Tropical Cyclones
Main Index Home Page Stock Weather Photos Australian Severe Weather Forum Storm News and Storm Chasing Reports Tropical Cyclones / Hurricanes / Typhoons Weather Data and Links Wild Fires / Bushfires Weather Observation Techniques Weather Picture Catalogue Tornado Pictures and Reports Stock Video Footage and DVDs for sale
Monthly Global Tropical Cyclone Summary September 2002
[Summaries and Track Data] [Prepared by Gary Padgett]

                              SEPTEMBER, 2002

  (For general comments about the nature of these summaries, as well as
  information on how to download the tabular cyclone track files, see
  the Author's Note at the end of this summary.)


                           SEPTEMBER HIGHLIGHTS
  --> Record number of September tropical storms form in Atlantic
  --> Two Caribbean hurricanes strike western Cuba and U. S. Gulf Coast
  --> Devastating hurricane strikes Mexico's Yucatan Peninsula
  --> Years 7th super typhoon forms in Northwest Pacific--strikes Japan
  --> Southwest Indian Ocean tropical depression damaging to Seychelles
  --> "Subtropical wanderer" becomes the Atlantic's third longest-lived
      tropical cyclone on record


              ***** Feature of the Month for September *****
                           2002 - 2003 SEASON

     The Australian Bureau of Meteorology maintains three Tropical
  Cyclone Warning Centres (TCWC):  Perth, Western Australia; Darwin,
  Northern Territory; and Brisbane, Queensland.  Each centre is allotted
  a separate list of tropical cyclone names for tropical cyclones forming
  within its area of responsibility (AOR).  In addition a TCWC located at
  Port Moresby, Papua New Guinea (PNG)--a former Australian territory--
  maintains a list of native names to assign to the very rare tropical
  cyclones which form within its AOR.

     The AORs of the respective centres are:

  (1) Perth - 125E westward to 90E and south of 10S.  Currently, and for
      at least the next few years, the Perth TCWC will issue warnings for
      any systems north of 10S and south and west of the Indonesian islands.

  (2) Darwin - 125E eastward to 138E and extending northward to the
      equator.  There is a little irregularity with the eastern border
      in the Gulf of Carpentaria.  The Darwin TCWC issues High Seas
      Warnings for the entire Gulf of Carpentaria, but Brisbane issues
      Tropical Cyclone Advices and names cyclones in the eastern portion
      of the Gulf.  Also, currently, and for at least the next few years,
      the Darwin TCWC will issue warnings for any systems west of 125E
      and within the Indonesian archipelago in the Banda, Flores, and
      Java Seas.

  (3) Brisbane - 138E eastward to 160E and generally south of 10S.  The
      northern border with the Port Moresby AOR is somewhat irregular.

  (4) Port Moresby, PNG - immediate vicinity of the island of New Guinea
      and eastward to 160E generally north of 10S although the southern
      border is somewhat irregular.

     Names for the 2002-2003 season (** indicates name has already been

          Perth          Darwin        Brisbane        Port Moresby

         Fiona          Craig           Erica            Epi
         Graham         Debbie          Fritz            Guba
         Harriet        Evan            Grace            Ila
         Inigo          Fay             Harvey           Kama
         Jana           George          Ingrid           Matere
         Ken            Helen           Jim              Rowe
         Linda          Ira             Kate             Tako
         Monty          Jasmine         Larry            Upia (See Note)
         Nicky          Kim             Monica
         Oscar          Laura           Nelson
         Phoebe                         Odette
         Raymond                        Pierre
         Sally                          Rebecca
         Tim                            Sandy
         Vivienne                       Tania

  NOTE: I must admit I really don't know what to expect with regard to
  naming of tropical cyclones by the Port Moresby TCWC.  The list given
  above is the one which has been advertized for years, and is still
  included in the WMO Region V Operational Plan.  However, in May, 2002,
  the first tropical cyclone in Port Moresby's AOR in nine years formed
  and was named Upia--the last name on the list.  Ostensibly, that name
  will not be used again, but I really don't know what the next storm
  to develop in that region will be named.    We'll just have to wait
  and see.

                      and the SOUTH PACIFIC OCEAN

     The Tropical Cyclone Warning Centre (TCWC) at Nadi, Fiji, has
  tropical cyclone warning responsibility for the South Pacific east of
  160E and from the equator to 25S.   The Meteorological Service of New
  Zealand at Wellington has warning responsibility for waters south of
  25S, but almost all tropical cyclones in this basin form north of 25S.
  When a rare cyclone forms in the Wellington area of responsibility
  (AOR), it usually will be assigned a name from the Fiji list (such as
  was done for Tropical Cyclone Gita in February, 1999.)

     Tropical cyclone warning responsibility for South Indian waters west
  of 90E are shared by several TCWCs.       The Regional Specialty
  Meteorological Centre (RSMC) for the region is the office of Meteo
  France on the island of La Reunion.  However, following a long-standing
  practice, the sub-regional centres at Mauritius and Madagascar share
  the responsibility for actually naming tropical storms with Mauritius
  naming systems east of 55E and Madagascar covering the area west of
  55E.   RSMC La Reunion issues warnings for the basin independently of
  these sub-regional centres, but only advises regarding when or when not
  to assign a name to a developing cyclone.

     Names for the 2002-2003 season (** indicates name has already been

       Southwest Indian                          South Pacific

     Atang **          Noe                   Yolande **    Ivy
     Boura **          Opanga                Zoe           Judy
     Crystal           Pale                  Ami           Kerry
     Delfina           Qacha                 Beni          Lola
     Ebula             Rita                  Cilla         Meena
     Fari              Serame                Dovi          Nancy
     Gerry             Tina                  Eseta         Olaf
     Hape              Ulysse                Fili          Percy
     Isha              Vicente               Gina          Rae
     Japhet            Winston               Heta          Sheila
     Kalunde           Xena
     Luma              Yves
     Manou             Zaitoune

                            ACTIVITY BY BASINS

  ATLANTIC (ATL) - North Atlantic Ocean, Caribbean Sea, Gulf of Mexico

  Activity for September:  1 tropical depression
                           4 tropical storms
                           4 hurricanes
                           1 hybrid system

                        Sources of Information

     Most of the information presented below was obtained from the
  various tropical cyclone products issued by the Tropical Prediction
  Center/National Hurricane Center (TPC/NHC) in Miami, Florida:
  discussions, public advisories, forecast/advisories, tropical weather
  outlooks, special tropical disturbance statements, etc.    Some
  additional information may have been gleaned from the monthly
  summaries prepared by the hurricane specialists and available on
  TPC/NHC's website.     All references to sustained winds imply a
  1-minute averaging period unless otherwise noted.

     I'd like to extend a very special thanks to Kevin Boyle and John
  Wallace for their assistance in writing storm reports.  Kevin wrote
  the summaries for Edouard, TD-07, Josephine and Kyle, and a big chunk
  of the reports on Hanna, Isidore and Lili.   John, in addition to his
  Eastern Pacific duties, authored the summary on Tropical Storm Fay.  A
  big thanks is due also to Chris Fogarty for his report on Hurricane
  Gustav, and to Chris and Jack Beven for the information and comments
  they provided regarding the Newfoundland hybrid system of 5 September.

                 Atlantic Tropical Activity for September

     A new record was set in the Atlantic basin during September, 2002.
  For the first time on record, eight tropical storms formed during a
  calendar month.  Seven have occurred on several occasions:  August,
  1933; September, 1949; September, 1988; August, 1995; and September,
  2000.  There have also been a few occasions when nine storms developed
  during a 30-day period or less, but never with eight systems reaching
  tropical storm intensity within one calendar month.   Of the eight
  tropical cyclones, four reached hurricane intensity and two became
  intense hurricanes.  Over the period 1950-2001, the September averages
  for "named" storms, hurricanes, and intense hurricanes are 3.40, 2.40,
  and 1.25, respectively.  In addition to the eight tropical storms and
  hurricanes, another tropical depression formed but was quite short-
  lived and did not reach tropical storm intensity.  As the month of
  September opened Tropical Storm Dolly was located far to the east of
  the Lesser Antilles, struggling against hostile upper-level shear.
  The storm subsequently turned northward and had dissipated by the 4th
  far to the southeast of Bermuda.

     The pattern of short-lived cyclones of higher-latitude origin seen
  during July and August continued well into September.  Of the eight
  named cyclones forming in September, six were of non-tropical origin.
  Only Hurricanes Isidore and Lili, plus the non-developing depression,
  formed in the tropics from African waves.   Tropical Storm Edouard
  formed during the first week of the month off the northeastern Florida
  coast, made a loop, then moved west-southwestward, making landfall as
  a very minimal tropical storm near Flagler Beach.  The weakening
  depression continued across the northern peninsula and moved into the
  northeastern Gulf of Mexico where it dissipated.  At about the same time
  that Edouard was dissipating, a disturbance off the Texas coast
  intensified into Tropical Storm Fay.  After remaining quasi-stationary
  for a day or so, Fay moved westward into the mid-Texas coast and soon
  weakened to a tropical depression.  The remnant depression meandered
  over southern Texas and northern Mexico for several days, dropping
  hefty amounts of rainfall which led to substantial flooding.

     As Fay's remnants were weakening over Texas and Mexico, a subtropical
  LOW began developing southeast of North Carolina and became Subtropical
  Storm Gustav--the first subtropical storm to be named under TPC/NHC's
  new operational policy which was adopted in late 2001.  After a couple
  of days Gustav had acquired predominant tropical characteristics and
  was reclassified as a tropical cyclone.  The storm recurved just east of
  Cape Hatteras, then accelerated rapidly to the northeast.  Gustav reached
  hurricane intensity on 11 September as it was speeding northeastward--the
  latest date for the first hurricane of the year since 1941.  Gustav
  began transforming into an extratropical cyclone as it neared eastern
  Nova Scotia, but made landfall as a nominal hurricane.   Around mid-month
  weak Tropical Storm Hanna formed in the north-central Gulf of Mexico and
  made landfall along the northern Gulf of Mexico coastline.

     Hurricane Isidore formed in the Northwest Caribbean shortly after the
  middle of September and became the first of the year's two major hurri-
  canes on the Saffir/Simpson Scale.  Isidore crossed western Cuba as a
  Category 1 hurricane, then deepened into a major Category 3 hurricane
  as it progressed westward across the southern Gulf of Mexico.   On the
  22nd the storm made an unexpected hard left turn and moved inland into
  Mexico's Yucatan Peninsula near Merida.  The storm meandered about over
  land for a couple of days, weakening into a tropical storm.   After
  moving back out into the Gulf, Isidore began a northward trek which
  eventually carried it into the Louisiana coastline just west of New
  Orleans as a strong tropical storm.    At about the time that Isidore
  was developing in the Caribbean, Tropical Storm Josephine formed well to
  the east of Bermuda.  Josephine soon began moving northeastward and
  within 36 hours had merged with a cold front and lost tropical character-

     A few days after Josephine's formation, a subtropical LOW began taking
  shape in the same general area.  This system was dubbed Subtropical Storm
  Kyle on the 21st and was reclassified as a tropical storm the next day.
  Kyle was destined to wander aimlessly around the Western Atlantic sub-
  tropics for the next three weeks until it had become the Atlantic basin's
  third longest-lived tropical cyclone on record, after Hurricane Inga of
  1969 and Hurricane Ginger of 1971, both of which were also "subtropical
  wanderers".  Kyle did crank up into a 75-kt hurricane at one point while
  located east-southeast of Bermuda.     Subsequently, the cyclone went
  through cycles of weakening and re-intensifying--counting its initial
  classification as a tropical storm, Kyle was upgraded to tropical storm
  intensity five times.  The storm eventually moved westward, then turned
  northward, grazing the North Carolina coastline before heading out to
  sea and becoming extratropical.   As an extratropical cyclone, Kyle
  remained on the charts for several days, and at one point looked as if
  it had ideas of taking on tropical characteristics again.  Finally, on
  18 October the remnants of Kyle were absorbed into a strong extratropical
  cyclone southwest of the British Isles.

     The month's final cyclone was Hurricane Lili, which turned out to be
  the year's most intense hurricane.  Lili's track was remarkably similar
  to that of Isidore.  Lili, however, reached tropical storm intensity
  east of the Windward Islands before entering the Caribbean Sea.  After
  nearing hurricane intensity, the storm weakened briefly into a strong
  tropical wave south of Hispaniola.   However, Lili soon recovered, and
  after flirting with Jamaica for a couple of days, began a steady west-
  northwestward track toward the western tip of Cuba, which had been
  visited by Isidore only a week earlier.   The storm crossed western Cuba
  as a Category 2 hurricane on 1 October and entered the Gulf of Mexico
  on a northwestward heading.  On the 2nd Lili intensified dramatically
  into a severe Category 4 hurricane as it took aim on the central
  Louisiana coast; then, just as suddenly, weakened in the final hours
  before landfall so that it came ashore as a minimal Category 2

                Newfoundland Storm System of 5 September

     One additional system needs commenting on--I received an e-mail
  from Chris Fogarty on the evening of 5 September regarding a system
  which earlier that day had crossed Newfoundland.  Chris' comments
  follow (slightly edited):

     "Another one of those 'curious critters' formed over the Northwest
  Atlantic early on 5 September.   The storm started out as a mass of
  convection well south of the Grand Banks of Newfoundland.  The storm
  moved across the Avalon Peninsula of Newfoundland, producing heavy rain
  and gusty winds.  A hook-and-eye type feature formed on radar.  Satellite
  imagery showed a round cloud shield.  It seems to me this was a case of
  rapid cyclogenesis aided by low static stability and convection over the
  cyclone center."  In a later message Chris stated that the system was
  similar to the storm on 7 July (see July summary) which passed near Sable
  Island--a hybrid system embedded on a front.

     Chris had originally sent the e-mail, along with some satellite and
  radar images as well as sounding data, to Jack Beven also.   After I
  had sent out Part 2 of the September summary, Jack replied that he had
  finally had time to take a closer look at the system.  His comments

  "1. Our surface analysis (NHC's) had this LOW at the western end of a
      front for most of its life.  That might disqualify it from being a
      subtropical cyclone, but looking at the details of the analysis and
      at satellite imagery, the strength of the "front" is somewhat
      dubious.  It seems to me that much of the apparent temperature
      gradient is due to the sea surface temperatures and not to true air
      mass contrast.

  "2. The system moved northward in the warm sector of a baroclinic
      system to the west. That is a good sign, suggesting that there was
      little or no cold air intrusion into the system until it was north
      of Newfoundland.

  "3. The radar hook is quite interesting.  However, it doesn't fit very
      well with the visible satellite image from 10 minutes earlier, which
      suggests an exposed low-level center southwest of the radar hook.

  "4. The biggest problem I see in calling this a subtropical storm is
      that I can't find evidence of 35-kt sustained winds.  I've checked
      our surface analyses, I've (partly) checked our ship report files,
      I've checked QuikScat, and I can't find clear proof that the cyclone
      was of subtropical storm intensity.  I've also looked for formal
      satellite intensity estimates, and there are none."

     And now, Chris has looked further and unearthed some more observations
  on the LOW.     An automatic weather station at Cape Race did record
  sustained winds of 35 kts, with a MSLP of 1001.1 mb, at 1500 UTC on
  5 September.  This suggests that the system at least briefly was of
  gale strength.  Whether it was tropical or subtropical, given the high
  latitude and cool temperatures, is something which will require more
  study.  This system will not be added to the 2002 roster of tropical
  and subtropical cyclones at this time, but will be given further
  analysis later and could possibly eventually be added to the Atlantic
  Best Track database during the on-going reanalysis.  (A special thanks
  to Chris and Jack for the information and comments on this interesting
  little system.)

                        TROPICAL STORM EDOUARD
                           1 - 8 September

  A. Storm Origins

     Tropical Storm Edouard originated from an area of cloudiness and
  thunderstorms initially located about 475 nm east of the southern
  Bahamas.  This disturbance was noted on a STWO issued by NHC on
  29 August at 2130 UTC.  The system drifted westward during the following
  two days with little change.   However, a reconnaissance plane sent to
  investigate the disturbance on the afternoon of the 31st found a broad
  area of low pressure centred approximately 130 nm northeast of the
  northern Bahamas with a few squalls of 20 to 25 kts.  Conditions were
  favourable for slow development, and on 1 September, 2100 UTC, Air Force
  Reserves reconnaissance, satellite imagery, radar and surface obser-
  vations indicated that the disturbance east of the Florida East Coast
  had developed into the fifth tropical depression of the season.  The
  centre of Tropical Depression Five was located near 29.0N, 79.2W, or
  about 100 nm east of Daytona Beach, Florida.  Reconnaissance at 01/1700
  UTC found 850-mb flight-level winds of 35 kts, a broad LLCC, and a
  surface pressure of 1014 mb.   Also, Melbourne Doppler radar had
  occasionally been indicating 36-40 kt winds between 3000 and 4000 metres.

     Satellite images showed that the main area of deep convection was
  east and southeast of the centre, a result of westerly shear over the
  area.  The shear was light enough for some strengthening, and at
  02/0600 UTC, after a Hurricane Hunter aircraft had reported 450 m flight-
  level winds of 47 kts east of the LLCC and a central pressure of 1007 mb,
  Tropical Depression Five was upgraded to a tropical storm in an inter-
  mediate advisory.   Tropical Storm Edouard was located about 85 nm east
  of St. Augustine, Florida, and was moving slowly toward the east, perhaps
  in response to a mid-level shortwave trough passing to the north and
  also due to the LLCC following the deep convective bursts.

  B. Track and Intensity History

     Edouard stalled early on 3 September near 30.3N, 70.6W.  A dropsonde
  released at the 850-mb center found a central pressure of 1003 mb and
  southerly 30-kt winds at the surface, indicating that the mid-level
  circulation was displaced east of the LLCC.    Despite the continued
  presence of westerly shear, Edouard maintained vigorous deep convection
  in the eastern quadrant.  The MSW was raised to 55 kts at 03/1500 UTC
  after a reconnaissance aircraft reported a visual estimate of surface
  winds of 55-60 kts from 300 m and found flight-level winds of up to
  71 kts.   Edouard was at the time nearly stationary about 165 nm east
  of Jacksonville, Florida.  This represents Edouard's peak intensity
  with weakening beginning soon afterward as the system began to suck in
  dry air and to suffer the affects of 50-kt west to west-northwesterly
  vertical shear.    Edouard began to slowly drift west-southwestward,
  but became stationary again at 04/0300 UTC roughly 100 nm east-northeast
  of Daytona Beach.  There was patchy, deep convection to the east of the
  LLCC at this time, and Edouard was depicted as a well-defined swirl of
  clouds with intermittent deep convection for the rest of its career.

     A low-level ridge to the north of Edouard began to guide the 35-kt
  tropical storm westward toward the Florida East Coast.   Edouard made
  landfall just north of Daytona Beach between 0000 and 0100 UTC on 
  5 September.   Surface observations indicated that the system had
  weakened to a depression by this time and the MSW was dropped to 25 kts
  at 05/0900 UTC.     Six hours later, Edouard had crossed the Florida
  Peninsula and moved into the warm waters of the Gulf of Mexico.  However,
  shear was too high for the system to regenerate, and furthermore, Edouard
  began to interact with the pre-Fay disturbance, a second negative factor.
  The final advisory on Edouard was issued by NHC at 06/1500 UTC, placing
  the dissipating centre off the Northwest Florida coast south of Ft.
  Walton Beach.  The remnants were eventually absorbed into Tropical Storm
  Fay located off the Texas coast.

  C. Damage and Casualties

     Aside from localised flooding caused by heavy rains in Florida, there
  were no casualties or damage associated with Tropical Storm Edouard.

  (Report written by Kevin Boyle)

                          TROPICAL STORM FAY
                           4 - 11 September

  A. Storm Origins

     The disturbance that became Fay was noted in the eastern Gulf of
  Mexico as early as 31 August.  However, it did not organize until
  4 September, when it began to consolidate southeast of the Texas coastal
  bend.    Though it was still ragged, the first advisory on Tropical
  Depression Six was issued at 2100 UTC on 5 September when it was located
  roughly 85 nm southeast of Houston.  The classification as a depression
  was based on reconnaissance and offshore oil rig data.   Caught in a
  persistent col, the depression was stationary.   In fact, the tropical
  cyclone was eventually listed as stationary in six of the eight forecast/
  advisories that the NHC issued on it.     A tropical storm warning was
  issued for the Texas and Louisiana coasts.

