Best Track Committee Re-Analysis Comments for 1933

General comments:

1. Several new acronyms are introduced in the metadata, including QJR, TAN, CHA, and VH. Please spell them out somewhere in the text.

QJR is Quarterly Journal of the Royal Meteorological Society. The first time that QJR is used in the 1933 metadata… this has now been spelled out. TAN is Tannehill. The one time the acronym CHA was used was deleted. CHA means Cape Hatteras NCDC data. The “CHA” was replaced by “NCDC” in Storm 8. VH is Virginia Hurricanes.

2. For storms 3 (formerly 2), 10, 11, and 12, the genesis time has been moved up (in one case by a week) without a good description of the basis for doing it or (in some cases) providing maps. Please take another look at these storms to a) better justify the earlier start, and b) ensure that the documentation is complete.

For Storm 3 (formerly Storm 2), moving the genesis up 3 days was due to the combination of an analyzed intensity of 70 kt at 12Z on 27 June and a west wind of 15 kt located a few hundred nm SE of a north wind of 15 kt on 24 June… indicating that if a tropical depression already existed by that time, that the analyzed position should be reasonably accurate. This cyclone can either be begun on the 27^th as a strong tropical storm or low end Category 1 hurricane or it can be begun on the 24^th as a tropical depression. The analyzed position on the 24^th based on the data that day indicates a very consistent speed and direction of motion from the 24^th until well beyond the 27^th, and this indicates that the circulation on the 24^th near 9.0N, 42.5W is the same circulation as the hurricane on the 27^th at 9.7N, 59.0W. The circulation on the 24^th was more likely a tropical depression than a monsoon circulation because the circulation does not appear to be broad/elongated. The two available observations indicate a more compact circulation, more like a tropical depression. Text has been added to the metadata, and no changes have been made to revised HURDAT.

Storm 10 is analyzed to have been a continuation of original Storm #7, which was removed (see additional note #4). This system is analyzed as a tropical depression from 14 August at 12Z – 20 August at 18Z. It is analyzed as an open wave from 21-24 August. It is analyzed to have become a tropical depression again on 25 August, and this is the date of genesis analyzed for Storm 10. The maps for the August 14-25 portion of this system are in the “removed TCs” portion of the binder. However, the August 24^th and 25^th maps as well as the observations from the Excel sheet from the 24^th-25^th have now been placed with the Storm 10 material. After a more careful analysis, it has been determined that there is probably not enough evidence of a closed circulation on the 25^th and early on the 26^th to justify a change to the timing of genesis. Therefore, the timing of genesis in the original HURDAT – 18Z on 26 August – is retained as the revised genesis time. The period from August 25^th 00Z – August 26^th 12Z, which had been recommended as an extension to the genesis by 42 hours, has now been deleted from the metadata. Also, the metadata summary paragraph has been slightly edited.

For Storm 11, after further analysis, it appears that although west winds are observed on 20-21 August, there is not enough evidence of a tropical cyclone on those days. However, on 22 August, there appears to be enough evidence that a tropical depression had formed by that day near 12.7N, 26.0W. The lowest pressures, which are sufficiently lower than other pressures farther away, are observed along with sufficiently cyclonic turning winds over a small enough area. The 20^th at 12Z through the 22^nd at 00Z on deleted. The cyclone is begun as a 25 kt tropical depression at 06Z on the 22^nd, increasing to 30 kt at 00Z on the 24^th. Everything else remains unchanged from our previous analysis. The main reason for lowering the winds for the 22^nd-23^rd is because the highest observed wind during that time was only 20 kt. The metadata summary paragraph has been edited.

For Storm 12, after further analysis, it is determined that although a tropical cyclone might have existed on 27 August near 14N, 28W, there is not nearly enough evidence to make this determination. The genesis time of this tropical cyclone originally listed in HURDAT (06Z on 31 August) should not be changed, as there is not enough evidence to begin the cyclone earlier. The 27^th through 00Z on the 31^st is deleted. The 55 kt intensity reanalyzed for 06Z on the 31^st is unchanged (which is an increase from the original HURDAT intensity of 45 kt). The metadata summary paragraph has been edited.

2a. On a related note, the earlier genesis of some of these systems seems to be based on the presence of a close circulation in the eastern Atlantic. While this is not a bad idea per se, in the current era we are able to see monsoon-type closed circulations that do not meet tropical cyclone criteria. Please re-examine these cases in light of that.

This comment has been addressed, where appropriate, in the response to comment #2.

3. There are several uses of the term “open low” in the metadata text. Has this term been used before? If not, please define it somewhere, as it is a bit confusing. It seems to lie somewhere between “open wave” and “closed low”.

“Open low” had referred to HWM plots with an “L” but with no closed isobars analyzed. A better term would be “spot low”. Everywhere in the metadata that said “open low” is now replaced with “spot low.”

4. Storms 14 and 15 both made landfall near Tampico, Mexico with central pressures of 960 mb and documented eye passages that lasted about 2 hours in each case. However, storm 14 is shown with a 95 kt landfall intensity while storm 15 is assigned 100 kt. The only apparent difference is that storm 14 had “lower than normal environmental pressures”. Please re-examine these landfalls and strengthen the explanations of the assigned intensities. Can RMW’s for the Tampico landfall’s be calculated from the available data.

Storm 14 did have a significantly lower than normal environmental pressure. The 960 mb central pressure occurred at Tampico at about 14Z on the 15^th. 11 hours earlier, at 03Z, the pressure at Tampico was 998 mb. 11 hours after the 960 mb, at 01Z on the 16^th, the pressure at Tampico was 1002 mb. The analyzed speed of the storm at landfall is 13 to 14 kt, which means that the cyclone would have been about 155 nm from Tampico 11 hours before and after landfall. Those numbers correspond to a pressure gradient of roughly 40 mb / 155 nm from the center to the 1000 mb isobar. Furthermore, an analysis of the September 15 map reveals an outer closed isobar (OCI) of around 1003-1004 mb with an ROCI of around 200 nm. For the RMW, a TC moving 13.5 kt with a calm center lasting for 2 hr means that the center is at least 27 nm wide, which means that the radius of the calm was at least 13.5 nm. This should correspond to an RMW of at least 20 nm. The climatological RMW for a central pressure of 960 mb at 22N is 16 nm.

For Storm 15, the OCI was around 1006 mb and the ROCI was around 200 nm. Although a ship was inside the eye for 2 hours about 6 hours prior to landfall… and RMW cannot be estimated from available data because it is unknown how the ship was moving during the time inside the eye, and no visual descriptions were given as to the size of the eye. The forward speed of the cyclone was 11 kt. It can be argued both ways to go 95 or 100 kt for this situation. Since the original HURDAT had a 95 kt intensity at the point before landfall, and since the environmental pressures are low, a 95 kt intensity is probably the right choice even though the pressure-wind relationship yields 102 kt before taking into account the effects of low environmental pressure. The intensity at 9/24 18Z and 9/25 00Z are both now shown to be 95 kt.

5. This is the first time we’ve seen coded ship reports in the data binders, and it’s an old code that no one on the committee can interpret. We need to find the decoder for this code, as well as the many other codes (updated ship/land station, SAO/METAR, aircraft, and radar) that we’re likely to encounter as we advance through the re-analysis.

The ship codes were found. After decoding all of the ship reports from the 6 pages of Bermuda ship obs in the 1933 binder… the peak intensity of Storm 8 was increased from 115 to 120 kt. A couple of changes were made to Storm 9 as well. The latitudes of Storm 13 were moved north by as much as 0.5 degrees on 13-14 September based on these ship obs.

1933 Storm #1 (new):

1. The committee concurs that this systems had a tight hurricane-like inner core. However, it does not concur with adding the system to HURDAT at this time. It notes that the temperature on Bermuda and at the nearby ship were 60-65F. It also notes that the Historical Weather Maps (HWM) show a significant temperature gradient around the precursor disturbance on 25 April. Please perform a detailed analysis of the temperature fields of this system, as there is a chance it may have been more of a frontal hybrid system than a tropical cyclone. If the system is eventually added, it may be necessary to add a couple of extratropical points at the beginning of the track.

Isotherms for the air temperature in increments of 5°F have been plotted on maps for 4/25 12Z and 4/26 12Z. There appears to be about a 5 degree temperature gradient across the developing circulation on the 25^th with temperatures between about 67-73 F. On the 26^th, there appears to be little in the way of any temperature gradient across the low. Uniform temperatures in the mid-upper 60s surround the low. Temperatures of 60-61 are observed within 50-100 nm of the center, and this could be due to rainfall cooling down the air temperatures. Pressure and anemometer traces from Bermuda shown in the QJR (1934) report indicate a significant probability that this was a hurricane. The QJR report is in Binder #2 for 1933. This cyclone is re-submitted to the committee for consideration that it was a tropical cyclone.

2. What is the relevance of the 39 kt ship report over the Gulf of Mexico on 25 April? Please better relate it to the cyclone or delete it.

Ship highlight in metadata on 25 April actually occurred at 77.2W, which is near the developing cyclone and is therefore a relevant observation. The position of the ob in the metadata is corrected from 87.2W to 77.2W.

3. Is there any information on the height of the anemometer on Bermuda? This may have an impact on the analyzed intensity of the system.

After an extensive search, this could not be obtained.

1933 Storm #2:

1. Give the number of ships near the system on 14 May, is the evidence sufficient to start this system as a 25 or 30 kt depression instead of as a tropical storm?

At 12Z on 14 May, a ship with an east wind of 30 kt with a 1006 mb pressure was located (according to COADS) at 14.0N, 77.4W. It is possible that the given position is reported to be in error 5 degrees too far east. If the ship was actually at 14.0N, 82.4W, it would make a lot of sense given the analyzed position, because that observation would be close to the center. It is therefore recommended to keep the 35 kt intensity at 12Z on the 14^th. Based on the analysis of the ship previously believed to be at 77.4W instead at 82.4W, along with careful analysis of other data… the reanalyzed HURDAT longitudes from 5/14 12Z through 5/15 18Z are changed to 82.0W, 82.7W, 83.4W, 84.1W, 84.9W, 86.0W.

2. Do the synoptic data at 1200 UTC 19 May justify the proposed southward shift from the original HURDAT track?

After further analysis, although the synoptic data at 1200 UTC 18 May does just some slight southward shift from the original HURDAT track, perhaps it does not justify quite as large or a change as we made initially. A solution halfway between the original HURDAT and our previous analysis is appropriate. A latitude of 18.0N is chosen for the position at 12Z on 19 May and latitudes have been appropriately changed in the metadata.

1933 Storm #3:

1. What is the basis for shifting the track northward over Trinidad? The Monthly Weather Review article states the center went south of Trinidad, and there is no data shown from Trinidad that justifies such a change.

The analyzed latitudes from 6/27 12Z – 6/28 06Z have been adjusted slightly: 6/27 12Z: Moved from 9.7N to 9.6N; 6/27 18Z: Moved from 10.0N to 9.9N; 6/28 00Z: Moved from 10.5N, to 10.3N; 6/28 06Z: Moved from 11.0N, to 10.9N. Available data indicates that the storm still tracked a few tenths of a degree north of the original HURDAT positions.

2. Please provide a basis for increasing the intensity to 80 kt at 0000 UTC 28 June. Given the ship report of 60 kt and 986 mb, wouldn’t 70 or 75 kt be more appropriate?

Agreed that a 75 kt intensity is more appropriate at 0000 UTC 28 June. The intensity at 18Z on the 27^th has also been accordingly lowered to 75 kt. The reason we analyzed an intensity of 80 kt before is because we assumed that since the 986 mb was simultaneous with 60 kt, that the central pressure was about 980 mb (if the ship was inside the RMW) and possibly significantly less than 980 mb (if the ship was outside the RMW). A 980 mb central pressure equals 78 kt according to the Brown et al. southern pressure-wind relationship.

