SECTION TWO HAWAIIAN HURRICANES THEIR HISTORY, CAUSES AND FUTURE 

INTRODUCTION

Hurricanes are one subclass of a category of phenomena known to the global meteorological community as tropical cyclones. The history of tropical cyclones in the Hawaiian Islands comprises three periods:

1. "Contact" until 1950.

2. 1950 until 1970.

3. 1970 until present.

During the first period there was no formal recognition that tropical cyclones approached the Hawaiian Islands. Efforts to document that era have been only subsequent to the recognition of tropical cyclones as a feature of Hawaiian climate.

The history of tropical cyclones in Hawaiian waters began with the recognition by Robert Simpson and the staff of the Weather Bureau Forecast Office at Honolulu that a low pressure system found east of the islands on August 12, 1950 possessed the spatial characteristic of a tropical cyclone rather than a mid-latitude or extratropical cyclone. The storm originally referred to as Hurricane ABLE became known as Hurricane Hiki. This was the first officially recognized hurricane in Hawaiian waters (Simpson, 1950).

Although the launch of the first weather satellite, TIROS 1, on April 1, 1960 improved detection of storms over open waters surrounding the State, it was only in 1969 when access to geostationary satellite imagery and operational TIROS class satellites insured that storms near Hawaii would not go undetected. Since 1969 was a quiet year we begin the quantitative record with 1970.

In this paper we first describe the essentials of tropical cyclones including definitions, structure, hazards and formation mechanisms. We then discuss the two stages of the modern era of Hawaiian hurricane climatology. We discuss notable storms, interesting additional near-misses, the role of the El Nino Southern Oscillation (ENSO) in storm formation, and the typical storm tracks.

Pre-modern history (period one above) will be summarized. We shall raise issues which preclude attaining definitive knowledge of that era. An example of these issues is the problem of differentiating between a November Kona storm and a late season hurricane (typical of ENSO years).

Lastly we shall discuss issues related to possible impacts of global environmental change on Hawaiian hurricanes. Issues include:

1. the anticipated impact of sea-level rise;

2. ENSO frequencies; and

3. sea surface temperature and its relationship to storm formation.

Definitions

Tropical cyclones are low pressure systems which form over the tropical oceans. They are smaller than mid-latitude cyclonic storms and are characterized by a core in which air temperatures are warmer than those of the surrounding environment. In contrast mid-latitude cyclones are colder than their surrounding environments. Meteorologists refer to tropical cyclones as warm core lows. Terminology varies among tropical cyclone basins. United States forecast centers classify tropical cyclones in the following categories:

Tropical Depression: A weak tropical cyclone with a surface circulation including one or more closed isobars (lines or curves of constant pressure) and highest sustained winds (measured over one minute or more) of less than 39 miles per hour. Tropical depressions are assigned a number denoting their chronological order of formation in a given year.

Tropical Storm: A tropical cyclone with highest sustained winds between 39 and 73 miles per hour.

Hurricane (or typhoon only west of 180^o longitude): A tropical cyclone with highest sustained winds between 74 and 149 miles per hour.

Super hurricane or super typhoon: A tropical cyclone with sustained winds of at least twice nominal hurricane intensity. This means sustained winds of 150 miles per hour or greater. Super hurricanes are rare; several super typhoons occur annually in the western North Pacific.

Operational forecast responsibilities for United States interests are divided among three forecast centers covering four tropical cyclone basins. The basins and forecast centers are:

North Atlantic Ocean: This region includes all waters north of the equator, the regions of primary activity are the tropical Atlantic between Africa and the Windward Antilles, the Caribbean Sea and the Gulf of Mexico. Forecast responsibility lies with the National Weather Service National Hurricane Center (NHC) in Miami, Florida.

Eastern North Pacific Ocean (Figure 1): This region extends from the west coasts of Mexico and Central America westward to 140^o West Longitude. Again forecast responsibility resides in Miami at the National Hurricane Center. This is the second most active tropical cyclone basin on the planet. An average of 16 tropical storms or hurricanes form in this basin every year. This is the primary source of tropical cyclones approaching the Hawaiian Islands (Figure 2).

Western North Pacific Ocean: This region extends from the International Date Line westward through the South China Sea. It is the most active tropical cyclone basin in the world . Forecast responsibility for American interests is assigned to the Joint Typhoon Warning Center (JTWC) on Guam. JTWC is jointly managed to the United States Air Force and Navy. In addition to western North Pacific responsibilities, JTWC also supports military assets in the Indian Ocean and western South Pacific west of the International Dateline.

Central North Pacific Ocean: Although we mention this basin last, it is the most important basin for Hawaiian concerns. Its boundaries are the equator, 140^o West Longitude and the International Dateline. Forecast responsibility resides at the Central Pacific Hurricane Center (CPHC), which is located within the National Weather Service Forecast Office at Honolulu. Subsequent discussions will focus on this basin.

