Reply to NHC Best Track Change Committee comments of
2 June 2005, 8 June 2005, 30 March 2006, 6 April 2006,
19 April 2006, 1 June 2006, and 26 July 2006.
Reply by Landsea indented – February 2008
Minutes of the best-track change committee meeting, 02 June, 2005
Members Jack Beven, Eric Blake, Jim Gross, Richard Pasch, Ed Rappaport, and
chair (Colin McAdie) present.
We note that prolonged early activity in the Atlantic has intervened since
The committee met to consider the most recent submission (1915-1922) from
Landsea et al.
By agreement, only 1915, 1917, and 1918 were to be considered at this meeting.
It was found that the metadata sheets in the notebooks differ from the emailed
version. We assume, but wish to confirm, that the emailed version supersedes
those found in the notebooks.
1. Yes, indeed the emailed versions supersede that found in
the notebooks. But hopefully, these are now identical.
It seems that most of the changes submitted revolve around a comparison of the
Historical Weather Map Series, COADs, and HURDAT. The HWM series are
obviously valuable in clarifying many situations, but are ultimately analyses.
As such, they are necessarily based upon sparse data in the vicinity of a
developing circulation. As noted previously, the committee feels that
observations form the strongest argument for changes. In particular the
frequently cited ‘lack of a closed circulation’ should be supported by
observations. Also, as noted previously, the terminology ‘a low of at
most 1010 mb’ could be a source of confusion to others readers of the
metadata. This phrase appears to refer to the value of the innermost closed
isobar drawn in the analysis.
2. It is agreed that the Historical Weather Map analyses are
subjective, based sometimes on quite sparse data and thus can
often be incorrect. It is agree that "lack of closed circulation"
needs to be supported by observations in order to make a significant
change to existing HURDAT positions.
Do those currently involved in the reanalysis have a standard set of criteria
for diagnosing subtropical systems? e.g. how is it determined that there is
no front? Could these criteria be forwarded to the committee?
3. Given that the 1910s and 1920s have only surface data available,
we are unable to adequately differentiate between tropical cyclones
and subtropical cyclones. (Though, occasionally, we may have enough
observations near the inner core to have evidence for a large RMW of
a cyclone, which might indicate a subtropical, rather than tropical,
cyclone. However, such high density of observations is relatively
It is recommended that until the geostationary satellite era and
the NCAR-NCEP/ECMWF reanalysis era (beginning in 1966) that
subtropical cyclones not be indicated in HURDAT. Currently, the first
subtropical storm listed in HURDAT is in 1968.
Concerning the list of U.S. landfalls, it was noted that the time of landfall
is given to the nearest 10 minutes. It should be noted that these times are
obtained through interpolation (if so) since this implies an unrealistic
precision. Also, with respect to system size, the location of the outermost
closed isobar is not well known, even today. Going forward, how will this be
integrated with the quadrant system used in current operational products?
4. It is agreed that the landfalls of U.S. hurricanes and tropical
storms be given to the nearest hour. The estimates of the outer closed
isobar and its radius are provided only for U.S. landfalling hurricanes
at the synoptic time closest to landfall. This is consistent with
current operational practices, but is not as extensive.
There was some discussion about the difficulty of not having the ship
observations plotted. We assume these have been plotted by those doing the
reanalysis, but they do not seem to be consistently shown in the notebooks.
If these are plotted elsewhere, could they be supplied to the committee?
5. All plots conducted by the reanalysis team are contained within
the notebooks provided. Typically, the synoptic analysis from the HWM
is combined with the COADS data once daily (around 12 UTC, using 10-14
UTC ship data). For fast moving cyclones over relatively data rich
areas, twice or even four times daily synoptic analyses are performed.
Numerous typos and spelling errors were noted. A spell-checker might serve to
filter many of these out.
6. Hopefully, spell check has caught the remaining typographical
1915 #1. The wind at Jacksonville suggests leaving the track as it was
originally, farther to the west.
As far as the intensification over land, this system does not appear to be
independent of the baroclinic system. That is, it is probably not a purely
tropical, intensifying system. Are there any sources indicating the system
might have been stronger at landfall (e.g. damage south of Jacksonville) ?
1915/01: It is agree to keep the track near Jacksonville farther to
the west as originally indicated. It is agreed that the cyclone was
reintensifying as a hybrid system and extratropical transition is
now indicated 6 hours earlier at 18 UTC on the 4th. Thank you for
pointing out the additional information on this cyclone on pages 484-5
of MWR, as we had unfortunately overlooked these. Based upon the
observations found here as well as obtaining the USWB Original Monthly
Records for Jacksonville, it did suggest that the system was likely
of minimal hurricane intensity at landfall (65 kt). Winds are
boosted upward accordingly.
1915 #2. There appears to be an inconsistency in the pressure at San Juan
on 8/11. Was the pressure 29.60 (page 5) or 29.77 (page 6) ?
It appears that the track should be closer to Dominica, and later, about 6Z.
Are there any observations from Guadeloupe?
Should the extratropical portion of the track be further delayed beyond the
12h extension currently proposed? Despite the classic-looking baroclinic
development depicted in HWM, it seems somewhat unlikely that a category 4
would completely transition this quickly, especially in mid-August.