  B. Track and Intensity History

     The depression was upgraded to Tropical Storm Fay at 0300 UTC on
  6 September after a reconnaissance plane found a 52-kt FLW--it was then
  located about 110 nm southeast of Galveston.   The Hurricane Hunters
  found no well-defined LLCC; Fay was upgraded largely as a cautionary
  measure.  Though conditions were favorable for strengthening, Fay's
  large circulation remained vague as it drifted ever so slowly westward
  south of a weak ridge.  One of the forecasters on duty noted similarities
  between Fay and Tropical Storm Frances of 1998, a storm that was
  similarly large and disorganized.  The amorphous storm intensified even
  while atmospheric conditions became less favorable, reaching 50 kts at
  1500 UTC on the 6th.  Amazingly, though reconnaissance aircraft measured
  winds as high as 68 kts and estimated a surface wind of 55 kts, they
  still found no closed center, making the cyclone's motion difficult to
  determine.  The NHC stated that satellite imagery indicated only a
  series of smaller vortices rotating inside a larger gyre.   Even so, a
  hurricane watch was hoisted for the Texas coast, and as Fay's upper-level
  situation temporarily improved, fears of a hurricane landfall increased

     At 0000 UTC on the 7th Fay's CP bottomed out at 998 mb, with the MSW
  estimated at 50 kts, while located 90 nm south of Galveston.  This was
  the peak intensity of Fay, which at this time looked more subtropical
  than tropical, an assertion corroborated by water vapor imagery taken
  near 1200 UTC on that day.  Fay did not as much make landfall as it did
  reform a new center that accelerated slightly and tracked west-northwest
  across the center of Matagorda Bay, near Palacios, Texas, around 1030
  UTC on the 7th.  The MSW was roughly 50 kts with a CP of 999 mb.  The
  Palacios weather station recorded a MSLP of 999 mb from roughly 0600
  UTC to 0900 UTC while the highest recorded MSW was only 32 kts at 0554
  UTC.   Well away from the center, Galveston experienced a peak MSW of
  31 kts from 0652-0752 UTC and a minimum SLP of 1004 mb at 0552 UTC.
  Port Lavaca, just across Matagorda Bay from Palacios, recorded a peak
  MSW of only 18 kts at 0551 UTC, and a MSLP of 1001 mb from 0936 UTC to
  1001 UTC.  Jamaica Beach recorded the only known storm surge measurement
  of 1.6 meters above the Mean Lower Low Water level.

     Fay quickly weakened to a depression after landfall, and turned more
  to the west as a ridge built to its north and west.  After 1500 UTC that
  day, advisories were issued by the Hydrometeorological Prediction Center.
  The ridge north of Fay pushed it to the southwest early on the 7th, but
  on the 8th it turned back to the west.    Fay's vortex became quasi-
  stationary southwest of San Antonio on the 8th, then began a slow,
  seemingly drunken wobble toward the U.S.-Mexico border.  The depression's
  center crossed the Rio Grande early on the 10th and continued on a rough
  south-southwesterly track.  Fay turned south-southeast later that day,
  executed a small cyclonic loop, then turned westward upon the issuance
  of the last advisory at 0300 UTC on 11 September.  The weak center was
  then located only about 15 nm south of Monterrey, Mexico.   By later
  that day, there was no trace of an organized circulation.

  C. Rainfall Reports

     The most noteworthy aspect of Fay by far was the rain it brought to
  central Texas, a region that had been hit hard by catastrophic floods
  only two months earlier.  Fay was not the disaster it was feared to be,
  but the rainfall amounts were robust, to say the least.   Totals were
  high across a large swath of central Texas.  The top five-day totals
  given in the final advisory from the Hydrometeorological Prediction
  Center are listed below (all figures are in centimeters and all
  locations are in Texas):

     Location                                  Amount (cm)

     Freeport/Dow Chemical                      34.44
     Jamaica Beach                              32.66
     Freeport                                   35.59 (as of 8 September)
     Pearsall 9 East                            30.48
     Mount Royal, San Antonio                   29.97
     Elk Runner, San Antonio                    28.17
     Kelly AFB, San Antonio                     27.58
     Zanzamora, San Antonio                     25.78
     Dilley                                     25.73
     Hondo                                      21.84
     Floresville                                19.13
     Galveston                                  18.87
     Granger                                    17.34

     The 90-hour storm total for San Antonio International Airport through
  0900 UTC on the 10th was only 12.73 cm; this is clearly unrepresentative
  of the rains most of South Texas experienced, let alone Bexar County
  (the county San Antonio is in).  According to the
  website, the town of Sweeny, Texas, in Brazoria County received 50.8 cm
  of rain.  Sweeny is about 100 km south of Houston.  Radar estimates from
  the National Climatic Data Center suggest isolated rainfall totals as
  high as 46 cm between Corpus Christi and Houston.

  D. Damage and Casualties

     Damage in Houston proper was apparently light, aside from a power
  outage affecting 27,000 people.    Galveston escaped with only a few
  downed limbs and signs.  In Surfside Beach, near Freeport in Brazoria
  County, Fay uprooted trees, mangled roofs, and damaged several homes
  and businesses in addition to causing flooding damage.

     Flood damage from Fay was substantial, and nine counties were declared
  federal disaster areas on 26 September.  That being said, the heavy rains
  were most welcome in the parched Rio Grande Valley, a region that has
  suffered a drought for many years.  Fay spawned at least three tornadoes
  according to the Austin American Statesman on 8 September--one in
  Wharton County destroyed a mobile home and severely damaged three others
  while one in Hungerford damaged a mobile home and a dumpster truck.
  Unfortunately, no specific monetary damage figures are available.  No
  casualties are known at present.

  E. References

  *FEMA Information>

  *90-Hour Rainfall Totals:>

  *Five-Day Rainfall Totals:>

  *Austin American Statesman>

  *HPC Tropical Archive>

  *September Monthly Summary (NHC)>


  *Rainfall Impact Info>

  *Radar Image>

  *Galveston Info (brief)>

  *Jamaica Beach Storm Info>

  (Report written by John Wallace)

                          TROPICAL DEPRESSION
                            7 - 8 September

     A small area of low pressure left the western coast of Africa in
  early September and moved west-northwestward over the open Atlantic for
  several days.  Tropical Weather Outlooks from NHC began mentioning the
  system around midday on 5 September when it was located about 1200 nm
  east-northeast of the Leeward Islands.  By early morning of the 7th the
  small but well-defined system was located approximately 1100 nm east-
  southeast of Bermuda.  The first advisory on Tropical Depression Seven 
  was issued at 1500 UTC, 7 September, when the system had developed
  enough convection near the centre to be classified as a tropical
  depression.  However, convective activity subsequently decreased and was
  sheared to the east of the LLCC.  TD-07 was a weak system, and after the
  initial three advisories was devoid of convection for the rest of its
  short life.  The tropical cyclone moved westward under a very weak mid-
  level ridge with a peak MSW of 30 kts.  The final advisory on TD-07 was
  issued by NHC at 1500 UTC on 8 September after new convection failed to
  develop and with the system forecast to turn northwards into a region
  of strong vertical wind shear.

  (Report written by Kevin Boyle)

                            HURRICANE GUSTAV
                            7 - 14 September

  A. Storm Origins

     A Tropical Weather Outlook (TWO) issued by TPC/NHC on the morning of
  6 September noted that an area of cloudiness and showers with a few
  embedded thunderstorms had developed and extended from the central
  Bahamas eastward and northeastward for several hundred miles.  The area
  of disturbed weather was associated with a large upper-level LOW and a
  surface trough.    The TWO noted that any tropical or subtropical
  development would likely be slow to occur.  The system appeared slightly
  better organized on the 7th and the TWOs from NHC began to emphasize
  the potential for tropical or subtropical cyclone development.  At 1800
  UTC a weak LOW center was located approximately 425 nm east of Miami.
  (This from the Navy JMV file sent to me by Michael Pitt.)  The LOW moved
  generally north-northeastward during the evening of the 7th and early
  morning of the 8th.

     Early on 8 September satellite imagery and surface observations
  indicated that the broad area of low pressure east of the northern
  Bahamas was becoming better organized and a reconnaissance plane
  was scheduled to investigate the system later in the day.  At 1200 UTC
  the broad center of the disturbance was located approximately 500 nm
  southeast of Cape Hatteras, and at 1500 UTC the first advisory on
  Subtropical Depression Eight was issued.  A change was made late in
  2001 to NHC's operational procedures to the effect that any subtropical
  depression or storm for which warnings were issued would be numbered
  sequentially with the next tropical cyclone number.   The second change
  made to the operational procedures was that when a system was designated
  as a subtropical storm (winds of gale force or higher), it would be
  assigned the next available tropical cyclone name.  The reconnaissance
  flight into the system found 40-kt winds at 450 m to the northeast of
  the center along with a 1006-mb central pressure, so the depression was
  upgraded to Subtropical Storm Gustav at 2100 UTC.  Since the forecast
  track brought Gustav's center almost to the Mid-Atlantic Coast within
  36 to 48 hours, a tropical storm watch was issued for portions of the
  North Carolina coastline.

  B. Track and Intensity History

     By 0000 UTC on 9 September Gustav's center was located about 335 nm
  south-southeast of Cape Hatteras, moving west-northwestward at 13 kts.
  As the day progressed Gustav continued to become better organized and
  very slowly began to look more like a tropical cyclone.  Baroclinic
  cloudiness to the east began to separate and move away from the cyclone
  and anticyclonic outflow became better defined.   Deep convection also
  became better established near the center with morning visible imagery
  revealing a partially-exposed center on the eastern edge of the deep
  convection.   Reconnaissance data, however, indicated that there was at
  best only a weak warm core at 850 mb and the radius of maximum winds
  remained on the order of 75-100 nm, so Gustav remained classified as
  a subtropical storm.   By 2100 UTC the storm's center had reached a
  point about 190 nm south-southeast of Cape Hatteras, and the west-
  northwestward motion had slowed to 8 kts.    An early afternoon
  reconnaissance flight found maximum flight-level winds of 53 kts, and
  reports from NOAA buoy 41002 indicated that the central pressure had
  dropped to 999 mb.  Based on this the MSW remained at 40 kts--where it
  had been pegged since 0900 UTC.

     Gustav had strengthened slightly by 10/0000 UTC--the MSW was upped
  to 45 kts based on a reconnaissance report of 55-kt winds at 450 m to
  the southwest of the center along with a central pressure of 996 mb.
  As of the 0900 UTC advisory Gustav still exhibited little evidence of
  a warm core and the system remained underneath an upper-level LOW.
  However, the discussion bulletin noted that there was evidence that the
  core was about to become better organized.   An intermediate advisory
  at 1200 UTC reclassified Gustav as a tropical storm based on aircraft
  reports that an inner core of strong winds was developing.  Tropical
  Storm Gustav was then located approximately 100 nm due south of Cape
  Hatteras, moving northward at 8 kts with maximum winds of 50 kts.  During
  the morning a reconnaissance plane found peak winds at flight-level of
  62 kts with a central pressure of 987 mb.    By afternoon convection
  had wrapped almost completely around the center and the central pressure
  had dropped to 984 mb.  The center of Gustav passed just east of Cape
  Hatteras during the afternoon and at 2100 UTC was located only about
  17 nm east-northeast of the cape.   The AWS at Diamond Shoals reported
  a SLP of 984.8 mb during the afternoon, and Cape Hatteras reported a
  wind gust of 68 kts around 2200 UTC.

     The 11/0300 UTC discussion noted that Gustav continued to appear
  better organized and looked more like a tropical cyclone.  A well-defined
  convective band had wrapped completely around the center, but there were
  not yet any indications of an eye forming.   The good news for the U. S.
  coast was that Gustav had become embedded in southwesterly flow ahead of
  a developing trough and was beginning to accelerate away from the main-
  land.  The center of Gustav at 0300 UTC was located approximately 105 nm
  northeast of Cape Hatteras and moving northeastward at 16 kts.  The MSW
  had increased to 55 kts and Gustav was still forecast to reach hurricane
  intensity before becoming extratropical.  The MSW was bumped up to 60 kts
  at 11/0900 UTC based on CI estimates of 65 and 55 kts from TAFB and SAB.
  However, shortly after the issuance of that advisory, a reconnaissance
  plane found 80-kt winds at 850 mb southwest of the center with a central
  pressure of 975 mb.  Visible imagery showed a well-developed tropical
  cyclone with convection wrapped all the way around the center.  Gustav
  was upgraded to a 65-kt hurricane at 1200 UTC, located about 225 nm
  southeast of New York City.   This is the latest date for the formation
  of the Atlantic season's first hurricane since 1941 when the first
  hurricane did not form until 18 September.     Other more recent late-
  appearing first hurricanes were Diana of 1984 (10 September) and Erin
  just last year (9 September).

     By the afternoon of 11 September Hurricane Gustav had accelerated and
  begun extratropical transition.  However, a reconnaissance aircraft
  around midday found a central pressure of 964 mb and peak flight-level
  winds of 104 kts, and the stepped-frequency microwave radiometer on the
  NOAA research aircraft measured 71-kt surface winds.  Based on this data,
  the MSW for Gustav was upped to 80 kts at 2100 UTC, and the forecaster
  writing the discussion remarked that this was possibly conservative.
  The hurricane was then located approximately 240 nm south-southwest
  of Halifax, Nova Scotia, and racing northeastward at 33 kts.   At 0300
  UTC on 12 September Gustav was near eastern Nova Scotia about 80 nm
  south-southwest of Sydney and moving northeastward at 40 kts.  The MSW
  was reduced to 70 kts, and the storm was beginning to look ragged as it
  began to lose its tropical characteristics.  The final NHC advisory on
  Gustav was issued at 12/0900 UTC.   The center was then estimated to be
  inland in Newfoundland about 60 km south-southwest of Stephenville.
  The storm was still of hurricane intensity, but was rapidly losing its
  tropical characteristics as it interacted with a non-tropical upper-level
  LOW.  The cloud top temperatures had warmed considerably and what little
  convection remained was displaced northeast of the LLCC.  The powerful
  storm subsequently turned northward and slowed as it moved off the east
  coast of Labrador.  By midday on the 14th the system had weakened into
  a 40-kt gale in the Labrador Sea about halfway between northern Labrador
  and southern Greenland.

  C. Additional Discussion

     Chris Fogarty, a Research Meteorologist with Environment Canada and
  currently a doctoral student at Dalhousie University, prepared and sent
  a report on Hurricane Gustav.  Portions of Chris' summary are included
  below.  A special thanks to Chris for sending me the summary and for
  permission to use it.

  (1) Storm Structure

     This event was handled rather well by the numerical models because 
  essentially it behaved like a strong extratropical system.    The
  approaching trough from the west was digging over New England and formed
  its own LOW over Maine.  The baroclinic energy from this was eventually
  transferred to Gustav, whose center remained intact as it made landfall.
  On a large scale the system appeared extratropical, yet had a tropical
  center which was rapidly becoming sheared and losing its warm core.
  Most of the rainfall was actually not related to the core rains of
  Gustav, but from the moist boundary between Gustav and the baroclinic
  development to its northwest.  Comparison of surface data, radar and
  satellite imagery suggest there was a large degree of tilt with Gustav
  such that the mid-level center was sheared to the east or northeast of
  the surface center.  This pattern also occurred during the similar 
  extratropical transition of Hurricane Michael in October, 2000.

  (2) Landfall Details
    At this time both agencies (NHC and CHC) are agreeing Gustav was 
  technically a hurricane at landfall.  There is some uncertainty as to 
  which community in southern Cape Breton should be classified as the 
  landfall point.  This owes to lack of surface data and difficulty 
  inferring the surface center from radar and satellite images.  There 
  was some degree of decoupling between the surface center and mid-level
  center.  The closest community would likely be St. Esprit at 1:30 am 
  ADT September 12th (12/0430 UTC) with a storm central pressure of 960 mb,
  moving northeast around 35 to 40 knots.   Hart Island on the southeastern
  tip of mainland Nova Scotia reported a sea level pressure of 961.4 mb
  at 45 minutes past midnight September 12th  (12/0345 UTC).  Hart Island
  was about 30 km north of the storm center.

  D. Meteorological Observations

  (1) Wind Observations
    Gustav was a very complicated system with strongest winds displaced
  well away from its center at the time of landfall.  Damaging winds
  occurred as far west as Maine where large trees were toppled.   Prince
  Edward Island seemed to take the brunt of the storm--not Cape Breton
  Island as one might expect.  There were also very high winds south of
  the storm track on Sable Island where gusts reached 66 knots with a
  sustained wind of 48 kts.  Over land, many stations recorded gusts over
  50 kts, and in a few cases gusts near 65 knots.  Winds did not appear
  to be very high near the center of the storm as it crossed Cape Breton.
  Sydney's winds did not really pick up until several hours after the
  storm passed, and barely gusted to 30 knots during its passage.  A
  Canadian buoy just north of the storm track reported gusts to 60 kts,
  and St. Paul's Island on the northernmost part of Nova Scotia recorded
  a peak gust of 66 kts.   Charlottetown's peak winds at the height of the
  storm were sustained 35 kts, gusting to 52 kts.  Finally, St. Lawrence,
  Newfoundland, recorded a peak gust of 70 kts.

  (2) Rainfall Observations
    There were very heavy rainfall amounts across the entire Atlantic 
  Canadian region.  Some of the highest amounts were along a swath which
  was left of and parallel to the storm track over Prince Edward Island
  and Central Nova Scotia.  A few localities received over 100 mm (4")
  of rain in a twenty-four hour period.  This is impressive considering
  how fast the storm was moving.  Below are some noteworthy storm-total
  rainfalls which fell over a 24-hour period:
  Lyon's Brook, Nova Scotia     108 mm - near New Glasgow
  Ashdale, Nova Scotia          105 mm - Hants County
  Liverpool, Nova Scotia        102 mm 
  Middleboro, Nova Scotia       100 mm - near Pugwash, northern Nova Scotia
  Halifax Airport, Nova Scotia   94 mm
  Charlottetown, PEI             70 mm
  Halifax (Shearwater)           56 mm
  Sydney, Nova Scotia            56 mm

  (3) Storm Surge

    Higher than normal water levels were experienced along all the
  coastlines of Prince Edward Island, northern and eastern Nova Scotia,
  and eastern New Brunswick.  Some localized coastal flooding was 
  reported from each of the three Maritime Provinces.  In particular, 
  the southern coast of Prince Edward Island narrowly missed a serious 
  surge event when the high point of the surge (over 1400 mm at 
  Charlottetown) occurred 3-4 hours after the high astronomical tide 
  of the day--at which point only a 700-mm surge was occurring.  Had the
  two events coincided, the total water level at Charlottetown would have
  likely matched the all-time record value and a significant flooding
  event would have resulted in downtown Charlottetown (similar 
  to or worse than the flood from the January, 2000, superstorm surge).

  E. Damage and Casualties
    Hardest hit was Prince Edward Island where whole trees were toppled
  and there was some local flooding.  Some docks were damaged in Eastern
  New Brunswick.  High-sided vehicles were prohibited from crossing the
  Confederation Bridge between Prince Edward Island and New Brunswick,
  and ferry service was suspended between Nova Scotia and  Newfoundland.
  There were scattered power outages from limbs and trees falling onto
  lines.  In the Halifax area there was only some leaf litter after the
  storm went through--winds did not gust much more than 35 kts.

