circulation and sufficiently organized convection to be classified as a tropical depression. The depression, initially centered about 120 n mi south of the western tip of Cuba, moved slowly northwestward, becoming a tropical storm at 0000 UTC 11 June about 60 n mi northeast of the northeastern tip of the Yucatan Peninsula. Strong southwesterly shear in the Gulf of Mexico kept the main area of convection displaced from the circulation center, and there was little change in Alberto’s strength until the following day. As the storm interacted with the deep warm waters of the Loop Current early on 12 June, convection increased. By late morning, a circulation center had reformed near the convection about 60 n mi to the northeast of its previous position (Fig. 4), and Alberto’s maximum winds increased abruptly to 60 kt. The system’s minimum pressure of 995 mb was reached a few hours later, near 0000 UTC 13 June, when Alberto was centered about 100 n mi south of Apalachicola, Florida.

Fig. 4.

Geostationary Operational Environmental Satellite-12 (GOES-12) visible satellite image of Tropical Storm Alberto at 1445 UTC 12 Jun 2006, near the time of the cyclone’s maximum intensity (image courtesy of University of Wisconsin—Madison).

Fig. 4.

Geostationary Operational Environmental Satellite-12 (GOES-12) visible satellite image of Tropical Storm Alberto at 1445 UTC 12 Jun 2006, near the time of the cyclone’s maximum intensity (image courtesy of University of Wisconsin—Madison).

Alberto’s increase in organization was short lived. While the cyclone moved northeastward in the eastern Gulf of Mexico, its cloud pattern became elongated and the convection began to diminish. Dry air overtook the cyclone early on 13 June and Alberto began to weaken as it moved toward Florida’s Big Bend region. Alberto made landfall near Adams Beach at about 1630 UTC 13 June with maximum winds of 40 kt. Alberto continued to weaken after it moved inland, becoming a depression early on 14 June and losing tropical characteristics later that day over South Carolina. The extratropical remnant of Alberto moved across North Carolina and into the Atlantic, where it became a powerful extratropical storm just south of Nova Scotia. The cyclone slowly weakened as it continued across the North Atlantic Ocean to the British Isles, where it was finally absorbed by a frontal system on 19 June.

The strongest wind reported in Alberto was a flight-level observation of 81 kt from a NOAA WP-3D aircraft, at an altitude of 5000 ft at 1657 UTC 12 June, although this observation was considered to represent a transient convective feature. Selected surface observations from land stations and data buoys are given in Table 2. Alberto produced torrential rains in western Cuba, where there was one report of 445 mm and several others in excess of 300 mm. The highest reported rainfall in the United States was 179.8 mm near Tarpon Springs, Florida. There were several reports of sustained tropical storm–force winds in Florida, the highest of which (42 kt) occurred at the Old Port Tampa National Ocean Service (NOS) site. There were seven tornadoes in South Carolina, most of them rated as F0.

Table 2.

Selected surface observations for Tropical Storm Alberto, 10–14 Jun.

Selected surface observations for Tropical Storm Alberto, 10–14 Jun.
Selected surface observations for Tropical Storm Alberto, 10–14 Jun.

Property losses associated with Alberto were relatively minor. In Florida, storm surge flooding near Homossassa in Citrus County put 1 m of water into a restaurant and damaged 20 homes. Numerous houses suffered flood damage in Levy County, and a few downed trees in Tallahassee caused power outages. There were no “direct” deaths associated with Alberto as a tropical cyclone. After the system became extratropical, however, an 8-yr-old boy was pulled into a drainage system and drowned near Raleigh, North Carolina. There was also a press report of four sailors missing about 200 n mi south of Nova Scotia when Alberto passed by during its extratropical storm stage.

b. Unnamed tropical storm, 17–18 July

As part of its routine postseason review, NHC occasionally identifies a previously undesignated tropical or subtropical cyclone based on new data or meteorological interpretation. This year’s review reclassified a short-lived system as a tropical storm.

The tropical cyclone originated along the tail end of a cold front that moved offshore of the northeastern United States late on 13 July and stalled over the western Atlantic Ocean. An extratropical low formed on 16 July along the decaying front when an upper trough approached from the west. The upper trough weakened, and the surface low moved slowly northeastward over warm 27°–28°C waters. Buoy and satellite data suggest that the front associated with the low dissipated late on 16 July, although the low lacked organized convection until early the next day, when a large burst of convection formed near the center. It is estimated that a tropical depression formed at 0600 UTC 17 July about 210 n mi southeast of Nantucket Island, Massachusetts.

Satellite intensity estimates indicate that the depression attained tropical storm strength 6 h later while accelerating toward the northeast. A large curved band of convection formed in the northern portion of the storm, with other banding features becoming more prominent throughout the day, and the system reached a peak intensity of about 45 kt near 1800 UTC 17 July. Shortly thereafter, the cyclone crossed the north wall of the Gulf Stream and encountered much cooler waters. Convection diminished significantly overnight, and by 1200 UTC 18 July the system became a nonconvective remnant low. The remnant low moved across Newfoundland later on 18 July, then turned toward the eastnortheast and dissipated on 19 July over the open waters of the North Atlantic Ocean. There were no reports of damage or casualties associated with this system.

While the low was considered to be a frontal cyclone operationally, postevent analysis of the system revealed no significant frontal boundaries. In particular, Canadian buoy 44137 indicated little or no surface temperature gradient across the cyclone as it passed close by, and a QuikSCAT pass at 2234 UTC 17 July showed no evidence of frontal structures in the wind field. The QuikSCAT data also showed that the system had a relatively small radius of maximum winds (about 30 n mi), and the buoy data indicated a steep drop in pressure and sharp increase in wind that are typical of a tropical cyclone. Figure 5 shows the system with a large burst of convection and the overall satellite appearance of a sheared tropical cyclone. AMSU observations, which were unavailable in real time, indicate that the system had a significant tropospheric warm core, and these observations are consistent with the Florida State University phase analysis (Hart 2003) of the system. Based on this evidence, the system is considered to have been a tropical cyclone.

Fig. 5.

GOES-12 visible satellite image of the unnamed tropical storm at 1245 UTC 17 Jul 2006 (image courtesy of University of Wisconsin—Madison).

Fig. 5.

GOES-12 visible satellite image of the unnamed tropical storm at 1245 UTC 17 Jul 2006 (image courtesy of University of Wisconsin—Madison).

c. Tropical Storm Beryl, 18–21 July

The genesis of Beryl can be traced to the same frontal system that spawned the unnamed tropical storm. The front stalled off the coast of North Carolina on 16 July, and over the next day or two the temperature gradient across the front gradually dissipated over sea surface temperatures of 26°–27°C. By 1200 UTC 18 July, a low pressure center formed near the southwestern end of the remnant trough, about 250 n mi east-southeast of Wilmington, North Carolina, with sufficient organized deep convection to designate the system as a tropical depression. Six hours later, as convective banding features became more prominent over the eastern semicircle of the circulation, the cyclone became a tropical storm.