Hurricane Structure

A hurricane is a compact system which consists of three major sections

(Figure 3):

1. The EYE: The eye is a relatively calm region in which winds increase from light to maximum strength over a radial distance of five to ten miles. The eye is never truly calm. The eye is free of rain but need not be cloud free. There are often low clouds and occasionally thin high clouds (cirrus).

FIGURE ONE: Tropical cyclone basins of the North Pacific Ocean.

FIGURE TWO: Number of storms passing within a 75 nautical mile of points in the Eastern Pacific region. The number of storms passing through a grid of overlapping equal-area circles was counted for the period 1966-1984 and normalized to give the frequency for a 100-year period. (Data compiled by C. Neumann of the National Hurricane Center and presented by W. Frank, 1987.)

FIGURE THREE: Idealized profiles of winds as a function of radius for a typical hurricane. Storm is moving into the paper. Winds are stronger in the right-hand semi-circle relative to the moving storm.

2. The CORE: This is the principal belt of convective clouds and intense winds. The core is typically five to 10 miles wide. The core is often called the eyewall.

3. The OUTER REGION: This region extends from the core outward to the surrounding environment. This region can extend to 150 miles from the eye. The winds gradually diminish to those of the remote environment. This region is punctuated by intense rainbands which can produce strong wind gusts in squalls.

Storm dimensions vary. Generally western North Pacific typhoons are larger than Atlantic hurricanes. East Pacific hurricanes are thought to be equivalent in dimension to Atlantic systems. Few central Pacific storms have been examined in sufficient detail to draw any conclusions as to relative size.

Figure 4 depicts a hurricane of typical dimension centered upon Maui, demonstrating the horizontal extent of the winds.

Damage Sources

Hurricanes produce damage through several mechanisms. Severe damages often occur far inland from the coastal zone which is the region where most attention historically has been focused.

Winds: Hurricanes are classified by the sustained winds. The major impacts are in the coastal zone at landfall. Damages include loss of roofs from single family dwellings and condominiums, broken glass, breached doors, and, in some instances, total destruction. Engineered structures usually suffer damage to trim but complete structural failure is unlikely. Increased friction over land and separation from the oceanic energy source cause the storm to rapidly weaken. The normal expectation is that this damage would be limited to the coastal zone.

The Andrew experience suggests that wind hazards can extend further inland than previously thought. Scientists are currently trying to determine the causes for Andrew's damage swath in south Florida.

The Kauai experience with both Iwa (1982) and Iniki (1992) demonstrates that mountainous terrain can cause unusual patterns of wind enhancements and wind reductions. Little research has been devoted to the effects of

Figure Four: Idealized moderate intensity hurricane centered on Lahaina Roadstead, Maui County. This shows that a storm of this size could impact aviation at all principal airports in the State.

hurricanes on mountainous islands. Taiwanese scientists have been the most active in this field (Wang, 1989).

Surge and overwash: A storm surge is a rise of sea level due to the effects of the very low barometric pressure in the storm center (the inverse barometer) and due to winds directed in the direction of storm movement. Overwash is the piling of water in the coastal zone due to wave action. Surges can exceed twenty feet; along coasts with extensive continental shelves such as the U.S. Gulf of Mexico and the northern Bay of Bengal, surges are a major threat.

The greatest loss of life in the United States occurred on Galveston Island during a 1900 hurricane. Six thousand people died in that storm surge. The greatest modern tropical cyclone disaster occurred in 1970 in Bangladesh. Three hundred thousand people died in a surge exacerbated by inland flooding due to torrential rains. The elevated sea level restricts stream run-off, causing waters to back up in the coastal zone. Bangladesh has experienced many such disasters.

The actual storm surge for Iniki on Kauai was only five feet. The effects of overwash produced high water levels of twenty feet and more along the south shore. The Hawaiian islands rise steeply from the ocean bottom and thus have no significant shelf. The effects of Iwa first gave meteorologists cause to reevaluate previous notions about the relative roles of surge and overwash in Hawaii.

Rains: Hurricanes concentrate substantial amounts of moisture. The amount of rain falling at a particular location is a function of the local topography as well as the translation speeds of the storm. Totals of over twenty inches in one day have occurred. The world record twenty-four hour rain is 73.6 inches at Reunion Island in Mauritius. This rain was associated with a south west Indian Ocean tropical cyclone.

Heavy rains associated with hurricanes can fall well inland several days after the storm has "died" upon landfall. Examples include 40 inches of rain in northern China in 1975; 100,000 died in the subsequent floods. The source was a typhoon which had moved inland over Shanghai. In the United States Hurricane Agnes caused $3 billion in damages primarily due to flooding in Pennsylvania.

