Best Track Committee Re-Analysis Comments for 1935
(Replies to the Comments in Bold Face and [Bracketed] – CWL March 2012)
1935 Storm #1 (new):
1. The committee concurs with adding this storm to HURDAT.
2. In the metadata summary, it says “A possible central pressure…”. What does “possible” mean? Please clarify if this is a central pressure or not.
[A ship with a 5 kt NNE wind and 1006 mb pressure is the basis for considering this a central pressure. “Possible” has been removed.]
3. The committee has concerns that the evidence is insufficient to call this a tropical storm on May 15. Perhaps it would be better to delay intensification until after the center moves north of the Dominican Republic?
4. (Richard Pasch) TC #1: The observations support adding this system to the database as a new tropical storm for 1935.
1935 Storm #2 (originally #1):
1. Is it possible that the track can be extended back another day or two? The committee notes that the Historical Weather Maps (HWM) for August 16 shows north winds at Barbados and Martinique and two ships near 10N 56W with southwest winds. The HWM for August 17 shows southwest winds at Martinique, south winds at Barbados, and north winds at Guadeloupe. This suggests a circulation was present east of the Lesser Antilles possibly as early as 16 August. Even if the track cannot be extended back, daily metadata should be put together for those days that includes these observations. (The map of 17 August is in the binder, but the map for 16 August is not.)
[After reexamination of the observations on the 16th and 17th, genesis for this system is extended back 48 hours.]
2. The committee has concerns that the intensity may be too strong on 18-19 August considering the lack of observations of hurricane-force winds. However, addressing that issue will have to wait until the genesis time issue is resolved.
[Agreed to further reduce down the intensity on the 18th. However, lack of sufficient inner core observations on the 19th preclude any further reduction of maximum winds on that date. Intensity had already been reduced by 10 kt at 00 and 06 UTC and 5 kt at 12 and 18 UTC on the 19th.]
3. Since the 955 mb observation on 21 August was clearly not in the eye, it suggests the possibility the central pressure was below 950 mb. Could the intensity at that time be stronger than 110 kt? (Richard Pasch) The observations from the ship Angelina early on 21 August, of a pressure of 955 mb in winds of hurricane force, suggests a central pressure that is probably in the low 940s. Recommend that the intensity be boosted even higher around this time, to Category 4.
[Agreed that the intensity should be raised and is now set at 115 kt on the 21st, which is the peak intensity for the cyclone.]
4. The committee has concerns about maintaining major hurricane strength on 24 August given the unclimatologically high latitude and the possibility that extratropical transition started that day. The data indicates a central pressure in the lower 950’s or upper-level 940’s mb if the 956 mb is correct, but at high latitude this doesn’t always produce 100 kt winds. Please re-examine this.
[Agreed to reduce down the intensity late on the 24th and early on the 25th. A ship reported 70 kt WNW winds with 956 mb pressure at 21 UTC on the 24th. This suggests winds of at least 93 kt from the Landsea et al. north of 35N pressure-wind relationship. As the system was near 42N at the time and that extratropical transition had begun, intensity is analyzed to be 95 kt at 18 UTC on the 24th and 85 kt at 00 UTC on the 25th. ]
5. What was the basis for delaying the time of extratropical transition? Please include it in the metadata summary.
[The structure of the cyclone at 12 UTC on the 24th was still fairly symmetric and two ships 200 nm north of the cyclone reported 79F air temperatures. The system still had another 18-24 hours of evolution before it became extratropical.]
6. Can the system be tracked as an extratropical low after 26 August?
[The cyclone became absorbed before 00 UTC on the 27th of August by a larger extratropical low that moved eastward across Canada. There is no evidence that the system had a separate closed low at 12 UTC on the 27th.]
7. (Richard Pasch) In the discussion of this hurricane, the changes to track and intensity were characterized as “minor”. However the increase in intensity late on 24 August in the reanalyzed best track from 80 kt, originally, to 100 kt is a change of 2 Saffir-Simpson categories is not a “minor” change.