3. Can additional observations be found for Trinidad, northeastern Venezuela, the Netherland Antilles, or Cuba by contacting the meteorological services of those countries? In Trinidad and Venezuela, the rarity of a hurricane strike suggests they would have some information on this special event.

Each of the weather services for these countries were contacted. Trinidad has no records before 1946.

4. Since Venezuela is a large country, it is a good idea to specific in the metadata descriptions which part of the country the storm was over. This is done in some places but not in others.

Ok, a more detailed description has been added as to location impacted.

5. The intensity is decreased from 80 kt to 60 kt during the passage across Trinidad and/or Venezuela. What is the basis for this, and can it be re-examined pending the resolution of point 2?

Instead of showing a decrease in intensity from 80 to 60 kt, a decrease from 75 to 65 kt is now shown. The intensity at 12Z on the 28^th is increased to 65 kt because there is no evidence that it continued to weaken even after the interaction with land was complete.

6. On 29 June – 1 July, the metadata summary mentions the ship reports of hurricane-force winds and pressures of 985/982 mb. However, how this translates to a 85 kt intensity is not adequately explained. Please clarify this.

The following has been added to the metadata: “Peripheral pressures of 985 and 982 mb at 11Z on the 29^th and 20Z on the 30^th, respectively, yield winds greater than 71 kt and greater than 75 kt according to the Brown et al. southern pressure-wind relationship. Both of these observations were peripheral pressures recorded simultaneously with hurricane force winds, so the central pressure was likely significantly lower than these values recorded by the ships. An 80 kt intensity is analyzed for the 29^th at 12Z and a 85 kt intensity is chosen for 18Z on the 30^th.”

7. Does Perez have any hard data from Cuba during the passage of this cyclone?

Nothing else, except that there were 22 deaths in Cuba and $4 million in damage. Also of interest from a Mike Chenoweth email from July 13^th St. Kitts-Nevis Daily Bulletin from telegram from Kingston, Jamaica: “Recent tropical storm which did heavy damage in Trinidad, Mexico, Cuba, also struck Grand Cayman destroying all cultivation. Four schooners wrecked. No lives lost.”

8. There are small changes from the original HURDAT track on 2-3 July. Are these supported by the data, or were these made for a smoother track? Please clarify this.

These were made for a smoother track.

9. What is the basis of keeping the 95 kt intensity from time of the 965 mb ship report on 5 July until final landfall? The committee notes that the original HURDAT weakens the storm before landfall. Is there some basis for weakening it, or was this one of the interpolation errors that plagues the original HURDAT during old landfalls?

It’s tough to make that call. We have data from one ship with readings of 50 kt and pressures in the 990s. But there is not a lot of data to justifying weakening. There is also not a lot of data to justify changing the HURDAT intensity. Usually, the interpolations were performed only between points every 24 hours. Since it is 2 days in this case, perhaps a solution to lower it some will be appropriate. The intensities at 06-12Z on the 6^th have been lowered by 5 kt to 90 kt… and the intensities at 18Z on the 6^th – 00Z on the 7^th have been lowered by 10 kt to 85 kt. After rerunning the Kaplan and DeMaria Inland Decay Model with an 85 kt landfall intensity… 60 and 41 kt are yielded for 06 and 12Z instead of 64 and 45 kt, which were previously yielded with a 95 kt landfall intensity. However, the previously analyzed intensities of 60 and 35 kt for those respective times are unchanged.

1933 Storm #4 (combination of former storm 3 and 4):

1. The committee concurs with merging these two tracks together. However, since the original storm #3 was supposed to have made landfall in the Veracruz/Tampico area, any observations from those stations to show it didn’t would be appreciated.

No observations from Veracruz, Tampico, or any other Mexican stations are currently available to us within 1 week before or after July 19-20 from any source other than HWM.

1a. The committee notes that this track merger was originally proposed by the late Stan Wright back in 1989!

Wow, that’s certainly very interesting. Is there is reference that should be cited from Stan Wright?

2. Do the observations on 14 July truly support a closed circulation? Pressure falls of 2-3 mb in 24 hours do not automatically mean the circulation was closed at the time. Please re-examine the genesis time.

There are no observations of west winds south of the center of 14 July. There were also no ships south of the center on that day. Given that observations by the 15^th indicated a closed circulation, combined with the chance that this may have been a tropical depression as early as the 13^th given the observations, there is not enough evidence to change the HURDAT genesis time forward or backward in this case.

3. A ship reporting west winds 15 kt at 1200 UTC 15 July suggests the proposed position is too far west. Please re-examine this.

An observed SSE wind of 15 kt at 12Z on the 15^th is listed in COADS at 16.8N, 68.6W. Based on the west wind of 15 kt farther east, agreed to move HURDAT position slightly east of previously analyzed position from 69.0W to 68.8W at 12Z. The longitudes at 7/15 06Z and 18Z are accordingly moved east by 0.1 degree.

4. Is any information available from Jamaica for this system?

Yes. The system reportedly dropped 9 inches of rain in Kingston. The rainfall damaged several bridges and roads. Mudslides and overflowing rivers flooded several towns with knee-deep waters. Moderate winds downed several banana trees across the island. All of this information is courtesy of Wikipedia:

http://en.wikipedia.org/wiki/1933_Atlantic_hurricane_season#cite_note-july181933-7

And the writer of the wiki entry received their information from the 2 sources below:

· ^ The Daily Gleaner (1933-07-18). "Torrential Rains Create Havoc in Kingston and St. Andrews". Retrieved 2006-09-29.

· ^ The Daily Gleaner (1933-07-20). "Burst Rivers Flooded Roads in Villages". Retrieved 2006-09-29.

5. In reference to the 995 mb ship report on 18 July, please provide better documentation that the ship pressure has a 5 mb low bias. Rapid pressure changes by themselves do not mean the report was wrong. How does it compare to other nearby observations before it enters the cyclone?

On 19 July at 12Z, when that ship is at 16.1N, 87.3W and reported a 1007 mb pressure, a ship at 16.1N, 87.5W reported a 1013 mb pressure. A ship log for the ship of interest is plotted on the back on the July 18^th HWM map in the binder. We are not disagreeing that this ship’s pressure dropped by 13 mb in a 25 hour period from 11Z on the 17^th to 12Z on the 18^th, but we do think that the 995 mb is too low. Instead of dropping from 1008 mb to 995 mb, we believe the pressure on the ship fell from about 1013 mb to 1000 mb.

6. Are any observations available from Belize City? The committee notes that a set of Belize City observations is available for September and October.

None could be found.

7. The HWM data suggests enough temperature gradient developed across the cyclone to consider it extratropical on 24 July. Please re-examine the issue of whether extratropical transition occurred.

Agreed. The period 7/24 06Z until dissipation is now analyzed an extratropical, meaning that no changes to extratropical transition are made from what was in HURDAT originally.

8. What is the basis for increasing the intensity over land on 25-26 July from 20 to 30 kt? Please clarify this.

On the 25^th, a 25 kt intensity is analyzed because of a 20 kt ship observation in the northern Gulf of Mexico that is close enough to be considered part of the circulation. On the 26^th, a 30 kt intensity is analyzed due to observed winds of 30 and 28 kt, respectively, from Mobile, AL and Montgomery, AL.

1933 Storm #5 (new):

1. The committee concurs with the addition of this storm, pending a more detailed analysis of the temperature gradient across the system.

Isotherms of air temperature have now been drawn in 5F increments on the HWM maps for 24-26 July.

2. Can maps for before 24 July and 27 July be added to the binder?

Done.

1933 Storm #6:

1. The committee notes that the center must have gone very near Antigua, as the data from the adjacent island of Montserrat does not suggest the center passed over than island.

2. According to the newspaper supplement from Mike Chenoweth, St. Croix and St. Kitts both reported peripheral pressures of 1000 mb on 25-26 July (exact times are uncertain.) This is likely worth noting somewhere in the metadata.

Based on the committee’s comments #1 and #2 above, along with considering information supplied by Mike Chenoweth and performing a more detailed analysis again with all data being considered, some alterations are made to the positions at all times from 7/25 06Z – 7/26 06Z. 7/25 06Z- changed from 15.5N, 58.9W to 15.7N, 59.0W; 7/25 12Z- changed from 16.1N, 60.4W to 16.5N, 60.7W; 7/25 18Z- changed from 17.0N, 62.1W to 17.3N, 62.4W; 7/26 00Z- changed from 17.8N, 63.6W to 17.8N, 63.7W; 7/26 06Z: changed from 18.6N, 65.0W to 18.5N, 65.1W. The reason why the track was shifted northward at 12Z on the 25^th is because Antigua recorded a minimum pressure of 1002 mb whereas Montserrat observations do not indicate that the cyclone passed close to that place. The track previously analyzed went halfway between Antigua and Montserrat. The new track is within one-tenth of a degree of Antigua- or almost on top of that island. The new track also shows that the storm passed about 10 nm north of St. Kitts, where a minimum pressure of 1000 mb was recorded, instead of right over them. The reason for the westward adjustment in the 12 and 18Z positions on the 25^th is because Antigua recorded their min pressure at 15Z and St. Kitts recorded their min pressure at 18Z, so the positions at 12Z and 18Z must also be moved west in addition to being moved north. One of the pieces of information from Mike Chenoweth states that the storm passed between St. Croix and St. Thomas. Although St. Croix recorded a minimum pressure of 1000 mb, the cyclone is analyzed to have been strengthening at that time, so that observation should be farther from the center than the 1000 mb recorded at St Kitts the day before. A 52 kt wind was recorded at St. Thomas. Besides the commentary, there is no concrete evidence in the terms of wind direction information that would indicate that the center passed south of St. Thomas. The 06Z position on 26 July is moved southwest by one-tenth of a degree to show that it passed only one-tenth of a degree north of St. Thomas instead of two or three tenths of a degree. No intensity changes have been made.

3. The 74 kt wind at Turks Island seems to be estimated. Is it known whether this was an estimated sustained wind or a gust?

Nowhere in the commentary was the word “gust” used. The quote is: “The lowest barometer reading at Turks Island was 29.37 inches, accompanied by a wind velocity estimated at 85 miles per hour.” One might assume that’s an estimated sustained wind. But little weight is placed on the quantitative value of estimated hurricane-force winds – regardless if it was supposed to be a gust or a 1 min wind, because of the huge uncertainties in such visual estimates of extreme winds.

4. The committee notes that before the cyclone hit Florida even the ships that reported hurricane force winds were reporting relatively high pressures. Is the data definitive enough to say that the current and proposed HURDAT intensities from 28-30 July are a little too high?

