Tropical Cyclones
Main Index Home Page Stock Weather Photos Australian Severe Weather Forum Storm News and Storm Chasing Reports Tropical Cyclones / Hurricanes / Typhoons Weather Data and Links Wild Fires / Bushfires Weather Observation Techniques Weather Picture Catalogue Tornado Pictures and Reports Stock Video Footage and DVDs for sale
Monthly Global Tropical Cyclone Summary August 2003
[Summaries and Track Data] [Prepared by Gary Padgett]

                               AUGUST, 2003

  (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.)


                             AUGUST HIGHLIGHTS

  --> Bermuda struck by most intense hurricane in many decades (actually
      in early September)
  --> Strong tropical storm strikes northeastern Mexico
  --> Category 2 hurricane strikes southern Baja California Peninsula
  --> Hurricane threatens Hawaii but passes to south
  --> Northwest Pacific active with four typhoons--several Asian nations
      adversely affected


               ***** Feature of the Month for August *****

                       FOR SEPTEMBER AND OCTOBER

     Last month I highlighted a technique developed by Eric Blake
  for forecasting Atlantic tropical cyclone activity for the month
  of August.  This month's feature concerns similar forecasts for
  the months of September and October developed by Philip Klotzbach,
  a research associate at Colorado State University (CSU) and
  currently a member of the CSU forecast team headed by Dr. William
  Gray.   Over the past 20 years Dr. Gray and his team have issued
  seasonal forecasts for the Atlantic basin which have shown considerable
  skill at predicting overall seasonal activity at lead times varying
  from 3 to 10 months.  However, significant month-to-month and multi-
  week variability exists within most seasons.  Very active hurricane
  seasons may contain periods of 2-4 weeks with little activity, while
  inactive years can exhibit short periods in which the tropics are
  quite busy.

     The monthly feature for July included several examples of this
  intraseasonal variability, so those won't be repeated here.  Also,
  in the June summary's monthly feature, definitions can be found
  for the terminology employed in the CSU forecasts (e.g., NS, H, IH,

     The success of Eric's initial August forecast in 2000 prompted
  Phil to investigate the feasibility of similar forecasts for the
  months of September and October.   No formal forecast was made
  in 2001, but a statement was included in the early August, 2001,
  CSU forecast release to the effect that September of 2001 would
  likely experience above-normal activity.  And since October tropical
  cyclone activity tends to correlate well with September activity,
  Phil indicated he felt that October, 2001, would also be an active
  month in the tropics.  Those general forecasts verified quite well:
  September's NTC was 58% and October's was 28%, well above the average
  NTC values for the two months of 48% and 17%, respectively.  (See the
  URL:> )

     September is the most active month for tropical activity in the
  Atlantic basin with an average of almost 50% of the annual NTC
  occurring during the month.   Phil based his prediction scheme for
  forecasting September tropical cyclone activity on reanalysis data
  from the National Centers for Environmental Prediction and the National
  Center for Atmospheric Research (NCEP/NCAR).  Based on hindcasting
  results from 1950-2000, from 30-75% of the variance of most tropical
  cyclone parameters can be hindcast by the end of July, with this
  hindcasting skill improving to 45-75% by the end of August.  Simple
  least-squares linear regression techniques were utilized to calculate
  hindcast skill, and variables were selected that explained the
  largest degree of variance when combined with the other predictors
  in the scheme.

     With regard to October-only forecasts, the approach is somewhat
  different.  Utilizing the NCEP/NCAR reanalysis data, four predictors
  were found which, in combination, explained about 50% of the variance
  of October NTC.  However, no combinations of predictors has yet been
  identified which can explain large amounts of variance in the
  individual tropical cyclone parameters (NS, NSD, etc).    So the
  procedure for the October, 2003, forecast was to forecast the overall
  NTC and then adjust the other parameters accordingly.

     The predictors used tend to be global in nature and include zonal
  and meridional winds at 200 and 1000 mb and sea level pressure
  measurements at various global locations.   Many of the predictors
  are strongly correlated with global modes such as ENSO and the
  Pacific Decadal Oscillation (PDO).   As was the case in describing
  Eric's August-only forecast scheme last month, the details of Phil's
  prediction schemes are better described with graphics, and charts
  and tables giving much more information may be found at the following
  links on the CSU website:>>>>>

     The full paper on the September-only forecast scheme will be
  published sometime in 2004 in _Weather and Forecasting_.  Phil
  sent me a copy of this highly interesting document, which is very
  detailed, in .pdf format, and he has indicated he would be happy
  to e-mail a copy to any interested person who would like to have
  one.  Phil can be reached at:  [email protected] .

     The remainder of this article will focus on the specifics of
  Phil's September forecast for 2002, and his September and October
  forecasts for 2003.  Since data through the end of July is needed
  for the August forecast, only one August-only forecast can be
  issued each year.   However, two September-only forecasts are made,
  the first utilizing data through the end of July, and the second
  based upon data collected through August.  In a similar fashion,
  three October-only forecasts are issued: one in early August, one
  around 1 September, and another around 1 October.

     The forecasts issued for September, 2002, as well as the observed
  statistics, are presented in the following table:


  Parameter      1 Aug Forecast        1 Sep Forecast       Observed

    NS                 3                     3                 8
    NSD               18                    13                37.25
    H                  2                     2                 4
    HD                 6                     7                 8
    IH                 1                     1                 1
    IHD                2                     2                 1.75
    NTC               30                    26                56

     The activity was significantly underforecast, but September of
  2002 presented a particularly difficult situation to have forecasted
  accurately.  With an El Nino in place and some additional negative
  atmospheric variables present, the 2002 Atlantic season had been
  rather quiet through the end of August (4 generally short-lived NS
  and no hurricanes), and indications were that this quiet trend would
  continue.  However, things turned around abruptly in September.  The
  total of 8 NS was the highest ever observed to develop in any calendar
  month, and conditions during the middle and latter portions of the month
  resembled a more typically active Atlantic season with the formation of
  intense Caribbean and Gulf of Mexico hurricanes Isidore and Lili.
  Also, the extremely long-lived Kyle added more than 15 days to the
  NSD total.

     No forecast was made for October in 2002, but the early August
  forecast release contained an implicit forecast for the October-
  November period for an NTC of 12%.  This was obtained by taking the
  seasonal forecast parameters for the remainder of the season after
  1 August, and simply subtracting the forecast values for August
  and September.  The NTC for the final two months of the hurricane
  season was further reduced to 8% on 1 September.  The NTC for October
  of 2002 was 17%--exactly average--but the month produced no tropical
  cyclones.  All the NTC was generated by Hurricane Lili and Tropical
  Storm Kyle, both inherited from September.

     The following table contains the September-only forecasts and
  verification for September, 2003 (based on operational intensities):

  Parameter      1 Aug Forecast        1 Sep Forecast       Observed

    NS                 4                     4                 4
    NSD               14                    18                29
    H                  2                     2                 3
    HD                 6                    11                22.25
    IH                 1                     1                 1
    IHD                1.25                  6.5              13.5
    NTC               33                    55                92

     The September NTC was once again underforecast, but also once again
  there was an anomalous occurrence which would have been very difficult
  to have predicted.  It is very unusual to see two intense hurricanes
  exist for as long as did Fabian and Isabel.  The two storms together
  generated 14.75 IHD (Hurricane Fabian was an IH for 1.25 days in
  August).  Those two great hurricanes alone generated an NTC of 95%!
  However, Phil's forecast for the numbers of storms (4 NS, 2 H, 1 IH)
  was almost exactly on target, spoiled only by Kate's six hours of
  hurricane intensity early on 30 September.

     The following table summarizes the three October-only forecasts
  along with the observed October activity (based on operational


  Parameter   1 Aug Fcst    1 Sep Fcst    1 Oct Fcst     Observed

    NS            3             3              3             3
    NSD          12            12             19            21.5
    H             2             2              2             0
    HD            6             6              7             5.5
    IH            0             1              0             1
    IHD           0             1              0             1.5
    NTC          19            30             21            29

     The situation in October was a little odd.  None of the three
  named storms which actually formed in the month reached hurricane
  intensity, but Kate (which had briefly reached hurricane intensity
  in September but weakened) regained hurricane intensity early in
  October and was a hurricane for 5.5 days.  However, under the rules,
  Kate counts as a September hurricane.  But since it initially
  reached Category 3 status in October, it counts as an October IH.
  So the 1 September forecast for October was right on target, but
  the final update on 1 October was still fairly good in that it 
  forecast above-normal tropical cyclone activity for the month.

     It is to be expected that as more experience is gained, the
  quality of the individual monthly forecasts will improve.  And if
  the monthly forecasts for the three primary months of the Atlantic
  hurricane season improve, then the overall seasonal forecasts should
  correspondingly become better.

     I'd like to thank Phil for giving me permission to feature his
  work and for sending me a copy of his research paper.  Again, as a
  reminder, persons interested in receiving a copy of the September
  forecast paper should contact Phil at the e-mail address given above.

                            ACTIVITY BY BASINS

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

  Activity for August:  1 tropical depression
                        2 tropical storms **
                        1 major hurricane

  ** - Tropical Storm Erika has been upgraded to minimal hurricane
       status in post-storm analysis

                        Sources of Information

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

                   Atlantic Tropical Activity for August

     Over the period 1950-2002, August has averaged 2.7 named storms, 1.5
  hurricanes and 0.6 intense hurricanes.  In terms of overall activity,
  August, 2003, was a rather average month with 3 named storms, 1 hurricane
  and 1 intense hurricane.  Tropical Storm Erika formed on 14 August in the
  eastern Gulf of Mexico from a disturbance of subtropical origin which
  had traveled westward for several days from the Atlantic north of
  Hispaniola.  Erica sailed westward at an unusually rapid speed for a Gulf
  of Mexico storm and made landfall in northeastern Mexico only two days
  after its formation southwest of Tampa.  Erica intensified fairly quickly
  as it neared landfall and may possibly have been a minimal hurricane when
  the center moved onshore.

     Tropical Storm Grace formed in the western Gulf of Mexico on 30 August
  and limped ashore in Texas as a minimal tropical storm the next day.  The
  main storm to form in August was mighty Hurricane Fabian.  This large,
  severe Cape Verde hurricane tracked westward for many days, eventually
  recurving along a smooth, parabolic track which took the center just west
  of Bermuda, bringing Category 3 hurricane conditions to the small island.
  Fabian was the most intense hurricane to strike the British colony in
  many decades and was rather destructive to the island.

     One non-developing depression formed during August in the northeastern
  Caribbean.  A strong tropical wave entered the Caribbean on 20 August,
  and the next day a ship reported a west wind south of the center.  There-
  fore, advisories were initiated on Tropical Depression 09 at 2100 UTC
  with the poorly-defined center located approximately 225 nm south of San
  Juan, Puerto Rico.    The depression was expected to develop into a
  tropical storm and be near hurricane strength in 96 hours.    On the
  morning of the 22nd, high-resolution visible satellite imagery indicated
  that the depression was becoming better organized with a well-defined
  cloud band to the north and east of the center and good outflow in all
  quadrants except the west.  However, a reconnaissance plane investigating
  the system south of Santo Domingo in the afternoon could not find a
  closed circulation, so the system was downgraded to a tropical wave.
  Even so, the system had a vigorous mid-level circulation and a well-
  defined cloud pattern in satellite imagery, and data T-numbers were 2.0
  and 2.5 from the various satellite analysis agencies.  Regeneration was
  considered a possibility, but this never materialized.

  NOTE: A short report on Tropical Depression 09, written by Lixion
  Avila, can be found on TPC/NHC's website at the following URL:>

  Also, additional reports on some of the other Atlantic tropical cyclones
  of 2003 have already been placed on TPC/NHC's website:>

                          TROPICAL STORM ERIKA
                             12 - 17 August

  A. Storm Origins

     The precursor of Tropical Storm Erika can be traced back to an area
  of disturbed weather which appeared in the central subtropical Atlantic
  almost a week before the storm was named.     On 8 August a small
  circulation accompanied by showers and gusty winds formed in the Atlantic
  between Bermuda and the Azores.  By the morning of the 9th the system was
  located about 800 nm east-southeast of Bermuda moving west-southwestward,
  and appeared to have become slightly better organized.   The motion
  became more westward at a rather quick pace of 17 kts over the next
  couple of days, but upper-level conditions had become less favorable and
  further development was hindered for the time being.  Tropical Weather
  Outlooks issued by TPC/NHC on 11 August indicated that the disturbed
  weather was due to the interaction of an upper-level LOW and a sharp
  surface trough.   By afternoon that day the disturbance was located about
  350 nm north of Puerto Rico, moving west at 17 kts.    The associated
  convection was still not organized very well.

     On 12 August satellite imagery indicated the formation of a weak
  surface LOW about 240 nm east of the central Bahamas, accompanied by an
  increase in shower activity near the center.  Also, it was noted that
  upper-level winds were becoming more favorable for tropical development.
  During the afternoon, however, convection diminished near a small, low-
  level swirl north of the Turks and Caicos, and a reconnaissance plane
  investigating the system found a mid-level circulation 260 nm to the
  northeast of the weak low-level swirl.  On the afternoon of the 13th
  a Special Tropical Disturbance Statement was issued by TPC/NHC.  A
  reconnaissance plane still found no closed surface circulation, but
  did measure winds to near tropical storm force to the north and north-
  east of the area of lowest pressure.  Concurrent satellite imagery
  indicated that the system's organization was improving.

     Little change was noted as the 13th rolled over into the 14th.  The
  disturbance continued westward, passing over the southern tip of the
  Florida Peninsula during the morning of 14 August.  Satellite images
  depicted a well-organized system, but surface observations from the
  Keys indicated only a broad low-pressure area without a well-defined
  closed circulation, and pressures were not falling significantly either.
  However, a reconnaissance mission into the disturbance in the afternoon
  found a small, poorly-defined circulation with 300-m winds of 47 kts to
  the north of the developing center.  A central pressure of 1011 mb was
  measured along with a northwest wind about 100 m above the surface.
  Based on this information, along with well-defined satellite and radar
  signatures, the system was upgraded to Tropical Storm Erika at 2100 UTC,
  located in the eastern Gulf of Mexico southwest of Ft. Myers, Florida,
  and about 300 nm southeast of the mouth of the Mississippi River.
  Erika was scooting rapidly westward at 18 kts, guided by a persistent
  strong HIGH over the south-central United States.  The initial advisory
  predicted landfall in southern Texas or northeastern Mexico in 36 to 48
  hours--a forecast which verified very well.

  B. Storm History

     A reconnaissance flight into the cyclone during the evening hours
  found a much better defined system.  The peak winds measured were about
  the same as previously, but the central convection had increased
  markedly.  Based on the improved organization, the MSW was upped to
  40 kts at 15/0300 UTC.  At 0900 UTC on the 15th Erika's center was
  located approximately 450 nm east of Brownsville, Texas, moving westward
  at the unusually rapid translational speed of 21 kts--very rapid at least
  for a westward moving system in the Gulf of Mexico.  A reconnaissance
  plane early in the morning found a CP of 1008 mb and a peak FLW of 50 kts
  about 60 nm northwest of the center.    A tropical storm warning had
  been issued for the southern Texas coast, and a hurricane watch was in
  effect for northeastern Mexico.  The intensity was increased to 45 kts
  at 1500 UTC, based on Dvorak ratings of T3.0 from the satellite analysis
  agencies, plus a 40-kt ship report at 1200 UTC from a location north of
  the center.  There appeared to be some easterly shear, but conditions
  overall favored some strengthening.

     At 15/2100 UTC Erika's center was located about 215 nm east of Browns-
  ville, moving west at 19 kts with the MSW now estimated at 50 kts.  Erika
  appeared to be better organized with large, curved bands, good outflow,
  and very strong convection near the center.  A reconnaissance plane found
  that the pressure had fallen to 1000 mb with a maximum 850-mb FLW of
  57 kts.  This would normally correspond to a surface MSW of about 45 kts,
  but with the fall in pressure, it was assumed that the plane had not
  sampled the actual maximum winds.   Erika was forecast to be a minimal
  hurricane at landfall.   During the evening of the 15th WSR-88D data from
  Brownsville revealed that Erika was sporting a 30-nm wide eye with
  asymmetric surrounding convection, primarily in the western semicircle.
  The minimum pressure measured was 992 mb, and the peak FLW measured by
  a reconnaissance plane was 67 kts at 700 mb.    Satellite intensity
  estimates were 65 kts from TAFB and SAB and 55 kts from AFWA, so the
  intensity was increased to 60 kts at 16/0300 UTC.

