Tropical Cyclones
[Index]
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 2001
[Summaries and Track Data] [Prepared by Gary Padgett]

                  MONTHLY GLOBAL TROPICAL CYCLONE SUMMARY
                      
                              SEPTEMBER, 2001


  (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.)

  SPECIAL NOTE:  Michael Bath, who lives near the northern coast of New
  South Wales, has agreed to be my proofreader for the summaries.  This
  should help speed up the preparation time for the summaries a good bit.
  A special thanks to Michael for agreeing to serve in this capacity.

  ***********************************************************************

                           SEPTEMBER HIGHLIGHTS
                      
  --> Atlantic produces four hurricanes--none make landfall as hurricanes
  --> Northeast Pacific very active with one intense Category 4 hurricane
  --> Northwest Pacific also very active with five typhoons

  ***********************************************************************

               ***** Feature of the Month for September *****
                         
                 SOUTHERN HEMISPHERE TROPICAL CYCLONE NAMES
                            2001 - 2002 SEASON

              TROPICAL CYCLONE NAMES for the AUSTRALIAN REGION
                          
     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.  Technically, Perth's AOR is south of
      10S with Indonesia being responsible for waters north of 10S, but
      I believe the plan is that any rare tropical cyclone which might
      form north of 10S would be named by Perth.

  (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.

  (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 2001-2002 season (** indicates name has already been
  assigned):

          Perth          Darwin        Brisbane        Port Moresby
  -----------------------------------------------------------------------

         Alex **        Bonnie          Bernie **        Epi
         Bessi **       Craig           Claudia **       Guba
         Chris **       Debbie          Des              Ila
         Dianne         Evan            Erica            Kama
         Errol          Fay             Fritz            Matere
         Fiona          George          Grace            Rowe
         Graham         Helen           Harvey           Tako
         Harriet        Ira             Ingrid           Upia
         Inigo          Jasmine         Jim
         Jana           Kim             Kate
         Ken                            Larry
         Linda                          Monica
         Monty                          Nelson
         Nicky                          Odette
         Oscar                          Pierre



           TROPICAL CYCLONE NAMES for the SOUTHWEST INDIAN OCEAN
                      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 2001-2002 season (** indicates name has already been
  assigned):

       Southwest Indian                          South Pacific
  -----------------------------------------------------------------------

     Andre **          Naly                  Trina **      Fili
     Bako **           Orelie                Vicky **      Gina
     Cyprien **        Parina                Waka **       Heta
     Dina **           Quincy                Yolande       Ivy
     Eddy **           Roger                 Zoe           Judy
     Francesca **      Solo                  Ami           Kerry
     Guillaume         Teddy                 Beni          Lola
     Hary              Usha                  Cilla         Meena
     Ikala             Vero                  Dovi          Nancy
     Jery              Wendy                 Eseta         Olaf
     Kesiny            Xora
     Lalita            Yolande
     Mirana            Zaza

  ***********************************************************************
  
                           ACTIVITY BY BASINS

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

  Activity for September:  1 tropical depression
                           4 hurricanes

  NOTE:  Much of the information presented below was obtained from the
  TPC/NHC discussion bulletins issued with every regular advisory.  All
  references to sustained winds imply a 1-min averaging period unless
  otherwise noted.   Some information was also obtained from the monthly
  summary for September prepared by the Hurricane Specialists and
  available on TPC/NHC's website.

     In addition, I received a good deal of information from Peter Bowyer
  of the Canadian Hurricane Centre on the effects several of the tropical
  cyclones had on Newfoundland and the offshore fishing banks.  Some of
  the information was contained in reports prepared by Chris Fogarty of
  the Newfoundland Weather Centre.  A special thanks to Peter for sending
  me these reports.


                  Atlantic Tropical Activity for September
                  ----------------------------------------

     In contrast to the month of August when no hurricanes formed in the
  Atlantic basin, four tropical storms formed during September and all
  reached hurricane intensity--two of these reached Category 3 on the
  Saffir/Simpson scale.  On the average the month of September produces
  3.4 tropical storms with 2.4 reaching hurricane intensity and 1.2
  attaining intense hurricane status.  (These averages based on the
  period 1950-2000.)   Erin, the first hurricane of the 2001 season,
  reached hurricane intensity on 8 September.  This is the latest date
  the first hurricane of the season has formed since 1984.  Hurricanes
  Erin and Felix both became Category 3 hurricanes, but fortunately both
  remained well out in the Atlantic, although Erin brushed Bermuda.

     Tropical Storm Gabrielle formed in the southeastern Gulf of Mexico
  around mid-month and crossed the Florida Peninsula from west to east,
  making landfall near Venice at just under hurricane intensity.
  Gabrielle re-intensified in the Atlantic as it moved away from the
  U. S. coast and reached hurricane intensity as it passed well west
  of Bermuda.  The final hurricane of the month, Humberto, formed in
  a residual trough left over from Gabrielle south-southwest of Bermuda.
  Humberto subsequently passed northward west of the island, eventually
  turning northeastward across the North Atlantic shipping lanes.

     A very interesting feature of the Atlantic tropical cyclones this
  season has been the double peak in intensity all eight of the named
  systems through the end of September have exhibited.   A brief synopsis
  follows:

  (1) Allison - peaked at 50 kts off Texas coast, weakened over land,
      then re-intensified to gale force as a subtropical cyclone over
      southeastern Louisiana and southern Mississippi.

  (2) Barry - reached an initial peak of 40 kts before weakening into a
      tropical depression, then re-intensified to near hurricane force
      as it moved inland in the Florida Panhandle.     (Operationally,
      Barry was maintained as a minimal tropical storm, but the weakening
      to depression status will be reflected in the Best Track.)

  (3) Chantal - actually a triple peak intensity of sorts.  Upgraded to
      a tropical storm east of the Windward Islands but weakened into an
      open wave later the same day.       Redeveloped in the central 
      Caribbean Sea and reached an intensity of 60 kts, weakened slightly
      due to shear in the western Caribbean, then re-intensified to
      60 kts just as it made landfall on the southern Yucatan Peninsula.
      (NOTE: In a telephone conversation with Jack Beven, Jack indicated
      that in the Best Track file Chantal will most likely not be
      considered a tropical storm east of the Antilles.)

  (4) Dean - developed suddenly just north of the Virgin Islands into a
      50-kt tropical storm, but was sheared apart and dissipated the next
      day.  The remnant LOW then moved northward and interacted with an
      old frontal trough, regaining tropical storm intensity north of
      Bermuda and almost reaching hurricane intensity before moving over
      colder SSTs in the North Atlantic.

  (5) Erin - initially developed into a 50-kt tropical storm in the
      tropical Atlantic east of the Lesser Antilles, then weakened into a
      tropical wave due to hostile shear.  A new center formed to the
      north 24 hours later and ultimately developed into Hurricane Erin.

  (6) Felix - became a tropical depression in the eastern Atlantic but
      weakened into a tropical wave 24 hours later.   Redeveloped a
      couple of days later and eventually became Hurricane Felix.  (Felix
      was the only one of the eight which did not have a double peak in
      intensity after becoming a named storm.)

  (7) Gabrielle - formed in the southeastern Gulf of Mexico, made land-
      fall near Venice, Florida, just shy of hurricane intensity, weak-
      ened over land but re-intensified after moving into the Atlantic,
      ultimately reaching hurricane force west of Bermuda.

  (8) Humberto - reaching an initial peak intensity of 85 kts while
      passing west of Bermuda, weakened to minimal hurricane intensity
      a day later, then very surprisingly underwent a rejuvenation with
      winds climbing to 90 kts north of the 40th parallel!

     Interestingly, none of the seven tropical cyclones forming after
  1 October displayed a double peak in intensity except for Hurricane
  Karen, and its initial peak was reached as an extratropical storm
  before undergoing an evolution into a subtropical and later tropical
  cyclone.

     In addition to the four hurricanes, the only other tropical cyclone
  during September was a short-lived tropical depression in the western
  Caribbean.   A tropical wave was noted approaching the Windward Islands
  on 15 September.  The wave moved through the islands into the eastern
  Caribbean on the 16th accompanied by showers, thunderstorms, and gusty
  winds.  The disturbance continued westward across the Caribbean during
  the next few days with little sign of development.  As it reached the
  western Caribbean on the 19th, early morning visible satellite imagery
  suggested that it had become better organized.   Advisories were
  initiated on Tropical Depression #9 at 2100 UTC.  The circulation was
  broad and there appeared to be two centers--one east of Costa Rica
  and one about 100 nm east-southeast of Puerto Cabezas, Nicaragua.
  The strongest and most persistent deep convection was associated with
  the northernmost center, so that center was considered to be the
  primary one.     Upper-level outflow was very circular and quite
  impressive.

     The depression moved steadily westward and by 20/0000 UTC was moving
  inland near Puerto Cabezas.   Although the surface circulation began
  to quickly weaken over the mountainous terrain, the cloud pattern
  remained well-organized.     The remnants of Tropical Depression #9
  subsequently moved across Central America and emerged into the Eastern
  Pacific where they eventually developed into Hurricane Juliette, which
  became the second major hurricane of the season in that basin.



                        Hurricane Erin  (TC-06)
                            1 - 17 September
                        -----------------------

  A. Origins
  ----------

     A vigorous tropical wave left the west African coast on 30 August,
  passing south of the Cape Verde Islands the night of the 30th.
  Environmental conditions were favorable and some of the numerical
  models were suggesting that tropical cyclone development might occur
  in a couple of days.  The wave remained poorly organized on the 31st
  but overnight the convective organization improved.  TPC/NHC initiated
  advisories on Tropical Depression #6 at 2100 UTC on 1 September with
  the center located roughly 650 nm west-southwest of the Cape Verdes.
  A drifting buoy west of the LLCC had reported a northwest wind of
  15 kts and a pressure of 1008.7 mb.  Deep convection had increased
  markedly and this trend continued through the night.  By early morning
  of the 2nd Dvorak numbers were T3.0 from SAB and T2.5 from TAFB, so
  the depression was upgraded to Tropical Storm Erin with 40-kt winds.


  B. Track and Intensity History
  ------------------------------

     Erin set out on a westward track, strengthening slightly to 45 kts
  as the 2nd progressed.  The storm remained poorly organized, even
  though shear was weak and outflow good.  Winds were upped to 50 kts
  at 03/0900 UTC when the cyclone was centered about 950 nm east of the
  island of Barbados.  However, the upper and lower circulation centers
  were still not vertically aligned.    Visible pictures during the
  morning of 3 September revealed an exposed LLCC southwest of the main
  CDO feature.   At 1000 UTC a drifting buoy about 60 nm west of the
  center reported 31-kt winds.  By 1100 UTC the buoy was reporting north
  winds of only 20 kts with an attendant SLP of 1008.4 mb.  The 1500 UTC
  discussion noted that tropical storm-force winds were likely confined
  to the northeast quadrant.

     As Erin moved westward it moved closer to an upper-level LOW which
  induced shear over the cyclone.  Several bursts of convection were
  noted, but the cold clouds had an amorphous shape.  The MSW was reduced
  to 40 kts at 2100 UTC.  However, by early on 4 September the shear had
  decreased and the storm appeared better organized.  NHC raised the MSW
  back to 45 kts at 0900 UTC based on a 0200 UTC buoy report of 41-kt
  winds about 40 nm northeast of the center.  Also, outflow had improved
  and the convective pattern was more circular in appearance.   The
  abatement in the shear, however, was temporary.  The first flight into
  the storm by the Hurricane Hunters on the morning of 5 September found
  a poorly-defined circulation with a minimum pressure of 1013 mb.   A
  peak flight-level wind of 49 kts was measured east of the center.
  Tropical Storm Erin was downgraded to a tropical wave at 05/2100 UTC
  about 250 nm northeast of Guadeloupe, having lost its circulation and
  deep convection due to a hostile environment.  The system appeared to
  have several weak centers of vorticity along the wave axis.

     On the morning of 6 September, a Tropical Weather Outlook issued by
  NHC noted that thunderstorms had increased overnight and that a LLCC
  might be forming on the southwestern side of the new patch of deep
  convection.  A reconnaissance flight was dispatched into the area
  during the afternoon and found a poorly-defined LLCC with winds of
  30-35 kts at flight level in rainbands.     The cloud pattern was
  gradually becoming better organized with an impressive curved band of
  convection to the north.   Advisories were re-initiated on Tropical
  Depression Erin at 2100 UTC, locating the new center about 650 nm
  southeast of Bermuda--well to the north of where the old one had
  dissipated the day before.  Visible satellite imagery on the morning
  of the 7th indicated that the LLCC and outflow had improved since the
  previous day.   Erin was likely a tropical storm at 1500 UTC, but the
  forecaster preferred to wait until the Hurricane Hunters investigated
  the cyclone during the afternoon before upgrading.  The reconnaissance
  crew found a SLP of 1004 mb and flight-level winds of 46 kts northeast
  of the center, so Erin was upgraded once more to tropical storm status
  at 2100 UTC.  Convection was not particularly deep, but the LLCC was
  well-defined with good banding and fair outflow.

     A shortwave trough passing to the north had weakened the steering
  currents and Erin began a slow drift to the north-northwest.  The CP
  had dropped to 999 mb by 08/0600 UTC with 49-kt winds found at the
  925-mb level northeast of the center.  Visible pictures during the
  morning of the 8th indicated a much better-organized storm than 24
  hours previously with good banding and deep convection near the center.
  Dvorak numbers had reached T3.5, so the MSW was upped to 45 kts at
  1500 UTC.   A strengthening ridge to the north was steering Erin on a
  slow northwesterly track.  Based on the official forecast track, a
  hurricane watch was issued for Bermuda.   During the afternoon visible
  imagery revealed a well-defined cloud band wrapping around the center,
  indicating that a banding eye was likely forming.  A reconnaissance
  flight found a pressure of 994 mb with flight-level winds of 76-80 kts,
  so Erin was upgraded to a hurricane at 2100 UTC when located roughly
  350 nm southeast of Bermuda.   A flight into the hurricane during the
  evening found winds of 84 kts and a CP of 992 mb, and concurrent
  satellite intensity estimates were 77 kts, so the intensity was bumped
  up to 75 kts at 09/0300 UTC.  Erin was sporting an eye 35-40 nm in
  diameter with good outflow, especially to the north of the center.

     A reconnaissance flight around 0600 UTC on 9 September found a
  flight-level wind of 101 kts at 700 mb with an attendant pressure of
  982 mb, so the MSW was increased to 90 kts at 0900 UTC, making Erin a
  Category 2 hurricane on the Saffir/Simpson scale.  The eye continued
  to become better defined, and the next flight into the storm at 1800
  UTC found that Erin had deepened into a major Category 3 hurricane.
  Peak winds at 700 mb were 118 kts while a GPS drop into the north-
  eastern eyewall showed a spot surface wind of 111 kts with mean
  boundary layer winds around 100 kts; the official MSW was set to
  105 kts as a compromise.  The CP had fallen to 968 mb--a little high
  for 105 kts, but environmental pressures were also rather high.  The
  center of Hurricane Erin was located about 100 nm east of Bermuda at
  the time it became the season's first major hurricane.

     The 105-kt intensity was maintained for 24 hours.  The cloud tops
  warmed a bit during the night, but the eye became even better defined.
  Erin continued moving to the north-northwest, and by 1500 UTC on the
  10th was centered approximately 150 nm north of Bermuda, and was
  displaying the classic signature of a major hurricane with a well-
  defined eye 30 nm in diameter.  A NOAA P3 research plane flew into
  Erin during the afternoon, and data suggested that the winds had likely
  weakened a bit to 100 kts or less.  Wind profile data indicated that
  Erin had formed concentric eyewalls.   During the evening stepped-
  frequency microwave radiometer data and GPS drops indicated that
  surface winds had decreased to around 80 kts, although one drop
  reported 89 kts at an altitude of 17 m.  Dvorak intensity estimates,
  however, were still 90 kts, so the MSW was lowered to 85 kts at 11/0300
  UTC.   By 0900 UTC on 11 September, Erin had reached the ridge axis
  and was preparing to recurve to the northeast.  The intensity had
  dropped to 80 kts, where it remained for 18 hours.

     By early on 12 September Erin had recurved and was moving in an
  eastward direction.  The intensity was reduced to 75 kts, but as the
  hurricane was moving over a ribbon of warm water, the weakening process
  was slow.  Large swells from Erin reached the northeastern U. S. during
  this period.   On the 13th the storm turned more to the northeast,
  although at a slow pace.  The eye remained large and well-defined with
  Erin still moving over 26 C SSTs.  The MSW was lowered to 70 kts at
  13/0600 UTC and remained pegged there for 24 hours.   Late on the
  13th the eye was still apparent in infrared imagery, but eyewall
  convection had become asymmetric, perhaps due to the storm's 
  encountering cooler waters.   

     By early on 14 September Erin had accelerated to 17 kts.  The eye
  was still visible but the MSW was lowered to 65 kts.  As the day
  progressed Erin's eye disappeared and the convection became displaced
  north of the center--signs that the storm was beginning extratropical
  transition.  Satellite intensity estimates remained at 65 kts however,
  so the cyclone was maintained as a hurricane.  Buoy 44625, located
  northwest of the center, reported a pressure of 988.2 mb at 1800 UTC.
  By 15/0300 UTC the remaining convection had become confined to the
  northwest quadrant.  TAFB was still reporting 65 kts, but SAB had
  classified the storm as extratropical, so the final NHC advisory was
  issued, downgrading Erin to a 60-kt storm that was rapidly becoming
  extratropical.   Erin at this time was located very near Cape Race,
  Newfoundland, and the remnant extratropical storm continued moving
  rapidly northeastward the next couple of days.  The last reference
  to the system in MPC's High Seas Forecasts was at 17/0000 UTC when
  it was a 40-kt gale near 65N, 35W, still moving northeastward.

     In summary, the peak intensity of Hurricane Erin (per TPC/NHC's
  operational advisories) was 105 kts from 09/1800 UTC through 10/1800
  UTC.    The minimum pressure measured by the Hurricane Hunters was
  968 mb at 09/1800 UTC.  Gale-force winds extended out from the center
  150 nm to the northeast and 100 nm elsewhere while hurricane-force
  winds reached outward 40 nm in the eastern semicircle and 30 nm in
  the western semicircle.


  C. Meteorological Observations
  ------------------------------

     In addition to the reconnaissance data and buoy reports incorporated
  into the above section, I have received some information from Peter
  Bowyer of the Canadian Hurricane Centre regarding observations made in
  southeastern Canada and in offshore waters.  The Laurentian Fan Buoy
  (WMO 44141) reported wave heights to 9 m as Erin passed by early on
  14 September.   As the storm passed a few miles south of the Nickerson
  Bank Buoy, seas built to 7 m with a maximum wave height of 14 m being
  reported at 15/0000 UTC.