Tornadoes and other small scale winds: The core and rain bands of hurricanes produce smaller scale wind systems. The existence of tornadoes associated with hurricanes has been known for many years. Analysis of damage swaths from Andrew and Iniki by T. Fujita (personal communication) reveals additional wind systems. They appear to be concentrated downdrafts. On Kauai twenty-six such features have been found. Most were associated with the core of Iniki but some lay outside the core. While the author (Schroeder) was in Key West, Florida in June 1972, Hurricane Agnes passed over 100 miles to the west but two waterspouts (tornadoes over water) struck the lower Keys.

Conditions Necessary for Hurricane Formation

Although scientists are still attempting to specify the details of storm formation, we understand the necessary background conditions. They are:

1. Minimum sea surface temperatures: Empirically we have determined that sea surface temperature (SST) should be at least 80^oF. The waters must possess sufficient heat that conduction and evaporation can supply sufficient heat to overcome the tendency of air to cool as it moves radially toward lower pressure. Radial movement to lower pressure is equivalent to rising from sea level to higher elevations (lower pressure). Air must expand and cool. If this were to occur the storm would lose its warm core and die.

Although a minimum SST must be present, efforts to relate SST to storm intensity and frequency of storm formation have been inconclusive. Ramage (1974) demonstrated that tropical cyclones can survive short-lived exposure to waters below the threshold.

2. Instability: The atmosphere must possess temperature and moisture structures (vertical) that will allow deep convective cloud systems to form and persist. The convection is the mechanism for warming the initial central regions of the developing disturbance.

3. Shear (actually lack of shear): The mechanism for pressure fall and, hence, wind acceleration is the accumulation of warm air in a column. Any combination of horizontal winds which can carry this warm air away will prevent storm development. We state this requirement as: The horizontal winds must possess little variation (shear) of either direction and speed with height.

4. Latitude: The system must develop sufficiently far from the equator that the earth's rotation (Coriolis Force) impart cyclonic spin on the individual air parcels. Nominally 5^o separation north or south of the equator is quoted though a few storms have formed slightly closer (3^oN).

5. Initial disturbance: Some initial disturbance must be present. There is considerable debate about the nature of the disturbance. The accepted sources of Atlantic disturbances are vortices originating over Africa and crossing the Atlantic in the trade winds. In the western North Pacific large eddies form in the pressure troughs associated with the Asian monsoon. Central Pacific disturbances probably arise from several sources.

An important issue is that all of the above conditions must be satisfied for a tropical cyclone to form. Tropical cyclones are rare. An average of eighty systems per year form in all the tropical cyclone basins combined. Assuming an average life cycle of five days only one storm exists on any given day in the global tropics.

THE MODERN ERA OF Hawaiian/CENTRAL PACIFIC HURRICANES

Samuel Shaw of the Central Pacific Hurricane Center (CPHC) prepared a comprehensive history of tropical cyclones known or inferred to pass near or over the Hawaiian Islands from 1832 through 1979 (Shaw, 1981). Subsequently CPHC has produced annual summary reports. The latter reports are the most detailed available for central Pacific hurricanes. Tracks and intensities for all storms since 1950 are available from the National Climatic Data Center.

The first "official" hurricane in Hawaiian waters was Hiki in August 1950. Hiki moved parallel to the windward coasts of the islands. The southern semicircle of a westward moving hurricane is the most benign sector. The islands experienced sustained winds of tropical storm force -- 68 mph at Kilauea Point, Kauai and 50 mph at Niihau. In addition to some damage to lightly-constructed housing, heavy rains caused substantial flooding of the Waimea River. Fifty-two inches of rain fell in 96 hours at Kanalohuluhulu Ranger Station. (Shaw, 1981).

Between 1950 and 1959 seventeen tropical cyclones (of all intensities) were identified in the central Pacific. Twelve occurred in two years (1957 and 1959). In contrast Shaw and his collaborators could only document nineteen storms between 1832 and 1949.

Shaw did an admirable job of searching for records of storms after the arrival of the missionaries. There are fundamental problems with the old records:

1. Initially there were few written records. Only the missionaries kept records and they (the missionaries) were not ubiquitous.

2. There were no measurements of the key parameter, wind speed.

3. Hawaiians had no word for hurricane. David Malo (1843) described a variety of Kona winds.

4. November hurricanes could be confused with the subtropical cyclone or Kona Low ( Simpson, 1952).

5. Hawaiian structures would blow down at relatively low wind speeds. This is consistent with the structures of their Polynesian brethren in the South Pacific.

6. Mission houses were based on New England architecture and could withstand most extreme winds.

7. Many written records remained unexplored. Tsunami researchers continue to search these records. Since Hawaiians also have no word for tsunami, tsunami researchers search for references to high waters, and then attempt to determine if they were due to storms or tsunamis.