    A report on Hurricane Gustav along with some pictures can be found
  on the website of the Canadian Hurricane Centre at the following link:>

  (Report written by Gary Padgett with significant contributions by
  Chris Fogarty)

                         TROPICAL STORM HANNA
                           12 - 16 September

  A. Storm Origins

     On 10 September surface observations indicated the presence of a
  broad area of low pressure located off the Texas coast.  The large area
  of disturbed weather covered much of the Gulf of Mexico and was first
  mentioned in a STWO issued at 10/1530 UTC.  Remarks in the 11/1805 UTC
  Tropical Weather Discussion noted that an upper-level LOW was combining
  with a 1008-mb LOW over the central Gulf of Mexico to produce an
  extensive area of showers and thunderstorms over the region which were
  spreading over the Florida Peninsula.  The area was considered a
  possibility for tropical development.  However, due to the proximity of
  the upper-level trough there was also the potential for subtropical
  cyclone development in a manner similar to the early stages of Hurricane

     The disturbance had formed into a depression by 12/0300 UTC while
  located near 26.5N, 86.5W, or about 240 nm south of Eglin AFB, Florida.
  An Air Force Reserves unit aircraft at this time found a CP of 1001 mb
  and peak 450 m flight-level winds of 40 kts.  However, the depression
  was not particularly impressive-looking in satellite imagery with little
  deep convection near the LLCC and little curvature in the convective
  bands to the south.  The initial movement was toward the north-northeast
  at 4 kts, but the system's motion was generally slow and erratic for
  several days.  A reconnaissance aircraft reported flight-level winds
  of 45 kts in the southeast quadrant shortly after 0300 UTC on the 13th,
  and a later pass through the system revealed 50-kt winds north of
  the centre.   These observations, along with two surface reports of
  35-kt winds, led to the upgrading of TD-09 to Tropical Storm Hanna at
  13/0900 UTC.   Hanna was then centred approximately 220 nm south-
  southwest of Pensacola, Florida, moving northwestward at 7 kts with
  a MSW of 40 kts.

  B. Track and Intensity History

     Early on 14 September Hanna was still meandering and the storm had
  changed little other than an increase in the MSW to 45 kts based on a
  reconnaissance report of 58-kt winds at 850 mb at 13/2346 UTC.  This
  wind observation was made about 85 nm southeast of the centre and was
  the only evidence for 45-kt surface winds.   This represents the peak
  intensity for Tropical Storm Hanna.   The storm existed in a sheared
  environment with most of the tropical storm-force winds located to the
  east and southeast of the exposed centre.     Forecast models were
  predicting Hanna to move toward the northeast in response to the south-
  westerly airflow of a mid-level trough, and by 14/0900 UTC Hanna had
  finally taken the anticipated northeastward track toward the
  Mississippi/Alabama coastline.  A reconnaissance aircraft reported a
  peak flight-level wind at 925 mb of 57 kts, so the MSW appeared to be
  holding steady at 45 kts, even though the exposed circulation was
  becoming deformed and elongated.

     Tropical Storm Hanna made landfall near the Mississippi/Alabama
  border around 1500 UTC on the 14th.  Shortly before landfall a
  reconnaissance plane measured a central pressure of 1002 mb and a peak
  flight-level wind of 59 kts just east of the centre, so the storm
  apparently maintained its 45-kt intensity until landfall.  Hanna was
  downgraded to a tropical depression at 1800 UTC while located in the
  vicinity of Mobile, Alabama, and the final advisory was issued at
  14/2100 UTC, placing the weakening centre about 65 km northeast of
  Mobile and moving northeastward at 11 kts.

     Responsibility for issuing information on Hanna's remnants was then
  assumed by HPC in Maryland.  The LOW was followed northeastward across
  the states of Alabama, Georgia, South Carolina and into eastern North
  Carolina.  The final storm summary from HPC was issued at 0900 UTC on
  16 September and placed the very weak centre near Hickory, North
  Carolina, or about 50 km northwest of Charlotte near 35.6N, 81.3W, and
  moving east-northeastward at 22 kts.

     What ultimately happened to the surface remnants of Hanna is somewhat
  of a mystery to the authors.  The MPC High Seas Forecast for 16/1800 UTC
  mentioned an inland 1014-mb LOW near 36N, 77W, moving east-northeastward
  at 20 kts.  Winds to 25 kts and seas to 8 ft were forecast within 240 nm
  in the south and southeastern quadrants.  Whether this was the same
  centre which had been Hanna or a new baroclinic development is uncertain,
  but the 1800 UTC position suggests it could have been Hanna.  However,
  the 17/0000 UTC forecast did not refer to this LOW, although it did
  forecast a LOW in 24-hours to be near 37N, 69W.  Because of this
  uncertainty, the track for Hanna in the accompanying cyclone tracks
  file ended with the final HPC summary.

  C. Meteorological Observations
     The Dauphin Island weather station reported a gust of 45 kts shortly
  before 14/1200 UTC, and Pensacola recorded a gust of 54 kts around 1500

     Hanna brought torrential rains to the Southeastern United States.
  Some selected precipitation totals from the HPC storm summaries are
  tabulated below:

  Location              Amount (mm)              Time Period

  Ocala, FL                91           36 hours ending at 15/0000 UTC
  Destin, FL               97                        "
  Valparaiso, FL          107                        "
  Panama City, FL         101                        "
  Hurlburt Field, FL       91                        "

  Donalsonville, GA       371           24 hours ending at 15/1100 UTC
  Blakely, GA             184                        "
  Newton, GA              152                        "
  Chipley, FL             206                        "
  Marianna, FL            128                        "

  Greenville, SC           63            3 hours ending at 15/2100 UTC
  Spartanburg, SC          60                        "

  Greenville, SC          111           24 hours ending at 16/0000 UTC
  Marianna, FL            104                        "
  Anderson, SC             95                        "
  Albany, GA               88                        "

  D. Damage and Casualties

     The Monthly Summary for September on TPC/NHC's website indicates that
  three deaths were attributed to Tropical Storm Hanna, but the location
  of these is not known to the authors.  Also, when the NHC summary was
  written in early October, the monetary damage estimate due to Hanna had
  not been determined.   The particulars on the fatalities and damage will
  be available later when the official NHC report for Hanna has been

  (Report written by Kevin Boyle and Gary Padgett)

                           HURRICANE ISIDORE
                           14 - 27 September

  A. Storm Origins

     The first mention of the pre-Isidore disturbance was in the STWO
  issued at 0930 UTC on 11 September (a tropical wave located 435 nm
  west-southwest of the Cape Verdes). The wave had likely left the African
  coast around 9 September.  This system moved westward across the tropical
  Atlantic and slow development was anticipated, as indicated in the STWOs.
  Despite being situated in a dry air environment, visible satellite images
  on the morning of 12 September indicated that the wave was slowly
  becoming better organized.  By 13/1530 UTC the disturbance had moved
  westward to a position approximately 520 nm east of the Windward Islands.
  Satellite images early on 14 September indicated that shower/thunderstorm
  activity had increased significantly, although at the time there was no
  evidence of a LLCC.  As the wave approached the Lesser Antilles a US Air
  Force Reserves' reconnaissance plane was sent to investigate the system
  and found a small closed circulation with 20-25 kt winds and a CP of
  1009 mb located near the southwestern tip of the island of Trinidad.
  High resolution visible images indicated that the system had a large
  envelope, curved bands, and an excellent outflow pattern.  Based on
  this information, the first warning on Tropical Depression Ten was issued
  at 14/2100 UTC.

     Surface observations and satellite pictures early on 15 September
  indicated that TD-10 was moving westward at 20 kts or more over northern
  Venezuela.    This overland trajectory resulted in a decrease in deep
  convection near the weakening LLCC.  Surface observations from the ABC
  Islands (Aruba, Bonaire, Curacao) showed only a slight wind shift from
  east-northeast to east-southeast.  The last advisory on this stage of
  TD-10 was issued at 15/2100 UTC, the system having degenerated into a
  tropical wave.  The remnants of TD-10 were monitored in the STWOs because
  of the strong possibility of a tropical depression reforming.   A plane
  investigating the strong tropical wave at 16/1405 UTC found no signs of
  a closed LLCC, although the vigorous tropical wave had winds of up to
  tropical storm force.  At 16/2130 UTC the system was located about
  130 nm east-southeast of Jamaica.  A Tropical Weather Discussion late
  on the 16th noted that "....the last few available images of visible
  imagery indicated that a LLCC may be trying to form approximately 130 nm
  southeast of southern Jamaica with low-level cloud elements beginning to
  wrap back around to the east near 15N, 76W.  Extrapolated inflow to the
  southeast of the main convection also indicates that a second low/mid-
  level center may be trying to form just south of Hispaniola."    A
  reconnaissance aircraft found a closed LLCC on the morning of the 17th
  and advisories were re-initiated on Tropical Depression Ten at 17/1500
  UTC, placing the centre about 125 nm south of Kingston, Jamaica, with
  the MSW estimated at 30 kts.

  B. Track and Intensity History

     Moving northwestward at 5 to 6 kts early on 18 September, TD-10 slowly
  strengthened into a 35-kt tropical storm.   The system was upgraded to
  Tropical Storm Isidore in an intermediate advisory at 0600 UTC.  The
  broad, elongated centre was estimated to be about 90 nm southwest of
  Kingston, Jamaica.  The upgrade was based on an aircraft report of 450-m
  flight-level winds of 46 kts about 65 nm east-northeast of the LLCC.
  By 18/2100 UTC the MSW had risen to 50 kts as Isidore continued to
  organise and strengthen.   The LLCC was relocated to 19.0N, 78.8W, at
  this time and aircraft reported a CP of 999 mb.   At 19/0800 UTC a
  reconnaissance plane reported a partial eyewall and found that the CP
  had fallen to 990 mb.  The MSW was raised to 55 kts, and to 60 kts six
  hours later.  Isidore became the second hurricane of the 2002 Atlantic
  season at 19/2100 UTC after an aircraft reported peak flight-level winds
  of 76 kts with estimated surface winds of 70 kts and a CP of 984 mb.  The
  19/2100 UTC advisory intensity was set at 65 kts with the centre located
  near 20.6N, 82.0W, or about 85 nm southeast of the Isle of Youth, Cuba,
  and moving northwestward at 8 kts as it was guided by a mid to upper-
  level ridge over Florida and the Bahamas.

     The MSW had risen to 75 kts (Dvorak T-numbers 4.5 and 5.0) by 0300 UTC
  on 20 September.  Isidore had expanded into a large system with tropical
  storm-force winds extending 125 nm from the centre.  Winds of up to
  25 kts were being experienced in the Florida Keys as the core of the
  hurricane was pounding western Cuba with winds of 85 kts.  The eye of
  Isidore, with a CP of 964 mb (as measured by a reconnaissance plane),
  moved inland very near La Fe, Cuba at 20/2100 UTC.    Six hours later
  Isidore had moved into the southeastern Gulf of Mexico with the MSW
  changing very little due to the less mountainous terrain of western Cuba.
  In fact, the storm had strengthened into a Category 3 hurricane with a
  MSW of 100 kts by 21/1500 UTC after it had begun a period of rapid
  intensification, accompanied by a more westward drift towards the Yucatan
  Peninsula.   Hurricane Isidore was centred at the time approximately
  65 nm west-northwest of Cabo San Antonio on the western tip of Cuba.
  The 100-kt intensity was supported by a 100-kt dropsonde report at
  21/1200 UTC plus Dvorak intensity estimates of 102 kts from TAFB, SAB
  and AFWA.

     (Editor's Note:  I have learned from NHC that Isidore's intensity
  over western Cuba has been lowered to Category 1 levels for the Best

     The MSW had reached 110 kts (the peak for the storm) by 2100 UTC on
  21 September, but strengthening leveled off--perhaps due to an eyewall
  replacement cycle.  Aircraft observed a smaller eye and a double maximum
  wind band structure, and both the Cancun radar and satellite data showed
  an outer convective ring surrounding the eye.  The plane reported a CP
  of 936 mb at 0700 UTC, 22 September, and 934 mb a few hours later.
  However, the MSW remained pegged at 110 kts when normally satellite
  estimates of T6.5 and a 934-mb CP would translate to a 125-kt to 130-kt
  system.  During the time of its peak intensity Isidore was moving west-
  ward just offshore and parallel to the northern coast of Mexico's Yucatan
  Peninsula.  On the afternoon of the 22nd, the eye of the storm made a
  sudden hook to the left and moved inland over the extreme northern
  Yucatan coast around 23/0000 UTC about 30 km east of Merida.  The MSW
  was still estimated at 110 kts at landfall.   The hurricane was moving
  on a southwesterly path at 4 kts at this time in response to mid-level
  ridging to the west-northwest.     Interaction with the land mass of
  Yucatan induced a rapid weakening and the MSW had dropped to 65 kts by
  23/0900 UTC, and to below hurricane strength six hours later.  Satellite
  imagery indicated that most of the deep convection was occurring mostly
  over the southern portion of the circulation with scattered and
  disorganized amounts elsewhere.   Tropical Storm Isidore performed a loop
  during 23 September inland over the Yucatan Peninsula, the initial
  southerly motion earlier in the day turning northerly by 23/2100 UTC.
  During this time the MSW weakened to 45 kts, and by 24/0300 UTC Isidore
  was barely a tropical storm with maximum winds estimated at only 35 kts.

     Broad mid and upper-level troughing over the central US and western
  Gulf of Mexico and a developing ridge extending from the northwestern
  Caribbean towards Florida meant a change to a north to north-northeast
  heading for the slow-moving Isidore.    Satellite images and surface 
  observations from Merida, Mexico, indicated that the poorly-defined LLCC
  was located just to the north of Merida at 0900 UTC, 24 September, and
  was about to move back over water.   Satellite imagery depicted Isidore
  as a very large swirl of low to mid-level clouds (that covered all the
  Gulf of Mexico) with the nearest convective bands far from the centre to
  the northwest and southwest of the LLCC.   However, the tropical cyclone
  lacked an inner core, and this was well backed up by a reconnaissance
  mission which indicated that the radius of maximum winds was nearly
  120-125 nm from the centre.   Based on the information from the plane,
  as well as CI estimates of 55 kts from SAB and AFWA, the MSW was
  increased to 50 kts at 24/1500 UTC.  Isidore at the time was centred
  approximately 70 nm north-northwest of Merida, or about 420 nm south
  of the Louisiana coast.
     Despite the centre of Isidore moving northward over the warm waters
  of the Gulf of Mexico under favourable developmental conditions, the
  storm failed to tighten up its inner core.   This was the reason why
  Isidore failed to intensify into a hurricane again.     Surface
  observations, Air Force Reserves' reconnaissance planes, and a NOAA G-IV
  jet indicated that the strongest winds remained located more than 100 nm
  to the northeast of the LLCC with a large area of light winds around the
  centre.  Water vapour imagery indicated that dry air was being entrained
  into the southern portion of Isidore's circulation on 25 September.
  Also, a developing upper-level LOW was centred to the storm's southwest,
  likely inhibiting further strengthening.    At 25/2100 UTC the MSW was
  upped slightly to 55 kts, based on a report from a ship (located well
  northeast of the centre) of sustained winds of 56 kts, gusting to 71 kts.
  The CP measured by a reconnaissance aircraft around this time was 989 mb.
     Due to the large circulation, tropical storm-force winds were felt on
  the southeastern Louisiana coast as early as 0300 UTC on 26 September.
  These strongest winds extended 90-100 nm from the LLCC.  Tropical Storm
  Isidore made landfall at 26/0900 UTC just west of Grand Isle on the
  Louisiana coast with a MSW of 55 kts.   This was the peak intensity of
  Isidore during the second part of its career in the Gulf of Mexico.  In
  satellite imagery Isidore was looking better-organized with strong
  convection near the LLCC, and was probably on its way to becoming a
  hurricane until it ran out of energy by moving inland.   The large
  circulation accelerated to the north-northeast, embedded in the south-
  westerly flow ahead of a trough.   A steady weakening occurred after
  landfall and Isidore became a 30-kt depression on the final NHC advisory,
  issued at 26/2100 UTC.  The depression was centred about 80 km north-
  northeast of Jackson, Mississippi, and moving northward at 22 kts.
  HPC assumed responsibility for issuing storm summaries on the remnants
  of Isidore as long as there remained a threat from flooding.  The fourth
  and final summary from HPC was issued at 2100 UTC on 27 September.  The
  depression had become extratropical over western Pennsylvania and was
  speeding northeastward at 35 kts.

  C. Meteorological Observations

     At Isidore's peak intensity (110 kts, 934 mb) on 22 September in
  the southern Gulf of Mexico, hurricane force winds reached outward
  from the centre 45 nm in the northern semicircle and 30 nm in the
  southern quadrants.   Gales extended out 175 nm north of the centre
  and from 100-125 nm to the south.   On 25 September, when Isidore was
  a large sprawling tropical storm churning northward through the central
  Gulf, gales reached out 275 nm to the northeast of the centre and
  200 nm to the southeast.

     As Isidore was crossing western Cuba, Cabo San Antonio reported
  maximum sustained winds of 73 kts at 20/2100 UTC.  On the afternoon
  of 22 September, a ham radio report from the Yucatan indicated
  sustained winds of 61 kts, gusting to 78 kts, near Merida.  During
  the afternoon of the 25th a ship located northeast of the centre
  reported sustained winds of 56 kts, gusting to 71 kts.   Around
  26/0000 UTC, the C-MAN station at Southwest Pass (near the mouth of
  the Mississippi River) recorded sustained winds of 43 kts with a
  peak gust of 52 kts.  (NOTE: Some reconnaissance reports were included
  in the narrative above.    Additional information reported by the
  reconnaissance flights can be found in the discussion bulletins issued
  with each regular advisory.  These are archived on NHC's website at
  the following URL:> )

     Following are some rainfall accumulations reported in the storm
  summaries issued by HPC:

  (1) 24-hour totals ending at 2300 UTC on 26 September
  McComb-Pike County, Mississippi          194 mm
  Tupelo, Mississippi                      176 mm
  Hattiesburg, Mississippi                 119 mm

  (2) 24-hour totals ending at 0000 UTC on 27 September
  New Orleans-Audobon, Louisiana           162 mm
  New Orleans-International Airport        143 mm
  Jackson, Tennessee                       148 mm
  NWS Memphis, Tennessee                    80 mm
  Evergreen, Alabama                        88 mm

  (3) Storm Totals
  New Orleans-Audobon, Louisiana           299 mm
  New Orleans-International Airport        292 mm
  McComb-Pike County, Mississippi          258 mm
  Tupelo, Mississippi                      224 mm
  Jackson, Tennessee                       210 mm
  Hattiesburg, Mississippi                 209 mm
  Rough River Lake, Kentucky               164 mm
  Woodbury, Kentucky                       157 mm
  Evergreen, Alabama                       147 mm
  Louisville APT, Kentucky                 134 mm
  Scottsburg, Indiana                      126 mm
  Leavenworth, Indiana                     124 mm
  NWS Memphis, Tennessee                   102 mm
  Cincinnati, Ohio                          91 mm

     There were some unofficial reports of over 500 mm of rain recorded
  in the New Orleans area.

  D. Damage and Casualties

     In Cuba some areas received in excess of 600 mm of rain during the
  first 24 hours of Isidore's passage, and coastal storm surges reached
  from 2.5-3.75 m above normal tide level in some areas.  Some 280,000
  people as well as thousands of cattle were evacuated as the storm
  approached.  Flooding from the heavy rains was particularly severe
  in the municipalities of San Juan y Martinez, Sandino, and Guane with
  reports of destroyed houses, roofs blown off, disrupted electricity
  supplies, and blocked roads.   Some 134 tobacco drying houses were
  affected, damaging valuable tobacco stocks, and at least 3000 tons of
  citrus fruit were reported lost.   Earlier, rains from Isidore had
  pounded Jamaica, no doubt causing some flooding, but no reports from
  that island on the effects of Isidore were available to the author.

     Mexico was especially hard hit by Isidore, as the storm landed on
  the northern Yucatan coast as a Category 3 hurricane.   The impact
  was most severe in the state of Yucatan, but the states of Quintana
  Roo, Campeche, and Chiapas were also significantly affected by the
  hurricane.  In Yucatan state alone over 500,000 persons were adversely
  affected by Isidore.  Almost 23,000 houses lost their roofs and 12,800
  houses were destroyed.  (Another report stated that 33,000 village
  houses were complete destroyed.)  Barns, warehouses and storage bins
  were impacted with 75% being destroyed or severely damaged.  One report
  indicated that three persons were killed as a result of the hurricane,
  and 240 injuries were reported.

     Agricultural losses in Yucatan state were staggering:  80% of maize
  (corn) production land (70,000 hectares) was destroyed and still flooded
  almost a month after the hurricane; 40,000 hectares of fruit trees were
  lost; 8.5 million poultry drowned, representing 90% of the poultry
  production in the state; and 80% of the pigs disappeared in the floods.

