Hurricane seasons reanalyzed: 1944-1953 + 1931-1935

[Ground Scouring]

This post contains old news, but since it has not been mentioned yet, I will be the first to bear the news. For fans of hurricane history, two big reanalysis bonanzas have come out from NOAA: the latest revisions to the best track for the 1931-1935 Atlantic hurricane seasons plus a gigantic, 851-page thesis on the 1944-1953 Atlantic seasons--spanning the very beginning of the modern reconnaissance period!--penned by hurricane researchers. For the first time, actual plane observations from early missions are being used in the Atlantic hurricane reanalysis effort.

A recent thesis by Andrew Hagen of the University of Miami, coauthored by Christopher Landsea and Donna Strahan (2012), has preliminarily reanalyzed the Atlantic hurricane seasons of 1944-1953. The published paper can be located here (with an attached appendix containing the revised track maps and changes here). If you wish to view the full text of the reanalysis, contained in the original thesis here, beware; though it contains all the storms and the revisions, it is fully 851 pages long, but is well worth even a partial read. Anyway, the published paper in the first link is the real meat to be dissected, as it distills all the important information stemming from the research. Just note that the research is not official until the NOAA reanalysis specialists review and approve the changes.

Among the highlights:

• 1950 was analyzed to have been much less intense than currently listed, with an accumulated cyclone energy index (ACE) of 210 rather than 243. Two major storms were downgraded. Additionally, the intense Hurricane Dog was downgraded from 165 kt (Category 5) to 125 kt. Notably, Hurricane King, with a tiny eye of perhaps less than 4 mi/3 nm wide, was upped to 110 kt--near-Category 4 intensity!--at landfall on Miami, FL, making it one of the most potent micro hurricanes to strike a major metropolitan area.
  
• From 1945 to 1950, South FL was hit by no less than four Category 4 hurricanes--three of them, in 1945, 1948, and 1949, currently listed as Category-3 strikes in the official list of U.S. landfalls, 1851-2011! If confirmed, the number of Category-4 landfalls would set a record for Category-4 landfalls in a single decade.
  
• Only two storms currently listed as a major U.S. landfall during the 1944-1953 period were downgraded: the 1944 Great Atlantic hurricane was lowered in every state and made a 95-kt landfall on Long Island, NY; the October 1944 TC was downgraded to a 90-kt Category 2 TC in SW FL, near Englewood. (Also, the 1947 Fort Lauderdale hurricane, while keeping its Category-4 South FL landfall, was reduced to a 95-kt Category 2 impact in LA and MS.)
  
• A new Category 5 hurricane was discovered in 1953: Hurricane Carol was upgraded from 130 to 140 kt over the open Atlantic Ocean, based upon a plane report of 929 mb along with an eye diameter of just 4 miles! (The 929-mb pressure was also the lowest reported by aircraft for this decade; the second lowest, 937 mb, was recorded in Hurricane Easy 1951, which was lowered in intensity from 140 to 130 kt.)
  
• One TS at U.S. landfall, Hazel 1953, was actually upgraded to a 65-kt/987-mb hurricane as it struck SW FL near Fort Myers.
  
• The October 1944 TC, upon striking W Cuba, was upgraded from 105 kt to 120 kt, based upon a (ship?) report of 937 mb, apparently in the eye near Havana. (The same TC later passed over the Dry Tortugas, FL, at 105 kt, where the eye reportedly lasted for two full hours with a central pressure of 949 mb.) In the Leeward Islands, Hurricane Dog 1950 was boosted from 105 to 125 kt, based upon a sustained wind measured on an island, west of the eye.
  
• A latte surprise: a new, 70-kt hurricane in the E Atlantic in October 1944 struck Portugal as a 50-kt TS, producing TS winds there and in nearby Spain--the first such occurrence in Atlantic-basin history. (Move over, Vince 2005.)
  

The other significant highlight is that the reanalysis of the 1931-1935 seasons has been released, with loads of interesting surprises.

The main dish:

• 1933 officially became, in terms of ACE, the most intense Atlantic season ever recorded, with an ACE of 259 beating the 250 of 2005. The rise was largely due to an impressive number of upgrades: the Cuba-TX TC (Storm 8) and the Caribbean cruiser (Storm 14) were upgraded to 140 kt in the Bahamas and the Caribbean, becoming two new Category 5 hurricanes; the two landfalling NC hurricanes were boosted to 120 kt over the open Atlantic, with the first, in August, found to have started near Africa and become a major hurricane four days earlier; and the unusual Trinidad hurricane in July became a TS two days earlier.
  