The revised landfall pressure (953 mb to 930 mb) is questioned. It appears
that the 28.06 pressure (950 mb?) observed at Velasco was inside the RMW.
Given this, 945 mb would be a better estimate.
Should Ho’s smaller RMW be used here (29 n mi) ?
It is not clear why the “Schloemer” equation is used. Please provide a
1915/02: The discrepancy in the San Juan pressure is mentioned in
the daily metadata writeup, but it is uncertain which is correct.
For the suggestion to move the track close to Dominica, the track
already essentially goes over the island. Given the observed minimum
pressure at Dominica and the rather good data coverage of this fast
moving hurricane, a close passage to the island around 18 UTC on the
10th appears quite reasonable. It would not fit any of the available
data to delay the track 12 hours for closest approach to Dominica at
06 UTC on the 11th. Unfortunately, there are no data or reports
available from Guadeloupe. It is agreed that the extratropical
transition should be later given the observations available. This is
now estimated to have occurred around 00 UTC on the 20th, 36 hours
that originally listed in HURDAT. It is agreed that the 930 mb central
pressure at landfall estimate was too low. However, 945 mb appears
to be too high, especially given the 955 mb pressure measured in
Houston (inland) at the same time as 70 kt of wind. A value of 940 mb
is proposed instead. This estimate is likely the highest that could
support a Category 4 winds, given the RMW of around 25-30 nmi. The
Schloemer equation's use was dropped.
1915 #3. The duration of the impact (described as three days) is questioned.
An examination of track and size suggests that a duration of 12h may be more
realistic. An extensive ship report in Bermuda appears in MWR, but is not
discussed in the metadata.
The ship Kilbride was at anchor through the event. Although well-removed from
the eye, a pressure of 28.85 in (977 mb) was reported. It is noted in MWR that
these pressures may be .2 inches too high (983 mb?) when compared to Hamilton.
In any event, this in conjunction with the damage strongly suggests that the
intensity should not be reduced, especially in light of the solid construction
typical of Bermuda.
Also note that relevant articles in MWR may appear in later editions. For
example, note the discussion of the storm of 1-2 August 1915 within the
September edition (pp 484-485).
1915/03: It is agreed that a better description is that the primary
impact occurred over 12 hours. The extensive ship report from the
Kilbride anchored in Bermuda was included and discussed, but the ship's
name was just not mentioned. All of the Kilbride's observations are
in the excel spreadsheet and the peak ob for each day had been
described in the metadata for the 2nd, 3rd, 4th, and 5th. It is
agreed to retain the cyclone as a Category 3 near Bermuda, so no
changes are made to HURDAT's winds from the 2nd until the 6th.
1915 #4. Concerning the removal of track on 31 Aug, does the data rule out a
closed low? Weaken, rather than remove. 30 kt would appear to be reasonable.
Are there any data from the Cayman Islands? A strong tropical storm on 1 Sept
would blend better with Perez’s category 2 at landfall.
Note last sentence - ‘genesis’ should be ‘lifetime.’
1915/04: It is agreed to keep the track of the system back to
the 31st of August, though as suggested weaker than in HURDAT
originally. Unfortunately, no data other than that in HWM are
available for the Cayman Islands. The cyclone is indicated to
be a strong tropical storm on the 1st as recommended. "Genesis"
change to "lifetime" as suggested.
1915 #5. (new) accepted.
1915 #6. On 9/22 is data only available from Antigua? Nothing from the
other islands? The metadata is inconsistent in that the data from Antigua
suggests a stronger system than indicated in the revision. In fact, the data
from Antigua suggest extending the track eastward. It is noted in the
metadata that the HWM does not analyze a closed low, however the wind at
Trinidad (HWM) does suggest that it could have been closed. Also note that
a SW wind in the ‘ABCs’ is very difficult to obtain, and supports
a stronger system.
p 21 wording 20 feet above ‘seal level’
p 22 ‘winds were significantly reduced from the 22nd to the 26th is at is
There are a number of questions concerning reduction of intensity at landfall.
Track not accepted as revised. Concerning the following, Jack Beven has
agreed to be the focal point.
The chief difficulty is that the data support a double eyewall, with the outer
wind max affecting Mississippi. The track of the eye is not clear. In the
discussion, one possibility is that the location of the Weather Bureau office
is in error. The pressure observed at Tulane was probably near the central
pressure, assuming a small 10 n mi diameter eye. See Sullivan for possible
category 2 conditions in Mississippi.
Committee also notes that the storm surge model may not adequately handle
multiple wind maxima.
1915/06: Unfortunately, no data is available from the other islands.
It is agreed to extend the track eastward to the 21st and to revise
the reanalyzed winds upward. The typographic errors are corrected.
The reanalysis team met with Jack Beven and arrived at the main
conclusions at landfall in Louisiana as shown in the metadata: 944 mb
hurricane at landfall with a double RMW of 20 and 55 nmi and peak
sustained winds of 110 kt at landfall. Mississippi's impact was
increased as suggested to Category 2.
Minutes of the best-track change committee meeting, 30 March, 2006
Members Jack Beven, Eric Blake, Richard Pasch, Ed Rappaport, and chair
(Colin McAdie) present.
The committee met to further consider the most recent submission (1915-1922)
from Landsea et al. With the time available, 1916 through early September of
that year (system 7) was discussed.
