Reanalysis questions

[1900hurricane]

Helene '88 was probably closest at 09Z to having a T7.0 (below). The eye didn't break WMG until that time when it reached 14.5ºC, and the white ring (or red colder than -70ºC/203K in my image) dissipated by 12Z. Even then though, the ring width falls short of the required 0.5º thickness in the northern part of the CDO, which can be seen since I have plotted every half-degree.



Just eyeballing, I might do 135 kt for 00Z (W/a deep OW), 130 kt for 06Z (B/a not so deep OW), and 135 kt again for 12Z (B/WMG following up B/WMG w/W ring form 09Z) before coming down.

[1900hurricane]

I've taken a look back this evening at the 1988 season after reading the personal analysis you both have presented, and I'm thinking there is a good chance both Helene and Joan did reach Category 5 status. I'm also struck by the similarities between the 1988 and 2007 seasons. Dean following a similar track to Gilbert and Felix following a Joan-esque one. Looking at some satellite images and loops of Joan, there are definitely some moments when I'm reminded of Felix, and I'm thinking based on the size of the storm and the pressures, Joan could quite possibly have peaked with 140-150kts, in fact, I think it's more likely than not.

I'd be interested to see some KZC analysis on Joan to see what the outputs would be. I know I've seen some KZC analysis done by some users on the forum, but it's beyond my skill level.

I use KZC quite a bit since I've written a few Python programs that can implement it pretty quickly. Probably the biggest issue with running it for Joan '88 would involve deriving the S parameter, which approximates storm size. The two most common ways to derive it are from TS wind radii and from the radius of the outermost closed isobar, both of which may be a pain digging up for Joan. I could probably figure something out though once I have some free time, even if it's just something as simple as eyeballing it and using a confidence interval.

[Chris90]

I've taken a look back this evening at the 1988 season after reading the personal analysis you both have presented, and I'm thinking there is a good chance both Helene and Joan did reach Category 5 status. I'm also struck by the similarities between the 1988 and 2007 seasons. Dean following a similar track to Gilbert and Felix following a Joan-esque one. Looking at some satellite images and loops of Joan, there are definitely some moments when I'm reminded of Felix, and I'm thinking based on the size of the storm and the pressures, Joan could quite possibly have peaked with 140-150kts, in fact, I think it's more likely than not.

I'd be interested to see some KZC analysis on Joan to see what the outputs would be. I know I've seen some KZC analysis done by some users on the forum, but it's beyond my skill level.

I use KZC quite a bit since I've written a few Python programs that can implement it pretty quickly. Probably the biggest issue with running it for Joan '88 would involve deriving the S parameter, which approximates storm size. The two most common ways to derive it are from TS wind radii and from the radius of the outermost closed isobar, both of which may be a pain digging up for Joan. I could probably figure something out though once I have some free time, even if it's just something as simple as eyeballing it and using a confidence interval.

Haha, yeah, you were who I was thinking of when I was thinking of the KZC analysis. I've read through a lot of your analysis in the intense tropical cyclones thread and read through your reanalysis of the 1979 typhoons.

I'll see what I might be able to dig up on Joan and if I find anything I'll post it here. I already did a quick Google search before on it, but data on the storm isn't exactly extensive. I did find a little article with some info that mentioned the recon flight measured 936mb towards the edge of the eye which is why they estimated 932mb as the lowest central pressure and it also mentioned flight level winds peaked at 143mph, which surprised me. I'll see if I can find info regarding the wind radii and OCI to make it a bit easier if you do happen to get the chance to take a quick look at it.