     At 16/0900 UTC the storm was centered only about 40 nm southeast of
  Brownsville, moving slightly south of due west at 16 kts.  The Browns-
  ville radar depicted a well-organized system with the eyewall open toward
  the northeast.  Doppler winds were 80 kts to the north of the center and
  almost as high to the south.   By 1500 UTC Erika's center was located
  inland in Mexico about 85 km southwest of Matamoros.  The discussion
  bulletin at 1500 UTC was very interesting, indicating that Erika possibly
  briefly reached hurricane intensity at landfall.  Dvorak ratings from the
  satellite agencies were T4.0 and T4.5 (65 and 77 kts).   Also, the high-
  resolution radar at Brownsville reported peak winds of 91 kts at 750 m
  in a small area to the southeast of the center along the coast.   In the
  opinion of the Hurricane Specialist who authored the discussion bulletin,
  this corresponds to at least 65 kts at the surface.  Minimal hurricane
  intensity is also supported by a pressure of 987 mb and a closed eyewall 
  as reported by a reconnaissance aircraft.

  (NOTE: I have learned from a telephone conversation with one of the NHC
  Hurricane Specialists that the issue of Erika's intensity at landfall is
  still being discussed, and there is a good possibility that in the Best
  Track Erika will be upgraded to a minimal hurricane at landfall.)

     Erika continued moving fairly quickly inland as it weakened.  By 2100
  UTC the cloud pattern was still quite organized with deep convection and
  excellent outflow, but the latest available radar data indicated that the
  central features were no longer well-defined.  Thus, Erika was downgraded
  to a 30-kt depression.  By 17/0300 UTC the depression was breaking up
  over the mountains of northeastern Mexico.  Satellite imagery and radar
  data from Brownsville showed a significant decrease in the intensity and
  coverage of convection, and surface observations from Monterrey and
  Saltillo indicated peak winds of 20 kts or less.  The final advisory on
  Erika, issued at 17/0300 UTC, placed the weakening depression about
  85 km south-southwest of Monterrey, moving west at 12 kts, and forecast
  to completely dissipate within the next 12-24 hours.

  C. Damage and Casualties

     Even though Erika made landfall in northeastern Mexico as a strong
  tropical storm (possibly as a minimal hurricane), I have been unable
  to locate any reports of damage or casualties resulting from the

  (Report written by Gary Padgett)

                            HURRICANE FABIAN
                        27 August - 9 September

  NOTE: The official TPC/NHC storm report on Hurricane Fabian is now
  available online at the following URL:>

     The very informative report was jointly authored by Richard Pasch,
  Eric Blake and Dan Brown, and I would encourage all readers to access
  this report.  However, the report was not available until after I had
  completed my summary of Fabian, so I am sending that one along as well.

  A. Storm Origins

     A large tropical wave moved off the west African coast on 24 August.
  The Tropical Weather Outlook issued by TPC/NHC at 0930 UTC on the 25th
  mentioned that the wave was located about 285 nm southeast of the Cape
  Verde Islands, moving west at 13 kts, and that conditions appeared to
  be favorable for further development.  By the morning of 26 August a
  well-defined surface LOW accompanying the wave was located approximately
  225 nm southeast of the Cape Verdes.  Convective activity had increased
  near the center overnight and atmospheric conditions were believed to
  be conducive for continued development.   Convection gradually increased
  in coverage and organization, and at 2100 UTC on the 27th, the first
  advisory was issued on Tropical Depression 10, located about 415 nm west
  of the southernmost Cape Verdes and moving westward at 17 kts with the
  MSW estimated at 25 kts.

     The intensity was upped to 30 kts at 28/0300 UTC based on T2.0 Dvorak
  ratings from TAFB and SAB, plus data from a scatterometer pass.
  Convection was beginning to take on a more banded appearance.   The
  depression remained somewhat disorganized early on the 28th--at least
  three low-level swirls were being identified as centers--so the MSW
  remained at 30 kts even though TAFB and SAB had bumped up their estimates
  to 35 kts by 0600 UTC.  During the day the cloud pattern's organization
  improved with more prominent banding features and an increase in central
  convection.   With T-numbers ranging from 35 to 45 kts, the depression
  was upgraded to Tropical Storm Fabian at 2100 UTC.  The new tropical
  storm was located about 870 nm west of the southernmost Cape Verdes and
  moving quickly westward at 19 kts.  The MSW was estimated at 40 kts,
  and continued intensification was forecast with Fabian progged to reach
  hurricane intensity in 72 hours.

  B. Storm History

     Tropical Storm Fabian intensified steadily as it moved rather quickly
  westward across the Atlantic.  By mid-morning on the 29th the cloud
  pattern consisted of a cyclonically curved band which wrapped around the
  center of circulation.    A 29/0930 UTC SSM/I pass showed a ring of
  convection associated with Fabian's center.  T-numbers from all satellite
  agencies had reached 3.5 by 1200 UTC, but the MSW remained at 50 kts on
  the 1500 UTC advisory due to the ragged cloud pattern.  During the after-
  noon the convective signature and outflow continued to improve, and
  Fabian was upgraded to a hurricane at 2100 UTC, only 24 hours after being
  named as a tropical storm.  Hurricane Fabian was located about 950 nm
  east of the Lesser Antilles, and the westward motion had slowed down a
  bit to 14 kts.

     The season's third hurricane continued to intensify at a steady pace.
  Early in the morning of 30 August Fabian was sporting cloud tops to
  -80 C near the center with evidence of outer banding in the eastern semi-
  circle.  Winds were upped to 70 kts at 0900 UTC and further to 85 kts
  at 1500 UTC, thus making Fabian a Category 2 hurricane on the Saffir/
  Simpson scale.  An eye feature was becoming apparent in visible images,
  and by afternoon the eye was well-defined and surrounded by very deep
  convection.  T-numbers had reached 5.5, so at 2100 UTC Fabian's MSW was
  increased to 100 kts, making the storm the first major hurricane of the
  season only 72 hours after being classified as a tropical depression.
  Hurricane Fabian was then located roughly 675 nm east of the Lesser
  Antilles, moving westward at about 12 kts.   During the evening the deep
  convection became even more symmetric around a very tight 10-nm diameter
  eye.  Dvorak satellite intensity estimates from TAFB and AFWA had reached
  6.0 by 31/0000 UTC, so the MSW was upped to 110 kts on the 31/0300 UTC
  advisory.  During the morning of the 31st the coldest cloud tops looked
  a little ragged and the eye became obscured by the cirrus canopy for
  a time, so the intensity remained pegged at 110 kts for 18 hours.  But
  by 2100 UTC the eye had become distinct again in both visible and infra-
  red images, embedded within a perfectly round CDO and surrounded by
  plenty of banding features.  Therefore, the MSW was upped to 115 kts,
  making Fabian a dangerous Category 4 hurricane on the Saffir/Simpson
  scale.  The hurricane was then centered about 435 nm east of the northern
  Leeward Islands, moving west at 10 kts.

     Monday, 1 September, was an exciting day in the life of intense
  Hurricane Fabian.  At 01/0300 UTC the MSW was raised slightly to 120 kts
  based on an improved satellite signature and CI estimates of 127 kts from
  TAFB and SAB, and 140 kts from AFWA.  The eye had cleared out and warmed
  significantly.    Then, the first Air Force Hurricane Hunters
  reconnaissance mission reached the hurricane around 0600 UTC.  Peak winds
  of 133 kts were measured at 700 mb, bolstering confidence in the 120-kt
  surface MSW estimate.  However, the lowest CP measured (949 mb) was not
  quite as low as had been estimated in the 0300 UTC advisory (940 mb).
  Another plane investigated Fabian in the afternoon around 1800 UTC and
  found that the CP had dropped to 942 mb.  Also, peak flight-level winds
  of 140 kts were indicated, corresponding to 125 kts at the surface.  The
  2100 UTC discussion bulletin noted that eyewall convection had become
  colder and more symmetric, and also that the outflow pattern had become
  more symmetric since the morning.  Fabian was now at its peak intensity--
  a large, severe Category 4 hurricane located 240 nm east-northeast of the
  island of Barbuda in the northern Leewards, moving west-northwestward
  at 10 kts.  Hurricane-force winds extended outward 45 nm in all quadrants
  except the southwest, and gales reached outward 120 nm in the northern
  quadrants and 75 nm to the south.

     During 2 September Fabian's track gradually bent more to the right--by
  2100 UTC the cyclone was moving between west-northwest and northwest at
  9 kts.  The intensity remained basically static--the MSW was lowered
  slightly to 120 kts at 0900 UTC based on reconnaissance data.  Satellite
  intensity estimates came down a bit during the day to 102-115 kts, but
  the 1800 UTC reconnaissance plane found peak 700-mb winds of 130 kts,
  which supported 115-120 kts at the surface.  Two NOAA P-3 aircraft found
  105 kts at the surface in the northwest eyewall with the Stepped
  Frequency Microwave Radiometer, but at advisory time had not sampled the
  stronger northeastern quadrant.  The eye diameter had increased to 30 nm
  from 20 nm the previous day.  Late on the 2nd Fabian was making its
  closest approach to the Leeward Islands, being centered at 2100 UTC about
  185 nm north-northeast of Barbuda.   A NOAA P-3 aircraft made an eye
  penetration around 02/2221 UTC and found a 700-mb FLW of 125 kts and a CP
  of 944 mb, so the MSW was reduced slightly to 115 kts at 03/0300 UTC.
  The inner-core convection had also weakened and become more ragged in
  appearance.   Some westerly shear was also beginning to affect Fabian as
  evidenced by a lack of cirrus outflow over the western semicircle.  The
  eye had also become somewhat ragged in satellite images and the Hurricane
  Hunters reported that it had taken on an elliptical shape, although it
  was still surrounded by a ring of deep convection.  At 03/2100 UTC the
  center of Hurricane Fabian was located about 575 nm south of Bermuda,
  moving northwest at 8 kts.  Based on lower T-numbers and findings from
  USAF and NOAA reconnaissance aircraft, the MSW was lowered to 105 kts,
  placing Fabian back in Category 3 status.

     For days Fabian had been moving along the southern periphery of the
  subtropical ridge, and by late on the 3rd was located on the southwestern
  edge of the ridge.  With a large trough over the eastern U. S., recurv-
  ature to the north and then northeast was forecast.  The projected track
  was pointing to Bermuda as a likely target of Fabian, and accordingly
  at 0900 UTC on 4 September, the Bermuda Weather Service issued a hurri-
  cane watch for the island.  At 0900 UTC the hurricane was centered about
  500 nm south of the island and moving north-northwest at 7 kts.   A NOAA
  P-3 aircraft made an eye penetration at 03/2245 UTC and found a 738-mb
  FLW of 127 kts in the north quadrant along with a CP of 939 mb.  Also,
  an eyewall dropsonde reported a wind speed of 142 kts at 28 m above the
  surface.  Based on this information, the MSW was upped to 110 kts in the
  04/0300 UTC advisory.  However, a few hours later data from a Hurricane
  Hunters' mission indicated some weakening, so the MSW was reduced to
  105 kts once more at 0900 UTC.

     During the morning of 4 September Fabian's outflow became better
  established as the upper-level LOW northeast of the Bahamas weakened.
  This LOW had been responsible for the unfavorable shear over the storm
  on the previous day.  The intensity was held at 105 kts during the day,
  but at 0000 UTC on the 5th was upped to 110 kts again.  Two eyewall drops
  between 04/2000 and 2200 UTC had indicated surface winds of 111 and 114
  kts, and the CP was measured at 940 mb.   At 05/0300 UTC Hurricane Fabian
  was centered 260 nm south of Bermuda, and the storm was now moving almost
  due northward at a faster pace of 15 kts.   The storm had continued to
  grow in size, and winds to hurricane force reached outward 100 nm to the
  northeast of the eye and 75 nm to the southeast and northwest while gales
  covered an area over 300 nm in diameter.  A reconnaissance plane in the
  storm around 0600 UTC reported a CP of 948 mb and 110 kts for a peak FLW,
  so the MSW was reduced back to 105 kts at 0900 UTC.  The eye was 40 nm
  in diameter and circular, but as the morning progressed the eye became
  elongated.   The 1200 UTC reconnaissance flight found a peak FLW of 118
  kts and a CP of 951 mb, so it appeared Fabian was maintaining its MSW
  of 105 kts.  During the afternoon the center of severe Hurricane Fabian
  passed just west of Bermuda, bringing the intense east eyewall directly
  over the island.     Observations from Bermuda (see Section C below)
  indicate that the hurricane struck the island as a major Category 3
  hurricane on the Saffir/Simpson scale--the strongest hurricane to
  strike the island since 1926.

     By 0300 UTC on 6 September Fabian was centered about 110 nm north-
  northeast of Bermuda and was moving northeastward at about 17 kts.
  Recent observations from Bermuda indicated winds in the 50-60 kt
  range (made with a hand-held anemometer due to the destruction of the
  normal wind measuring equipment).  By 0900 UTC winds had dropped to
  just below gale force on the island as Fabian continued to pull away
  toward the northeast.  The MSW was dropped to 95 kts at 1500 UTC, but
  even during the afternoon Fabian remained a well-organized hurricane
  with a distinct eye surrounded by deep convection and good outflow.
  The MSW remained at 95 kts in the 2100 UTC advisory, and the forecaster
  noted that this could be conservative.  However, Fabian's intensity
  began to steadily decline on the 7th.   The 07/0900 UTC discussion noted
  that Fabian was beginning to feel the effects of the upper-level
  westerlies with some shearing evident.  Satellite intensity estimates
  were in the 80-90 kt range, so the intensity was adjusted to 85 kts.

     Weakening continued with Fabian crossing the 25 C isotherm during the
  late morning of the 7th.  Central convection was less intense and the
  hurricane displayed a southeast-northwest elongated appearance in
  satellite imagery.  By late afternoon a small core of deep convection
  remained near Fabian, but satellite imagery revealed a frontal tail
  forming on the southeastern side of the cyclone along with cold-air
  stratocumulus clouds wrapping into the southwest quadrant, signifying
  that Fabian was beginning extratropical transition.  The MSW was further
  decreased to 75 kts at 07/2100 UTC, and Fabian was forecast to be fully
  extratropical in 24 hours.  The weakening hurricane was then located
  approximately 250 nm southeast of Cape Race, Newfoundland, and moving
  quickly northeastward at 27 kts over SSTs below 21 C.

     Convection continued to decrease in coverage and intensity early on
  8 September as the low-level circulation gradually became exposed.
  Enhanced infrared imagery and a 08/0008 UTC SSM/I pass depicted an
  asymmetric structure with a very small area of deep convection confined
  to the northwest quadrant.  By 0900 UTC the LLCC had become fully-exposed
  and the MSW was reduced to 65 kts, based on an intensity estimate from
  TAFB.  (AFWA and SAB had estimated the intensity as an extratropical
  system.)   By 1500 UTC satellite images indicated there was no deep
  convection below -50 C near the center of Fabian, and cold-air strato-
  cumulus clouds had infiltrated the core of the cyclone.  TPC/NHC issued
  the final advisory on Fabian at this time, placing the center almost
  600 nm east-northeast of Cape Race and moving northeastward at 34 kts.
  Water temperatures were near 15 C and the storm was rapidly becoming
  fully extratropical.  The final advisory intensity was kept at 65 kts
  due to a late-arriving buoy report of 70-kt winds at 0700 UTC.  The
  extratropical Fabian continued moving northeastward, gradually turning
  northward toward Greenland on the 9th.  After 11/0000 UTC the LOW began
  moving toward the east, crossing southern Iceland on the 12th.  It was
  last noted moving northward once again to the east of Iceland later
  on 12 September.

  C. Meteorological Observations

     Other than the buoy report referenced in the final paragraph in
  Section B, all the surface observations from Hurricane Fabian were
  reported from Bermuda.  At 05/1800 UTC the Bermuda Weather Service
  relayed an observation of a sustained wind of 75 kts and a peak gust of
  87 kts from Bermuda Harbor Radio.   Around 05/2000 UTC the Bermuda Air-
  port estimated sustained winds at 91 kts with a peak gust of 116 kts.
  The highest reported wind was an unofficial report from Bermuda Harbor
  Radio around 05/2100 UTC of a sustained wind of 102 kts and a peak gust
  of 110 kts.  This latter observation was made shortly before the 76 metre
  instrumented mast blew down.   The Bermuda Airport also recorded a SLP
  of 965.5 mb during the afternoon of the 5th.