     Rain began falling in Newfoundland late on the 13th from a quasi-
  stationary frontal zone extending northeast of Erin.  The heaviest
  rain was concentrated along the southern coast of Newfoundland west
  of the Burin Peninsula where more than 100 mm was recorded at Sagona
  Island and Burgeo.  A second wave of rain, associated with Erin's
  cloud shield, brought amounts near 75 mm from Bonavista to the Burin
  Peninsula.   A few of the higher storm totals are:

     Sagona Island - 131 mm
     Burgeo - 102 mm
     Bonavista - 78 mm
     St. Lawrence - 66 mm
     Lethbridge - 63 mm

     No hurricane force winds were reported on land.    Some exposed
  coastal stations reported gusts of 58 kts in northwesterlies behind
  Erin while the strongest winds in the easterlies ahead of the storm
  were at Cape Race, where a peak gust of 52 kts was recorded at 2124
  UTC on the 14th.  Both Cape Race and Bonavista reported gusts to 58 kts
  early on the 15th.    St. John's and Grates Cove reported gusts to
  45 kts and 49 kts, respectively, also early on the 15th.


  D. Damage and Casualties
  ------------------------

     No reports of any damage or casualties resulting from Hurricane Erin
  have been received.



                         Hurricane Felix  (TC-07)
                              7 - 19 September
                         ------------------------

  A. Origins
  ----------

     A tropical wave left the African coast on 3 September and marched
  westward across the tropical Atlantic.  This wave never developed into
  a tropical cyclone, but did influence and interact with the following 
  wave which eventually became Felix.     The cyclone tracks file for
  September contains a track of this leading disturbance which was sent
  to the author by Roger Edson.  No advisories were issued by NHC on this
  system--it was treated as a tropical wave--although some of the
  Tropical Weather Discussions indicated that a broad circulation
  was associated with the wave.    Roger estimated that the LOW was
  accompanied by winds up to 30 kts, which is plausible--many tropical
  waves are accompanied by winds of 30 kts or stronger without forming a
  well-defined low-level circulation.

     The next tropical wave to emanate out of Africa entered the Atlantic
  on 5 September.     A 1009-mb LOW was noted on the 6th as the well-
  organized disturbance passed south of the Cape Verdes.   By the after-
  noon of the 7th, the system had become better organized with banding
  features and good outflow and was classified as a tropical depression
  at 07/2100 UTC when located about 350 nm west of the Cape Verdes.
  During the night of 7 September Tropical Depression #7 displayed a band
  of strong convection, but no deep convection was located near the
  center.   By the morning of the 8th the depression consisted of a
  broad circulation with only a few bands of convection, and by afternoon
  no evidence of a LLCC could be seen in satellite imagery, so the system
  was downgraded to a tropical wave and advisories discontinued.

     In the meantime, the tropical LOW preceding TD-07 had become quasi-
  stationary in mid-ocean, and by the morning of 9 September the remnants
  of the depression were combining with the leading wave to produce a
  broad low-pressure area about 1100 nm east of the Lesser Antilles.
  Advisories on TD-07 were re-initiated at 0900 UTC on 10 September with
  the center located about 525 nm west of the earlier final position at
  08/1800 UTC, or about 1000 nm east of Martinique.  Deep convection had
  persisted near the LLCC while outflow had become more pronounced.
  Also, some nearby buoys had reported pressures around 1008 mb.  Visible
  pictures on the 10th brought a complication, however.  Two distinct
  LLCCs were evident, one consistent with the previous west-northwestward
  motion while a weaker circulation to the southeast was associated with
  the weaker tropical wave that had preceded the depression.  The deepest
  convection was concentrated east of the northern vortex, which was
  considered to be the primary center.

     There was some question during the night of the 10th as to whether
  or not the system was maintaining a closed circulation, but depression
  status was maintained pending the availability of visible images the
  next morning.  Organization improved, however, and at 11/0900 UTC all
  three satellite estimates were near tropical storm strength.  By 1500
  UTC the LLCC had become better defined under deep convection and the
  outflow had gradually improved, so the depression was upgraded to
  Tropical Storm Felix at this time.    The newly-christened tropical
  storm was located about 800 nm east-northeast of Guadeloupe and was
  moving north-northwestward around the periphery of a subtropical ridge.


  B. Track and Intensity History
  ------------------------------ 

     For the first 24 hours of its life as a tropical storm, Felix 
  remained at minimal tropical storm intensity.  Convection, while deep,
  was displaced to the northeast of the LLCC.   Morning visible images
  on 12 September indicated that the cyclone had become better organized
  with the center embedded in deep convection and an eye-like feature
  trying to develop.  The MSW was increased to 55 kts at 1500 UTC and to
  60 kts at 2100 UTC.    Felix was upgraded to a 65-kt hurricane at 
  13/0300 UTC based on Dvorak estimates, partially-contaminated QuikScat
  data, and a doughnut-shaped feature in SSM/I imagery.   The storm by
  this time was moving north, or slightly east of due north, and 
  strengthened quickly after reaching hurricane intensity.  The advisory
  issued at 13/1500 UTC noted that Felix had developed an eye embedded
  within the CDO with plenty of banding and fair outflow.  Satellite
  intensity estimates from TAFB and SAB had reached T5.0, so the MSW
  was increased to 90 kts.  By 1800 UTC the satellite presentation of the
  storm had continued to improve--the eye was distinct and surrounded by
  very deep convection with plenty of curved bands.  TAFB had increased
  their Dvorak estimate to T5.5, so the MSW was upped to 100 kts, making
  Felix the second Category 3 hurricane of the season.  The storm at this
  time was centered roughly 1200 nm southwest of the Azores, moving
  northward.

     Felix remained at its peak intensity of 100 kts for 18 hours, then
  began to slowly weaken.     By 1500 UTC on the 14th the eye had 
  disappeared in infrared imagery and was not well-defined in visible
  pictures.  Also, water vapor imagery indicated that some southwesterly
  shear was impinging on the cyclone.  The MSW was reduced to 95 kts
  at 1500 UTC and to 90 kts at 2100 UTC.  Felix was by this time moving
  northeastward at 13 kts, and this motion had become east-northeasterly
  by 15/0300 UTC.   By the morning of 15 September the hurricane was
  moving eastward and continuing to show signs of weakening.  Visible
  imagery suggested that mid- to upper-level winds were undercutting the
  impressive cirrus outflow and inducing mid-level shear.  The MSW was
  left at 90 kts, although the discussion noted that this might be on
  the high side.  Around 1500 UTC a drifting buoy (41644) reported a
  pressure of 975.8 mb in the northwest quadrant of the eye.  Although
  deep convection persisted near the LLCC, the areal coverage had shrunk
  and, based in part on the buoy report, the MSW was lowered to 85 kts
  at 2100 UTC.  At 16/0300 UTC satellite intensity estimates from TAFB,
  SAB and AFWA were 90 kts, 77 kts, and 65 kts, respectively, and the
  eye had reappeared in METEOSAT-7 infrared imagery, so the intensity 
  was left at 85 kts.

     On the 16th the cloud pattern associated with Felix became more
  ragged with cloud tops warming and convection eroding over the south-
  western semicircle--an upper-level trough appeared to be impinging
  on the western side of the tropical cyclone.  The MSW was lowered to
  80 kts at 0900 UTC.  About this time, drifting buoy 44765 reported a
  SLP of  981.4 mb from a position about 60 nm northeast of the center.
  The NHC discussion noted that central convection had decreased, but
  that the eye had expanded to a diameter of 80 nm.    There were
  indications during the afternoon that the shear over Felix might be
  lessening--there was a burst of deep convection noted in the north-
  east quadrant.  Drifting buoy 44765 was in the eye of Felix for most
  of the morning and reported pressures ranging from 973.8 mb at 1100 
  UTC to 976 mb at 1500 UTC.  The intensity was reduced to 75 kts at
  2100 UTC and to 65 kts at 17/0300 UTC.  Satellite intensity estimates
  at this time were below hurricane force, but the earlier buoy data had
  suggested that Dvorak estimates were a little low.

     The storm's eastward motion had slowed to 8 kts or less by the
  morning of 17 September as a mid-latitude trough had bypassed the area.
  At 1500 UTC Felix was located about 250 nm south of the northwestern-
  most Azores and was drifting to the southeast.  Northwesterly shear had
  displaced the convection to the southeast of the LLCC and the cyclone
  was downgraded to a 60-kt tropical storm.  SSTs in the area were about
  24-25 C.   Felix continued to slowly weaken as it remained quasi-
  stationary or drifted slowly southward.  Winds were down to 45 kts by
  0300 UTC and continued to drop as the day progressed.   By 0900 UTC
  little if any deep convection remained as the storm drifted south-
  southwestward, retracing its path over waters that had been cooled by
  upwelling.   Felix was being steered by flow on the eastern side of
  a mid-level subtropical ridge.   No deep convection was observed at
  1500 UTC and the MSW was reduced to 35 kts.  Drifting buoy 44765, which
  had passed through the eye of Felix two days earlier, again encountered
  the storm and reported a pressure of 1001.2 mb.  Felix was downgraded
  to a tropical depression at 2100 UTC, based on nearby ship reports of
  25 kts.   As the 19th began Felix consisted of a broad but rather well-
  defined circulation devoid of any deep convection.  A QuikScat pass
  indicated maximum winds of 25 kts, and ship KAKF confirmed these winds
  at a location well south of the center.  Since models indicated no
  chance of regeneration,  NHC issued the final advisory on the former 
  Category 3 hurricane at 0300 UTC on 19 September with the remnant
  LOW drifting slowly southward, well south of the Azores. 

     Hurricane Felix was at its estimated peak intensity of 100 kts from
  1800 UTC on 13 September through 1200 UTC on the 14th with the central
  pressure estimated at 965 mb during this period.  At its peak Felix was
  a large, severe hurricane.    Gales covered an area between 300 and
  400 nm in diameter and hurricane-force winds reached outward from the
  center 30-40 nm in all directions.


  C. Meteorological Observations
  ------------------------------

     No reconnaissance missions were flown into Hurricane Felix.  The
  only surface reports available to the author were the buoy and ship
  observations mentioned in the preceding section.


  D. Damage and Casualties
  ------------------------

     No damage or casualties are known to have resulted from Hurricane
  Felix.



                      Hurricane Gabrielle  (TC-08)
                            11 - 21 September
                      ----------------------------

  A. Origins
  ----------

     A Tropical Weather Outlook issued by TPC/NHC on the afternoon of
  8 September noted that disorganized cloudiness and thunderstorms
  extended from the eastern Gulf of Mexico across the Florida Peninsula
  into the northern Bahamas and adjacent Atlantic waters.  This area of
  disturbed weather was associated with a weak trough of low pressure.
  The trough and unsettled weather persisted in the area for the next
  few days, and on the morning of 11 September surface data revealed the
  existence of a definite closed circulation over the southeastern Gulf
  of Mexico.   The deep convection did not show much curvature, but was
  near enough to the LLCC to warrant classifying the system as a tropical
  depression.  The first advisory on TD-08 was issued at 2100 UTC with
  the center located about 125 nm west-northwest of Key West, Florida.
  The maximum winds were estimated at 25 kts, and with northerly shear
  over the area and an upper-level vortmax moving south-southeasterly
  toward the depression, strengthening was forecast to be slow.

     By the early morning of the 12th the depression was a little better
  organized.  Buoy 42003 in the eastern Gulf reported a pressure below
  1009 mb with a 24-hour drop of 4 mb.  Deep convection remained to the
  south of the broad and ill-defined center, but convective banding had
  increased in the eastern semicircle across southern Florida.  As the
  day wore on an upper-level shortwave trough near the north-central
  Gulf Coast modified the upper flow over the system, resulting in a
  decrease in the shear.   Satellite imagery revealed increasing
  organization of the depression and buoy data indicated pressures were
  continuing to fall, so the MSW was bumped up to 30 kts at 2100 UTC.
  A reconnaissance plane visited the area during the evening and found
  a very broad circulation with convection not very well organized near
  the center, but rather oriented in a linear band extending from the
  northern Yucatan coast to the southwest Florida coast.

     Deep convection began to persist near the center on the morning of
  13 September, and ship WPKD, located about 70 nm southeast of the
  center, reported 32-kt sustained winds at 0600 UTC with a pressure of
  1006 mb.     Outflow was good in the southeastern semicircle but
  restricted to the northwest due to upper-level shear.   Around 1200
  UTC ships DCUW and WCZ5238 reported winds of 35 and 40 kts,
  respectively, and both TAFB and SAB had assigned a T2.5 Dvorak rating,
  so the depression was upgraded to Tropical Storm Gabrielle with 40-kt
  winds at 1500 UTC, located about 200 nm west-northwest of Key West.
  The storm was still experiencing some shear--the center was partially-
  exposed on the western edge of the deep convection.


  B. Track and Intensity History
  ------------------------------

     A reconnaissance flight during the afternoon found the center well
  northeast of the previous position.   It was not clear whether the
  center had reformed or if the storm had been moving northeastward.
  WSR-88D animation suggested the latter, so warnings were ordered up
  for portions of the Florida coast.    Gabrielle was located in a 
  shearing environment due to an upper-level trough over the eastern 
  Gulf of Mexico.  However, the SHIPS model indicated that the storm 
  could become a hurricane before moving inland.  The reconnaissance
  flight at 14/0000 UTC found a pressure of 996 mb and flight-level winds
  of 52 kts.    The system was somewhat sheared with bursts of deep 
  convection forming east of the center.  Heavy rainbands from Gabrielle
  were spreading onshore over the southern Florida Peninsula and the
  Keys.   During the night the center of Gabrielle appeared to keep
  reforming to the northeast under the convection.  The reconnaissance
  plane at 0600 UTC found a central pressure of 992 mb but encountered
  winds of 70-75 kts at flight level near the Florida coast.  The storm
  at that time was moving northeastward at 9 kts.

     The center of Gabrielle reached the southwestern Florida coast near
  Venice around 1200 UTC on 14 September.  The storm almost reached
  hurricane intensity just before making landfall.  Reconnaissance data
  indicated that the pressure fell to 980 mb, and 700-mb winds as high
  as 81 kts were measured at flight level.  These, however, were near the
  Venice C-MAN station, which at the time was reporting gusts to only
  40 kts.  It appears that the higher winds at flight-level were not
  mixing down to the surface.  Gabrielle spent the next 24 hours crossing
  the Florida Peninsula, emerging into the Atlantic just northeast of
  Cape Canaveral by around 1200 UTC on the 15th.    The convection
  decreased markedly after the storm made landfall with the strongest
  remaining convection found in bands well-removed from the center to the
  north and east.  The MSW was maintained at 40 kts during Gabrielle's
  overland trek, being supported by synoptic observations of 35-40 kts
  near the Florida East Coast, particularly the St. Augustine C-MAN
  station.

     While over Florida Gabrielle began to take on the appearance of an
  occluded frontal LOW more so than a tropical cyclone.  The circulation
  was large with a non-symmetrical wind field and devoid of any deep
  convection near the center.  The fact that the storm maintained a CP
  less than 999 mb and winds of 40 kts while over land suggests that its
  nature was not purely tropical.  By 15/1500 UTC the center had moved
  offshore into the Atlantic near Cape Canaveral.  The NHC discussion
  noted that it resembled a subtropical cyclone with the strongest winds
  and convection removed from the center and an upper-level cyclone
  superimposed over the surface center.  A reconnaissance flight found
  winds near 60 kts in the southeast and northwest quadrants with a CP
  of 998 mb, so the MSW remained at 40 kts.   Two ship reports of 43 and
  44 kts at 1800 UTC led to the intensity being raised slightly to 45 kts
  at 2100 UTC.  Convection had moved a little closer to the center, but
  Gabrielle still had a "frontal" look about it.   SAB had classified
  the system as extratropical, and TAFB was calling it a subtropical
  storm.  The cyclone was moving northeastward roughly parallel to the
  southeastern U. S. coast and by 2100 UTC had accelerated to 13 kts.

     The intensity was increased to 50 kts at 16/0300 UTC based on ship
  reports of 50-kt winds in the northwest quadrant near the coast and
  supported by reconnaissance data and GOES cloud drift winds which
  indicated winds to 50 kts in other quadrants.  Six hours later it
  appeared that Gabrielle was about to become extratropical--a
  reconnaissance mission reported that the center had become more
  elongated with little or no core temperature rise at 850 mb.  However,
  the storm was still moving over warm waters and a slight increase in
  convection near the center was noted.   By 1500 UTC reconnaissance
  reports and satellite imagery indicated that Gabrielle had become
  better organized and had strengthened some.  At 1140 UTC a pressure of
  995 mb was measured and at 1328 UTC a plane found winds of 67 kts
  northwest of the center, so the MSW was increased to 55 kts.  The
  temperature of the core had also warmed and satellite imagery indicated
  increasing outflow in the northwest semicircle, suggesting that
  Gabrielle was still a warm-core tropical cyclone as it sped northeast-
  ward at 17 kts.

     During the 16th the storm continued to develop deep convection on
  the northern side of the ragged, exposed LLCC.  Satellite intensity
  estimates ranged from 35 kts from SAB to 55 kts from KGWC.  At 0000
  UTC on 17 September a GPS drop made by a NOAA research aircraft
  reported winds of 68 kts near the center just above the surface, so
  Gabrielle was upgraded to a 65-kt hurricane at 17/0300 UTC when
  centered approximately 325 nm west of Bermuda.  Even so, the system
  still appeared to be sheared and asymmetric with the deep convection
  to the north and east of the center.  A dry slot wrapping around into
  the southern semicircle produced a comma-cloud signature more typical
  of subtropical storms than tropical cyclones.  A reconnaissance flight
  into Gabrielle around 1200 UTC measured winds to 85 kts at 1500 m along
  with a pressure of 983 mb, so the MSW was bumped up to 70 kts in the
  1500 UTC advisory.
  
     By afternoon of the 17th Gabrielle was an impressive system in
  satellite imagery despite the southwesterly shear affecting the storm.
  Visible pictures indicated that a banding eye might be trying to form
  as moderate convection had wrapped 3/4 of the way around the LLCC.
  Also, the storm had good outflow to the north.  However, at 18/0300 UTC
  Gabrielle was downgraded to a 55-kt tropical storm.  The upper-level
  center was being displaced to the northeast of the LLCC by the south-
  westerly shear.  By 0900 UTC some deep convection had redeveloped just
  north of the LLCC, so the intensity was maintained at 55 kts.  However,
  Gabrielle was well northeast of Bermuda at this time and beginning to
  move over progressively cooler SSTs.   Six hours later some deep
  convection still persisted near the center, but the cloud pattern was
  beginning to look more like an evolving extratropical low-pressure
  system.  The MSW of 55 kts in the 1500 UTC advisory was based on
  tropical and extratropical satellite intensity estimates of 50 and 
  55 kts from TAFB and KGWC, respectively.

     Deep convection continued to persist near Gabrielle's center through
  the afternoon, but the appearance of the storm was more that of an
  extratropical LOW.  The MSW was increased to 60 kts at 2100 UTC based
  on a report from ship WMLH, located about 90 nm northwest of the 
  center.   Observations from the ship in conjunction with AMSU data
  indicated that Gabrielle was still a warm-core system.   NHC issued
  their final advisory on Gabrielle at 19/0300 UTC with the storm
  centered roughly 375 nm southwest of Cape Race, Newfoundland.  A
  recent QuikScat pass had indicated windspeeds near 60 kts, and earlier
  a buoy had reported a pressure of 977.6 mb, but all cloud tops were
  warmer than -50 C and there was no deep convection anywhere near the
  LLCC, so Gabrielle was declared extratropical.  At 1200 UTC on the 19th
  the center of Post-Tropical Storm Gabrielle (the term used by the
  Canadian Hurricane Centre) passed about 50 nm south-southeast of Cape
  Race.  The last reference to the system in MPC's High Seas Forecasts
  placed the center near 59N, 31W, at 1800 UTC on 21 September, still
  generating 50-kt winds.
 