     The capital city of Yucatan state, Merida, was badly damaged with
  many trees uprooted, roofs ripped from houses, 70% of the power lines
  down, and 80% of the city flooded by waters from 0.2 to 0.3 m deep.
  Some 70,000 persons were evacuated from high risk coastal areas.  The
  city of Puerto Progreso was also severely impacted by Isidore.  A week
  after the storm water supply systems in the impacted areas were
  functioning at only 50% of their total capacity.

     In the state of Campeche from 15,000 to 20,000 village houses were
  destroyed either by direct impact of Isidore or by the following floods.
  Over 30,000 head of cattle were lost; 63,500 honey bee production
  apiaries were destroyed; and 100,000 hectares of agricultural land

     New Zealand-based storm chaser (or more appropriately, natural
  disaster chaser) Geoff Mackley was in Progreso as Isidore made landfall.
  A brief description of Geoff's experiences plus some striking photos
  showing the damage caused by Isidore may be found at the following URL:

     < >

     In the United States there were four drowning deaths caused by
  Isidore, and damage has tentatively been estimated at $200 million.
  Losses to agricultural crops were substantial in some areas of the
  Southeast as the storm's heavy rains came during or just before
  harvest time.  Storm chaser Jim Edds from the Florida Keys was in
  the Mobile, Alabama, area during Isidore and has placed on his website
  some pictures depicting flooding along the Mobile Bay Causeway due to
  the storm surge.  The URL is:  < >

     Many additional reports, articles, and press releases on Isidore's
  effects can be found on the ReliefWeb homepage at the following URL:>

  E. Additional Discussion

     Hurricane Isidore's peak MSW of 110 kts vs a minimum CP of 934 mb
  serves as a good illustration of the individuality of storms and the
  fact that there is not a one-to-one correspondence between the MSW
  and CP--a notion which seems to be fairly prevalent among the public
  and media.     In the Atlantic basin a 934-mb central pressure would
  normally support a MSW of 125-130 kts.  However, the maximum intensity
  of a hurricane depends on several factors, including the central
  pressure, the environmental pressures in the vicinity of the storm, and
  the distance over which the maximum pressure fall occurs.  Additionally,
  peculiarities of the internal structure of a given hurricane can play a
  role in determining the maximum intensity, and also the storm's
  translational speed can help to augment the wind field on the right-hand
  side of the cyclone (with respect to its direction of motion).  Isidore
  was a relatively large, slow-moving hurricane located in a region of
  lower-than-normal sea level pressures.   By way of contrast, Hurricane
  Georges in September, 1998, boasted 135-kt winds with a CP of 937 mb
  at its peak intensity well east of the Lesser Antilles.  Georges was a
  smaller storm in areal extent than Isidore, and was zipping along in
  the easterlies at twice the speed of Isidore with a rather strong ridge
  of high pressure to its north.

     Another factor which plays a role in the strength of the surface
  winds is the intensity of the core convection.   Intense convection
  helps to bring the extreme winds occurring at higher elevations down
  to the surface.  With other factors being equal, a strengthening
  hurricane with intense convection can have a significantly higher MSW
  than a steady-state or weakening storm with the same central pressure
  but which has weaker convection.  According to Rich Henning, a member
  of the 53rd Weather Reconnaissance Squadron (the Hurricane Hunters),
  Isidore's central convection was not particularly impressive, especially
  around the time it made landfall.  Rich was on a flight into the storm
  as the centre moved onshore in the Yucatan Peninsula, and he stated that
  while the eye was extremely well-formed with excellent outflow and no
  sign of vertical shear (i.e., an optimum dynamical environment), there
  was very little deep convection near the eye, most of it being well to
  the east over the peninsula and west over the Bay of Campeche.   Rich
  attributes this in part to upwelling just off the northern Yucatan
  shoreline, leading to SSTs a few degrees cooler than farther east in
  the Yucatan Channel.  (Thanks to Rich for sharing his observations and
  thoughts on Isidore.)

  (Report written by Kevin Boyle and Gary Padgett)

                       TROPICAL STORM JOSEPHINE
                           17 - 19 September

     An area of low-pressure centred about 740 nm east of Bermuda was
  first mentioned in a Tropical Weather Outlook (TWO) from TPC/NHC at
  0930 UTC on 16 September.  The LOW was located in an environment of high
  surface pressures, and slow development of the system was forecast.  The
  TWO issued at 1530 UTC 17 September indicated that the LOW was acquiring
  tropical characteristics about 600 nm east of Bermuda.  Satellite images
  showed that the LOW contained enough convection to qualify as a tropical
  cyclone, and advisories were initiated on Tropical Depression Eleven at
  2100 UTC, 17 September.    The system was embedded in an upper-level
  trough and moving slowly northward at 7 kts at the time.  The depression
  was initially not forecast to reach tropical storm intensity; however,
  based on an 18/0300 UTC report of winds of 37 kts from ship PDKK, Quik-
  Scat and SSM/I winds of 35 kts, and Dvorak satellite estimates of T2.5
  from TAFB, TD-11 was upgraded to Tropical Storm Josephine at 0900 UTC
  when located approximately 650 nm east-northeast of Bermuda.   The
  system was then moving north-northeastward at 12 kts.

     Josephine reached a peak intensity of 35 kts on the 18/0900 UTC
  warning and, although there were still intermittent bursts of convection
  near the LLCC, dry air intrusion at mid-levels combined with increasing
  westerly and southwesterly upper-level shear prevented further develop-
  ment.  However, the pressure gradient remained very tight on the east
  side as Josephine moved generally northeastward on the western side of
  a large HIGH centred to the northeast and in advance of an approaching
  mid-latitude trough.  Josephine accelerated northeastwards between these
  features towards cooler waters and increasing southwesterly shear, and
  was absorbed into the frontal system well to the east of Newfoundland
  on 19 September.  A surprise came when ship C6LV3 reported 50-kt winds
  at 1200 UTC about 85 nm southeast of the centre.   This was the reported
  MSW on the final advisory issued by NHC at 1500 UTC, which placed the
  centre of the now-extratropical Josephine at 42.5N, 42.5W, or about
  550 nm west-northwest of the northwesternmost Azores, moving northeast-
  ward at 25 kts.    This ship report suggests that Josephine was very
  possibly stronger during its tenure as a tropical storm than estimated
  in the advisories, especially considering its rather rapid translational
  speed.  The post-Josephine extratropical LOW had been absorbed into
  another LOW to the north by the 20th.

  (Report written by Kevin Boyle)

                            HURRICANE KYLE
                       20 September - 18 October

     Even though Kyle was a hurricane for only 2.5 days, it persisted over
  the subtropical Atlantic for 22.25 days, becoming the third longest-lived
  Atlantic basin tropical cyclone on record.  Hurricane Ginger (1971) holds
  the Atlantic longevity record of 27.25 days, followed by Hurricane Inga
  (1969) with a lifespan of 24.75 days.  Kyle's track was very similar to
  Hurricane Olga of 2001 (as well as to the aforementioned Inga and Ginger)
  in that it wandered aimlessly for days over the subtropics with several
  fluctuations in intensity.  Kyle was upgraded to tropical storm intensity
  five separate times during its life.

  A. Storm Origins

     Long-lived Kyle began as a non-tropical LOW that formed east-southeast
  of Bermuda near 28.4N, 51.7W, on 20 September.   The LOW had developed
  some organized convection, but a TRMM overpass at 20/1601 UTC indicated
  that the maximum winds were at least 100 nm from the centre.   Based on
  this information TPC/NHC issued the first warning on Subtropical
  Depression Twelve.  Initially moving northward at 10 kts, the subtropical
  depression became better organized and was upgraded to Subtropical Storm
  Kyle at 21/0900 UTC when centred about 700 nm east of Bermuda.  Kyle was
  situated in a favourable environment with 26-27 C SSTs and light vertical
  shear with a good anticyclonic outflow at upper levels.   Continuing in
  a northward direction, Subtropical Storm Kyle decelerated to a virtual
  standstill in response to increased ridging to the north.

  B. Track and Intensity History

     Kyle had fully transformed into a tropical storm by 2100 UTC,
  22 September, as deep convection had properly developed over the centre
  and the circulation had become more detached from a frontal zone.  Kyle
  was then centred approximately 775 nm east of Bermuda and stationary.
  The cyclone phase analysis from Pennsylvania State University plus AMSU
  data showed that the system had developed a warm core.  The MSW was
  estimated at 40 kts but was lowered to 35 kts at 23/0300 UTC as
  T-numbers from TAFB and SAB were 2.5 at this time.  Kyle began to move
  toward the southwest on 23 September as pressure built to the north of
  the system.  This southwestward heading was maintained for several days
  on the east side of the northeast-southwest oriented mid-level ridge
  near Bermuda.  Kyle had been undergoing shear which was pushing the
  associated convection to the west of the LLCC, but these conditions
  began to relax on 24 September.    In response to the more favourable
  conditions, the MSW rose to 55 kts, convection wrapped around the centre
  with cooling cloud tops, and a banding-type eye feature appeared.
  Hurricane intensity was reached late on 25 September based on the banding
  eye and Dvorak intensity estimates of T4.0 or 65 kts.   Kyle by this time
  had reached a point approximately 525 nm east-southeast of Bermuda.

     Kyle had only a day or so to strengthen before northwesterly shear
  began to affect the system.     However, the cyclone reached a peak
  intensity of 75 kts at 0900 UTC on 26 September when located roughly
  450 nm east-southeast of Bermuda, based on Dvorak satellite intensity
  estimates of T4.5 from TAFB and SAB.  Water vapour imagery at 27/0300 UTC
  showed the northerly wind shear about to impact the hurricane.  The eye
  became less distinct and satellite estimates began to drop slightly early
  on the 27th.  Kyle appeared to be fighting off the shear for awhile and
  the MSW remained at 75 kts.  However, the circulation became distorted
  east-west, and coupled with the advection of dry air into Kyle from the
  north, the MSW began to fall steadily.  A microwave pass at 27/1247 UTC
  showed that the upper-level centre was displaced about 15 nm south-
  southwest of the LLCC.   The shearing conditions were now pushing Kyle
  on a more west-southwestward track, still trapped between a mid to upper-
  level ridge located between Bermuda and the southeastern US and a second
  ridge lying along 17N in the tropical Atlantic.   Kyle was downgraded to
  a tropical storm at 28/1500 UTC when located about 375 nm south-southeast
  of Bermuda.  The storm had reached the southernmost point of its track
  and had begun a slow westward drift.  The completely-exposed LLCC then
  turned to the northwest during 29 September with the MSW dropping to
  40 kts.  Late on the 29th Kyle began a slow northerly crawl, and this
  continued on the 30th before the system came to a halt later that day
  near 29.8N, 64.8W, or approximately 150 nm south of Bermuda.

     CIMSS analyses at 30/0900 UTC indicated that the northeasterly shear
  that had plagued Kyle had begun to ease significantly and deep convection
  correspondingly redeveloped over the LLCC.  This development proved to
  be temporary, however,  and Kyle became elongated northeast-southwest,
  subsequently splitting into two LLCCs. The northeastern centre moved away
  rapidly to the east-northeast and soon merged with a cold front while
  the southwestern LLCC, which remained classified as Kyle, remained
  stationary.  At 30/2100 UTC the system was downgraded to a 30-kt tropical
  depression.   (NOTE:  Due to the difficulty in positioning the poorly-
  defined LLCC of Kyle, there are uncertainties in its track and motion at 
  this time.)

     Tropical Depression Kyle was undergoing northerly shear on 1 October
  which was displacing the deep convection to the southern portion of the
  system.  However, visible images, a QuikScat pass, and Dvorak estimates
  suggested a stronger system than a depression, and at 01/1500 UTC Kyle
  was upgraded back to a 35-kt tropical storm, centred about 300 nm south-
  southwest of Bermuda.  Despite the shear Kyle's large circulation
  (albeit slanted) was able to continuously produce deep convection and
  the MSW was raised to 40 kts on the next advisory at 01/2100 UTC.  Kyle
  continued to strengthen and became vertically-aligned again, reaching a
  secondary peak intensity of 60 kts at 02/2100 UTC while creeping a little
  (or reforming) to the east.  For several days Kyle had been meandering
  around 27-29N, 64-66W, on the southern flank of a HIGH to its north.
  Another HIGH building over southern Georgia meant a change to a north-
  westerly heading late on 3 October.  This HIGH introduced another episode
  of northeasterly shearing conditions for Kyle to endure, and by 04/0900
  UTC the MSW was hovering at 35 kts.    The LLCC had become partially-
  exposed to the northeast of the deep convection, and by 05/0300 UTC it
  had become completely detached 130 nm away from the deep convection.
  Kyle was downgraded to a depression at 05/0900 UTC with little associated
  deep convection and a poorly-defined LLCC, located approximately 350 nm
  west of Bermuda.

     A slow crawl to the north-northeast began late on 5 October, and this
  had changed to an east-northeasterly motion by 06/0300 UTC.   However,
  Kyle soon came to a halt again as it got stuck in a col with a ridge to
  the northeast and southwest, a large upper-level LOW to the east-
  southeast, and a broad frontal trough to the northwest over the east-
  central US.  Meanwhile, Kyle was upgraded to tropical storm status for
  the third time at 1500 UTC on 6 October when located roughly 250 nm
  west-northwest of Bermuda.     This upgrade was based on Dvorak classi-
  fications and several uncontaminated 35-kt wind speed vectors over the
  southern semicircle from a QuikScat pass.  A deep convective bursting
  pattern occurred also on 6 October.   The approaching upper-level LOW
  from the east-southeast began to influence Kyle early on 7 October and
  a slow south to southwestward motion began.   The circulation became
  elongated west-northwest to east-southeast with one circulation centre
  near 33.5N, 71.5W, and a second vortex/vortmax around 32N, 70W, near the
  main convection.  The centre was relocated to near 32.8N, 70.7W, which
  was a mean position between the two vorticity centres.  Although the main
  convection was nearer the eastern centre, some small bursts had occurred
  near the western one.

     On 8 October the synoptic situation became more complicated.   In
  addition to the large upper-level cold LOW to the southeast, there was a
  small mid-to upper-level LOW just east of Florida and a short-wave trough
  moving through the Carolinas.  However, this pattern simplified as the
  cold LOW turned northward and weakened and the LOW near Florida moved
  southwards.  The short-wave trough moved away to the northeast with an
  anticyclone following in its wake.  Again, Kyle found itself moving on a
  curving southwesterly track around the southern flank of a HIGH.

     Late on 8 October Kyle consisted of basically a low-level cloud-swirl
  with very little convection.  This led to its subsequent downgrade to a
  tropical depression at 08/2100 UTC, located about 425 nm east-southeast
  of Charleston, South Carolina.  Subsidence associated with the HIGH was
  the prime culprit in suppressing associated deep convection, and Kyle
  continued as a swirl of low clouds through 9 October.  A US Air Force
  Reserves reconnaissance mission investigating the system later that day
  found a MSW of only 25 kts.   Kyle turned more to the west and began to
  look a little healthier early on 10 October as convection developed,
  first in the southeast and east quadrants, and then over the LLCC.
  However, northerly shearing quickly blew this convection into the
  southern portions of the storm.  Early on 11 October deep convection
  increased markedly and the Jacksonville radar showed improved banding

     An Air Force Reserves' aircraft flew into the system around 0700 UTC
  on 11 October and found 45-kt flight-level winds about 40 nm northeast
  of the centre along with a central pressure of 1008 mb.    Kyle was
  upgraded to a tropical storm once again with a MSW of 35 kts at 0900 UTC,
  centred approximately 100 nm south-southwest of Charleston.  The storm
  began to accelerate toward the north-northeast parallel to the Carolina 
  coastline ahead of a mid-level trough and associated cold front.  Only
  the western part of Kyle's circulation moved overland, and a
  reconnaissance plane indicated that most of the tropical-storm force
  winds were occurring to the east of the LLCC over water.  A later flight
  into the storm around 11/2100 UTC found a maximum flight-level wind of
  only 29 kts, so Kyle was downgraded to a depression in a special advisory
  issued at 12/0000 UTC when the centre was located very near Cape Fear,
  North Carolina.

     Tropical Depression Kyle moved through the Outer Banks during the
  night and emerged into the Atlantic very early on 12 October.  A
  reconnaissance mission at 12/0600 UTC found winds of 49 kts at 450 m
  along with a pressure of 1009 mb over eastern Pamlico Sound.  Also, the
  Diamond Shoals C-MAN station reported 8-min avg winds of 38 kts with
  gusts to 43 kts at 0700 UTC.   Based on these observations, Kyle was
  upgraded back to tropical storm intensity for the fifth time at 12/0900
  UTC.  The storm was located about 50 nm northeast of Cape Hatteras, and
  had accelerated east-northeastward at 20 kts as it interacted with a
  mid to upper-level shortwave trough.  The final advisory--the 89th--was
  issued by NHC at 12/1500 UTC when the wind-field of Kyle began to expand
  and its associated convection took on the appearance of a frontal band. 

     The extratropical Kyle was tracked across the Atlantic, executing its
  third anticyclonic loop on 14 October, then heading east-southeastward
  and southeastward.  On the 17th Kyle's remnants turned east-northeastward
  and passed just north of the Azores Island of Faial.  The island reported
  30-kt winds and dew points of 20 C as the LOW approached.  There was a
  notable wind shift associated with this system, signifying the presence
  of fronts.  Kyle finally ended its long, long trek on 18 October as it
  was absorbed into an extratropical LOW centred to its northwest.   The
  merger of this LOW with the remnants of Kyle resulted in a large, deep
  LOW which moved northeastward to affect the British Isles during 20-23

  C. Meteorological Observations

     On 30 September Bermuda experienced wind gusts to tropical storm
  force.  This was caused by the combination of the outer part of Kyle's
  circulation and a cold front.  The only significant weather report due to
  Kyle in the US was from the Diamond Shoals C-MAN station which reported
  8-min avg winds of 38 kts with gusts to 43 kts at 12/0700 UTC.

  D. Damage and Casualties

     The deep depression associated with the remnants of Kyle caused one
  death in stormy seas southwest of the British Isles.  There were no other
  reports of casualties or damage as Kyle remained over the open waters of
  the Atlantic.

  (Report written by Kevin Boyle)

                             HURRICANE LILI
                        21 September - 4 October

  A. Storm Origins

     The humble beginnings of Hurricane Lili were first mentioned in a
  STWO issued at 1530 UTC on 20 September.  An area of cloudiness and
  showers associated with a westward-moving tropical wave was located
  about 1400 nm east of the Windward Islands.  This wave had likely left
  the African coast around the 16th or 17th.  Upper-level conditions were
  forecast to be favourable for development, and STWOs issued during the
  20th and 21st September indicated that the disturbance was becoming
  better organized.  Ship reports and buoy data showed a broad LLCC, and
  at 2100 UTC, 21 September, the first advisory was issued on Tropical
  Depression Thirteen, located near 10.4N, 45.7W, or about 825 nm east-
  southeast of Barbados.  The initial MSW on this advisory was 30 kts.
  For the following two to three days, the depression was forecast to
  move on a predominantly westward track, steered by a mid-level ridge
  to the north.

     The MSW remained at the 30-kt threshold through the 22nd as the storm
  moved smartly westward toward the Windward Islands.  Despite the impres-
  sive convective cloud pattern, visible satellite images revealed that
  the LLCC was displaced west of the CDO.  This was confirmed by a 21/1728
  UTC TRMM pass.   At 23/1200 UTC the centre was positioned near 12.7N,
  58.7W,  or about 50 nm east-southeast of Barbados, moving westward at
  20 kts.  Surface observations at 23/1500 UTC suggested that the LLCC was
  elongated west to east, and advisory positions at this time were a full
  degree or more west of the satellite fixes.  The 23/1500 UTC advisory
  also noted that Barbados had reported winds of 36 kts, gusting to 48 kts,
  during a light shower.  These winds were possibly caused by downdrafts
  bringing the 40-kt gradient-level winds down to the surface.  An Air
  Force Hurricane Hunter plane investigating the system found 58-kt winds
  at 300 m, equivalent to a surface MSW of 45 kts.  Based on this finding,
  Tropical Depression Thirteen was upgraded to Tropical Storm Lili in a
  special advisory at 23/1530 UTC.  The centre was located about 40 nm
  south-southwest of Barbados, moving slightly north of due west at
  17 kts.