  D. Damage and Casualties

     Damage in Bermuda was extensive.  A preliminary estimate placed the
  total at around $300 million (US).  The hurricane was also responsible
  for eight deaths:  four fatalities on Bermuda, three fishermen who
  drowned off Newfoundland, and a rip current drowning near Cape Hatteras.

  (Report written by Gary Padgett)

                          TROPICAL STORM GRACE
                         30 August - 2 September

  A. Storm Origins

     A tropical wave left the coast of western Africa during the third
  week of August.  A Tropical Weather Outlook issued by TPC/NHC at 1530 
  UTC on 21 August indicated that an area of disturbed weather associated
  with a tropical wave had developed in the Atlantic midway between Africa
  and the Lesser Antilles.  The system was moving westward at 13 kts and
  conditions were only marginally favorable for further development.  By
  the morning of the 23rd the wave was 825 nm east of the Windward Islands,
  and satellite imagery indicated that showers had increased some since
  the previous day.  Also, upper-level conditions were becoming a little
  more favorable for development.  A Tropical Weather Outlook on the
  morning of 24 August described the system as a large, complex tropical
  wave located about 520 nm east of the Lesser Antilles.  The southern
  portion of the wave had become a little better organized, and convection
  was seen to increase even more during the afternoon.

     Early on the 25th the wave was moving across the Lesser Antilles and
  into the extreme eastern Caribbean Sea.  By afternoon of the 26th the
  system was producing showers and thunderstorms over Hispaniola and
  eastern Cuba.  Surface pressures were low, but upper-level winds were
  not favorable for development.   On 27 August cloudiness and showers
  associated with the wave diminished and the system consisted of a very
  broad area of low pressure extending from the northwestern Caribbean
  across Cuba and into the Bahamas.   A Tropical Weather Outlook from
  TPC/NHC around midday on the 28th noted that cloudiness and showers
  had become more concentrated over the northwestern Caribbean.   By
  late afternoon the area had become better organized with a large area
  of showers spreading over the Cayman Islands, western Cuba and the
  Yucatan Channel.  The system was forecast to move into the Gulf of
  Mexico in a day or two where upper-level conditions would be more
  conducive to tropical cyclone development.

     By 29 August the disturbance had moved into the southern Gulf of
  Mexico with little change in organization as it moved slowly northwest-
  ward.  At 1315 UTC on the 30th, a Special Tropical Disturbance State-
  ment was issued by TPC/NHC.  Surface observations and high resolution
  visible images indicated that a tropical depression appeared to be
  forming in the Gulf.  The first advisory on Tropical Depression 11 was
  issued at 1500 UTC, placing the poorly-defined center about 300 nm east-
  southeast of Corpus Christi, Texas.  TD-11 was moving northwestward at
  10 kts, and the initial intensity was set at 30 kts.  Tropical storm
  warnings were issued from High Island, Texas, to Corpus Christi.

     An intermediate bulletin at 1800 UTC indicated that TD-11 was nearing
  tropical storm strength, and that a reconnaissance plane was approaching
  the system.   At 2100 UTC the depression was upgraded to Tropical Storm
  Grace, located about 245 nm east-southeast of Corpus Christi and moving
  west-northwestward at 8 kts.  The maximum 450-metre FLW reported was
  43 kts, corresponding to about 34 kts at the surface, and there were also
  some reports from ships and oil platforms of tropical storm-force winds
  well to the northeast of the broad center.  The discussion bulletin noted
  that upper-level outflow was good in the eastern semicircle but somewhat
  restricted to the west.

  B. Storm History

     A reconnaissance flight into Grace during the evening hours found a
  very poorly-defined system.  Indeed, as the 31/0300 UTC discussion
  bulletin noted, it was a stretch to call it a tropical cyclone.  Grace
  was more of a north-south oriented trough with pressures 150 nm north
  of the center about the same as near the supposed center.   The plane
  found a FLW of only 40 kts at 450 metres, but a ship reported surface
  winds of 36 kts at 31/0000 UTC about 180 nm northeast of the center.
  The 0000 UTC reconnaissance flight found multiple, poorly-defined
  centers within a larger elongated circulation.  The 0900 UTC advisory
  position, about 100 nm southeast of Corpus Christi, was a compromise
  between aircraft and surface observations as well continuity from the
  previous advisory.  An upper-level LOW in the vicinity of Brownsville
  was inhibiting strengthening of the tropical cyclone.

     Grace was downgraded to a tropical depression at 31/1500 UTC with the
  center of the elongated trough near the Texas coast near Port O'Connor.
  The highest wind observed shortly before advisory time was 30 kts at
  the Sabine Pass CMAN Station.  At 1300 UTC the lowest reported MSLP was
  1008.6 mb near Galveston, but at 1200 UTC Buoy 42019 reported 1008.7 mb.
  The center position was kept south near the buoy, but the advisory noted
  that it could just as well have been placed 100 miles further north near
  Galveston.    The center position was not particularly important since
  the strongest winds and heaviest rains extended well north and east of
  the center.  The 1500 UTC advisory was the final one issued by TPC/NHC.
  Future releases on the weakening remnants of Grace became the
  responsibility of the Hydrometeorological Prediction Center (HPC),
  located at Camp Springs, Maryland.   Tropical Depression Grace moved
  slowly northwest over eastern Texas, later turning toward the northeast.
  By the time of the issuance of the final HPC summary at 0900 UTC on
  2 September, Grace's remnant LOW had merged with a stationary front
  and was located about 120 km west-northwest of Hot Springs, Arkansas.

  C. Rainfall Measurements

     Some selected rainfall amounts resulting from Tropical Storm Grace
  and/or the combination of Grace's remnants and the stationary front

  Station                         Rain (mm)        Time Period (UTC)
  Beaumont/Port Arthur, TX           72        24 hrs ending Aug 31 1800
  Galveston, TX                      69        Total through Sep 01 0000
  Beaumont/Port Arthur, TX          124        Total through Sep 01 0600
  Houston Clover Field, TX          107        Total through Sep 01 0600
  Houston Clover Field, TX           85        24 hrs ending Sep 01 1200
  Houston Hobby, TX                  85        24 hrs ending Sep 01 1800
  Terre Haute, IN                   119        24 hrs ending Sep 02 0000
  Huntsville, TX                     84        30 hrs ending Sep 02 0600
  Jefferson City, MO                 99        30 hrs ending Sep 02 0600
  Indianapolis, IN                  187        30 hrs ending Sep 02 0600
  Lafayette, IN                      92        30 hrs ending Sep 02 0600

     The above rainfall information was obtained from the HPC storm

  D. Damage and Casualties

     No reports of damage or casualties resulting from Tropical Storm Grace
  have been received.

  (Report written by Gary Padgett)


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

  Activity for August:  1 tropical depression
                        2 tropical storms
                        2 hurricanes

                        Sources of Information

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

               Northeast Pacific Tropical Activity for August

     Over the period 1971-2002, August has averaged 4.0 named storms, 2.4
  hurricanes and 1.2 intense hurricanes.  August, 2003, was pretty close
  to normal with 4 named storms, 2 hurricanes and no intense hurricanes.
  Tropical Storm Ignacio reached hurricane intensity on 24 August--the
  latest starting date for the first hurricane of the season since the
  beginning of the modern satellite era.  Tropical Storms Guillermo and
  Hilda formed during the second week of the month well southwest of
  Mexico and were rather insignificant.   Ignacio strengthened into a
  Category 2 hurricane and moved up into the Gulf of California, passing
  east of Cabo San Lucas.  The hurricane grazed the coast near La Paz
  and later moved inland on the peninsula northwest of La Paz, subsequently
  dissipating over the mountainous terrain.  Hurricane Jimena, forming
  far to the west, also became a Category 2 hurricane and was beginning to
  pose a threat to the Big Island of Hawaii, but as it approached the
  island, weakened and turned to the west-southwest.

     The only non-developing depression during the month formed in the
  Honolulu AOR west of 140W.  A disturbance which had been followed for
  several days had acquired enough convective organization to be classified
  as a tropical depression by the 15th.  Advisories were initiated on
  Tropical Depression 01C at 2100 UTC, placing the center approximately
  475 nm southeast of Hilo.  Almost immediately after being classified,
  convection died down rather significantly.  The discussion bulletin at
  16/0900 UTC noted that the depression was a marginal system with an
  elongated shape and several vorticity centers.   By 17/0000 UTC the
  depression had become too weak to be properly classified and was down-
  graded to a disturbance at 0300 UTC when located roughly 350 nm south-
  southwest of Hilo.  An exposed vortex center was visible, but there was
  no evidence of westerly winds on the south side of the system.  Deep
  convection was located far to the east and the vortex was forecast to
  move into an even more hostile environment.

     Additional reports on some of the other Eastern Pacific tropical
  cyclones of 2003 have already been placed on TPC/NHC's website:>

                         TROPICAL STORM GUILLERMO
                               7 - 13 August

  A. Storm Origins

     An area of disturbed weather formed on 4 August about 630 nm south-
  southeast of the southern tip of the Baja California Peninsula.  The
  system was moving west-northwestward, and a Tropical Weather Outlook
  issued by TPC/NHC indicated that there was some potential for
  development over the next couple of days.  The disturbance was better
  organized the next day as it continued moving west-northwestward, and
  by the evening of the 6th had reached a point about 535 nm southwest
  of the tip of the Baja Peninsula.  Convection and upper-level outflow
  had increased significantly during the afternoon, and it was expected
  that a tropical depression would likely form later in the evening or 
  by the next day.

     The first advisory on Tropical Depression 07E was issued at 1500 UTC
  on 7 August, locating the center approximately 525 nm southwest of Cabo
  San Lucas.  Initially, the depression was not forecast to reach tropical
  storm intensity before reaching colder waters.   However, by 2100 UTC 
  TD-07E was better organized and near tropical storm strength.  Outflow 
  was excellent, some banding features were evident, and the center 
  appeared to be near the northwest portion of the deepest convection.  At 
  0300 UTC on 8 August the cyclone was located about 560 nm southwest of 
  Cabo San Lucas, and appeared much better organized than six hours 
  previously.  A small CDO feature had formed over the LLCC, and TAFB and 
  SAB were both rating the system at T2.5+, or about 40 kts.  Therefore, 
  TD-07E was upgraded to Tropical Storm Guillermo with an initial intensity
  of 40 kts.

  B. Storm History

     For the remainder of the 8th Guillermo continued to intensify.  By
  1500 UTC the storm had developed a persistent cold overcast with cloud 
  top temperatures to -80 C near the center.  During the afternoon SSM/I 
  data showed tightly-curved convective bands and a possible eye.  However,
  the cloud pattern as revealed in visible imagery had decayed slightly,
  possibly due to southerly shear caused by an upper-level LOW to the
  west of Guillermo.  Nonetheless, the MSW was upped to 50 kts--the peak
  for the storm's history.   During the early morning of 9 August the
  cloud pattern's appearance was not very impressive.  Infrared imagery
  showed irregularly-shaped bursts of deep convection and the banding
  had become ill-defined.  However, microwave data suggested that the LLCC
  was still embedded in the deep convection, and a QuikScat pass supported
  an intensity of 50 kts; therefore, the official MSW remained at 50 kts.

     By 09/1500 UTC Guillermo's cloud tops had warmed, but the MSW was
  maintained at 50 kts based on CI estimates of 55 kts from TAFB and 45 kts
  from SAB.  Some binary interaction with the large developing TD-08E
  located about 575 nm to the east, which was moving 3-4 kts faster than
  Guillermo, was considered a possibility.   However, by afternoon the 
  convection had become disorganized and ragged-looking, and the outflow 
  pattern had become severely restricted in the eastern semicircle due to
  outflow from TD-08E--soon to become Tropical Storm Hilda.  The MSW was
  decreased to 45 kts, based on CI estimates ranging from 45 kts to 30 kts.
  A burst of deep convection near the center during the evening hours led 
  to the MSW being maintained at 45 kts in the 10/0300 UTC advisory.  Also,
  at 0300 UTC, TD-08E was upgraded to Tropical Storm Hilda.       East-
  southeasterly shear from Hilda's outflow continued to affect Guillermo, 
  and at 1500 UTC the intensity was reduced to 40 kts.  A mid-level vortex 
  in the convective debris could be seen moving away from Guillermo to the

     By afternoon, in addition to the unfavorable effects of Hilda, upper-
  level westerlies from a disturbance located to the southwest began to
  impinge on Guillermo.   The cyclone's intensity was reduced to 35 kts at
  2100 UTC; it was still generating bursts of convection but no banding
  nor any other signs of organization were evident.  At 0300 UTC on the
  11th Guillermo was downgraded to a tropical depression when centered
  about 1175 nm west-southwest of Cabo San Lucas.  The system was largely
  devoid of any central deep convection.  Tropical Storm Hilda was also
  showing signs of weakening, so the likelihood of any interaction between
  the two cyclones was considered slim.  A QuikScat pass at 0300 UTC showed
  winds still very close to tropical storm strength, but shortly afterward
  convection began to decrease, and continued weakening was forecast as
  Guillermo's predicted track would carry it into a cooler and more stable

     Guillermo was maintained as a tropical depression for a couple more
  days as it continued moving westward, mainly due to occasional small
  bursts of deep convection.  The system crossed 140W into Honolulu's
  AOR shortly after 1800 UTC on the 12th, and at 13/0300 UTC the CPHC
  issued the final advisory on Tropical Depression Guillermo, placing
  the convection-free swirl about 800 nm east-southeast of Hilo.

  C. Damage and Casualties

     No damage or casualties are known to have occurred as a result of
  Tropical Storm Guillermo.

  (Report written by Gary Padgett)

                           TROPICAL STORM HILDA
                              9 - 13 August

     The Tropical Weather Outlook issued by TPC/NHC for the Eastern North
  Pacific at 1600 UTC on 7 August noted that a broad area of disturbed
  weather had developed about 460 nm south of Manzanillo, Mexico, and
  was moving westward.  Slow development of the system was considered
  a possibility.   By early on the 8th the disturbance was located about
  475 nm south-southwest of Manzanillo, but had changed little during
  the night.  By 0900 UTC on 9 August a circular-shaped area of strong
  convection had developed near or over the estimated center of
  circulation with banding features present in the southern semicircle.
  Impressive outflow was also observed over the western quadrants of
  the system.  Based on this improved structure, advisories were initiated
  on Tropical Depression 08E at 0900 UTC with an initial intensity of
  25 kts.  The depression was then centered about 600 nm south-southwest
  of Cabo San Lucas on the southern tip of the Baja California Peninsula.

     The MSW was increased slightly to 30 kts at 1500 UTC, but during the
  afternoon inner-core convection became less defined, and microwave and
  conventional satellite imagery suggested that there were multiple
  circulations rotating around a mean center.   By 10/0300 UTC deep
  convection had redeveloped to the south of the center, but the overall
  cloud pattern had improved only slightly.  However, Dvorak T-numbers
  from TAFB and SAB had reached T2.5, so TD-08E was upgraded to Tropical
  Storm Hilda at 0300 UTC, located roughly 575 nm south-southwest of
  Cabo San Lucas.  By 0900 UTC the outflow pattern was looking more
  symmetrical, and vertical shear was forecast to remain light for the
  next few days.  Hilda was moving west-northwestward at 12 kts, but a
  complication to the forecast track was anticipated if the cyclone began
  to undergo an interaction with Tropical Storm Guillermo, then located
  about 10 degrees to the west.

     Hilda remained poorly-organized, and the MSW was never estimated any
  higher than 35 kts.  By 0300 UTC on 11 August the LLCC had become exposed
  on the north side of the deep convection, but the convection to the south
  of the center was strong enough to warrant CI estimates of 35 kts.  By
  0900 UTC the LLCC was located 75 nm to the north of the deep convection.
  Hilda was moving away from its upper-level anticyclone and had become
  isolated from any moist southerly inflow.  Guillermo had by this time
  been downgraded to a tropical depression, and with both systems
  weakening, the possibility of any binary interaction between the two
  cyclones seemed remote.