[1900hurricane]

I will say one thing about Joan '88 that likely held it down when approaching final landfall after looking at some HURSAT data. It looks like Joan was going through eyewall replacement. The microwave pass from 0010Z on October 22 clearly showed concentric eyewalls (below). Eyewall replacement became apparant in conventional imagery not long afterwards (also below) The microwave pass is within an hour of the 937 mb dropsonde near the edge of the eye from 0055Z, as mentioned in this portion of the report. Interestingly, with regards to the lowish flight level winds of of 124 kt/143 mi/hr at that time, it was actually recorded at 850 mb rather than a more standard 700 mb for more intense systems. Per Franklin, Black, and Valde, the recommended reduction factor from 850 mb to surface winds is 0.8, and using this reduction factor for the 124 kt at 850 mb yields a surface wind estimate only of about 100 kt, which is surprisingly low to me. I might chalk it up to undersampling if anything, and reduction factors aren't always consistent from storm to storm anyway.

I'd probably put the floor from 00Z to 06Z at 115 kt, which is near the AH77 value for 932 mb when adjusting the assumed environmental pressure upwards some for the Atlantic. The ceiling is probably in the vicinity of 140 kt, which would be right at 00Z October 22 as eyewall replacement was beginning, and this is based on a very impressive satellite presentation between 21Z-00Z.

[CrazyC83]

When dealing with the 1980s, the planes were not as accurate as they are now.

The P-3 was very reliable since 1977, and that should get at least strong weighting in the winds, but the AF planes weren't as trustworthy (often with a low bias) until the current models in 1999.

[1900hurricane]

That makes sense. It also explains why most of the reported flight level winds were so low in many of the recon era Pacific Typhoons.

[CrazyC83]

Revisiting Hugo, here is how I would reanalyze it (reasons below) - green indicates weaker, red indicates stronger, blue indicates new BT point.