     Hurricane Gabrielle's peak intensity of 70 kts was attained at
  17/1200 UTC and was maintained for twelve hours with an attendant
  minimum CP of 983 mb.  However, in the Gulf of Mexico phase of
  Gabrielle, a reconnaissance plane measured a pressure of 980 mb at
  14/1009 UTC, shortly before the tropical storm made landfall near
  Venice, Florida.  At the time of Gabrielle's peak intensity on the
  17th, gales covered an area about 450 nm in diameter with the radius
  of hurricane-force winds 25 nm in all directions.   Shortly before
  the storm made landfall in Florida, gales extended outward from the
  center 125 nm to the northeast and 100 nm to the southeast.


  C.  Meteorological Observations
  -------------------------------

     As Tropical Storm Gabrielle approached the west coast of Florida
  during the early morning of 14 September, Dry Tortugas reported winds
  of 38 kts, gusting to 44 kts.  Around mid-morning on the 14th the
  AWS at Venice reported sustained winds to 50 kts with a peak gust of
  63 kts.  During the afternoon, St. Petersburg on the west coast and
  St. Augustine on the Atlantic coast both reported gusts to 47 kts.
  The St. Augustine AWS (elevation 16.5 m) reported sustained winds of
  48 kts at 14/2200 UTC and 41 kts at 15/0100 UTC while Gabrielle's
  center was over the peninsula.  At 15/1800 UTC, ships DHER and OZQP2
  reported sustained winds of 44 and 43 kts, respectively, while north-
  west of the center.     According to the NHC monthly summary for
  September, rains in excess of 250 mm fell over portions of west-central
  Florida and a storm surge of 1.5 m above normal affected the west coast
  of the peninsula.

     Bermuda reported wind gusts to 42 kts early on 17 September as
  Gabrielle passed slightly more than 300 nm west of the island.  The
  storm had just been upgraded to a hurricane at that time.  At 1800 UTC 
  on 18 September, when Gabrielle was southwest of Newfoundland, ship
  WMLH, located about 90 nm northwest of the center, reported a pressure
  of 984.5 mb with the air temperature 23.9 C and the dew point 22.2 C.
  This observation, in conjunction with AMSU data, helped to establish
  that Gabrielle was still a warm-cored system.  Gabrielle's center
  passed over or very near the Laurentian Fan Buoy (44141) just before
  midnight on 19 September.   Wave heights exceeded 9 m and the reported
  SLP was 977.6 mb and falling.    The Nickerson Bank Buoy (44251),
  located about 80 nm northwest of Gabrielle, reported seas of 5 m.

     In Newfoundland light rain began falling on the 18th over the
  southern Avalon and Burin Peninsulas, well north of the storm's center.
  The rain became heavier on the 19th, and between 0600 UTC and 1200 UTC
  extremely heavy rain fell across the eastern Avalon Peninsula as
  Gabrielle approached to within around 70 nm due south of Cape Race.
  Much of the rain was due to embedded thunderstorms in the circulation
  around Gabrielle.  Cape Race recorded 48 mm in one hour between 0900
  and 1000 UTC, and St. John's West measured 27 mm between 0600 and 0700
  UTC.   Some storm totals include:

     Cape Race - 162 mm
     St. John's West - 127 mm
     St. John's Airport - 119 mm

  Rainfall amounts west of the Avalon Peninsula were less than 35 mm.

     Strong northerly winds were experienced over eastern sections of
  Newfoundland with gusts to 54 kts or higher recorded on all three
  peninsulas:  Avalon, Bonavista, and Burin.  Some of the peak winds
  reported were:

     Cape Race - 70 kts at 19/1257 UTC (maximum sustained 57 kts at
                 1200 UTC)
     St. John's West - 48 kts at 19/1404 UTC
     St. John's Airport - 47 kts at 19/1114 UTC
     Grates Cove - 59 kts at 19/1816 UTC
     Bonavista - 56 kts at 19/1900 UTC
     Gander - 34 kts at 19/1600 UTC
     Sagona Island - 54 kts at 19/1247 UTC
     St. Lawrence - 56 kts at 19/1227 UTC


  D. Damage and Casualties
  ------------------------

     The heavy rains of Gabrielle led to major river flooding in west-
  central Florida and on the lower St. Johns River.   There was also
  wind damage to roofs, mobile homes and trees.    Total damage was
  estimated at $230 million.  A drowning death along the Alabama coast
  was attributed to Gabrielle.

     In Newfoundland extensive damage due to flooding occurred in St.
  John's.  Numerous basements were flooded and some streets were under-
  mined.  Local brooks and lakes swelled beyond capacity, inundating
  parks and backyards.  (No dollar amount of this damage was available.)



                       Hurricane Humberto  (TC-10)
                            21 - 28 September
                       ---------------------------

  A. Storm Origins
  ----------------

     After Tropical Storm Gabrielle had crossed Florida and moved on to
  the northeast and strengthened as a sort of quasi-subtropical storm,
  a residual trough, somewhat resembling a front, continued advancing
  southeastward, reaching the general area between Puerto Rico and
  Bermuda by 19 September.  On that date a Tropical Weather Outlook
  issued by NHC at 1530 UTC noted that a weak LOW accompanied by some
  scattered convection was located about 475 nm northeast of Puerto
  Rico.  Afternoon satellite pictures indicated a circulation might be
  forming, although environmental pressures were quite high.  By midday
  on the 20th convection was better organized around a well-defined
  surface LOW about 350 nm northeast of Puerto Rico.

     The LOW gradually became better organized through the night, and
  advisories were initiated on Tropical Depression #10 at 1500 UTC on
  21 September.   The system had moved west-northwestward to a position
  roughly 450 nm south of Bermuda by this time.  Both TAFB and SAB were
  estimating winds at 25 kts and deep convection was developing near the
  center.  The depression had a very impressive satellite signature with
  plenty of banding, convection and cirrus outflow, but a reconnaissance
  flight during the afternoon did not find a well-enough defined LLCC to
  make a formal fix.      Wind observations, however, did show a weak
  surface circulation about 40 nm west-southwest of the upper-level
  circulation center.  An upper-level LOW to the southwest of TD-10 was
  still inducing some shear over the system.

     At 22/0000 UTC the intensity estimates from TAFB and SAB were both
  35 kts, but ship KAKF, located near the depression's center, reported
  winds of only 22 kts with a pressure of 1011.8 mb.  By 0600 UTC, SAB's
  estimate had reached 45 kts, and the system displayed a nicely curved
  convective band in the eastern semicircle.  However, multispectral
  infrared imagery showed some exposed low-level cloud lines west of the
  main convection.  The 1200 UTC Dvorak estimates from all three agencies
  were T3.0 (45 kts), and the cloud pattern was well-organized with
  curved bands and fair outflow, but since on the previous day the system
  was not as intense as satellite estimates had suggested, the forecaster
  chose to maintain depression status pending the next reconnaissance
  mission.   The winds, however, were increased to 30 kts.

     An early afternoon reconnaissance flight into the cyclone found a
  fairly strong tropical storm with flight-level winds of 57 kts in the
  west quadrant and a central pressure of 998 mb.   A special Tropical
  Update at 1830 UTC upgraded the depression to Tropical Storm Humberto,
  located about 250 nm south-southwest of Bermuda and moving north-
  westward at 10 kts.  The initial intensity was set at 50 kts, and the
  2100 UTC discussion noted that Humberto had a well-defined, tightly-
  wound surface center.   The storm was forecast to gradually turn to
  the north around the periphery of the subtropical ridge to the east.


  B. Track and Intensity History
  ------------------------------

     By 0300 UTC on 23 September a CDO feature had developed, although
  the MSW remained at 50 kts.  A drifting buoy about 40 nm northwest of
  the center had reported winds of 25 kts and a SLP of 1011.9 mb at
  23/0000 UTC, indicating that Humberto was a rather small tropical
  cyclone.   The 0600 UTC reconnaissance flight found a pressure of
  994 mb with 65-kt winds at 450 m both east and west of the developing
  25-nm diameter eye.  Convection had consolidated near the center and
  outflow was good, so the intensity was upped to 55 kts.  By mid-morning
  on the 23rd Humberto displayed a cold symmetric CDO with good banding
  north and east of the center, and an 1100 UTC SSM/I pass revealed a
  doughnut-like eye feature.  Estimates from TAFB and SAB had reached
  65 kts, but the MSW was increased to only 60 kts since another plane
  was due to reconnoiter the storm at 1800 UTC.

     The reconnaissance flight at midday on the 23rd found a central
  pressure of 984 mb and peak flight-level winds of 86 kts at the 700-mb
  level northeast of the center, but a concurrent GPS drop near the eye-
  wall measured only 56 kts at the surface.  However, the mean boundary
  layer windspeed was 79 kts, so the MSW was increased to 70 kts, making
  Humberto the fourth hurricane of the season and also for the month of
  September.   Humberto was then centered approximately 125 nm west-
  southwest of Bermuda, moving northward at 9 kts.  Observations during
  the evening from a multi-aircraft mission indicated that Humberto
  had strengthened further.  Eyewall dropsonde data from a NOAA research
  airplane showed winds of 87 kts at the surface in the northeast
  quadrant.  Stepped-frequency microwave radiometer data also indicated
  similar winds, so the MSW was increased to 85 kts at 24/0300 UTC,
  making Humberto a Category 2 hurricane on the Saffir/Simpson scale.
  The minimum pressure was 983 mb--a little high for 85 kts--but the
  cyclone was embedded in a higher-than-normal sea-level pressure
  environment.

     Humberto's tenure as a Category 2 hurricane, however, was quite
  brief.  The reconnaissance plane at 0600 UTC found that the pressure
  had risen to 989 mb, and the highest flight-level wind encountered
  was 68 kts.   Satellite imagery suggested some southwesterly shear
  impacting the hurricane.  The MSW was lowered to 80 kts, and was
  further reduced by 5 kts on each subsequent advisory until Humberto
  was a minimal 65-kt hurricane at 25/0300 UTC.  Oddly, during the
  morning of the 24th, Objective Dvorak numbers were slightly higher
  than during the night.  Cloud tops warmed significantly during the
  day, and another multi-aircraft mission into the storm during the
  evening indicated that Humberto had weakened into a minimal hurricane.
  GPS drops measured surface winds around 62-64 kts and radar
  observations indicated that the eyewall was open over most of the
  southern semicircle.   The weakening trend halted, however, and
  throughout the 25th Humberto held on to hurricane intensity, displaying
  an intermittent eye-like feature surrounded by deep convection.  The
  hurricane had recurved and was moving northeastward in the general
  direction of Newfoundland.

     Weakening and eventual extratropical transition were forecast, but
  Humberto had one more trick up its sleeve.   The NHC discussion at
  0900 UTC on 26 September noted that an eye surrounded by deep
  convection had persisted throughout the night, and with TAFB rendering
  a Dvorak number of T4.5 (77 kts), the MSW was increased to 70 kts.
  This improvement in Humberto's appearance coincided with a jog to the
  north, probably in response to ridging near the cyclone downstream of
  a large cutoff LOW over the Great Lakes.  By 1500 UTC the 15-nm eye
  had become better defined and cloud tops had cooled even more.
  Dvorak numbers from both TAFB and SAB had reached T5.0, so Humberto's
  winds were bumped up to 90 kts, making the storm once more a
  Category 2 hurricane.  The hurricane at this time was north of the
  41st parallel about 300 nm southeast of Halifax, Nova Scotia.  This
  strengthening at such a northerly latitude was most unusual.  Humberto
  was located over SSTs of 25-26 C just south of a sharp SST gradient,
  and a building upper-level ridge near and north of the storm had
  apparently sheltered it from stronger westerlies.

     However, like its first flirtation with Category 2 intensity,
  Humberto was not able to maintain the 90-kt intensity for very long.
  By afternoon the eye was becoming indistinct and the MSW was lowered
  to 80 kts at 2100 UTC.  The convective pattern was becoming asymmetric
  with most of the cold clouds shifting to the northeast semicircle.
  Humberto was likely feeling the effects of cooler SSTs and stronger
  westerly flow, its forward speed having increased to 13 kts.   The
  Canadian Laurentian Fan Buoy (44141) reported a pressure of 990 mb
  at 2100 UTC, and a SSM/I overpass at 2310 UTC showed a partial eye-
  wall to the north.  Although Dvorak intensity estimates remained at
  90 kts, Humberto's organization was decreasing and the intensity was
  reduced to 65 kts at 27/0300 UTC.   Based on buoy reports, SSTs along
  and near the forecast track of Humberto were ranging from 15-20 C.

     On the 27th Humberto's track took an almost due eastward turn.
  The storm was downgraded to a 60-kt tropical storm at 0900 UTC.  The
  convection was showing less banding and was becoming amorphous as
  the cyclone accelerated eastward.  Satellite intensity estimates were
  still at hurricane intensity, but given the stabilizing cool SSTs, it
  was deemed unlikely that any hurricane-force winds were making it down
  to the surface.  During the morning westerly flow aloft increased, and
  by 1500 UTC Humberto was sailing eastward along a sharp SST gradient
  at 28 kts.   During the afternoon the cloud pattern deteriorated with
  the convective cloud mass becoming elongated and distorted and removed
  from the LLCC.  NHC issued the final tropical advisory on Humberto
  at 2100 UTC with the storm located about 500 nm east-southeast of Cape
  Race and racing slightly south of due east at 27 kts.  The storm was
  rapidly becoming extratropical and was forecast to be absorbed into
  a very large cyclonic circulation over the North Atlantic during the
  next day or two.

     Hurricane Humberto reached its first peak in intensity at 0300 UTC
  on 24 September when the MSW and central pressure, based upon data
  from a reconnaissance flight, were 85 kts and 983 mb, respectively.
  Hurricane-force winds extended out 25 nm northeast and southeast of
  the center and 15 nm elsewhere.   Gales reached outward 90 nm in the
  eastern semicircle and 35 nm in the western semicircle.   Humberto's
  primary peak occurred after the surprise re-intensification on the
  morning of 26 September, but the 90-kt MSW and 970-mb central pressure
  were only estimates based on satellite imagery.   At the time of this
  second peak in intensity, winds to hurricane force extended outward
  30 nm from the center in all directions while 34-kt winds reached out
  60 nm in the northern quadrants and 120 nm to the southeast.


  C. Meteorological Observations
  ------------------------------

     I have received no synoptic reports associated with Humberto other
  than the buoy observation referenced in the above section.


  D. Damage and Casualties
  ------------------------

     No reports of damage or casualties resulting from Hurricane Humberto
  have been received.   According to information from Peter Boywer of the
  Canadian Hurricane Centre, Humberto was exclusively a marine event and
  had no impact on inland areas in the Canadian Maritimes.   Like its
  predecessors, however, Humberto did pass over the Laurentian Fan and
  Southwestern Grand Banks marine areas, being the fourth tropical
  cyclone in less than a month to bring storm or hurricane-force winds
  to those areas.



                       ADDENDUM TO AUGUST SUMMARY
          Impact of Tropical Storm Dean in Southeastern Canada
          ----------------------------------------------------

     Since the August summary was issued, Peter Bowyer of the Canadian
  Hurricane Centre has sent me some information on the effects of
  Tropical Storm Dean in Newfoundland and on the Canadian marine banks.
  As Dean crossed the southwestern Grand Banks on 28 August, many ships
  and buoys to the right of the storm track and out to 200 nm reported
  gale-force winds.  There were no confirmed gales to the left of the
  track, however.   Hibernia (WMO 44145) reported south-southeasterlies
  of 47 kts at an anemometer height ot 139 m just before Dean passed to
  the north.

     A large area of 4 to 6-metre seas was produced to the southeast of
  the track across the southern Grand Banks.  Ship PDHY (41.7N, 49.5W)
  at 28/1800 UTC reported a wave of 5.7 m while Tail-of-the-Bank Buoy
  (44140) experienced seas of 5.7 m at 1600 UTC approximately 120 nm
  to the right of the track.  At about the same distance to the left of
  the storm's track, seas built to 3.5 m at the Nickerson Bank Buoy
  (44251) at 1800 UTC.

     Only light southeasterly winds of 10-15 kts were experienced over
  Newfoundland.  A band of rain displaced well to the north (250 to
  300 nm) of Dean's track remained quasi-stationary over the eastern
  portion of the island.    Heavy rains occurred between midnight and
  midday on the 28th from the Bonavista Peninsula to the Burin Peninsula.
  Some of the highest storm totals include:

     Bonavista - 107 mm
     Lethbridge - 76 mm
     St. Lawrence - 72 mm

  The capital of Newfoundland, St. John's, recorded a much lower 24 mm
  while Gander netted only 19 mm.

  ***********************************************************************

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

  Activity for September:  2 tropical depressions
                           2 tropical storms
                           3 hurricanes
                           
  NOTE:  Much of the information presented below was obtained from the
  TPC/NHC discussion bulletins issued with every regular advisory (CPHC
  for locations west of 140W.)  All references to sustained winds imply
  a 1-min averaging period unless otherwise noted.

     The narratives for all of the named storms were written by John
  Wallace of San Antonio, Texas.   Because of the interaction between
  Gil and Henriette, those cyclones are covered in a single report.
  A very special thanks to John for devoting much of his school holiday
  time to working on the summaries.



             Northeast Pacific Tropical Activity for September
             -------------------------------------------------

     The month of September was quite active in the Northeast Pacific
  basin.  Five tropical storms/hurricanes formed during the month, well
  above the annual average of about three.  Three of the storms reached
  hurricane intensity with one becoming a major hurricane.  The averages
  for September are two hurricanes with one reaching major hurricane
  intensity (Category 3+ on the Saffir/Simpson scale).   Early in the
  month Hurricane Gil and Tropical Storm Henriette moved in close
  proximity to each other and engaged in a binary interaction.  Around
  mid-month Tropical Storm Ivo formed near the Mexican coast and prompted
  the issuance of tropical storm warnings for portions of the Mexican
  coastline.  The storm, however, remained offshore with minimal impact
  onshore.

     The most significant cyclone of the month was Hurricane Juliette.
  USAF Reserves Hurricane Hunters flew some missions into the hurricane,
  and on one of these, measured a central pressure of 923 mb--the second
  lowest pressure ever actually measured in an Eastern Pacific hurricane.
  Juliette later brought tropical storm and hurricane conditions to
  portions of the Baja California Peninsula.  In its last gasp, Juliette,
  after weaking to below depression status, exceptionally re-intensified
  over the northern reaches of the narrow Gulf of California.    The
  system never regained tropical storm status but did generate some
  well-organized, albeit shallow, convection.    The final named cyclone
  was short-lived Hurricane Kiko, which was a minimal hurricane for only
  six hours.

     In addition to the five systems mentioned above, two tropical
  depressions formed in the Central North Pacific west of 140W.  The
  first of these, TD-01C, was very brief, being in advisory status for
  only 18 hours on 11 September.  The first advisory, issued by the CPHC
  at 11/0000 UTC, placed the center about 400 nm southeast of Hilo.  The
  system was moving westward fairly quickly and was not forecast to
  reach tropical storm intensity.  By 1200 UTC the depression had lost
  all its convection and the final advisory was issued at 1500 UTC,
  placing the dissipating center about 400 nm south-southeast of South
  Point on the Big Island.