     QuikScat and SSM/I data indicated that winds were still 35 kts prior
  to 1600 UTC, but the cloud pattern had decayed considerably by 2100 UTC
  with the LLCC completely-exposed well to the northwest of the remaining
  deep convection.  Hilda was downgraded to a tropical depression at this
  time, located approximately 785 nm west-southwest of Cabo San Lucas.
  Like Guillermo, Hilda continued to produce occasional bursts of deep
  convection, usually not near the center, during its decaying phase.
  By 1500 UTC on 13 August Hilda was a large swirl of low clouds with a
  small area of convection well-removed from the center.  Since upper-
  level winds were forecast to remain quite hostile for the next couple
  of days, re-intensification was considered very unlikely.  The final
  advisory from TPC/NHC was issued at 13/1500 UTC, placing the weak 25-kt
  center about 1200 nm west-southwest of Cabo San Lucas.

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

  (Report written by Gary Padgett)

  NOTE: After writing this report on Hilda, I discovered that the official
  TPC/NHC storm report on this cyclone, authored by Lixion Avila, is now
  available on NHC's webiste at the following link:>

                          HURRICANE IGNACIO
                            22 - 27 August

  NOTE:  The official TPC/NHC storm report on Hurricane Ignacio,
  authored by Miles Lawrence, is now available at the following

     Since the official storm report is already available online, there
  is no need of my writing a preliminary report on the system, so my
  comments here will be brief.

     Tropical Depression 09E formed on 22 August southeast of the tip
  of the Baja California Peninsula, was upgraded to Tropical Storm
  Ignacio the next day, and at 0600 UTC on 24 August became the first
  hurricane of the 2003 Eastern North Pacific season.  This represents
  the latest date for the appearance of the first hurricane in this
  basin since the beginning of the modern satellite era in 1966.  The
  only other year in which the first hurricane did not put in its
  appearance until August was in 1968, when Hurricane Fernanda became
  that season's first hurricane on 8 August (according to the Best
  Track file).  Also, 2003 was the first year not to produce a July
  hurricane since 1972.

     Ignacio followed a slow north-northwesterly track which carried
  it just east of Cabo San Lucas on the tip of the Baja and later inland
  on the Peninsula just north of La Paz.  Heavy rainfall, enhanced by
  the storm's slow movement, was the primary destructive agent in
  association with Hurricane Ignacio.  Miles' report indicates that
  two persons were drowned as a result of the hurricane.

     For more information on Ignacio, please see the report at the
  link given above.

                           HURRICANE JIMENA
                       28 August - 5 September

  A. Storm Origins

     A Tropical Weather Outlook issued by TPC/NHC at 1100 UTC on 27 August
  mentioned a broad area of disturbed weather located roughly 1130 nm
  southwest of the southern tip of the Baja California Peninsula (Cabo
  San Lucas).  The disturbed weather was poorly organized but there was
  some potential for further development.  By very early on the 28th the
  system was located about 1215 nm west-southwest of Cabo San Lucas and
  had become much better organized.   The first advisory on Tropical
  Depression 10E was issued at 0900 UTC, locating the center approximately
  1250 nm west-southwest of Cabo San Lucas and moving westward at 10 kts.
  Convection was concentrated near the center and outflow was good in all
  directions.   Six hours later satellite classifications had reached a
  consensus of T2.5, or 35 kts, but since a TRMM microwave pass at 28/0841
  UTC had suggested that the LLCC might be displaced north of the center
  position inferred from infrared imagery, the forecaster opted to hold
  off upgrading the cyclone for the moment.

  B. Storm History

     By 2100 UTC, however, CI estimates from TAFB and SAB had reached 45
  kts, so TD-10E was upgraded to Tropical Storm Jimena.  The new tropical
  storm was then located about 1350 nm west-southwest of Cabo San Lucas,
  or roughly 1375 nm east-southeast of Hilo, Hawaii.   Jimena continued
  to intensity steadily as evidenced by increased banding of cold
  convection and the occasional appearance of an eye-like feature.  At
  29/0000 UTC satellite intensity estimates were 55 kts from TAFB, 45 kts
  from SAB and Honolulu, and 35 kts from AFWA; hence, at 0300 UTC the
  MSW was jumped up to 50 kts.  A 29/0556 UTC SSM/I pass revealed a well-
  defined eye under the concentrated cold convection, so the MSW was upped
  again to 55 kts at 0900 UTC.  Six hours later, Jimena was upgraded to the
  second hurricane of the 2003 season based on CI estimates ranging from
  55 kts to 77 kts, plus a 29/0932 UTC TRMM pass which had revealed a
  small distinct eye embedded in the deep convection.  Newly-upgraded
  Hurricane Jimena was located about 1125 nm east-southeast of Hilo, moving
  west-northwestward at 14 kts.     The west-northwesterly motion was
  attributed to a weakness in the subtropical ridge induced by a mid to
  upper-level trough north of Hawaii.  This ridge was expected to lift out
  and allow the ridge to become re-established, leading to a more westward
  track for Jimena in about 24 hours.

     The last few visible images on 29 August showed a very tight eye only
  10 nm in diameter which was partially cloud-filled.  Based on Dvorak
  intensity estimates of 90 kts from TAFB and AFWA, Jimena's intensity was
  upped to 85 kts at 30/0300 UTC.  Since the hurricane was moving along the
  26 C SST isotherm and would soon be entering cooler waters, it was felt
  that the intensity would level off for 24 hours or so with a gradual
  weakening beginning thereafter.   The final TPC/NHC advisory issued at
  0900 UTC maintained the 85-kt MSW, based on intensity estimates ranging
  from 77 kts to 102 kts.   Since the hurricane's center had moved west
  of 140W, the Central Pacific Hurricane Center (CPHC) located in Honolulu
  assumed responsibility for issuing advisories on Jimena.   In spite of
  its passage over SSTs slightly cooler than 26 C, persistent Dvorak
  estimates of 5.0 from Honolulu and even higher AFWA estimates prompted
  CPHC to raise the MSW to 90 kts at 2100 UTC--the peak for the storm's
  history.  (The minimum estimated CP for Jimena was 970 mb.)  At the
  time the center of Jimena's well-defined eye was located approximately
  735 nm east-southeast of South Point on the Big Island of Hawaii, moving
  almost due westward at 17 kts.   An immediate slow weakening was fore-
  cast, yet Jimena maintained the 90-kt peak intensity for 24 hours.

     With the hurricane's persistent rather rapid westerly motion in the
  direction of Hawaii, a hurricane watch was issued for the Big Island in
  a special advisory at 31/0000 UTC, followed by a tropical storm warning
  for the island three hours later.  As the 31st progressed Jimena began
  to encounter some slight south-southwesterly shear which caused the storm
  to gradually show signs of weakening.   The MSW remained at 90 kts
  through the 1500 UTC advisory based on the higher of the available Dvorak
  estimates, but at 2100 UTC the intensity was reduced to 85 kts.  The
  hurricane's eye was not apparent in visible imagery, and the system was
  becoming somewhat elongated.  The weakening trend became more pronounced
  on 1 September with Jimena being downgraded to a 60-kt tropical storm at
  1500 UTC when located about 150 nm south of Hilo.  The highest FLW found
  by a reconnaissance aircraft was 55 kts.  The hurricane watch for the
  Big Island was also cancelled at this time.  Jimena's track had also
  taken a turn to the west-southwest, meaning that the likelihood of the
  storm's striking the Big Island was further reduced.    The closest
  approach of Jimena to the island occurred around 01/1200 UTC when the
  center was located about 100 nm southeast of South Point.

     Jimena quickly fell apart after passing south of the Big Island--by
  1500 UTC on 2 September the system had weakened to minimal tropical storm
  intensity of 35 kts.  Sporadic bursts of convection were occasionally
  seen, but the general trend was that of weakening.  By 0900 UTC on the
  3rd Dvorak T-numbers were 2.0 or lower from all the applicable agencies,
  so Jimena was reduced to tropical depression status.  The depression's
  center was placed approximately 300 nm east-southeast of Johnston Atoll.
  The MSW was decreased further to 25 kts at 03/1500 UTC, where it remained
  for 24 hours.  Jimena continued to maintain a good peripheral cloud
  circulation and some convection near the center.  As the depression moved
  farther west, shear decreased and Jimena began to show signs of life once
  more.  The MSW was boosted slightly to 30 kts at 04/1500 UTC as the
  system exhibited some intense deep convection in the central core region.

     With most Dvorak classifications at 2.5 at 1200 UTC, Jimena was
  upgraded once again to minimal tropical storm status at 04/2100 UTC when
  located about 375 nm west-southwest of Johnston Atoll.  The center was
  relocated somewhat further to the south of the previous advisory's
  position.  Most of the intensity models were rather enthusiastic for
  some continued modest strengthening over the next 48 hours, but this was
  not to be.   By 05/0300 UTC Jimena's cloud and convective structure had
  become quite ragged, and the convection had weakened with cloud tops
  10 C warmer than six hours previously.   All Dvorak classifiers rendered
  a final T-number of 2.0, but due to constraints, Jimena was retained as
  a tropical storm for one more advisory cycle.  A developing upper-level
  LOW was inducing increased shear over the tropical cyclone.  By 0900 UTC 
  the weakening cyclone was moving across the International Dateline and 
  CPHC downgraded it back to a 30-kt tropical depression and issued their 
  final warning.

     JTWC issued the final two warnings on Jimena.  The MSW was further
  lowered to 25 kts at 05/1200 UTC.  A 05/0914 UTC SSM/I pass showed that
  the LLCC was completely exposed with convection occurring far to the
  south of the circulation center.    A 05/1727 UTC TRMM pass indicated
  that the LLCC had nearly dissipated.  The final JTWC warning was issued
  at 1800 UTC, placing the dissipating center about 500 nm northeast of
  Majuro Atoll in the eastern Marshall Islands, or approximately 715 nm
  east-southeast of Wake Island.

  C. Damage and Casualties

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

  (Report written by Gary Padgett)


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

  Activity for August:  4 tropical depressions **
                        1 tropical storm
                        4 typhoons ++

  ** - one classified as a tropical depression by some of the Asian TCWCs
       but not by JTWC, and as a tropical storm by PAGASA; another
       referred to as a tropical depression only in a CPHC satellite
       bulletin; two others classified as depressions by CWBT (one of
       these by JMA also)

  ++ - one of these classified as a typhoon by JTWC only

                        Sources of Information

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

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

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

     The reports on Typhoon Morakot/Juaning, Typhoon Krovanh/Nina,
  Tropical Storm Vamco/Manang, and Typhoon Dujuan/Onyok were written
  by Kevin Boyle.  Significant portions of these reports and the report
  on Typhoon Etau/Kabayan, as well as most of the information on Tropical
  Depression Lakay, were based on information supplied by Huang Chunliang.
  A special thanks to Kevin and Chunliang for their contributions.

               Northwest Pacific Tropical Activity for August

     August was a very active month in the Northwest Pacific basin.  Five
  tropical cyclones were named by JMA--four of these became typhoons per
  JTWC's analysis.  Another system was named by PAGASA and was upgraded to
  a minimal tropical storm by that agency, although no other warning center
  did so.   Typhoon Morakot formed early in the month east of Luzon and
  followed a northwesterly track which carried it across southern Taiwan
  and into China.  JTWC was the only agency which classified Morakot as a
  typhoon.  NMCC upgraded it to severe tropical storm status (winds > 47
  kts), but none of the other TCWCs estimated the MSW any higher than 45
  kts.  Also during the first week of August, much-stronger Typhoon Etau
  formed well to the east of the southern Philippines and followed a
  typical recurving track to the east of Taiwan which carried it over
  Okinawa and later over Japan, first striking Shikoku and later Honshu.

     Typhoon Krovanh, known as Nina ("NEEN-yah") in the Philippines, took
  shape very slowly around mid-month in the Philippine Sea.   Once the
  depression finally developed into a tropical storm, it followed a some-
  what unusual west-southwesterly track before turning westward across
  northern Luzon as a typhoon of moderate intensity.  After emerging into
  the South China Sea, Krovanh turned to the west-northwest and regained
  intensity, passing just off northeastern Hainan Dao before striking
  the Leizhou Peninsula.  Krovanh emerged into the Gulf of Tonkin and
  made a final landfall in extreme northern Vietnam.   Minor Tropical
  Storm Vamco formed east of Taiwan and moved northwestward into China,
  passing just off the northern tip of Taiwan.

     Finally, late in the month, Typhoon Dujuan (named Onyok by PAGASA)
  formed west of Guam and followed a fairly straight west-northwesterly
  track which carried it through the Luzon Strait south of Taiwan and
  into southern China near Hong Kong.   Dujuan became a strong typhoon
  and neared super typhoon intensity (per JTWC) as it passed just off
  Taiwan's southern tip.

     Another South China Sea system was classified as a tropical depression
  by several of the Asian TCWCs--though not by JTWC--and was assigned the
  name Lakay by PAGASA.   Since Lakay was treated as a minimal tropical
  storm by PAGASA, and since it caused some enhanced rainfall in China
  as it dissipated, a short report will be included on this system.

     Huang Chunliang has advised me of two additional tropical disturbances
  which were treated as weak tropical depressions by some of the Asian
  TCWCs.  A disturbance which took shape during the final two days of
  July several hundred miles east-northeast of Saipan was carried as a
  weak tropical depression (MSW of 25 kts or less implied) by JMA in the
  summary portion of that agency's High Seas Forecasts.  JTWC gave this
  system a fair potential for development, but dissipated it in their STWO
  on 2 August.  The CWB of Taiwan also upgraded this LOW to depression
  status.  Another disturbance located to the west of the pre-Etau system
  was briefly classified as a tropical depression by CWBT on 2 August.
  Apparently, at one point this system and the one which ultimately became
  Etau were competing for inflow, but the eastern system won and developed
  into Typhoon Etau.

                            TYPHOON MORAKOT
                     (TC-10W / TS 0309 / JUANING)
                              1 - 5 August

  Morakot: contributed by Thailand, means 'emerald'

  Juaning: PAGASA name, is a Filipino male nickname

  A. Storm Origins

     Typhoon Morakot began as an area of convection near 8.5N, 150.0E,
  or approximately 130 nm northwest of Chuuk, and was noted in a STWO
  issued at 0600 UTC on 30 July.  Animated multi-spectral satellite
  imagery revealed a broad region of cycling convection.  Upper-level
  analysis indicated marginal wind shear conditions over the area such
  that the potential for development within the next 24 hours was regarded
  as poor.  This was upgraded to fair at 30/0800 UTC when multi-spectral
  satellite imagery showed convection beginning to wrap and consolidate
  around a well-defined LLCC.  Numerous low-level cloud lines were also
  observed.   Upper-level analysis depicted a slightly more favourable
  environment, i.e., weak vertical shear and moderate diffluence aloft.
  The fair development potential was maintained in STWOs through the rest
  of 30 July and into the 31st.  Based on improving organisation of the
  deep convection, which had increased and consolidated about the LLCC,
  the potential for development was upgraded to good at 31/1400 UTC and
  accordingly a TCFA issued.  At this time, the disturbance was located
  approximately 275 nm east of Luzon, Philippines (near 15.5N, 127.2E).

     The 1st day of August saw several TCWCs opening their account in
  quick succession on this system, beginning with PAGASA issuing their
  initial warning on Tropical Depression Juaning at 01/0600 UTC.  JTWC
  followed suit at 01/1200 UTC, and NMCC at 1800 UTC.  Six hours later,
  at 02/0000 UTC, JMA started writing bulletins, HKO at 02/0600 UTC, and
  finally CWB of Taiwan at 0900 UTC.  (Note: Both HKO and CWB upgraded to
  tropical storm status at 02/0600 UTC.)     Tropical Depression 10W's 
  initial position was approximately 175 nm east of Luzon, Philippines 
  (near 16.6N, 125.5E), based on JTWC's first warning, with the storm 
  moving slowly north-northwestward at 3 kts.  The Prognostic Reasoning 
  Message issued at 01/1200 UTC indicated that TD-10W was forecast to 
  continue north-northwestward during the near term toward a weakness 
  caused by a longwave trough over Japan, then turn more westward towards 
  China as the ridge rebuilt behind the trough.    The depression was 
  embedded in an area of shearing easterlies, but these were not too 
  strong and some slow intensification was forecast.