From 18–20 July, the tropical storm moved mainly northward, paralleling the coast of the mid-Atlantic states along the western periphery of a subtropical high pressure area. Vertical shear was not strong and upper-level anticyclonic outflow became established over the system, allowing Beryl to strengthen to its maximum intensity of 50 kt by 1800 UTC 19 July. Beryl maintained this intensity for about a day, but by 1800 UTC 20 July a slow weakening trend began as Beryl passed over cooler waters. Steering winds ahead of a midtropospheric trough moving through the Great Lakes caused the storm to turn toward the north-northeast and northeast, with a gradual increase in forward speed. The center of Beryl passed over Nantucket around 0645 UTC 21 July, at which time the cyclone’s maximum sustained winds were estimated to be 45 kt. The cyclone continued to accelerate northeastward, and Beryl lost tropical characteristics shortly after 1200 UTC 21 July. The system crossed Nova Scotia and moved over Newfoundland, where it merged with another extratropical cyclone after 1200 UTC 22 July.

Selected surface observations associated with Beryl from land stations and data buoys are given in Table 3. Wind gusts to tropical storm force were reported on Nantucket. There were no reports of casualties or damage.

Table 3.

Selected surface observations for Tropical Storm Beryl, 18–21 Jul.

Selected surface observations for Tropical Storm Beryl, 18–21 Jul.
Selected surface observations for Tropical Storm Beryl, 18–21 Jul.

d. Tropical Storm Chris, 1–4 August

The tropical wave that spawned Chris moved across the west coast of Africa early on 26 July, and maintained vigorous convection for the next couple of days as it moved westward at 15–20 kt. By late on 28 July, the convection weakened considerably and became less organized as the system moved into a region of large-scale subsidence and dry air. The wave continued to propagate westward at a slightly slower forward speed, and convection began to increase on 30 July several hundred miles east of the southern Leeward Islands. The convection became better organized over the next day or so, and a closed surface wind circulation developed late on 31 July. It is estimated that a tropical depression formed at 0000 UTC 1 August about 205 n mi east-southeast of Barbuda.

The depression moved west-northwestward at about 10 kt, and despite moderate upper-level westerly wind shear, became a tropical storm around 0600 UTC 1 August. Chris reached its peak intensity of 55 kt early on 2 August about 55 n mi east-northeast of Anguilla in the northern Leeward Islands, but began to weaken the following day when northwesterly shear increased and dry air became entrained into the circulation. Chris passed about 100 n mi north of Puerto Rico and Hispaniola and weakened to a tropical depression at 1800 UTC 3 August, when the cyclone was centered about 195 n mi east-southeast of Grand Turk Island. Strong westerly shear continued, and the convection associated with Chris diminished as the system moved westward toward the southeastern Bahamas. Chris degenerated into a nonconvective remnant low pressure system by 0600 UTC 4 August about 80 n mi east-southeast of Grand Turk Island. The remnant circulation skirted the northern coast of Cuba before it dissipated on 6 August near Havana.

Direct effects from Chris in the Lesser and Greater Antilles were mainly confined to brief periods of heavy rainfall and some localized floods. News reports indicate rainfall from the storm caused the Fajardo River in Puerto Rico to overflow its banks. Rainfall reached up to 50 mm across portions of Hispaniola, the Turks and Caicos, the Bahamas, and eastern Cuba. In some of the mountainous areas of Hispaniola, there were unofficial reports of near 100 mm of rain. Although the center of Chris passed about 100 n mi to the north of the Dominican Republic, the cyclone still caused heavy rains and significant floods in and around Santo Domingo, where many roads were impassable due to the combined effects of high water and mudslides. Overall, however, damage was minor and no casualties were reported.

e. Tropical Storm Debby, 21–26 August

Debby formed from a vigorous tropical wave that moved across the west coast of Africa on 20 August. Almost immediately after moving offshore, the wave developed convective banding and a broad closed circulation. The first Dvorak classification was made at 1200 UTC 21 August, and by 1800 UTC a tropical depression had formed about 225 n mi south-southeast of Praia in the Cape Verde Islands.

The depression initially moved west-northwestward to the south of the subtropical ridge. Around 1200 UTC 22 August, the center of the cyclone passed about 100 n mi to the southwest of the southernmost Cape Verde Islands, bringing thunderstorms and gusty winds to the southern islands of Fogo and Brava. The depression strengthened as it moved away from the islands, becoming a tropical storm around 0000 UTC 23 August. By 1200 UTC that day, Debby’s sustained winds reached 45 kt, but there was little or no change in strength for the next two days while the cyclone moved between west-northwestward and northwestward at 15–20 kt over the open waters of the eastern Atlantic. Intensification during this period appeared to be limited by a dry and stable air mass surrounding the cyclone, along with marginal sea surface temperatures. On 25 August, southerly shear increased in association with an upper-level trough, displacing the deep convection to the north of the center. Debby began to weaken and became a depression around 0600 UTC 26 August. By 1200 UTC that day there was no deep convection within about 150 n mi of the center, and Debby degenerated to a remnant low about 1225 n mi east-southeast of Bermuda. The remnant low turned gradually northward ahead of an approaching frontal trough, generating intermittent convection before dissipating just east of the trough early on 28 August.

f. Hurricane Ernesto, 24 August–2 September

Ernesto, with landfalls as a tropical storm in Cuba, Florida, and North Carolina, was fleetingly a hurricane over the central Caribbean Sea. The storm also produced torrential rainfall and floods in portions of Hispaniola, while gale-force winds and heavy rains associated with the extratropical remnants of Ernesto affected portions of Virginia, Maryland, Delaware, and New Jersey. Ernesto was directly responsible for five fatalities in Haiti.

1) Synoptic history

Ernesto formed from a tropical wave that emerged from the west coast of Africa on 18 August and moved steadily westward across the tropical Atlantic during the following several days. On 23 August, convection associated with the wave increased, triggering the first Dvorak classifications at 1200 UTC, when the wave was located about 500 n mi east of the Lesser Antilles. As the wave approached the Lesser Antilles, a surface low formed near the southern Windward Islands. The system soon acquired a well-defined surface circulation and became a tropical depression at 1800 UTC 24 August, while centered about 40 n mi north-northwest of Grenada.