After Turks Island recorded a minimum pressure of 995 mb and estimated maximum winds of 74 kt on 27 July, on the 28^th, a ship recorded a minimum pressure of 1002 mb and encountered maximum winds of 60 kt. It is tough to say for sure how close this ship was to the center, but peripheral observations on the 28^th through the 30^th indication that the OCI and environmental pressure were very high with this cyclone, and the cyclone was also very small and likely compact. On the 29^th, the peak ob was an HWM ob – a ship that reported 45 kt with 1007 mb at 12Z. Unfortunately, the peak observations from this ship are not available as only a 12Z ob was provided. There was only one observation of a hurricane force wind from genesis through Florida landfall (aside from the estimated wind at Turks Island) - a ship at 09Z on the 30^th reported hurricane force winds simultaneously with 998 mb. On the 30^th at 12Z, a 45 kt wind with simultaneous 992 mb pressure means the central pressure was about 988 mb (if the ship was inside the RMW – which is not certain), which would yield 62 kt according to the Brown et al. north of 25N pressure-wind relationship. But the environmental pressure was high and observations indicate the storm was small. For the Florida landfall, which is analyzed at 16Z on the 30^th, besides the observation from Fort Pierce, there are not many available observations from the stations in Florida. However, since there were about 20 observations within 100 nm of the center from the 28^th-30^th, there is enough evidence to lower the previously analyzed intensity by 5 to 10 kt from 18Z on the 27^th through the Florida landfall at 16Z on the 30^th. It is noted that the number of reporting stations on the east coast of Florida in 1933 were few and far between within 50-75 nm of where this cyclone made landfall. Only the 12Z ob is available from West Palm Beach. The next observation site north of that is Fort Pierce, and the one north of that is Titusville. There is not enough evidence to remove this as a hurricane for Florida. However, there appears to be enough evidence to decrease the landfall intensity to 65 kt. The intensities from 18Z on the 27^th to landfall are changed to the following (kt) in the format: new analyzed intensity (previously analyzed intensity): 27/18Z: 75 (80); 28/00Z: 75 (80); 28/06Z: 75 (80); 28/12Z: 70 (80); 28/18Z: 70 (80); 29/00Z: 70 (80); 29/06Z: 70 (80); 29/12Z: 70 (80); 29/18Z: 70 (75); 30/00Z: 70 (75); 30/06Z: 70 (75); 30/12Z: 65 (70); LANDFALL (30^th/16Z): 65 (70). In addition, the landfall central pressure is changed from 985 to 988 mb. A 988 mb central pressure is also added to HURDAT at 12Z on the 30^th because it is assumed that the ship that recorded 992 mb with 45 kt at 12Z was inside the RMW (it was 8 nm from the analyzed center position – this value of 8 nm is uncertain and should not be used as a guess of the size of the RMW).

5. Please review the location of the Florida landfall. The ship report at 1200 UTC 30 July may not totally support the proposed southward shift in the track. On a related noted, can it be determined how close Fort Pierce was to the landfall RMW? The maximum winds were only 52 kt, which suggests the combination of landfall location and RWM size kept the RWM away from the station. This could also feed back on the landfall location.

The track of Connor (1956) as well as the MWR track show a landfall location near 27.0N. Willing to move landfall location back to 27.1N, but not farther north than that because believe that Fort Pierce did not experience the RMW (RMW size is unknown, but is likely not very large- probably either average or smaller than average). The latitudes at 12 and 18Z on 30 July are moved 0.1 degree N from our previous analysis and the same is true for the landfall point at 16Z. The main reason for this is we are edging it back slightly closer to the original HURDAT location because it was determined that there is only enough evidence that is passed 0.3 degrees south of Fort Pierce and not likely more than that.

Based on the answers to comments #4 and #5, a re-run on the Kaplan and DeMaria inland decay model using a 65 kt landfall intensity yields 55, 50, 45, and 37 kt for 18Z on the 30^th, and 00, 06, and 12Z on the 31^st, respectively. The 60 kt intensity that we previously analyzed at 18Z on the 30^th is unchanged because this was only 2 hours after landfall, and it is possible that the back side of the RMW was just moving inland. The only additional intensity change over FL is to lower the intensity at 00Z on the 31^st from 55 to 50 kt.

6. The committee has concerns about the final landfall intensity in Mexico based on the Brownsville data. It is unclear from the available data whether Brownsville was inside the RMW, and it is noted that the maximum wind was from the northwest – a direction that looks a little strange given the proposed track in reference to Brownsville. Please get the data from the Brownsville OMR to better assess the landfall characteristics of this storm.

The Brownsville OMR is unavailable. After re-assessing all of the available data along with recommendations and analyses from Ho et al. (1987), Schwerdt et al. (1979) and Connor (1956), it is decided to make no additional changes to any of the landfall parameters for this last landfall. According to Schwerdt et al. (1979), the landfall point was 25.7N, 97.1W and the system was moving towards the WSW or 250 degrees at landfall. The closest approach to Brownsville, according to the reanalyzed track, which uses the same landfall point as Schwerdt et al., is 20 nm. According to Schwerdt et al., the lowest pressure recorded at Brownsville was 981.4 mb, the RMW was 25 nm, and Brownsville experienced the RMW. According to Ho et al. (1987), the central pressure of 975 mb was computed based on the 981.4 mb min pressure at Brownsville and other data they may have had available for Brownsville.

Also, it is worthy to note that in the 1933 MWR page 244, it says the maximum wind at Brownsville was 66 mph N, whereas in MWR page 362, it says the maximum wind at Brownsville was 72 mph NW. In 1933, Brownsville had two regular reporting stations - a WBO and a WBAS. The WBAS (25.9N, 97.4W) was 4 nm east of the WBO (25.9N, 97.5W). It is uncertain whether both of these reports (the 66 mph max and the 72 mph max) were different reports from the same station or whether these two reports were the maximum winds experienced at both stations, respectively. The anemometer height at the WBO is 29m and the anemometer height at the WBAS is 10m. It is known that the 66 mph max report is a 5-min max and was experienced at the WBO. This converts to a 10m 1-min max of 56 kt. For the 72 mph report, it is uncertain which station it is at (anemometer height either 29m or 10m) and it is uncertain if the averaging time was 5-min or 1-min. It is possible that the 72 mph was the fastest mile wind at the same station the 66 mph max 5-min was recorded. It is also possible that the 72 mph report is the max 5-min wind at the WBAS station (where the anemometer height was 10m). If the latter scenario is correct, then the max wind experienced at the WBAS after converting to 10m 1-min was 66 kt. If the former scenario is correct, then the max wind recorded at either station after converting to 10m 1-min was 58 kt. The latter scenario (66 kt) appears more believable in this case.

1933 Storm #7 (originally #6):

1. The committee has concerns about the revised genesis time. While a west wind on 11 August supports an earlier genesis, there is no other evidence of a closed circulation for several more days. This includes an apparent lack of westerly winds in the Lesser Antilles, as highlighted by east winds at Grenada as the system passed. Please re-examine the initial genesis time and provide a better justification for the earlier start.

There are no observations south of the analyzed center location at 12Z on 12 and 13 August. On the 12^th, the system hadn’t reached the Windward Islands yet, and on the 13^th, it had already passed the islands. On the 14^th, although there is a 10 kt E wind about 160 nm SSW of the center (from an island off the coast of South America near 12N, 68.5W), that’s not quite close enough for evidence that a closed circulation did not exist. However, a 30 kt intensity is analyzed during that time, which was down from 45 kt in HURDAT originally. On the 15^th at 12Z, there is a 15 kt ENE wind 140 nm SSW of the analyzed center, but that was the closest observation south of the center on the 15^th. At 22Z on the 15^th, a ship near the center recorded a minimum pressure of 1007 mb with simultaneous 25 kt winds, and this ship reported a wind shift from SE to W. This data at 22Z on the 15^th indicates the likely existence of a closed tropical cyclone near that location. A 35 kt intensity and a pretty reliable position were analyzed on the 15^th – early 16^th. Given that there are no ship observations closer than 140 nm SW through SE of the center from the 12^th – late on the 15^th to confirm whether or not the circulation was closed, there is not enough evidence to delay the genesis time shown in HURDAT originally (which was at 06Z on the 12^th). As it is right now, the 10 kt WNW wind on the 11^th is 160 nm SSW of our analyzed position. We have now made a further adjustment to the latitudes from 8/11 12Z – 8/12 12Z so that the new analyzed position at 8/11 12Z is now only 115 nm NNE of the 10 kt WSW wind observation on the 11^th. The revised (previously analyzed) latitudes are: 8/11 12Z: 11.0N (11.7N); 8/11 18Z: 11.4N (11.9); 8/12 00Z: 11.8N (12.2); 8/12 06Z: 12.2N (12.4N); 8/12 12Z: 12.6N (12.7N). The genesis time at 8/11 12Z from our previous analysis is not changed.

2. If the data does support genesis on 11 August, there needs to be a daily metadata write-up for that date. Likewise for the extra day added to the life of the cyclone on 21 August.

Done.

3. Please note in the metadata that the ship reporting the wind shift from southeast to west on 15-16 August is found in the report table of the Monthly Weather Review in August. Also, given the significance of the observation, please reference it in the daily metadata section as well.

Done.

4. Is there any additional data from Cuba to help with the intensity analysis?

According to Perez’s reanalysis report, it maintained tropical storm intensity of 35 kt during its passage over Cuba.

5. There is reference to a 45 kt ship report in the 18 August daily metadata that is apparently not used for the best track intensity. Yet, there is no information in the metadata as to why. Please clarify this.

That ship, located about 3 degrees east of the center, is likely to be biased about 10 kt too high for wind. There are other obs around there, and none of them, including the ones that are closer to the center, have winds that strong. However, in weighting this somewhat more in the reanalysis, it is decided to keep the original HURDAT intensity of 40 kt at 8/18 12Z and 18Z.

6. Is any data available for 22 August?

On 22 August, Storm 8 is analyzed to be at 32.8N, 71.1W with a 105 kt intensity. This cyclone (Storm 7) likely dissipated on the 21^st. If it hadn’t dissipated on the 21^st, it would have been absorbed by Storm 8 on the 22^nd. The COADS has now been obtained for 22 August. Observations between 12Z on the 21^st and 12Z on the 22^nd indicate that although there is a slight chance that the low could have been closed until about 00Z on the 22^nd as a remnant circulation of this cyclone, the data is not definitive enough to extend the HURDAT track for 6 or 12 additional hours.

1933 Storm #8:

1. The committee concurs with the revised genesis and early portion of the track.

Ok.

2. Please note somewhere in the metadata that the log of the encounter of the ship Tuna on August 18 (947 mb) is available in the 1933 data binder. Is it possible to estimate an RMW from this data?

That the ship log for Tuna is inside the 1933 binder has now been mentioned in the summary paragraph in the metadata. The ship recorded a central pressure of 947.5 mb (with calm) at 1730Z on the 18^th at an estimated position of 23.0N, 54.5W. This ship had hurricane force winds from 14Z-17Z, calm at 1730Z, and hurricane force again from 18Z-22Z. It’s too difficult to estimate an RMW from the ship data both the timing of the RMW is too uncertain and the duration of the calm is unknown. A central pressure of 948 mb equals 115 kt according to the southern-intensifying subset of the Brown et al. pressure-wind relationship. So the 115 kt at 18Z on the 18^th is unchanged. Previously, although the 948 mb pressure on this day was available, it was not treated as a central pressure nor was the time entered. Previously, a 45 kt intensity at 8/14 12Z and a 115 kt intensity at 8/18 12Z were analyzed, and interpolated to obtain the intensities between these 2 times, showing a very gradual strengthening. Now, the same thing is done except to go from 45 kt at 8/14 12Z to 115 kt at 8/18 18Z. Therefore, all intensities from 8/15 18Z to 8/18 12Z are lowered by 5 kt from what was analyzed before. These new intensity analyses are all 5 kt closer to the original HURDAT intensities.

3. Given the lack of core observations before 18 August, perhaps the development rate could be adjusted slightly to match Dvorak’s one T-number per day?

That sounds good. This has now been done (so a large portion of our response to #2 has been further changed). 115 kt at 8/18 18Z corresponds to a Dvorak T-number of 6.0. A 5.0 at 8/17 18Z, a 4.0 at 8/16 18Z, and a 3.0 at 8/15 18Z corresponds to 90, 65, and 45 kt, respectively. 90 and 65 kt are chosen at 18Z on the 17^th and 16^th, and 50 kt is chosen for 18Z on the 15^th, since we analyzed 45 kt at 8/14 12Z.

4. The 23 August metadata says that “Norfolk (VH) has 76 kt winds, while the metadata summary says Norfolk had 61 kt 1-minute winds. Please resolve this discrepancy.