[It is agreed that the 80 to 100 kt change should have been described as a major change, as at least 20 kt is the threshold for “major”. However, with the revision as requested above, the revised intensity at 18 UTC on the 24th is 95 kt, a 15 kt increase.]
1935 Storm #3 (originally #2):
1. In the 29 August metadata, it is mentioned that the HWM does not analyze a closed low. However, the northwest wind on the HWM near 23N 70W and a southerly wind near 22N 68W suggests a low is there even if the HWM didn’t analyze it. Please note this in the metadata.
[We now indicate a trough from the HWM in the daily metadata and that there is enough evidence that a low existed in the summary paragraph.]
2. The evidence for calling this system a tropical storm on 29-30 August is a little thin. Is it possible the intensity could be reduced some on those days? It is noted than on the 31 August HWM there is a report of SW 30 mph south or southeast of the center, which gives a higher confidence that the system was a tropical storm on that day.
[Agreed to begin the system as a tropical depression on the 29th and 30th. Agreed to reference the SW wind on the 31st for justification for tropical storm intensity on that date.]
3. The 1 September metadata states that the HWM low was near 23.7N 77.3W. It looks like a typo, as this is the same as the HURDAT position and the actually HWM center is closer to 22N. Please correct this.
4. Is the 924 mb central pressure measured at 2100 UTC 2 September close enough in time to 1800 UTC to be used as a central pressure in HURDAT? It seems unlikely in light of how quickly the system was intensifying.
[Agreed to remove this central pressure from HURDAT, as it was not representative of what was occurring at the synoptic time.]
5. The committee concurs with the increase in peak intensity to 160 kt. Since the Florida Keys landfall was at 0200 UTC 3 September, should this intensity apply at 0000 UTC as well as at 0600 UTC? Also, since there are a number of pressure readings from the Florida Keys, is it possible to perform a gradient wind speed calculation to help estimate the intensity?
[Agreed to indicate this intensity of 160 kt at 00 UTC on the 3rd as well. Unfortunately, there do not exist either the detailed time series of sea level pressure from a single station or enough simultaneous pressure measurements near the center of this tiny hurricane to be able to perform a meaningful gradient wind speed calculation.]
6. In this 4 September metadata, there is reference to a ship report of 35 kt and 941 mb at 0000 UTC near 26N 79.3W. This pressure looks very low given the distance from the center. Please check this report and remove it if it is incorrect.
[The report is indeed incorrect and has been removed.]
7. The committee does not concur with the assessment of a 960 mb central pressure for the landfall in northwest Florida. The metadata states this came from Ho et al. and is based on the 980 mb peripheral pressure at Egmont Key, Florida. First, the lowest pressure at Egmont Key occurred at 0300 UTC 4 September, which is 15 hours before landfall. Second, Cedar Key, Florida, which is much closer to the location of the landfall than Egmont Key in both space and time, reported a minimum pressure of 985 mb just before landfall. Please re-assess this landfall intensity, including finding wind data for the Cedar Key station to aid in interpreting the pressure data. (Richard Pasch) In comparing the re-analyzed best track to the original HURDAT for the passage of this, the infamous “Labor Day” hurricane, across the Straits of Florida, the Florida Keys and Florida Bay, the new track is more consistent with the known time of the arrival of the small eye over Craig Key. While I agree with the boosted intensities over the Keys and the extreme eastern Gulf of Mexico, shouldn’t the 985 mb pressure from Cedar Key (which was not in the center) be used to estimate the pressure at the second landfall in the Big Bend area of Florida rather than the Egmont Key observation, which was near Tampa? This would probably not support a pressure quite as low as 960 mb for the second landfall.