8. Storms passing near but not over the islands may or may not have been encountered by ships.

TIROS 1 was launched on 1 April 1960. During the 1960's a number of storms were identified in post analyses by research scientists such as Col. James Sadler of the U.S. Air Force and University of Hawaii. Between 1960 and 1969, 34 tropical cyclones were identified in the central Pacific. By 1970 operational agencies were routinely using satellite observations to identify tropical cyclones. Between 1970 and 1979, 34 storms again entered or formed in the central Pacific. During the 1980's the number increased to 54 storms. During the first three years of the 1990's the annual storm numbers have been consistent with the 1970's and 1980's. Between 1970 and 1992, 106 tropical cyclones have affected the central Pacific hurricane basin. The annual average is 4.5 storms (see Table One).

Examination of decadal data since 1950 suggest the following:

1. A doubling of storm reports after the launch of the first meteorological satellite in 1960.

2. Stable numbers of storm reports for two decades.

3. A 60 per cent increase in storms in the 1980's.

The 60 per cent increase in the 1980's is puzzling. The first three years of the 1990's continue the trend (18 storms). Possible explanations are either (1) climate fluctuations or (2) changes in analysis procedures. Technology and procedures have changed.

Notable Events

The most notable modern hurricanes for Hawaii have been:

1. Nina, November 29-December 1, 1957

2. Dot, August 4-6, 1959

3. Iwa, November 23, 1982

4. Estelle, July 23-25, 1986

5. Iniki, September 10-11, 1992

Table Two lists characteristics of these five as well as other storms mentioned in the text.

Nina (Shaw,1981)

Normal hurricane season runs from June through October. Nina was an "off-season" (November 29) hurricane. It formed in an unusual location (near Palmyra Island in the Line Islands south of Hawaii). Nina moved north within 120 miles west-southwest of Kauai (Figure 5) before turning westward. Nina produced heavy rains and floods with winds of up to 92 mph at Kilauea Point, Kauai. Damages were approximately $100,000 (1957 dollars) primarily due to high surf on the south shore of Kauai. The only other island affected was Oahu which had strong trades exceeding gale force. Honolulu International Airport recorded an all-time record sustained wind of 65 mph on 30 November.

Dot (Shaw,1981)

Dot (the Statehood hurricane) was first discovered by merchant ships on 1 August, 1959. At 1400 Hawaiian standard time the S.S. Sonoma reported extraordinarily low surface air pressure and winds of 90 mph. Aircraft monitored Dot as it approached the islands. Peak winds were between 150 and

Figure Five: Tracks of the four most important hurricanes in Hawaii since 1950. Tracks are limited to the vicinity of the islands.

165 mph on the afternoon of 2 August. The variations were due to estimation

techniques. Either value makes Dot the most intense storm in the modern history of the central Pacific hurricane basin. On 5 August Dot passed 90 miles south of South Point with peak winds of 132 mph (Figure 5). Dot turned to a northwest track and eventually moved near due north over Kauai on the night of 6 August.

Hawaii, Oahu and Kauai were all affected, with the most damage on Kauai. The estimated damage on Kauai was $6 million (1959 dollars), primarily to agriculture. Sustained winds at Kilauea Point lighthouse were 81 mph with gusts of 103 mph. Damage to single family residences in Kilauea, Lihue and Lawai ranged from minor to severe. Agricultural damages were due to wind (sugar and macadamia) and flooding (pineapple). Wave damage was limited. Hawaii suffered damage due to flooding and surf. Oahu suffered flooding and spot wind damages.

Dot had a large eye of 35 to 40 miles diameter. The best track placed the eye west of Lihue. The total area of the eye (using 35 mile diameter) was 962 square miles. Kauai by comparison is only 553 square miles.

Iwa (Chiu et al, 1983)

Although hurricanes sporadically brushed the Hawaiian Islands between 1959 and 1982, no major damages occurred until 1982 when Hurricane Iwa formed in mid-November west of the Line Islands and followed a course parallel to that of Nina only to turn northeast and accelerate, brushing Kauai on the afternoon of 23 November (Figure 5) and producing a strong squall line which struck Oahu that evening. At its peak intensity Iwa had sustained winds of 92 mph. As it moved northeastward it was weakening and we (Chiu et al) found no observed winds meeting the definition of "hurricane" on either Kauai or Oahu. Peak sustained wind at Lihue was 73 mph (adjusted to a 30 foot reference height from 64 mph at 20 feet).

Nevertheless the damages were far greater than with Dot twenty-three years earlier. Final damage estimates were $239 million. Primary losses were no longer suffered by agriculture but by the tourism industry which had developed since statehood and by the residents of the substantially larger 1982 population. Surge and overwash on the south shore of Kauai exceeded the 100-year inundation levels developed for tsunamis. Heaviest hit areas were the Poipu resort area and the Princeville Resort on the north shore. Princeville suffered from enhanced winds funneling through the mountain ridges immediately inland.