• New hurricanes discovered: a fishy TS in September 1931 was boosted to 65 kt; a totally new 65-kt fish was discovered in September 1933; and another, totally new, 80-kt hurricane was found off the East Coast in July 1934.
  
• The 1935 Labor Day hurricane was greatly increased in intensity from 140 kt to a whopping 160 kt/892 mb at landfall in the FL Keys--which, given its tiny 6-mi-wide eye and tight environmental gradient, was no surprise. The storm deepened 40 kt in the last 24 hours before landfall; in the five hours before striking Craig Key, FL, the storm deepened 22 mb, as a ship reported 924 mb in the eye--confirmation that the storm was rapidly deepening. (In other news, the other FL TC that year, the weird Yankee hurricane that hit Miami in November, was found to be 85 kt at landfall and 90 kt in the Bahamas.)
  
• The early 1930s were a monstrous time in the Caribbean. Fully six major hurricanes struck the Bahamas, Mexico, Central America, and Cuba from 1931-1935. All of them were raised in intensity: a September 1931 TC that struck Belize to 115 kt; the 1932 Puerto Rican Huracán San Ciprian to 125 kt; the 1933 Cuba-TX TC to 105 kt in Cuba (and 140 kt in the Bahamas), the later September Caribbean cruiser to 120 kt in the Yucatán; and the 1935 Cuba-Bahamas hurricane to 120 kt in the Bahamas. These storms alone killed at least 3,046 total people along their paths--not including the 1935 Haitian-Nicaraguan hurricane, which claimed as many as 3,000, or an August 1933 TS that killed 70 in Jamaica, among others. (See this official list.)
  
• A late-season monster in November 1932. In keeping with the theme above, in November 1932 Cuba suffered its deadliest hurricane disaster--a 130-kt beast with a Katrina-sized, 35-40-mi radius of maximum winds that made landfall N of Santa Cruz del Sur. The 21-foot storm surge virtually destroyed the town and killed up to 3,500 people in Cuba. Days earlier, the storm had become a 150-kt hurricane based upon an incredible ship report of 70-kt winds with a pressure of 915 mb...outside the eye! The storm is now the only November Category 5 hurricane in Atlantic history, beating 135-kt Lenny 1999 in intensity, and maintained that intensity for three full days--the longest such duration in Atlantic annals, if the reanalysis experts decide to downgrade Hurricane Dog 1950.
  
• A big Texas micro hurricane in 1932. The August 1932 hurricane that hit near Freeport, TX, kept its Category-4 ranking but was raised to 130 kt/935 mb at landfall--with an eye less than 12 mi wide at landfall! The storm intensified extremely rapidly prior to landfall, from a 55-kt TS to a 130-kt monster in just a little more than 24 hours. East Columbia, TX, reported 942 mb in the eye...miles inland. Compare this storm to Celia 1970, only at a higher level. It is yet another reminder that not all N-Gulf systems weaken prior to landfall.
  

As usual, comments and questions are welcome!

[phil882]

Good stuff snowflake.... the recon data is particularly interesting with the various increases and decreases in intensity. I like your cliff-notes version of the findings, since I probably don't have enough time to read a 800 page report

[andyhb]

Very interesting data, Josh would've dreamed to live in the 1930s...intense Caribbean Cruisers galore (and chasing that Labor Day ferocity wouldn't hurt).

Although I dare say the Labor Day storm would be too intense, there were multiple reports of 190-200+ mph sustained winds with that thing (which fits with upgrade to 160 kts).

[Ground Scouring]

Good stuff snowflake.... the recon data is particularly interesting with the various increases and decreases in intensity. I like your cliff-notes version of the findings, since I probably don't have enough time to read a 800 page report

Ha...thanks!

I think the reanalysis offers a very good overview of the types of resources that are used and how little data the original HURDAT may have relied upon in the 1960s...for example, the current 165-kt classification of Dog 1950 seems to have been based solely upon the 1950 Monthly Weather Review summary, which stated that reconnaissance estimated such winds while reporting 100-foot wave heights--as we know, aside from pressure readings, such data from the early reconnaissance era are notoriously suspect. The wave data might be useful if more data were known about the size of the storm and of its fetch and gradient. As Dog was apparently a fairly large hurricane at the time, the fetch over the Sargasso Sea would have easily created such wave heights at an intensity well below 165 kt.