[Second landfall occurred near 22 UTC on the 4th northwest of Cedar Key, near 29.6N 83.4W. Ho et al. estimated a central pressure of 960 mb, based upon a peripheral pressure reading of 980 mb from Egmont Key about 15 hours before landfall. However, Cedar Key measured 985 mb at 1750 UTC just a few hours before landfall, but did not measure winds. The closest full weather station for the second landfall was at Apalachicola, which measured a minimum pressure of 997 mb at 2120 UTC and maximum 5 min winds of 25 kt from the NW at 20 UTC. Apalachicola was about 85 nm west of the point of landfall. Cedar Key’s 985 mb was measured about 20 nm east of the hurricane’s center, as the cyclone was moving northward about four hours before landfall. The central pressure at landfall is roughly estimated to 965 mb, based primarily upon the Cedar Key information. This suggests an intensity of 90 kt from the Brown et al. north of 25N pressure-wind relationship and 86 kt from the subset of weakening cyclones. The Ho et al estimate of about a 20 nm RMW is close to near average (23 nmi) from central pressure and landfall latitude climatology (Vickery et al), as is the translational velocity (9 kt) and outer closed isobar (1012 mb). Thus 85 kt at the second landfall in Florida is analyzed here, making this a Category 2 hurricane impact with a 10 kt increase in intensity compared to HURDAT originally in the 18 UTC slot. However, this does keep the Category 2 impact status for northwest Florida (“AFL2”) as originally found in HURDAT.]
8. Is there an OMR available for Raleigh, North Carolina, which is closer to the center track than Greensboro?
[While additional OMR reports are unavailable, the North Carolina Local Climatological Report was accessed. This did indicate that Raleigh experienced 996 mb on the 6th – the lowest for that station for the month. All of the stations available from the Florida, Georgia, South Carolina, North Carolina, and Virginia Local Climatological Reports have been plotted. Analysis of these indicates that the revised track needs no further revision. However, given that there were only five reports of tropical storm force winds from the 18 stations in the vicinity of the storm - the largest of which was 42 kt from Savannah – the intensities are lowered slightly late on the 5th and early on the 6th. Additionally, the Raleigh 996 mb sea level pressure likely was a central pressure, given the track of the cyclone, and it is added into HURDAT at 06 UTC on the 6th.]
9. Are there good data for the 80 kt intensity on 7 September as an extratropical cyclone, or is this based on maintaining what was previously in HURDAT?
[The 80 kt on the 7th as an extratropical cyclone is supported by a 60 kt N/964 mb around 12 UTC and a 70 kt WSW/979 mb at 17 UTC. The cyclone underwent rapid deepending as an extratropical cyclone on the 7th.]
10. In the 9 September metadata, 28.73 inches should be expressed in millibars.
11. Would the GA1 status need to include an inland designation?
1935 Storm #4 (originally #3):
1. The committee concurs with the proposed changes in the track and the genesis time. (Richard Pasch) Not sure that the surface observations of 30 August are that conclusive that a closed circulation did not exist over the extreme northern Yucatan Peninsula, so I recommend that the system be retained as a tropical depression in the best track for that day.
[Additional observations were obtained from Mexico from their daily synoptic maps. These do suggest that a closed – though weak – cyclone was in existence by 12 UTC on the 30th of August. Thus genesis is retained as beginning on this date, though significantly farther south than originally shown in HURDAT. The additional observations also indicate a position farther west on the 31st.]
2. The committee has concerns over the proposed intensity near Veracruz. The Veracruz area is well known for strong orographically-aided northerly winds, and while this occurs most often in the winter it is possible in tropical cyclones as well. Since the pressure in Veracruz was relatively high (1004 mb), the committee would feel more comfortable with a 45-50 kt peak intensity.
[The additional Mexican observations also uncovered 1001 mb at Puerto Mexico at 12Z on the 1st, which is a 10 mb drop from the 31st. This along with Veracruz observations is consistent with a high end tropical storm.]
3. On a related note, the Mexican data for Veracruz shows the wind report as a “maximum velocity”. Is this a peak sustained wind or a peak gust?
[These reports were for a 5 min averaging time.]
4. The 35 kt wind ob mentioned in the 1 September metadata is not in the surface observation table. Please add this.
1935 Storm #5 (originally #4):
1. The committee concurs with the proposed intensity changes on 23-24 September. Is it possible the system should start as a 25 kt depression?