Oahu suffered from surge and overwash primarily along the leeward coast and from winds in central and windward Oahu. The winds were topographically enhanced by the Waianae Mountains and Kolekole Pass in the first instance and the Koolau Mountains in the second. As the squall line generated by Iwa traversed Oahu, it produced peak winds sequentially at Barbers Point, Honolulu International Airport, Wheeler Field and Kaneohe Marine Corps Air Station within a 45 minute period. Iwa's most remembered impact on Oahu was the island-wide power blackout due to the loss of the major transmission lines in the Koolaus.

Estelle (CPHC Staff, 1987)

Estelle entered the Central Pacific on 21 July 1986 and followed a steady track of west northwest until it reached 17^oN and then moved directly west, passing within 120 nm of South Point on 23 July. Peak winds were 86 mph near the center. Winds along the Puna and Ka'u coasts were northerly gusting to gale force. Estelle gradually weakened as it moved west. Juxtaposition of copious moisture brought to the islands by Estelle and a trough in the upper troposphere yielded heavy rains on Oahu on 24 and 25 July.

Estelle's major impact was on the Puna coast of Hawaii. The combination of storm motion and wind-generated swell caused heavy surf along the southeast coasts of Hawaii and Maui. At Vacationland Estates near Kapoho five houses were totally destroyed and others severely damaged. The total cost exceeded two million dollars, making Estelle the most expensive tropical storm to date for the Island of Hawaii.

Iniki (still being analyzed)

Iniki was the most intense storm to strike the Hawaiian Islands in the modern era. Iniki formed from tropical depression 18-E near 140^oW on 6 September (Figure 5). It passed 250 miles south of South Point with sustained winds of 100 mph. It deepened and began to assume a north northwest course as it came under the influence of a mid-tropospheric trough west of the islands. By the morning of 11 September Iniki had sustained winds of 145 mph and was moving north toward Kauai, Oahu and the leeward Hawaiian Islands. After undergoing an apparent wobble of its eye (leading to a heightened warning level for Oahu), the eye of Iniki went inland at Waimea, Kauai at 1530 Hawaiian Standard Time, exiting near Haena.

At landfall Iniki was compact and intense. Peak winds were 130 mph with gusts of 160 mph. Observed winds at Lihue peaked at sustained winds of just below 100 mph (again adjusted to 30 feet). The eye was 10 miles in diameter (contrast to Dot above). Kauai was devastated. Whereas Iwa produced pockets of damage, Iniki produced uniformly widespread damage. Surge and overwash once more produced excessive high water levels along the south shore. Although the actual surge (measured by a tide gage at Port Allen) was only five feet, high water marks reached over twenty feet with the highest at twenty-eight feet.

Oahu suffered from overwash and some limited winds. Damages were primarily to the leeward coasts, with pockets elsewhere. The relative lack of wind on Oahu was due to the compactness of Iniki. The radius of hurricane force winds was 50 nautical miles (57.5 statute miles).

Economic losses were staggering. Insured losses were $1.6 billion, while government and agricultural losses increase the estimates to over $3 billion. Kauai is far from recovery as we write this report.

The four storms described above were the major events of the past 43 years gleaned from what we consider "reasonably" reliable records. During this period the meteorological community gradually became more aware of and responsive to the threat to the islands. Economic losses are not a good yardstick for comparison of storms. Kauai in 1959 was quite different from Kauai in 1982 or 1992. Populations and infrastructure had changed. Simple inflation affects comparisons. The Iwa losses project to $500 million in 1992 dollars.

Additional Events (Shaw,1981; CPHC annual reports)

We shall next briefly mention recent storms which point to the variety of effects weak storms and near-miss storms have on the State.

Fico (1978)

Fico set a record by maintaining Hurricane Intensity for seventeen days and was tracked by meteorologists for 5000 nautical miles. Fico passed 175 miles south of South Point on 20 July and tracked west northwest away from the State (Figure 6). Its peak intensity was 115 mph as it passed South Point. The strong pressure difference over the islands due to Fico's low central pressure, juxtaposed with the normal July tradewind high north of the islands, caused trades to gust to near 60 mph. The winds downed trees and caused power outages. Fico also brought damaging surf to the Puna district of Hawaii. Fico was one of a bumper crop of central Pacific storms in 1978. A total of thirteen storms or remnants formed in or entered the basin.

Susan,(1978)

While Fico was interesting, Susan was an even greater threat to the islands and also suffered a spectacular demise. Susan formed southeast of Hawaii (Figure 6) on 18 October. By 21 October Susan had winds of 138 mph which placed it among the three strongest central Pacific hurricanes to that time. As Susan approached Hilo on 23 October, hurricane warnings were posted for the Big Island. At this time Susan was the most intense storm to threaten the islands. (Note: Dot was more intense but not while in the vicinity of the islands.) Susan collapsed rapidly as she entered a region of strong winds aloft (shear). The circulation literally split. In satellite imagery the upper-level cirrus blew off to the northeast leaving an exposed lower circulation which filled rapidly and drifted west. The spectacular end of Susan points the role that upper-level winds near the Hawaiian Islands often play in weakening storms approaching from the east.