Convection increased over the low-level center of the depression as it moved toward the west-northwest along the southern periphery of a midlevel ridge over the western Atlantic. The cyclone became a tropical storm at 1200 UTC 25 August while centered about 280 n mi south of San Juan, Puerto Rico. Ernesto turned toward the northwest and continued to intensify as it moved into the central Caribbean Sea on 26 August. The next day at about 0600 UTC, while Ernesto was centered about 70 n mi south of the southern coast of Haiti, it was very briefly of hurricane strength, with maximum sustained winds of 65 kt and a minimum pressure of about 992 mb. Shortly thereafter, the small inner core of the storm deteriorated as the circulation interacted with the mountainous terrain of Haiti, and Ernesto quickly weakened back to a tropical storm. The center of circulation passed offshore very near the southwestern tip of Haiti at about 0000 UTC 28 August, by which time the storm’s intensity had decreased to 40 kt.

Under southwesterly shear associated with an upper-level low over the Bahamas, Ernesto continued to weaken as it moved northwestward toward Cuba. The storm had maximum sustained winds of 35 kt when its center made landfall along the southeastern coast of Cuba, near Playa Cazonal and just west of Guantanamo Bay, at 1115 UTC 28 August. Ernesto turned northwestward and its center stayed inland over Cuba for about 18 h, but it remained a tropical storm during that time. The center emerged off the north-central coast of Cuba by 0600 UTC 29 August and the intensity increased slightly to 40 kt. A midlevel high pressure area over the southeastern United States was migrating eastward during this period, allowing Ernesto to continue northwestward. The storm began to traverse the warm waters of the Florida Straits late on 29 August, and convection gradually increased. Apparently still hindered by disruption of its inner core over Cuba, and possibly influenced by moderate easterly wind shear, Ernesto did not gain any additional strength between Cuba and southern Florida.

Ernesto made landfall at Plantation Key, Florida, around 0300 UTC 30 August. Two hours later, a second landfall occurred on the Florida mainland in southwestern Miami–Dade County. At both landfalls Ernesto had maximum sustained winds of 40 kt and a minimum central pressure of 1003 mb. Thereafter, Ernesto weakened only slightly, and it remained a tropical storm throughout its passage over Florida. The storm moved northward along the center of the Florida Peninsula and within a weakness in the midlevel ridge, its center passing over Lake Okeechobee around 1800 UTC 30 August. Ernesto turned north-northeastward, and its center emerged over the Atlantic Ocean near Cape Canaveral, Florida, very early on 31 August.

Fueled by the warm waters of the Atlantic, convection increased over the center of the cyclone, and Ernesto intensified to a strong tropical storm as it continued north-northeastward ahead of a deep-layer trough approaching from the west. It reached an intensity of 60 kt by 1800 UTC 31 August while centered about 150 n mi south-southwest of Wilmington (Fig. 6). The central pressure continued to gradually fall, and an eye was becoming discernible in satellite imagery as the storm center approached the coast. The center came ashore at 0340 UTC 1 September on Oak Island, North Carolina, a few miles south-southwest of Wilmington and just west of Cape Fear. At the time of landfall, Ernesto was very near the threshold between tropical storm and hurricane status, with an intensity of 60 kt and a minimum pressure of 985 mb. Ernesto weakened as it moved across eastern North Carolina, where it became a tropical depression by 1200 UTC 1 September.

Fig. 6.

GOES-12 visible satellite image of Hurricane Ernesto (as a tropical storm) at 1815 UTC 31 Aug 2006 (image courtesy of University of Wisconsin—Madison).

Fig. 6.

GOES-12 visible satellite image of Hurricane Ernesto (as a tropical storm) at 1815 UTC 31 Aug 2006 (image courtesy of University of Wisconsin—Madison).

Even before Ernesto had reached North Carolina, the pressure gradient between the tropical cyclone and high pressure centered over southeastern Canada was producing gale-force winds near the coasts of Virginia, Maryland, Delaware, and New Jersey. These gales continued even after Ernesto weakened to a tropical depression. By the time the center of circulation reached the North Carolina–Virginia border at about 1800 UTC 1 September, Ernesto had interacted with a preexisting frontal zone and become extratropical. The extratropical remnant of Ernesto moved slowly northward over Virginia and Maryland on 2 September, with peak winds of about 40 kt. By 1800 UTC that day, when the center of circulation was near Washington, D.C., the system began to weaken. On 3 September, it was no longer producing gale-force winds as it accelerated across Pennsylvania and New York into southeastern Canada, and it was absorbed into a larger extratropical low pressure system the next day.

2) Meteorological statistics

Ernesto was briefly near the threshold between tropical storm and hurricane strength over the central Caribbean Sea early on 27 August. A dropsonde at 0832 UTC measured an instantaneous wind of 64 kt at 10 m, but poststorm analysis of the profile suggested that this observation was most likely representative of a gust rather than a sustained wind; the surface estimate derived from the mean wind over the lowest 150 m of this sounding was only 56 kt. The strongest flight-level wind measurement that day was 78 kt at 0732 UTC, corresponding to about 62 kt at the surface using the average adjustment from 850 mb (Franklin et al. 2003). One hour later, at 0834 UTC, the aircraft measured a flight-level wind of 69 kt at 700 mb, which also corresponds to about 62 kt at the surface. Given the inability of the aircraft to sample the entire circulation, it is possible that during this period somewhat stronger winds occurred elsewhere that would be consistent with the operational estimate of Ernesto briefly having 65 kt winds near 0900 UTC 27 August.

Ernesto’s estimated Florida landfall intensity of 40 kt is based on aircraft observations. The strongest sustained surface wind measured in Florida was 35 kt, obtained by an instrument on Lake Okeechobee operated by the South Florida Water Management District (SFWMD), at 2030 UTC 30 August (Table 4). The Lake Worth Coastal-Marine Automated Network (C-MAN) station, on the Atlantic coast in Palm Beach County, also reported a sustained wind of 35 kt that day. These data suggest that Ernesto remained a tropical storm throughout its entire path over Florida, although the only areas to receive sustained winds of tropical storm force likely were near or over water.

Table 4.

Selected surface observations for Hurricane Ernesto, 24 Aug–1 Sep 2006. The table includes observations indirectly associated with Ernesto as well as during Ernesto’s extratropical stage.

Selected surface observations for Hurricane Ernesto, 24 Aug–1 Sep 2006. The table includes observations indirectly associated with Ernesto as well as during Ernesto’s extratropical stage.
Selected surface observations for Hurricane Ernesto, 24 Aug–1 Sep 2006. The table includes observations indirectly associated with Ernesto as well as during Ernesto’s extratropical stage.