2. Can it be determined if the 959 mb pressure at Cienfuegos, Cuba was in the eye or not? It is noted that the 28 September metadata summary states “PROVIDE MORE DETAILS FROM PEREZ”. Please include said details rather than that statement. J
[Perez has been contacted again to determine whether the 959 mb is an eye reading or a peripheral pressure reading. While waiting for a response, we are assuming that it is a central pressure.]
3. The committee concurs with the proposed change in the peak intensity. (Richard Pasch) The observation from the Ship A.C. Bedford of 70 kt west-northwest winds with a pressure of 986 mb at 0300 UTC 26 September seems to support Category 2 intensity around that time, so the decreases of intensity in the re-analyzed track at 0000 and 0600 UTC on that date do not appear to be justified. The increase of intensity to 120 kt on 29 September looks reasonable given the observations from Bimini.
4. The committee has concerns with the addition of hurricane impacts in southeastern Florida. Before doing this, it is recommended that local newspapers be checked for damage accounts consistent with those of hurricane conditions. (Richard Pasch) The change to show Category 1 impact in extreme southeastern FL, even in the vicinity of Elliot Key, is questionable given that: 1) the “visual” (not estimated, but read off the anemometer dial?) observation of 70 kt at Fowey Rocks is rather elevated and needs to be reduced to 10 m and, 2) that station is located several miles offshore.
[The Fowey Rocks lighthouse keeper’s observation of 70 kt is based upon a visual estimate of the sea state and not from an anemometer. Given the uncertainty in this visual estimate, and that it is at an offshore location a few miles from any coastal islands, and that there were no reports of significant wind-caused damage, the analysis of peak impact for southeast Florida has been reduced to a high end tropical storm.]
1935 Storm #6 (originally #5):
1. The committee concurs with the revised genesis time. Is it possible the system should be started as a 25 kt depression?
2. The committee has concerns that the proposed intensities are too high on 19-21 October. During this time, there are no reports of pressures below 1000 mb, no reports of winds above 40 kt, and the system has a large envelope with relatively low environmental pressures. Given these things, a more gradual intensification to hurricane strength would be better.
3. On a related note, what data is available from Jamaica for this system? Are there any anecdotal accounts of this system affecting Jamaica as a hurricane? It is recommended that the Jamaican Meteorological Service be contacted for whatever information they can provide.
[Newly available from the Data Rescue at the National Climatic Data Center, archives of the Jamaica Weather Report were consulted for both this system (and storm #5). The numerous observations and reports of impacts do indicate that the system made a landfall in eastern Jamaica as a tropical storm around 13Z the 21st, a few hours earlier than indicated previously.]
4. Is the evidence conclusive that the center actually made landfall in Cuba, or did it just brush the coast? What does Perez say about that?
[Perez considers this a hurricane impact for Cuba with the track either brushing the coast or barely making a landfall.]
5. It should be noted in the metadata that the 60 kt wind in Santiago is from an incomplete record, as the Monthly Weather Review (MWR) states the anemometer blew down after that observation.
6. The committee has concerns that even the proposed revised intensities over Central America are too high. Is there any information from recent landfalling storms in Central America that could be used to refine this part of the track?
[The two best analogues for this system during the last two decades – as far as track goes – are 2007’s Felix and 2005’s Beta. Felix made landfall with 140 kt, weakened to 85 kt in six hours, 50 kt in 12 hours, and 25 kt in 18 hours. Beta made landfall with 90 kt, weakened to 55 kt in six hours, and 20 kt in 12 hours. These were utilized to refine the intensity as the cyclone decayed over Central America, with even weaker intensities than earlier estimated.]
7. (Richard Pasch) I question keeping the system as a tropical cyclone until as late as 1200 UTC 27 October. Although the surface map for that date and time suggests a center near the Pacific coast of Mexico, it is dubious that the original surface circulation center could interact for that long with the mountainous terrain of Central America and remain intact. This was more likely a new center that was forming in the remnant disturbance.