The max 1-minute wind listed in the Norfolk, VA OMR is 61 kt. The metadata write-up indicates that there must have been another observation where 76 kt was reported. There are two sources found that list wind speeds in this range. One of them in the Monthly Aerological Record from the Norfolk, NAS, which states that the maximum gust was 84 kt there. The other is from Chronological List of Tropical Storms that have Affected the Hampton Roads Area of Virginia (Virginia Hurricanes) which states that the maximum wind at Cape Henry, VA was 71 kt NE.

61 kt NE (max 1-min) (62m AGL) at 8/23 ~1130Z at Norfolk [WBO] (36.8N, 76.3) (OMR, MWR). This equals 53 kt NE at 10m.

71 kt NE (max w) (16m AGL) at Cape Henry, VA (CHY) (36.9N, 76.0W) (Virginia Hurricanes). This equals 69 kt at 10m. Unsure whether the reported value of 71 kt is a max 1-min or max 5-min wind. If it was a 5-min wind, one would need to multiply 69 kt * 1.06, which equals 73 kt. The true max 10m 1-min wind at Cape Henry was somewhere in the range of 69-73 kt, but it was essentially about 71 kt NE.

84 kt NE (max gust) at 8/23 0915Z at Norfolk NAS (36.9N, 76.3W) (Monthly Aero Record)

Originally, it was listed in the metadata and the Excel spreadsheet that the source “Virginia Hurricanes” listed a max wind for Norfolk of 76 kt. Actually, the observation in “Virginia Hurricanes” says that a max wind of 71 kt was recorded at Cape Henry and the max wind at Norfolk was 61 kt, just as in the other sources. The observation in the metadata is changed from 76 kt (max w) at Norfolk to 71 kt (max w) at Cape Henry.

5. What is the basis for the 75-80 kt intensities at the U. S. landfalls? Is it the observed winds on the weaker side of the hurricane and the likelihood that stronger winds existed elsewhere? If so, please state this.

In response to the answer for the previous comment (#4) on this storm, the highest observed wind on land, after reducing to a 10m 1-min value was 71 kt NE at Cape Henry, VA. Although the time of this observation is not known, it likely occurred around 12Z and other observations indicate the strongest winds around that time were northeast of the center. The analyzed intensity at 12Z is 75 kt. The 80 kt analyzed intensity at the 10Z landfall was chosen because the pressure-wind relationship at landfall yields about 88 kt, but the RMW was large and the environmental pressure was low. Also, there were no coastal land/station observations to the right of the storm track until Cape Henry recorded the 71 kt, which was likely a couple of hours after the first landfall in North Carolina.

6. What is the basis for the possible high bias in the Atlantic City anemometer? Please state it in somewhere in the metadata.

The Atlantic City anemometer is 52m AGL. After converting to 10m, the max 1-min 10m wind at Atlantic City was 58 kt. This has now been stated in the metadata. There probably was not a high bias- other than the reanalysis previously didn’t take into account the height of the anemometer. Even though the highest observed wind around 18Z on the 23^rd is now 58 kt (instead of 66 kt), and even though Kaplan and DeMaria inland decay model yields 52 kt for 18Z on the 23^rd, the analyzed intensity of 65 kt is maintained because Atlantic City was 160 nm NE of the center at 1730Z when it recorded its highest wind. The RMW probably became larger from the 40 nm RMW analyzed at landfall 8 hours earlier, but it wasn’t 160 nm. Stronger winds likely existed somewhat closer to the center.

7. The 25 August data on the HWM suggest the frontal zone may have been a little more southwest than actually analyzed. Is it possible that the extratropical transition was a little earlier than proposed, or does the rest of the data on that look good?

It appears as though temperatures are in the mid to upper 60s surrounding the cyclone. These observations are all in the early morning (12Z) and they are pretty far north. However, the uni-directional NE winds seen north of the center on the HWM map on the 25^th probably indicate that extratropical transition is near completion. After carefully reviewing the HWM maps from 25 and 26 August, the data indicates that the system likely became extratropical around 18Z on the 25^th, instead of 06Z on the 26^th, which we had previously analyzed. This has now been changed.

8. The 28 August HWM show an occluded front with a circulation at its northern end. Could this be the remains of the hurricane?

Yes, we analyzed a 12Z position on the 28^th at 44.5N, 52.5W.

9. Is there a better term than “intercoastal waterways” to describe the sounds of North Carolina in the metadata summary?

Yes. This has been changed.

10. In the last paragraph, is the section about the possible alternative intensities and the cyclone possibly being a Category 5 hurricane necessary? It would probably be better just to say that the date, time, and location of the observation are not known and thus it cannot be used to determine the best track intensity.

Once we realized that the 947 mb observation on the 18^th was a central pressure rather than a peripheral pressure, this sentence should certainly be deleted, and it has been. Too add to this, after decoding the Bermuda ship reports, a ship observation of 940 mb with simultaneous hurricane force was found at 21Z on 19 August at 28.3N, 60.2W from the ship Antiochia. MWR page 363 also indicates that this ship recorded a 940 mb pressure, which was the lowest pressure recorded during the lifetime of the cyclone. Since the central pressure was 947.5 mb on the 18^th at 1730Z, the cyclone continued to intensify because the central pressure on the 19^th at 21Z was likely well below 940 mb, and may have been near 930 mb. A peripheral pressure of 940 mb yields winds of at least 115 and 120 kt according to the Brown et al. north of 25N pressure-wind relationship and its intensifying subset. If the central pressure was 933 mb, this would equal 121 and 127 kt, respectively, according to the same pressure-wind relationships. A peak intensity of 120 kt is now analyzed from 06Z on 19 August through 00Z on 20 August (an increase in the winds by 35-40 kt above the original HURDAT).

11. Please check with Hugh Cobb, who done much work on this storm, to ensure he is comfortable with the proposed changes near the U. S. landfalls.

The reanalysis was provided to Hugh Cobb on Aug. 2, 2011 and he replied with his comments on Aug. 5, 2011. It is pasted below (in non-bold). Responses to Hugh’s comments are in bold.

Andrew and Chris,

I have looked over your write up and have the following recommendations and/or suggestions:

1 - Is the pressure of 971 MB from Quantico an actual MSLP and reduced to sea level or is it a station pressure? The reason I ask this is that Quantico is only about 50 nm south of Washington DC and DC recorded a pressure of 980 mb just a few hours later which suggests a quite rapid and perhaps uneven rate of filling for the tropical cyclone. I took this into consideration when coming up with the 973 mb pressure for 18Z and came up with a more even rate of filling.

Not sure, but most meteorological reports were good about listing the sea-level pressure. Even if it wasn’t an SLP, the elevation of Quantico is only about 4m above sea-level, so low that one would only need to add 1 mb at most to correct it to sea level. Also, it is assumed that this must have been a central pressure, and it occurred around 22Z. A station farther away to the right of the center- the Naval Proving Ground- at 38.3N, 77.0W, recorded 975 mb with 21 kt SE at 2030Z. The 00Z position is changed to 38.7N, 77.2W, so now the center is shown to have gone farther from Washington, D.C. The 18Z central pressure has been changed from 970 to 971 mb.

2 - Norfolk recorded their pressure of 971 mb at 1420Z with light winds which suggested a pressure of 969 mb at 1420Z. Thus I was too high on the central pressure for 12Z. However backtracking 2 hours to 12Z and interpolating from the landfall pressure of 963 mb at 10 UTC, perhaps a pressure of 966-67 mb is more appropriate for 12Z?

Agreed to go with a central pressure of 966 mb instead of 965 mb for 12Z central pressure. Would rather not go 967 mb because the hurricane was over water most of the time between 10Z-12Z, whereas after 12Z, it was over land.

Other than this, I am good with the Re-Analysis. You mentioned the possibility of hurricane force winds for Maryland, Delaware and New Jersey. Given the track of the storm placed Maryland closer to the RMW and higher winds may have occurred between the center and 160 nm out at ACY, Maryland appears to be the best candidate for Category 1 impacts as well.

Agreed. A Category 1 impact for Maryland is added to HURDAT.

1933 Storm #9:

1. On 23 August, the HWM shows a ship with a west wind north of Puerto Rico, and a southwest wind in the Virgin Islands. Could this mean that genesis occurred earlier than currently proposed, especially in light of the large westward shift of the genesis position? Even if this is not the case, there should likely be a daily metadata entry for 23 August.

Yes. Genesis is now indicated at 12Z on 23 August as a 25 kt tropical depression. A metadata paragraph for August 23^rd has been added.

2. Is any data available from Bermuda for this system?

No observations beyond the once daily HWM obs are available

3. What is the basis for making this system a tropical storm on 25 August? Please clarify this in the metadata summary.

It was made a tropical storm on 25 August because since there were no observations on this day closer than 150 nm from the center, there is not enough evidence to lower the original HURDAT intensity. Text has been added to the metadata.

4. The committee notes that no gales were observed until 28 August. Are the ship reports for 25-27 August close enough to the center to rule out the possible of gales on those days? If so, the time the system became a tropical storm may need revision. (Also see points 3, 5, and 6.)

On the 25^th, the closest ship was 150 nm from the center. On the 26^th, the closest ship report was 50 nm from the center. This ship reported 30 kt with 1010 mb at that time. That was the only 30 kt report on the 26^th. On the 27^th, the highest observed wind was 25 kt from a ship 25 nm south of the analyzed center position at 00Z. A 12Z, a 20 kt ship observation occurred only about 40 nm NNW or NW of the analyzed center position. Because a 30 kt wind was observed on the 26^th, and because a more intense storm is seen on the 28^th, the analyzed intensity of 35 kt on the 27^th is maintained.

5. In the 1933 data binder, there is a section marked “NCDC miscellaneous”, which contains a variety of reports. One collection is a group of ship reports transmitted to Bermuda during August and September. There is a report from 28 August from the S. S. Caroline Maersk sailing from Tampico to Germany which states at 1620 GMT it passed the center of a “tropical disturbance” near 34 deg 16 min N 66 deg 19 min W, which puts it very close to the proposed track of the storm. The ship reports winds S by E force 10 (50 kt), seas very rough and confused, barometer 749 mm (998.6 mb), heavy showers, and no visibility. This observation suggests that while the 28/18Z position is reasonably good, the intensity needs some revision upward.

The following observation is added to the metadata and Excel spreadsheet:

50 kt S around ~1620Z at 34.3N, 66.3W

999 mb around ~1620Z at 34.3N, 66.3W.

It appears as though there is a chance that 999 mb was a central pressure and that the 999 mb pressure and the 50 kt winds were not simultaneous observations. A central pressure of 999 mb equals 45 and 50 kt according to the Brown et al. (2006) and the Landsea et al. (2004) pressure-wind relationships, respectively. A 50 kt intensity is chosen for HURDAT at 12 and 18Z on the 28^th. A 45 kt intensity is chosen for 06Z on the 28^th. A 40 kt intensity is chosen for 00Z on the 28^th. The reason why such a fast intensity increase from 35 kt at 18Z on the 27^th is chosen is because data on the 27^th indicates a weak cyclone with an intensity likely not above 35 kt. Since we chose 50 kt beginning at the 29^th at 12Z due a ship report of 45 kt with 1001 mb, the 50 kt intensity is now analyzed from the 28^th at 12Z through the 30^th at 06Z.

A central pressure of 999 mb is added to HURDAT at 18Z on the 28^th.

6. This section contains several more ship reports that likely pertain to this storm, but they are written in an old marine weather code. A copy of the code is needed to decode them.

After decoding the additional ship reports pertaining to this cyclone, the only important observation found occurred at 8/26 2030Z at 28.5N, 69.3W – 30 kt SW with 1008 mb. Based on this observation, the positions at 8/26 18Z and 8/27 00Z were changed from the revised positions just very slightly by 0.1-0.2 degrees so that the cyclone wasn’t centered southeast of a SW wind observation.