  B. Track and Intensity History

     The MSW was increased to 50 kts at 23/2100 UTC based on an aircraft
  report of 64 kts at 450 m at 1712 UTC.   Satellite images and surface
  observations at this time indicated that Tropical Storm Lili was
  continuing to become better organized and that deep convection was
  beginning to wrap around the north side of the circulation.  Lili had
  begun crossing the Windward Islands, and at 24/0000 UTC was located
  near 12.7N, 62.0W, or about 40 nm west of the Grenadine Islands.  Wind
  gusts to 65 kts and 47 kts had been reported from Barbados and St Lucia,
  respectively.  At 24/1500 UTC the MSW was increased to 60 kts based
  on an aircraft report of 244-m flight-level winds of 76 kts in the
  northeast quadrant.  The westerly heading had slowed to 9 kts by
  25/0300 UTC with Lili having weakened dramatically to a minimal tropical
  storm with a MSW of 35 kts.  This weakening was due to southerly shearing
  conditions over the system.  In fact, an Air Force reconnaissance flight
  during the early evening was unable to find a closed LLCC.  Radar data
  from San Juan indicated the possible presence of a circulation near
  15.7N, 65.5W, but the plane could only find easterly winds of 35-45 kts
  at flight-level and easterly surface winds of 37 kts in that area.  This
  suggested that the circulation was at mid to upper-levels only.

     A reconnaissance flight around 25/1200 UTC indicated the presence of
  at least a weak LLCC near 13.7N, 68.7W, but not defined enough for a
  vortex message.  Flight-level (450 m) winds of 47 kts and 45 kts were
  found to the northeast of the centre, respectively.  Based on these
  wind reports, the MSW was increased to 40 kts in an intermediate advisory
  at 1200 UTC.  Due to the uncertainty of the centre position, the LLCC
  was relocated a little to the south to near 14.1N, 69.3W, or about 260 nm
  south-southeast of Santo Domingo at 25/1500 UTC.  The MSW was reduced
  back to 35 kts at 26/0300 UTC.  Lili appeared very ragged in infrared
  satellite images at this time, and most of the deep convection was
  located northeast of the weak and broad LLCC.   An Air Force Hurricane
  Hunters' plane reported that the strongest winds in the northern semi-
  circle were well away from the centre.     The deep convection had 
  decreased in areal coverage, and since reconnaissance flights had
  struggled to find a LLCC for almost 48 hours, Lili was downgraded to
  a tropical wave and advisories discontinued at 26/1500 UTC. 
     Due to the likelihood of regeneration NHC continued to monitor Lili's
  remnant LOW.  At 2130 UTC the system was located about 200 nm southeast
  of Jamaica, and conditions were expected to become more favourable for
  development during the following two days.    A Special Tropical 
  Disturbance Statement (STDS) was issued at 27/0235 UTC indicating that
  deep convection was reforming near the LLCC and that the system was
  getting better organized with a well-defined LLCC.  The STDS also noted
  that advisories would resume at 27/0300 UTC.  Lili was restarted as a
  tropical depression with a MSW of 30 kts at this time, located about
  205 nm southeast of the eastern tip of Jamaica.   A 26/2309 UTC micro-
  wave pass indicated that the tight LLCC was located just southwest of 
  the deep convection, indicating that some wind shear was still present.
  Cloud top temperatures had cooled to -80 to -85 C over the previous few

     Although shear was still affecting the system, Lili regained tropical
  storm status at 27/1500 UTC with the MSW increasing to 40 kts six hours
  later.  At 27/2100 UTC the LLCC began to be "smothered" by the CDO
  generated by the deep convection as the shear conditions lessened.  Lili
  was moving toward the northwest at 7 kts early on the 28th, but had
  slowed to 3 kts by 28/2100 UTC with the storm centred roughly midway
  between the eastern tip of Jamaica and the western tip of Haiti.  An Air
  Force Reserves' reconnaissance plane found flight-level westerly winds
  of 51 kts south of the centre along with a minimum central pressure of
  999 mb at 28/0900 UTC.  The centre was relocated to near 18.6N, 75.8W,
  or about 100 nm south-southwest of Guantanamo Bay, Cuba, at 2100 UTC
  after aircraft reported that the centre had reformed to the northeast of
  the previous one.  The MSW had been upped slightly to 45 kts at 28/0000
  UTC, but was reduced back to 40 kts at 29/0300 UTC based on a CP of
  1002 mb and peak 850-mb flight-level winds of 50 kts in the northeast
  quadrant.   However, the reconnaissance mission around 29/1200 UTC found
  peak 850-mb winds of 57 kts with a CP of 994 mb, so the MSW was raised
  once more to 45 kts at 29/1500 UTC, and to 50 kts six hours later.   The
  reconnaissance crew also noted the presence of an 8-11 nm diameter eye
  feature which persisted through the 29th.    At 2100 UTC on 29 September
  the centre of Tropical Storm Lili was moving westward at 4 kts just
  north of western Jamaica near 18.7N, 77.8W, or only about 20 nm north-
  northeast of Montego Bay.

     The cyclone gradually began to pull away from Jamaica to the west-
  northwest and steadily intensified.  The MSW was upped to 55 kts at
  30/0300 UTC--an earlier GPS dropsonde just northeast of the centre had
  revealed average boundary layer winds of 70 kts, and the aircraft had
  reported the existence of a closed eye wall 20 nm in diameter.   Lili
  finally reached hurricane intensity at 1500 UTC on 30 September when
  located just north of Little Cayman Island.  The discussion bulletin
  well illustrates the manner in which the forecasters have to digest
  and draw conclusions from the data they receive: "Although flight-
  level winds do not quite support hurricane intensity, as the peak
  850-mb wind reported by the reconnaissance aircraft was 75 kts--of
  which 80% is 60 kts--a GPS dropsonde in the northeast eyewall supports
  upgrading Lili to a hurricane.  This drop had 81 kts at 14 m elevation,
  and a low-layer mean wind of 90 kts, which adjusts to 73 kts at the
  surface.  The advisory intensity is set to 65 kts, and is a compromise
  between the adjusted values from the aircraft and dropsonde."

     Hurricane Lili began to draw a bead on western Cuba as its forward
  motion slowly increased.  The storm's west-northwestward track was
  caused by a deep-layer high centred near 30N, 80W, and this feature was
  forecast to steer Lili in the same direction for several more days.
  The storm continued to strengthen and peak winds were up to 75 kts and
  the CP had fallen to 975 mb by the time the centre reached the Isle of
  Youth south of western Cuba during the early morning of 1 October.
  Lili's eye was crossing the western tip of Cuba around 1800 UTC.  The
  storm was upgraded to a Category 2 hurricane with 85-kt winds at this
  time, based on reconnaissance reports, even though the storm "sputtered"
  a bit due to land interaction (part of the eyewall weakened and cloud
  tops warmed some).  By 2100 UTC Lili's eye had cleared Cuba and entered
  the southeastern Gulf of Mexico, being located about 45 nm north of Cabo
  San Antonio, Cuba, and moving northwestward at 13 kts.  Satellite images
  at this time indicated a healthy, organized system with a well-defined
  eye seen on radars at Casablanca, Cuba, and Key West, Florida.  The
  MSW was increased to 90 kts at 2100 UTC based on CI estimates of 90 kts
  from TAFB, SAB and AFWA, plus earlier flight-level winds of 102 kts
  reported by a reconnaissance aircraft northeast of the centre while
  Lili's eye was still over Cuba.

     Once Lili had entered the Gulf of Mexico, the storm steadily
  strengthened as it tracked northwestward at 14 kts.  During the after-
  noon and evening of 1 October Lili went through an eyewall replacement
  cycle as the eye diameter shrank from 35 nm to 17 nm.  The 90-kt MSW
  was maintained for the 02/0300 UTC advisory, although the discussion
  noted that the eye had cleared out and became better defined with the
  eye temperature rising from -55 C to -9 C in just two hours.   The
  CP had fallen to 955 mb by 02/0600 UTC and the MSW was increased to
  95 kts, putting Lili on the brink of major hurricane status.  The
  hurricane was then centred a little over 400 nm south-southeast of
  New Orleans.  A reconnaissance plane around 1200 UTC found 125-kt
  flight-level winds 10 nm northeast of the centre along with a CP of
  953 mb.  On this basis Lili's MSW was upped to 105 kts at 1500 UTC,
  making the storm a Category 3 hurricane on the Saffir/Simpson scale
  as it continued moving at 13 kts toward the Louisiana coast.

     An aircraft around 1700 UTC reported that the central pressure had
  fallen from 954 mb to 941 mb in five hours.  A peak flight-level wind
  of 131 kts was also measured; hence, a special advisory package was
  issued at 1800 UTC upgrading Lili to a Category 4 hurricane with 115-kt
  winds.  The storm's centre was then located about 285 nm almost due
  south of New Orleans.   The CP continued to fall, but more slowly,
  reaching a minimum of 938 mb at 02/2000 UTC.   The MSW was increased
  to 120 kts at 2100 UTC based on a flight-level wind of 136 kts.  Lili's
  peak intensity of 125 kts occurred around 03/0300 UTC when the storm
  was centred about 190 nm south-southeast of Marsh Island on the Louisiana
  coast.  Even though the CP had risen to 942 mb, a peak 700-mb wind of
  141 kts was observed by an Air Force Reserves' reconnaissance plane,
  and the stepped-frequency microwave radiometer on the NOAA research
  plane indicated surface winds of 125 kts.

     Lili maintained its 125-kt peak intensity for only a short while
  before a rapid weakening trend set in.    By 03/0900 UTC the MSW had
  fallen to 105 kts, the eyewall had collapsed into fragments, and water
  vapour imagery suggested that the system was being sheared from the
  southwest.    Radar images at this time indicated possible dry air
  intrusion into the southwestern part of the circulation.   Rapid
  weakening continued and the MSW had fallen to 85 kts when Lili's eye
  made landfall near the western edge of Vermillion Bay around 03/1400 UTC.
  At the time of landfall Lili was moving due northward, and by 03/2100 UTC
  the weakening cyclone was moving northward at 16 kts and had weakened to
  a 45-kt tropical storm located very near Alexandria, Louisiana.  Turning
  northeastward, Lili was downgraded to a tropical depression at 04/0900
  UTC when centred near 34.3N, 90.9W, or 120 km southwest of Memphis, 
  Tennessee, moving north-northeastward at 18 kts.  HPC issued only two
  storm summaries on the remnants of Lili, the second at 2100 UTC stating
  that no closed circulation had been evident since 1800 UTC and that the
  remnants of the storm were being absorbed into a low-pressure system
  in the northern Great Lakes area.

  C. Meteorological Observations

     Lili reached its peak intensity of 125 kts with a CP of 938 mb during
  the afternoon of 2 October in the north-central Gulf of Mexico.  By this
  time the storm had grown into a fairly sizable hurricane.   Sustained
  winds of hurricane force extended outward from the eye about 45 nm in
  the northeast quadrant and 20 nm to the southwest.   Gales reached out
  170 nm to the northeast and about 100 nm in the southwest quadrant.

     Early in its career as Lili was passing through the southern Wind-
  wards, Barbados recorded a peak sustained wind of 41 kts, gusting to
  65 kts, while St. Lucia reported peak gusts of 43 kts.   On the morning
  of 30 September, Little Cayman reported peak gusts in the 61-70 kt range.
  Punta del Este on the Isle of Youth measured gusts to 94 kts around
  1200 UTC on 1 October, while La Fe reported sustained winds of 55 kts.

     Shortly before 2100 UTC on 2 October, the centre of Lili passed only
  about 10 nm from NOAA buoy 42001 (25.9N, 89.7W).   The pressure fell
  44 mb in three hours to a minimum of 956 mb.  The buoy's anemometer
  recorded an 8-min avg sustained wind of 91 kts with a peak gust of
  129 kts, accompanied by 10-m seas.

     As Lili's centre was nearing the Louisiana coast (around 03/0900 UTC),
  Grand Isle and Southwest Pass reported winds gusting to 69 kts and
  70 kts, respectively.    Storm chaser Tim Marshall was situated near
  Delcambre, Louisiana, on Route 14.  Tim's instrumentation recorded the
  maximum winds at 1508 UCT:  a sustained wind of 54 kts with a peak gust
  of 84 kts.  The minimum SLP of 977.6 mb was measured six minutes later.

     NOTE: Some reconnaissance reports were included in the narrative
  above.  Additional information reported by the reconnaissance flights
  can be found in the discussion bulletins issued with each regular
  advisory.  These are archived on NHC's website at the following URL:

     < >

     Lili's circulation was not nearly as large as that of Isidore, and
  the storm moved through fairly quickly; hence, rainfall accumulations
  were significantly less than with the earlier cyclone.   A power plant
  near Alexandria, Louisiana, recorded 126 mm during the 24 hours ending
  at 04/1200 UTC, and a station in Franklinton, Louisiana, measured 119 mm
  during the same period.  The highest storm total rainfall amounts
  reported in the HPC storm summaries were (all sites in Louisiana):

     Buras            213 mm
     Boothville       183 mm
     Slidell          164 mm

  D. Damage and Casualties

     As Lili passed through the Windward Islands, heavy rains triggered
  mudslides with left four persons dead on the island of St. Vincent.
  There were also four flood-related deaths in Jamaica where severe
  flooding destroyed about 30% of the island nation's sugarcane crop.
  Ham radio reports from Cayman Brac indicated that some damage occurred
  there in the form of downed trees, power lines and poles.  Also, several
  homes experienced roof damage and there was some minor beach erosion.

     In Cuba 190,000 persons were evacuated in advance of the second
  hurricane to affect the western provinces in less than two weeks.
  One report stated that 600,000 persons were evacuated, but that could
  have been the total from both Lili and Isidore.   Lili followed so
  closely on the heels of Isidore on an almost identical track that it
  is difficult, if not impossible, to accurately assess the damage caused
  by each hurricane.  Over 50,000 homes were significantly damaged by
  both hurricanes--32,000 with structural damage and 18,000 with damaged
  roofs.  The two storms caused extensive damage to coffee and banana
  plantations, and the rice harvest was severely impacted.  Also, high
  winds from the hurricanes downed power lines and poles, leading to a
  loss of electricity in many rural and coastal towns.

     Total damage in the United States has been tentatively estimated
  at $800 million.  The Louisiana sugarcane crop was especially hard hit.
  A report from the United Nations Environment Programme (UNEP) stated
  that the total damage estimate from Lili was $2 billion (presumably
  in USD).  Storm chaser Mike Theiss from Key Largo, Florida, has placed
  on his website quite a few pictures depicting flooding and other damage
  from Louisiana.  The address is:  < >

     Many additional articles, reports, and press releases on the effects
  of Hurricane Lili may be found on the ReliefWeb homepage at the
  following URL:>

  E. A Bullet Dodged

     Regarding Lili's dramatic weakening in the hours before landfall in
  Louisiana, many media reports, in describing the storm, used the old
  cliche that the U. S. really "dodged a bullet".  Louisiana was very
  fortunate indeed that Lili weakened before marching ashore.  Given that
  the force of the wind varies with the square of the velocity, 125-kt
  winds would have had over twice the force of winds of 85 kts, not to
  mention the additional destruction that a much higher storm surge would
  have caused.      No doubt many research papers will be written
  investigating the factors that led to Lili's rapid intensification and
  equally rapid demise.

     Just to speculate a little--it may well be that the U. S. dodged two
  bullets.  Had Isidore not made the sudden turn southward into the Yucatan
  Peninsula and subsequently weakened, it would have eventually moved into
  the very warm waters of the Bay of Campeche and away from the influence
  of the landmass to its south.   Some of the models were forecasting
  Isidore to become a Category 4 hurricane, and had this happened, a large
  severe hurricane akin to Hurricane Carla (1961) could have borne down on
  the Texas or western Louisiana coastline.    And to continue the
  speculation a bit further, had the scenario with Isidore described above
  panned out, its stirring of Gulf waters farther east likely would have
  been less, meaning that Lili would have had better thermodynamics to
  accompany its excellent dynamics, which in turn would mean that the
  storm very possibly could have made landfall as a significantly stronger
  hurricane.   And if Lili had bent a little farther east, New Orleans
  or perhaps Biloxi, Mobile, or Pensacola could have had to deal with a
  Category 3 or 4 hurricane.

     Just as Hispaniola's high mountains have saved the U. S. from many
  destructive hurricanes, so Mexico's Yucatan Peninsula in this case
  helped to prevent at least one, possibly two, intense hurricanes from
  striking the United States.

  (Report written by Kevin Boyle and Gary Padgett)


  NORTHEAST PACIFIC (NEP) - North Pacific Ocean East of Longitude 180

  Activity for September:  1 tropical depression
                           2 tropical storms

                       Sources of Information

     Most of the information presented below was obtained from the
  various tropical cyclone products issued by the Tropical Prediction
  Center/National Hurricane Center (TPC/NHC) in Miami, Florida (or the
  Central Pacific Hurricane Center (CPHC) in Honolulu, Hawaii, for
  locations west of longitude 140W):  discussions, public advisories,
  forecast/advisories, tropical weather outlooks, special tropical
  disturbance statements, etc.  Some additional information may have
  been gleaned from the monthly summaries prepared by the hurricane
  specialists and available on TPC/NHC's website.  All references to
  sustained winds imply a 1-minute averaging period unless otherwise

     The following summaries for the three September systems were authored
  by John Wallace of San Antonio, Texas.    A special thanks to John for
  his assistance.

             Northeast Pacific Tropical Activity for September

     In contrast to an active August, the month of September was unusually
  quiet across the Northeast Pacific basin.    Only two tropical storms
  formed, and neither reached hurricane intensity, making this the first
  time since 1972 that no hurricanes formed during the month of September.
  The averages for the entire NEP basin over the period 1971-2001 are
  3.6 tropical storms/hurricanes, 2.2 hurricanes, and 1.1 intense

     On the first day of the month, Hurricane Hernan, which had formed
  in late August, became the second Category 5 hurricane of the season
  while located about 350 nm south-southwest of Cabo San Lucas.  Hernan
  continued to move away from the coast and had dissipated far to the
  west of the Baja by the 6th.   Also, on the 1st, former Hurricane
  Fausto, which had weakened into a tropical depression, briefly regained
  tropical intensity in the central Pacific well to the northwest of
  Hawaii.  (Fausto was not re-upgraded operationally, but the end-of-season
  summary from TPC/NHC stated that it regained tropical storm intensity.)
  The complete reports on Hernan and Fausto can be found in the August

     Early in the month Tropical Depression 11E moved closely enough to
  the Mexican coast that tropical storm warnings were hoisted for the
  southwestern Baja California Peninsula, but the system neither affected
  the coast nor reached tropical storm intensity.   Around mid-month
  Tropical Storm Iselle cranked up into a 60-kt tropical storm while
  located south of the Baja Peninsula, but did not affect the Mexican
  coast.   Short-lived Tropical Storm Julio on 26 September became the
  first NEP cyclone of the year to make landfall in Mexico, moving inland
  near Zihuatanejo.  Julio's rains brought some significant flooding to
  the region.

                          TROPICAL DEPRESSION
                            5 - 8 September

     The disturbance that become Tropical Depression Eleven-E was first
  noted in the NEP on 31 August.  The system tracked westward and slowly
  organized as it trailed powerful Hurricane Hernan--by 5 September it
  had become organized enough to warrant upgrading to Tropical Depression
  Eleven-E at 2100 UTC, located roughly 175 nm west of Manzanillo, Mexico.

    The depression's northwesterly track around the periphery of a
  subtropical ridge to its north threatened to take it close to Baja
  California.  Weighed with uncertainty in the intensity forecast, the
  Mexican government issued a tropical storm watch for the coast north 
  of Cabo San Lucas to Cabo San Lazaro on the west coast of the Baja
  Peninsula which was later upgraded to a tropical storm warning and
  hurricane watch.