Uleki (1988)

Uleki formed in the central Pacific in late August 1988 (Figure 7) and moved along a standard west northwest track until it passed South Point. On the morning of 2 September Uleki slowed and turned north and even briefly northeast. After this hesitation it resumed a west northwest course. At the time of its turn Uleki had winds of 120 mph and was aimed at Oahu. The threat extended to Maui. Uleki was a serious though short-lived threat to the central main islands. Uleki's stall and turn is typical of a hurricane which briefly feels the effects of a mid-latitude trough passing to its north. Hurricane Elena (1985) behaved similarly, threatening the Gulf of Mexico from Houston to Tampa over Labor Day weekend of 1985 (Sparks et al, 1991).

Fefa (1991)

Fefa was an eastern North Pacific hurricane which died as it approached Hawaii in August 1991. It was interesting because the decaying circulation reached the Big Island and produced strong thunderstorms which caused local flooding and rare injuries due to lightning strikes.

Figure Six: Tracks of Susan and Fico, 1978. Fico set a record for duration at hurricane intensity. Susan posed a serious threat to Hilo prior to a dramatic demise. Solid-to-short-dot transition reflects change from hurricane to depression as Susan encountered strong shear of winds aloft.

Figure Seven: Track of Uleki, 1988. Uleki posed a serious threat to Maui through Kauai during a brief recurvature.

Fabio (1988)

Fabio was a weakening eastern Pacific storm which passed south of the Big Island on 3 August 1988. Fabio is interesting because even though the center was far south southwest of the island of Hawaii, heavy showers erupted along the Hilo and Hamakua coasts causing local flooding. The message here is that a storm center need not be strong or even close for the islands to experience impacts.

ISSUES AND INTERPRETATION

We have presented a history of the modern era of Hawaiian hurricanes. Our findings are that the numbers of reported storms have increased in part as a result of technology (e.g. weather satellites) and partially due to increased awareness. If we consider 1980 through 1992 as representative of current activity, an annual average of 5 storms form in or enter the central Pacific hurricane basin. Some threaten Hawaii; a few actually strike.

We shall briefly discuss interannual variability in genesis regions and storm frequency. We specifically shall address the role of the El Nino Southern Oscillation (ENSO) in Hawaii's hurricane activity.

1. Genesis of storms. The primary source of central Pacific hurricanes are disturbances which form in the eastern North Pacific and move westward steered by the winds in their surrounding environment (Figure 2). Occasionally storms form near 140^oW and have a shorter approach to the islands. Examples are Kate (1976), Susan (1978) and Iniki (1992).

The unusual episodes are storms such as Nina (1957) and Iwa (1982) which form south and west of the islands then move north. They form later in the year (November) and nearer to the equator. Both 1957 and 1982 were the onset years of warm episodes of ENSO. As a warm phase of ENSO develops, equatorial wind patterns change and a shear zone forms between equatorial westerlies and subtropical easterlies. This shear zone is a source of cyclonic disturbances which can grow into hurricanes. During ENSO onsets tropical storms gradually form further and further eastward from the west Pacific into the central Pacific. Kate (1976) occurred under similar circumstances. In the winter of 1992 a weak storm (Ekeka) struck the island of Palmyra. The essential feature of these episodes is NOT THE WARM WATERS NORMALLY ATTRIBUTED TO ENSO but rather the CHANGES IN ATMOSPHERIC CIRCULATIONS. The waters south and east of the islands are nearly always warm enough for tropical cyclone formation. The central Pacific normally lacks an efficient source of initial disturbances.

ENSO warm phases have corresponded to some of the largest annual storm counts in the central Pacific but the relationship is not unique. 1972, 1982, and 1992 were warm phase years and major storm years. 1978 was not a warm phase year but still had as many storms as the warm phase years. 1977 was a warm phase year; the central Pacific storm total was zero.

2. Storm tracks. The most typical storm track near Hawaii is toward the west northwest. This is consistent with the large scale winds which steer the storms. Hawaii lies at a longitude near to that of the center of the subtropical high which drives the trades. As storms pass Hawaii they will naturally curve to the northwest unless the high extends unusually far to the west. In the trades, winds turn to the south with height contributing to a southeasterly steering wind. In the upper troposphere the winds over the islands are southwesterlies and contribute to north turns as well. Thus the winds generally favor the "typical" track.