  B. Storm History

     At 02/0000 UTC JTWC upgraded the system to a minimal tropical storm,
  i.e. 35 knots.  TS-10W was named Morakot at 02/0600 UTC when JMA raised
  the MSW to 35 kts (10-min avg).  By 02/0600 UTC all warning agencies
  were classifying the system as a tropical storm.     Tropical Storm
  Morakot was then moving northwestward at 7 to 8 kts, accelerating to
  14 kts late in the day while continuing to strengthen.  The MSW was
  upped to 45 kts at 02/0600 UTC.  (JMA's 10-min avg at this time was also
  45 kts, and this turned out to be their peak intensity.)  Multi-spectral
  and water vapour imagery showed good equatorward outflow aloft and
  increasing amounts of cycling deep convection in the northern semi-
     By 0000 UTC on 3 August Tropical Storm Morakot/Juaning had moved
  northwestward to a position approximately 225 nm south of Taipei, Taiwan
  (near 20.8N, 122.0E).  The MSW was still 45 kts at this time.  However,
  animated multi-spectral satellite imagery at 03/0600 UTC revealed the
  formation of a large, ragged eye.  Based on this and CI estimates of 55
  to 65 kts, Morakot was upgraded to a typhoon.  (JTWC was the only agency
  to rank Morakot as a typhoon.  All the Asian warning agencies except
  NMCC regarded Morakot as a 45-kt system.  NMCC's peak MSW was 50 kts:  a
  severe tropical storm.   All the intensity estimates from the Asian
  centres represent a 10-min average.)  Events began to unfold quite
  quickly late on the 3rd and into the 4th.  After a brief jog toward the
  west-northwest, Typhoon Morakot turned back to the north-northwest and
  made its first landfall near Taitung, Taiwan, at 03/1500 UTC.  Prior to
  landfall on the southern tip of Taiwan, a SSM/I pass at 03/0952 UTC
  depicted a symmetrical 45-nm diameter eye. 

     At 0000 UTC on 4 August Typhoon Morakot was located approximately
  145 nm southwest of Taipei, Taiwan (near 23.0N, 120.0E), and had
  decelerated quite markedly to 6 kts as it turned to the west.   (This
  sudden change in heading to the west and sudden deceleration may possibly
  have been due to the mountainous terrain of Taiwan.)  The system had by
  then moved beyond PAGASA's boundaries and the last warning was issued by
  that agency.  Animated water vapor imagery (and a 03/2223 UTC SSM/I pass)
  indicated that dry air entrainment was evident in the northwestern
  quadrant.  The MSW was still estimated at 65 kts at 04/0600 UTC when the
  typhoon was about to make landfall in the vicinity of Xiamen City, China,
  moving northwestward at a faster pace of 14 kts.  The 04/1200 UTC warning
  (#13) indicated that the storm had made its second landfall to the north-
  east of Xiamen, China, and was beginning to weaken rapidly overland.  At
  this time Morakot was downgraded to a tropical storm.  The weakening
  tropical cyclone continued to move northwestward and further inland, and
  the final warning was issued at 05/0000 UTC.  (NMCC and CWB issued their
  final bulletins at 04/1500 UTC, and HKO likewise ended warning coverage
  three hours later.  JMA had ceased writing statements even earlier at 
  04/0900 UTC.)

  C. Meteorological Observations

     All the observations in this section were obtained from a report sent
  by Huang Chunliang of Fuzhou City, Fujian Province, China.  A special
  thanks to Chunliang for compiling and sending the information.

  (1) Landfalls

     According to the CWB warnings, Tropical Storm Morakot made its first
  landfall near Dawu Town, Taitung County, Taiwan, around 03/1350 UTC with
  the MSW estimated at 45 kts and a pressure of 990 hPa.  Afterward, the
  storm entered the waters of Taiwan Strait from Jiangjyun Town, Tainan
  County, around 03/2030 UTC.

     According to the NMCC warnings, Tropical Storm Morakot made a second
  landfall near Weitou Town, Jinjiang City (a sub-city of Quanzhou City),
  Fujian Province, around 04/1130 UTC with a MSW of 35 kts and a pressure
  of 998 hPa.

  (2) Rainfall Observations

  Province/        Station                     Period           Rainfall
   Region                                       (UTC)             (mm)

  Taiwan      Ta-chi Mt, Taitung County     02/1600-03/1500       150
  Taiwan      Lanyu, Taitung County         02/1600-03/1500       140
  Taiwan      Chih Pen, Taitung County      02/1600-04/1500       653
  Taiwan      Che Chen, Taitung County      02/1600-04/1500       379
  Taiwan      Mu Cha, Hualien County        02/1600-04/1500       296
  Taiwan      Hengchun, Pingtung County     02/1600-04/1500       242
  Taiwan      Zuojhen, Tainan County        02/1600-04/1500       196
  Taiwan      Zuoying, Kaohsiung City       02/1600-04/1500       194
  Fujian      Nan'an, Quanzhou City         04/0000-05/0000       252
  Fujian      Nan'an, Quanzhou City         04/0000-05/0600       381.2
  Fujian    * Jinjiang, Quanzhou City       04/0000-05/0000       254
  Fujian    * Jinjiang, Quanzhou City       04/0000-05/0600       344.2
  Fujian      Licheng, Quanzhou City        04/0000-05/0600       244.4
  Fujian      Shishi, Quanzhou City         04/0000-05/0600       251.1
  Fujian      Anxi, Quanzhou City           04/0000-05/0600       229
  Fujian      Yongchun, Quanzhou City       04/0000-05/0600       186
  Fujian      Dehua, Quanzhou City          04/0000-05/0600       161
  Zhejiang    Wencheng                      03/2100-05/0900       103
  Guangdong   Shantou                       04/0000-05/0000       102.7

  Note (*): Rains in the range of 233 to 544 mm fell on Jinjiang City,
  where Morakot made its second landfall, during the 18-hr period ending
  at 05/1000 UTC.   This was the most torrential rainfall recorded since
  meteorological records began in the city.  Zimao Town, Jinjiang City,
  reported an 8-hr total of 334 mm that day.

  (3) Wind Observations

  (a) Taiwan Region

       Station                           Peak Gusts

  Lanyu, Taitung County               Beaufort Force 14
  Dawu, Taitung County                Beaufort Force 11
  Hengchun, Pingtung County           Beaufort Force 11
  Taitung City                        Beaufort Force 10
  Tainan                              Beaufort Force  9

  (b) Fujian Province

     The coastal county of Pingtan reported the maximum gusts among all the
  stations in Fuzhou.  A peak gust up to Beaufort Force 8 was recorded at
  that station at 03/1921 UTC.

  (c) Guangdong Province

     Affected by the peripheral convection, many places in eastern
  Guangdong and the region of the Zhujiang River Delta also received some
  rain and gusty winds from the thunderstorms triggered by Morakot.  An
  automatic weather station located in Guanyao Water Conservancy Bureau,
  Nanhai District, Foshan City, reported a peak gust of 68 kts at 04/1342
  UTC.  Stations Yantian and Longgang, both located in Shenzhen City,
  reported peak gusts to 49 kts and 45 kts, respectively.  Guangzhou City,
  the provincial capital, also recorded 33.4 mm of rains with a peak gust
  to Beaufort Force 8.

  D. Damage and Casualties

     A total of 99 towns of Quanzhou City, Fujian Province, where Morakot
  made its second landfall, suffered from the storm, which caused 240
  million yuan of losses to the city.   At least one death and another
  missing person were reported there.

     In several districts of Heyuan City, Guangdong Province, which were
  affected by the periphery of the storm, power was cut off due to the
  torrential rain, which lasted about 27 minutes around 04/0900 UTC.   In
  addition, two workers were reported to have been blown down from a
  45-meter high falsework and lost their lives.

  E. Artificial Rain

     The following paragraphs appear exactly as sent by Huang Chunliang,
  with some very minor editorial changes.

     "To stimulate more rains artificially from the mouth of a storm?
  Sounds crazy, but it was definitely true in the provinces of Fujian,
  Zhejiang and Jiangxi during the storm.

     "Many southern provinces of China are suffering from the most severe
  drought in decades this summer.  Take Fuzhou for example.  Little rain
  had fallen in the City for more than a month before the storm!  What's
  worse, the long running heatwave also hit us with a peak temperature of
  41.7 C being recorded in the afternoon of 26 July.  (According to the
  database, the urban area of Fuzhou had not recorded a temperature
  exceeding 40 C since meteorological records began 120 years ago until
  15 July, when a record value of 41.1 C was reported.)     No wonder
  Morakot was called a "long-awaited storm" here.   Thanks to the storm,
  we were brought a breath of fresh air in the grip of drought, though
  the relief was fleeting.

     "We had no choice but to fetch more rain on our own initiative from
  the mouth of Morakot.   So several missions were operated by the local
  meteorological services of Fujian, Zhejiang and Jiangxi to trigger more
  rainfall to the dry earth.   As a result, the drought was effectively
  relaxed in these provinces.

     "After Morakot's landfall in Fujian, 703 artificial rainfall missions
  were carried out in 120 cities/counties of the province with 1027 rockets
  and 14,700 cannonballs carrying silver iodide cannisters being fired (as
  of the 6th).  As a result, 1350 towns received rainfall from the missions
  which brought rainfalls in the 40-60 mm range over an area totaling
  138,500 square kilometres in the province.  The highest amounts measured
  in Zhejiang Province were 103 mm in Wencheng, 86 mm in Yongjia, and 61 mm
  in Yuhuan.  The artificial rains in Zhejiang were triggered by silver
  iodide dispensed from aircraft rather than by the firing of cannons and

     "All in all, what I want to emphasize in this part is that Morakot
  turned out to be more of a beneficial tropical cyclone than a baleful
  one, even for the city of Quanzhou, which 'suffered most badly' from
  the storm."

  (Report written by Kevin Boyle and Huang Chunliang)

                               TYPHOON ETAU
                       (TC-11W / TY 0310 / KABAYAN)
                              3 - 11 August

  Etau: contributed by the United States, is a Palauan word meaning
        'storm cloud'

  Kabayan: PAGASA name, is a Filipino term meaning 'same citizenship'

  A. Storm Origins

     The precursor to the large, severe Typhoon Etau was an area of
  convection which developed on 31 July approximately 60 nm north-
  northwest of Chuuk.  An interim STWO issued by JTWC at 31/2300 UTC
  noted that the convection had persisted for 12 hours, and that various
  satellite sensors indicated the presence of a weak LLCC to the east
  of the deepest convection.  An upper-level analysis depicted weak
  diffluence and weak to moderate vertical shear over the region.  The
  potential for development was upgraded to fair at 0600 UTC on 1 August.
  A recent QuikScat pass depicted the weak LLCC to the east of the deep
  convection, and vertical shear had lessened somewhat.  By 1400 UTC the
  disturbance was located around 400 nm east of Yap.  The convective
  organization had continued to improve with much energy being absorbed
  from a weaker LLCC to the west.  (This secondary system was briefly
  classified as a tropical depression by the CWB of Taiwan.)

     At 01/2300 UTC JTWC issued a TCFA for the system, which was by then
  located approximately 375 nm east of Yap.   The convection had become
  better organized, and an upper-level analysis indicated favorable
  outflow on the equatorward side.   A QuikScat pass early on the 2nd
  depicted a broad and elongated LLCC, but the CIMSS relative vorticity
  product showed that the low-level vorticity was increasing.   A second
  TCFA was issued at 2200, followed by the first JTWC warning on Tropical
  Depression 11W at 0000 UTC on 3 August.  The warning placed the center
  of the depression roughly 300 nm west-southwest of Guam, moving north-
  westward at 11 kts.  The MSW was estimated at 25 kts, and the system
  displayed improving organization with the convection consolidating near
  the LLCC.  TD-11W was tracking northwestward around the southwestern
  periphery of a low to mid-level ridge extending westward from the
  Marianas.  The cyclone was forecast to track toward a weakness in the
  ridge south of the Ryukyu Islands.

  B. Storm History

     The depression's organization continued to improve, and at 03/0600 UTC
  JTWC, NMCC and JMA all upgraded to tropical storm status with JMA naming
  the system Tropical Storm Etau.   Etau was located about 325 nm west of
  Guam, moving northwestward at 11 kts.  The system exhibited good outflow
  to the south.   By 04/0000 UTC the cyclone had reached a position about
  510 nm west of Saipan.  Animated multi-spectral imagery and a 03/2303 UTC
  SSM/I image depicted the LLCC just to the north of the deep convection.
  At 0600 UTC animated enhanced water vapor imagery indicated that Etau had
  linked up with a TUTT cell to the northeast with enhanced poleward
  outflow a good possibility for the next 12-24 hours.  The MSW by this
  time had increased to 55 kts as Tropical Storm Etau continued tracking
  northwestward along the southern periphery of a mid-level ridge located
  to the north over Japan.

     By 04/1800 UTC satellite CI estimates were in the 65-90 kt range.  At
  this time all three warning centers upgraded Etau to typhoon status,
  located approximately 600 nm south-southeast of Naha, Okinawa.  The storm
  was plodding northwestward at a slightly slower speed of 8 kts.   Early
  on 5 August enhanced outflow to the east-northeast and equatorward was
  noted, and convection was increasing in organization.     By 1800 UTC
  Typhoon Etau had reached a position 350 nm south-southeast of Naha.
  Winds had reached 90 kts, and the system appeared well-organized with a
  banding eye feature evident.   Motion was still northwestward, but the
  typhoon was forecast to turn more poleward as the steering ridge weakened
  with the approach of a shortwave trough currently located over Asia.

     At 0000 UTC on 6 August Etau was located a little less than 300 nm
  south-southeast of Naha and had turned to a north-northwesterly heading
  at 13 kts.  The MSW was upped to 95 kts based on CI estimates of 90 kts
  and an improving trend noted in microwave imagery.  Banding features were
  becoming more tightly wrapped--indicative of continued intensification.
  Twelve hours later (1200 UTC) Etau was located 140 nm south-southeast of
  Okinawa and was by now moving northward along the western periphery of
  the mid-level ridge to the east-northeast.  The typhoon was passing the
  ridge axis and was forecast to begin recurving toward Japan within the
  next 12-18 hours.  By 1800 UTC the storm was centered only about 65 nm
  south-southeast of Naha, and peak winds had increased to 100 kts around
  a 35-nm diameter eye.

     A portion of Etau's large eye was over the island of Okinawa at 0000
  UTC on 7 August with the storm moving northward at 12 kts.  The MSW was
  increased to its peak of 110 kts based on CI estimates of 102 kts.  The
  storm was also at its peak intensity per JMA's and NMCC's warnings at
  this time:  100 kts per NMCC and 80 kts per JMA (both representing 10-min
  averages).  The minimum CP estimated by JMA was 945 mb.  A QuikScat pass
  at 06/2109 UTC revealed that the 35-kt and 50-kt wind radii were almost
  twice as large as had been indicated in the previous warning.  Etau was a
  large severe typhoon with gales covering an area about 470 nm in diameter
  and storm-force winds extending over a zone 170 nm in diameter.  By 0600
  UTC the typhoon was moving north-northeastward at 10 kts from a position
  75 nm northeast of Kadena AB, Okinawa.   It had by now linked up with a
  shortwave trough approaching from Korea and had begun to recurve along
  the Ryukyu Island chain.   At 1800 UTC Typhoon Etau was located 215 nm
  south-southeast of Sasebo, Japan, moving northeastward at 16 kts. 
  Animated water vapor imagery indicated that the system was experiencing
  northwesterly shear.

     By 0600 UTC on 8 August the slowly-weakening typhoon was centered
  about 140 nm south-southeast of Iwakuni, Japan, with the MSW estimated
  at 90 kts.    A 08/0600 UTC synoptic observation at Murotomisaki on the
  eastern shoreline of the island of Shikoku reported sustained winds of
  82 kts (10-min avg) for two consecutive hours.  Typhoon Etau made land-
  fall on Shikoku around 1200 UTC about 15 nm west-northwest of Muroto-
  misaki.  The cape experienced sustained typhoon-force winds for eight
  hours from 0400-1100 UTC with a minimum hourly SLP of 951.7 mb.   The 
  storm made landfall on Honshu near the Osaka International Airport around
  08/1900 UTC.  Tokushima reported a surface pressure of 964 mb at 1620
  UTC.   The lowest SLP reported at the Osaka Airport was 973 mb.   At
  0000 UTC on 9 August Etau was inland about 330 km west of Tokyo with the
  MSW down to 65 kts.   Animated multi-spectral imagery depicted strati-
  form clouds in the western semicircle with bands of deep convection east
  of the LLCC.