     Eleven-E flirted with tropical storm strength on the 6th; its official
  peak MSW was 30 kts, with a minimum CP of 1006 mb at 0900 UTC on the 7th
  when centered about 115 nm southwest of Cabo San Lucas.  The system then
  began a weakening trend due to dry air entrainment, though its CP oddly
  reached 1006 mb again on the 8th during a brief burst of convection.
  Dissipation was rapid, and the final advisory was issued at 1500 UTC on
  8 September with the weakening center about 225 nm west-southwest of
  Cabo San Lucas.  The remnant LOW was no longer evident by the 11th.

     No casualties or other damages are known to have resulted from
  Tropical Depression Eleven-E.

  (Report written by John Wallace)

                         TROPICAL STORM ISELLE
                           15 - 20 September

     The disturbance that became Iselle developed suddenly on 14 September
  as a large cyclonic disturbance south of Acapulco, Mexico.   By 0900 UTC
  on the 15th, satellite data and a report of a 1008-mb SLP from ship WTEE
  at 0100 UTC warranted its upgrade to Tropical Depression Twelve-E.  The
  depression, located approximately 250 nm south of Manzanillo, initially
  tracked west-northwestward, roughly paralleling the coast.

     The depression organized steadily, and at 0000 UTC on 16 September
  became Tropical Storm Iselle, centered about 190 nm south-southwest of
  Manzanillo, Mexico.     The upgrade was based largely on surface
  observations and a ship report.  Concurrent with its upgrade, a tropical
  storm warning was issued for the coast from Lazaro Cardenas to Puerto
  Vallarta.  Easterly shear kept the intensification rate slow for the next
  two days, while a decreased threat prompted the Mexican government to
  drop the warnings for the coast after 1500 UTC on the 16th.  

     Iselle's MSW peaked at 60 kts at 2100 UTC on 17 September, with a CP
  of 990 mb, when it was located roughly 250 nm west-southwest of Cabo San
  Lucas.  Dry air entrainment weakened Iselle thereafter, and its forward
  speed decreased as the ridge to the north weakened due to an approaching
  trough.    Iselle was already well on the way out by the 18th, but a
  renewed burst of convection early on the 19th enabled Iselle to "feel"
  the mid to upper-level trough that had previously left it in a weak
  steering flow.  The storm turned abruptly northeastward toward the Baja
  coast in response, though its forward speed remained little more than a
  drift.  Shear added to the overall hostility of the synoptic environment,
  and Iselle was downgraded to a depression at 2100 UTC on the 19th.  This
  development was interesting, because a ship report of a sustained wind
  as high as 43 kts was discounted based on the weight of satellite data
  and two other ship reports.

     Iselle made an eleventh-hour north-northwestward turn away from the
  coast under the influence of an upper-level LOW to its west.  Weakening
  continued, and the final advisory on Tropical Depression Iselle was
  issued at 1500 UTC on 20 September when the dying circulation was located
  approximately 200 nm west-northwest of Cabo San Lucas.   A cyclonic 
  remnant was no longer identifiable the following day.
     No casualties or damages are known to have resulted from Tropical
  Storm Iselle.

  (Report written by John Wallace)

                         TROPICAL STORM JULIO
                           25 - 26 September

     The progenitor disturbance for Julio was first definitively noted on
  24 September; it may have been extant earlier, but was disrupted by the
  large circulation of Hurricane Isidore to the northeast.  The system
  organized quickly beginning at 0600 UTC on the 25th, and was subsequently
  upgraded to Tropical Depression Thirteen-E at 1500 UTC, located roughly
  90 nm south-southwest of Zihuatanejo, Mexico.

     An upper-level LOW southwest of Baja California steered Julio slowly
  northward, presenting an immediate threat to the coast.  A tropical storm
  warning was issued with the first advisory, extending from Zihuatanejo
  to Punto San Telmo.     Though it was under some easterly shear, the
  depression strengthened to Tropical Storm Julio at 2100 UTC when it was
  located some 30 nm west of Zihuatanejo.  Proximity to land and shear kept
  Julio a minimal tropical storm with a MSW of 35 kts and associated CP of
  1000 mb.  This was Julio's peak intensity.

     (Editor's Note:  The NHC discussion for the 26/0300 UTC advisory noted
  that based on Dvorak intensity estimates of 45 kts from TAFB and SAB,
  plus an earlier 35-kt report from Zihuatanejo, the synoptic intensity
  at 0000 UTC had been raised to 40 kts, but as the storm's center had
  made landfall, the 0300 UTC advisory intensity was brought back down
  to 35 kts.  The end-of-season report on TPC/NHC's website lists the peak
  MSW for Julio at 40 kts.)

     Julio made landfall just west of Lazaro Cardenas, Mexico, near 0300
  UTC on 26 September.   The cyclone made an expected northwestward turn
  once inland, as the upper LOW to its west moved away, and a new ridge
  built in over Mexico.    On its new track Julio roughly straddled the
  coast, in a fashion similar to that of Hurricane Pauline in 1997.  Julio
  was no Pauline, of course, and the rugged terrain of the Mexican coast
  had weakened it to a depression by 0900 UTC on the 26th.  The final
  advisory was issued at 1800 UTC that same day as a closed circulation
  was no longer evident.

     There are conflicting reports of damage from Julio.  The NHC lists
  no significant damages or casualties.  An unofficial source, however,
  states that damage in Zihuatanejo was substantial:  500 houses were
  damaged and 1,000 trees uprooted, in addition to disruption of the
  power grid (1).  A more reliable source states that flash flooding from
  Julio damaged or destroyed 100 homes in Acapulco and Zihuatanejo (2).
  No deaths were reported in either source.    (Editor's Note:  A brief
  mention was made of Julio in a release archived on the ReliefWeb homepage
  which stated that 5000 persons were affected by the heavy rains and
  flooding in Guerrero State.)




  (Report written by John Wallace)


  NORTHWEST PACIFIC (NWP) - North Pacific Ocean West of Longitude 180

  Activity for September:  1 tropical depression **
                           3 tropical storms ++
                           1 super typhoon

  ** - classified as a tropical depression by JMA only

  ++ - one of these was classified as a tropical storm by JMA and NMCC
       but not by JTWC

                         Sources of Information

     Most of the information presented below is based upon tropical
  cyclone warnings and significant tropical weather outlooks issued
  by the Joint Typhoon Warning Center of the U. S. Air Force and
  Navy (JTWC), located at Pearl Harbor, Hawaii.   In the companion
  tropical cyclone tracks file, I normally annotate track coordinates
  from some of the various Asian warning centers when their center
  positions differ from JTWC's by usually 40-50 nm or more.   All
  references to sustained winds imply a 1-minute averaging period
  unless otherwise noted.

     Michael V. Padua of Naga City in the Philippines, owner of the
  Typhoon 2000 website, normally sends me cyclone tracks based upon
  warnings issued by the Japanese Meteorological Agency (JMA) and the
  Philippines' Atmospheric, Geophysical & Astronomical Services
  Administration (PAGASA).  Also, Huang Chunliang of Fuzhou City, China,
  sends me each month tracks obtained from warnings issued by the
  National Meteorological Center of China (NMCC), the Central Weather
  Bureau of Taiwan (CWBT) and the Hong Kong Observatory (HKO).  A very
  special thanks to Michael and Chunliang for the assistance they so
  reliably provide.

     In the title line for each storm I have referenced all the cyclone
  names/numbers I have available:   JTWC's depression number, the 
  JMA-assigned name (if any), JMA's tropical storm numeric designator,
  and PAGASA's name for systems forming in or passing through their
  area of warning responsibility.

     The summaries for Tropical Storms Hagupit and Mekkhala were written
  by Kevin Boyle of Stoke-on-Trent, UK, and with significant contributions
  by Huang Chunliang of Fuzhou City, Fujian Province, China (currently a
  student in Beijing).  A special thanks to Kevin and Chunliang for their

             Northwest Pacific Tropical Activity for September

     As the month of September opened, the remnants of the destructive
  Typhoon Rusa were weakening over the Korean Peninsula and the Sea of
  Japan.  Far to the south and east, powerful Typhoon Sinlaku was taking
  aim on the island of Okinawa, the eye crossing directly over the island
  on the 4th and 5th.  Late on the 7th a somewhat weaker Sinlaku made
  landfall on the Chinese mainland near Wenzhou City.  Typhoon Sinlaku
  was rather destructive with losses estimated at upwards of 4 billion
  yuan.  Sinlaku was also deadly as 32 persons lost their lives due to
  the storm.   Much farther to the east, intense Typhoon Ele had entered
  the Northwest Pacific basin from the Central North Pacific, crossing
  the International Dateline around 0000 UTC on 30 August.  Ele moved
  generally on a slow northward track for several days as it gradually
  weakened, becoming a tropical depression near the 40th parallel late
  on 9 September.  Typhoon Ele's entire track remained east of 165E.
  Complete reports on Rusa, Sinlaku, and Ele can be found in the August
  summary (with Ele's report in the Northeast Pacific basin portion of
  the summary).

     Tropical cyclogenesis waned in the Northwest Pacific basin during
  September after a very active July and August.  Three tropical storms
  formed (one of these not recognized as a tropical cyclone by JTWC) plus
  the year's seventh super typhoon (per JTWC nomenclature).  The active
  pattern in the South China Sea seen during August continued into
  September with two tropical storms forming and making landfall in
  southern China.  Tropical Storm Hagupit formed between Luzon and Hong
  Kong during the second week of the month and moved westward, making
  landfall in Guangdong Province southwest of Hong Kong on the 12th.
  The weak remnants of Hagupit drifted around in the northern South China
  Sea/southern China area for nearly a week following the cyclone's land-
  fall.   A couple of weeks later, Tropical Storm Mekkhala formed south
  of Hainan Dao and moved northward across the island and into the Gulf of
  Tonkin.   Mekkhala reached the Chinese coast in the Guangxi Autonomous
  Region but retreated back out into the Gulf of Tonkin where it eventually

     In between Hagupit and Mekkhala, a tropical depression moved northward
  toward Japan from the Marianas Islands.  South of Honshu the system
  intensified and was named Tropical Storm Changmi by JMA.  NMCC also
  issued tropical storm warnings on the system.  JTWC, however, did not
  regard Changmi as a tropical cyclone and did not issue warnings.  The
  system probably was some sort of baroclinic/hybrid development.   The
  final named storm of the month formed on the 27th east of the Marianas
  and ultimately developed into a super typhoon.  The storm, Higos, passed
  through the northern Marianas and as it began to weaken, got caught in a
  strong southerly steering current and raced northward toward Japan,
  passing over or very near Tokyo at full typhoon intensity.

     A tropical disturbance north of Palau was increasing in organization
  on 13 September and elicited a TCFA from JTWC, but this was cancelled
  early on the 14th after the convection dissipated.  Another disturbance
  was classified as a weak tropical depression by JMA in the Summary
  portion of their High Seas bulletins (implying 25-kt winds).  The system
  was located approximately 625 nm east-northeast of the northernmost
  Marianas Islands around 1800 UTC on 21 September and moved generally
  northward.  The final bulletin referencing this system as a tropical 
  depression was issued at 22/1800 UTC and placed the center about 725 nm
  east-northeast of Iwo Jima.  An interim STWO from JTWC at 0200 UTC on
  the 22nd assigned this disturbance a fair potential for development
  as satellite imagery showed a partially-exposed LLCC with a small but
  persistent area of deep convection over the western portion of the LLCC.
  However, only a few hours later (at 0900 UTC) another STWO was issued
  downgrading the development potential to poor.    Satellite imagery
  depicted an exposed LLCC with minimal deep convection being sheared to
  the west of the LLCC.  (A track for this system was included in the
  September cyclone tracks file based on JMA's bulletins.)

                         TROPICAL STORM HAGUPIT
                          (TC-23W / STS 0218)
                            9 - 17 September

  Hagupit: contributed by the Philippines, means to 'lash' or 'flog'

  A. Storm Origins

     At 0600 UTC, 8 September, JTWC issued a STWO for an area of convection
  located near 19.4N, 126.4E, or approximately 240 nm east-northeast of
  Luzon, that had developed several hours earlier.  This area of convection
  was the seedling that became Tropical Storm Hagupit.    Animated multi-
  spectral imagery at this time indicated a broad LLCC with cycling deep
  convection.  An analysis at 200 mb indicated that the area was under
  weak to moderate wind shear.  The development potential was rated poor
  on this STWO, but the disturbance subsequently began to show signs of
  increased organization.   A TCFA was issued at 09/1900 UTC and JTWC
  issued the first warning on Tropical Depression 23W at 10/0000 UTC,
  placing the center approximately 320 nm east-southeast of Hong Kong, or
  near 19.5N, 119.0E.  JMA was also classifying the system as a tropical
  depression at this time, positioning the center near 19.0N, 118.0E at
  10/0000 UTC.  The system continued to consolidate and was upgraded to
  tropical storm status on the JTWC advisory at 10/1200 UTC.

  B. Track and Intensity History

     The 10/1200 UTC warning relocated the LLCC about 100 nm northwest of
  the previous warning position so that the system was consolidated about
  180 nm east-southeast of Hong Kong, or near 20.6N, 116.9E.  Tropical
  Storm 23W continued on its westward track at around 10 kts, and was
  assigned the name Hagupit at 11/0000 UTC when JMA upgraded the system
  to 35 kts (10-min avg), fixing the centre near 20.4N, 114.1E.  At this
  time Hagupit's outflow in the northern quadrants was being restricted
  by the upper-level ridge which was governing the movement of the cyclone.
  Tropical Storm Hagupit peaked at 11/0600 UTC with its MSW estimated at
  45 kts, and then slowly began to weaken as it approached the Chinese
  coastline.  (NOTE:  Both HKO and NMCC considered Hagupit a much stronger
  system with a 10-min avg MSW of 60 kts at 11/0600 UTC and 11/1200 UTC,
  respectively.  JMA also analyzed Hagupit as a more intense tropical
  cyclone than did JTWC with an estimated 10-min avg MSW of 50 kts.)

     Landfall occurred at 1930 UTC on 11 September approximately 110 nm
  west-southwest of Hong Kong.  JTWC had by this time downgraded Severe
  Tropical Storm Hagupit to a depression.  The system continued to weaken
  as it moved further inland and JTWC issued the final warning on Hagupit
  at 12/0600 UTC.  HKO followed suit six hours later, but NMCC continued
  to issue bulletins through 13/0600 UTC.   Shangchuan Dao, Guangdong
  Province, recorded sustained 10-min avg winds of 49 kts, gusting to
  68 kts, in Hagupit.  Pinggang Town, Yangjiang City, where Hagupit made
  landfall, reported winds of 58 kts during the storm's landfall.  (It is
  uncertain if this is a sustained wind or a gust measurement.)

     The remnants of Tropical Storm Hagupit meandered and continued to
  plague the southern coastline of China for several days.  The broad LLCC
  moved back into the South China Sea and attempted to reform.  Conditions
  for redevelopment were good with excellent diffluence aloft aided by an
  upper-level LOW to the east and a passing trough poleward.  There was
  very little wind shear over the area.  JTWC issued a TCFA at 15/0100
  UTC, upgrading the potential for development to good, but the development
  potential was downgraded to poor on the next STWO issued at 16/0600 UTC
  since the LLCC had lost its deep convection by this time.  After another
  STWO was issued, the system was finally considered properly dissipated
  on the final STWO issued on Hagupit at 18/0030 UTC.  (NOTE: JMA continued
  to issue bulletins on the remnant LOW throughout this period until
  17/1800 UTC.)

  C. Meteorological Observations

     The rainfall reports below were sent by Huang Chunliang of Fuzhou
  City, Fujian Province, China.  A special thanks to Chunliang for
  compiling and sending the information.   Most of the amounts tabulated 
  are 24-hour totals, and I have included only those exceeding 100 mm.

  (1) Daily Rainfall Reports from the Meteorological Service
      of Guangdong (GRMC)

        10/0000--11/0000 UTC:   No amounts exceeding 100 mm

        11/0000--12/0000 UTC: 

           Cities           Stations         Rainfall (mm)
           Yangjiang        Yangjiang          118.0
           Jiangmen         Taishan            107.1
           Jiangmen         Enping             117.4
           Jiangmen         Shangchuan         118.3

        12/0000--13/0000 UTC:

           Cities           Stations         Rainfall (mm)
           Jiangmen         Enping             160.0
           Yunfu            Xinxing            142.1
           Shanwei          Haifeng            119.1

        13/0000--14/0000 UTC: Data missing

        14/0000--15/0000 UTC: Data missing

        15/0000--16/0000 UTC:

           Cities           Stations         Rainfall (mm)
           Foshan           Shunde             118.1
           Zhongshan        Zhongshan          256.8
           Jiangmen         Heshan             174.1
           Shanwei          Shanwei            112.8
           Shanwei          Haifeng            107.9
        16/0000--17/0000 UTC: Data missing

        17/0000--18/0000 UTC: No amounts exceeding 100 mm

  (2) Storm Total Rainfall Reports from the Meteorological Service
      of Guangdong (GRMC) from 11/0000--18/0000 UTC (only amounts
      exceeding 200 mm are listed):

           Cities        Stations      Rainfall (mm)
           ----------    ----------    -------------
           Zhanjiang     Zhanjiang       344.0
           Zhanjiang     Suixi           332.5
           Zhanjiang     Lianjiang       314.2
           Zhanjiang     Wuchuan         247.7
           Zhanjiang     Leizhou         327.5
           Zhanjiang     Xuwen           291.6
           Jiangmen      Heshan          323.7
           Jiangmen      Kaiping         277.9
           Jiangmen      Xinhui          282.5
           Jiangmen      Enping          510.3
           Jiangmen      Taishan         351.9
           Jiangmen      Shangchuan      524.7
           Huizhou       Huiyang         240.1
           Huizhou       Huidong         244.4
           Shenzhen      Shenzhen        346.8
           Shanwei       Shanwei         279.3
           Shanwei       Haifeng         321.8
           Foshan        Shunde          257.2
           Shantou       Shantou         223.3
           Shantou       Chaoyang        200.4
           Shantou       Chenghai        214.8
           Zhuhai        Zhuhai          407.8
           Zhuhai        Doumen          423.3
           Zhongshan     Zhongshan       493.1
           Yangjiang     Yangjiang       567.5
           Yangjiang     Yangchun        322.8
           Maoming       Maoming         292.8
           Maoming       Gaozhou         249.2
           Maoming       Huazhou         272.0
           Maoming       Dianbai         312.5
           Jieyang       Jiexi           328.3
           Jieyang       Puning          299.7
           Jieyang       Huilai          222.3
           Yunfu         Yunfu           258.6
           Yunfu         Luoding         207.1
           Yunfu         Xinxing         318.0

  (3) Miscellaneous Meteorological Observations

      (a) Guangdong Province
          (1) Shangchuan Dao reported sustained winds of 49 kts, gusting
          to 68 kts in the storm.

          (2) A automatic weather station in Pingsha Town, Zhuhai City
          reported the highest daily rainfall amount of 198.2 mm during
          the period from 11/0000 through 12/0000 UTC.

          (3) Pinggang Town, Yangjiang City, where Hagupit made landfall,
          reported winds of 58 kts when the landfall occurred.   It is
          uncertain if this is a sustained wind or a gust measurement.

      (b) Yulin City, Guangxi Region

          12/0000--16/0000 UTC:

          Stations                 Rainfall (mm)
          ---------------------    -------------
          Yulin(the urban area)      312
          Luchuan                    427
          Beiliu                     238
          Bobai                      134
          Rongxian                   114

      (c) Hainan Province

          Qionghai and Haikou recorded rains of 62 mm and 51 mm,
          respectively, during  the period from 18/0000 through
          18/2100 UTC.

      (d) Jiangxi Province

          Suichuan, Jiangxi recorded rains of 156 mm during the 24-hour
          period ending at 14/1200 UTC.

  D. Damage and Casualties

     (1) Guangdong Province

         According to the statistics on the 17th of September, more than
         866,700 people in 131 townships of six cities were affected,
         330 houses destroyed and 43,880 hectares of crop lands flooded.
         The total economic loss caused by Hagupit was more than
         270 million yuan (US$32.53 million).

         Also, eleven fishermen who were reported missing when Hagupit hit 
         Guangdong were feared dead.  This is the only casualties report
         that has been officially confirmed.