1933 Storm #10:

1. The committee notes that the new proposed genesis point on 25 August is over Belize. Also, is the circulation actually closed on 26 August? Please re-examine the genesis based on these issues.

After a more careful analysis, it has been determined that there is not enough evidence of a closed circulation on the 25^th and early on the 26^th to justify changing to the timing of genesis from the original HURDAT. Therefore, the timing of genesis in the original HURDAT – 18Z on 26 August – is retained as the revised genesis time.

2. Two typographical errors in the daily metadata – “sunsettled” in the 26 August and “72 mph” (instead of 22) in the 28 August.

These typos have now been fixed.

3. The committee has concerns about whether this system actually reached tropical storm strength. There are no reports of gales, and while the central pressure is low, so are the environmental pressures. Are there enough ship reports to state conclusively that there were no gales associated with the system? (A first glance at the HWM suggests there are *not* enough.) If so, you might want to consider downgrading this to a depression. If not, you might want to add a statement to the metadata summary stating the possibility that this system did not actually become a tropical storm.

With a central pressure likely around 1002 mb and environmental pressures around 1008-1009 mb, and also taking into account the sparse observational coverage, it is likely that this cyclone contained 35 kt winds. Four observations at 12Z on the 29^th are closely inspected. These observations are:

15 kt SE with 1007 mb at 22.3N, 97.7W (SHIP) (***BAD OB***)

15 kt SW with 1005 mb at 21.2N, 97.0W (SHIP)

15 kt N with 1003 mb at 22.3N, 97.9W (Tampico)

25 kt SE with 1007 mb at 24.7N, 95.0W (SHIP)

After plotting the above observations on a synoptic map, it has been determined that the 1^st ob listed above is a bad observation. Given that the other three obs are the only observations during that entire day (plus one other ob at 14Z south of the center), along with the 25 kt with 1007 mb far from the center, there could have easily been stronger winds closer to the center on the right (north) side of the storm. We have no observations from the right (north) side of the storm except one observation from Tampico, which is a couple of hours before the cyclone makes it closest approach. There’s a huge area from 22.2-23.5N, and from 96W to the Mexican coast that could have experienced 35 kt winds between 12Z and landfall. Therefore, it is not recommended to change the peak intensity of 35 kt. A statement has been added to the metadata summary that there is a possibility that it was only a tropical depression.

4. While the southward acceleration shown in the original HURDAT is likely unrealistic, other cyclones in this region have had a southward component of motion near or after landfall. Please re-examine the last two or three track position to see if the proposed motion is justifiable.

After reanalyzing all available observations in the Excel spreadsheet and the HWM maps from the 29^th and 30^th, it has been determined that none of that data provides any aid in determining whether or not a southward turn occurred. Instead of showing a due a westward motion, a slight turn to the west-southwest will now be shown after landfall, not as large of an acceleration as originally shown in HURDAT.

1933 Storm #11:

1. The description of the storm structure given for the encounter with the ship Jamaica Pioneer on 30 August (Monthly Weather Review 1933 page 274) is that of a small and concentrated vortex. The eye passage described over Cuba seems to be significantly larger, as if the hurricane underwent an eyewall replacement cycle and/or implosive filling. Perhaps the proposed intensities on 30 August - 1 September could be re-examined in light of this?

After a reassessment of the data, and consideration of the committee’s comment above, the intensities at 8/31 18Z and 9/1 00Z are lowered by 5 kt from what we had before. In the response to comment #2 on this storm, the reasoning for keeping the analyzed intensities on 9/1 from 06Z-12Z is provided. However, the intensity at 9/1 18Z is increased to 100 kt after running the Schloemer equation with the 979 mb observed at Havana. Data and commentary suggests that Havana never experienced a lull. The reanalyzed track now takes the storm 15 nm north of Havana at closest approach. Data and commentary from a few hours earlier on the 1^st indicated that the RMW was likely in the range of 12-15 nm. If one assumes that Havana was at the RMW, at the time of the 979 mb minimum pressure, then the Schloemer equation yields a central pressure of 962 mb. But if one assumes that the RMW is 12 nm when the storm passed 15 nm from Havana, then the central pressure is 955 mb. A central pressure of 962 mb yields 99 kt according to the southern pressure-wind relationship, and 955 mb yields 106 kt. The 18Z point is changed from 90 kt to 100 kt. The pressure information, combined with the 82 kt winds recorded at Havana (on the left side of the storm) are reason enough to assign a 100 kt intensity. There is no data that indicates the cyclone was weaker. Perez analyzed a 110 kt intensity from 06Z-18Z on the 1^st. The new intensity of 100 kt splits the difference between the original HURDAT and Perez, and also is the best choice given the available data.

2. Are there any pressures available from Cuba other than the Havana report?

Yes. A 951 estimated central pressure (from Perez) at 12Z on 1 September is available based on an observed 959 mb minimum pressure observation at Isabela de Sagua, Cuba (22.9N, 80.0W). This note from Perez may suggest that two different methods of calculating the central pressure at that time yield 951 and 954 mb, respectively. The 954 mb value appears more reasonable, however, because Perez notes that the station was located 0.5*RMW from the center when it measured 959 mb… meaning that it was inside the RMW and halfway between the RMW and the center. A 954 mb central pressure yields 107 kt according to the Brown et al. (2006) southern pressure-wind relationship and 104 kt according to its weakening subset. Since the hurricane made its closest approach to Isabela de Sagua around 10Z according to the re-revised track, the 110 kt intensity that we had chosen for 06Z and the 105 kt intensity that we chose for 12Z seem reasonable. A 954 mb central pressure is added to HURDAT at 12Z on 1 September

3. As with storm #6, the Brownsville OMR is needed to better assess what happened at the Texas landfall.

The Brownsville OMR is not available. Given that the 20 nm RMW is near the climatological average of 17 nm RMW and that the forward speed of the storm was slightly slower than average (8 kt), the landfall intensity is chosen to be at 110 kt (right at that suggested by the Brown et al. north of 25N pressure-wind relationship). Below is some more information on the landfall:

According to Connor (1956), the tide at Brownsville (or probably at the coast near Brownsville) was estimated at 12 to 15 ft; tide at Corpus Christi estimated 8 ft.

According to the table in the 1933 MWR p. 363, 90 mph (78 kt) winds were estimated at Brownsville, TX.

Schwerdt et al.: 949 estimated central pressure; 97 kt calculated/converted max 10m 1-min wind; 20 nm RMW; 1012 mb OCI; speed 8 kt. Landfall pt 26.2N, 97.1W.

Ho et al.: 948.9 mb central pressure based on 950.6 mb at Brownsville; 20 nm RMW; 8 kt speed. Landfall pt 26.0N, 97.2W.

Dunn and Miller: “Major” – winds 101-135 mph, pressure 949-982 mb. 40 killed, $12 million damage.

1933 Storm # 12:

1. What is the basis for the 95 kt intensity at 0600 UTC 2 September? The ship data suggests a central pressure in the lower 970 mb’s. So perhaps 85-90 kt is more appropriate?

The two sources with information on the ship Gulfwing are MWR and Tannehill. Information from the MWR September table of gales does not indicate that the ship definitely passed inside the RMW. The ship’s maximum wind of 60 kt W was reported at the time of the minimum pressure of 978 mb, according to MWR. According to Tannehill, the ship “reported a barometer reading of [981 mb] and wind 80 mph.” Neither source contains any information with any indication that ship experienced a lull associated with the RMW. Therefore, the 95 kt intensity is retained. The wording in the metadata summary has now been clarified so that there is no confusion on this point.

2. In the back of the 1933 data binder, there is a description of this storm at Harbour Island taken from Neely’s book on major hurricanes of the Bahamas. It mentions that the pressure at Harbour Island fell to 27.90 inches (944.8 mb). Is this any way to assess the quality of this value, and if it is correct, should the intensity for this part of the track be adjusted upward?

New observations have been added to the September 3^rd metadata and Excel sheet:

“945 mb (min p) at 11Z at Harbour Island (25.5N, 76.6W) (Neely); 109 kt around or at 11Z at Harbour Island (Neely); 122 kt at Spanish Wells (25.5N, 76.8W) (Neely).

MWR states that a 30 minute calm was experienced, but it is not perfectly clear whether the 30-min calm occurred on Eleuthera or Harbour Island and if it occurred at the same location the 945 mb pressure was measured, although it is probably more likely than not that 945 mb was the central pressure or at least very near that value. Tannehill indirectly suggests that the 945 mb pressure at Harbour Island was a central pressure. After consideration of all data and accounts from all sources, it is believed that the 945 mb pressure at Harbour Island was a central pressure, and this value is added to HURDAT at 12Z on the 3^rd. A 945 mb is a central pressure equals 116 and 110 kt, respectively according to the Brown et al. (2006) southern and north of 25N pressure-wind relationships. It also equals 118 and 115 kt according to the intensifying subsets of those pressure-wind relationships. The storm was smaller than normal and the speed was 14-15 kt. From this, an intensity of about 115-120 kt should be selected. Since the original HURDAT has 120 kt, no change is made to the 120 kt intensity at 12Z. The intensities at 9/3 06Z and 18Z are changed from 110 kt 115 kt. Due to this newly considered information, the positions at all times on 3 September are adjusted slightly.

3. Is it possible to calculate an RMW from the description of the eye passage over Harbor Island?

If the middle of the eye passed over them, since the calm lasted 30 minutes, and the speed of the storm was 15 kt, the calm eye had to have been at least 7.5 nm in diameter, which is 3.75 nm for calm eye radius, which would be an RMW of roughly 6 nm for the lower limit of what the size of the RMW could have been. The geometric center of the eye likely passed slightly south of Harbour Island since the wind shifted from N to SE and our revised track shows the center passing just south of there. The climatological RMW for a 945 mb central pressure at this latitude is 16-17 nm. Since, the longest lull described in Neely was 30 minutes even though the geometric center likely passed over some part of Eleuthera, chances are that the RMW was smaller than the climatological value, but larger than estimate of 6 nm if one assumes that Harbour Island was in the geometric center. This data as well as the data mentioned in the response to comment #2 all supports that the original HURDAT intensity of 120 kt at 12Z on the 3^rd looks appropriate and should not be changed.

4. For the Florida landfall, the metadata summary states that the system had a small RMW and a small ROCI. Yet, the proposed landfall intensity is only 3 kt above the climatological intensity. Perhaps a 115 kt landfall intensity is better? On the other hand, is a Category 4 status consistent with the damage reports in Florida?

Considering that the 13 nm RMW is only less than 25% smaller than the climatological RMW, and the environmental pressure is somewhat higher than normal, but not drastically higher, 5 kt was added to the pressure-wind relationship, which yields 112 kt. This is a situation where retaining the original HURDAT, whether it said 110 or 115 kt would be most appropriate. Additionally, the damage reports are more consistent with a Category 3 hurricane. Since HURDAT originally lists as a 110 kt Category 3, and since there is not evidence to indicate it is more likely to have been a Category 4 than a 3, the 110 kt intensity is HURDAT is retained, keeping this as a Category 3 for southeast Florida.

1933 Storm #13:

After decoding the Bermuda ship reports, the latitudes on 9/13-9/14 are revised by as much as 0.5 degrees north of the previously analyzed positions due to a 978 mb pressure observation at 21Z on 13 September. Before decoding the Bermuda ship reports located in the binder, there were no pressures below 1000 mb on the 13^th or 14^th. Now that there is a 978 mb pressure, a more accurate placement of the position is possible.

1. If the Monthly Weather Review Statement that the center of the hurricane passed just west of Hatteras is incorrect, that should be stated somewhere in the metadata summary.