     Tropical Depression 02C formed about 750 nm southeast of Hilo on
  23 September.  Unlike its predecessor, TC-02C was forecast to intensify
  into a tropical storm; however, this did not happen.  The depression
  sailed west-northwestward at a low latitude, passing about 500 nm
  due south of South Point at 1200 UTC on the 24th.   The forecast for
  strengthening did not verify due to shear associated with an upper-
  level trough west of the Hawaiian Islands, along with interference
  from a weak disturbance near Johnston Island.  The final CPHC advisory
  was issued at 0900 UTC on 25 September and placed the dissipating
  center about 650 nm south-southwest of Honolulu.



                        Hurricane Gil  (TC-08E)
                            4 - 9 September
                        -----------------------

                                 AND

                   Tropical Storm Henriette  (TC-09E)
                            4 - 8 September
                   ----------------------------------

  A. Storm Origins
  ----------------

     The track and development history of Hurricane Gil and Tropical 
  Storm Henriette is a classic case of direct tropical cyclone 
  interaction, also known as binary interaction, or the Fujiwhara effect.
  Because of their interaction and almost identical life spans, the two
  systems will be covered in a single summary.

     The origin of Gil appears to be unrelated to a tropical wave.  A 
  cyclonic disturbance first became apparent along the ITCZ on 
  2 September and steadily organized.  By 0900 UTC on 4 September, the 
  LOW's organization on satellite imagery was sufficient to warrant its 
  upgrade to Tropical Depression Eight-E about 1075 nm west of 
  Manzanillo, Mexico.    The tropical cyclone tracked slowly west-
  northwestward, turning more westward as it was upgraded to Tropical
  Storm Gil at 1800 UTC on 4 September, located some 1125 nm west of
  Manzanillo.

     Henriette's origin can apparently be traced back to a vigorous 
  tropical wave that exited the African coast on 24 August.  It quickly 
  weakened, however, and tracked uneventfully across the basin.  A LOW 
  formed along the wave axis on 1 September as it entered the tropical 
  Eastern Pacific and slowly organized.  By 1500 UTC on 4 September 
  Dvorak estimates justified its upgrade to Tropical Depression Nine-E,
  just six hours after Eight-E's first advisory, while located
  approximately 275 nm west-southwest of Manzanillo.  Its initial west-
  northwesterly track turned more westward on the 5th.  Easterly shear
  from the same ridge that was its primary steering influence hindered
  intensification.  Nine-E was poorly-organized, and its future was in
  doubt until a QuikScat pass supported its upgrade to Tropical Storm
  Henriette at 2100 UTC on 5 September about 700 nm west of Manzanillo.
  At this time, Gil and Henriette presented an odd pair--Henriette was
  the larger circulation, but Gil was much better organized.


  B. Track and Intensity History
  ------------------------------

     Gil's westward motion slowed on the 5th, and it briefly stalled 
  while Nine-E/Henriette tracked steadily westward.  At 0300 UTC on 
  6 September, Gil was upgraded to hurricane status while Henriette 
  remained a 40-kt tropical storm centered east-northeast of Gil.

     Exactly when Gil and Henriette first interacted is uncertain, but 
  the interaction manifested itself in earnest beginning on the 6th. 
  Henriette's track changed from westerly to west-northwesterly while 
  accelerating slightly around Gil's northeast flank, in spite of being 
  the larger cyclone.  Gil tracked slowly due west (perhaps "held back"
  by Henriette) and strengthened, reaching its peak MSW of 85 kts and
  minimum CP of 975 mb at 2100 UTC on 6 September roughly 1175 nm west-
  southwest of Cabo San Lucas.   Henriette intensified as well on the
  6th and reached its peak MSW of 55 kts, with a CP of 994 mb, at 0300
  UTC on 7 September while centered roughly 825 nm west of Cabo San
  Lucas.  At the time of its peak, Henriette was 460 nm northeast of Gil.
  The overall track of both systems remained westerly, since it appears
  that they were both embedded in a larger-scale gyre south of the
  subtropical ridge.

     Henriette weakened soon after its peak, in spite of excellent out-
  flow, after crossing into unfavorably cool waters.  It continued to 
  track west-northwestward at a brisk pace on the 7th to finally overtake
  Gil in longitude on the 8th.  Some detail should be given to the events
  on the 7th and 8th, as these were arguably the most interesting days of
  the interaction between the two systems.

     Gil weakened on the 7th as well, due to interference from
  Henriette's outflow.  It maintained hurricane strength while it made 
  an abrupt turn to the northwest, bearing north-northwestward by the end
  of the day.    The magnitude of Henriette's influence on Gil is
  interesting, considering that by 2100 UTC on the 7th, Henriette was
  generating only weak convection while Gil remained a solid hurricane.
  This suggests that the size of a tropical cyclone's circulation is more
  important than its intensity in binary interaction.   Perhaps Gil's
  influence was dominant, however, and the necessarily reciprocal nature
  of binary interaction only made it appear as if Henriette was the 
  dominant circulation.   Given Henriette's ultimate fate, this is a 
  possibility.  Perhaps more likely is that Henriette was the dominant 
  circulation at first but that it was eventually usurped by Gil by 
  virtue of its deterioration.

     A weak Henriette overtook Gil in longitude early on the 8th.  Gil's
  forward motion, deliberate for most of its life, accelerated briefly
  but dramatically to the north-northwest, peaking at 22 kts at 1500 UTC
  on the 8th, at which time it weakened to tropical storm strength.
  Gil's acceleration coincided with Henriette's comparably dramatic 
  acceleration and turn to the west-southwest as the two systems 
  co-revolved.  The final advisory on Henriette, as a tropical
  depression, was issued at 1500 UTC on 8 September with the center
  located about 1125 nm east of Hilo, Hawaii.  By the time of its last
  advisory, Henriette was little more than a convectionless vortmax
  lacking a closed circulation, centered some 250 nm west of Gil.

     Gil scarcely fared better on the 8th.  Shear, probably coupled with
  cool SSTs, took a heavy toll on the tropical storm, and by late that 
  day Gil was a weakly convective whorl.  Gil turned west-northwestward
  on the 8th as part of the continued co-revolution with Henriette's 
  remnant LOW.    On the 9th Gil was downgraded to a depression as it 
  turned west-southwestward.  It feebly held on to tropical cyclone
  status as stable air intruded into the system, further eroding its
  already weak convection.  The final advisory on Tropical Depression
  Gil was issued at 2100 UTC on 9 September, placing the center about
  970 nm east of Hilo.  At the time of the final advisory, Gil's and
  Henriette's remnants comprised two co-revolving LOWs embedded in a
  larger cyclonic gyre.     Henriette's  remnant LOW apparently had
  dissipated, or had been absorbed by Gil's remnant LOW, by the 10th.
  Gil's remnant, however, remained identifiable until the 15th, after
  which it dissipated in the Pacific well north of Hawaii.


  C. Damage and Casualties
  ------------------------

     No casualties or damage are known to have been caused by Gil or 
  Henriette as the two cyclones performed their do-si-do far removed
  from any populated shores.



                       Tropical Storm Ivo  (TC-10E)
                             10 - 14 September
                       ----------------------------

  A. Storm Origins
  ----------------

     Ivo developed from a tropical wave--possibly of African origin--that
  was first noted in the eastern Caribbean on the 3rd of September.  
  It tracked uneventfully across the Caribbean and Central America and
  entered the eastern tropical Pacific on the 7th.  A LOW formed along
  the wave axis on the 10th and organized enough to warrant its upgrade
  to Tropical Depression Ten-E at 2100 UTC that day, centered roughly
  100 nm south-southwest of Acapulco, Mexico.

     The cloud envelope of the incipient disturbance and Ten-E itself 
  were massive, but it was only superficially healthy.  Initial Dvorak 
  ratings were low for a depression while easterly shear interfered 
  with its circulation.  Ten-E tracked westward at first, south of a 
  mid-level ridge.  At 0600 UTC on the 11th, ship ZDBE2 reported 37-kt 
  easterly winds and a 1006.5 mb pressure while located 135 nm north-
  northeast of the center.  This data, along with a Dvorak rating of 
  2.5 from TAFB,  supported Ten-E's upgrade to Tropical Storm Ivo at 
  0900 UTC on 11 September about 225 nm south-southeast of Manzanillo.


  B. Track and Intensity History
  ------------------------------

     Ivo's track bent to the west-northwest after its upgrade, around 
  the periphery of the ridge to its north.  This prompted the government
  of Mexico to issue a tropical storm warning for the Pacific coast from
  Acapulco to Cabo Corrientes at 0900 UTC on the 11th.  The warnings 
  were discontinued east of Lazaro Cardenas at 1500 UTC and east of 
  Manzanillo at 2100 UTC the same day as Ivo paralleled the coast.  
  At 0300 UTC on the 12th, a tropical storm warning was issued for Baja 
  California from Cabo San Lucas to Cabo San Lazaro.  Mexico dropped 
  the watches for the coast from Manzanillo to Cabo Corrientes at 
  0900 UTC on the 12th, roughly coinciding with Ivo's turn to the 
  northwest.

     Ivo peaked at 1800 UTC on 12 September about 175 nm south of Cabo 
  San Lucas with a minimum CP of 998 mb and a 45-kt MSW.  This slight 
  strengthening occurred in spite of continued mild northeasterly shear.
  Ivo maintained this intensity for little more than a day before 
  weakening due to shear and cooler SSTs.  At 2100 UTC on the 13th, all 
  warnings for Baja California were dropped.  Ivo weakened quickly as 
  its convection collapsed--a ship just east of its center at 0600 UTC 
  on the 14th reported southeasterly winds of only 20 kts.  This report 
  justified its downgrade to a depression at 0900 UTC that day with the
  cyclone's track bending back toward the west as low-level steering 
  winds became dominant.  The final advisory on the convection-free 
  vortex of Tropical Depression Ivo was issued at 2100 UTC on 
  14 September and placed the dissipating center about 425 nm west-
  northwest of Cabo San Lucas.  QuiKScat data indicated that Ivo's winds
  were no higher than 25 kts at this time.  Ivo's remnants were 
  indiscernible by late on the 16th.

  
  C. Casualties and Damage
  ------------------------

     Though rainbands from Ivo impacted the Mexican coast, no casualties
  or significant damage are known at this time.



                      Hurricane Juliette  (TC-11E)
                        21 September - 3 October
                      ----------------------------

  A. Storm Origins
  ----------------

     The origin of Juliette is uncertain.  Ostensibly, Juliette was a 
  redevelopment of Caribbean Tropical Depression Nine.  There are a few
  people in the meteorological community, however, who feel that 
  Juliette only formed within the larger convective envelope of the 
  post-Nine tropical disturbance and is thus only indirectly related
  to TD Nine.  For simplicity and clarity, in this summary it will be 
  presumed that Juliette was at least loosely related to Tropical 
  Depression Nine.

     Tropical Depression Nine may have developed from a strong tropical
  wave that exited the African coast around 12 September, though the 
  evidence linking it to TD Nine is tenuous.  The wave that definitively 
  spawned TD Nine was first noted just east of the Lesser Antilles on 
  the 15th.  On the 19th, a LOW formed along the wave axis--later that 
  day, satellite and surface analyses supported its upgrade to Tropical 
  Depression Nine at 2100 UTC.  Even then its outflow was excellent and
  its cloud pattern robust.  Tropical storm watches were briefly issued
  for the entire Caribbean coast from Nicaragua to Belize.  They were 
  dropped when the depression made landfall on Nicaragua's sparsely-
  populated Mosquito Coast near Puerto Cabezas at 0300 UTC on the 20th.
  This was also the last advisory issued on Tropical Depression Nine.

     The central convection of the remnant disturbance weakened over 
  Central America on the 20th.  South of the center, however, convection
  flared up over the extreme eastern Pacific as early as 1800 UTC that 
  day.  By 0000 UTC on the 21st, this new convective center intensified 
  south of Guatemala and had a distinct cyclonic signature.   Convection
  was strong, but apparently the system's organization was insufficient 
  to issue warnings on.    It was with some surprise that a Hurricane 
  Hunter mission into the system later that day found flight-level winds
  of 52 kts at ~1520 m, corresponding to a surface wind of 40 kts.  
  Stepped-frequency radiometer data indicated a limited surface MSW of 
  50 kts.   On the basis of this data, the disturbance was upgraded
  directly to Tropical Storm Juliette in a special 1800 UTC advisory 
  on 21 September while located some 190 nm southeast of Puerto Angel, 
  Mexico.  Juliette tracked west-northwestward, south of a mid-level 
  ridge.  Though outflow was not impressive, very warm SSTs favored 
  further strengthening, though no one forecast the actual course of 
  events.


  B. Track and Intensity History
  ------------------------------

     Juliette's proximity to land prompted the hoisting of watches and 
  warnings for the southwestern Mexican coast, beginning with Juliette's
  first advisory.  These extended roughly from the Gulf of Tehuantepec 
  to Manzanillo and shifted westward on the 21st and 22nd.  The watches 
  and warnings were dropped at 1500 UTC on the 22nd as Juliette pulled 
  away from the coast.  The storm failed to thrive on the 22nd--deep 
  convection weakened and became asymmetric in spite of moderately 
  favorable upper-level conditions.   A surprising development was an 
  abrupt southward shift of the center, while the overall heading 
  shifted to due west.  This coincided with Juliette's weakening 
  convection.  NHC staff stated that it was probably due to downslope 
  flow from the Isthmus of Tehuantepec.

     A sustained intensification trend began late on the 22nd as 
  outflow improved and convection re-intensified over the center.  The 
  first hints of an eye were apparent in satellite imagery by 2215 UTC 
  that day.   Juliette's intensity took off on the 23rd--satellite 
  microwave imagery indicated that an eyewall was forming by about 
  0300 UTC, while at 1500 UTC satellite data warranted Juliette's 
  upgrade to hurricane status when located about 280 nm south-southeast
  of Manzanillo.   Juliette reached major hurricane intensity only six 
  hours later.  A hurricane watch was concurrently issued for the south-
  west coast of Mexico.

     Juliette continued to explode in intensity on the 24th.  At 0300 
  UTC the three-hour average Objective Dvorak T-number (ODT) was 7.0, 
  equivalent to a MSW of 140 kts.  The constraints of the Dvorak 
  technique, however, meant that the official intensity was "only" 
  115 kts--the wind field needed time to "catch up" to the satellite 
  estimates.  Juliette had a pinhole eye surrounded by a compact, 
  extremely cold CDO at this time.   Its eye diameter was only 5 nm, 
  though its circulation was large and imposing.  Juliette reached its 
  125-kt first peak in intensity at 0900 UTC on the 24th with an 
  estimated CP of 936 mb when located some 260 nm south-southeast of 
  Manzanillo.   Juliette's CP dropped 53 mb in 24 hours for an average 
  rate of 2.2 mb per hour; this qualifies as rapid deepening.  Between 
  2100 UTC on the 23rd and 0900 UTC on the 24th, the CP fell at an 
  average rate of 2.41 mb per hour, just short of the 12-hour mininum 
  criterion for explosive deepening (2.5 mb/hr for 12 hrs).

     Juliette's intensity and organization flagged after its peak, 
  apparently due to an eyewall replacement cycle.  A Hurricane Hunter
  mission found flight-level winds (FLW) of only 90 kts in its southeast 
  quadrant at 1800 UTC and a 94-kt FLW in its northeast quadrant later 
  that day.  Its estimated MSW dropped first to 100 kts, then to 90 kts 
  by 2100 UTC while its central pressure rose accordingly.   The 
  hurricane soon re-intensified, however, and regained major hurricane
  strength by 0300 UTC on the 25th.  By 0900 UTC the three-hour ODT 
  was once again 7.0, and this time it was supported by a T7.0 from 
  the TAFB.  Once again, however, the NHC intensity estimate remained 
  conservative until the Hurricane Hunters reached the system.  The 
  conservatism was also based on an earlier upward bias of satellite 
  estimates, as determined by reconnaissance on the 24th.

     The Hurricane Hunters finally reached Juliette late on the 25th 
  and found a Category 4 hurricane that was more intense than expected.
  They recorded a FLW (at 700 mb) of 131 kts, and measured a 
  surprisingly low CP of 923 mb in Juliette's 9 nm-wide eye.  Though 
  lower CPs have been estimated in NEP storms before (e.g., the 902 mb 
  CP of Hurricane Linda (1997)), this is the second-lowest CP actually 
  measured by reconnaissance, after 1973's Hurricane Ava (915 mb). 
  Although the conversion factor for FLW to surface MSW (0.9) suggested 
  surface winds of 120 kts, the 923 mb CP justified the NHC's increase 
  of the official MSW to 125 kts.  This was Juliette's "true" peak 
  intensity, attained 415 nm south-southeast of Cabo San Lucas at 
  2100 UTC on 25 September.  Dvorak intensity estimates from TAFB and 
  the AFWA were as high as 140 kts by 0300 UTC on the 26th while SAB 
  estimated 127 kts; the intensity remained 125 kts, however.

     Juliette was a large hurricane, particularly for the NEP basin.
  Hurricane-force wind radii were an impressive 60 nm, and 12 ft+ seas 
  extended as far as 325 nm from the center in its southern quadrant
  at 0300 UTC on the 26th.  Ship data from around 0000 UTC that day 
  revealed 34-kt wind radii as large as 225 nm in its eastern semi-
  circle, placing storm-force winds perilously close to the Mexican 
  coast (18 nm offshore of Manzanillo).  Tropical storm warnings were 
  accordingly re-issued for the southwest Mexican coast at 0300 UTC.

     After tracking slowly west-northwestward for over two days, Juliette
  made a long-forecast northwestward turn on the 26th, due to the ridge 
  to its north shifting eastward while a trough eroded its western limb.
  The first watches were issued for Baja California at 0600 UTC that day.
  Juliette weakened from its earlier peak, though it remained a major 
  hurricane throughout the 26th.  Satellite imagery indicated, and a
  reconnaissance mission found, concentric eyewalls on the 26th.  A
  flight late that day found an amazing three concentric eyewalls, though
  they found no FLW higher than 93 kts.    For a hurricane to have three
  eyewalls is not unprecedented--three were found in the Atlantic's
  Hurricane Edouard in 1996 (1)--but they are definitely one of the
  curios of tropical meteorology.  In any event, on the basis of their
  data, Juliette was downgraded below major hurricane intensity at 0300
  UTC on the 27th.  The estimated MSW of 95 kts was considered generous
  given the Hurricane Hunters' findings, but it was rather low for a
  central pressure that was still as low as 942 mb.   According to the
  NHC's official statements, the discrepancy was probably due to the
  broad central windfield of Juliette which didn't generate a strong
  pressure gradient.

     NOTE--A nice radar picture of the three concentric eyewalls can be
  found at:  http://www.hurricanehunters.com/juliette.htm>

     The Mexican government issued the first hurricane warning for Baja 
  California at 0900 UTC on the 27th as Juliette turned due north.  In 
  spite of persistent forecasts of a westward turn, Juliette continued
  on this heading.  The NHC considered its motion anomalous; indeed, one
  forecaster said that Juliette seemed to be moving straight through a
  ridge.  Juliette slowly weakened as it tracked straight towards Baja
  California.  By 0300 UTC on the 28th, Juliette was only 45 nm 
  west-southwest of Cabo San Lucas, putting the region well-within the
  radii of hurricane-force winds.