It was noted that most of the changes submitted revolve around a comparison of
the Historical Weather Map Series, COADs, and HURDAT. (please see full
discussion on previous minutes).
Also, as noted previously, the terminology ‘a low of at most 1010 mb’
could be a source of confusion to others readers of the metadata. This phrase
appears to refer to the value of the innermost closed isobar drawn in the
Numerous typos and spelling errors were noted.
1916 #0. Possible new system – please investigate. May 12-17, 1916.
MWR gives substantial evidence of a system moving northward out of the tropics
and affecting the Florida peninsula.
1916/00: Agreed - system added in as new tropical storm (as new
1916 #1. What observations support the new positions on 6/28?
On p. 2 (July 4) note that the first sentence appears to refer to another
system. As far as the intensity at landfall, was the publication
“Hurricanes and Tropical Storms in the Gulf of Mexico consulted? There is
a cluster of observations plotted in the area of landfall. Based on the
discussion in MWR, it appears that the RMW may have been smaller than 26 n mi.
Was the Ft. Morgan observation used explicitly?
The pressure observation in Jackson on the 6th is used to derive a wind of
56 kt, apparently with a pressure-wind relationship. As noted, this is well
inland. Would a conventional pressure-wind relationship apply in this case?
Climatology would suggest a weakening before landfall. Similar cases in this
area suggest that the system would be stronger over open water, then weakening
15-20 kt before landfall. Final sentence. The statement about unexpected
landfall is a commentary on the forecast, which seems out of place in this
1916/01 (now 1916/02): Observations from HWM and COADS support
beginning the cyclone a day earlier on the 28th. The July 4th
description is corrected. The observations from the "Hurricanes
and Tropical Storms in the Gulf of Mexico" were utilized, but it
is agreed that the RMW was smaller than Ho et al. indicated. An
estimate of 15-20 nmi is accepted for use here.
As we do not know what the intensity was over the open Gulf of Mexico,
it may be too speculative to assume that the cyclone was significantly
weakening before landfall - especially with it retaining a tight RMW
at landfall. However, the new Brown et al. north of 25N pressure-wind
relationship is more conservative than the Gulf of Mexico pressure-wind
relationship from Landsea et al. and the Brown et al. one is utilized
here. The wind derived from the central pressure in Jackson explicitly
takes into account the difference between marine and over land
exposures, as the 56 kt directly from the relationship is reduced to
1916 #2. Where is the 50 kt ship noted on the 15th? By our interpretation,
it seems to be quite distant (100 n mi) from the track. Track and intensity
changes seem very inconsistent with this observation. Track seems to have
been moved away from the ship. Please reanalyze.
Note that MWR gives a maximum wind for July for San Juan which was likely
induced by this system.
Committee notes that the 980 mb ship observation does seem to be
representative of the central pressure, supporting reduction in intensity for
that portion of the track.
1916/02 (now 1916/03): The 50 kt ship was found in the COADS dataset.
The track and intensity were revised to have the cyclone closer to
the high wind reports and to retain the original intensity (hurricane
force). The MWR maximum wind for San Juan has been added into the
reanalysis, and does indicate a tropical storm for the northeast
1916 #3 (A number of spelling errors are noted in this discussion.) Would
not the 971 mb peripheral pressure offshore imply a central pressure of
possibly 960 mb? It is not clear why the hurricane is weakened as it makes
landfall. Rather, given the peripheral pressure, small size, and slow forward
motion, it seems possible that this was a major hurricane at landfall. Is
there any information from Cary Mock on this system?
It seems that the 961 mb pressure reported by the U.S.S. Hector on the 13th
should appear in the HURDAT file. The track appears to go directly over the
1916/03 (now 1916/04): It is agreed that the central pressure of
the hurricane at landfall is stronger than originally estimated.
960 mb is now suggested with winds of 95 kt - at the border of
Category 2 and 3. Prof. Cary Mock's analysis from the impacts of
the hurricane (storm surge and wind caused damage) are also in
agreement with about that intensity. The 961 mb pressure reported
by the USS Hector was accompanied by 70 kt ENE wind at 1930 UTC on
the 13th at 31.8N 78.9W - suggesting a position just south of
the ship. This is what is indicated in HURDAT.
1916 #4 Several important points are noted with this potential category 4
landfall. The 28.00” (948 mb) barometer reading appears to be well north
of the center position. Metadata does state that this was inland, but equally
important is that it was not a central pressure. Given the distance inland,
an extremely rapid filling rate is implied here. The question is, what was
the central pressure at the time of the 948 mb observation? Since the RMW is
used to decrease the wind, what is the confidence in this estimate? Also note
that the track plot color coding does not agree with the numerical data.
On p. 14, end of paragraph: it is stated that the intensity was underestimated
by 20 kt at landfall, rather than 10 kt. Committee requests run by Storm Surge
group for consistency with observed surge.
1916/04 (now 1916/06): It is agreed that the 948 mb pressure value
from Kingsville is not a central pressure reading. It is estimated
that the central pressure at the time of closest approach was about
940 mb. The Ho et al. inland pressure-decay model with a 1.5 transit
overland suggests a 932 mb central pressure at landfall. Given
the storm surge runs performed with 932 mb central pressure and
RMW of 25 nmi match the observed storm tides, this gives some
confidence in the revision to a Category 4 hurricane. The track map
is corrected for this hurricane as well as the description of the
change in intensity.