Done. According to the all of the numerous available observations of wind and pressure, it clearly passed east of Cape Hatteras. There is no data that offers a contradictory solution. One possible explanation for why numerous sources had the track of this hurricane passing west of Cape Hatteras could be because of the account that this hurricane was the worst ever experienced in the Morehead City/New Bern area up until that time. However, upon further inspection of the commentary, a meteorologist noted that most of the damage caused there was due to the storm surge rather than the winds. Given the revised track, and the shape of the coastline in that area, this makes perfect sense because the easterly winds north of the cyclone would have driven a high storm surge onto land in those places likely peaking around 06Z on the 16^th.

2. What is the basis for specifically stating that winds of 90 kt affected the North Carolina coast? What data or damage supports this value? (Also see points 2 and 3.)

The original reanalysis of 946 mb for the central pressure at 12Z on 16 September was likely too low. That value was chosen because of a report from Diamond Shoals of hurricane force with 954 mb at 1020Z. However, the winds at Diamond Shoals shifted from east to west, and at 12Z, it was reporting 20 kt W with 955 mb. The center (or at least the area inside the RMW very close to the geometric center had to have passed over Diamond Shoals. Diamond Shoals reported a minimum pressure of 953 mb. A better estimate of the central pressure at the 12Z closest approach is 952 mb, and this value is added to HURDAT at 12Z. Using the 10 kt per mb rule may not have been the right choice when we had hurricane force with 954 mb. There is some possibility that hurricane force winds were just beginning to decrease when the 954 mb pressure was measured. The other observations outweigh the option of assigning a 946 mb central pressure based on the 10 kt per mb rule. This central pressure of 952 mb at 12Z is now consistent with the minimum pressure of 957 mb observed at Cape Hatteras at 11Z, which was simultaneous with winds of about 35 kt from approximately the NNE inside the RMW. September 16^th at 12Z is also the analyzed time of the closest approach of the center of the hurricane to the U.S. coast. A 952 mb central pressure equals 103, 99, and 96 kt according to the Brown et al. (2006) north of 25N, north of 25N and weakening, and the Landsea et al. (2004) north of 35N pressure-wind relationships, respectively. The newly estimated RMW of 37 nm (see response to comment #3 below) is about 50% larger than the climatological RMW value of 25 nm. The speed of the storm was 8 to 9 kt, the analyzed OCI was 1013 mb (but Schwerdt et al. says 1017 mb), and the ROCI was a large 275 nm. Given all of those factors, a 90 kt intensity is chosen for 12Z on the 16^th. The parametric model from Schwerdt et al. (1979) was then run to obtain the highest wind speed experienced on the coast when the back RMW was touching the coast at a time between 12Z-18Z, and it yields 83 kt for highest winds experienced on the North Carolina coast, which is a Category 2, although it is essentially right at the borderline of the Category 1/2 threshold. However, given the impacts described in coastal North Carolina, it is more likely that Category 2 impacts were experienced there. Therefore, an 85 kt Category 2 impact is analyzed for North Carolina.

3. Do the data from Cape Hatteras and Diamond Shoals support a 40 n mi RMW? A different size of RMW would affect the expected winds on the North Carolina coast at the very least.

The data from Cape Hatteras alone indicates an RMW of greater than 34 nm. Cape Hatteras was inside the RMW for approximately 7.9 hr on 16 September from 0638Z-1430Z. The forward speed of the storm was about 8-9 kt. But Cape Hatteras was not in the exact center. Assuming that the RMW is a perfect circle, that the hurricane passed 12 nm east of Cape Hatteras, and that a 68 nm (34*2) chord passed over Cape Hatteras inside the RMW, this yields a calculated RMW of 36 nm. If the hurricane had an RMW of 40 nm, then the cyclone would have passed 21 nm east of the station for a 68 nm chord of Hatteras being inside the RMW to have occurred. Given the uncertainties, the RMW of this hurricane at the time of closest approach to Cape Hatteras was about 35-40 nm. Using the Diamond Shoals data to aid in the RMW calculation is problematic for observations after 1020Z because the position of the Lightship is unknown since it detached from its mooring. However, the data from Diamond Shoals is valuable in determining the 12Z position, and thus the RMW at 12Z.

4. The Monthly Weather Review and the Cape Hatteras data record both state that the 79 kt maximum wind was an estimate due to the anemometer being damaged. First, this should be explicitly noted in both the daily metadata and the metadata summary. Second, what impact does this have on the estimated intensity at closest approach and the estimated wind speeds over the North Carolina coast?

Done. This has now been stated in the metadata.

The anemometer at Cape Hatteras was damaged just before the highest winds were recorded in association with the back RMW of the cyclone. The highest wind measured by the anemometer in association with the first RMW – after converting to 10m and 1-min – was at least 64 kt around 0425Z on the 16^th (exact value uncertain due to the only once daily reporting of the fastest mile wind). The highest wind associated with the first passage of the RMW may have been as high as 70 kt, but probably not any higher than that. The highest estimated wind associated with the 2^nd passage of the RMW at Cape Hatteras after converting to 10m 1-min was 75 kt. We previously stated that this value was 79 kt. It shouldn’t matter too much since this wind was estimated anyway. According to the revised track and the path that the hurricane took relative to the coast, one might assume that the maximum wind experienced on the North Carolina coast should have only been about 5 kt higher than the maximum wind recorded at Cape Hatteras. The 83 kt obtained by the run of the Neumann model is 8 kt higher than the estimated 75 kt maximum at Cape Hatteras.

5. It is noted that the closest approach of the center to southeastern Virginia is about 60 n mi. As best as can be determined, do the 64-kt winds extend far enough out to the northwest to justify the VA1 status?

The closest approach of the center to southeastern Virginia is 50 nm at 18Z on the 16^th. A run of the Schwerdt et al. (1979) model to obtain the highest winds experienced on the Virginia coast (using a 40 nm RMW and an 85 kt intensity) at 18Z on the 16^th yields 69 kt winds for the coast of Virginia at its border with North Carolina. It is noted in Virginia Hurricanes that 76 kt NE winds were recorded at Cape Henry and 65 kt winds were recorded at Norfolk. However, in the MWR table of highest winds for the month, the highest 5-min wind at Cape Henry (anemometer 16m AGL) was 57 kt. This converts to a maximum 10m 1-min wind of 58 kt. In the same MWR table, the highest 5-min wind at Norfolk (anemometer height 62m AGL) was 49 kt. This converts to a maximum 10m 1-min wind of 45 kt. These values of 58 and 45 kt are much lower than the 76 and 65 kt listed in Virginia Hurricanes. Based on the revised track, the Schwerdt et al. (1979) model is run again to determine what the likely maximum winds would have been at Cape Henry according to the model. According to the reanalyzed track, the center was 64 nm from Cape Henry at time of closest approach (around 20Z on the 16^th). A 40 nm RMW and an 85 kt intensity are used. The model yields 59 kt winds for Cape Henry, just 1 kt higher than the 58 kt listed in MWR table (after converting to 10m 1-min). The accuracy of the model in calculating the winds at Cape Henry increase the confidence that this was a Category 1 for Virginia farther south since the model yielded 69 kt there. Based on that, a Category 1 impact is added to HURDAT for Virginia. However, the winds listed in Virginia Hurricanes to have been recorded at Cape Henry and Norfolk are disregarded.

6. Recent large hurricane at high latitudes have had maximum winds lower than their central pressures might otherwise suggest (e. g. Igor of 2010). In light of that, is it possible that the assigned intensities on 17 September are too high? Please re-examine this, and at the least provide a stronger justification for the increased intensities on 17 September.

Given the available data, more central pressure estimates should have been made at times after 12Z on the 16^th. At 18Z on the 16^th, a 954 mb central pressure is added to HURDAT based on 40 kt W with 958 mb at 17Z from the drifting Diamond Shoals Lightship observation which likely occurred inside the RMW. At 00Z on the 17^th, a 959 mb central pressure is added to HURDAT based on a 50 kt with 964 mb from a ship at 23Z located about 0.3 degrees from the center and analyzed to be inside the RMW since the winds had decreased from 60 kt before the minimum pressure was recorded to 50 kt at the time of the minimum pressure. At 18Z on the 17^th, a 965 mb central pressure is added to HURDAT based on information from two ships that indicate a central pressure of about that value. At 12Z on the 19^th, during the extratropical phase, a 965 mb central pressure is added to HURDAT based on good evidence of this central pressure from ship data. The intensity at all times from 9/16 06Z – 9/19 00Z are lowered by 5-10 kt from our previous analysis. From 9/16 18Z – 9/18 00Z, the intensities are lowered by 10 kt from our previous analysis. At 9/16 18Z, a 954 mb central pressure equals 94 kt according to the Landsea et al. (2004) north of 35N pressure-wind relationship. An 85 kt intensity is selected for 18Z. At 9/17 00Z, a 959 mb central pressure equals 91 kt north of 35N. An 80 kt intensity is selected from 00Z. At 9/17 18Z, a 965 mb central pressure equals 86 kt north of 35N. A 75 kt intensity is selected for 18Z. The last observation of a hurricane force wind occurred at 9/17 19Z. The last observed 60 kt report occurred at 9/18 00Z. A 70 kt intensity is chosen for 9/18 00Z. The landfall intensity for the Canadian landfall is 9/18 11Z is now analyzed at 65 kt, and 65 kt is also the 12Z intensity on the 18^th. The cyclone is analyzed to have weakened to 60 kt by 18Z on the 18^th. This 60 kt intensity is held through 12Z on the 19^th, mainly because of an analyzed central pressure of 965 mb at 12Z on the 19^th and because 45-50 kt winds were observed by ships.

7. How certain is the timing of extratropical transition, especially when the nearly coincident landfall in Nova Scotia is considered? Please double check this.

According to Environment Canada, the cyclone made landfall in Nova Scotia near dawn on September 18^th as a 60 kt tropical storm. The 60 kt intensity appears to be taken from the original HURDAT intensity, but I’m not sure what source they got the “tropical” part from.

8. The last paragraph of the metadata summary is rather poorly written. It needs to be re-written to better state the 946 mb pressure estimate at landfall and to avoid being repetitious about the center of the eye not crossing the coast.

Done. It has been rewritten.

1933 Storm #14:

1. In the 10 September metadata, please attribute the quote about the weather conditions over the Caribbean to the Monthly Weather Review. Also, please correct “Tula” to “Tela”.

Done.

2. The downgrade of the Yucatan landfall to a tropical storm requires more justification. First, the Monthly Weather Review states that the center went inland north of Paya Obispo (Chetumal). How far north? The old and proposed HURDAT tracks have it very close to the city, but how certain is this? Second, the committee notes that additional observations were located for Tampico for this storm. Can a similar record be found for Payo Obispo/Chetumal? The observation of 45 kt/999mb at this station at 1200 UTC 13 September does not prove hurricane status, but it is 10 hours after the estimated landfall time and thus does not rule it out. Finally, while the decrease in intensity to 60 kt at 1200 UTC 12 September looks reasonable, the subsequent lack of strengthening looks less so. Other recent storms in this area have intensified up to landfall.

A time series of observations from Chetumal, Mexico was plotted on the back of the September 12 HWM map. The report at 03Z on 13 September was 25 kt NW with 998 mb, and at 13Z, it was 25 kt S with 999 mb. In between 03Z – 13Z on the 13^th, there was a ten-hour period with no available observations from that station. Assuming that the system passed north of the station, given that the wind shifted from NW to S, the wind should have backed by 135 degrees. Pressure at Chetumal likely got lower between the 03Z and the 13Z observations, and it is likely that the center passed north of the station by 15 to 30 nm, but this is an estimate. Since we had adjusted the longitudes as well as the latitudes, our track is farther south than the original HURDAT track. This is partially based on a ship observation at 07Z on the 13^th of 40 kt ESE with 997 mb at 18.6N, 87.0W. According to the original HURDAT track, the cyclone passed the longitude of Chetumal at 14Z on the 13^th, but according to the revised track, the cyclone passed the longitude of Chetumal at 08Z on the 13^th- halfway between the time of the 03Z and 13Z observations. Although this data may provide somewhat more justification that the HURDAT Yucatan landfall intensity was too high – it perhaps does not provide enough justification that it did not make landfall as a hurricane. The intensity is 9/13 00Z and Yucatan landfall is changed to 65 kt (we had 60 kt previously).