     Early on the 28th, the Hurricane Hunters found a minimal hurricane
  whose CP had risen approximately 1 mb per hour since their last fix.
  Ship 4XFD reported 67-kt winds in the same region where reconnaissance
  had reported FLW of 79 kts (the ratio of the surface winds to FLW, 
  0.85, agreed very well with the classic formulae for the conversion).
  Again, reconnaissance found a central pressure that seemed to be too 
  low (971 mb) for the winds they found, supporting the NHC's conjecture 
  that the windfield was too broad to generate a higher MSW.

     Juliette slowed on the 28th.  Its motion became erratic, and it had 
  stalled by 2100 UTC as it weakened to tropical storm-strength some
  65 nm west-northwest of Cabo San Lucas.  The Mexican weather service
  reported 52-kt sustained winds at Cabo San Lucas at this time as well
  as a 93-kt gust at 0100 UTC the previous evening.   An automatic 
  station at Cabo San Lucas reported a 76-kt wind at 0100 UTC on the 
  28th, though it is unclear if this were a sustained wind or a gust 
  (probably the latter)(2).  The erratic motion was probably due to 
  Juliette finally "feeling" a ridge over the American Southwest 
  while the trough off California weakened and a ridge built in.  This 
  left the storm in a col, a weak steering environment where it 
  lingered near the coast.  It was the shape of things to come, as it 
  turned out.

     Juliette surprised forecasters by re-intensifying to hurricane 
  strength at 0300 UTC on 29 September, 65 nm northwest of Cabo San 
  Lucas, while remaining stationary.  Satellite estimates suggested that
  the storm intensified even as it was mostly over land, though that
  fact was key to the NHC's conservatism in the intensity estimate.
  Both Cabo San Lucas and La Paz reported only light rain and below-
  storm-force winds at this time.   Apparently most of Juliette's rain 
  and winds were over remote, mountainous regions.  Juliette weakened 
  again to a tropical storm at 1500 UTC on the 29th, apparently just 
  over the west coast of Baja California near Todos Santos.  Weakening 
  began apace on the 29th, due to dry air entrainment, land interaction,
  and wind shear.  Juliette's center passed very close or over the 
  island of Santa Margarita, just off the west coast of Baja California, 
  at 0000 UTC on the 30th as it finally moved out of its stall and 
  assumed a meandering, roughly northwestward track offshore.

     Juliette weakened to a depression as it turned north again on the 
  30th, making landfall at Cabo San Lazaro at approximately 0800 UTC.  
  It tracked northward, straddling the coast, and moved inland over the 
  peninsula.  Steering currents remained light as the system was 
  apparently embedded in a mid-level ridge.    Though its central 
  convection was weak, convection persisted over the Gulf of California 
  and the adjacent land areas.  A 40-kt gust recorded at Santa Rosalia 
  was discounted by the NHC as spurious, since it was discordant with 
  earlier data from that station and inconsistent with satellite data.
  At 2100 UTC on the 30th, Juliette's center re-formed over the Gulf of 
  California, roughly 15 nm southeast of Tiburon Island.   Even so, 
  satellite and surface data discrepancies made pinpointing its 
  location and intensity difficult.  Juliette was embedded in a larger 
  cyclonic gyre that had a significant influence on its motion.

     The final advisory on Juliette was issued at 0300 UTC on 1 October
  after its low-level circulation decoupled from the rest of the cyclone.
  Its poorly-defined center was placed inland over the Sonoran coast, 
  some 20-25 nm north of Tiburon Island.   In yet another surprise for 
  forecasters, the remnant LOW's convection and circulation regenerated
  over the Gulf of California on the 1st, warranting its re-upgrade 
  to a depression at 0300 UTC on 2 October some 55 nm south-southeast 
  of Puerto Penasco, Mexico.    Althought Juliette was almost 
  unclassifiable by the Dvorak technique, WSR-88D data from Yuma, 
  Arizona, revealed good banding and winds as high as 35-50 kts at 
  0900 UTC.  The latter winds, however, were well above sea level, and 
  it was deemed unlikely that the high winds would have worked down to
  the surface due to the system's weak convection.

     Juliette meandered toward the northwest over the far northern Gulf 
  of California on the 2nd before weakening once again.   It turned 
  southwestward late on the second, as the deep-layer LOW it was embedded
  in was deflected southwards by a ridge over the Pacific.  The second
  final advisory on Tropical Depression Juliette was issued at 0300 UTC 
  on 3 October as it made landfall about 80 nm south-southeast of San 
  Felipe, Mexico.


  C. Meteorological Observations
  ------------------------------

     Juliette lingered off southern Baja California for days, resulting 
  in some impressive rainfall totals for a usually arid region.  One 
  government official was quoted as saying that Cabo San Lucas received 
  457 mm of rain in four days, equivalent to the total of four average 
  rainy seasons.  The highest 24-hour totals were:  136 mm at Santiago 
  (Baja California Sur) on the 27th, 167.5 mm at Enpalme (Sonora) on 
  the 30th, and 207.2 mm at San Felipe (Baja California) on 2 October 
  (3).


  D. Damage and Casualties
  ------------------------

     The estimated damage in Baja California was minor, roughly 
  $12 million according to press reports.  The Associated Press stated 
  that in Cabo San Lucas, a resort city of 25,000, Juliette smashed 
  docks, uprooted trees and power lines, and broke the city's water 
  mains.  Many tourists were stranded in Cabo San Lucas for the storm's
  duration as floodwaters washed out the main bridge to the mainland 
  and blocked smaller roads.    Damage to the homes of the poor was 
  significant, while most luxury hotels received only minor damage. 
  One hotel, however, suffered 90% damage according to AP sources.

     In southwestern Mexico, heavy rains from Juliette flooded 200 homes 
  in the state of Michoacan and downed many trees.    There was minor
  flooding and tree damage in Acapulco, while rain-swollen rivers knocked
  out two bridges in the state of Oaxaca.  

     Three deaths resulting from Juliette are known at present.  One man
  drowned when his boat capsized near the Mexico mainland while a fisher-
  man from Acapulco was lost in high seas.  A Colorado surfer drowned in
  rough waves and was the only American casualty.


  E. References
  -------------

  (1.)  Weatherwise.  February/March 1997, p. 40.
  (2.)  "Ciclon Tropical Juliette."  Servicio Meteorol�gico Nacional
         http://smn.cna.gob.mx/ciclones/tempo2001/pacifico/juliette/
         juliette.html>
  (3.)  Ibid.



                        Hurricane Kiko  (TC-12E)
                             22-25 September
                        ------------------------

  A. Storm Origins
  ----------------

     Kiko developed from a tropical wave that was first noted in the
  Eastern Pacific on 19 September.  By late on the 21st, a definite
  cyclonic disturbance had formed along the wave axis.  The organization
  of this disturbance improved, warranting its upgrade to Tropical 
  Depression Twelve-E at 0300 UTC on 22 September about 800 nm west-
  southwest of Manzanillo, Mexico.

     Although conditions seemed to be favorable for significant 
  strengthening at first, easterly shear slowed intensification while an
  approaching trough restricted its outflow.  This combination of factors
  made positioning the center difficult.   Early on, an initial west-
  northwestward track was found to be northwesterly when the first good
  visible imagery became available on the 22nd.  Difficulties aside, the
  depression became organized enough to be upgraded to Tropical Storm
  Kiko at 2100 UTC on 22 September when located about 950 nm west of
  Manzanillo.  About this time Kiko's track became due westerly as it
  moved south of the climatological subtropical ridge.  The NHC stated
  that the westward turn may have been due to the realignment of Kiko's
  LLCC with its deep convection as outflow improved.


  B. Track and Intensity History
  ------------------------------

     Kiko intensified slowly through the 23rd, perhaps due to the
  entrainment of stable air into its western semicircle late that day.
  Even so, microwave satellite imagery revealed eye formation late on
  the 23rd. Therefore, at 0900 UTC on 24 September, Kiko was upgraded
  to hurricane status roughly 1225 nm west of Manzanillo with 65-kt MSW
  and a CP of 987 mb.  This was Kiko's peak intensity.  At this time,
  microwave imagery revealed a 20-nm eye, supported by infrared imagery
  and Dvorak intensity analyses.  Kiko's upgrade coincided with a brief
  west-northwestward jog before a westward track resumed late on the
  24th.  Interestingly, the NHC also speculates that the decrease in
  Kiko's forward motion commencing on the 23rd may have also been due
  to the re-organization and regeneration of its central convection.

     Kiko weakened to tropical storm-strength soon after its peak due to
  cooler waters and increased shear.  Kiko's deterioration was relatively
  rapid and it was downgraded to a depression at 1500 UTC on the 25th.
  The final advisory on Tropical Depression Kiko was issued at 2100 UTC
  on 25 September when it was located some 1425 nm west of Manzanillo.

     Kiko was arguably more remarkable in death than it was in life.
  The storm's remnant was recognizable long after the final advisory was
  issued and generated intermittent deep convection until the 28th as it
  drifted westward across the Pacific.  The low-level vortmax persisted
  longer still, and its convection even flared up once again as an extra-
  tropical cyclone approached.  A cold front associated with the LOW
  finally absorbed Kiko's remnant by 0600 UTC on 2 October, well north-
  east of Hawaii.


  C. Damage and Casualties
  ------------------------

     No casualties or damage are known to have resulted from Kiko.

  ***********************************************************************

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

  Activity for September:  2 tropical depressions **
                           5 typhoons 

  ** - Both of these systems were treated as tropical depressions by
       JMA and some of the other Asian TCWCs, but not by JTWC

  NOTE:  Most of the information on each cyclone's history presented in
  the narrative will be based upon JTWC's advisories, and references to
  winds should be understood as a 1-min avg MSW unless otherwise noted.
  However, in the accompanying tracking document I have made comparisons
  of coordinates with some of the Asian warning centers when their
  positions differed from JTWC's by usually 40-50 nm or more.  A special
  thanks to Michael V. Padua, owner of the Typhoon 2000 website, for
  sending me the PAGASA and JMA tracks.   Also, a special thanks to
  Huang Chunliang of Fuzhou City, China, for sending me tracks based on
  warnings from the National Meteorological Center of China (NMCC), the
  Hong Kong Observatory (HKO), and the Central Weather Bureau of
  Taiwan (CWBT).

     In addition, Roger Edson of the University of Guam sent me tracks
  for two tropical depressions which were warned on by some of the Asian
  TCWCs but not by JTWC.  A special thanks to Roger for the information
  he sent.

     In the title line for each storm I plan to reference 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 responsibility.


            Northwest Pacific Tropical Activity for September
            -------------------------------------------------

     The Northwest Pacific basin was very active during the months of
  September.  Five tropical cyclones were named by RSMC Tokyo and all
  reached typhoon intensity.  Typhoon Danas clipped the southeastern
  corner of Japan and caused a few fatalities and some flooding.  Two
  of the storms, Nari and Lekima, brought destructive floods to Taiwan
  with Nari being especially deadly.  Typhoons Vipa and Francisco both
  recurved northeastward into the North Pacific with no effects on land.

     Two additional systems were tracked as tropical depressions by JMA
  and some of the other warning centers.  Roger Edson also provided me
  with tracks for these systems.  Since these depressions were operating
  simultaneously and, based on Roger's assessment, there is a possibility
  they could have reached minimal tropical storm intensity, in the tracks
  file I identified them with the Greek letters "Rho" and "Sigma".  The
  first system, "Rho", was active in some form or another from the 5th
  until the 12th of September.  It originated as a weak LOW in the Gulf
  of Tonkin east of Hanoi, drifted very slowly eastward across the
  Luichow Peninsula, reaching a point about 225 nm south-southwest of
  Hong Kong by 1200 UTC on 8 September, then drifted southward and
  eventually back to the west.  The system became slightly better
  organized on the 9th, and on 10 September made landfall around 0200
  UTC near Lingshui Li Autonomous County of Hainan Province.  The
  depression continued westward across Hainan Dao, re-entering the
  Gulf of Tonkin once more, and eventually moved inland into southern
  China where it dissipated on the 12th.

     Qiongshan City recorded 160 mm of rain from the night of the 7th
  through the afternoon of the 8th, and a three-day storm total in
  excess of 210 mm.  Wenchang City recorded 115 mm in an eight-hour
  period on the 8th and 217 mm in a 54-hour period ending at 0600 UTC
  on 10 September.  Haikou City measured 127 mm of rain during the
  same 54-hour period and 155 mm for a three-day total.  One person
  was injured and 15 houses destroyed in a tornado which lasted a few
  minutes in Xiuying District, Haikou City around 0840 UTC on the 10th.
  (The above information was sent to me by Huang Chunliang.)

     The second depression, "Sigma", formed south of Taiwan on the 7th
  and moved generally northeastward over the next several days.  JMA
  elevated the 10-min avg winds to 30 kts for a 24-hour period beginning
  at 09/0000 UTC when the system was located about 100 nm east-southeast
  of the southern tip of Taiwan.  It was at this juncture that Roger
  Edson assigned a 1-min avg MSW of 35 kts.  The most notable aspect
  of this depression was its interaction with Typhoon Nari which was
  likely one of the factors contributing to Nari's highly erratic track.
  The depression continued moving northeastward, passing south of
  Okinawa on the 10th, and had weakened into a weak LOW south of Japan
  by the 12th.  Both "Rho" and "Sigma" were mentioned by JTWC in their
  daily STWOs, but no warnings were issued on either system.



                     Typhoon Danas  (TC-19W / TY 0115)
                              2 - 14 September
                     ---------------------------------

  Danas: contributed by the Philippines, is a Tagalog word meaning
         "to feel" or "to experience"


  A. Storm Origins
  ----------------

     JTWC issued a STWO at 0200 UTC on 2 September which noted that an
  area of convection had formed about 450 nm west of Wake Island.  A
  LLCC with persistent convection was developing in a col area south of
  a TUTT cell.  Diffluence southeast of the trough was creating a fairly
  favorable environment for development.  Although the first JTWC warning
  was not issued until 03/0000 UTC, the JMV file (a sort of working Best
  Track) indicates that the system had become a 25-kt depression at
  0600 UTC when located about 500 nm west of Wake Island.  A TCFA was
  issued at 2130 UTC as an increase in the areal coverage of deep
  convection had been noted during the previous few hours.

     JTWC issued the first warning on Tropical Depression 19W at 0000 UTC
  on 3 September.  The system was centered about 600 nm west of Wake
  Island or about 570 nm east of the Northern Marianas.  Maximum winds
  were estimated at 25 kts and the depression was moving westward at
  5 kts.  Convective organization was increasing, and at 1200 UTC the
  MSW was bumped up to 30 kts with deep convection building over the
  LLCC.     At 0000 UTC on the 4th both JTWC and JMA upgraded the 
  strengthening tropical cyclone to tropical storm status with JMA
  assigning the name Danas.  The newly-christened Danas was then located
  about 450 nm east of the Northern Marianas, moving westward at 10 kts.


  B. Track and Intensity History
  ------------------------------

     Initially steered westward by a subtropical ridge to the north,
  Danas gradually slowed and turned northward as a baroclinic trough
  east of Honshu moved eastward.    The cyclone intensified rather
  quickly after reaching tropical storm strength.    The intensity
  estimates at 0600 and 1200 UTC on 4 September are good illustrations
  of the situation often faced by forecasters when they must choose
  between widely varying satellite intensity estimates.   At 04/0600 
  UTC both JTWC and JMA reported 45 kts (1-min and 10-min averages,
  respectively).  Satellite CI estimates at the time were 30, 45, and
  55 kts.  Six hours later JTWC's 1-min avg MSW was 60 kts and JMA's
  10-min mean MSW was 55 kts--satellite estimates ranged from 45 to
  65 kts.  JTWC upgraded Danas to typhoon status at 1800 UTC with the
  center located approximately 350 nm east of the Northern Marianas.
  The intensity was upped to 80 kts at 05/0000 UTC while at the same
  time JMA classified the storm as a 70-kt typhoon.   Danas by this
  time was moving northward at 6 kts.

     JTWC increased the MSW to 85 kts at 05/1800 UTC.  A 05/2006 UTC
  SSM/I pass depicted a small eye feature with good convective banding
  in the eastern semicircle.  Upper-level outflow was good, especially
  toward the southeast.  By 0600 UTC on the 6th satellite imagery
  revealed a 15-nm round cloud-filled eye in the center of the CDO.
  The trough which had brought about the northward turn began to fill
  and the ridge to the north of the typhoon began to strengthen--by
  1200 UTC Danas was beginning to move in a north-northwesterly
  direction which gradually became northwesterly.  The storm reached
  its peak intensity of 115 kts (per JTWC's warnings) at 0000 UTC on
  7 September when the eye was centered approximately 400 nm east-
  northeast of Iwo Jima.  The MSW was based on CI estimates of 115 and
  127 kts.  Danas maintained its peak intensity for 18 hours.  A SSM/I
  pass at 07/1023 UTC revealed a tight, symmetrical system with an
  8-nm diameter cloud-filled eye and a primary banding feature over
  the southeast quadrant.

     By 08/0000 UTC Typhoon Danas was moving due westward as it was
  being steered by an intense mid-level HIGH situated east of Japan.
  The intensity had come down a bit to 105 kts at 1800 UTC on the 7th,
  although one satellite intensity estimate was still at 127 kts.  JTWC
  lowered the MSW to 100 kts at 0600 UTC, but raised it back to 105 kts
  at 1200 UTC based on Dvorak intensity estimates ranging from 90 to
  115 kts.   At 08/1800 UTC Danas had turned to the northwest as it
  passed about 235 nm due north of Iwo Jima.  The storm was also located
  about 430 nm south-southeast of Tokyo and was beginning to pose a
  distinct threat to Japan.   Typhoon Danas continued moving northwest-
  ward toward southwestern Honshu until it reached its point of
  recurvature around 0000 UTC on 10 September about 235 nm southwest of
  Tokyo.    An approaching mid-latitude trough had weakened the ridge
  which had been steering the storm, allowing it to recurve rather
  abruptly to the north-northeast.   Fortunately, Danas began to weaken
  steadily as it approached Honshu.  The MSW was down to 85 kts by 0000
  UTC and continued to drop as the typhoon approached the Tokyo area.

     At 1200 UTC the 35-nm diameter banding eye of Danas was located
  about 150 nm southwest of Tokyo and moving north-northeastward at
  7 kts.    The MSW was 75 kts at that time and had dropped to an
  estimated 70 kts by the time the eye reached the coast just southwest
  of Yokosuka at 11/0035 UTC.  (Yokosuka is a short distance south of
  Tokyo on a point of land extending out into Tokyo Bay.)    The
  NAVPACMETOCCEN Yokosuka reported sustained winds of 65 kts from the
  southeast gusting to 75 kts as the storm made landfall.  (JMA's 10-min
  avg MSW at 0000 UTC was 55 kts.)    Danas moved across the peninsula
  southeast of Tokyo and by 0600 UTC was moving out into the Pacific
  about 60 nm northeast of Tokyo.  JTWC downgraded Danas to a 60-kt
  tropical storm at this time.  (Interestingly, one satellite intensity
  estimate was still as high as 77 kts.)  After crossing extreme south-
  eastern Honshu the former typhoon accelerated northeastward and
  weakened as it began to transition into an extratropical cyclone.
  The final JTWC warning on Danas was issued at 12/0600 UTC.  The storm
  had completed extratropical transition and was located well east of
  Hokkaido and south of the Kuril Islands, racing northeastward at
  22 kts.  The extratropical gale continued to moved northeastward, then
  curved eastward as it approached the 49th parallel.   By 0600 UTC on
  14 September the system was still a 45-kt gale south of the south-
  western Aleutian Islands.