1916 #5 If the 35 kt southerly wind at Miami is correct, the track should be
pushed westward over the peninsula, rather than kept offshore. It would be
consistent with this observation to add 5 kts to the intensity of the storm at
1916/05 (now 1916/07): It is agreed to place the track to the west
of Miami and boost the landfall intensity to 40 kt.
1916 #6 The eastward extension of this track does not appear to be justified
based on the HWM.
Although ‘available observations’ are noted in the discussion, the
reduction of wind on the 27th-28th does also not appear to be justified.
MWR notes a small hurricane of great intensity on the 28th.
It is likely with a small system that the winds would have been undersampled
by ship observations.
Page 17, last paragraph, change 27th and 18th to 27th and 28th. Page 18,
first paragraph, suggested wording, “Barometric pressure increased at Swan
1916/06 (now 1916/08): It is agreed to remove the eastward extension
of the track. It is agreed to retain the original intensity on
the 27th and 28th. Wording changed as suggested.
1916 #7 Given the 35 kt ship observation on the 5th, the wind should be
higher than 35 kt.
With an accompanying pressure of 1017mb, it seems that the ship was not near
the center. Are there enough observations to add two initial positions? HWM
indicates a open wave.
1916/07 (now 1916/09): It is agree that the cyclone was stronger than
originally indicated and winds are raised to 45 kt. It is agreed that
the system dissipated earlier than indicated and the last position is
now estimated to have occurred at 00 UTC on the 7th.
Minutes of the best-track change committee meeting, 06 April, 2006
Members Jack Beven, Eric Blake, Lixion Avila, Richard Pasch, Ed Rappaport, and
chair (Colin McAdie) present.
The committee met to continue discussion of 1916, part of the most recent
submission (1915-1922) from Landsea et al. With the time available, 1916 was
It was noted that the storm surge group has agreed to run 1916 #4 (category 4
landfall along south Texas coast). If this run supports the intensity at
landfall, the committee will finalize its comments and submit the change for
After serving on the committee since its inception, Ed Rappaport has decided
to step down. The committee will discuss Ed’s replacement at the next
1916 #9 becoming #8. Ship observations well away from the center indicate
tropical storm force winds, or greater. Given the distance (as much as three
degrees latitude) these observations from the 18th through the 20th in fact
support hurricane intensity, rather than a reduction to tropical storm. In
particular the 45 kt ship 125 n mi to the southeast of the center is more
consistent with the original HURDAT hurricane status, rather than a reduction
to a 50 kt tropical storm.
1916/09 becoming 08 (now 1916/10): It is agreed that intensity should
not be dropped below hurricane force for this cyclone, given the gale
force winds at large radii on more than one day.
1916 #10 becoming #9. It is not clear why the peak intensity was reduced by
10 kt. The effects on Bermuda in fact seem to support major hurricane status.
The HWM series is used as a source. Given this, a key question in this case
is why did the HWM analyst draw for a 995 mb low, especially given the usual
lack of observations? It seems that the analyst may have consulted other
sources, including existing track maps, or the predecessor to HURDAT.
1916/10 becoming 09 (now 1916/11): It is agreed that the cyclone be
retained as a major hurricane as it was making its closest approach
to Bermuda. It is unknown exactly what data sources the HWM analyst
was using to come up with his minimum pressure of 975 mb on the 22nd,
though it is possible that the impact in Bermuda was well-known.
1916 #12 becoming #10. 963 mb is mentioned as a ‘possible’ central
pressure. Given this, should it appear in HURDAT? This system may have been
a major hurricane during its passage through the Virgin Islands. Please
reanalyze with this in mind, especially if the 963 mb was likely a central
1916/12 becoming 10 (now 1916/13): It is agreed to use the 963 mb
as a likely central pressure and it is included into HURDAT as such.
Using the low latitude pressure-wind relationship, this suggests
about 95 kt - at the border of Category 2 and 3. It is possible that
the cyclone was of major hurricane intensity during its passage through
the Virgin Islands.
1916 #13 becoming #11. Please change central pressure on 10/18 to 970 mb.
The 85 kt wind estimate at Swan Island suggests increasing the winds near the
center. Also, do the wind observations at Swan Island need to be corrected?
Committee concurs that category 3 intensity is not justified.
1916/13 becoming 11 (now 1916/14): It is agreed to change the central
pressure at U.S. landfall to 970 mb. The 85 kt at Swan Island is a
visual estimate and would be quite uncertain (with no current method
to correct the wind). However, it is agreed to keep the winds as
originally shown in HURDAT (95 kt) during the hurricane's closest
approach to Swan Island.
1916 #14 becoming #12. It was noted that this system is reminiscent of H.
Floyd in 1987. Acceptance of the revised track (the extratropical portion)
hinges on the central question of whether the front could cause winds as strong
as observed, or should they be attributed to the system itself. Please note
the apparent cyclonic circulation to the south on 15 November. Is this the
tropical system? Note that an examination of local data available in the
library (Miami Annual and Monthly Local Climatological data 1911-1961) is
not consistent with a tropical cyclone passage.