Ernesto was again near the threshold between tropical storm and hurricane status near the time of landfall in North Carolina. The strongest flight-level (850 mb) wind on 31 August was 73 kt at 1947 UTC, corresponding to about 58 kt at the surface. The SFMR instrument measured 60–65 kt near 1800 UTC, but the higher end of this range resulted from isolated peak values that appeared slightly inflated by rain and were therefore discounted. Radar velocities measured from the Weather Surveillance Radar-1988 Doppler (WSR-88D) at Wilmington reached 80–84 kt at an altitude of 1500 ft just west of the circulation center less than 3 h prior to landfall, corresponding to about 60–63 kt at the surface, but the highest values were associated with a transient, mesoscale feature. Although the best-track intensity during this period was set to 60 kt, it is possible that the maximum wind was not sampled and that Ernesto might have reached hurricane strength near North Carolina.

The lowest central pressure of 985 mb occurred near the time of final landfall, as evidenced by a surface observation of 985.4 mb at Wilmington. The central pressure remained relatively low as Ernesto proceeded inland. Six hours after landfall, when the center was about 90 n mi inland, a pressure of 988.5 mb was measured at Kinston, North Carolina. The strongest sustained wind measured by an official surface-based anemometer in North Carolina was 50 kt at the Wrightsville Beach NOS station at Johnny Mercer Pier, where a gust to 64 kt was also reported. An unofficial gust to 79 kt was reported at Cedar Island, just northeast of Cape Lookout. The storm surge produced by Ernesto was just less than 1 m, although a total water rise (storm tide) of nearly 2 m was reported in a few locations. The surge affected a long stretch of the North Carolina coastline, including the bays, harbors, and rivers adjacent to Pamlico Sound as far north as Dare County (Table 4).

The interaction of Ernesto with high pressure centered over southeastern Canada produced sustained gale-force winds and heavy rains over and near the coasts of Virginia, Maryland, Delaware, and New Jersey. Some of the observations included in Table 4 are associated with this interaction and are therefore not reflected in the best-track intensity of the tropical cyclone. The combined effects of the two systems, as well as the extratropical remnant of Ernesto, resulted in storm surge flooding along the western shores of Chesapeake Bay and the adjacent rivers, where storm tides of up to about 2 m were reported.

Storm-total rainfall amounts exceeded 125 mm throughout a broad swath across eastern South Carolina, North Carolina, and Virginia, as well as southern Maryland (Table 4). More than 250 mm of rain fell in several locations in North Carolina and Virginia, including a maximum of 371 mm at Wrightsville Beach, North Carolina. The heavy rains also led to river flooding for several days after Ernesto’s landfall. In particular, the Northeast Cape Fear River crested at 5.7 m at Chinquapin, well north of Wilmington, and remained in major flood stage from 2–7 September.

The large metropolitan areas of southeastern Florida escaped with no more than 25–50 mm of rain from Ernesto. The storm dropped 75–150 mm of rain, however, in many areas near the path of the storm’s center from the Cape Canaveral area to Lake Okeechobee, in portions of southwestern Florida, and in isolated spots in the upper Florida Keys. A storm-total maximum of 221 mm was reported at the South Golden Gate Estates SFWMD station located east of Naples. In the Caribbean, 75–125 mm of rain were common over much of eastern Cuba as Ernesto moved over the island, with a maximum of 189 mm reported at Nuevitas, Camaguey. In the Dominican Republic, 178 mm of rain fell at Barahona as the center of Ernesto passed to the southwest.

A total of five weak tornadoes were reported in association with Ernesto. Two of these touched down in Osceola County in central Florida on the afternoon of 30 August, and the other three were reported in eastern North Carolina on the evening of 31 August.

3) Casualty and damage statistics

Ernesto was directly responsible for five fatalities, all in Haiti. In addition, the extratropical remnant of Ernesto was responsible for two deaths that occurred in Gloucester, Virginia, when a tree fell on a residence.

In the Caribbean, Ernesto was associated with damage in Haiti, the Dominican Republic, and Cuba. Ernesto caused mudslides and flooded many residences across the mountainous terrain of Hispaniola. At least six homes were destroyed with dozens more damaged in Haiti, and a bridge was destroyed in Port-au-Prince. More than 400 homes were flooded in the city of Santo Domingo in the Dominican Republic. Only minor damage was reported in Cuba.

The estimate of insured losses in the United States, provided by the Property Claim Services of the Insurance Services Office, is $245 million. Total U.S. property damage is estimated at roughly $500 million. In eastern North Carolina, heavy rains resulted in the flooding of several homes, and other homes were damaged by strong winds. For days following landfall, rain-induced river flooding inundated several homes along the Northeast Cape Fear River north of Wilmington. Storm surge caused minor coastal flooding and beach erosion along the immediate Atlantic coastline. The surge along bays and rivers, including the Pamlico and Pungo Rivers in Beaufort County and Collington Harbor in Dare County, flooded several homes and businesses. Minor property damage was caused by the three tornadoes in eastern North Carolina. In Virginia, storm surge along the western shores of the Chesapeake Bay and into tidal sections of adjacent rivers flooded several homes, and some piers and boats were significantly damaged. Strong winds downed trees and power lines in coastal areas of North Carolina, Virginia, Delaware, and New Jersey. Ernesto’s heavy rains in southwestern Florida caused flooding in several homes in Glades County.

g. Hurricane Florence, 3–12 September

Florence had a complex genesis that involved two tropical waves, the first of which moved across the west coast of Africa on 29 August. This wave moved slowly westward, showing some signs of convective organization on 31 August, when a second wave moved westward into the Atlantic at a faster forward speed. By 2 September, the two waves combined to form a large area of disturbed weather over the eastern tropical Atlantic. The convective organization of the combined disturbance increased, and it is estimated that a tropical depression formed near 1800 UTC 3 September about 855 n mi west of the Cape Verde Islands.

With multiple vorticity maxima rotating around a mean center of circulation for several days, the depression’s initial structure did not favor rapid development. In addition, southwesterly vertical shear associated with an upper-level trough to the system’s west hindered development, and the depression took two days to reach storm strength as it moved west-northwestward. However, the shear decreased on 8 September when an upper-level ridge developed over Florence, and late that day convection began to consolidate around a vorticity center on the western side of the large cyclonic envelope. Florence strengthened slowly on 9 September and more rapidly the following day, becoming a hurricane early on 10 September about 340 n mi south of Bermuda.

Florence turned northward and reached its estimated peak intensity of 80 kt late on 10 September, but had weakened slightly when its center passed about 50 n mi west of Bermuda on 11 September. The hurricane recurved northeastward into the westerlies the next day, maintaining hurricane strength until it became extratropical early on 13 September about 420 n mi south-southwest of Cape Race, Newfoundland.