3. Was the inland decay model run for the system while it was over Yucatan?

We did not previously perform a run of the Kaplan and DeMaria inland decay model for the Yucatan landfall because we analyzed it as a tropical storm. But now, a run of the Kaplan and DeMaria inland decay model has been included in the metadata.

1933 Storm #15:

1. What is the data that confirms the closed, but weak circulation in the Windward Islands? Is it the northwest wind at Port of Spain? If so, please note this explicitly in the metadata. The committee notes that the data log from Grenada does not show any west winds when this disturbance was near, so perhaps the track should be south of there?

The northwest wind at Port of Spain indicates that there is a possibility that a closed circulation exists (this has now been stated in the metadata). Therefore, the timing of genesis in HURDAT should probably not be changed. The observations from Grenada were overlooked previously during the analysis. Based on these observations, the positions of the TC at all times from 06Z on 16 September to 06Z on 18 September are adjusted south and/or east of the previously selected positions at those times by as much as seven tenths of a degree on the 17^th. The new track shows that the center of the depression passed just south of Grenada around 00Z on the 17^th. Since Grenada reported an observation at 22Z on the 16^th of an east wind of 9 kt, the system was likely very weak, which is why a tropical depression is analyzed.

2. On a related note, the original HURDAT starts this on 16 September despite the Monthly Weather Review saying no definitive center could be detected. Perhaps the genesis of this system could be delayed a day?

Based on our response to comment #1 (above) we feel that the NW wind at Port of Spain indicates the possibility of a closed circulation, and that the 16^th should not be removed from HURDAT; instead the intensity should be lowered to a tropical depression on that day.

3. Is there a possibility that some of the large westward adjustments on 19 September are too far west?

Yes, it is possible that our large adjustments on 19 September are too far west, as there is a lack of data on that day. We have shifted the 12Z position on the 19^th eastward by 1 degree from 76.7W to 75.7W. Due to this adjustment, the longitudes from 9/18 12Z – 9/20 18Z are also shifted east. Eastward shifts equal to or exceeding 1 degree from our previous analyzed positions were made from 9/19 12Z – 9/20 00Z. At 9/21 00Z, no change is made because the position is known well since a ship in the eye recorded a 929 mb central pressure at 0130Z on the 21^st.

4. Can an RWM be calculated on 21 September based on the available ship data and the known forward speed of the hurricane?

The forward speed of the hurricane at the time was 11 kt. The data from the ship Virginia suggests an RMW of at least 5.5 nm. Given that the wind shifted nearly 180 degrees, and also taking into consideration the uncertainty in the exact time the RMW was experienced by the ship, the RMW may have been 7-8 nm. The climatological RMW for this latitude and central pressure is 11 nm. Given this smaller than normal RMW, 5 kt should be added to the pressure-wind relationship to determine the intensity (133 + 5 = 138 kt). However, this hurricane was extremely compact. Before the ship reached the eye of the hurricane, the ship’s pressure decreased by 75 mb in 3.5 hours from 1004 to 929 mb. After the calm, the pressure rose by 75 in 3.5 hr from 929 to 1004 mb. Although the ship traveled a distance of about 35 nm during each of those 3.5 hr periods, if one assumes the ship was stationary, the radius of the 1004 mb isobar was only 38 nm. That is a pressure gradient of 75 mb over a distance of 38 nm. The radius of the 996 mb isobar is estimated to be about 23 nm. That’s a pressure gradient of 67 mb / 23 nm. The radius of the 969 mb isobar is estimated to be 6 nm. That’s a pressure gradient of 40 mb / 6 nm. Given this the extent of the hurricane is tiny in addition to the small RMW, 140 kt seems like the right choice for the intensity.

5. What is the rationale for lowering the intensity 5 kt every 6 hr from the peak intensity (140 kt) until landfall on Yucatan (120 kt)?

Since there is not enough data to determine the landfall intensity in Yucatan with any accuracy, this solution works best rather than letting the 140 kt intensity ride until landfall or keeping the original HURDAT landfall intensity of 90 kt, which would mean that a rapid weakening would have to be shown while the system is over water of 50 kt in a 24-hr period.

1933 Storm #16 (new):

1. The committee concurs with the addition of this system.

Ok.

2. The committee would like to see data from before 24 September to better assess the genesis of the system.

The COADS data and the HWM maps for 21-23 September have now been placed in the binder.

3. Does the 60 kt northwesterly wind on 26 September allow for the possibility the system was a hurricane at that time? In theory, that should not have been the strongest side of a tropical cyclone moving north-northeast.

That’s a good point. The intensities from 00Z-12Z on the 26^th have been increased from 60 to 65 kt, making this a new hurricane. For the ob of 50 kt with 996 mb, if this observation occurred inside the RMW (which is not known), then it would indicate a central pressure of 991 mb using the 10 kt per mb rule inside the RMW. A central pressure of 991 mb would equal 61 kt according to the north of 35N Landsea et al. (2004) pressure-wind relationship.

4. Is it possible extratropical transition occurred somewhat earlier than currently proposed, given the cooler temperatures being advected into the cyclone on 27 September?

Yes, observations at 12Z on 27 September indicate the system was already extratropical by that time. Extratropical transition is moved up by 12 hours from 18Z to 06Z on the 27^th.

1933 Storm #17:

1. Overall, the evidence for this system being a tropical storm is very weak, and it could be removed from HURDAT if better evidence for it being a tropical storm is not found. (See the following points.)

We decided to eliminate 9/26-9/30, but there is good evidence that it was a tropical cyclone from 1-4 October, reaching a peak intensity of 40 kt late on the 2^nd. We have therefore left in the 1-4 October portion of this tropical cyclone in HURDAT.

2. San Juan did not report gales in its Monthly Weather review record as this system passed by. Is any OMR available that could provide a minimum pressure?

There were no gales recorded at San Juan. An OMR is not available for San Juan to provide a minimum pressure.

3. On 28 September, San Juan or some other station on Puerto Rico had southwesterly winds on the Historical Weather maps, which is consistent with the proposed revised position. However, that map also shows a vorticity center south of the Virgin Islands closer to the original HURDAT position. Please clarify this situation.

There is no closed circulation associated with the vorticity center south of the Virgin Islands. Discussion has been added to the metadata about this.

4. What data is available from Santo Domingo?

We have once daily observations from Santo Domingo at 12Z each day. If the time series of these observations is the only data being considered, it is inconsistent with the original track and indicates the disturbance that was farther east was the more pronounced one. The wind shifts indicate that the vort max could have passed north of Santo Domingo late on the 30^th instead of south of the station early on the 29^th.

5. Does the data after September 30 allow smoothing of the rather severe kink in the proposed track?

Yes, the data allow for some smoothing to this kink in the track. Positions from 00Z on the 1^st through 00Z on 4 October are changed by as much as 0.9 degrees (from our previously analyzed position) on 1 October and 0.6 degrees from our previously analyzed position on 2 October to remove some of the kinkiness in the track. There is still a kink in the track, but it is now less severe.

6. The gale observations used to declare this system a tropical storm on 2 October are several hundred miles from the center. This suggests the winds may be due more to the gradient north of the cyclone than to the cyclone itself and that these obs are not sufficient to call this a tropical storm.

These gales are about 250-270 nm from the newly proposed position. The 1006 mb pressure observation on the 3^rd indicates it is possible that there might be some validity to those gales and that they might possibly be part of the system. The SW wind at Turks Island on the 2^nd indicates the system might be closed and located to the north or NNW of that island by a somewhat uncertain distance, but closer to Turks Island than where the gales occurred. In the area where the gales were reported, the synoptic pressure gradient does not appear strong enough to produce gales. For those gales to be real, there would likely have to be a low with <1006 mb pressure to the SE of the gales. Also, this system may have been a subtropical cyclone, as suggested by the large RMW.

7. Why is there a reference to Storm #18 for the 2 October gale observations? Are these winds even farther from that system than they are from Storm #17?

The reference to Storm #18 has been removed as it is not necessary.

1933 Storm #18:

1. The Historical Weather Maps show west winds and low pressures at Cabo Gracias a Dios on 30 September and 1 October, which are not consistent with a genesis position of 12N on 1 October. Please re-examine this and at the very least provide an explanation of why the Cabo Gracias obs were disregarded.

The positions from 10/1 06Z – 10/2 12Z have been adjusted. On 10/1 at 06Z, the revised position is moved from 12.2N, 80.5W to 15.3N, 81.4W

2. The committee has issues with the use of 976 mb as the Cuban landfall pressure in the metadata summary. The lowest pressure in Havana apparently occurred after the eye passed, which suggests that this value is not representative of the landfall pressure on the other side of the island. Indeed, it more suggests that the system deepened significantly after passing Havana into the Straits of Florida. Please re-examine this. Does Perez have any information, such as a more detailed time series of data from Havana?

After obtaining Perez’s recommended positions and intensities every 6 hours, it is noted that Perez’s 12Z intensity is 90 kt with a 973 mb central pressure for the Cuban landfall. Assuming the central pressure was 973 mb at 12Z, this yields a wind speed of 86 kt according to the Brown et al. (2006) southern pressure-wind relationship. It is decided to make the 12Z intensity 90 kt instead of 95 kt like previously analyzed. Perez lists a 976 mb central pressure at 18Z, as in HURDAT, and notes that a calm at Havana lasted for 7 or 8 hours while the system performed a small loop near or just north of Havana.

3. What is the justification for the 110 kt intensity at 1200 UTC 5 October, given that the 966 mb ship report suggests 96 or 93 kt winds? Even of the central pressure is near 960 mb that does not by itself support 110 kt. Please clarify this.

It appears as though there is not crystal clear evidence that the ship traveled inside the RMW of the cyclone. If the central pressure at 12Z on 5 October was 960 mb, this would suggest winds of 102 and 100 kt according to same pressure-wind relationships. However, since we aren’t certain that the ship was inside the RMW, the central pressure could have been a little lower – in the 950s by that time also. On the other hand, this system had become very large with low environmental pressure. It is agreed that a 100 kt intensity would be more appropriate for 12Z on the 5^th than the 110 kt that we previously analyzed.

4. On a related note, what is the basis for the peak intensity of 115 kt in the Bahamas? The 959 mb ship report suggests a central pressure in the lower 950’s. However, that by itself would not support an intensity that high. Please clarify this.

A time series of this ship derived from the MWR table of gales indicates that there is a chance that this ship passed just barely inside the RMW, but that the winds did not decrease below hurricane force. It is also possible that the ship was not inside the RMW. A run of the Schloemer equation, using an environmental pressure of 1007 mb, and assuming this ship was located at the RMW when it recorded its minimum pressure of 958 mb, yields a central pressure of 930 mb, which yields 124 kt according to the north of 25N Brown et al. pressure-wind relationship. However, since the data indicates it is possible that the ship had winds at the RMW of intense hurricane force, and then winds may have decreased to 70 kt or so inside the RMW, it is possible that the ship went a couple of miles inside the RMW where the winds decreased from perhaps 100+ kt to 70 kt. Given this possibility, the Schloemer equation is run again assuming that the ship reached a distance of 0.75 RMW from the center (1 RMW being the RMW). It yields a central pressure of 940 mb in this scenario, which yields 115 kt according to the Brown et al. north of 25N pressure-wind relationship. It is unlikely that the ship was closer than 0.75 RMW from the center. This estimate of 940 mb is probably the best estimate of the central pressure around 08Z on the 6^th. Since the system was large and with low environmental pressure, a peak intensity of a 110 kt Category 3 hurricane is more appropriate and is now utilized in HURDAT.