[Agreed to dissipate the system over Central America by 12 UTC on the 27th, as originally indicated in HURDAT.]
1935 Storm #7 (originally #6):
1. The HWM indicate this system originated as a frontal low possibly as early as 25 October. If possible, please include best track positions for these days (to conform to current practice) along with a daily metadata write-up.
[After obtaining the HWM and COADS observation and analyzing each day back to the 25th, it does not appear that a closed low occurred until the 30th. These analyses are now included in the binder and additional dates are added to the metadata writeup.]
2. The committee has some concerns about the possible frontal nature of the cyclone even on 30-31 October. If possible, please provide an analysis of the temperature field near the center on those days to ensure it had become non-frontal on those days, and to better assess when the cyclone most likely became subtropical and then tropical.
[Unfortunately, on the 30th, very little temperature data is available near the cyclone. The few observations available do, however, suggest a small to minimal temperature gradient across the system. On the 31st, the more numerous data show uniform temperatures near 75F across the cyclone. Thus the system is retained as a tropical cyclone on both days.]
3. It is noted that the 35 kt ship report in the 30 October metadata is 300 n mi from the center if the position is given correctly. Is this ship the basis for the proposed intensity? If so, please re-examine the intensity.
[An observation at Bermuda of 30 kt N and 1017 mb at 12 UTC on that day is consistent with the pre-existing 35 kt in HURDAT at that time. ]
4. There is a typo in the 3 November metadata: “973 mb” should read “964 mb”.
5. The committee has concerns that the proposed 95 kt peak intensity may be too high given the baroclinic origin and large-size nature of the cyclone. While an increase over the original HURDAT intensity appears warranted, an 85-90 kt peak intensity might be better.
6. On a related note, the landfall intensity of 90 kt in Miami may also be too high. There are no wind observations that support an intensity this high, and the central pressure suggests the cyclone was filling before it reached Miami. Please consider using an 80-85 kt intensity for the Miami landfall and the possibility this was a category 1 hurricane. (Richard Pasch) The proposed increases in intensity in the revised best track on 3-4 November seem a little too much given the estimated central pressures, and the observed 973 mb and 977 mb pressures at Miami and Ft. Lauderdale respectively which suggests that the eye was rather broad at landfall in FL. Suggest an overall maximum intensity of 90 kt, with an intensity of 85 kt at landfall in Miami. This would still be consistent with a Category 2 landfall in southeast Florida.
[Agreed to reduce down the peak intensity to 90 kt and the intensity at landfall to 85 kt, with the mention that this may have been a Category 1 at landfall given the uncertainties.]
7. In regards to the newly-introduced extratropical transition on 8 November, are the data sufficient to determine that a closed circulation still existed? This is not clear either from the daily metadata or the binder map. If the circulation still existed, then the transition looks correct. If not, perhaps a revision to show absorption by the front might be best.
[While somewhat ambiguous, it appears that a closed circulation still existed at 12 UTC on the 8th. Thus the extratropical transition is retained.]
1935 Storm #8:
1. The committee concurs with the addition of this system.
2. The genesis of this system needs more analysis. The low is quite well developed in the 1200 UTC 3 November data. However, no low is present on the HWM for the previous day – just an open trough. Please re-examine the data to refine the genesis time.
[The HWM and COADS data were obtained and examined for the 2nd of November. The is no indication that a closed low developed on this date in the context of numerous observations in the area. Genesis is retained for 00 UTC on the 3rd.]
3. There were several surges of cooler air into the system during its lifetime, and at least twice it acquired the elongated structure more characteristic of a decaying baroclinic low. The data need to be re-examined to see if the system fell below subtropical/tropical cyclone criteria around 9 November and again on 13-14 November. If so, the best track should probably call it a low instead of a tropical depression.