     Typhoon Danas was at its peak intensity of 115 kts from 07/0000
  through 07/1800 UTC.  The minimum central pressure estimated by JMA
  was 950 mb.   Danas was an average-sized typhoon.    At its peak
  intensity storm-force winds extended outward from the center 80 nm
  in the eastern semicircle and 50 nm to the west while gales covered
  an area about 300 nm in diameter.    By the time the typhoon was 
  approaching Japan the gale-force radii had expanded--gales extended 
  outward 190 nm to the northeast and southeast and 160 nm to the north-
  west of the center.


  C. Meteorological Observations
  ------------------------------

     The only report available was the observation from Yokosuka which
  was referenced in the above section.


  D. Comparison Between JTWC and Other Centers
  --------------------------------------------

     Center position coordinates between the various warning centers
  were in very good agreement for Typhoon Danas.  Intensity estimates
  between JTWC and JMA were in good agreement except for the storm's
  most intense phase.   JMA's peak 10-min avg MSW was 80 kts, and was
  kept at this value for a full four days, from 06/0000 through 10/0000
  UTC, except for a six-hour period on the 8th when it was lowered to
  75 kts.   JTWC's 1-min avg MSW estimates were at 100 kts or greater
  for most of this period, peaking at 115 kts on the 7th.   NMCC's
  peak estimated intensity (10-min avg) was 90 kts, from 0600 UTC on
  8 September through 0000 UTC on the 10th.  During the time when JTWC's
  MSW was at 115 kts, NMCC's 10-min mean wind was 80 kts.


  E. Damage and Casualties
  ------------------------

     According to press reports, five persons died in Japan as a result
  of Typhoon Danas.  Four of the deaths were due to mudslides--the other
  fatality was a 50-year-old man who fell from the roof of his house
  while carrying out repairs.     In addition, fifteen persons were
  injured, mainly by flying debris.

     The town of Nikko, north of Tokyo, recorded in excess of 870 mm of
  rain over a four day period.  Flooding was reported by 148 homeowners
  in several prefectures, and over 220 households were advised to
  evacuate due to the threat of floods and mudslides.  Several trees
  were blown down across streets in the downtown Akasaka area, crushing
  one car, but no injuries were reported.

     Ferry services between Honshu and outlying islands were disrupted,
  more than 140 domestic and international flights cancelled, and
  services on a number of train lines in and around Tokyo cut back.
  Danas' high winds and heavy rains also shut down some of the high-speed
  bullet trains, leaving thousands of passengers stranded.



                  Typhoon Nari  (TC-20W / TY 0116 / Kiko)
                              5 - 21 September
                  ---------------------------------------

  Nari: contributed by South Korea, is the lily, a wildflower commonly
        observed in the summer in Korea


  A. Storm Origins
  ----------------

     The disturbance which became the long-lived and erratic Nari had its
  origin in the Philippine Sea in late August.   An area of convection 
  had formed by 0600 UTC on 31 August roughly midway between northern
  Luzon and the Northern Marianas.  A broad cyclonic circulation with
  westerly winds equatorward of the center was located in the region, and
  there was some isolated convection noted along the southern periphery
  of the circulation.  A 200-mb analysis and CIMSS products indicated
  moderate vertical shear over the region.  The disturbance remained
  quasi-stationary and by 2 September had weakened and was no longer
  considered a suspect area for tropical cyclogenesis.  On the 3rd the
  area of convection redeveloped farther west in association with a
  monsoon trough in the region.  A QuikScat pass on the 4th depicted a
  weak LLCC located about 150 nm north of the previous day's position.

     Things got interesting on the 5th!  Based on JTWC's STWO, the 
  disturbance moved northwestward to a position east of southern Taiwan.
  JMA, however, classified the system as a weak tropical depression at
  1200 UTC, located about 265 nm west-southwest of Okinawa.  At 1800 UTC
  winds were increased to 30 kts (10-min avg).  JTWC had assessed the
  potential for development to be fair at 1200 UTC, and at 1930 UTC a
  TCFA was issued, noting an improvement in the coverage and organization
  of deep convection.  At 0000 UTC on 6 September JMA upgraded the system
  to Tropical Storm Nari, located about 200 nm west-southwest of Okinawa.
  At the same time JTWC initiated warnings on Tropical Depression 20W
  with 25-kt winds.  JTWC's position was about 55 nm west-northwest of
  JMA's, or about 100 nm east of Taipei, Taiwan.   The JTWC warning noted
  that the center was fully-exposed to the east of a line of deep
  convection.

     JMA's intensity remained at 35 kts throughout the 6th while JTWC
  kept the system at depression status, based on CI estimates of 25 and
  30 kts.  Enhanced infrared imagery depicted a very weak, diffuse system
  with a partially-exposed center.   Several bursts of deep convection
  were noted around 1800 UTC, so JTWC increased the MSW to 30 kts.  Radar
  imagery revealed a well-defined mid and lower-level circulation center
  near the bursts of deep convection.  An amended 1800 UTC warning was
  soon issued upgrading Nari to tropical storm status based on synoptic 
  reports from Okinawa which showed rapid pressure falls and winds near
  gale force.  (Satellite CI estimates at the time were still 25 and
  30 kts.)  Nari was centered at this time only about 50 nm west of
  Okinawa, moving east-northeastward at 7 kts.


  B. Track and Intensity History
  ------------------------------

     The track of Tropical Storm/Typhoon Nari over the next 5 or 6 days
  after it was named was extremely erratic and convoluted with twists,
  turns, and loops within loops.  It remains to be seen if the final
  "best track" for Nari will smooth out any of the kinks, but during this
  period the storm was within easy radar coverage range from Okinawa
  and the coordinates from JTWC, JMA, and NMCC were in excellent agree-
  ment.  Later, as the storm approached Taiwan, the Taipei radar helped
  to pinpoint the eye, so again there was little discrepancy between
  the warning centers as to the exact location of Nari's center.   In
  order to try to present the Nari saga as it unfolded in a clear and
  orderly manner, I am going to adopt an approach which I've considered
  on several occasions in the past but have never implemented--a daily
  diary format.  (Dates represent day boundaries in UTC.)

  September 7th:  At 0000 UTC Nari's center was located approximately
  13 nm south of Kadena AB, moving slowly eastward at 7 kts.  The MSW
  was estimated at 45 kts, based mainly on synoptic reports--satellite
  intensity estimates were lower.  The storm continued drifting eastward
  and by 0600 UTC had reached a point about 25 nm east of Kadena where
  it remained essentially stationary for several hours before reversing
  its motion and drifting back toward the west.  Nari also intensified
  as bursts of deep convection developed near the center, leading to the
  MSW being upped to 50 kts at 1200 UTC.  By 1800 UTC the storm's center
  had drifted back westward and was located essentially over Kadena AB.
  Based on synoptic reports JTWC upgraded Nari to a 70-kt typhoon at
  this time while JMA and NMCC had increased their 10-min avg MSW
  estimates to 55 kts and 50 kts, respectively.  Nari was a small cyclone
  with 50-kt winds confined to a zone within 20 nm of the center.

  September 8th:  For most of 8 September Nari drifted initially north-
  westward, later turning to a westerly track.  By 1800 UTC the typhoon's
  center had reached a point about 140 nm west-northwest of Kadena AB.
  JTWC brought the MSW down to 65 kts at 1200 UTC where it remained for
  24 hours while JMA's and NMCC's intensity estimates remained at 60 kts.
  Nari had entered Taiwan's AOR by 1800 UTC and CWBT assessed the 10-min
  avg MSW at 60 kts also.   The storm remained small with gales covering
  an area no more than 100 nm in diameter.

  September 9th:  Nari's westward motion came to a halt early on the 9th
  and the storm sat essentially stationary at a position about 150 nm
  west-northwest of Kadena AB.  The storm's organization improved during
  the day with satellite intensity estimates reaching T4.5 (77 kts).  An
  08/2151 UTC TRMM pass had depicted tightly-wrapped but narrow
  convective banding.  JTWC increased the MSW to 75 kts at 0600 UTC and
  NMCC upgraded Nari to typhoon status at the same time.   At 1200 UTC
  NMCC increased their 10-min avg MSW estimate to 70 kts while HKO and
  CWBT both upgraded Nari to a typhoon.  JMA had dropped their reported
  intensity early in the day, but bumped it back up to 60 kts at 1200
  UTC.  Nari's areal dimensions had increased somewhat by 1800 UTC with
  gales covering an area over 150 nm in diameter.

  September 10th:  Nari commenced a southward drift early on the 10th
  which later became east-southeastward.  By 1800 UTC the storm had
  moved to a point roughly 85 nm west-southwest of Kadena where it then
  remained quasi-stationary for about 48 hours.   Nari weakened some,
  likely due to the upwelling of cooler water in the area where it had
  been loitering.  JTWC reduced Nari's intensty to 65 kts at 0000 UTC
  while NMCC made a more drastic reduction from 70 to 50 kts.  JMA's
  10-min avg MSW estimate remained at 60 kts through the 10th.  As the
  storm pulled away from the area where it had sat stationary on the 9th,
  it began to show signs of intensifying once more.  NMCC upped their
  intensity to 60 kts at 1200 UTC, and JTWC increased their 1-min avg
  MSW to 75 kts once again at 1800 UTC.

  September 11th and 12th:  On these two days Typhoon Nari was quasi-
  stationary approximately 40-50 nm west of Kadena.  The storm wiggled
  about some, seeming to describe a very tight clockwise loop within
  a larger counterclockwise loop.   The re-intensification trend noted
  on the 10th continued--at 11/0000 UTC satellite imagery indicated that
  a 5-nm irregular eye had developed with an increase in deep convection
  over the eastern quadrant.  The eye diameter had increased to 10-nm
  by 0600 UTC and CI estimates had reached 102 kts.  At 1200 UTC
  satellite intensity estimates were 102 and 115 kts, so JTWC upped the
  MSW to 100 kts, which was the peak intensity for Typhoon Nari.  The
  storm's eye at this time was only about 35 nm west of Kadena and 50-kt
  winds reached outward 35 nm from the center.  A gradual weakening
  trend began afterward, once more likely due to upwelling of cooler
  waters, and winds had come down to 75 kts by 1800 UTC on the 12th (per
  JTWC).  A SSM/I pass at 12/1236 UTC indicated that the convection
  associated with the vortex had expanded and weakened.  JMA's peak
  intensity for Nari was 85 kts while NMCC's was slightly lower at
  80 kts.  Both these warning centers had reduced Nari to a 60-kt
  tropical storm by 1800 UTC on the 12th.   The area covered by gales
  had expanded to cover a zone almost 200 nm in diameter.

  September 13th:  By 13/0000 UTC Nari was located about 70 nm west of
  Kadena and had begun to track slowly toward the west-northwest.  At
  1200 UTC the typhoon was about 120 nm west-northwest of Kadena and had 
  crossed its track of five days earlier, completing a counterclockwise
  loop which contained a tiny clockwise loop.  The storm had briefly
  moved northwest to this point, then moved little for the remainder of
  the 13th.  Nari's intensity remained basically static--JTWC decreased
  the MSW to 65 kts at 0600 UTC but bumped it back up to 70 kts at 1800
  UTC.  JMA's and NMCC's 10-min avg MSW remained pegged at 60 kts, and
  after Nari re-entered CWBT's AOR at 0600 UTC, that agency estimated
  the intensity at 60 kts.

  September 14th:  On 14 September Nari weakened slightly, being down-
  graded to a 60-kt tropical storm by JTWC at 1800 UTC.   The other
  warning centers' 10-min avg MSW estimates likewise remained under
  typhoon intensity, including that from HKO (Nari had re-entered that
  center's AOR by 14/0600 UTC).   The slow northwestward motion seen the
  previous day came to a halt about 235 nm northeast of Taipei, and by
  1800 UTC a slow southwestward drift had begun.

  September 15th:  Nari turned more to the south-southwest as the 15th
  began, and for much of the day moved in that direction at a steady
  6 kts, but slowed and turned slightly back to the west-southwest late
  in the day.  By 1800 UTC the storm's 22-nm diameter eye was centered
  approximately 115 nm east-northeast of Taipei.  The weakening observed
  on the previous day proved to be temporary--late on the 14th animated
  satellite imagery revealed a redeveloping banding eye feature and CI
  estimates were 65 kts, so JTWC upgraded Nari back to typhoon status
  after only six hours as a tropical storm.  As the day progressed the
  eye became more distinct and the areal coverage of deep convection
  increased.  JTWC had increased the MSW back to 80 kts by 1800 UTC,
  and all the other agencies had re-upgraded Nari to typhoon status by 
  1200 UTC.

  September 16th:  By 16/0000 UTC Nari was located about 100 nm east-
  northeast of Taipei and drifting westward at 3 kts.  This motion later
  became southwesterly at a slightly faster pace as the day wore on.
  JTWC increased the MSW to 90 kts at 0000 UTC based on CI estimates of
  77 and 102 kts.  Gales reached out about 100 nm from the 25-nm round
  eye.  By 1200 UTC Nari's eye was nearing the coast of northeastern
  Taiwan and made landfall around 1500 UTC just east of the city of Lo
  Tung.    JTWC estimated the maximum winds at landfall to be 85 kts
  sustained, gusting to 105 kts.  By 1800 UTC the center was inland and
  weakening as it tracked to the southwest at 8 kts.    JMA did not
  increase Nari's intensity (10-min avg) above 65 kts during this
  secondary peak, but both NMCC and CWBT raised their estimates to
  80 kts, and HKO was reporting 75 kts just before landfall.

  September 17th and 18th:  Nari had weakened to a tropical storm by
  17/0000 UTC (according to all warning centers) as it drifted slowly
  south-southwestward over Taiwan.  Whereas the center coordinates from
  the various warning agencies agreed exceptionally well during the time
  Nari was fairly well-organized and within radar range of Kadena and/or
  Taipei, during the 17th and 18th the position estimates became rather
  divergent--not unexpected with a weakening system over land.  But in
  general all the TCWCs except JMA tracked Nari's center down the
  western side of the island.  JMA, however, tracked the storm's center
  southward just off the eastern coast of Taiwan.   JTWC's positions at
  0600 and 1200 UTC on 18 September abruptly jumped eastward to near the
  eastern side of the island, but the 1800 UTC position was back to just
  west of southwestern Taiwan.   All the warning agencies maintained
  Nari at tropical storm intensity throughout the 17th and 18th except
  for JTWC and NMCC, who downgraded the storm to depression status at
  18/1800 UTC.   While the exact location of the center may not be known
  with certainty during these two days, the storm's slow drift down the 
  backbone of Taiwan led to very heavy rainfall over the island which 
  had disastrous consequences.  More on this below.    (By 16/1800 UTC 
  Nari had entered PAGASA's AOR and that agency began issuing warnings 
  on the cyclone, naming it Kiko--a Filipino nickname.)

  September 19th:  After all the surprises which this long-lived and
  erratic cyclone had brought, Nari still had one more trick up its
  sleeve.  After clearing Taiwan Tropical Depression Nari turned westward
  across the Taiwan Strait.   Satellite CI estimates were initially at
  30 kts, but the JTWC warning issued at 1200 UTC noted the receipt of
  a 35-kt CI estimate.   By 1800 UTC intensity estimates had reached
  45 kts, so JTWC upgraded Nari to a 45-kt tropical storm located about
  225 nm east of Hong Kong.  A 19/1628 UTC SSM/I pass depicted the
  possible development of a banding eye.   The cyclone had by this time
  exited the AORs of Taipei and Manila, so those TCWCs were no longer
  players.  NMCC also upgraded Nari back to tropical storm status at
  19/1800 UTC with the 10-min avg MSW estimated at 40 kts.  HKO, which
  had never downgraded Nari, raised its intensity estimate back to
  40 kts.

  September 20th:  Early on the 20th animated visible and infrared
  imagery revealed a developing banding eye and improving organization
  of the deep convection.   JTWC increased the MSW to 55 kts at 20/0000
  UTC with Nari centered about 155 nm east of Hong Kong.  HKO upped their
  10-min avg MSW estimate to 50 kts, and NMCC raised theirs to 60 kts,
  which, following the standard 10-min to 1-min conversion rule, would
  imply a 1-min avg MSW of about 70 kts.   Nari's westward motion picked
  up some as it neared the Chinese coast.   At 0600 UTC the center of
  the storm was just making landfall about 80 nm east-northeast of Hong
  Kong with 55-kt winds based on JTWC's warning.  HKO's and NMCC's 10-min
  avg MSW estimates had fallen to 40 kts, but both those agencies were
  placing the center to the north and west of JTWC's position and hence
  a little farther inland.  As Nari continued westward into China the
  convection began to weaken rapidly, and by 0000 UTC on 21 September
  the system was a weak 20-kt LOW located about 150 nm west-northwest
  of Hong Kong.  Interestingly, during this final encore to Nari's long
  performance, JMA never increased the intensity above 30 kts, somewhat
  surprising considering the consensus among the other warning centers
  that the cyclone re-intensified markedly prior to its final landfall
  in China.


  C. Steering Influences on Typhoon Nari
  --------------------------------------

     With Nari following such an extremely erratic track--which looked
  rather like a strand of cooked spaghetti which had been dropped onto
  a platter--I thought it would be interesting to take a look at some 
  of the circulation features which helped to steer Nari on its long
  and winding road.  The initial JTWC warning on TD-20W noted that the
  system had formed within a weakness in the subtropical ridge with weak
  steering currents.  Slight interaction with a weak ridge building over
  China or else an indirect interaction with Typhoon Danas to the east
  were considered possibilities.  Nari's initial eastward motion which
  carried it across Okinawa the first time was due to the influence of
  a mid-level ridge to its southeast, between the storm and Typhoon
  Danas.  As the larger and more intense Danas approached, this ridge
  began to erode and a mid-level ridge began to build near the Chinese
  coast, leading to Nari's turning back to the west-northwest and
  crossing Okinawa a second time.

     However, the Chinese ridge did not build far enough east at that
  time to induce Nari to follow a more protracted westerly or south-
  westerly course.  The typhoon remained within a weakness in the sub-
  tropical ridge, resulting in slow, erratic motion.  The move back to
  the southeast which brought the cyclone again near Okinawa as it
  reached its peak intensity was in part apparently due to interaction
  with an exposed circulation southeast of the typhoon on 9 and 10
  September.  This system was carried as a tropical depression by JMA,
  and the author also received a track from Roger Edson.  (This was the
  depression which I dubbed "Sigma" in the tracks file for September.)

     Another player in Nari's eastward motion toward Okinawa was the
  mid-level Chinese ridge, which had become poleward-oriented and
  extended northeastward from the South China Sea.  By the 11th this
  ridge was weakening and Nari's eastward motion came to a halt.  The
  storm was still located within a weakness in the subtropical ridge,
  but an extension of the ridge was forecast to build slightly over the
  East China Sea over the next couple of days as a baroclinic system
  over the Sea of Japan lifted out toward the north.  By 14 September
  the tropical cyclone had come under the influence of a mid-level
  ridge over central China and was tracking westward.  As this ridge
  built to the north of Nari, the storm turned to the southwest toward
  Taiwan.