AL111989, HUGO, 65,
19890910, 1200, , TD, 13.2N, 20.0W, 25, 1010, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890910, 1800, , TD, 13.3N, 21.8W, 25, 1010, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890911, 0000, , TD, 13.2N, 23.7W, 30, 1009, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890911, 0600, , TD, 13.0N, 25.5W, 30, 1007, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890911, 1200, , TD, 12.8N, 27.3W, 30, 1005, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890911, 1800, , TS, 12.5N, 29.2W, 35, 1003, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890912, 0000, , TS, 12.5N, 31.0W, 40, 1002, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890912, 0600, , TS, 12.5N, 32.9W, 45, 1000, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890912, 1200, , TS, 12.5N, 34.8W, 45, 998, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890912, 1800, , TS, 12.6N, 36.7W, 50, 996, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890913, 0000, , TS, 12.6N, 38.2W, 55, 994, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890913, 0600, , TS, 12.7N, 40.0W, 55, 992, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890913, 1200, , TS, 12.8N, 41.8W, 60, 990, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890913, 1800, , HU, 12.8N, 43.5W, 65, 987, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890914, 0000, , HU, 12.9N, 44.9W, 70, 984, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890914, 0600, , HU, 13.0N, 46.3W, 80, 980, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890914, 1200, , HU, 13.2N, 47.8W, 85, 975, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890914, 1800, , HU, 13.6N, 49.1W, 95, 967, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890915, 0000, , HU, 13.8N, 50.5W, 105, 959, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890915, 0600, , HU, 14.0N, 51.9W, 120, 946, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890915, 1200, , HU, 14.2N, 53.3W, 140, 930, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890915, 1800, , HU, 14.6N, 54.6W, 150, 918, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890916, 0000, , HU, 14.8N, 56.1W, 140, 921, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890916, 0600, , HU, 15.1N, 57.3W, 130, 927, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890916, 1200, , HU, 15.4N, 58.4W, 120, 940, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890916, 1800, , HU, 15.8N, 59.4W, 120, 941, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890917, 0000, , HU, 16.1N, 60.4W, 125, 940, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890917, 0500, L, HU, 16.3N, 61.3W, 125, 939, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890917, 0600, , HU, 16.4N, 61.5W, 125, 941, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890917, 1200, , HU, 16.6N, 62.5W, 125, 946, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890917, 1800, , HU, 16.9N, 63.5W, 130, 942, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890918, 0000, , HU, 17.2N, 64.1W, 135, 934, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890918, 0600, L, HU, 17.7N, 64.8W, 130, 938, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890918, 1200, L, HU, 18.2N, 65.5W, 120, 943, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890918, 1300, L, HU, 18.3N, 65.6W, 115, 944, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890918, 1800, , HU, 19.1N, 66.4W, 95, 958, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890919, 0000, , HU, 19.7N, 66.8W, 90, 959, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890919, 0600, , HU, 20.7N, 67.3W, 85, 962, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890919, 1200, , HU, 21.6N, 68.0W, 80, 965, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890919, 1800, , HU, 22.6N, 68.6W, 75, 968, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890920, 0000, , HU, 23.5N, 69.3W, 80, 961, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890920, 0600, , HU, 24.4N, 70.1W, 85, 957, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890920, 1200, , HU, 25.2N, 71.0W, 90, 955, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890920, 1800, , HU, 26.3N, 72.2W, 95, 950, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890921, 0000, , HU, 27.2N, 73.4W, 100, 948, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890921, 0600, , HU, 28.0N, 74.9W, 105, 946, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890921, 1200, , HU, 29.0N, 76.1W, 115, 945, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890921, 1800, , HU, 30.2N, 77.5W, 120, 940, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890922, 0000, , HU, 31.7N, 78.8W, 125, 933, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890922, 0400, L, HU, 32.8N, 79.8W, 130, 930, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890922, 0600, , HU, 33.5N, 80.3W, 90, 948, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890922, 1000, T, HU, 35.1N, 81.4W, 70, 967, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890922, 1200, , TS, 35.9N, 81.7W, 60, 972, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890922, 1800, , TS, 38.5N, 81.8W, 50, 987, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890923, 0000, , EX, 42.2N, 80.2W, 45, 985, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890923, 0600, , EX, 46.0N, 74.5W, 45, 985, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890923, 1200, , EX, 49.0N, 69.0W, 45, 988, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890923, 1800, , EX, 51.0N, 65.0W, 40, 993, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890924, 0000, , EX, 52.0N, 62.0W, 40, 994, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890924, 0600, , EX, 52.5N, 60.5W, 40, 993, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890924, 1200, , EX, 53.0N, 59.5W, 40, 991, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890924, 1800, , EX, 53.5N, 58.5W, 40, 989, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890925, 0000, , EX, 54.0N, 57.0W, 45, 983, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890925, 0600, , EX, 56.0N, 52.0W, 50, 979, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19890925, 1200, , EX, 58.0N, 46.0W, 55, 974, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,

1) Peak intensity increased upwards to 150 kt, based on the likelihood of undersampling due to a plane in distress at its peak, as well as the P-W relationship for 918mb. Intensities before that flight were also increased accordingly.

2) Landfall at Guadeloupe was increased to 125 kt, based on Recon reports of 137 kt (unclear the height but assuming at 700 mb) and a slightly lower pressure confirmed by surface data.

3) Considered an increase to 140 kt at 18/0000Z based on 151 kt flight level winds, but it appeared too transient (I'd think it would have been a cat 5 if we had SFMR data to back it up). I went with 135 kt there for its Caribbean peak.

4) Landfalls at St. Croix, Vieques and mainland Puerto Rico were all increased, based on surface pressure observations and the lack of reliable Recon winds in that time period. I went with 130 kt, 120 kt and 115 kt, respectively, for those three landfalls.

5) Intensities are decreased for the most part on the 19th and most of the 20th, as structural changes were underway. Recon did not support the intensities in HURDAT, even if they are biased low they never supported more than a cat 1. By late on the 20th, the data caught up as it became much better organized.