The unusual storm tracks correspond to breakdowns in the standard wind patterns. The typical mechanism is a trough deepening at 20,000 feet above sea level to the west of the islands. This was the case for Iwa and Iniki. Here the storms took northward tracks toward Kauai. Iniki drifted northward until its demise; Iwa rode along in the mid-latitude westerlies and caused showers to San Francisco Bay.

3. Storm intensity. Storm intensity responds to changes in development conditions. Specifically these are changes in sea surface temperature (SST), stability and shear.

SST is HIGHLY OVERSOLD as a control on intensity of storms approaching Hawaii. The normal SST pattern has isotherms running almost exactly east to west across the Pacific to Hawaii and then curving northward west of the State. If storms drift sufficiently far north, SST can become a factor. However in one instance a storm reintensified well north of the islands (Cochran, 1976). SST is a necessary but not sufficient condition for hurricane intensification. Other than as a minimum condition for cyclogenesis, SST has no known correlation with any tropical cyclone parameter.

If a storm drifts northward into the trades, it will encounter cooler air (northeast winds) and hence more stable air which can be entrained into the storm core and weaken the circulation. This can often happen near the islands. Generally, storms need to get well north of the islands for this effect to be dramatic.

The most common factor contributing to the demise of storms near the islands is vertical shear of the horizontal winds. The story of Susan (1978) is a prime example of the role of shear. Winds of 35 mph aloft are sufficient to ventilate and weaken a hurricane. Geographically, Hawaii is fortunate in that winds aloft are generally westerly and near the right magnitude. Storms which are on tracks further south have lower probabilities of encountering unfavorable shear.

IS ANY ISLAND IMMUNE?

Our experience indicates that hurricane strikes on the Hawaiian Islands seem to be quite rare. Near misses with varying degrees of impact are more frequent (Figure 8). Some common arguments that circulate outside the scientific community include:

1. "Storms from the east are no threat."

2. " The Big Island diverts storms."

3. "Island ________ has never been hit."

4. "Kauai is the most likely target."

We shall discuss each of these.

1. Storms from the east are a threat. Storms from the east must fight a hostile environment to reach the islands. The standard defense for the state is the shear in the upper-level westerlies in the central Pacific. However storms do make it. An unnamed tropical storm moved over the Big island on 7 August 1958; the remnant of Tropical Storm Irah moved southwest through the Molokai Channel on 17 September 1963 as Tropical Depression #31; a storm which Shaw (1981) named the Kohala cyclone crossed the Kohala Mountains of the Big Island and recurved over Maui on 9 August 1871.

2. The Big Island may divert storms, however no physical mechanism or documentary proof has ever been offered. The Big Island represents a target only 60 miles wide to a storm on a southeast to northwest track so it is a very small target. Many weaker systems have moved over the Big Island from the

Figure Eight: Diagram capturing all tracks of tropical cyclones passing within 3^o Latitude of the islands between 1950 and 1992. Major hurricanes are named.

east. The Kohala cyclone was described as " a tornado" by eyewitnesses. The reality is that steering currents can drive a storm directly into the Big Island as well as any other Hawaiian Island. Larger islands such as New Caledonia, Luzon, Mindanao and Taiwan do not divert tropical cyclones. The belief that the Big Island diverts storms is exactly that , a belief, with no support in fact. It is a poor basis for public policy planning.

3. The record indicates that every island has felt the impacts of tropical cyclones. Since the history is so fragmentary it is inappropriate to assume any island is at less risk.

4. Kauai has had some very bad luck. Luck is an appropriate term. Consider the case of Iniki. If Iniki had started its northward turn 6 hours earlier Oahu would have been devastated. The turn was due to the arrival of the trough at 20,000 feet west of the State. A delay of six hours in the turn would have led to a storm track well west of Kauai.

The conclusions are:

1. Every island has been affected. Some have had recent direct hits.

2. No island is without risk from hurricanes.

3. The randomness of nature plays a role. Uleki(1988) was poised to hit Oahu or Maui; Iniki could have hit Oahu or missed all the islands.

Once every three years the center of a storm of hurricane intensity approaches within a day of one of the islands. Often the storm is passing south of the Big Island and actually poses little threat, but recurvature, such as with Iniki, or threatened recurvature, such as with Uleki, makes the threat greater. Meteorologists and civil defense planners emphasize recurving storms because such storms have the greatest potential for major surge and wave damage to vulnerable leeward coastal resorts and major population centers.

THE FUTURE

The current interest in global environmental change especially that due to Greenhouse warming leads to concerns about frequencies and intensities of tropical cyclones everywhere. The issues include:

1. Sea level rise and surge;

2. ENSO frequencies; and

3. the effects of a warming ocean.

The following paragraphs briefly comment on these issues.

1. Sea level rise. Estimates of sea level rise have dropped consistently as knowledge of the behavior of ice and the oceans has grown. Any rise in sea level is serious, however, in that the overwash region will extend further inland. The extent of increased inland penetration will depend upon slope (Fletcher,1992).