The extratropical remnant of Florence maintained a large circulation and hurricane-force winds as it approached Newfoundland. The center passed near Cape Race late on 13 September, then moved east-northeastward over the open North Atlantic before turning eastward and weakening on 14 September. After making a large cyclonic half-loop from 16–19 September southwest of Iceland, the remnant extratropical low was absorbed by a developing extratropical system to its south.

Florence brought hurricane conditions to Bermuda on 11 September (Table 5). An automated station at St. David’s (elevation 48 m) reported sustained winds of 71 kt at 1340 UTC with a gust to 97 kt. The Bermuda Maritime Operations Centre reported a wind gust of 100 kt at 1001 UTC. The Bermuda Airport reported a peak gust of 78 kt at 1555 UTC along with 34 mm of rain. The extratropical remnant of Florence brought hurricane-force winds to portions of Newfoundland on 13 September, where Sagona Island reported sustained winds of 66 kt with a gust to 81 kt at 1500 UTC.

Table 5.

Selected surface observations for Hurricane Florence, 3–12 Sep.

Selected surface observations for Hurricane Florence, 3–12 Sep.
Selected surface observations for Hurricane Florence, 3–12 Sep.

Media reports indicate that Florence caused no deaths on Bermuda and only a few minor injuries. The hurricane caused minor wind damage and power outages on the island. Minor wind damage and power outages also occurred in Newfoundland.

h. Hurricane Gordon, 10–20 September

Gordon, a category 3 hurricane at its peak, affected the Azores as a hurricane, and as an extratropical low brought heavy rains and high winds to parts of western Europe. The cyclone originated from a tropical wave that left the west coast of Africa on 2 September. This wave was initially well organized, with an associated surface low along the wave axis and some convection. However, vertical wind shear from an upper-level trough east of then Tropical Storm Florence hindered development for about a week as the wave and trough both progressed westward. By 9 September, the vertical shear in the vicinity of the tropical wave lessened slightly, and convection increased around the surface low. It is estimated that a tropical depression formed at 1800 UTC 10 September about 470 n mi east-northeast of the Leeward Islands.

The depression moved west-northwestward initially, and became a tropical storm the next day. Gordon turned toward the northwest and slowed on 12 September as it headed into a weakness in the subtropical ridge left behind Florence. The upper trough, which had continued to produce moderate southwesterly shear over the cyclone during this time, weakened and moved to the southwest, allowing the shear to abate. The storm turned toward the north and intensified into a hurricane, forming a large ragged eye early on 13 September. Gordon then rapidly intensified, its peak winds increasing by 50 kt in the 30-h period from 1800 UTC 12 September until 0000 UTC 14 September, and reached its estimated peak intensity of 105 kt while centered about 500 n mi east-southeast of Bermuda.

Gordon turned toward the northeast and briefly accelerated on 14 September, but an upper trough that had been steering the system bypassed the hurricane late that day. Midtropospheric ridging rebuilt to the north of Gordon and its forward motion stalled on 15 September, about 575 n mi east-southeast of Bermuda. Increasing vertical wind shear and the upwelling of cooler waters then caused Gordon to weaken to a category 1 hurricane on 16 September as it drifted northwestward.

Early on 17 September, building high pressure to the east of Gordon caused the hurricane to begin moving northeastward. Gordon accelerated to the northeast on 18 September, and turned to the east-northeast later that day around the strengthening subtropical high (Fig. 7). Vertical wind shear relaxed during this time and convection redeveloped around the eye. Despite moving over relatively cool waters of about 25°C, the cyclone reached a second peak intensity of 90 kt on 19 September about 420 n mi west-southwest of the Azores. Below-normal upper-tropospheric temperatures over the subtropical Atlantic likely countered the effects of the cool waters for a time, but later that day wind shear increased and Gordon began to weaken. The hurricane moved a little south of due east early on 20 September, its center passing between the islands of São Miguel and Santa Maria in the Azores at about 0900 UTC. Hurricane-force wind gusts were experienced on Santa Maria, but Gordon’s strongest winds remained south of the islands. Later that day, Gordon began interacting with a cold front and weakened, and the system became extratropical at 0000 UTC 21 September about 240 n mi west of the west coast of Portugal.

Fig. 7.

GOES-12 visible satellite image of Hurricanes Gordon and Helene at 1745 UTC 18 Sep 2006 (image courtesy of University of Wisconsin—Madison). Gordon is located at the top of the image.

Fig. 7.

GOES-12 visible satellite image of Hurricanes Gordon and Helene at 1745 UTC 18 Sep 2006 (image courtesy of University of Wisconsin—Madison). Gordon is located at the top of the image.

The extratropical remnant of Gordon was a vigorous system, maintaining winds of 55 kt or greater for almost two days after losing tropical characteristics. The extratropical low turned northeastward and was then driven northward at almost 50 kt on 21 September around a larger midlatitude low. Gordon’s remnant remained distinct from the larger system, bringing heavy rains and winds gusts of hurricane-force to Spain and the United Kingdom, and on 22 September sustained winds with the system increased to hurricane force. Gordon’s extratropical remnant became the dominant low pressure area and made a large cyclonic loop while slowly weakening over the far northeastern Atlantic Ocean. The system completed the loop just south of Ireland on 24 September and the low dissipated between Ireland and England later that day.

Observations of interest in Gordon include a peak wind gust of 71 kt, recorded at Santa Maria in the Azores on 20 September. The extratropical remnant of Gordon produced numerous gusts of hurricane force in Spain on 21 September, including one gust to 99 kt on the northwest coast at Punta Candieira. The British Broadcasting Corporation reported that wind gusts to 70 kt were measured in southwestern England, but no exact location was given.

There were no reports of casualties associated with Gordon as a tropical cyclone. The BBC reported only minor damage in the Azores, consisting mostly of fallen trees and power outages. After Gordon became an extratropical low, however, four injuries due to falling debris from high wind were reported in Spain. Power was also reported out for 100 000 customers in Spain. The extratropical cyclone brought high winds and rain that affected practice rounds at the Ryder Cup golf tournament in Ireland. About 126 000 homes were without power after the storm in Northern Ireland and one injury was reported.

i. Hurricane Helene, 12–24 September

Helene developed from a vigorous tropical wave and broad area of low pressure that emerged from the coast of Africa on 11 September. Convective activity associated with the low quickly increased, and by 1200 UTC 12 September the system had developed into a tropical depression while centered approximately 200 n mi south-southeast of the Cape Verde Islands. Under easterly shear, the sprawling depression was slow to develop initially. On 13 September, however, convective banding began to increase over the northwestern semicircle of the circulation and the depression became a tropical storm at 0000 UTC 14 September about 425 n mi west-southwest of the Cape Verde Islands.