5. Given the lack of major impact in the Florida Keys, is a BFL1 status justifiable? Please do a better job explaining how this came about. Certainly saying that 70 kt winds affected the Keys without any supporting observations is getting too precise. The committee notes that the metadata summary is rather contradictory on this point, as it states “A tropical storm impact is analyzed for southeast FL and the Keys.” before talking about the BFL1 status.

Further investigation has been conducted. After converting to a 10m 1-min wind, the maximum recorded at Key West was 38 kt, and the maximum recorded at the Miami WBO was 30 kt. There were no available observations between these sites. According to the analyzed track, the place where the cyclone would have made its closest approach was Long Key in the Florida Keys. A run of the Schwerdt et al. model yields 79 kt for the winds at should have occurred at Long Key. However, the Schwerdt et al. model was run at the time of closest approach to Key West and yielded 64 kt for Key West… whereas the maximum winds there were only 39 kt. The Schwerdt et al. model was run again at the time of closest approach to Miami and yielded 53 kt for Miami… whereas the maximum winds there were only 30 kt. In terms of absolute numbers (kt), these winds recorded at these stations were 23 kt and 26 kt lower than the values yielded by the Schwerdt et al. model. In terms of percentage, the actual values recorded were only 57% and 59% of the wind value computed by the Schwerdt et al. model. Utilizing a similar ratio to reduce the 79 kt computed by the model for Long Key, it is estimated that winds of 55 kt occurred in the Keys at Long Key – only a tropical storm impact.

6. Is it possible that extratropical transition was complete sometime on 7 October?

On 7 October, there are no observations in the NW quadrant within a few hundred nm of the center. The obs located at least a few hundred miles NW of the center are cool- in the 50s. On 7 October, there is no evidence that the strongest winds are not coincident with the lowest pressures relatively close to the center. Since there is a possibility that temperatures NW of the center near the center might not be cold yet, it appears that there is not quite enough evidence to make this extratropical by 12Z on the 7^th. Extratropical transition at 00Z on the 8^th is retained.

7. In the 9 October daily metadata, please attribute the quote to the Monthly Weather Review. Also in the quote “central of the British Isles” should read “central west of the British Isles”.

Done.

1933 Storm #19:

1. Please justify the removal of 25 October from the proposed track. While the 1200 UTC Historical Weather Maps do not show a well-defined circulation, low pressures are already present. If the data cannot conclusively justify this change, please use the original HURDAT entries for 25 October.

Agreed. After a re-assessment of the data, there is not enough evidence to remove 25 October. The system is analyzed as a 25 kt tropical depression from 06Z-18Z on 25 October (down from 35 kt originally).

2. The committee has some concerns with the peak intensity. Some feel the original intensity is correct, while others think the proposed reduced intensities are correct or even too high. Please better justify the proposed choice of 80 kt. Can an RMW be calculated from the Jamaican data to help calibrate the intensity?

When the cyclone passed over Torrie, the max wind before the lull occurred 4 1/3 hours before the max wind after the lull. The wind shifted 180 degrees there in 5 1/2 hours (it shifted about 160 degrees in the 4 1/3 hours between peak winds on either side of the RMW). The speed of the storm was about 4 kt. This means the cyclone would have traveled about 17 nm in 4 1/3 hours. This means that the RMW was at least 8-9 nm. But the data indicates that the center of the storm still probably passed slightly east of the station. It is estimated that the RMW around 22Z-23Z on the 29^th was about 11 nm. The climatological RMW is 15 nm. This smaller RMW warrants going 5 kt above the pressure-wind relationship. However, due to the slow forward speed of the cyclone, this would cancel that out. Therefore, after considering both factors, it is probably best not to make any adjustment to the Brown et al. pressure-wind relationship when a central pressure was available at 20Z at Savannah-La-Mar. The 982 mb observed central pressure at 20Z from Savannah-La-Mar yields 75 kt according to the Brown et al. southern pressure-wind relationship. Thus 75 kt is chosen at 12 and 18Z (originally 85 kt in HURDAT) as well as at landfall at 20Z.

3. The Calabash Bay Station is at 17.9N 77.7W if the notes in the Jamaican report are correct. At 1200 UTC 29 October, the station was reporting northeasterly winds at hurricane force even though the best track position is 17.8N 77.8W, just southwest of the station. Should this position be nudged a tenth of a degree or two to the east?

Agreed. 12Z position is moved one-tenth farther east to 17.8N, 77.7W.

4. “The cyclone was likely smaller than average”. Was it? Please clarify this and better explain the relevance to the peak intensity determination (see point 2).

Please see response in point 2 (above).

5. Does Perez have any additional information on this system in Cuba? There is a concern that the landfall intensity in Cuba is too high.

Perez lists this cyclone as a tropical storm for Cuba.

6. Please better explain why the intensity was kept at 45 kt for most of 1-4 November. On a related note, is the temporary weakening from 45 kt to 40 kt on 1 November correct?

While over Cuba, the cyclone is analyzed to have weakened from 55 kt at 12Z on 31 October to 40 kt at 06Z on 1 November. It is analyzed that the cyclone strengthened again to 45 kt after emerging back over water. Therefore, no temporary weakening is shown while over water. 35 kt ship observations occurred at 12Z on the 1^st and again at 18Z on the 1^st. However, upon further inspection of the data coverage and rather low environmental pressures on the 2^nd and 3^rd of November, it is agreed to lower the intensity on these 2 days. The intensities from 12Z on 2 November through 18Z on 3 November are lowered by 5 kt from the previous analysis.

1933 Storm #20:

1. The committee has issues with the thermal structure of this system. The Historical Weather Maps show a frontal structure for the life of this system, and in this case the data suggests those features are at least somewhat justified. Please make a detailed analysis of the thermal structure of this system to better determine when (and if) it was actually a tropical or subtropical cyclone.

Isotherms were drawn in 5F increments at 12Z each day from 26-28 October. The analyses indicate that there was only a slight temperature gradient across the low (~5 degrees) on the 26^th and 27^th at 12Z, but on the 28^th at 12Z, the temperature gradient across the low is almost twice the magnitude as on the previous two days. On the 26^th and 27^th, no fronts would be reanalyzed as extending through or within a couple hundred nm of the cyclone; however, on the 28^th at 12Z, the cyclone is becoming extratropical, although the cyclone likely didn’t develop a warm front and cold front until later on the 28^th. Taking into account the 36F temperature gradient across the low on the 29^th at 12Z, along with that at 12Z on the 28^th, it was close to becoming extratropical, but may have not completed transition yet, the analyzed timing of extratropical transition of 18Z on the 28^th looks appropriate.

2. In the metadata summary, “the winds were oriented unidirectional on both sides of the front”. Please clarify what this means.

Wording has been changed/clarified in the metadata.

3. If the temperature data confirms this system did become a tropical or subtropical cyclone, it may be necessary to have a couple of extratropical low points at the beginning of the record – possibly before the current genesis time.

Ok. We have an east wind north of the center at 00Z on the 26^th, so we will begin the cyclone at this time as extratropical and 35 kt… becoming a 40 kt tropical cyclone at 06Z on the 26^th, because by 12Z, the temperature gradient is minimal. It was likely closed by 00Z and tropical by 06Z.

1933 Storm #21:

1. The committee concurs with the proposed changes.

Ok.

1933 Additional Notes:

1. In addition to the disturbances noted here, the June Monthly Weather Review (pg. 178) discusses a gale occurrence from 9-15N at about 25W on 26-27 June. Climatology is against this being a tropical cyclone, but this needs to be investigated.

The COADS data was obtained for 26-27 June from 5-20N, 16-40W. No closed circulation was found. Highest observed winds in this box on these 2 days from COADS were 25 kt and lowest pressure was 1009 mb. Although no closed circulation was found, there are a couple of observations around 06Z-12Z on the 26^th near 10N, 25W that show some turning of the winds with winds up to 25 kt and pressures as low as 1009 mb. There is clearly not nearly enough evidence from the COADS data and the info provided in MWR to warrant farther investigation of this suspect until more observations from more sources are obtained. Nevertheless, we have written a short paragraph about this suspect in the additional notes.

2. The committee concurs with the decision to leave out the disturbances of 1-6 June, 8-12 July, and 17-19 July.

Ok.

3. 1933 Former storm #7: The committee is in favor with removing this system, although there is a concern about the lack of data it passed through the Lesser Antilles. There is a data log from Grenada which if the committee is interpreting it properly suggests 40 mph winds and lower pressures on 15-16 August. Please re-examine this and the other Lesser Antilles data. The committee does not favor merging this system with storm #10, so if this system is removed please change that explanation to a possible alternative scenario.

Ok. After looking at the Grenada data, it is still decided to remove this system from HURDAT. However, the 25 kt intensity for 17 August is changed to 30 kt. The positions from 16^th-18^th are changed by up to half of a degree- most southwest of the previously analyzed position to show a track south of Grenada, and that it made its closest approach there just before 00Z on the 17^th. At 00Z on the 17^th, Grenada reported 31 kt SE with 1010.5 mb. Upon inspection of the wind reports for the entire month, it seems as though there is a high bias of about 10 kt in the reported wind speeds, because judging from climatology, one is not going to have winds above 20 mph at 13Z on 24 of the 31 days of the month and winds above 30 mph at 13Z 10 of the 31 days. It seems too high. The discussion about storm #10 has been removed.

4. 1933 Former storm #17: The committee is in favor of removing this system. However, it would like a re-examination of the data before doing this, paying particular attention to the distance of the obs from the two possible tracks. The data shown on the 27 September map in the binder is far enough from the center that it does not preclude the system being a tropical storm at that time. In addition, even if the more southward track through Central America is accepted (similar but south of that of Tropical Storm Matthew of 2010), there was possibly time for the system to have been a tropical storm over water before it made landfall.

On 27 September, the closest observation from the center to scenario #1 is 130 nm away – 10 kt S with 1010 mb at 12Z. Although the information on the 27^th is not enough to remove the system from HURDAT, observational coverage on the 28^th and 29^th is sufficient enough to remove the system from HURDAT. Given that along with that we often see cyclonic turning in that area of the SW Caribbean, it is still recommended to remove this system from HURDAT.

On 28 September, under scenario #1, there are 3 observations within 100 nm of the center, the closest being 55 nm from the center – 20 kt SSE with 1008 mb observation. Under scenario #2 on the 28^th, that same ship is 95 nm from the center, and that is the closest observation. On the 29^th, under scenario #1, there are about 3 obs within 100 nm from the center, the closest being 60 nm from the center. All 3 of the obs were winds of 5 kt with pressures 1008-1009 mb. Under scenario #2 on the 29^th, there is only 1 or 2 obs within 100 nm of the center – Tela, Honduras and also Belize City. On the 30^th, under scenario #2, the closest obs are: Frontera (110 nm), Rapachula (130 nm), Belize City (190 nm). Under scenario #1, it becomes an open wave on the 30^th.

The day that provides the best argument for the existence of a tropical cyclone is on the 28^th. If perhaps the one observation of 20 kt SSE with 1008 mb at 13.2N, 83.2W was too far west by 60-90 nm, then a position near 12.6N, 82.5W, would appear reasonable. Given that there are 3 observation within 100 nm- 2 of them being within about 55-60 nm from the center, and given that the highest measured winds are 20 kt and that you often see cyclonic turning in that area of the southwestern Caribbean when there is not a TC present, there is enough evidence to remove this system from HURDAT.