2. ENSO frequencies. Some General Circulation Model (GCM) simulations of a Greenhouse-warmed world suggest that ENSO warm phases would be more frequent. While all such models are still very primitive, increased frequency of warm events is significant for Hawaii. Storms forming near the Line Islands in off seasons would be more frequent. These storms have a good chance of reaching the islands.

3. Sea surface temperature (SST) effects. As stated before, this is an oversold concept. Efforts to relate SST to genesis have failed. Several studies have tried (Evans,1990;Raper,1993). The best air-sea interaction model of cyclogenesis (Emmanuel,1986) does state that the theoretical maximum intensity attainable by a hurricane does depend upon SST as one constraint. Very few hurricanes reach their theoretical maximum intensity because the other factors involved (see earlier discussion) are absent or diminished.

If SST does increase in the Greenhouse world of the future, the theoretical limit on hurricane intensity should increase. This is not equivalent to a prediction of stronger storms; it is only a statement of potential. Hurricane formation requires that a number of conditions be met. Hurricane formation is rare. Intense hurricanes are even rarer. Iniki did not reach its theoretical potential intensity. Dot in 1959 approached this limit early in its life.

CONCLUSIONS

We should assume that hurricane activity near Hawaii will persist at the levels seen in the 1980's. Hurricane threats will be frequent; actual strikes rare. The consequences of a direct strike on major population centers such as Oahu necessitates planning based on the eventuality of an Iniki type storm. Research on ENSO cycles, global warming and hurricane interaction with islands (Smith and Smith,1993) will provide additional information in the next decade. This information will assist in long-term planning.

REFERENCES

Chiu, Arthur N.L., Luis E. Escalante,J.Kenneth Mitchell, Dale C. Perry, Thomas A. Schroeder and Todd Walton,1983. Hurricane Iwa, Hawaii, November 23, 1982. Committee on Natural Disasters, National Research Council, National Academy Press, Wash,D.C., 129pp.

Cochran, Donald R., 1976. Unusual tropical development from a mid-Pacific cold low. Satellite Applications Information Note 2/76-1, Feb. 1976, 4pp.

Emmanuel, Kerry A., 1986. An air-sea interaction theory for tropical cyclones. Part 1: steady-state maintenance., J.Atmos.Sci., 43(6), 585-604.

Evans, Jenni L.,1990. Envisaged impacts of enhanced greenhouse warmings on tropical cyclones in the Australian region. CSIRO, Div. of Atmos. Res., Tech. Paper 20, 31pp.

Fletcher, C.H. III, 1992. Sea-level trends and physical consequences: applications to the U.S. Shore. Earth-Science Reviews, 33, 73-109.

Frank, William M. , 1987. Tropical cyclone formation. In, A Global View of Tropical Cyclones, R.Elsberry, Ed., U.S. Naval Postgrad. School, Monterey, Ca., pp. 53-90.

Malo, David, 1843. Hawaiian Antiquities, Honolulu, Bishop Museum Press

Raper, S.C.B.,1993. "Observational data on the relationships between climatic change and the frequency and magnitude of severe tropical storms." In, Climate and Sea Level Change, Warrick,R.A., E.M.Barrow and T.M.L.Wigley, eds. Cambridge, Cambridge University Press, pp. 192-212.

Ramage, C.S., 1974. The typhoons of October 1970 in the South China Sea: intensification, decay and ocean interaction. J.Appl.Meteorol., 13, 739-751.

Shaw, Samuel L. ,1981. A History of Tropical Cyclones in the Central North Pacific Ocean and the Hawaiian Islands 1832-1979. National Oceanic and Atmospheric Administration, 137 pp.

Simpson, Robert H., 1950. Hiki--Hawaii's first hurricane of record. Weatherwise, 3(6):127-134.

Simpson, Robert H. , 1952. Evolution of the Kona Storm, a subtropical cyclone. J.Meteorol., 9(1),24-35.

Smith, Ronald B. and David F. Smith, 1993. Hurricanes and Mountainous Islands. Preprints, Ninth Conference on Atmospheric and Oceanic Waves and Stability, San Antonio, American Meteorol. Soc., 37-40.

Sparks, Peter, Earl J. Baker, James Belville, and Dale C. Perry, 1991. Hurricane Elena, Gulf Coast August 29-September 2, 1985. Natural Disaster Studies, Vol.Two, National Academy Press, Washington, D.C., 121pp.

Staff, Central Pacific Hurricane Center, 1980-1992. Tropical Cyclones-Central North Pacific. A series of annual reports highlighting each year's tropical cyclone activity.

Wang, Shih-Ting, 1989. The characteristics of typhoon wind field and methods of gust wind prediction in Taiwan. National Science Council Technical Report 77-58, Taipei, Taiwan.