Moving west-northwestward over the tropical Atlantic Ocean, Helene steadily intensified and became a hurricane at 1200 UTC 16 September about 1000 n mi east of the northern Leeward Islands. The next day Helene turned northwestward and slowed down in response to a weakness in the subtropical ridge created by Hurricane Gordon, which at that time was located a little over 1000 n mi northwest of Helene. Helene continued to strengthen, attaining category 3 status at 0000 UTC 18 September, and 6 h later it reached its estimated peak intensity of 105 kt. At this time, Helene and Gordon were at nearly the same longitude, with Gordon centered about 875 n mi north of Helene.

Helene had developed a concentric eyewall structure, and it weakened to a category 2 hurricane when the inner eyewall began to deteriorate late on 18 September (Fig. 7). As Gordon moved east of Helene’s longitude, a narrow mid- to upper-level ridge built to the north of Helene, which caused the hurricane to turn westward on 19 September. The following day, Helene turned northward ahead of a large deep-layer trough along the East Coast. Moderate southwesterly shear ahead of this trough induced additional weakening, and Helene was a category 1 hurricane when it passed about 475 n mi east of Bermuda early on 21 September. Thereafter, Helene turned east-northeastward over the open waters of the central Atlantic and retained hurricane strength until it became extratropical on 24 September about 275 n mi northwest of the Azores.

On 25 and 26 September, the extratropical cyclone moved northeastward and weakened to a gale center before passing very near the west coast of Ireland on 27 September. The system produced gale-force wind gusts across much of Ireland and northwestern Scotland on 27–28 September. A wind gust of 49 kt was reported at the Valentia Observatory on the southwestern coast of Ireland, and in Scotland a wind gust of 64 kt was reported on South Uist Island in the Outer Hebrides. Just before reaching northwestern Scotland, the extratropical remnant of Helene merged with a larger extratropical low on 28 September.

The strongest winds measured in Helene were from a NOAA reconnaissance aircraft, which reported 111 kt at a flight level of 850 mb in the eastern eyewall at 1831 UTC 17 September. The standard adjustment ratio for this flight level (Franklin et al. 2003) yields a surface wind of 89 kt, which was in good agreement with an SFMR maximum wind of 88 kt on this flight. QuikSCAT data on 22–23 September were particularly helpful in establishing Helene’s status as a hurricane until its extratropical transition on 24 September. A few hours before Helene became extratropical it passed within about 20 n mi of drifting buoy 44613 (located at 40.3°N, 39.6°W), which reported a minimum pressure of 971.8 mb at 0900 UTC 24 September.

j. Hurricane Isaac, 27 September–2 October

Isaac developed from a tropical wave that crossed the west coast of Africa on 18 September. Disturbed weather associated with the wave began to show signs of organization about 900 n mi west of the Cape Verde Islands on 23 September, when Dvorak classifications were initiated. As the large disturbance moved west-northwestward over the next few days, convection was intermittent under the influence of moderate to strong upper-level winds. By 27 September the upper-level winds had relaxed, and the system developed into a tropical depression about 810 n mi east-southeast of Bermuda.

The depression strengthened to a tropical storm early the next day, although the cyclone was slow to strengthen further as it moved over cool waters upwelled by Hurricanes Gordon and Helene. In addition, mid- to upper-level dry air had become entrained into Isaac’s core and the cyclone took on a subtropical appearance on 28 September, with little convection near the core. Initially, Isaac moved northwestward around an upper-level low to its west-southwest, but on 29 September the cyclone turned west-northwestward in response to midlevel ridging over the central Atlantic, away from the upwelled waters and into an environment of lower shear. Deep convection redeveloped near the center, the cloud pattern became more consolidated, and Isaac began to strengthen, becoming a hurricane near 1200 UTC 30 September. Isaac reached its peak intensity of 75 kt around 0000 UTC 1 October, while centered about 285 n mi east of Bermuda.

Later that day, Isaac began to recurve around the western periphery of the subtropical ridge with an increase in forward speed, moving quickly to the north-northeast the next day ahead of an approaching deep-layer trough. During this time, Isaac encountered increasing southwesterly shear and cooler waters. By 1200 UTC 2 October, it weakened to a tropical storm. While racing toward the northeast at around 35 kt, Isaac’s center passed about 35 n mi southeast of the Avalon Peninsula of Newfoundland late in the afternoon. Tropical storm–force winds were felt across portions of the southern Avalon Peninsula; at Cape Race sustained winds reached 40 kt and gusts reached 52 kt. Reported rainfall amounts were generally less than 25 mm. As it passed Newfoundland, Isaac maintained a core of strong winds and convection near the center, but it lost tropical characteristics quickly thereafter. By 0000 UTC 3 October Isaac had become extratropical, and it was absorbed within a larger extratropical low by 1800 UTC that day.

There were no reports of damage or casualties associated with Isaac.

3. Forecast verifications and warnings

For all operationally designated tropical (or subtropical) cyclones in the Atlantic and eastern North Pacific basins, the NHC issues an official forecast of the cyclone’s center position and maximum 1-min surface wind speed. Forecasts are issued every 6 h and contain projections valid at 12, 24, 36, 48, 72, 96, and 120 h after the forecast’s nominal initial time (0000, 0600, 1200, or 1800 UTC). At the conclusion of the season, forecasts are evaluated by comparing the projected positions and intensities to the corresponding poststorm derived “best track” positions and intensities for each cyclone. A forecast is included in the verification only if the system is classified in the best track as a tropical (or subtropical) cyclone at both the forecast’s initial time and at the projection’s valid time. All other stages of development (e.g., tropical wave, remnant low, extratropical) are excluded. For verification purposes, forecasts associated with special advisories3 do not supersede the original forecast issued for that synoptic time; rather, the original forecast is retained. All verifications reported here include the depression stage.

It is important to distinguish between forecast error and forecast skill. Track forecast error is defined as the great-circle distance between a cyclone’s forecast position and the best-track position at the forecast verification time. Skill, on the other hand, represents a normalization of forecast error against some standard or baseline. By convention, tropical cyclone forecast skill is positive when forecast errors are smaller than the errors from the baseline. Particularly useful skill standards are independent of operations (and hence can be applied retrospectively to historical data), and provide a measure of inherent forecast difficulty. For tropical cyclone track forecasts, the skill baseline is the Climatology and Persistence (CLIPER5) model, which contains no information about the current state of the atmosphere (Neumann 1972; Aberson 1998). The version of CLIPER5 presently in use is based on developmental data from 1931 to 2004 for the Atlantic and from 1949 to 2004 for the eastern Pacific.