As the reanalysis paper notes in an excerpt from a Navy tropical-cyclone report on Dog:

‘‘As in previous flights into this storm, no penetration was planned because of the severity of the turbulence. . .it was considered desirable and adequate to circumnavi- gate at approximately the 70 kt wind circle. Features of this flight include the observation of the extremely large swells ahead of the hurricane, and the extent of hurricane winds over a very large area. It is believed that highest winds near the center were probably in excess of 150 kt’’

Another interesting fact is that in seven of the ten reanalyzed years, the ACE totals actually dropped...despite having accounted for an average 2.0 new TCs found each year (21 new tropical cyclones were found). The intensities of many well-surveyed TCs like Dog were decreased mainly due to peripheral reports and reconnaissance data, mainly pressures, that suggested the storms were not as intense as currently indicated in the best track. (Note that the 1945 TX hurricane is listed as a 120-kt landfall in the best track but is just a 95-kt/963-mb Category 2 hurricane after reanalysis--in line with the official Category-2 NHC designation for the storm in TX. The best-track winds were based upon reports of 120-kt gusts near the center around Port Lavaca.) However, given what we know now about the frequency of small systems, the possibility is high that some of the smallest known systems during 1944-1953, including some of the newly documented TCs, attained higher peak intensities than indicated after reanalysis, mainly because documenting the intensity of small cyclones is difficult...even with satellites and advanced data more than 70 years later. (See Tracy 1974 for some interesting debate on whether it was equal to a Saffir-Simpson-level Category 4 or 3.)

[Ground Scouring]

Very interesting data, Josh would've dreamed to live in the 1930s...intense Caribbean Cruisers galore (and chasing that Labor Day ferocity wouldn't hurt).

Although I dare say the Labor Day storm would be too intense, there were multiple reports of 190-200+ mph sustained winds with that thing (which fits with upgrade to 160 kts).

Also, do not forget about the unprecedented 110-kt U.S. landfalls occurring nearly simultaneously--within 24 hours of each other--at Brownsville, TX, and Jupiter, FL, on September 4-5, 1933. Such a close proximity of major U.S. landfalls--with two small, intense storms within 830 miles of each other, one moving WNW in the Gulf, the other moving WNW toward S FL--has never happened before or since in the Atlantic. Both of those storms may have been 115 kt at landfall...the reanalysis write-up on the TX hurricane (which was a long-lived Cape Verde TC that on August 31 made a 140-kt landfall on Mayaguana in the Bahamas) notes this possibility due to uncertainty regarding conditions at Brownsville at the time of closest approach:

If the 949 mb at Brownsville was indeed recorded with simultaneous

70 kt winds, it would suggest a central pressure at the time of well below 949 mb – possibly around

941 mb. If that value is subtracted back to the coastline, it could suggest a landfall central

pressure of 937 mb. A run of the Schloemer equation using 20 nmi for the RMW and a 10 nmi distance

between the 949 mb ob and the center of the hurricane yields a central pressure of 939 mb, and when

extrapolated back to the coast, that yields a central pressure of about 935 mb. It is estimated

that the landfall central pressure was approximately 940 mb. A 20 nmi RMW is chosen based on the

Ho’s analysis (slightly larger than the 16 nmi climatological RMW) and the forward speed of the

storm was 8 kt. A 940 mb central pressure equals 115 kt according to the north of 25N pressure-wind

relationship, but 110 kt is chosen for the landfall intensity because of the large RMW and slow speed,

retaining the hurricane as a Category 3 for south Texas. It should be noted that it remains uncertain how close the center was to Brownsville.

If the center made landfall farther north than currently shown and the RMW was smaller, then it

is possible that this hurricane was a category 4 at landfall with Category 4 conditions north of Brownsville on Padre Island.

Also, the FL TC was rather small, with an eye less than 17 miles wide and a relatively fast 14-kt movement at the time...plus a slightly tighter-than-usual pressure gradient. Some damage descriptions from that storm in the Hobe Sound-Stuart area (second-story floors blown off, 75%+ of the Stuart roofs blown off or severely damaged, and other details I have seen in newspapers and books) suggest a Category-4 landfall of 115 kt at Jupiter...but then again, the integrity of many of the warehouse-type structures at Hobe Sound was not great. Anyway, 110 kt/948 mb seems just as reasonable as 115. The storm was definitely a big hit for the Stuart area.

[thewxmann]

Wow, this is a really great summary snowflake... thank you for compiling and sorting all that information. It's good to reminisce on a time where there were intense MDR storms...

[HurricaneJosh]

From 1945 to 1950, South FL was hit by no less than five Category 4 hurricanes--four of them, in 1945, 1948, 1949, and 1950, currently listed as Category-3 strikes in the official list of U.S. landfalls, 1851-2011! If confirmed, the number of Category-4 landfalls would set a record for Category-4 landfalls in a single decade.