     JTWC downgraded Etau to a tropical storm at 09/0600 UTC--the storm was
  then located approximately 165 km northwest of Tokyo, moving northeast-
  ward at 20 kts.  (JMA had downgraded Etau six hours earlier.)  The storm
  was beginning to transform into an extratropical cyclone as its forward
  speed increased--at 1200 UTC it was centered over land 145 km south-
  southwest of Misawa, racing northeastward at 31 kts.   JTWC and NMCC
  issued their final warnings on Etau at 1800 UTC with JTWC declaring the 
  cyclone extratropical.  The storm had moved back over the Pacific and was
  located about 90 nm (165 km) east-northeast of Misawa.  Animated water
  vapor imagery showed that Etau had completed transition to an extra-
  tropical cold-core system.  JMA followed suit and classified the storm
  extratropical six hours later.   The remnants of Etau continued moving
  northeastward, turning to the north on 11 August.  At 1200 UTC on the
  11th the system was a stationary 35-kt gale in the western Bering Sea
  just east of the Kamchatka Peninsula.

  C. Meteorological Observations

  (1) Landfalls

     Based on JMA warnings, there were five distinct landfalls of Typhoon

     (a) Northern Okinawa Island, Okinawa Prefecture, around 07/0100 UTC
     (b) Amami Oshima, Kagoshima Prefecture, around 07/1200 UTC
     (c) Near Muroto City, Kochi Prefecture, around 08/1230 UTC with
         a MSW (10-min avg) of 78 kts and a CP of 950 hPa
     (d) Near Nishinomiya City, Hyogo Prefecture, around 08/2100 UTC
     (e) Near Cape Erima, Hokkaido, around 09/1700 UTC

  (2) Rainfall

     The following table contains the top ten storm totals for the
  96-hour period 06/1500 through 10/1500 UTC.

  Ranking   Prefecture       Station                Rainfall (mm)
  01        Tokushima        Asahimaru                  683
  02        Kochi            Yanase                     639
  03        Miyazaki         Mikado                     539
  04        Tokushima        Fukuharaasahi              533
  05        Tokushima        Kitou                      507
  06        Kochi            Funato                     500
  07        Mie              Miyagawa                   494
  08        Nara             Mt. Hidegatake             483
  09        Nara             Kamikitayama               476
  10        Miyazaki         Nakagoya                   446

  (3) Peak Wind Observations

  Prefect.   Station            MSW/Dir    Date/   Pk Gust/Dir   Date/
                                (kts)     Time UTC   (kts)      Time UTC
  Okinawa    Minami-daito Is.   40.8/S    07/0520    68.0/SSE   07/0519
  Okinawa    Nago               58.1/W    07/0320    94.7/W     07/0310
  Okinawa    Naha               53.3/WNW  07/0020    98.0/N     06/2044
  Okinawa    Kumejima           33.6/WNW  07/0050    70.6/NW    07/0147
  Kagoshima  Okinoerabu         64.0/SE   07/0410    94.1/ESE   07/0250
  Kagoshima  Naze               29.9/ESE  07/1000    78.1/E     07/0922
  Kagoshima  Tanegashima        38.9/NNW  07/2250    78.9/E     07/1803
  Kagoshima  Yakushima          42.2/E    07/1700    74.4/WNW   07/2302
  Kagoshima  Makurazaki         34.2/NNW  08/0200    66.3/NNW   08/0154
  Miyazaki   Aburatsu           54.8/ESE  07/2010    88.1/ESE   07/1950
  Kochi      Murotomisaki       96.8/SW   08/1340   134.5/WSW   08/1329
  Kochi      Shimizu            38.9/E    08/0330    86.5/E     08/0333
  Ehime      Uwajima            31.7/W    08/1600    66.7/SE    08/0146
  Tokushima  Tokushima          33.6/SE   08/1540    70.0/SE    08/0415
  Wakayama   Shionomisaki       42.0/SSE  08/1240    81.6/SSE   08/1056
  Wakayama   Wakayama           42.4/SSW  07/1930    72.9/SW    07/1920
  Mie        Yokkaichi          18.5/ESE  08/1250    71.5/SE    08/1323
  Mie        Owase              29.9/S    08/1600    72.9/S     08/1600
  Mie        Tsu                40.6/SSE  08/1420    68.0/SE    08/1338

     Only stations which either reported sustained winds (10-min avg)
  of gale force or greater, or else gusts greater than typhoon force, are
  included in the above table.

     Karl Hoarau sent me some information which included three stations
  that recorded SLPs in the 950-952 hPa range while in the eye of Etau:

     Okinoerabu (27.4N, 128.7E) - 950 hPa at 07/0600 UTC
     Naze (28.2N, 129.5E) - 951.6 hPa at 07/1200 UTC
     Murotomisaki (33.2N, 134.2E) - 951.7 hPa at 08/1200 UTC

  NOTE: These SLPs are regular hourly observations.  It cannot be said
  with certainty that they represent the true minimum SLP.

  (4) Focus on Murotomisaki, Kochi Prefecture (WMO 47899)

     The station on Murotomisaki (Cape Muroto) recorded the highest winds 
  in Japan associated with Typhoon Etau.  The station's coordinates are
  33.25N, 134.18E.   The altitude of the station is 185 metres, so this
  should be taken into account when trying to compare the winds recorded
  there with other localities, or with the official MSW estimates.  For
  U. S. NWS and military tropical cyclone warnings, the standard height
  above sea level for which the reported MSW values apply is 10 metres.

     The Murotomisaki station recorded sustained winds (10-min avg)
  exceeding gale force for an extended period, beginning at 07/2100 UTC
  and continuing through 09/0800 UTC.   Sustained typhoon-force winds
  developed around 08/0400 UTC and continued unabated until after 1600
  UTC except for a two-hour period (1200-1300 UTC) when the eye was
  nearest the station.  Winds remained in excess of gale force during
  the closest approach of the eye.  As noted in the table above, the
  peak MSW of about 97 kts occurred at 08/1340 UTC, but the hourly
  observation with the highest MSW was at 1400 UTC with a sustained
  wind of 89 kts.   Sustained winds exceeded 80 kts for two different
  two-hour periods: 0500-0600 UTC and 1000-1100 UTC.  The minimum hourly
  pressure reported by the station was 931.8 hPa at 1200 UTC, which
  equates to a SLP of 951.7 hPa.  Also, as noted above, the peak gust
  of 135 kts was measured at 08/1329 UTC, but according to one of the
  local JMA warnings, the station also recorded a gust of 119 kts at 
  1132 UTC.

     In the history of the Murotomisaki station, only three typhoons
  have brought higher sustained winds, and only two storms have produced
  gusts higher than Etau since January, 1961.  These are tabulated below:

       Sustained Wind (kts)          Date          Name
             135.7               10 Sep 1965       Shirley
             129.7               16 Sep 1961       Nancy
             100.5               21 Aug 1970       Anita
              96.8               08 Aug 2003       Etau

       Peak Gust (kts)               Date          Name
             164.3               16 Sep 1961       Nancy
             149.9               10 Sep 1965       Shirley
             134.5               08 Aug 2003       Etau

  (5) Acknowledgements

     This report was a compilation based on a special report prepared
  by JMA with translation and processing performed by Huang Chunliang.
  Some information was also distilled, translated and edited from the
  raw data on the official web pages of JMA (Japanese version):


     The special JMA report mentioned above can be downloaded at the
  following link:>

     (Size: 885 Kbytes         Language: Japanese)

     A very special thanks to Chunliang for compiling, translating and
  sending along this report, and a thanks to Karl for the information
  he sent.

  D. Damage and Casualties

     No reports of damage or casualties from Okinawa have been received.
  On the main Japanese islands, press reports indicate 3 persons were
  killed as a result of Typhoon Etau with 3 additional persons missing.
  A total of 57 people sustained injuries as a result of the storm.  The
  National Police Agency reported that 15 homes were damaged and 754
  others were flooded.  Torrential rains were responsible for 106 land-
  slides.  There were 464 cancelled airline flights with 58,000 people
  inconvenienced as a result.  In western Japan 18,000 homes were without
  electrical power after the typhoon's passage.

  (Report written by Gary Padgett and Huang Chunliang)

                           TYPHOON KROVANH
                      (TC-12W / TY 0312 / NINA)
                            14 - 26 August

  Krovanh: contributed by Cambodia, is the name of a type of tree
           and also the name of a mountain

  Nina: PAGASA name, is a female name (Spanish for small girl)

  A. Storm Origins

     At 2300 UTC on 13 August an area of convection had persisted near
  6.9N, 155.3E ,or approximately 200 nm east of Chuuk, and this activity
  was first noted on a STWO issued by JTWC at this time.  The chances for
  development were assessed as fair, based on low-level cloud lines
  consolidating near a possible LLCC.  An upper-level LOW was providing
  good outflow aloft and wind shear over the area was light.  Convection
  associated with the suspect area continued to organise through the 14th,
  although the aforementioned upper-level LOW was suppressing activity in
  the northwestern quadrant.     However, this feature was expected
  to fill during the next 24 hours, further increasing the opportunity for
  the development of a tropical cyclone.  This was indicated in a TCFA
  issued by JTWC at 14/2000 UTC.  By this time the disturbance had moved
  northwestward to a position approximately 85 nm northeast of Chuuk (near
  6.8N, 154.7E).  The potential for development remained good at 14/2100
  UTC, and the STWO written at that time stated that the first warning
  would soon follow.  (Note: JMA began classifying the system as a 30-kt
  (10-min avg) tropical depression at 14/0600 UTC, the time of their
  first bulletin.)

     The first warning on Tropical Depression 12W at 15/0600 UTC quickly
  superceded the STWO issued at that time.  The newly-formed tropical
  depression was then located 370 nm east-southeast of Guam and moving
  west-northwestward at 13 kts toward the island.  This first warning
  mentioned a second partially-exposed LLCC located to the southeast of
  the deep convection and the primary centre being tracked.  There was no
  further mention of this attendant circulation, and deep convection waned 
  somewhat during the next six hours. Most of the convective activity had 
  been occurring in the northern semicircle with short bursts over the 
  LLCC.   An increase in convection over the southern quadrants was noted 
  in multi-spectral satellite animations at 16/0600 UTC.    The warning 
  position at this time was based on enhanced infrared satellite imagery 
  which located the centre approximately 160 nm east-southeast of Guam 
  (near 11.6N, 148.0E.)
     By 16/1800 UTC Tropical Depression 12W was approaching Guam, being
  located 90 nm to the east-southeast.  Six hours later the system had
  made its closest approach to Andersen AFB and was centred 30 nm to the
  northwest.  The depression continued to organise during this period,
  and a 16/2332 UTC SSM/I pass depicted deep cycling convection beginning
  to consolidate over the LLCC.  However, there were no other indications
  that strengthening was occurring.  In fact, a QuikScat pass at 17/0833
  UTC showed no evidence of a LLCC but rather an area of convergence near
  the deep convection.   Animated satellite imagery suggested that the
  convection was associated with a mid-level circulation.   Moving away
  from Guam on a west-northwesterly to northwesterly track, TD-12W failed
  to show any signs of an identifiable LLCC, and the 17/1800 UTC position
  (approximately 195 nm northwest of Andersen AFB, Guam) had to be
  determined by extrapolation of previous positions plus the apparent
  centre of the deep cycling convection.  Based on the lack of a proper
  LLCC, JTWC opted to discontinue warnings at 18/0000 UTC.  (JMA issued
  warnings for a further 18 hours before that agency also discontinued
     JTWC included the remnants of Tropical Depression 12W in their STWOs
  through the 18th, assessing redevelopment potential as poor at 0600 UTC,
  then upgrading to fair at 1000 UTC.  UW-CIMSS chart analysis showed
  light wind shear conditions and fair to moderate diffluence aloft.
  Early on the 19th, animated multi-spectral satellite imagery revealed
  an exposed LLCC north of the deep convection, which had increased in
  areal extent.   This was a step in the right direction for the tropical
  cyclone, and at 19/0600 UTC JTWC indicated in an STWO that the system
  would be re-designated as a tropical depression and warnings restarted
  within 6 to 12 hours.  Warnings were initiated at short notice, (i.e.,
  at 19/0900 UTC) and the regenerated depression's initial position was
  approximately 530 nm southeast of Okinawa, or 20.9N, 135.5E.  Deep
  convection had increased over the LLCC, although a 19/1155 UTC SSM/I
  pass showed decreased convection in the northwestern quadrant of TD-12W,
  possibly due to some kind of interaction with the developing Tropical
  Storm Vamco located to the northwest.   (JMA began re-issuing warnings
  at 19/1800 UTC although there was a reference to the system in the 
  19/1200 UTC bulletin.)
     At 0000 UTC on 20 August Tropical Depression 12W was centred approxi-
  mately 440 nm southeast of Okinawa, Japan, and moving westward at 7 kts.
  Recent microwave satellite imagery revealed a partially-exposed LLCC
  north of the deep convection.  The heading turned west-southwesterly at
  0600 UTC as the system became a 35-kt tropical storm, and toward the
  southwest six hours later as TS-12W became a named tropical cyclone when
  JMA upgraded the system to a 35-kt tropical storm (10-min avg).  Newly-
  christened Tropical Storm Krovanh reverted back to a westerly heading
  as it strengthened to 45 kts at 1800 UTC.
     (Krovanh also had another name:  Nina. This was PAGASA's internal
  name, used after that agency began writing warnings at 20/0000 UTC.
  NMCC began issuing statements at 20/1200 UTC while the remaining Asian
  warning agencies--HKO and CWBT--began writing warnings at 22/0000 UTC.)
  B. Storm History
     The MSW had increased to 60 kts by 0000 UTC on the 21st and Krovanh
  was on the verge of becoming a typhoon.  A banding eye was seen but the
  deepest convection was occurring mainly in the southern sectors.  A
  Prognostic Reasoning message issued at this time suggested further
  intensification at a climatological rate until the storm made its
  initial landfall on Luzon.  The extended forecast also suggested a
  possible landfall north and east of Hong Kong within approximately
  84 hours.   Krovanh became a 65-kt typhoon at 21/0600 UTC, although an
  earlier 21/0142 UTC AMSU pass showed that the mid-level circulation
  was slanted to the west of the LLCC.  Also, animated satellite imagery
  depicted the deep convection being sheared to the southwest due to an 
  upper-level LOW to the northeast propagating west toward Krovanh and
  interfering with the poleward outflow pattern.  Undaunted, the cyclone
  strengthened further with the MSW increasing to 75 kts at 1800 UTC.
  Microwave imagery at this time revealed a developing asymmetric eye
  averaging 15 nm in diameter.