Table 6 presents the results of the NHC official track forecast verification for the 2006 season, along with results averaged for the previous 5-yr period of 2001–2005.4 Mean track errors ranged from 30 n mi at 12 h to 265 n mi at 120 h. It is seen that mean official track forecast errors were smaller in 2006 than during the previous 5-yr period (by roughly 15%–20% out to 72 h), and at forecast projections through 72 h the errors established new all-time lows. Since 1990, 24–72-h official track forecast errors have been reduced by roughly 50% (Franklin 2007). Fairly substantial vector biases at the longer ranges were noted in 2006; at 120 h the official forecast bias was 82 n mi to the west of the verifying position. This bias, about 35% of the mean error magnitude, was about twice as large as the average bias in the leading dynamical track models, and was most pronounced for storms south of 20°N, suggesting a tendency to inadequately consider forecast guidance calling for recurvature.

Table 6.

Homogenous comparison of official and CLIPER5 track forecast errors in the Atlantic basin for the 2006 season for all tropical and subtropical cyclones. Long-term averages are shown for comparison.

Homogenous comparison of official and CLIPER5 track forecast errors in the Atlantic basin for the 2006 season for all tropical and subtropical cyclones. Long-term averages are shown for comparison.
Homogenous comparison of official and CLIPER5 track forecast errors in the Atlantic basin for the 2006 season for all tropical and subtropical cyclones. Long-term averages are shown for comparison.

While the track forecasts at each time period were more accurate in 2006 than they had been over the previous 5-yr period, only the forecasts from 12–72 h were also more skillful. The improved skill at 12–72 h in 2006 occurred despite the fact that CLIPER5 errors during 2006 were also below average, indicating below-average forecast difficulty. It is worth noting that the 96- and 120-h CLIPER5 errors and sample sizes were anomalously low in 2006, so the loss of skill at these time periods is likely of no long-term significance. An examination of annual skill trends (not shown) suggests that shorter-range forecast skill continues to increase, while no clear trend is apparent for 72 h and beyond.

Forecast intensity error is defined as the absolute value of the difference between the forecast and best-track intensity at the forecast verifying time. The skill of intensity forecasts is assessed using a modified version of the Statistical Hurricane Intensity Forecast (SHIFOR5) climatology and persistence model (Jarvinen and Neumann 1979; Knaff et al. 2003). The modified model, known as DSHIFOR5, is constructed by taking the output from SHIFOR5 and applying the decay rate of DeMaria et al. (2006). The application of the decay component requires a forecast track, which here is given by CLIPER5. The use of DSHIFOR5 as the intensity skill benchmark is new for 2006. On average, DSHIFOR5 errors are about 5%–15% lower than SHIFOR5 in the Atlantic basin from 12 to 72 h, and about the same as SHIFOR5 at 96 and 120 h.

Table 7 presents the results of the NHC official intensity forecast verification for the 2006 season, along with results for the preceding 5-yr period. Mean forecast errors in 2006 were very close to the 5-yr means, with errors ranging from about 7 kt at 12 h to just below 20 kt at 96 and 120 h. Forecast biases, however, were large and positive—near 5 kt at 72 h and over 7 kt at 96 h. In contrast, intensity biases for the period 2001–05 were near zero. It is interesting that these large positive biases occurred in a year for which there were very few instances of rapid strengthening (only 3.3% of all 24-h intensity changes qualified),5 but which followed a season that featured many such cases (2005 featured a 7.1% occurrence rate), and indeed the 2001–05 period as a whole had numerous strong and rapidly deepening storms. The lack of such storms in 2006 led to DSHIFOR5 errors that were considerably below normal (i.e., this year’s storms should have been relatively easy to forecast). However, these low DSHIFOR5 errors, coupled with the tendency to overforecast intensification, resulted in strongly negative official forecast skill in 2006.

Table 7.

Homogenous comparison of official and DSHIFOR5 intensity forecast errors in the Atlantic basin for the 2006 season for all tropical and subtropical cyclones. Long-term averages are shown for comparison.

Homogenous comparison of official and DSHIFOR5 intensity forecast errors in the Atlantic basin for the 2006 season for all tropical and subtropical cyclones. Long-term averages are shown for comparison.
Homogenous comparison of official and DSHIFOR5 intensity forecast errors in the Atlantic basin for the 2006 season for all tropical and subtropical cyclones. Long-term averages are shown for comparison.

Additional information on the 2006 forecast verification, both for official NHC forecasts and the objective guidance models, is given by Franklin (2007).

NHC defines a hurricane (or tropical storm) warning as a notice that 1-min mean winds of hurricane (or tropical storm) force are expected within a specified coastal area within the next 24 h. A watch indicates that those conditions are possible within 36 h. Table 8 lists lead times associated with those tropical cyclones that affected the United States in 2006. Because observations are generally inadequate to determine when hurricane or tropical storm conditions first reach the coastline, for purposes of this discussion, lead time is defined as the time elapsed between the issuance of the watch or warning and the time of landfall or closest approach of the center to the coastline. Such a definition will usually overstate by a few hours the actual preparedness time available, particularly for tropical storm conditions. The table includes only the most significant (i.e., strongest) landfall for each cyclone, and only verifies the strongest conditions occurring on shore. The issuance of warnings for non-U.S. territories is the responsibility of the governments affected and is not tabulated here. The table shows that warning goals were generally met in 2006, although no watch preceded the warning issued for Alberto.

Table 8.

Watch and warning lead times (see text) for tropical cyclones affecting the United States in 2006. For cyclones with multiple landfalls, the most significant is given. If multiple watch or warning types (TS or H) were issued, the type corresponding to the most severe conditions experienced over land is given.

Watch and warning lead times (see text) for tropical cyclones affecting the United States in 2006. For cyclones with multiple landfalls, the most significant is given. If multiple watch or warning types (TS or H) were issued, the type corresponding to the most severe conditions experienced over land is given.
Watch and warning lead times (see text) for tropical cyclones affecting the United States in 2006. For cyclones with multiple landfalls, the most significant is given. If multiple watch or warning types (TS or H) were issued, the type corresponding to the most severe conditions experienced over land is given.

Acknowledgments

The cyclone summaries are based on tropical cyclone reports written by the authors and their NHC hurricane specialist colleagues: Lixion Avila, Jack Beven, Richard Pasch, Richard Knabb, Stacy Stewart, Eric Blake, Jamie Rhome, and Michelle Mainelli. The authors thank Chris Velden and David Stettner of the University of Wisconsin—Madison Cooperative Institute for Meteorological Satellite Studies (CIMSS) for the satellite images presented here. Tropical Prediction Center colleague Ethan Gibney produced the track chart. Additional figures were provided by Joan David and Eric Blake. Much of the local impact information contained in the individual storm summaries was compiled by local NWS Weather Forecast Offices in the affected areas.