Good summary, snowflake. I've read much of the Hagen/Landsea paper and the analyses of each storm are very interesting.

Re: the spate of strong FL hurricanes in the 1945-1950 period... Their ultimate verdict for King 1950 was Cat 3 (110 kt). In the first version of the thesis, Hagen/Landea put it at Cat 4 (115 kt), but then revised it down a bit in a rewrite. So King should be considered a 3, not a 4.

[Ground Scouring]

Good summary, snowflake. I've read much of the Hagen/Landsea paper and the analyses of each storm are very interesting.

Re: the spate of strong FL hurricanes in the 1945-1950 period... Their ultimate verdict for King 1950 was Cat 3 (110 kt). In the first version of the thesis, Hagen/Landea put it at Cat 4 (115 kt), but then revised it down a bit in a rewrite. So King should be considered a 3, not a 4.

Thanks for the clarification...the appendix seemed to have been issued at the same time as that of the edited, published report, so I was unaware that it was not the final assessment. I have edited the blurb on King accordingly.

Still, the 1945-1950 period in S FL was just incredible...especially since the average return rate for a major hurricane in SE FL is about 13 years.

[dan11295]

The new value for Dog in 1950 makes a lot more sense. Was always very suspicious of the 165 kt previously assigned to it. There was no other recorded storm outside of the Caribbean, Gulf of Mexico or Gulf stream in the S Florida vicinity with a peak intensity above 145 kt (i.e Isabel)

[Sickman]

A 50 kt TS in Portugal? Wow. That's even more impressive considering that Vince barely made it to the Iberian Peninsula as a weakening depression.

[JPFLL]

This is great info to read through. Considering the majors that came through Florida during those periods.. is there anything I can read that discusses why Florida was the target during this period? thanks for posting this.

[aslkahuna]

A 50 kt TS in Portugal? Wow. That's even more impressive considering that Vince barely made it to the Iberian Peninsula as a weakening depression.

That would be equivalent to the 45kt TS that hit SoCA in 1939 (or the apparent Cat 1 in San Diego in 1858).

Steve

[HurricaneAndrew]

Among the highlights:

• Only two storms currently listed as a major U.S. landfall during the 1944-1953 period were downgraded: the 1944 Great Atlantic hurricane was lowered in every state and made a 95-kt landfall on Long Island, NY; the October 1944 TC was downgraded to a 90-kt Category 2 TC in SW FL, near Englewood. (Also, the 1947 Fort Lauderdale hurricane, while keeping its Category-4 South FL landfall, was reduced to a 95-kt Category 2 impact in LA and MS.)
  
• The October 1944 TC, upon striking W Cuba, was upgraded from 105 kt to 120 kt, based upon a (ship?) report of 937 mb, apparently in the eye near Havana. (The same TC later passed over the Dry Tortugas, FL, at 105 kt, where the eye reportedly lasted for two full hours with a central pressure of 949 mb.) In the Leeward Islands, Hurricane Dog 1950 was boosted from 105 to 125 kt, based upon a sustained wind measured on an island, west of the eye.
  

The October, 1944 hurricane is still a major hurricane for the United States because it was analyzed as a Category 3 for SW FL (Dry Tortugas counts as SW FL).

The October 1944 hurricane was increased from 105 to 120 kts for Cuba due to a 30-sec averaged wind of 122 kts at Havana as well as a 937 mb surface ob in Cuba at a sugar mill closer to the north coast of Cuba about 2 to 3 hours after landfall.

[LongBeachSurfFreak]

So Carol as reused in 53/54? Interesting that it was a major East Coast storm in bother years!

[Ground Scouring]

The October, 1944 hurricane is still a major hurricane for the United States because it was analyzed as a Category 3 for SW FL (Dry Tortugas counts as SW FL).

The October 1944 hurricane was increased from 105 to 120 kts for Cuba due to a 30-sec averaged wind of 122 kts at Havana as well as a 937 mb surface ob in Cuba at a sugar mill closer to the north coast of Cuba about 2 to 3 hours after landfall.

Thanks for the corrections and clarifications! Of course the western Florida Keys count under BFL (SW FL) according to the regional classification. I i I was also unaware that the 937-mb pressure was recorded at a sugar mill on land, though I knew it was probably after landfall due to its placement in the revised time-slots. (Check your messages.)

[thegreatdr]

I might know something about the changes suggested for 1951. Back in 2004, I wrote up a series of reports for possible new TCs for the Atlantic hurricane database. The best candidates were during the off-season of 1951 in January and December, which made the Hagen paper. I do not believe the February system made the cut, however.