Wording, p. 30, last paragraph before additional notes: for clarity replace
“at least 982 mb” with “higher than 982 mb.”
1916/14 becoming 12 (now 1916/15): It is likely that the strong
winds observed can be attributable to the cyclone after it
became extratropical. This is now more clearly stated in the metadata.
It should be reiterated though that the 62 kt observed at Sand Key
adjusts to 51 kt 1 min true. It is likely that the apparent cyclonic
circulation to the south on the 15th is ficticuous - note the 1017 mb
pressure with the ship with the north winds.
Additional system #2. Are there any tropical storm force winds observed
after the system lost extratropical characteristics? It was also noted that
there is no mention of this system in MWR.
1916 Additional system #2: Yes, there are gale force wind reports
after the system had lost extratropical characteristics. But the
large distance to the center from these observations suggest that
they cyclone was a large occluded low, rather than a tropical (or
Additional system #4. (existing storm #11) Please retain this storm. This
is a possible subtropical system. Note that although HWM shows an
extratropical system, the WNW winds are not consistent with an extratropical
1916 Additional system #4: Agreed - system retained as a tropical
storm (as 1916/12).
Additional system #5. Please add to the tropical depression file.
Additional system #6. How close to gale-force were the “numerous
1916 Additional system #6 (now Additional Notes #5): There were
several 25-30 kt reports from the 25th to the 27th, which is now
clarified in the writeup.
Note: The August 1916 MWR contains the following: 8 pm, Aug 5. A small
disturbance off the Rio Grande moved over Mexico. Storm force winds noted
along the Texas coast.
1916 Note: This August Gulf of Mexico system was investigated and
it is suggested to be added into the database as a new 50 kt tropical
storm (new 1916/05).
Minutes of the best-track change committee meeting, 08 June, 2005
Members Jack Beven, Eric Blake, Jim Gross, Richard Pasch, Ed Rappaport, and
chair (Colin McAdie) present.
The committee met to continue consideration the most recent submission
(1915-1922) from Landsea et al.
Only 1917 was considered at this meeting.
As in the previous discussion, numerous typos and spelling errors were noted.
1917 #1. Is it reasonable to upgrade this system to a tropical storm just
before landfall in Belize based on one observation (1006 mb peripheral
pressure)? At least two observations would strengthen the argument.
The positions of ships given in MWR for 7-8 July are extremely general.
Observations seem quite limited in this area. Can these ships be positioned
Wording - last sentence “almost four inches of rain were reported in
Veracruz in the 24 hour period ending on the morning of the 14th.”
1917/01: It is agreed that not enough definitive evidence was
available to analyze the cyclone as a tropical storm making landfall
in Belize. It is kept instead as a tropical depression at landfall.
Unfortunately, the two ships reporting west, gale force winds did not
come with an exact location in MWR and these were not found in COADS.
Wording on the rainfall changed as suggested.
1917 #2. It is not clear how a decision can be made on tropical depression
or tropical storm status based on the information given. That is, given
essentially identical information on August 6, 7, and 8, the system is
downgraded to a depression on the 6th. Is there a difference in central
The discussion quoted from MWR seems totally unrelated. What is the
significance of the statement about a system that formed east of the Virgin
The system is made extratropical too soon. It should be remain a tropical
storm through the 10th, increase intensity to 55 kt. What is being analysed
here is not the low shown in MWR. This should be characterized as a major
change to intensity.
1917/02: It is agreed that the observations available on the 6th, 7th
and 8th were nearly identical and thus now no downgrade in intensity to
a tropical depression on the 6th is indicated. The quote from the MWR
is indeed for this cyclone, but it appears to be an incorrect analysis.
This is now clearly stated in the metadata. It is agreed to continue
the cyclone as a tropical storm an additional 12 hours. Because of its
continuance as a tropical storm at its closest approach to
Massachusetts, the 994 mb peripheral pressure suggests winds of at
least 58 kt - 60 kt chosen for HURDAT. This is now indicated to be a
"major" change to the intensity. It is agreed that the MWR Tracks of
Lows has the cyclone far from the correct location on the 11th. This
is now clarified in the metadata writeup.
1917 #3. The reduction in intensity does not seem justified. The effects
described in the text are consistent with a major hurricane, based on how this
information would be interpreted today.
1917/03: It is agreed that this system was likely a major hurricane.
Original intensities are now retained from HURDAT.
1917 #4. Can it be determined that the 939 mb pressure was in fact peripheral?
Likewise, can the source of the 928 mb be determined? The track appears to
go directly over the 939 mb observation.
Concerning the U.S. landfall, how is the RMW determined? The metadata is
somewhat confusing, as it seems that Pensacola was not in the center of the
system. The RMW is quite important in this case, as it is further used to
adjust the intensity.
It seems unlikely that this system would have dissipated this rapidly. Are
there any COADS data or other data that would support extending the track?