REFERENCES

REFERENCES
Aberson
,
S. D.
,
1998
:
Five-day tropical cyclone track forecasts in the North Atlantic basin.
Wea. Forecasting
,
13
,
1005
1015
.
Aberson
,
S. D.
, and
J. L.
Franklin
,
1999
:
Impact on hurricane track and intensity forecasts of GPS dropwindsonde observations from the first season flights of the NOAA Gulfstream-IV jet aircraft.
Bull. Amer. Meteor. Soc.
,
80
,
421
427
.
Bell
,
G. D.
,
E.
Blake
,
C. W.
Landsea
,
M.
Chelliah
,
R.
Pasch
,
K. C.
Mo
, and
S. B.
Goldenberg
,
2007
:
State of the Climate in 2006—Atlantic basin.
Bull. Amer. Meteor. Soc.
,
88
.
(Suppl.), doi:10.1175/BAMS-88-6-StateoftheClimate
.
DeMaria
,
M.
,
J. A.
Knaff
, and
J.
Kaplan
,
2006
:
On the decay of tropical cyclone winds crossing narrow landmasses.
J. Appl. Meteor.
,
45
,
491
499
.
Dvorak
,
V. E.
,
1984
:
Tropical cyclone intensity analysis using satellite data. NOAA Tech. Rep. NESDIS 11, National Oceanic and Atmospheric Administration, Washington, DC, 47 pp
.
Franklin
,
J. L.
,
cited
.
2007
:
2006 National Hurricane Center Forecast Verification Report. [Available online at http://www.nhc.noaa.gov/verification/pdfs/Verification_2006.pdf.]
.
Franklin
,
J. L.
,
M. L.
Black
, and
K.
Valde
,
2003
:
GPS dropwindsonde wind profiles in hurricanes and their operational implications.
Wea. Forecasting
,
18
,
32
44
.
Hart
,
R. E.
,
2003
:
A cyclone phase space derived from thermal wind and thermal asymmetry.
Mon. Wea. Rev.
,
131
,
585
616
.
Hawkins
,
J. D.
,
T. F.
Lee
,
J.
Turk
,
C.
Sampson
,
F. J.
Kent
, and
K.
Richardson
,
2001
:
Real-time Internet distribution of satellite products for tropical cyclone reconnaissance.
Bull. Amer. Meteor. Soc.
,
82
,
567
578
.
Hebert
,
P. J.
, and
K. O.
Poteat
,
1975
:
A satellite classification technique for subtropical cyclones. NOAA Tech. Memo. NWS SR-83, 25 pp. [Available from National Weather Service, Fort Worth, TX 76102.]
.
Herndon
,
D. C.
, and
C.
Velden
,
2004
:
Upgrades to the UW-CIMSS AMSU-based tropical cyclone intensity estimation algorithm. Preprints, 26th Conf. on Hurricane and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 118–119
.
Hock
,
T. F.
, and
J. L.
Franklin
,
1999
:
The NCAR GPS dropwindsonde.
Bull. Amer. Meteor. Soc.
,
80
,
407
420
.
Jarvinen
,
B. R.
, and
C. J.
Neumann
,
1979
:
Statistical forecasts of tropical cyclone intensity for the North Atlantic basin. NOAA Tech. Memo. NWS NHC-10, 22 pp
.
Kaplan
,
J.
, and
M.
DeMaria
,
2003
:
Large-scale characteristics of rapidly intensifying tropical cyclones in the North Atlantic basin.
Wea. Forecasting
,
18
,
1093
1108
.
Kistler
,
R.
, and
Coauthors
,
2001
:
The NCEP–NCAR 50-Year Reanalysis: Monthly means CD-ROM and documentation.
Bull. Amer. Meteor. Soc.
,
82
,
247
267
.
Knaff
,
J. A.
,
M.
DeMaria
,
B.
Sampson
, and
J. M.
Gross
,
2003
:
Statistical, five-day tropical cyclone intensity forecasts derived from climatology and persistence.
Wea. Forecasting
,
18
,
80
92
.
Neumann
,
C. B.
,
1972
:
An alternate to the HURRAN (hurricane analog) tropical cyclone forecast system. NOAA Tech. Memo. NWS SR-62, 24 pp
.
Saffir
,
H. S.
,
1973
:
Hurricane wind and storm surge.
Mil. Eng.
,
423
,
4
5
.
Simpson
,
R. H.
,
1974
:
The hurricane disaster potential scale.
Weatherwise
,
27
,
169
.
186
.
Tsai
,
W-Y.
,
M.
Spender
,
C.
Wu
,
C.
Winn
, and
K.
Kellogg
,
2000
:
SeaWinds of QuikSCAT: Sensor description and mission overview. Proc. Int. Geoscience and Remote Sensing Symp., Vol 3., Honolulu, HI, IEEE, 1021–1023
.
Uhlhorn
,
E. W.
, and
P. G.
Black
,
2003
:
Verification of remotely sensed sea surface winds in hurricanes.
J. Atmos. Oceanic Technol.
,
20
,
99
116
.
Velden
,
C. S.
, and
K. F.
Brueske
,
1999
:
Tropical cyclone warm cores as observed from the NOAA polar orbiting satellite’s new Advanced Microwave Sounder Unit. Preprints, 23rd Conf. on Hurricane and Tropical Meteorology, Dallas, TX, Amer. Meteor. Soc., 182–185
.

Footnotes

Corresponding author address: James L. Franklin, National Hurricane Center, NOAA/NWS, 11691 SW 17th Street, Miami, FL 33165-2149. Email: james.franklin@noaa.gov

1

This total includes Alex and Ophelia, which brought hurricane-force winds to the coastline but whose centers remained offshore.

2

Tabulations of the 6-hourly best-track positions and intensities can be found in the NHC tropical cyclone reports (see Web address in text). These reports contain storm information omitted here because of limitations of space, including additional surface observations and a forecast and warning critique.

3

Special advisories are issued whenever an unexpected significant change has occurred or when U.S. watches or warnings are to be issued between regularly scheduled advisories. The treatment of special advisories in forecast databases has not been consistent over the years. The current practice of retaining and verifying the original advisory forecast began in 2005.

4

It has been traditional to use a 10-yr sample to establish representative NHC official forecast error characteristics. Given the increase in storm activity in recent years, as well as the significant improvements in track forecast accuracy, it is now felt that a 5-yr sample is more representative of the state of the science.

5

Following Kaplan and DeMaria (2003), rapid intensification is defined here as a 30-kt-or-greater increase in maximum winds in a 24-h period, and corresponds to the 5th percentile of all intensity changes in the Atlantic basin.