     At 0000 UTC on 22 August Typhoon Krovanh was located approximately
  220 nm east of Luzon, having moved back to a west-southwesterly
  heading.  Recent 21/2350 UTC SSM/I imagery showed a ragged 20-nm
  asymmetrical eye, although this feature was not apparent in visible or
  infrared satellite imagery.  The intensity had increased to 90 kts with
  a 10-min avg MSW of 65 kts as estimated by JMA resulting in an upgrade
  to typhoon status by that agency.     By 0600 UTC all agencies were
  monitoring Krovanh/Nina as a typhoon.   During the 22nd Typhoon Krovanh/
  Nina steadily approached Luzon and made landfall at approximately
  22/1115 UTC just north of Palanan.  The MSW was estimated at 80 kts with
  gusts to 100 kts.   During its passage across Luzon, Krovanh weakened
  to 70 kts, and by 1800 UTC was about to move off the west coast of Luzon.
     Six hours later, at 23/0000 UTC Typhoon Krovanh/Nina was tracking
  west-northwestward over the warm waters of the South China Sea, centred
  at this time approximately 65 nm west of Luzon.  The Prognostic Reasoning
  message issued at that hour forecast a continued west-northwesterly
  heading under the influence of a low to mid-level steering ridge to the
  north, and the typhoon was expected to make landfall on China's Leizhou
  Peninsula in the 48-54 hour time frame.   Krovanh was becoming better
  organized but the MSW was not raised until 24/0000 UTC.   By this time
  the typhoon had moved to within approximately 220 nm south of Hong Kong
  after trudging its way west to west-northwestward across the South China
  Sea.  (Krovanh/Nina had by now moved out of PAGASA's AOR.   Both CWBT
  and PAGASA ceased to issue warnings at 23/0600 UTC and 23/1200 UTC,
  respectively.)  Tightly-curved banding features as well as a banding-
  type eye were once again evident in satellite images.  Intensity 
  fluctuated during the next day or so with the MSW reaching 85 kts at 
  24/1200 UTC.  By this time the typhoon was closing in on land, and 
  24/0600 UTC microwave imagery revealed that convection was decreasing 
  in the northwestern quadrant as a result of interaction with the Chinese
  terrain.  However, a well-defined eye was still evident and this clipped 
  the far northeastern coastline of Hainan at 24/1800 UTC.
     At 0000 UTC on 25 August Krovanh was crossing the Leizhou Peninsula 
  of China with a slightly weakened maximum intensity of 80 kts.   The
  banding eye feature remained, although the central core temperature had
  cooled significantly with the deep convection north of the circulation
  (overland) continuing to wane.  At 25/0600 UTC Krovanh was making a last
  stand as it crossed the Gulf of Tonkin on its way to Vietnam.  The MSW
  reached 90 kts for the second time in the typhoon's history during this
  period.  After transiting the Gulf of Tonkin, Krovanh made landfall at
  approximately 25/1500 UTC just north of Cam Pha, Vietnam.  Upon reaching
  the Vietnamese coast the typhoon weakened quickly with winds dropping
  markedly to 65 kts.  The system continued to weaken significantly as it
  moved further into the continent.    Both JTWC and JMA issued final
  warnings at 26/0000 UTC, with JTWC's final position located 145 km north-
  northwest of Hanoi, Vietnam.  NMCC dropped Krovanh from their statements
  at 26/0300 UTC as HKO as done six hours earlier.
  C. Meteorological Observations

     The following are a few rainfall totals I managed to obtain from
  Luzon, Philippines.  However, I was unable to track down the exact
  date, time, and rainfall period.

     Laoag     110 mm
     Iba       203 mm
     Dagupan   342 mm
     Baguio    263 mm

     Huang Chunliang has passed along a wealth of information on the 
  effects of Typhoon Krovanh and also some rain and wind observations 
  which follow.

  (1) Landfalls

     According to the NMC warnings, Typhoon Krovanh made landfall near
  Wengtian Town, Wenchang City, Hainan Province around 24/2010 UTC with
  a MSW of 74 kts and a minimum SLP of 965 hPa.

     According to the NMC warnings, Typhoon Krovanh made landfall near
  Qianshan Town, Xuwen County, Guangdong Province around 24/2215 UTC with
  a MSW of 68 kts and a minimum SLP of 965 hPa.

  (2) Rainfall Measurements

  Province/    Station                   Period           Rainfall        
  Region                                  (UTC)             (mm)
  Hainan       Dongfang             24/0000 - 25/0000       124.7 
  Hainan       Dongfang             24/0000 - 26/0000       197   
  Hainan       Baisha               24/0000 - 25/0000       199  
  Hainan       Baisha               24/0000 - 25/1800       322   
  Hainan       Wenchang             24/0000 - 25/0000       189   
  Hainan       Changjiang           24/0000 - 25/0000       175   
  Hainan       Changjiang           24/0000 - 26/0000       314.7 
  Hainan       Nanfeng              24/0000 - 25/1800       394   
  Hainan       Baoqiao              24/0000 - 25/1800       305   
  Hainan       Jinjiang             24/0000 - 25/1800       392   
  Hainan       Lin'gao              24/0000 - 26/0000       309.7 
  Hainan       Danzhou              24/0000 - 26/0000       246.1 
  Guangxi      Beihai               24/1200 - 25/1200       159.6 
  Guangxi      Beihai               25/0000 - 26/0000       215.7 
  Guangxi      Qinzhou              25/0000 - 26/0000       271   
  Guangxi      Qinzhou              25/0000 - 27/0000       299.4 
  Guangxi      Hepu                 24/1200 - 25/1200       118   
  Guangxi      Fangchenggang        24/1200 - 25/1200       108   
  Yunnan       Jiangcheng           26/0000 - 27/0000       222

  (3) Wind Observations

  (a) Guangdong Province

     (1) Gust reports: Zhanjiang Harbor (87 kts); Shangchuan Dao (78 kts);
         the urban area of Zhanjiang City (72 kts); Haian Town, Xuwen
         County (68 kts); Baoxi Salt Field (66 kts); Jizhao Town, Wuchuan
         City (68 kts); Leizhou City (72 kts).

     (2) Danzao, Nanhai District, Foshan City, reported a peak sustained
         wind of 26 kts with a peak gust to 43 kts before the evening of
         24 August (locally).

  (b) Guangxi Zhuang Autonomous Region

     (1) Weizhou Dao reported a peak gust of 103 kts around 25/0900 UTC.
         This peak turned out to be the maximum value ever recorded in the
         Guangxi Region since 1982.    (The eye of Typhoon Krovanh was
         observed in Weizhou Dao around 25/0540 UTC, when the storm's 
         centre was located 18 km south of the island in the Gulf of 

     (2) Fangchenggang reported a peak gust of 72 kts on 25 August

  D. Damage and Casualties

  (1) Philippines

     News reports indicate that Typhoon Krovanh caused serious damage
  when it crossed the island of Luzon, but these were unspecific.
  According to HKO's report on Krovanh, heavy rains were responsible for
  the death of a girl and 1000 families were displaced.  (See link to
  HKO report below.)

  (2) Guangdong Province, China

     At least two people have been confirmed dead from the typhoon in
  Guangdong.  Direct economic losses in western Guangdong are estimated
  at 1.2 billion yuan.

  (3) Hainan Province, China

     The typhoon toppled 1700 houses and devastated 16 small reservoirs
  in Hainan, affecting the lives and livelihoods of more than 1,884,000
  residents in 129 towns of 11 cities/counties.   However, no casualties
  were reported in the province.  Direct economic losses there are
  estimated at 683 million yuan.  In addition, all shipping services were
  suspended for two days in south China's Qiongzhou Straits between the
  island of Hainan and the mainland.

  (4) Guangxi Zhuang Autonomous Region, China

     Beihai City was most seriously affected in Guangxi.  About 50% of
  the city's area suffered from a lack of power after 25/0700 UTC, and
  the water supply for the whole city had been cut off by 25/0900 UTC.
  Direct economic losses in the city were estimated at 988 million yuan.
  No casualties were reported in Guangxi Region.

  (5) Vietnam

    One death and five injuries were attributed to Krovanh in Vietnam
  where the typhoon destroyed 1000 homes, leaving several thousand people

  E. Additional Information

     Additional information on Typhoon Krovanh can be found in the
  report prepared by the HKO:>

  (Report written by Kevin Boyle and Huang Chunliang)

                       TROPICAL DEPRESSION LAKAY
                            18 - 21 August

     A monsoon trough became rather active after the middle of August
  from the northern South China Sea eastward into the Pacific with the
  axis lying approximately around 20N.  Several circulations formed
  along the trough and were treated in various ways by the several
  TCWCs.  On 18 August PAGASA named one of these disturbances Tropical
  Depression Lakay (a Filipino term for "husband" or "male").  (This
  was the disturbance with the temporary number 98W on NRL's website.)
  JTWC never issued any warnings on the system, but JMA classified it
  as a weak tropical depression in their bulletins.   However, on the
  19th both JTWC and JMA relocated 98W well to the east of their 
  previous positions.  This disturbance subsequently became Tropical 
  Storm Vamco (TC-13W per JTWC's numbering system).

     PAGASA, however, "transferred" Lakay to another disturbance to
  the west, one that was listed on NRL's website as 99W.  JTWC issued
  a TCFA for this disturbance, and JMA monitored the system as a 30-kt
  tropical depression in their bulletins on the 19th and 20th.  JTWC
  began mentioning yet another disturbance (91W) still further west
  in the South China Sea at 20/0000 UTC and assigned a fair potential
  for development.  It appears that by the end of Tropical Depression
  Lakay's life, the 91W center was the one being monitored by PAGASA.
  JTWC issued a TCFA for this third disturbance also, but it was 
  cancelled at 21/0030 UTC as the system had moved inland into China.
  The earlier TCFA for system 99W was cancelled also, at 20/1530 UTC.
  JMA dissipated their tropical depression (which may have been either 
  99W or 91W) at 21/0000 UTC.

     The Chinese warning agencies (NMCC, GRMC and CWBT), like JMA,
  mentioned only one South China Sea circulation in their respective
  bulletins and classified it as a tropical depression.  Some of their
  locations seem to be associated with 99W and some with 91W; others
  were not near where JTWC was locating either of those two systems.  
  The HKO also tracked only one circulation, but didn't even accord 
  depression status to it, classifying it as only a low-pressure area.   
  PAGASA, on the other hand, elevated Lakay to tropical storm status 
  with 35-kt winds (10-min avg) for a period of 24 hours on 19 and 20 

     If all this sounds confusing--don't worry, it truly is--and probably
  was even more so to the harried forecasters who were trying to make
  sense out of the monsoon "mess" as well as keep tabs on two other named
  tropical cyclones (Krovanh and Vamco).

     Huang Chunliang sent me a detailed analysis of the systems in
  question, and the above discussion is based on the material he sent.
  A special thanks to Chunliang for taking the time to attempt to untangle
  the confusing scenario in the South China Sea during those days in
  August and for sending the information to me.  The disturbance which
  moved inland brought some enhanced rainfall to southern China, and
  Chunliang included some of the more notable amounts in his report,
  which are tabulated below:

  Station/                Province/           Period          Rainfall
  Number                  Region              (UTC)             (mm)
  Xiamen (WMO 59134)      Fujian         20/0000 - 21/0000      82.4
  Xiamen (WMO 59134)      Fujian         21/0000 - 22/0000      34.4
  Hong Kong (WMO 45007)   Hong Kong      20/0000 - 21/0000      75.0
  Hong Kong (WMO 45007)   Hong Kong      21/0000 - 22/0000      51.4
  Shanwei (WMO 59501)     Hong Kong      21/0000 - 22/0000      50.2

     Also included was some information from Foshan City in Guangdong
  Province.  The tropical depression triggered a thunderstorm which
  required some weather alerts.  According to the final alert, the
  automatic weather stations located in the districts of Sanshui, Nanhai
  and Shunde reported gusts of Beaufort Force 8 to 9 during the storm.
  Rainfall amounts of 30.4 mm, 42.7 mm and 38.8 mm were recorded in the
  districts of Chancheng, Nanhai and Shunde, respectively, during a
  one-hour period.

  (Report written by Gary Padgett, based primarily on information sent
  by Huang Chunliang)

                         TROPICAL STORM VAMCO
                     (TC-13W / TS 0311 / MANANG)
                            19 - 20 August

  Vamco: contributed by Vietnam, is the name of a river in the western
         part of southern Vietnam which originates in Cambodia

  Manang: PAGASA name, is a Filipino female nickname, used primarily
          for an older, unmarried woman

  A. Storm History

     Short-lived Tropical Storm Vamco began as an area of convection
  located roughly 400 nm east of Luzon, Philippines.  The STWO issued at
  0600 UTC on 18 August relocated the area to a position near 20.0N,
  122.3E, or approximately 120 nm south-southeast of Taiwan.  Animated
  enhanced satellite imagery depicted disorganised convection on the
  southern side of a LLCC.  Upper-air analysis showed weak shearing
  conditions over the area and the development potential on this state-
  ment was upgraded to fair.  This was further upgraded to good at 19/0000
  UTC based on improved organisation during the previous six hours.  This
  STWO was issued in anticipation of a tropical depression, and the first
  warning on Tropical Depression 13W was released at 19/0300 UTC.  (JMA, 
  NMCC and CWBT issued their first bulletins at 19/0000 UTC.)

     At 0000 UTC on 19 August Tropical Depression 13W was located approxi-
  mately 270 nm east-southeast of Taiwan, tracking toward the north-
  northwest at a leisurely 5 kts.   Even though the system was rapidly
  organising, the Prognostic Reasoning message indicated only marginal
  strengthening due to the presence of less-than-ideal upper-level
  conditions, interaction with land, and possible interference from the
  remnants of Tropical Depression 12W.   By 0600 UTC the forward speed had
  more than trebled with the storm temporarily moving on a northward path.
  At this time JMA and NMCC upgraded the depression to tropical storm
  intensity with JMA assigning the name Vamco.   (PAGASA issued the first
  of three warnings at this time on Manang (their internal name), and
  after the storm had skirted through that agency's AOR, the third and
  final warning was issued at 19/1800 UTC.   HKO also began issuing
  warnings on Vamco at 19/0600 UTC but never classified the system as a
  tropical storm.)

     Movement for the rest of Vamco's life was toward the northwest, and at
  19/1800 UTC was centred only 30 nm northeast of Taipei, Taiwan.  (JTWC
  upgraded Vamco to tropical storm status at this time.)  However, recent
  microwave imagery indicated that the LLCC was partially-exposed to the
  east of the deep convection.  After its close flirt with Taiwan, the
  centre made landfall near Fanshan, China, at approximately 0030 UTC on
  20 August and was downgraded to a 25-kt tropical depression at 20/0600
  UTC.  Vamco was dissipating as a significant tropical cyclone over land
  by 20/1200 UTC, the time of the final JTWC warning.   (All agencies, bar
  CWBT, had discontinued warnings by this time:  JMA at 20/0300 UTC, HKO
  at 20/0600 UTC, and NMCC at 20/0900 UTC.    CWBT issued their last state-
  ment at 20/1800 UTC.)

  B. Meteorological Observations

     Huang Chunliang has sent some rainfall observations and wind reports.
  (Thanks to Chunliang for sending this information.)

  1. Landfall
     According to the NMCC warnings, Tropical Storm Vamco made landfall
  near Aojiang Town, Pingyang County, Wenzhou City, Zhejiang Province,
  around 20/0210 UTC with a MSW (10-min avg) of 45 kts and a minimum SLP
  of 990 hPa.

  2. Rainfall Data

  Provinces/     Stations                          Periods       Rainfall
   Regions                                           UTC            mm
  Zhejiang       Pingyang, Wenzhou City       19/0000 - 20/0000     70
  Zhejiang       Pingyang, Wenzhou City       19/0600 - 20/0600    101
  Fujian         Fuding, Ningde City          19/0000 - 20/0000     52
  Taiwan         Jhuangwei, Ilan County       18/1600 - 19/2100     69
  Taiwan         Luodong, Ilan County         18/1600 - 19/2100     61
  Taiwan         Suao, Ilan County            18/1600 - 19/2100     50
  Taiwan         Tatunshan, Taipei City       18/1600 - 19/2100     50
  Taiwan         Rueifang, Taipei County      18/1600 - 19/2100     51

  3. Gust Observations from Zhejiang Province

  Station               Peak Gust
  ---------             ---------
  Pingyang                49 kts
  Nanji Dao               54 kts

  D. Casualties and Damage
     Preliminary statistics indicated that some 326,000 residents of
  Pingyan County, Zhejiang Province, were affected by the storm, though no
  casualties were reported. A total of 984 houses collapsed and 4896 were
  partially damaged in the county.  Moreover, 1286.67 hectares of paddy-
  fields were ruined with 653.33 hectares of economic crops being damaged.
  Direct economic losses there were estimated at 38.56 million yuan.
     Like the earlier Severe Tropical Storm Morakot, Tropical Storm Vamco
  turned out to be more of a beneficial tropical cyclone than a baleful
  one, too.  Thanks to the storm, the drought in southern Zhejiang and
  northern Fujian was eased once again.

     No other reports of damages associated with Tropical Storm Vamco
  were received.

  (Report written by Kevin Boyle and Huang Chunliang)

                            TYPHOON DUJUAN
                      (TC-14W / TY 0313 / ONYOK)
                        28 August - 3 September

  Dujuan: contributed by China, is the name for the flowering plant
          'azalea'; the word also means 'cuckoo' in Chinese

  Manang: PAGASA name, is a Filipino male name which also means 'midget'

  A. Storm Origins

     Typhoon Dujuan originated from an area of deep convection that was
  noted in GOES-9 infrared satellite images as early as 25 August.  (All
  the STWOs that included the pre-Dujuan disturbance were unavailable
  to the authors.)  The slow-moving suspect area could be tracked for
  several days moving westward with variable convective amounts, and in
  fact by the end of the 25th there was very little associated convection.
  This theme continued into the 26th, but increasing organisation was
  evident by late on the 27th.  JTWC issued a TCFA on 27 August, locating
  the system near 18.4N, 138.1E.  The first warning was issued at 28/0600
  UTC, followed by an upgrade to tropical storm status on the second
  warning six hours later.  (The tropical cyclone remained unnamed until
  JMA upgraded it to tropical storm intensity at 29/1800 UTC.)   Tropical
  Storm 14W continued to organise while moving southwestward to west-
  southwestward, and at 29/0000 UTC was located approximately 750 nm east
  of the Philippines (near 16.1N, 135.3E.)  Animated multi-spectral imagery
  revealed enhanced equatorward outflow and satellite CI estimates were
  30 and 35 kts.