     The equatorward flow associated with the mid-level ridge situated
  by now over eastern China was the primary influence which steered Nari
  into Taiwan and southward down the backbone of the island.  The fact
  that the system was weakening, plus the distortion due to terrain 
  effects, led to considerable uncertainty at times in the location of 
  the center.  As the system re-organized in the Taiwan Strait, it came 
  under the influence of a building low and mid-level ridge over southern
  China which steered it westward and into the Chinese mainland.


  D. Comparisons Between JTWC and Other Centers
  ---------------------------------------------

     Normally I try to report on the significant differences in intensity
  between JTWC and other warning agencies in a separate section, but due
  to the manner in which I detailed Nari's history, I have incorporated
  those comparisons into the daily diary of the storm in Section B above.


  E. Meteorological Observations
  ------------------------------

  (1) Okinawa

     I have no rainfall amounts from Okinawa, and the only wind and SLP
  report available was one referenced in a e-mail from Mark Lander.  As
  Nari moved westward over the island the second time, Kadena Air Base
  recorded wind gusts in excess of 87 kts with a minimum pressure of
  974 mb.

  (2) Taiwan

     I have available some extensive rainfall records from Taiwan--far
  too many to present here.   These were sent to Huang Chunliang by
  Chun-Chieh Wu of the National Taiwan University, and Chunliang
  forwarded them to me.   Below are some of the totals for various
  periods at several different stations (all amounts are in mm).

     (a) For the 24-hour period ending at 16/1500 UTC:

     Chutzhu                        638
     Shun-Chi, Taipei County        546
     Tien-Mu, Taipei City           438
     Pai-Lan, Hsinchu County        425
     Kuang-Wu, Miaoli County        419
     Kao-I, Taoyuan County          419
     Pei-Kuan, Ilan County          321
     Nei-Hu                         305
     Sheuilin, Taichung County      174

     (b) For the 48-hour period ending at 17/1500 UTC:

     Tu-Chang, Ilan County         1051
     Chutzhu                       1012
     Ku-Lu, Ilan County             942
     Da-Jan-S, Taipei County        925
     Shun-Chi, Taipei County        867
     Neo-Tsu, Hsinchu County        850
     Nan-Kang, Taipei City          834
     Ta-Chi, Taoyuan County         799
     Tien-Mu, Taipei City           764
     Fong-Mei, Miaoli County        657
     Tsaoling, Yunlin County        514
     Fenchihu, Chiayi County        505
     Shui-Nan, Taichung City        458

     (c) Storm totals, from 15/1600 UTC through 17/2100 UTC:

     Tu-Chang, Ilan County         1255
     Ku-Lu, Ilan County            1174
     Chutzhu                       1145
     Da-Jan-S, Taipei County       1024
     Nan-Kang, Taipei City          915
     Neo-Tsu, Hsinchu County        912
     Shun-Chi, Taipei County        892
     Tien-Mu, Taipei City           883
     Ta-Chi, Taoyuan County         816
     Fong-Mei, Miaoli County        676
     Hen-Shan, Taichung County      550
     Fenchihu, Chiayi County        544
     Tsaoling, Yunlin County        535

     (d) Storm totals, from 17/1600 UTC through 18/2100 UTC:

     Chiayi City                    775
     Kuanzlin, Tainan County        743
     Shangder, Pingtung County      510
     Niu-Ton, Ilan County           436
     Chupei, Hsinchu County         397
     Pa-The, Taoyuan County         360
     Yueh-Mei, Kaohsiung County     358
     Chunan, Miaoli County          353
     Pei-Kang, Yunlin County        303
     Tung-Tou, Nantu County         277
     Anpu, Taipei City              259

     (e) Storm totals, from 18/1600 UTC through 19/1200 UTC:

     Chunjih, Pingtung County       217
     Fang-Lia, Pingtung County      146
     Hengchun, Pingtung County       97

  (3) Chinese Mainland

     Haifeng City and Shanwei City, located in Guangdong Province,
  recorded 241.9 mm and 307.9 mm, respectively, during the period from
  20/0800 through 21/0800 Beijing time (which is UCT plus 8 hours).
  Several stations recorded in excess of 100 mm during this period with
  Guangzhou, the provincial capital, reporting 77.9 mm.

     A special thanks to Mark, Chunliang, and Chun-Chieh for providing
  me with meteorological observations from Typhoon Nari.


  F. Damage and Casualties
  ------------------------

     I received no reports of any damage or casualties from Okinawa.
  Taiwan was especially hit hard by Typhoon Nari with the storm's rains
  causing some of the worst flooding ever experienced on the island.
  According to some information sent to me by Huang Chunliang, there
  were 93 deaths in Taiwan with 11 missing and 208 injured.  Thousands
  of homes and buildings were damaged by flooding.  Water supply to over
  1,200,000 households was cut off, electricity to 500,000 homes was
  disrupted, and 300,000 telephones were disabled.  Direct economic
  losses were estimated to be at least NT$ 30 billion.  (Some of this
  information was taken from the HKO's report on the storm, available
  at the following URL:

     http://www.weather.gov.hk/informtc/nari/report.htm>

     Many bridges and railway lines were washed away, and Taipei's
  Deputy Mayor stated that it might take months to get the city's Mass
  Rapid Transit system fully operational again.  The 93 fatalities
  resulting from Typhoon Nari rank this storm as one of the deadliest
  tropical cyclones on record in Taiwan.  From some information sent to
  me by Chunliang, the deadliest typhoons to affect the island since
  1958 have been:

     Ranking    Name       Date           Deaths     Missing
     -------    ----       ----           ------     -------
        1      Gloria    09 Sep 1963        204         88
        2      Pamela    12 Sep 1961        158        121
        3      Vera      31 Jul 1977        104         10
        4      Toraji    31 Jul 2001        103        111 (See Note)
        5      Shirley   31 Jul 1960        102         81
        6      Elsie     25 Sep 1969         93         12
        6      Nari      16 Sep 2001         93         11
        7      Fran      06 Sep 1970         89         41


  Note:  In the summary for July, 2001, I reported the death toll in 
  Taiwan from Typhoon Toraji at 35 with 108 persons missing at the time 
  of the report I had obtained, which was made two days after the storm 
  had struck.  The figures above reveal that Toraji turned out to be much
  more deadly than the preliminary report had indicated.

     In Guangdong Province, China, where Nari made its final landfall,
  three persons were reported killed.  Direct econmic losses were
  estimated to have been approximately 60 million yuan in the province.
  In Huilai County, 154 houses collapsed and 267 were partially damaged.
  Also, 3400 hectares of farmland were damaged by Nari in the province.



                     Typhoon Vipa  (TC-21W / TY 0117)
                             17 - 23 September
                     --------------------------------

  Vipa: contributed by Thailand, is a Thai woman's name


  A. Storm Origins
  ----------------

     Typhoon Vipa had its origin in a monsoon gyre--a relatively rare
  mode of the monsoon circulation in the Northwestern Pacific.  A long-
  lived upper-level LOW (TUTT cell), which had originated earlier near
  Wake Island, had by mid-September migrated westward and was located
  south of Japan.  The LOW had pulled the monsoon trough northward such
  that its associated cloud band was wrapped around the southern and
  eastern periphery of the large circulation--the signature of a monsoon
  gyre.  Typhoon Nari at the time was approaching northern Taiwan, and
  numerical model guidance had been consistently forecasting the
  development of a large low-level circulation east of the typhoon which
  would normally be regarded as the next tropical cyclone.  Instead, the
  forecast large LOW turned out to be the gyre.  The above information
  on the monsoon gyre was taken from an e-mail sent by Mark Lander, and
  Mark noted that a tropical cyclone or two forming in the monsoon cloud
  band was probably a good bet.

     An area of convection developed in the monsoon cloud band associated
  with the gyre on 16 September southwest of Iwo Jima and west-northwest
  of the Northern Marianas.     Deep convection was associated with a
  possible LLCC, and a 200-mb analysis indicated that an extension of an
  upper-level ridge was providing good outflow for the disturbance.  JTWC
  upgraded the development potential to fair at 1500 UTC, and a TCFA was
  issued at 2230 UTC.  Deep convection was increasing over the LLCC which
  was moving north-northeastward at 9 kts.  Early morning visible imagery
  showed low-level cloud lines wrapping in from the north toward the
  center, which appeared to be located on the northern edge of the
  convection.

     The first warning on Tropical Depression 21W, issued at 17/0000 UTC,
  placed the center about 300 nm west-northwest of the northernmost
  Marianas Islands.  The initial intensity of 25 kts was based upon 
  satellite CI estimates of 25 kts.  The MSW was increased to 30 kts six
  hours later based on synoptic reports.  A SSM/I pass at 17/0820 UTC
  revealed distinct rainbands and a more symmetric appearance to the
  depression, but the LLCC was still partially-exposed on the northern
  edge of the deep convection.   JTWC upgraded TD-21W to a 35-kt tropical
  storm at 1800 UTC when the center was located about 85 nm west-
  southwest of Iwo Jima, moving northward at 6 kts.     The system's
  organization continued to improve, and JMA upgraded the depression to
  Tropical Storm Vipa at 18/0000 UTC as it was passing about 85 nm west
  of Iwo Jima.


  B. Track and Intensity History
  ------------------------------

     The track of Typhoon Vipa was about as smooth and well-behaved as
  any forecaster could want--starkly contrasting with the incredibly
  convoluted track of Typhoon Nari.  The storm moved north-northwestward
  from its point of origin, executed a smooth recurvature south of Japan,
  and then accelerated northeastward on a very straight trajectory.
  After being upgraded to a tropical storm, Vipa remained at minimal
  tropical storm intensity for about a day, then intensified rather
  quickly to typhoon force.  The MSW (per JTWC) was still 35 kts at
  0600 UTC on 18 September, but was upped to 50 kts six hours later
  based on satellite CI estimates of 45 kts and a 18/1118 UTC SSM/I
  pass.  At 1800 UTC Vipa became a typhoon about 300 nm northwest of
  Iwo Jima or about 415 nm south-southwest of Tokyo.  Satellite imagery
  depicted a small, cloud-filled eye.   The cyclone was being steered
  by a subtropical ridge to the east, and was forecast to recurve to
  the northeast ahead of a baroclinic zone approaching from the west.

     By the time Vipa was upgraded to typhoon intensity, its forward
  motion had become northward, and by 19/0600 UTC the storm was moving
  north-northeastward.  Vipa reached its peak intensity of 75 kts at
  19/0000 UTC and remained at that level for 18 hours.  At 1800 UTC
  the storm was located roughly 200 nm south of Tokyo and was moving
  northeastward at 10 kts.  Enhanced infrared imagery depicted a slight
  erosion of the deep convection over the western semicircle--Vipa had
  begun to track beneath the upper-level southwesterlies.  JTWC reduced
  the MSW to 70 kts and forecast continued weakening and extratropical
  transition.  By 0600 UTC on 20 September Vipa was located about 135 nm
  east-southeast of Tokyo and had accelerated slightly to 15 kts.  The
  intensity was lowered to 65 kts and the JTWC warning noted that Vipa
  appeared to be rapidly transitioning into an extratropical cyclone.

     However, tropical cyclones are rather famous for producing the
  unexpected, and Vipa was no exception.  At 20/1200 UTC a 13-nm diameter
  eye had re-appeared in satellite imagery and CI estimates were back up
  to 75 kts, so JTWC raised the MSW back to that value where it remained
  for another 18 hours.    One CI estimate at 1800 UTC was as high as
  90 kts.  By 21/0600 UTC Vipa was located about 350 nm south of the
  Kuril Islands and racing northeastward at 39 kts.  Although one CI
  estimate was still 90 kts, the irregular eye had filled and JTWC
  decreased the intensity to 70 kts.  A 21/0731 UTC SSM/I pass revealed
  erosion of the eyewall and dry air entrainment associated with extra-
  tropical transition.  The MSW was further reduced to 55 kts at 1200
  UTC (although one CI estimate was 75 kts), and at 1800 UTC Vipa was
  declared extratropical about 400 nm east-northeast of the Kuril
  Islands.  The remnant extratropical storm continued northeastward,
  crossing the Aleutian Islands and entering the Bering Sea where it
  was still producing 45-kt gales at 23/0000 UTC. 

     In summary, Typhoon Vipa was at its peak intensity of 75 kts (per
  JTWC) from 0000 through 1800 UTC on 19 September, and also from 20/1200
  through 21/0600 UTC.  The estimated minimum central pressure reported
  in JMA's warnings was 970 mb at 19/1800 UTC.  At the time of Vipa's
  first peak intensity, gales extended outward 95 nm in the eastern
  semicircle and 55 nm to the west, while the radius of 50-kt winds was
  25 nm.  The storm was only slightly larger in areal extent during its
  second peak in intensity.


  C. Meteorological Observations
  ------------------------------

     I have received no surface observations taken from within or near 
  Typhoon Vipa.


  D. Comparisons Between JTWC and Other Centers
  ---------------------------------------------

     JMA and NMCC were the only two Asian warning centers issuing
  warnings on Typhoon Vipa.  JMA upgraded Vipa to typhoon intensity for
  only one six-hour period beginning at 19/1800 UTC.  The 10-min avg
  MSW was reduced to 55 kts at 20/0000 UTC and further to 50 kts at
  0600 UTC where it remained during the cyclone's second intensification
  phase.  NMCC, however, estimated Vipa's peak intensity at 70 kts during
  the storm's second intensity peak on 20 and 21 September.    This
  compares well with JTWC's estimated peak MSW of 75 kts.


  E. Damage and Casualties
  ------------------------

     I have received no reports of any damage or casualties resulting
  from Typhoon Vipa.



                  Typhoon Francisco  (TC-22W / TY 0118)
                            18 - 27 September
                  -------------------------------------

  Francisco: contributed by the United States, is a Chamorro man's
             name (Spanish form of Francis)


  A. Storm Origins
  ----------------

     JTWC issued a STWO at 15/0600 UTC which noted that an area of
  convection was developing approximately 200 nm east-northeast of
  Kwajalein.  A recent QuikScat pass had indicated the possible existence
  of a LLCC.   Diffluent flow overlay the region, southeast of a TUTT
  cell--a scenario somewhat similar to that preceding the development
  of Typhoon Danas earlier in the month.  The disturbance drifted west-
  ward and by 2300 UTC on the 16th was located 260 nm north-northwest of
  Kwajalein.  Animated infrared and visible imagery indicated that a LLCC
  was developing with increased convection and good outflow, therefore,
  the potential for development was upgraded to fair.

     The STWO issued at 17/0600 UTC relocated the system to a point about
  350 nm north of Kwajalein.    At 18/0000 UTC JMA began tracking the
  system as a 30-kt tropical depression located about 275 nm south-
  southeast of Wake Island.  JTWC issued a TCFA at 1200 UTC but placed
  the partially-exposed center about 100 nm east of JMA's 1200 UTC
  position.   Warnings on Tropical Depression 22W were initiated at
  19/0000 UTC with the system centered 390 nm southwest of Wake Island,
  moving west at 7 kts.  The initial MSW of 30 kts was based on satellite
  CI estimates of 30 kts.   The depression's center was located beneath
  the western edge of the deep convection.

     TD-22W was upgraded to a tropical storm at 19/1200 UTC.  The system
  had not moved very much and was located about 400 nm southwest of Wake
  Island.   By 20/0000 UTC the cyclone was moving west-northwestward from
  a point about 880 nm east-northeast of Guam.    CI estimates had all
  reached 35 kts by this time, so JMA upgraded the system to tropical
  storm status, assigning the name Francisco.


  B. Track and Intensity History
  ------------------------------

     Tropical Storm Francisco tracked steadily west-northwestward on
  20 September, reaching a point approximately 670 nm east of Saipan by
  1800 UTC.    For the first 24 hours or so after being upgraded to
  tropical storm status Francisco did not intensify, remaining at minimal
  tropical storm intensity.  However, by 20/1200 UTC deep convection had
  persisted and increased in areal coverage, so the MSW was increased to
  45 kts at this time and raised further to 50 kts at 1800 UTC based on
  CI estimates of 45 and 65 kts.  A 20/1032 UTC SSM/I pass depicted a
  banding feature developing over the northeastern quadrant.   Visible
  imagery early on the 21st resulted in a relocation of Francisco's
  center of over 100 nm to the northwest of the previous warning
  position.  The tropical cyclone continued to intensify as the 21st
  progressed.   The MSW had reached 60 kts by 0600 UTC, although multi-
  spectral imagery revealed that the LLCC became partially-exposed for
  a few hours.  Deep convection began to build once more over the vortex
  center, however, and at 1200 UTC JTWC upgraded Francisco to typhoon
  status based on CI estimates of 65 kts.  The storm was then centered
  approximately 300 nm east of the Northern Marianas Islands and moving
  northwestward at 12 kts.

     Typhoon Francisco was steered on a fairly smooth track along the
  southern and western periphery of a large mid-level HIGH centered
  well to the east of Japan.  As a major short-wave in the westerlies
  approached, the cyclone turned to the north and eventually northeast.
  Francisco continued to intensify, reaching an intensity of 90 kts by
  22/1200 UTC, however, the MSW plateaued at that level for about 24
  hours.   At 22/1200 UTC the storm was centered about 435 nm east-
  southeast of Iwo Jima and had begun moving to the north-northwest as
  it tracked around the periphery of the aforementioned HIGH.  A ragged
  eye 17-nm in diameter was becoming visible by 23/0000 UTC and persisted
  throughout the day.  Francisco passed about 350 nm east of Iwo Jima
  around 0600 UTC and by 1200 UTC was moving northward as it reached its
  peak intensity of 100 kts.  The typhoon appeared to be well-organized,
  sporting a 27-nm cloud-filled irregular eye, and animated water vapor
  imagery indicated a well-defined trough off the east coast of Japan
  that was helping to create a poleward outflow channel for Francisco.
  The minimum CP estimated by JMA was 945 mb at 24/0000 UTC.

     Francisco remained at its peak intensity for 18 hours, then began
  to weaken.  The storm was an average-sized typhoon--gales covered an
  area about 280 nm in diameter and storm-force winds extended outward
  about 50 nm from the center.  By 1200 UTC on 24 September Francisco
  was tracking north-northeastward at 16 kts from a point about 700 nm
  south-southeast of the Japanese island of Hokkaido.  Animated satellite
  imagery depicted a rapid weakening of the deep convection.    JTWC
  lowered the MSW to 90 kts based on CI estimates of 77 and 102 kts.
  A 24/1124 UTC SSM/I pass depicted tight but discontinuous convective
  banding.  The storm had begun to show signs of extratropical transition
  by 1800 UTC and this trend continued on 25 September.

     An SSM/I pass at 24/2056 UTC depicted an eroded eyewall with
  suppressed convection in the southern semicircle, and JTWC had reduced
  Francisco to minimal typhoon intensity by 25/0000 UTC when the storm
  was located about 400 nm east of Tokyo.  Francisco had temporarily
  moved due northward on the 24th but began to accelerate northeastward
  on the 25th as it got caught up in mid-latitude westerlies.  The storm
  had weakened to 60 kts by 0600 UTC and the intensity continued to drop
  throughout the day.   Dry air began to be entrained into the system and
  at 1800 UTC, JTWC declared Francisco to be extratropical and issued
  their final warning, placing the cyclone approximately 550 nm south of
  the Kamchatka Peninsula and racing northeastward at 36 kts.  The extra-
  tropical storm continued rapidly east-northeastward and by 0000 UTC on
  27 September was a 35-kt gale located south of the Aleutians and just
  east of the Dateline.