1917/04: According to Perez (personal communication) the Nueva Gerona
pressure observation of 939 mb was obtained when the winds were still
substantial (estimated to be storm-hurricane force), though no
anemometer was present at the site. Perez et al. (2000) utilized the
Schloemer equation to estimate the 928 mb central pressure. This is
now clarified in the metadata writeup. The track certainly did come
very close to the 939 mb reading, as the measurement was inside the
RMW. With regards to the U.S. landfall, the RMW was directly estimated
by analyzing the Pensacola wind record relative to the track of the
hurricane. This is now so stated in the metadata. With regards to
the decay, the cyclone's track was extended 12 hours (through 06 UTC
on the 30th) to accommodate a realistic weakening. As was noted, a
100 kt landfalling hurricane climatological decays to a minimal
tropical storm in about 24 hr, as per the Kaplan-DeMaria inland wind
decay model. The COADS data were obtained (and added into the
database). These data - strongest wind being 25 kt - reconfirmed that
no closed circulation was observable by 12 UTC on the 30th, as the
cyclone had been absorbed into a strong cold front
Additional system #2. Committee agrees with the assessment that this system
might be a tropical storm. The 35 kt ship is about 115 n mi away using the
positions given. Moving the track closer to the ship may be justified.
1917 - Additional Notes #2: It is agreed that a position of 20N 52W
may be a better estimate on the 15th.
Minutes of the best-track change committee meeting, 19 April, 2006
Members Jack Beven, Eric Blake, Lixion Avila, Richard Pasch, and chair
(Colin McAdie) present.
The committee met to discuss 1918, part of the most recent submission
(1915-1922) from Landsea et al. With the time available, 1918 was completed.
The committee agreed that Hugh Cobb should be asked to replace Ed. He has
worked on reanalyzing storms affecting the mid-Atlantic states, and this work
will complement our present capabilities.
It was also agreed that the HURDAT entry for Helene (1958) was in error, and
should be changed to reflect category 3 conditions along the North Carolina
coast. The coastal-crossing index will be entered as zero (XING=0) indicating
that the center did not cross the coast, while SSS=3 will indicate that
category 3 conditions were created over the coast.
1918 #1. The removal of positions on 8/1 and 8/2 does not seem justified.
Further, this results in an extremely unlikely genesis position in the central
Caribbean. A small system could easily have slipped through the islands.
After some discussion, it was decided that the increase in landfall intensity
is justified. First paragraph of discussion, change “Regarding the
probably…” to “Regarding the probable….” Last paragraph
1918/01 - Agreed to keep in the positions/intensities on the 1st and
2nd as originally indicated in HURDAT. Wording changed as suggested.
1918 #2. Note that changes to this system are not minor. On the 22nd, the
42 kt at Barbados supports the HURDAT intensity of 60 kt, rather than
supporting a decrease. There is a dilemma with the use of the 50 kt ship
report (S.S. Mohegan) on the 23rd. Apparently the track has been shifted
northward because of the ship report, but the intensity has also been reduced
– something of a self full-filling prophesy. That is, how do we know the
ship was near the center? Is there anything to support this? Some track
movement north is OK, but the committee feels that 60 kt is more appropriate
for HURDAT, barring any further information. The intensity increase off the
coast of Honduras is found to be justified.
1918/02 - Wording used for "minor" changes are generally track
adjustments of 120 nmi or less and intensity adjustments of 20 kt or
less. Larger ones have been called "major" changes. Agreed to keep
the 60 kt intensity near Barbados on the 22nd and to have smaller
north adjustment and 60 kt on the 23rd.
1918 #3. There is evidence that this system was a hurricane at the time of
landfall. Note in MWR, p. 378 a report of 64 mph at Cape Hatteras on the 25th.
Also, considerable damage is also noted in MWR, which seems consistent with
previous hurricane occurrences in this area. What is the minimum pressure
noted in OMR?
1918/03 - It is agreed that this cyclone was a Category 1 hurricane
at landfall in NC. More detail is included into the metadata regarding
the impacts in the region. The minimum pressure recorded was 1002 mb
at Wilmington, though this was not likely the central pressure at
1918 #4. Accepted, but please smooth out this track, given the paucity of
data. Track suggests much more detail than is known. There is some
inconsistency in the use of HWM. While in this case a temperature
gradient is noted across the front, similar evidence was used in 1916 to
conclude a system was baroclinic (additional #1). It seems that similar
conditions have been used to draw an opposite conclusion.
1918/04 - It is agreed to smooth the track of this new tropical storm.
While it is difficult to objectively analyze the baroclinicity of this
and other cyclones because of the lack of data, it appears that this
one had a smaller temperature gradient associated with it.
1918 #5 (formerly #4). Note that the comments for Sept 3 are for some other
system. Was the pressure at Bermuda 972 or 978 mb? In either case, the track
should be moved closer to Bermuda, or the intensity should be increased
1918/05 - The comments on the 3rd are for this system. The official
reading from Bermuda was 978 mb, but unofficial though potentially
valid reading of 972 mb was also described. The track is moved a bit
closer to Bermuda and the intensity is increased somewhat.
1918 #5 (original). Do not drop this system. Observations in the Windward
Islands clearly indicate the existence of a tropical system. It is likely
that the system progressed westward into the Caribbean, rather than turning
northward. The west winds in western Cuba are diurnal, and are considered
irrelevant in this case. Please note the ship observations in the western
Caribbean on the 13th, as they appear in the HWM analysis. Lack of data
cited does not justify dropping this system.
1918/06 (was 05 original) - It is agreed to retain this system. It is
also agreed that the cyclone turned northward, rather than progressing
westward into the western Caribbean.