  B. Storm History
     Animated multi-spectral satellite imagery at 29/0600 UTC showed that
  the northern portions of Tropical Storm 14W were being sheared as a
  result of a nearby TUTT cell located just north of the system.  Feeling
  the benefit of equatorial outflow, a second channel was opening as the
  upper-level LOW moved in tandem with the cyclone.  At 29/1200 UTC the
  warning intensity was raised to 45 kts, in line with satellite CI
  estimates, and PAGASA named the system Onyok as the storm entered their
  AOR.  The internationally recognized name Dujuan came into force when JMA
  upped their 10-min avg MSW to 35 kts at 29/1800 UTC.  The warning issued
  by JTWC at this time indicated that Dujuan was on the verge of becoming
  a typhoon with the MSW estimated at 60 kts and water vapour imagery
  indicating increased outflow in all quadrants.
     At 0000 UTC on 30 August the MSW reached 65 kts based on a compromise
  of satellite intensity estimates ranging from 55 to 77 kts.  At this
  time Typhoon Dujuan was located approximately 690 nm east of the
  Philippines, moving on a slow westerly heading of 3 kts.  Animated multi-
  spectral and microwave imagery showed a tightly-curved band forming near
  the centre of the system.     As Dujuan turned northwestward and
  accelerated, the MSW was increased further to 75 kts, and a tightly-
  curved banding eye feature was noted at 30/1200 UTC in microwave imagery.
  After briefly turning poleward, Dujuan reverted back to its westerly to
  west-northwesterly heading the next day.  (JMA upgraded Dujuan to typhoon
  status at 30/1800 UTC, and NMCC and PAGASA did so six hours later.)
     By 31/0000 UTC 31 Typhoon Dujuan/Onyok was positioned approximately
  595 nm east-southeast of Taiwan with a MSW of 80 kts.  This intensity
  was upped further to 90 kts at 0600 UTC, and raised to 100 kts twelve
  hours later.  A 31/1214 UTC 85 GHz microwave image showed the presence
  of tightly-wrapped banding features but no indications of an eye.
  However, in contrast, a symmetrical eye was analysed on the 37 GHz
  channel.  Continuing on a west-northwesterly path, Dujuan had moved to
  within 260 nm east-southeast of Taiwan by 0000 UTC 1 September.  The MSW
  had risen to 120 kts, equivalent to a Category 4 hurricane, at this time
  with an upper-level LOW east of Dujuan aiding outflow.  A 31/2235 UTC
  SSM/I pass depicted concentric eyewalls.  By 01/1200 UTC the southern
  part of Taiwan lay within the radius of 50 kt winds as the eye of Dujuan
  was centred approximately 65 nm southeast of the southern tip of the
  island.  At this time the intensity was peaking at 125 kts, but fell
  back slightly to 120 kts at 1800 UTC.  The symmetrical concentric eyes,
  noted in an earlier SSM/I pass, had disappeared, although an eye feature
  was evident in a 01/2232 UTC TRMM image.    (The peak 10-min avg MSW
  estimated by JMA was 80 kts.  Among the Asian TCWCs, the highest 10-min
  avg MSW of 100 kts was assigned by HKO.)

     Even though Typhoon Dujuan was past its prime and was now on a
  definite weakening trend, the storm's glory days were not over.  By 0000
  UTC on 2 September, Dujuan was beginning its approach to Hong Kong,
  China, with the maximum intensity still at 120 kts.   At this time, the
  central core of Dujuan lay approximately 220 nm east of Hong Kong,
  closing to 150 nm by 0600 UTC.  Another six-hour increment saw Dujuan
  making landfall on the Chinese coast with a maximum intensity of 100 kts.
  (The exact time of landfall as indicated in JTWC's 02/1800 UTC warning--
  #22--was at 1400 UTC.)   Now moving westward across the Chinese mainland,
  the typhoon weakened quickly and by 1800 UTC the MSW was barely at
  typhoon strength.  The last warning issued by JTWC at 03/0000 UTC
  located the dying storm near Yulin, China (22.6N, 110.5E).  Dujuan was
  downgraded to a 40-kt tropical storm on this last statement, based on 
  synoptic reports and CI estimates of 65 kts.   NMCC downgraded Dujuan to 
  30 kts at 03/0000 UTC, and issued their final warning three hours later.

  C. Meteorological Observations
  (1) Landfall (based on the NMC warnings)

     Typhoon Dujuan made its first landfall near Gangkou Town, Huidong
  County, Guangdong Province, around 02/1150 UTC with a MSW of 78 kts, 
  gusting to 107 kts, with a CP of 960 hPa.  Crossing the Gulf of Daya, 
  the storm made a second landfall near the coastal area of eastern 
  Shenzhen City, Guangdong Province, around 02/1250 UTC.  Dujuan entered 
  the waters of the mouth of Zhujiang River after rampaging across 
  Shenzhen.  Then, the third and final landfall occured near Nanlang Town,
  Zhongshan City, Guangdong Province, around 02/1515 UTC with a MSW of 
  64 kts, gusting to 78 kts and with a CP of 975 hPa.

  (2) Rainfall Measurements

  Station                     Province/        Period      Rainfall
                              Region            (UTC)        (mm)
  Hengchun (WMO 46759/59559)  Taiwan     31/1600 - 01/1600   106.0
  Hengchun (WMO 46759/59559)  Taiwan     01/1600 - 02/1600   108.0
  Chengkung (WMO 46761)       Taiwan     01/1600 - 02/1600   152.5
  Lanyu (WMO 46762/59567)     Taiwan     01/1600 - 02/1600   133.0
  Taichung (WMO 46766/59562)  Taiwan     01/1600 - 02/1600   146.0
  Tu-chang, Ilan County       Taiwan     31/1600 - 01/1500   197
  Tan-ta, Nantou County       Taiwan     31/1600 - 02/0000   205
  Hsi-ceiu, Hsinchu County    Taiwan     31/1600 - 02/0900   229
  Liyutan, Hualien County     Taiwan     31/1600 - 02/0900   511
  Yuli, Hualien County        Taiwan     31/1600 - 01/2100   304
  Fangshan, Pingtung County   Taiwan     31/1600 - 01/2100   271
  Chih Pen, Taitung County    Taiwan     31/1600 - 01/2100   153
  Guangfu, Hualien County     Taiwan     31/1600 - 02/0000   350
  Shihding, Taipei County     Taiwan     31/1600 - 02/0000   114
  Maubitou, Pingtung County   Taiwan     31/1600 - 02/0900   628
  Taipingsha, Ilan County     Taiwan     31/1600 - 02/0900   385
  Luye, Taitung County        Taiwan     31/1600 - 02/0900   341
  Ao-wan-ta, Nantou County    Taiwan     31/1600 - 02/0900   261
  Kaokou, Chia-I County       Taiwan     31/1600 - 02/0900   191
  Mu Cha, Taipei County       Taiwan     31/1600 - 02/0900   152
  Nansha, Guangzhou City      Guangdong  02/0000 - 03/0300   145.9
  Puning City                 Guangdong  02/0000 - 03/0000   131
  Puning City                 Guangdong  02/0000 - 04/0000   183.0
  Lishui, Nanhai District,    Guangdong  02/0000 - 03/0200   100.9 
     Foshan City                                             

  (3) Hourly Sustained Wind and Pressure Observations

  NOTE:  All MSW values are 10-min avg winds.

     (a) Station: Lanyu, Taiwan (WMO 46762/59567),
         Lat. 22.0N,  Lon. 121.6E,  Alt. 325 m

        Winds reached sustained gale force at 01/0100 UTC, then
     dropped below gale force during the next two hourly observations.
     At 01/0400 UTC winds had risen above gale force and so remained
     through the 1700 UTC observation.  Winds were 60 and 62 kts at
     0500 and 0600 UTC, respectively, and the peak hourly MSW of 89 kts
     occurred at 1400 UTC.  The 1500 UTC MSW was 84 kts, and the minimum
     hourly pressure of 926.6 hPa occurred at 1300 UTC.

     (b) Station: Dongshi, Taiwan (WMO 46730)
         Lat. 23.3N,  Lon. 119.7E,  Alt. 45 m

        The maximum hourly MSW of 41 kts occurred on 2 September
     from 0100 through 0400 UTC.  The minimum pressure hourly reading 
     of 982.4 hPa was taken at 1700 UTC on 1 September.

     (c) Station: Hengchun, Taiwan (WMO 46759/59559)
         Lat. 22.0N,  Lon. 120.8E,  Alt. 24 m

        The maximum hourly MSW of 41 kts occurred at 1700 UTC on
     1 September while the minimum hourly pressure of 963.6 hPa was
     measured at 01/1600 UTC.

  (4) Peak Sustained Wind and Gust Observations from Taiwan
  NOTE:  All MSW values are 10-min avg winds.

     Only those stations that reported peak gusts of typhoon force
  are given (wind direction is in degrees/dates are local dates):

                   Peak Sust. Wind        Peak Gust
  Station         (kts / dir / date)  (kts / dir / date)
  Lanyu            94.9/  40 / 1st    146.2/  20 / 1st
  Dawu             30.7/ 110 / 2nd     71.1/ 100 / 1st
  An Bu            28.6/ 180 / 2nd     65.1/ 200 / 2nd 
  Keelung          22.4/ 100 / 2nd     64.9/ 200 / 2nd
  Hengchun         45.3/ 110 / 2nd     92.7/ 160 / 2nd
  Sun Moon Lake    22.6/ 120 / 2nd     66.5/ 120 / 2nd 
  Dongshi          47.2/ 150 / 2nd     65.7/ 150 / 2nd 

  Note 1: Stations Taitung, Taipei, Hsinchu, Penghu, Kaohsiung, Chengkung,
  Ilan and Suao all reported peak gusts of Beaufort Force 10.

  Note 2: Yu Shan, a weather station located in the mountainous region of 
  Nantou County, reported a peak gust of 99.3 kts at 01/1911 UTC.

  Note 3: It should be noted that the anemometer in Lanyu was destroyed 
  around 01/1300 UTC--the third time since the weather station was built
  56 years ago.

  (5) Wind Observations from Fujian Province

  Station Peak Gust
  The urban area of Fuzhou City      52 kts
  Licheng, Quanzhou City             51 kts
  Yongchun, Quanzhou City            52 kts
  Chongwu, Quanzhou City             51 kts
  Jiuxianshan, Quanzhou City         78 kts
  Dongshan, Zhangzhou City           64 kts

  Note: Gusts of Beaufort Force 7 or higher were recorded in 30 cities/
  counties of the province during the period from 01/0600 though 02/1200

  (6) Wind Observations from Guangdong Province

  Station Peak Gust
  Nansha Power Factory, Guangzhou City   91 kts
  Nansha Headquarters, Guangzhou City    75 kts
  Fangcun District, Guangzhou City       42 kts
  Nanshan District, Shenzhen City        97 kts

  Note: Jiujiang, Nanhai District, Foshan City, reported a peak 10-min
  avg sustained wind of 37.9 kts at 02/1622 UTC with a peak gust of 59.6
  kts at 02/1612 UTC.

  D. Damage and Casualties

  (1) Taiwan Region

     In southern Taiwan three people were killed, one was missing and
  eight were injured during the approach of Dujuan.  The electricity
  supply to about 590,000 families was interrupted.  All transportation
  services in southern Taiwan were temporarily suspended.  As of 04/0800
  UTC the agricultural loss had reached NT$1.417 billion with another
  NT$46.9 million losses in stockbreeding.

  (2) Fujian Province

     Over 500 trees were uprooted or partly damaged in the urban area of
  Fuzhou City.  However, no casualties were reported in Fujian province.

  (3) Guangdong Province

     About 6,410,000 residents in Guangdong Province were affected by the
  typhoon.  Highways, infrastructure for telecommunications, water and
  power supplies, irrigation systems, and crops suffered serious damage
  as Dujuan, the most powerful typhoon that has affected the Zhujiang
  River Delta since 1979's Typhoon Hope, swept through.   Some 54,000
  houses collapsed, and 139,000 hectares of crops were damaged.  Direct
  economic losses in the province were estimated at 2.287 billion yuan. 

     Power supplies in 90% of the area in Shenzhen were interrupted as
  the typhoon struck the city. 

     Dujuan also turned out to be the most deadly typhoon for China so
  far this season.  In Zhujiang River Delta, at least 46 people lost
  their lives and about 1000 were injured in the storm.

  (4) Hong Kong Special Administrative Region

     In Hong Kong Dujuan disrupted public transport systems, closed
  the Hong Kong stock exchange, and caused the cancellation of 221
  flights while delaying 139 others.  More than 2000 people were
  stranded.  Injuries in association with Dujuan were slight (only
  22 reported), but four fishermen were missing after the storm's
  passage and feared drowned after a rescue search failed to find 
  their boat.

     See the report on Typhoon Dujuan prepared by the HKO at the
  following link:>

  (5) Macao Special Administrative Region

     The Macao Meteorological and Geophysical Bureau (SMG), the local
  meteorological service, raised storm signal No. Eight, the third-highest
  storm warning signal as the typhoon approached the south China coast
  Tuesday evening.  The public traffic was suspended at 9:00 pm on Tuesday.
  Over 30 flights through the Macao Airport were delayed, and two bridges
  linking the Macao Peninsula with Taipa Island were closed as the
  No. Eight warning was raised.  However, no big losses were reported.

  E. Double-eyed 'Cuckoo'

     Both the Taiwan and Hong Kong radars captured the 'cuckoo' with
  double (concentric) eyes.  A HKO report on this interesting phenomenon
  can be found at the following link:>

  Some CWB radar imageries described a possible anticlockwise movement
  of the inner eye, with the outer one being the frame of reference,
  while the typhoon was heading for the southern tip of Taiwan Island.

  (Report written by Kevin Boyle and Huang Chunliang)


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

  Activity for August:  1 tropical depression **

  ** - classified as a tropical depression by IMD only

               North Indian Ocean Tropical Activity for August

     I received some information from Huang Chunliang regarding a system
  in the Bay of Bengal which was classified as a depression by the Indian
  Meteorological Department (IMD).  This system was mentioned on 27 August
  as a low-pressure area over the northwestern Bay of Bengal off the
  Orissa-West Bengal coast.   During the evening it strengthened into a
  depression and moved in a westerly direction, crossing the Orissa coast
  early on the 28th.  The system continued moving farther inland and had
  weakened into a low-pressure area by the 29th.  The bulletins from the
  IMD indicate that the axis of the monsoon trough at sea level passed
  through the center of the depression, so the system is probably best
  classified as a monsoon depression.  No warnings were issued on this
  system by JTWC; in fact, it was never mentioned in that agency's STWOs
  for the North Indian Ocean.  The depression did enhance monsoonal rain-
  fall in portions of eastern India.  (A special thanks to Chunliang for
  sending me the relevant IMD bulletins.)


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

  Activity for August:  No tropical cyclones



  Activity for August:  No tropical cyclones



  Activity for August:  No tropical cyclones


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

  Activity for August:  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
  from the archive sites listed below.  (Note: I do have a limited e-mail
  distribution list for the track files.    If anyone wishes to receive
  these via e-mail, please send me a message.)

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

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


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

     The URL is:>

     Also, TPC/NHC has available on its webpage nice "technicolor"
  tracking charts for the 2002 Atlantic and Eastern North Pacific
  tropical cyclones; also, storm reports for all the 2002 Atlantic
  and Eastern North Pacific cyclones are now available, as well as
  track charts and reports on storms from earlier years.

     The URL is:>

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


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

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

  Huang Chunliang (Assistance with Western Northwest Pacific, South
                   China Sea)
  E-mail:  [email protected]

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


Document: summ0308.htm
Updated: 26th October 2006

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