  C. Comparisons Between JTWC and Other Centers
  ---------------------------------------------

     JMA's intensity estimates compared fairly well with JTWC's after
  adjusting for the different averaging periods.  JMA was a little on
  the low side during the period when Francisco was intensifying to
  typhoon force--JMA reported the 10-min avg MSW at 45 kts at 21/1800 UTC
  when JTWC was estimating 75 kts (NMCC was 60 kts at this point), and
  JTWC had increased the MSW to 90 kts by the time JMA upgraded Francisco 
  to a 65-kt typhoon.  However, JMA "caught up" and their peak 10-min avg
  MSW of 85 kts agrees very well with JTWC's peak of 100 kts.  NMCC's
  estimates tended to be higher than JMA's with a peak estimated 10-min
  avg MSW of 90 kts from 23/0000 through 24/0000 UTC.

     
  D. Damage and Casualties
  ------------------------

     No damage or casualties are known to have resulted from Typhoon
  Francisco.



                 Typhoon Lekima  (TC-23W / TY 0119 / Labuyo)
                              21 - 30 September
                 -------------------------------------------

  Lekima: contributed by Vietnam, is the name of a tree whose fruit has
          only one seed surrounded by a yellow pulp which resembles an
          egg yolk


  A. Storm Origins
  ----------------

     A STWO issued by JTWC at 2300 UTC on 19 September noted that an
  area of convection associated with a monsoon trough had developed a
  few hundred miles east of Luzon in the Philippine Sea.  The convection
  was disorganized, although there were indications of a possible
  developing LLCC.    The disturbance drifted northward over the next
  24 hours and at 21/0000 UTC, NMCC elevated the system to depression
  status with 30-kt winds located about 275 nm east-southeast of the
  southern tip of Taiwan.  JTWC's daily STWO issued at 0600 UTC indicated
  that convection was increasing around a broad LLCC.  A TCFA was issued
  at 1900 UTC and the first warning on Tropical Depression 23W followed
  at 22/0000 UTC.  The initial MSW was set at 30 kts based on satellite
  CI estimates of 25 and 35 kts.

     JTWC upgraded the depression to tropical storm status on the second
  warning at 22/0600 UTC based on CI estimates of 30 and 35 kts.  The
  cyclone was then located about 250 nm southeast of the southern tip of
  Taiwan and was tracking toward the west-southwest at 7 kts.  By 1200
  UTC all the satellite intensity estimates had reached 35 kts and JMA
  upgraded the system to tropical storm status, assigning the name
  Lekima.  (PAGASA had also upgraded the system to tropical storm status
  at 22/0600 UTC and assigned the name Labuyo--the name of a species of
  extremely hot Filipino pepper.)


  B. Track and Intensity History
  ------------------------------

     Tropical Storm Lekima continued to drift west-southwestward on the
  22nd, becoming quasi-stationary for several hours approximately 115 nm
  east-northeast of Luzon.   The storm resumed a very slow westerly to
  west-northwesterly motion on 23 September which, by 1800 UTC, had
  brought the center to a position about 55 nm north of Luzon or about
  175 nm south-southeast of the southern tip of Taiwan.  Lekima remained
  a minimal tropical storm for the first 24 hours after it was upgraded,
  but a steady intensification trend began on the 23rd which resulted in
  JTWC's upgrading the cyclone to typhoon status at 1800 UTC.  The storm
  intensified slightly early on the 24th with winds reaching 75 kts by
  0600 UTC, but then the strengthening trend levelled off and Lekima's
  intensity remained static for about 24 hours.  All the Asian warning
  centers had upgraded Lekima to typhoon status by 24/1200 UTC with
  the CWB of Taiwan being the last to do so.

     Early in Lekima's life steering currents were weak, the primary
  influence being a mid-level subtropical ridge over China.  The typhoon
  didn't move all that much on the 24th--mainly drifting slowly toward
  the west-northwest. However, at 1800 UTC JTWC relocated the storm
  farther to the north (based on microwave imagery) to a point roughly
  110 nm south-southeast of Taiwan.   Typhoon Lekima was still located
  within a weakness in the mid-level subtropical ridge situated over
  China, but a sub-equatorial ridge to the southeast was complicating
  the steering pattern.   At 24/1800 UTC the storm was drifting very
  slowly north-northeastward, but on the 25th curved back to the north
  and later northwest toward Taiwan.  By 1800 UTC on 25 September Lekima
  was centered approximately 60 nm southeast of the southern tip of
  the island.  A slight weakening noted late on the 24th proved to be
  temporary and by 25/0000 UTC the storm's organization appeared to be
  improving.  A TRMM pass at 24/2002 UTC had revealed a small eye feature
  and a primary band wrapping around the western semicircle.  JTWC upped
  the MSW to 80 kts at 25/0600 UTC and to 90 kts at 1200 UTC.  A small
  banding eye was centrally located within organized convective spiral
  bands.  The intensity remained at 90 kts for the 1800 UTC warning,
  based on satellite CI estimates ranging from 77 to 102 kts.

     Based on JTWC's warnings, Typhoon Lekima reached its peak intensity
  of 95 kts at 26/0000 UTC when it was only about 25 nm off the coast of
  southeastern Taiwan.  The MSW was based on satellite CI estimates of
  90 and 102 kts.  Gales extended out from the center 140 nm in the
  southwestern quadrant and around 130 nm in the eastern semicircle.
  The radius of storm-force winds was estimated to be 60 nm.  At 0600
  UTC the eye was just about onshore about 170 nm south of Taipei and
  moving north-northwestward at 6 kts.  The storm made landfall around
  0900 UTC near the city of Tawu packing MSW of 90 kts, gusting to
  110 kts (per JTWC).   At 1800 UTC, with the center inland over south-
  eastern Taiwan, JTWC downgraded Lekima to a 50-kt tropical storm.
  All the Asian warning centers had also downgraded the cyclone by this
  time.

     By 27/0000 UTC the center of Tropical Storm Lekima had crossed
  over to the southwestern coast of Taiwan, and during the 27th drifted
  slowly northward along the west coast of the island.    The storm
  slowly weakened, and at 1800 UTC JTWC downgraded Lekima to a 30-kt
  tropical depression.  Interestingly, all the other warning centers were
  still classifying Lekima as a tropical storm at this point.  Exactly
  what Lekima did on 28 September depends on which TCWC's track one
  follows.  JTWC's operational track takes Lekima on an approximate
  counterclockwise loop over northern Taiwan.   JMA's track also shows
  some erratic motion and possibly describes a tight loop.  However,
  neither NMCC or HKO show this jaunt to the east over northern Taiwan.
  CWBT's track ends with the downgrading of Lekima to a depression at
  28/0000 UTC, but the final position was in the Taiwan Strait west of
  the northern end of the island.  JTWC re-upgraded Lekima to a minimal
  tropical storm for a 12-hour period beginning at 28/0600 UTC based on
  some synoptic reports of 35 kts.

     By 29/0000 UTC, JTWC, JMA, HKO and NMCC were all in agreement that
  the center of the weakening cyclone was over water west of Taiwan.
  Lekima continued drifting northward, brushing the Chinese coast before
  turning northeastward into the East China Sea.    HKO had downgraded
  Lekima to a depression at 28/0600 UTC, but JMA maintained it as a
  minimal tropical storm for another 30 hours, and NMCC did not downgrade
  Lekima until 29/1800 UTC.  The final NMCC warning, issued at 0000 UTC
  on 30 September, placed the 30-kt depression well east of the coast of
  China, roughly midway between northern Taiwan and Kyushu.


  C. Comparisons Between JTWC and Other Centers
  ---------------------------------------------

     Some of the major differences between the various warning agencies
  in Lekima's track during its weakening phase have already been noted.
  With regard to intensity comparisons, the storm's history can be looked
  at in three stages.  From Lekima's inception up through around 1800 UTC
  on 24 September, the MSW estimates (10-min avg) from the Asian warning
  centers were basically in good agreement with each other and, after
  adjusting for the different averaging time, agreed well with JTWC.
  And, as was noted earlier, during the storm's decaying phase, JMA and
  NMCC continued to classify Lekima as a tropical storm for 24 or more
  hours after JTWC's final downgrade of the system to tropical depression
  status.

     The period from around 25/0000 UTC until Lekima had made landfall
  in Taiwan at 0900 UTC on the 26th is very interesting.  JTWC increased
  the typhoon's intensity from 75 kts at 25/0000 UTC to the peak of
  95 kts at 26/0000 UTC.  JMA estimated their peak 10-min avg MSW of
  75 kts at 24/1800 UTC (when JTWC was reporting 70 kts), then decreased
  Lekima's intensity to 65 kts during the period when JTWC was
  intensifying the storm.  NMCC peaked the storm at 75 kts at 25/1200
  UTC, then reduced the winds slightly to 70 kts until landfall.
  PAGASA's peak 10-min avg MSW of 75 kts was reached at 25/1200 UTC, but
  that agency also weakened Lekima to 65 kts before the storm made land-
  fall.    HKO and CWBT both estimated a peak intensity for Lekima of
  70 kts around the time of JTWC's 95-kt peak MSW.  (Anyone interested in
  trying to sort out these differences in further detail should consult
  the track for Typhoon Lekima in the cyclone tracks file for September.)


  D. Meteorological Observations
  ------------------------------

     As was the case with Typhoon Nari, Chun-Chieh Wu of the National
  Taiwan University sent some extensive rainfall records to Huang
  Chunliang, who then forwarded them on to me.  Below is a summary of
  some of the rainfall totals for various stations for several different
  periods of time (all amounts are in mm).

     (a) For the 24-hour period ending at 23/1500 UTC:

     Taipin-S, Ilan County          328
     Ku-Lu, Ilan County             294
     Hsi-Ceiu, Hsinchu County       229
     Tatung, Taipei County          156
     Chutzhu                        139
     Yu-Li, Hualian County          138

     (b) For the 48-hour period ending at 24/1500 UTC:

     Ku-Lu, Ilan County             430
     Taipin-S, Ilan County          412
     Chutzhu,                       273
     Yu-Li, Hualian County          272
     Hsi-Ceiu, Hsinchu County       266
     Da-Jan-S, Taipei County        249

     (c) For the 72-hour period ending at 25/1500 UTC:

     Ku-Lu, Ilan County             912
     Taipin-S, Ilan County          792
     Hsi-Ceiu, Hsinchu County       594
     Da-Nung, Hualian County        458
     Hualian City                   432
     Chutzhu                        426
     Taipin, Taipei County          348
     Chengkung, Taitung County      234

     (d) Storm totals, from 22/1600 UTC through 26/0500 UTC:

     Ku-Lu, Ilan County            1167
     Taipin-S, ILan County          973
     Hsi-Ceiu, Hsinchu County       695
     Da-Nung, Hualian County        590
     Chutzhu                        582
     Hualian City                   494
     Taipin, Taipei County          483
     Chengkung, Taitung County      405

     (e) For the 24-hour period ending at 26/1500 UTC:

     Tai-An, Hualian County         508
     Lu-Yeh, Taitung County         460
     Ceietuan, Ilan County          437
     Ku-Lu, Ilan County             338
     Kuang-Gu, Hualian County       308
     Anpu, Chutzhu                  297
     Taitung City                   295
     Fushan, Taipei County          238
     Mu-Tan, Pingtung County        223
     Ken-Ting, Pingtung County      216

     (e) For the 48-hour period ending at 27/1500 UTC:

     Pu-Lo-Wan, Hualian County      616
     Ceietuan, Ilan County          529
     Ku-Lu, Ilan County             498
     Lu-Yeh, Taitung County         482
     Tung-Men, Hualian County       425
     Shi-Nan, Kaohsiung County      375
     Fushan, Taipei County          365
     Mu-Tan, Pingtung County        364
     Anpu, Chutzhu                  344
     Taitung City                   310
     Ken-Ting, Pingtung County      264

     Patrick Hoareau sent me some observations from Basco in the Batan
  Islands where the eye of Lekima passed very near to.  At 24/2100 UTC
  Basco was reporting a SLP of 974 mb with north-northeast winds of
  51 kts.  At 25/0100 UTC the wind report was north-northwest at 51 kts
  with a SLP of 971 mb.

     In Fujian Province on the Chinese mainland, seven counties reported
  rainfall amounts exceeding 100 mm for the period 28/0000 UTC through
  29/0300 UTC.  Among these, Xiapu County reported the highest rainfall
  accumulation with 268 mm being recorded.


  E. Damage and Casualties
  ------------------------

     Being a slow-moving typhoon and coming right on the heels of the
  destructive Nari, the potential for another disaster on Taiwan was
  very real, but it seems that Lekima was much less destructive than
  its predecessor.  The storm did bring some very heavy rains, but the
  author has located only scanty references to damage caused by the
  storm.  Landslides blocked some highways, and 5600 persons were with-
  out power at one point.   The only Taiwanese fatalities I could find
  were two fishermen who drowned when their boat capsized in heavy seas
  generated by the storm.  A third person on board was missing and
  presumed drowned.   One man perished on Luzon in a river swollen by
  heavy rains spawned by Typhoon Lekima as it brushed that area.  On
  the mainland a landslide on 29 September triggered by the heavy rains
  of Lekima was responsible for one death (with another missing) near
  Wenzhou in Cangnan County, Zhejiang Province.

     If any further reports of damage or fatalities associated with
  Typhoon Lekima become available later, I shall include them in a
  future summary.

  ***********************************************************************

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

  Activity for September:  1 tropical cyclone of gale intensity

  NOTE:  The tracking and intensity information for North Indian Ocean
  Basin tropical cyclones is based primarily upon operational warnings
  from the Joint Typhoon Warning Center of the U. S. Air Force and Navy
  (JTWC) at Pearl Harbor, Hawaii.  Occasionally some information may
  be gleaned from the daily Tropical Weather Outlooks and other bulletins
  issued by the Indian Meteorological Department (IMD), which is the 
  WMO's RSMC for the basin.
          
     The MSW are based on a 1-min averaging period, which is used by
  all U. S. civilian and military weather services for tropical cyclone
  warnings.  For synoptic observations in the North Indian region,
  both 10-min and 3-min average winds are employed, but IMD makes no
  attempt to modify the Dvorak scale for estimating tropical cyclone
  intensity; hence, a 1-min avg MSW is implied.  In the North Indian
  basin JTWC usually does not initiate warnings until a system is
  well-organized and likely to attain tropical storm status within
  48 hours.



                        Tropical Cyclone  (TC-02A)
                             25 - 28 September
                        --------------------------

  A. Storm Origins
  ----------------

     An area of convection developed on 24 September about 200 nm west-
  southwest of Bombay, India.  Satellite imagery depicted improving
  organization with weak convective banding.  A 24/0038 UTC QuikScat pass
  indicated the existence of a well-defined LLCC on the eastern edge of
  the convection.  The disturbance was located under a 200-mb subtropical
  ridge axis in an environment of weak vertical shear.  The MSW was
  estimated at 25 kts (which would have qualified it for tropical
  depression status in the NWP basin).  The development potential was
  considered to be fair.   JTWC issued a TCFA at 1100 UTC based on a
  24/0514 UTC SSM/I pass which indicated developing banding west of the
  LLCC.

     The Indian Meteorological Department (IMD) classified the system as
  a depression on the 25th, the center being located about 325 nm west-
  southwest of Bombay at 0300 UTC.  JTWC issued a second TCFA at 25/1100
  UTC.  Earlier TRMM and SSM/I passes had failed to clearly define a
  LLCC, but multi-spectral animated imagery depicted weak cumulus lines
  outlining a partially-exposed LLCC just inside the eastern edge of
  the convection.  The LOW continued to move west-northwestward across
  the Arabian Sea, and by 26/0000 UTC animated visible imagery indicated
  that the LLCC had moved under the deep convection.  Satellite intensity
  estimates were 30 kts, but synoptic reports warranted upgrading the
  system to a 35-kt tropical cyclone.  The first JTWC warning placed the
  center of TC-02A about 350 nm west-southwest of Bombay, moving west-
  northwestward at 6 kts.   (NOTE: Information available to the author
  from IMD is quite sketchy, but that agency also classified TC-02A as
  a cyclonic storm (i.e., tropical storm) on 26 September.)


  B. Track and Intensity History
  ------------------------------

     Initially the cyclone was forecast to strengthen into a 50-kt system
  and make landfall on the southern end of the Arabian Peninsula, but
  this scenario never materialized.  Easterly shear inhibited significant
  strengthening of the tropical cyclone and eventually brought about its
  demise.  By 1800 UTC on the 26th the LLCC was exposed approximately
  45 nm east of the convection.   The JTWC warning at 27/0600 UTC still
  anticipated some slight strengthening, but by 1200 UTC the official
  forecast was calling for no additional intensification and gradual
  dissipation.  The center of TC-02A was then located about 230 nm east-
  southeast of Masirah Island off the coast of Oman.

     At 1800 UTC on the 27th the MSW was lowered to 30 kts, based on
  satellite CI estimates of 25 and 30 kts.  Animated infrared satellite
  imagery depicted weakening convection 35 nm west of the LLCC.  The
  system drifted westward during the next 24 hours as it slowly weakened.
  By 28/0600 UTC satellite imagery depicted a fully-exposed LLCC with no
  associated deep convection.  JTWC issued the final warning on this
  system at 1800 UTC, placing the dissipating center about 100 nm east-
  southeast of Masirah Island.


  C. Damage and Casualties
  ------------------------

     No damage or casualties are known to have resulted from this weak
  tropical cyclone.
 
  ***********************************************************************

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

  Activity for September:  No tropical cyclones

  ***********************************************************************

  NORTHWEST AUSTRALIA/SOUTHEAST INDIAN OCEAN (AUW) - From 90E to 135E

  Activity for September:  No tropical cyclones
                       
  ***********************************************************************

  NORTHEAST AUSTRALIA/CORAL SEA (AUE) - From 135E to 160E

  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:  hrd-type42.nhc.noaa.gov [140.90.176.206]
       (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:   sep01.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:  sep01.sum, for
  example.

    Back issues can also be obtained from the following websites
  (courtesy of Michael Bath, Michael V. Padua, Tom Berg, Michael
  Pitt, and Rich Henning):

    http://australiasevereweather.com/cyclones/>
    http://www.typhoon2000.ph> OR http://66.40.4.61>
    http://www.hurricanealley.net/>
    http://www.qisfl.net/home/hurricanemike>
    http://www.met.fsu.edu/gsc/Docs/Grads/henning/cyclones/>

  NOTE:  The URL for Michael V. Padua's Typhoon 2000 website has
  changed.

     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:

    http://www.met-office.gov.uk/sec2/sec2cyclone/sec2cyclone.html>


                   TROPICAL CYCLONE REPORTS AVAILABLE

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

     The URL is:  http://199.10.200.33/jtwc.html>

     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:  http://www.nhc.noaa.gov>


  Prepared by: Gary Padgett
  E-mail:  [email protected]
  Phone:  334-222-5327 (nights & weekends) / 850-882-2594 (weekdays)

  ***********************************************************************
  ***********************************************************************

Document: summ0109.htm
Updated: 27th December 2006

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