     (2) Guangxi Region

         The torrential rains in Yulin, Guangxi, triggered the most serious
         flooding in 30 years.

     (3) Jiangxi Province 

         The death toll from flooding and landslides caused by torrential
         rains associated with Hagupit was placed at 25 in Suichuan County
         of Jiangxi Province, East China.   Around 180,000 local people
         were affected by the disaster.  During the 24 hours beginning at
         0600 UTC on 13 September, rainfall amounts of 150-200 mm fell on
         the county.  The water level of Suichuan River rose for a time to
         99.24 meters, 0.24 meters above the warning line.    A total of
         159 villages were inundated, 110,000 hectares of farmland
         submerged, 400 water facilities destroyed, and 3800 houses and
         180 bridges ruined.

  E. Fuzhou Rainfall Event

     It was somewhat unusual that Hagupit became a large-sized overland 
  depression on the 12th.   Looking back at the satellite images, one
  finds that the easternmost rainband extended as far as the southeast
  Chinese coast, though the western semicircle of the system was quite

     A thunderstorm, which lasted from 2300 on the 12th through 0530 on 
  the 13th Beijing Time (12/1500 - 12/2130 UTC), was considered the
  strongest one in Fuzhou in recent years.  Station WMO 58847 (in Fuzhou
  City) recorded 124.9 mm of rains from the storm on the noisy night.

     In the urban area of Fuzhou, over 1000 hectares of farmland were 
  inundated, resulting in economic losses of 30 million yuan.

     Luoyuan, the northernmost county of Fuzhou, was stricken by rare 
  torrential rains. The county recorded 184 mm of rains during the period
  from 16/0425 through 16/1010 UTC.

     According to the analysis from NSMC/CMA (the National Satellite 
  Meteorological Center, a sub-agency of China Meteorological Agency),
  the Luoyuan rainfall event was caused by the periphery of Tropical
  Depression Hagupit.  (NOTE:  The formal NMCC track for Hagupit ended at
  09/1306 UTC.  This TCWC, however, did mention the system as a tropical
  depression in several Weather Forecasts during the re-generating stage
  of Hagupit.  Two other TCWCs, the Central Weather Bureau of Taiwan and
  the Japanese Meteorological Agency, did not downgrade Hagupit from
  tropical depression status until 17/0000 UTC.)

     Hundreds of houses were inundated with 1700 residents being affected
  by the torrential rains in Luoyuan County.

  (Report written by Kevin Boyle with significant contributions by
  Huang Chunliang)

                         TROPICAL STORM CHANGMI
                                (TS 0219)
                            18 - 25 September

  Changmi: contributed by South Korea, is the name for the flower 'rose'

     Tropical Storm Changmi was the second NWP system of 2002 to be named
  as a tropical storm by JMA, but which was not recognized as a tropical
  storm by JTWC.  The other was Tropical Storm Kalmaegi, which was classi-
  fied as a tropical depression by JTWC.  With Kalmaegi, it was a matter
  of intensity but with Changmi, it was a matter of cyclone type.  JTWC
  regarded Changmi as an extratropical LOW, hence, did not issue tropical
  cyclone warnings on the system.   As was the case with Kalmaegi, NMCC
  did regard Changmi as a tropical storm and issued five warnings on the

     Whether a tropical storm or not, Changmi's roots did lie in the
  tropics.  JTWC issued a STWO at 1600 UTC on 15 September, noting that an
  area of convection had developed approximately 135 nm north-northeast of
  Chuuk.    Animated infrared satellite imagery showed weak cycling
  convection associated with a possible weak LLCC.  A 200-mb analysis
  indicated that the disturbance had good divergence aloft by virtue of
  being situated in the southwestern quadrant of a subtropical ridge.
  Vertical shear was moderate with strong westerlies to the south.  At
  0600 UTC on the 16th the disturbance was located about 240 nm north of
  Chuuk, but by 17/0100 UTC the main area of interest had shifted to a
  position approximately 330 nm east-northeast of Guam.  Cycling deep
  convection was noted in association with a weak LLCC embedded in the
  monsoon trough.

     The potential for development was upgraded to fair at 0000 UTC on
  18 September--a partially-exposed LLCC was located about 440 nm north-
  northeast of Guam with cycling deep convection.   A TCFA was issued for
  the disturbance at 1200 UTC as deep convection had increased and CIMSS
  products indicated good divergence aloft with favorable vertical shear.
  At the same time, JMA classified the system as a 30-kt tropical
  depression; however, it appears that JMA was following another center.
  The JTWC position was approximately 380 nm southeast of Iwo Jima, or
  near 20.1N, 146.0E, while the JMA position was at 18.7N, 145.0E, or
  near Agrihan Island.   As the system moved northward, the northernmost
  (JTWC) center became fully-exposed and the southernmost center (JMA)
  became the dominant one.  JTWC cancelled the TCFA at 1100 UTC on the
  19th, noting that animated infrared imagery indicated that the system
  was rotating cyclonically around and beginning to merge with the second
  circulation to the south.   At 19/1200 UTC, JMA downgraded the tropical
  depression they had been following to a low-pressure area.

     The weak system continued northward with little change on the 20th,
  passing about 175 nm due west of Iwo Jima at 21/0000 UTC.  The STWO from
  JTWC at 21/0600 UTC noted that multiple exposed LLCCs were evident south
  of an area of deep convection.   A 200-mb analysis indicated moderate
  shear associated with an upper-level ridge northeast of the area.  A
  second TCFA was issued at 1000 UTC, locating a partially-exposed LLCC
  with increasing deep convection about 250 nm northwest of Iwo Jima.
  (JMA at this time was classifying the system as a weak (i.e., 25-kt)
  depression.)  At 0200 UTC on 22 September the LOW was located approxi-
  mately 340 nm north-northwest of Iwo Jima.  Multiple exposed LLCCs were
  seen southeast of the deep convection, and a 200-mb analysis indicated
  increasing vertical shear affecting the system as it interacted with
  mid-latitude flow.

     In the meantime, at 22/0000 UTC, JMA suddenly upgraded the system to
  Tropical Storm Changmi with 40-kt (10-min avg) winds, located roughly
  375 nm south-southwest of Tokyo.  (NMCC also initiated warnings on
  Changmi as a 40-kt tropical storm at the same time.)  JTWC, however,
  cancelled the TCFA at 0400 UTC and dropped the system from their tropical
  weather products.  The TCFA cancellation stated that extensive dry air
  entrainment as the system accelerated to the northeast was indicative
  of a developing extratropical LOW.   JMA upped Changmi's intensity to
  45 kts at 22/0600 UTC while NMCC maintained their MSW estimate at 40 kts
  throughout the period for which they issued warnings on the storm.

     Changmi scooted on east-northeastward and by 23/0000 UTC had become
  extratropical approximately 300 nm east of Tokyo.  The storm intensified
  as it became extratropical with JMA upping the winds to 50 kts at 0000
  UTC.  The storm continued northeastward over the next few days and
  intensified into a deep extratropical cyclone with hurricane force winds.
  The storm crossed the Dateline a little after 0600 UTC on the 25th and
  at 1800 UTC, when last referenced in JMA's bulletins, was a 958-mb
  storm with 65-kt winds in the Bering Sea.

     It isn't exactly clear to the author just what the character of this
  cyclone was on 22 September when it was named by JMA.  It seems possible
  that it was some sort of subtropical or hybrid system.  JMA and NMCC
  would likely not have classified it as a tropical storm had it already
  merged with a cold front--the primary signature of a classic extra-
  tropical cyclone.  Infrared satellite imagery around the time it was
  named reveals an exposed center with all the convection to the north,
  so it possibly could be classified as a sheared tropical cyclone.

     No damage or casualties are known to have resulted from Tropical Storm

  (Report written by Gary Padgett)

                         TROPICAL STORM MEKKHALA
                           (TC-24W / TS 0220)
                            23 - 28 September
  Mekkhala: contributed by Thailand, means 'angel of thunder'

  A. Storm Origins

     On 21 September, at 0600 UTC, JTWC issued a STWO for a new area of
  convection located in the South China Sea approximately 375 nm east-
  southeast of Hue City, Vietnam.   The system was depicted in multi-
  spectral imagery as an elongated trough with deep convection associated
  with convergence to the south.  The development potential was rated as
  poor on this STWO, based on moderate wind shear conditions and weak
  divergence aloft.   However, during the next 24 hours conditions became
  more conducive for development and based on this, the 22/0600 UTC STWO
  upgraded the potential for development to fair.  Animated multispectral
  imagery depicted improved organization of the deep convection at this
  time.    Conditions continued to become more favourable with lighter
  vertical wind shear and improved poleward outflow, and three hours later,
  JTWC issued a TCFA.   A 22/0620 UTC AMSU sensor image indicated a well-
  defined LLCC.   The development potential remained good on the 23/0600
  UTC STWO (a second TCFA was issued at the same time) despite the
  disorganized appearance of the deep convection.   QuikScat imagery in
  combination with synoptic reports revealed weak winds near the centre
  with 20 to 25-kt winds to the south and north.

     The first warning on Tropical Depression 24W was issued by JTWC at
  1200 UTC, 23 September, locating the center near 15.1N, 112.1E, or about
  275 nm east-southeast of Hue City,  Vietnam.   (HKO also began issuing
  bulletins at this time). TD-24W was moving north-northwestward at 4 kts
  with a MSW of 25 kts.   Animated enhanced infrared imagery indicated
  that the deep convection continued to cycle and consolidate near the
  LLCC.  However, a 23/1415 UTC SSM/I pass suggested that the mid-level
  circulation centre was displaced just northeast of the LLCC.  A later
  SSM/I pass (at 23/2252 UTC) showed convective rainbands curving into
  the LLCC.  The MSW was increased a little to 30 kts at 24/0000 UTC as
  the system continued to move north-northwestward.  (NMCC began releasing
  warnings at 24/0600 UTC.)      At 24/1200 UTC, TD-24W was centred
  approximately 130 nm south-southeast of Hainan Dao or near 16.2N, 109.9E.
  Animated satellite imagery indicated that deep convection had
  consolidated over the LLCC by this time.

  B. Track and Intensity History

     JTWC upgraded the system to tropical storm status at 24/1800 UTC
  based on the appearance of a banding eye feature and deep convection
  south of the LLCC which was revealed by a 24/1359 UTC SSM/I pass.
  Satellite CI estimates rose to 3.0/3.5 at 25/0000 UTC as the tropical
  cyclone continued moving north-northwestward to a position approximately
  35 nm south-southeast of Hainan Dao.   Animated water vapour imagery
  showed poleward outflow being enhanced by a passing shortwave trough.
  At 25/0300 UTC JMA upgraded the system to tropical storm status,
  assigning the name Mekkhala.  (HKO and NMCC had upgraded the system to
  a tropical storm at 25/0000 UTC.)   Tropical Storm Mekkhala reached
  its peak intensity of 55 kts at 25/1200 UTC as it moved over Hainan Dao.
  (JMA and NMCC both estimated the peak 10-min avg MSW at 45 kts while
  HKO's peak intensity was 40 kts.)  The mid-level ridge that was moving
  Mekkhala on a north-northwesterly course began to weaken as a developing
  longwave trough moved eastward into it.  This caused the cyclone to move
  more in a northerly direction and slow to a speed of 4 to 5 kts.   A
  weakening trend had begun as unfavourable conditions and land effects
  started to take their toll.  By 26/1800 UTC the MSW had fallen to 35 kts.
  Convection began to decrease, covering mainly the southern quadrant of
  the partially-exposed LLCC, as revealed by 26/1307 UTC SSM/I imagery.

     Mekkhala was downgraded to a tropical depression by JTWC at 27/0000
  UTC while located over the Gulf of Tonkin near 21.4N, 108.4E.  However,
  JMA and NMCC continued to classify Mekkhala as a tropical storm (MSW
  10-min avg) until 28/0000 UTC while HKO did not downgrade the system
  until 28/0300 UTC (on the final warning).  At 27/0600 UTC the LLCC of
  Mekkhala had moved slowly northward onto the coast of southern China near
  the city of Gang.  JTWC ceased issuing advisories on the system at this
  time.  Mekkhala came almost to a standstill as it meandered off the coast
  of China.  Associated convection decreased during 28 September and the
  Asian TCWCs had ceased issuing warnings by 28/0600 UTC.   The low-level
  remains were no longer discernible in satellite imagery by 30 September.

     Editor's Note - An infrared image of Mekkhala taken at 1833 UTC on
  27 September was recently sent to me.  This was taken 12 hours after
  JTWC's final warning, but distinctly shows an eye-like feature on
  the coast in the northern Gulf of Tonkin region.  JMA and HKO were
  still estimating the intensity (10-min avg) at 35 kts at that time,
  and NMCC was reporting 45 kts.   JTWC's final warning had been issued
  at 27/0600 UTC with the MSW (1-min avg) estimated at 25 kts, but
  Weizhou Dao in the northern Gulf of Tonkin had reported sustained
  winds between 41 and 47 kts, gusting to hurricane force, around
  0400 UTC.  Also, Beihai on the coast reported sustained gale-force
  winds gusting to storm force (time unknown).  These observations, plus
  the image noted above, suggest that the Asian TCWCs were correct in
  maintaining Mekkhala as a tropical storm through early on the 28th.

  C. Meteorological Observations

     The rainfall reports below were sent by Huang Chunliang of Fuzhou
  City, Fujian Province, China.     A special thanks to Chunliang for
  compiling and sending the information.  Most of the amounts tabulated
  are 24-hour totals, and I have included only those exceeding 100 mm.

  (1) Rainfall Reports from the Meteorological Service of Guangdong

      In Guangdong Province, three counties reported rainfall amounts
      that exceeded 100 mm.

          27/0000--28/0000 UTC:

          Cities              Stations      Rainfall (mm)
          --------            --------      -------------
          Qingyuan            Qingyuan        107.0

          28/0000--29/0000 UTC:

          Cities              Stations      Rainfall (mm)
          ---------           ---------     -------------
          Jiangmen            Shanchuan       163.9
          Zhanjiang           Leizhou         148.1

  (2) Rainfall Reports from Sanya City, Hainan Province

          Periods                       Rainfall (mm)
          --------------------          -------------
          24/0000--25/0000 UTC            104.5
          25/0000--25/0600 UTC            124.0
          25/0000--26/0000 UTC            343.1

  (3)  Rainfall Reports from the Meteorological Services of Hainan and

       23/0000--25/0000 UTC:  

          Qiongzhong, Hainan   114 mm

       24/0000--25/1200 UTC:

          Lingshui, Hainan     230 mm
          Baoting, Hainan      201 mm

       24/0000--26/1200 UTC:

          Sanya, Hainan        478.9 mm
          Lingshui, Hainan     327.9 mm
          Baoting, Hainan      295.9 mm
          Wuzhishan, Hainan    287.2 mm
          Ledong, Hainan       103.4 mm

       25/0000--26/0000 UTC:

          Lingshui, Hainan     227 mm

       26/0000--27/0000 UTC:

          Dongfang, Hainan     164 mm

       27/0000--28/0000 UTC:

          Beihai, Guangxi        316 mm
          Hepu, Guangxi          342 mm
          Lingshan, Guangxi      105 mm
          Weizhou Dao, Guangxi   280 mm

  (5) Wind Observations from Guangxi 

     Weizhou Dao reported sustained winds of Beaufort Force 9 (41-47 kts),
     gusting to force 12 (64 kts) before noon on the 27th (locally).
     Meanwhile, Beihai reported sustained winds of Beaufort Force 8
     34-40 kts), gusting to force 10 (48-55 kts).

  D. Damage and Casualties

  (1) Hainan Province

     A press report indicated that more than 6500 residents who were
  trapped in the floods were rescued by the local government.  Also,
  84 fishermen were rescued at sea, though another one was reported

     Five cities/counties recorded rains exceeding 200 mm during the
  3-day period ending at 27/0000 UTC with Sanya reporting the highest
  amount of 478.9 mm.

     According to the preliminary statistics on 28 September, some
  938,800 residents were seriously affected by the storm with 2500 houses
  collapsing.  Direct economic losses in the province were estimated to
  have been approximately 666.7 million yuan.

  (2) Guangxi Region

     Approximately 1,248,000 residents in Beihai and Qinzhou Cities were
  affected by storm damage.

     In Beihai City, 335 houses collapsed in the storm.  According to the 
  preliminary statistics, the direct economic losses were estimated to
  have been approximately 182.2 million yuan in the city.

  (3) Guangdong Province

     At around 26/0300 BJT (or 25/1900 UTC), Jinhe Town, Xuwen County was
  struck by a tornado in which 95 houses were partly damaged, resulting
  in three injuries.  Also, Taiping Town, Mazhang District was hit by
  another tornado around 26/0500 BJT (or 25/2100 UTC).  Both the tornadoes
  caused severe economic losses in the towns, though no monetary estimates 
  were specified in the official report.

  (Report written by Kevin Boyle with significant contributions by
  Huang Chunliang)

                            SUPER TYPHOON HIGOS
                             (TC-25W / TY 0221)
                         26 September - 5 October

  Higos: contributed by the United States, is the Chamorro word for the
         fruit 'fig'

  A. Storm Origins

     An area of convection developed on 25 September approximately 190 nm
  north of Eniwetok.  Animated visible satellite imagery revealed weak
  cycling convection associated with a weak LLCC while an upper-level
  analysis indicated that the system had good equatorward and poleward
  outflow.  The disturbance was located in the southwest quadrant of the
  subtropical ridge where vertical shear was weak.  By 0600 UTC on the
  26th deep convection was increasing near a partially-exposed LLCC
  and the potential for development had been upgraded to fair.  However,
  at 26/0630 UTC JTWC issued a TCFA for the system, upgrading the
  development potential to good, and at 0900 UTC the first warning on
  Tropical Depression 25W was issued.  The initial MSW was estimated at
  25 kts and the center was located approximately 560 nm east of Saipan,
  moving westward at 13 kts.

     A 26/0953 UTC SSM/I pass depicted a well-defined LLCC with a curved
  rain band in the southern semicircle, and the MSW was upped to 30 kts
  at 1200 UTC.  The depression was tracking westward along a low to mid-
  level ridge situated to its north, and this motion was forecast to
  continue for another 24 to 36 hours.   Convective organization continued
  to increase and by 27/0000 UTC, CI estimates had reached 35 and 45 kts,
  so JTWC upgraded TD-25W to tropical storm status.  The system was then
  located approximately 370 nm east of Saipan, moving west-northwestward
  at 14 kts.

  B. Track and Intensity History

     The MSW was increased to 45 kts at 27/0600 UTC.   At the same time
  both JMA and NMCC upgraded the system to tropical storm status with JMA
  assigning the name Higos.   A 27/1132 UTC SSM/I pass indicated a banding
  eye with increased symmetric deep convection around the LLCC.    The
  intensity was upped to 55 kts at 1200 UTC, and at 1800 UTC JTWC upgraded
  Higos to a 65-kt typhoon, based on CI estimates of 65 and 77 kts.  Higos
  was then located approximately 150 nm north-northeast of Saipan, tracking
  west-northwestward at 18 kts.  Around 28/0000 UTC Typhoon Higos passed
  just south of the island of Pagan.  The maximum sustained wind recorded
  was 51 kts with peak gusts of 99 kts.  The storm continued tracking west-
  northwestward at a pretty good clip and steadily intensified.  An 8-nm
  diameter eye was noted at 1200 UTC, and by 1800 UTC satellite CI esti-
  mates were 90, 102, and 115 kts.  JTWC increased the MSW to 105 kts in
  the 1800 UTC warning, locating the center about 315 nm south of Iwo Jima.
  (JMA and NMCC both had upgraded Higos to typhoon status at 28/1200 UTC.)