Additional system #1. Agree this is a TD.
Additional system #2. Do not add to depression file
Additional system #3. At least a TD, possibly a weak TS. Well-documented in
MWR. Please add to the tropical depression file. Sharp pressure falls are
noted. The HWM series analysis is very shaky and should not be used. Use an
independent analysis to determine status of this system.
Additional system #4. System does not appear in Connor. Note that Pensacola
and Sand Key observations appear to be irrelevant. What does “verifying
velocity” mean as used by Isaac Cline?
Minutes of the best-track change committee meeting, 01 June, 2006
Members Jack Beven, Eric Blake, Hugh Cobb, Richard Pasch, and chair
(Colin McAdie) present.
The committee met to discuss 1919, part of the most recent submission
(1915-1922) from Landsea et al. With the time available, discussion of 1919
1919 #1. Please retain an initial position as a 25 kt depression. The
development along a frontal zone would argue for slow development, and the
data does not rule out a small closed low at 12Z. Was there a simultaneous
wind observation to accompany the 999 mb pressure at landfall? The 999 mb
was apparently west of the center and therefore would support a central
pressure of 995 mb and a corresponding wind of 55 kt. This would be somewhat
more consistent with the damage noted.
1919/01 - Agreed to retain initial (00 UTC) position as 25 kt tropical
depression. While there was not a wind available right at 1115 UTC,
peak 5 min winds were 43 kt (36 kt true) both between 10-11 UTC and
11-12 UTC. This would suggest a central pressure around 995 mb
and a corresponding wind of 55 kt, which is now utilized for the
1919 #2. The large pressure drop observed at San Juan on 9/2 suggests
retaining a position at 12Z. The 18Z position might be moved a bit east to
accommodate this. The large pressure falls at San Juan and St. Thomas during
9/2 – 9/3 and the 3 year summary also suggest moving the track closer to
Although the large reduction in intensity over the coast of Hispanola appears
to be justified, by the 7th the revised intensities are lagging and should be
increased. This is also supported by the 998 mb peripheral pressure at Nassau.
This pressure suggest a large, powerful hurricane.
There are a number of significant pressure observations on the 10th. Can
these be pieced together in a coherent way? During the same period, the 927 mb
central pressure would support a wind of 135 kt. During transit across the
mid-Gulf (10/10 – 10/13) the system is reanalyzed to have a constant
intensity of 130 kts. Recent analogs (e.g. Rita) suggest this is unlikely and
somewhat unrealisitic. Although it may be difficult to explicitly describe
this with the data available, it is considered much more likely that the
The proposed intensity at landfall seems to depend upon the 950 mb observation,
which may not have been a central pressure. If it was a central pressure, this
is not reflected in the track. If there is no reason to believe this was a
central pressure, the hurricane should be left as a category 4 at the time of
Also, please examine the possibility that the remnants of the system passed
over New Mexico.
1919/02 - Agreed to retain a position at 12 UTC on the 2nd, to move
the track slightly eastward on the 2nd, and to locate the cyclone
closer to Puerto Rico on the 3rd. Agree to boost the intensities
on the 7th and 8th. The 10th has three amazing central pressure
readings from the Dry Tortugas and two nearby ships. While it is
uncertain, the lowest value (927 mb) may have sampled the central
pressure and is thus retained as the landfall pressure (already
found in Jarrell et al.) and in HURDAT. After reconsideration,
the 944 and 942 mb pressure readings on the 12th are likely
central pressure values due to the consistency of values between
different ships. This brings the intensity down to a Category 3
on the 12th. The 950 mb pressure reading on the 14th would appear
to be a central pressure value (and was assumed to be so in
the Ho et al. and Jarvinen et al. studies). However, the exact
location of this ship was fairly ambiguous and the track/timing
was more definitively determined by landfall data in Texas. Agreed
to extend the track to 12 UTC on the 16th to near New Mexico.
1919 #3 (new). Accepted, however please check for the possibility of a
1919/04 (new): Agreed. It is possible - for at least a portion of
the lifetime of this cyclone - that it was a subtropical storm and this
is now mentioned in the metadata writeup. Official use of this
designation is begun in the 1960s, at the advent of satellite imagery.
1919 #4. This drastically revised track is not accepted and needs work.
The observations from Bermuda are central to this reanalysis. The winds at
Bermuda seem to support a system just ahead of the front, which is also
consistent with the existing transition to extratropical. Note that
although the HURDAT position on the 13th may be too far north, this drastically
revised track does not seem consistent with HWM.
1919/05 (was old 1919/03): The system has been reworked, after
discussing the cyclone with Jack Beven. It was a difficult one
to analyze because of sparseness of data, some erroneous ship
observations, and interaction with a strong cold front. Mark
Guishard (Bermuda Weather Service) was able to provide additional
observations which did suggest that the cyclone stayed south and
east of Bermuda, instead of passing just west and then north of
the island as shown in HURDAT originally.
1) This is a broad system in the western Caribbean and the pressure-wind
relationship may not be strictly applicable. In this regard, 1004 mb does
not necessarily imply a TS.
1919/Additional System 1: Agreed that the pressure-wind relationship
may not be as applicable because of the low environmental pressures.