[The first usage of “Low” in the Atlantic best track is very recent – Arlene in 1987. The key criterion that we use to differentiate tropical cyclones from (remnant) lows is the presence of organized deep convection. Given that we are unable to examine this aspect in the pre-satellite era, it would not be appropriate to use the “Low” designation. However, discussion is added into the metadata writeup that indicates that the system might have been given this categorization for a couple of days in the satellite era.]
4. (Richard Pasch) On the track map, the arrow signifying the end of the system’s life is placed on the wrong end of the track (it’s at the beginning). It is questionable that this system was a tropical or even a subtropical storm from 3-6 November as the surface observations indicate that a temperature gradient existed across the cyclone. By 7 November, however, the surface data indicated fairly uniform temperatures around the low. Suggest starting the best track on that date.
[The track for this system is being corrected. After re-examination of the thermal structure on the 3rd through the 7th, it appears that the main baroclinicity remains north of the system. Thus the tropical (or perhaps subtropical) status is retained on those dates.]
1935 Additional Notes:
1. System #1: The committee concurs with leaving this out of HURDAT. (Richard Pasch) Surface data does indicate that there was a TC over the northern Gulf in late August, but there were no available observations to support tropical storm strength. Agree with not including this system into HURDAT.
2. In the MWR article on storm #6, there is mention of a minor disturbance over the Caribbean Sea from September 23-26 (the article gives the month incorrectly) that eventually merged with the hurricane. The report states this system caused local gales in Puerto Rico and southern Santo Domingo. The HWM do not analyze a low, but they show what might be a trough/tropical wave over the Mona Passage on 25 September and near the Dominican Republic on 26 September. Please investigate this system.
[MWR, HWM, and COADS observations indicate that the system did not obtain a closed circulation and remained a tropical wave moving across the Caribbean before being absorbed by the hurricane. Strongest observations obtained were 25 kt. As it did not have a closed circulation and that the gale force winds were not confirmed, this system is not added into HURDAT.]
3. The MWR chart of lows shows a low pressure area over the northwestern Caribbean on 30 September and over the southeastern Gulf of Mexico on 1-2 October. The data in the HWM suggest this system is weak. Please check the COADS data for this system to confirm this.
[The Monthly Weather Review, Historical Weather Maps, and COADS observations indicate a trough or possible tropical depression meandered near western Cuba from September 30 to October 2. No winds greater than 25 kt were recorded. It is unlikely that this reached tropical storm intensity and is thus not added into HURDAT.]
4. The HWM show a low over the eastern tropical Atlantic on 2 October with pressures below 1005 mb. The HWM analyze a nearly stationary and weakening low in this area for several days after this. Please check the COADS data to see if there are any significant observations near this system.
[HWM, COADS, and MWR data indicate that the highest observed wind was 30 kt and lowest pressure was 1006 mb, both on the 2nd. The system likely was at least a tropical depression and may have been a tropical storm. But without direct evidence of tropical storm intensity, this system is not added into HURDAT.]
5. The November MWR mentions on page 329 a ship reporting force 11 winds and a pressure of 29.66 inches in the southwestern Caribbean Sea (13N 82W) on 24-25 November. The HWM do not analyze a low in this area, but the plotted data suggest a possible cyclone in the area. Please investigate this.
[Obtaining the Historical Weather Maps, COADS ship data and the MWR information provides a somewhat ambiguous picture. The western Caribbean was swept by a strong cold front on the 23rd and 24th of November, as the temperatures were 40 F in Tampa, 52 F in Progresso, Mexico, and 60 F in western Cuba on the 24th. Strong cold air advection was occurring along the east coast of Central America on both dates, which in itself would suggest that a tropical cyclone did not occur. However, the presence of a closed circulation on the 25th (and possibly the 24th) in the southwestern Caribbean and the 1005 mb/60 kt NNE observations from the ship suggest that some type of cyclonic system did exist. However, none of the other observations ever indicated more than 25 kt occurred. It is judged that the system was caused by a combination of forcing from cold air advection and funneling along the east coast of Central America. Thus it is likely not a tropical cyclone (or a subtropical cyclone) and thus is not added into HURDAT.]