     Typhoon Higos intensified rather rapidly on 29 September--the MSW was
  raised to 120 kts at 29/0000 UTC, and at 0600 UTC the intensity was upped
  to 130 kts, making Higos the seventh super typhoon of the year.   JTWC's
  estimated peak intensity for Higos was 135 kts at 29/1200 UTC, based on
  CI estimates of 127 and 140 kts.   Storm-force winds reached outward from
  40-60 nm around the 20-nm eye, and gales extended out up to 150 nm in the
  northeast quadrant and to 110 nm in the southern semicircle.   JMA's and
  NMCC's peak 10-min avg MSW estimates for Higos were 90 kts and 100 kts,
  respectively, and the minimum central pressure estimated by JMA was
  935 mb.  Super Typhoon Higos was located approximately 340 nm southwest
  of Iwo Jima at the time of its peak intensity, and had turned more toward
  the northwest as the subtropical ridge to the north weakened and shifted
  eastward.  By 1800 UTC some erosion of the eyewall had occurred in the
  southwest quadrant so the MSW was reduced slightly to 130 kts.

     Early on 30 September a longwave trough extended from the Yellow Sea
  into the East China Sea and was moving eastward.  This trough was fore-
  cast to interact with Higos and recurve and accelerate the typhoon toward
  the northeast.   At 30/0000 UTC the super typhoon was located roughly
  330 nm west-southwest of Iwo Jima and had already turned to the north
  into a weakness in the subtropical ridge.  Poleward outflow was being
  enhanced by the approaching trough but eyewall erosion continued in the
  western semicircle, suggesting that the system was being sheared from
  the west.   Dry air entrainment continued to increase also and the MSW
  was reduced to 120 kts at 0600 UTC, and further to 110 kts at 1200 UTC.
  Typhoon Higos by then was located about 650 nm south-southwest of Tokyo,
  moving north-northeastward at 15 kts.  The western quadrants continued
  to weaken while enhanced infrared imagery revealed an 18-nm diameter eye.

     On 1 October the typhoon continued to accelerate north-northeastward
  in the southwesterly flow from the longwave trough to the west.   At
  01/0000 UTC Higos was located approximately 380 nm south-southwest of
  Tokyo, moving north-northeastward at 26 kts with the MSW still estimated
  at 110 kts.  (JMA and NMCC were reporting 80 and 90 kts, respectively,
  at 0000 UTC.)  By 0600 UTC the center of Higos was about 185 nm south-
  southwest of the Japanese capital, racing north-northeastward at 32 kts.
  JTWC decreased the MSW slightly to 95 kts on this warning.   The center
  of the typhoon roared inland near Yokosuka at approximately 1100 UTC
  and by 1200 UTC was inland very near Tokyo.  Yokosuka reported peak
  sustained winds of 51 kts (presumably 10-min avg) with gusts to 65 kts
  as Higos made landfall.    Station Irozaki (WMO 47666) in Shizuoka
  Prefecture recorded a peak gust of 103 kts shortly before 0900 UTC.
  This is the highest wind measurement of which the author is aware.  Even
  though Higos was in the process of extratropical transition as it made
  landfall, the very rapid translational speed allowed the storm to move
  to very high latitudes while retaining some tropical characteristics.
  Several stations in Japan reported peak wind gusts exceeding 80 kts, and
  station Urakawa (WMO 47426) on Hokkaido reported a peak gust of 87 kts
  at 01/1943 UTC.

     At 01/1800 UTC, JMA and NMCC downgraded Higos to a tropical storm
  while JTWC maintained minimal typhoon status through 02/0000 UTC.  At
  1800 UTC the storm was located over northeastern Honshu and tracking
  north-northeastward at 51 kts.   Satellite imagery, radar and synoptic
  data indicated that the LLCC had become decoupled west of the mid-level
  vortex.  The extremely rapid translational speed resulted from strong
  poleward steering flow caused by the combination of the longwave trough
  to the west and a mid-level high center to the east.   JTWC downgraded
  Higos to tropical storm status at 02/0000 UTC.  The cyclone was then
  over western Hokkaido about 55 nm north of Sapporo, tracking northward
  at 33 kts.  The final JTWC warning was issued at 0600 UTC on the 2nd,
  placing the by now extratropical cyclone's center approximately 200 nm
  north of Sapporo.   NMCC also issued their final warning on Higos at
  02/0600 UTC, although locating the center about 70 nm west of JTWC's

     The 02/0600 UTC position (from JTWC) was also very near the southern
  tip of Russia's Sakhalin Island, and the cyclone moved northward over
  the island as a slowly weakening extratropical gale.  JMA declared the
  system extratropical at 1200 UTC when the storm was centered along the
  western coast of southern Sakhalin.  The storm had reached the central
  Sakhalin area by 03/0000 UTC and thence turned eastward into the Sea of
  Okhotsk.  The LOW continued eastward, crossing the southern tip of the
  Kamchatka Peninsula around 04/0000 UTC and entering the Bering Sea as
  a 35-kt gale.  By 0600 UTC on 6 October the former super typhoon had
  weakened to a 1000-mb LOW with winds below gale force, located just
  west of the Dateline in the extreme southwestern Aleutian Islands.

  C. Meteorological Observations

     Huang Chunliang has sent me some rainfall and wind observations
  from Japan which he compiled from information given on JMA's website.
  A special thanks to Chunliang for preparing the report and passing it

  (1) Rainfall Observations - The amounts tabulated represent 24-hourly
      totals for the indicated dates.  The dates are local Japanese time
      (UTC+9), so the UTC times would span from 1500 UTC of the previous
      date to 1500 UTC of the indicated date.  Only amounts greater than
      100 mm are listed.

                                   1 October

      Prefecture    Station       WMO Code    Altitude (m)  Rainfall (mm)

      Tokyo         Oshima        47675           74           131.0
      Saitama       Chichibu      47641          232           208.0
      Tochigi       Nikko         47690         1292           181.5
      Yamanashi     Kawaguchiko   47640          860           152.5
      Shizuoka      Ajiro         47668           67           102.5
      Shizuoka      Mishima       47657           21           157.5
      Shizuoka      Shizuoka      47656           14           105.0
      Shizuoka      Hamamatsu     47654           32           118.0
      Shizuoka      Omaezaki      47655           45           131.0
      Aichi         Irako         47653            6           142.0
      Fukushima     Wakamatsu     47570          212           155.0
      Fukushima     Shirakawa     47597          355           102.0
      Iwate         Miyako        47585           43           105.0

                                   2 October

      Prefecture    Station       WMO Code    Altitude (m)  Rainfall (mm)

      Hokkaido      Hiroo         47440           32           112.5

     The following table contains rainfall amounts from some additional
  JMA stations (non-WMO stations).  Only amounts greater than 200 mm are

                                   1 October

      Prefecture    Station       JMA Code    Altitude (m)  Rainfall (mm)

      Kanagawa      Hakone        46161          850           342
      Saitama       Tokigawa      43162          295           203
      Saitama       Urayama       43157          400           215
      Tochigi       Shioya        41181          123           255
      Tochigi       Happagahara   41081         1087           223

  (2) Maximum Sustained Wind and Pressure Observations - The values listed
      represent the peak hourly 10-min mean winds and minimum pressure
      observations for the indicated stations.  If the minimum pressure
      occurred at a different time than the peak sustained wind, the time
      of the minimum pressure is given in parentheses.  The pressure
      measurements have been adjusted to mean sea level.

      Station       WMO Code    MSW (kts)  Time (UTC)    Min SLP (mb)

      Tateyama      47672       35 SSW     01/1100       967.9
      Katsuura      47674       50 SSW     01/1200       974.4 (1100 UTC)
      Choshi        47648       50 SW      01/1200       970.5 (1100 UTC)
      Chiba         47682       37 SW      01/1300       956.8 (1000 UTC)
      Oshima        47675       54 SSW     01/1300       977.5 (1200 UTC)
      Hachijo Jima  47678       45 W       01/0900       981.1 (0800 UTC)
      Ishinomaki    47592       45 ESE     01/1400       973.7 (1500 UTC)
      Miyako        47585       38 SE      01/1500       990.1
      Ofunato       47512       41 SE      01/1600       975.7
      Kushiro       47418       46 SSW     02/0300       991.2 (2200 UTC)
      Nemuro        47420       37 ESE     01/2100       996.0 (2300 UTC)
      Urakawa       47426       46 ESE     01/2000       982.3
      Nii Shima     44206 **    51 W       01/1100       N/A
      Haneda        44166 **    45 ESE     01/1000       N/A
      Shinkiba      44136 **    47 W       01/1200       N/A

      ** - JMA station codes

  (3) Gust Observations - The values given represent the peak wind gust
      recorded in association with the passage of Typhoon Higos.

      Prefecture      Station      Code        Time (UTC)   Gust (kts)

      Tokyo           Kozu Shima   44211 ++     01/0700          58
      Shizuoka        Irozaki      47666        01/0900         103 **
      Chiba           Choshi       47648        01/1200          98
      Fukushima       Iwaki        N/A          01/1310          93
      Miyagi          Ishinomaki   47592        01/1320          80
      Iwate           Ofunato      47512        01/1530          83
      Iwate           Miyako       47585        01/1610          80
      Hokkaido        Urakawa      47426        01/1943          87

      ++ - JMA station code - all others are WMO codes
      ** - second strongest gust recorded in station's history

  D. Damage and Casualties

     Damage in Japan due to Typhoon Higos appears to have been relatively
  light.  There were some houses unroofed, and 80,000 households were
  without electrical power at one point.     A 56,835-ton cargo ship
  registered in the Bahamas and carrying automobiles ran aground on Oshima,
  an island south of Tokyo, but none of the crew of 22 were injured.
  Ferries heading south of Tokyo were cancelled, along with about 200
  evening flights into and out of Tokyo airport.  There was concern that
  the apple and rice crops would be heavily damaged since the storm struck
  just before harvest time, but the author has not learned of any effects
  the storm had on those crops.

     Press reports indicate that five persons lost their lives in Japan due
  to Typhoon Higos.  Two persons died in Chiba Prefecture, near Tokyo, when
  they were electrocuted by electrical cables, and a security guard was
  killed when a plate glass window fell on him in Yokohama, about 30 km
  from Tokyo.  An Indonesian man drowned while trying to secure a fishing
  boat in Miyagi in northern Japan, and the body of a missing woman was
  recovered after being swept away by waves on the coast near Yokohama.

  (Report written by Gary Padgett with significant contributions by Huang


  NORTH INDIAN OCEAN (NIO) - Bay of Bengal and Arabian Sea

  Activity for September:  No tropical cyclones


  SOUTHWEST INDIAN OCEAN (SWI) - South Indian Ocean West of Longitude 90E

  Activity for September:  1 tropical depression **

  ** - treated as a minimal tropical storm by JTWC

                        Sources of Information

     The primary sources of tracking and intensity information for
  Southwest Indian Ocean tropical cyclones are the warnings issued by
  the Tropical Cyclone Warning Centre on La Reunion Island, part of
  Meteo France (MFR), and the Regional Specialised Meteorological Centre
  for the basin.    However, tropical cyclones in this region are named 
  by the sub-regional warning centres on Mauritius and Madagascar with
  longitude 55E being the demarcation line between their respective
  areas of warning responsibility.  The La Reunion centre only advises
  these agencies regarding the intensity of tropical systems.  References
  to sustained winds imply a 10-minute averaging period unless otherwise

     In the companion tropical cyclone tracks file, I occasionally
  annotate positions from warnings issued by the Joint Typhoon Warning
  Center (JTWC) of the U. S. Air Force and Navy, located at Pearl
  Harbor, Hawaii, when they differ from MFR's coordinates by usually
  40-50 nm or more.  The JTWC warnings are also the source of the
  1-minute average maximum sustained wind values included in the
  tracks file.    Additionally, information describing details of
  satellite imagery and atmospheric circulation features included in
  the narratives is often gleaned from the JTWC warnings.

     A special thanks to Kevin Boyle for writing the report on Tropical
  Depression 01S.

         Southwest Indian Ocean Tropical Activity for September

     The first September tropical system to warrant any warnings in
  four years formed early in the month at a very low latitude east of
  the Seychelles Islands.   In September, 1998, La Reunion issued
  warnings on two tropical disturbances, neither of which were deemed
  to have reached tropical depression status (30 kts).   However, the
  second of those systems was classified as a minimal tropical storm
  by JTWC.  This year's system was classified as a tropical depression
  by MFR and as a minimal tropical storm by JTWC.   However, the damage
  reports from the Seychelles certainly suggest that it could have been
  significantly stronger.

                         TROPICAL DEPRESSION
                          (MFR 01 / TC-01S)
                           5 - 8 September

  A. Storm History

     The first mention by JTWC of the disturbance that was to become
  TC-01S was in a STWO issued at 1800 UTC on 27 August.  The disturbance
  was located approximately 80 nm east-northeast of Diego Garcia near
  6.3S, 77.0E, and was accompanied by a weak LLCC and scattered deep
  convection.  The potential for development of the disturbance was
  noted as poor in this STWO.  The development potential remained 
  poor on the next STWO issued at 28/0500 UTC. (The STWOs for 29 and 30 
  August are missing.)  The 31/1800 UTC STWO noted convergence-induced
  cycling deep convection around a weak LLCC with the system embedded in
  a near-equatorial trough.  The development potential remained rated as
  poor in the 02/1800 UTC STWO.  The disturbance remained disorganized and
  upper-level northeasterlies were shearing the deep convection.   The
  disturbance weakened and was no longer considered a suspect area for
  further development by 04/1800 UTC.

     However, a 04/1241 UTC QuikScat pass continued to show a weak
  circulation near 3.6S, 71.4E, or about 240 nm north-northwest of Diego
  Garcia, and there was still surface convergence into the region.  An
  intense burst of convection signaled the start of new development and
  at 05/1500 UTC,  JTWC issued an interim STWO upgrading the development
  potential to fair.  Animated infrared imagery and a mid-level analysis
  indicated a well-marked mid-level circulation, but there was, at this
  time, no evidence of a LLCC.  Also, MFR had begun issuing warnings on
  the disturbance at 05/1200 UTC, indicating 25-kt winds (10-min avg) near
  the centre with 30-kt winds in isolated locations in the southern semi-

     MFR upgraded the disturbance to tropical depression status with 30-kt
  winds at 1200 UTC on 6 September, and JTWC issued their first warning on
  Tropical Cyclone 01S at 06/1800 UTC.  The system was centred near 4.0S,
  57.8E, or approximately 150 nm east-northeast of Victoria in the
  Seychelles.  SSM/I and animated satellite imagery at this time showed a
  small system of approximately 100 nm with a banding feature wrapping into
  the centre from the southeast.  TC-01S was moving west-southwestward at
  11 kts with the MSW estimated to be 35 kts. (This was the peak intensity
  for JTWC--Reunion's peak was 30 kts 10-min avg, which corresponds to a
  1-min avg MSW of 35 kts, so the two TCWCs were in good agreement.)
  The heading turned westerly and satellite imagery and a SSM/I pass at 
  07/0529 UTC depicted deep convection sheared southwest of the LLCC.
  This was caused by upper-level easterlies which began to inhibit further
  development.   The weakening system continued westward at 8 kts until
  the LLCC lost identity and all that remained was cycling deep convection.
  MFR downgraded the system from tropical depression status back to 25 kts
  at 07/1200 UTC, and JTWC decreased the 1-min avg MSW to 30 kts six hours
  later.  The dominant steering flow for this system was a low to mid-
  level ridge extending from the East Coast of Africa.  Both JTWC and MFR
  issued their final warnings on TC-01S at 08/0600 UTC.  The weakening
  centre was then located approximately 100 nm west-northwest of the

  B. Meteorological Observations

     According to press reports heavy rainfall and strong winds badly 
  affected the Seychelles Islands.  The islands of Mahe and Praslin
  recorded rainfall amounts of up to 327 mm and 125 mm, respectively.

     Winds of up to 71 kts (likely gusts) were recorded on Praslin.
  This suggests that TC-01S was possibly significantly stronger than
  the maximum intensity reported by both JTWC and MFR.  Gusts of this
  magnitude near sea level could imply 10-min avg winds of 50 kts.
  However, the altitude at which the observations were made isn't
  known.   The strong winds could have been very localised and not
  representative of the strength of the overall circulation.

  C. Damage and Casualties

     Close to 10 per cent of the Seychelles' population of 75,000 was 
  directly affected by the tropical depression, and 375 families were 
  forced to abandon their homes and seek refuge either with friends or 
  relatives.  The strong winds caused damage to housing, public buildings,
  roads, power supplies and telecommunications.  

     Flooding was severe in Grand Anse, Anse Kerlan, Baie St Anne, Cote
  D'or and Mont Plaisir, where some villages and settlements were
  literally submerged by water.  Many hectares of cultivated land were

     There are no reports of casualties.

  (Report written by Kevin Boyle)



  Activity for September:  No tropical cyclones



  Activity for September:  No tropical cyclones


  SOUTH PACIFIC (SPA) - South Pacific Ocean East of Longitude 160E

  Activity for September:  No tropical cyclones


                               EXTRA FEATURE

     In order to shorten the amount of typing in preparing the narrative
  material, I have been in the habit of freely using abbreviations and
  acronyms.   I have tried to define most of these with the first usage
  in a given summary, but I may have missed one now and then.  Most of
  these are probably understood by a majority of readers but perhaps a
  few aren't clear to some.  To remedy this I developed a Glossary of
  Abbreviations and Acronyms which I first included in the July, 1998
  summary.  I don't normally include the Glossary in most months in
  order to help keep them from being too long.  If anyone would like to
  receive a copy of the Glossary, please e-mail me and I'll be happy
  to send them a copy.


  AUTHOR'S NOTE:  This summary should be considered a very preliminary 
  overview of the tropical cyclones that occur in each month. The cyclone
  tracks (provided separately) will generally be based upon operational
  warnings issued by the various tropical cyclone warning centers.  The
  information contained therein may differ somewhat from the tracking and
  intensity information obtained from a "best-track" file which is based
  on a detailed post-seasonal analysis of all available data. Information
  on where to find official "best-track" files from the various warning
  centers will be passed along from time to time.

    The track files are not being sent via e-mail.  They can be retrieved
  in the following manner:

       (a) FTP to: []
       (b) Login as: anonymous
       (c) For a password use your e-mail address
       (d) Go to "data" subdirectory (Type: cd data)
       (e) Set file type to ASCII (Type: ascii)
       (f) Transfer file (Type: get remote_file_name local_file_name )
           (The files will be named with an obvious nomenclature--using
           September as an example:   sep02.tracks)
       (g) To exit FTP, type: quit

    Both the summaries and the track files are standard text files
  created in DOS editor.  Download to disk and use a viewer such as
  Notepad or DOS editor to view the files.

     The first summary in this series covered the month of October,
  1997.   If anyone wishes to retrieve any of the previous summaries,
  they may be downloaded from the aforementioned FTP site at HRD.  The
  summary files are catalogued with the nomenclature:  sep02.sum, for

    Back issues can also be obtained from the following websites
  (courtesy of Michael Bath, Michael V. Padua and Michael Pitt):>> OR>>

     Another website where much information about tropical cyclones may
  be found is the website for the UK Meteorological Office.  Their site
  contains a lot of statistical information about tropical cyclones
  globally on a monthly basis.  The URL is:>


     JTWC now has available on its website the complete Annual Tropical 
  Cyclone Report (ATCR) for 2001 (2000-2001 season for the Southern 
  Hemisphere).  ATCRs for earlier years are available also.  Recently
  added was the report for the Southern Hemisphere 2001-2002 season.

     The URL is:>

     Also, TPC/NHC has available on its webpage nice "technicolor"
  tracking charts for the 2001 Atlantic and Eastern North Pacific
  tropical cyclones; also, preliminary storm reports for all the 2001
  Atlantic and Eastern North Pacific cyclones are now available, as
  well as track charts and reports on storms from earlier years.

     The URL is:>

     A special thanks to Michael Bath of McLeans Ridges, New South Wales,
  Australia, for assisting me with proofreading the summaries.


  Gary Padgett
  E-mail:  [email protected]
  Phone:  334-222-5327

  John Wallace  (Eastern North Pacific, North Indian Ocean, Western
                 Gulf of Mexico)
  E-mail:  [email protected]

  Kevin Boyle  (Eastern Atlantic, Western Northwest Pacific, South
                China Sea)
  E-mail:  [email protected]

  Simon Clarke  (Northeast Australia/Coral Sea, South Pacific)
  E-mail:  [email protected]


Document: summ0209.htm
Updated: 27th December 2006

[Australian Severe Weather index] [Copyright Notice] [Email Contacts] [Search This Site]