3) This system may have been a tropical storm. Committee rejects HWM analysis.
The 5 mb pressure falls and gradient support a short-lived tropical storm. A
recent analog would be TS Earl. A tropical storm would also be more
consistent with the ship running aground.
1919/Additional System 3: Agree that this may have been a tropical
storm and additional discussion about this is included. Position
5) The committee cannot discount the Tannehill interpretation, and feels that
this system may have been a hurricane ahead of the front. Please note
discussion in 1919 MWR pp. 660-661. The 28.85 ship report cannot be assumed
to be central, given a small system.
1919 - was Additional System 5, now is new storm #3: Agreed that there
is sufficient evidence to add this system in as new tropical cyclone
(hurricane). The MWR ship observations were crucial to it being
6) This system may also have been a tropical storm, given the 9/10-9/11
pressure falls at Bermuda. Please re-analyze.
1919/Additional System 6: Agreed that this system may have been
a tropical storm. However, there are no data that explicitly support
this. More discussion has now been included into the metadata
Additional possible systems:
Aug 13-14. Note rain and wind observations, in particular the 62 mph
observation at Cape Henry, VA. Analyzed as a frontal low, but may have
become subtropical of the U.S. east coast.
1919/Additional System 4 (new - Aug. 13-14): Cyclone has been
examined, but fairly conclusive that it was extratropical for most or
all of its lifecycle. However, it is possible it was subtropical on
the 13th. This system is now included in the metadata writeup.
Nov. 17-19. Note that warnings were issued for the Florida coast, and a low
has been analyzed in the Bahamas on the HWM. There appears to be evidence
for a possible subtropical system.
1919/Additional System 8 (new - Nov. 17-19): System has been examined,
and it appears to be an extratropical cyclone throughout its lifetime.
A discussion of it has now been added into the metadata.
Minutes of the best-track change committee meeting, 26 July, 2006
Members Jack Beven, Eric Blake, Hugh Cobb, Richard Pasch, and chair (Colin
The committee met to discuss 1920 and 1921, part of the most recent submission
(1915-1924) from Landsea et al. With the time available, discussion of 1920
was completed. The meeting concluded after discussion of the first (existing)
storm of 1921.
It should be noted again that available observations should support proposed
changes. A lack of observations is an entirely different and insufficient
argument for change.
1920 #1. In the discussion (metadata) for September 10, please reference the
name of the ship which was the source of the 70 kt wind report. Please smooth
the track in the vicinity of this observation (around the 10th). The proposed
shift in the track in response to this observation seems warranted, but the
resulting kink is not realistic.
Also, please retain the existing intensity on the 12th. This would place the
peak intensity near the inflection point in the track. Reduction on the 13th
and thereafter OK.
Spelling: search for ‘caost’, ’23 miles and hour’
1920/01: All of the suggested changes were implemented.
1920 #2. The pressure drop in Belize suggests shifting the track westward
on the 19th. As the system reaches the mid-Gulf, the multiple 70 kt ship
reports suggest increasing the intensity on the 21st to 60, 65, and 70 kt at
0, 6, and 12Z respectively.
At the time of landfall, note that the surface pressures at Morgan City and
Houma (given the track) do not appear to be consistent. Are there any
observations from Baton Rouge? It would seem that the RMW should have passed
almost directly over them.
1920/02: Agreed to shift track westward on the 19th. Agreed to
boost on the winds on the 06 and 12 UTC of the 21st, but the 999 mb
central pressure at 03 UTC precludes going as high as 60 kt at 00 UTC.
The track is shifted slightly east for a better match to the lower
pressure in Houma than in Morgan City. No observations of winds
or pressures were available for Baton Rouge.
1920 #3. Data supports moving the tropical storm phase ahead to 12Z on the
20th. The reductions in intensity at 0Z and 6Z on the 22nd are accepted,
however please increase intensity to 70 kt at 12Z and 75 kt at 18Z based on
70 kt ship report.
The reduction to a tropical storm at the time of landfall was discussed, and
accepted. The system may in fact have been small, as reported, but the
preponderance of evidence seems to indicate that it was not a hurricane at the
time of landfall.
1920/03: Agreed to all minor changes suggested.
1920 #4. This new system is accepted
1920 #5. On the 25th, the numerous ship reports do not appear to rule out
a depression on this date, and it is suggested that this 24-hour period not be
dropped as proposed. Metadata for the 26th notes a “closed, but
ill-defined storm” but proposed intensity reduces system to a depression at
this time. An intensity of 35 kt at 12Z on the 27th would reflect a more
reasonable development. Suggest increasing peak intensity to 75 kt at 18Z on
the 29th based on available ship reports. Spelling: search for
1920/05 (new): Agreed to all minor changes suggested.
A possible addition to the additional systems:
Please see p. 610 MWR. Note discussion of a disturbance on Oct 9th affecting
extreme south Florida and the Florida straits. Evidence is not overwhelming,
but this might be a “possible.”
1920 Additional System: This was actually two separate disturbances.
The system of October 7-11 off of the Atlantic seaboard was
investigated, but it appears to have been extratropical for the
duration of its lifecycle. A second disturbance near South Florida
was also analyzed, though this appears to have been an open trough.
Both of these have been added into the metadata writeup.
Return to Overview