6) At SC landfall, the pressure was lowered to 930 mb based on a report of 933 mb in Mt. Pleasant which was just outside the eye (confidence pretty good on that). The intensity is a tough one to determine due to conflicting factors: that would naturally be 125-130 kt, it was moving fast (calls for higher) but it was also large (calls for lower). Also, the area that saw peak winds is very sparsely populated so there were no observations supporting such (strongest was in Sanpit River at 104 kt but surely it was higher in the uninhabited areas). The last Recon flight had 140 kt winds at 12,000 feet which is a non-standard elevation. For those reasons, I went with 130 kt for landfall, with uncertainty +/- 10 kt. Helene in 1958, with the same pressure in a similar location, was also assessed at 130 kt.

7) I added a BT point at 1000Z to confirm the NC1 intensity in HURDAT (cat 1 impact for inland NC).

[galaxy401]

These HURDAT reanalysis are going to get more interesting now that we are in the satellite era. If a storm like Hugo, Felix, or Helen existed now, what intensity would they be assessed? Jose is the closest example we have since then it seems.

[CrazyC83]

These HURDAT reanalysis are going to get more interesting now that we are in the satellite era. If a storm like Hugo, Felix, or Helen existed now, what intensity would they be assessed? Jose is the closest example we have since then it seems.

We learned a lot even in the last few years. We could probably do a reanalysis almost right up to recent seasons, as the SFMR became operational in 2007 (it was experimental from 2003-06) and SATCON came around when, 2010?

[1900hurricane]

I happen to look at re-analysis from 1931 to 1943.

Documentation for 1931 to 1943
http://www.aoml.noaa.gov/hrd/Landsea/la ... e-2014.pdf

I noticed the 1935 Labor Day Hurricane had a large radius of outer closed isobars (ROCI). It has a ROCI of 300 nautical miles, which is quite large. If I recall, the hurricane prior to hitting Florida Keys had hurricane force winds extending up to 15 miles.

Just saw this. Here's what I get when I plug that ROCI of 300 nm and OCI of 1010 mb into KZC when coupled with the 892 mb pressure, 7.4 kt forward speed, and 24.8ºN landfall latitude.

>>> from KZCeq import KZC, KZCroci
>>> v = 100
>>> roci = 300
>>> oci = 1010
>>> p0 = 1000
>>> c = 7.4
>>> p = 892
>>> l = 24.8
>>> while p0 > p:
v = v + 0.1
p0 = KZC(KZCroci(v, c, roci, l), oci)


>>> print("%3.0f" % v)
167

That 167 kt is more intense than in best track. When rounded to 165 kt, that KZC output would tie Allen '80 for the highest LMI for the NAtl.

[1900hurricane]

These HURDAT reanalysis are going to get more interesting now that we are in the satellite era. If a storm like Hugo, Felix, or Helen existed now, what intensity would they be assessed? Jose is the closest example we have since then it seems.

We learned a lot even in the last few years. We could probably do a reanalysis almost right up to recent seasons, as the SFMR became operational in 2007 (it was experimental from 2003-06) and SATCON came around when, 2010?

SATCON archives date back to 2005 on the CIMSS page, although it wasn't in its current form until late 2013. Interestingly, it appears someone re-ran a recent version of SATCON for Katrina '05. Looks like it did ok.

[Shell Mound]

I happen to look at re-analysis from 1931 to 1943.

Documentation for 1931 to 1943
http://www.aoml.noaa.gov/hrd/Landsea/la ... e-2014.pdf

I noticed the 1935 Labor Day Hurricane had a large radius of outer closed isobars (ROCI). It has a ROCI of 300 nautical miles, which is quite large. If I recall, the hurricane prior to hitting Florida Keys had hurricane force winds extending up to 15 miles.

Just saw this. Here's what I get when I plug that ROCI of 300 nm and OCI of 1010 mb into KZC when coupled with the 892 mb pressure, 7.4 kt forward speed, and 24.8ºN landfall latitude.

>>> from KZCeq import KZC, KZCroci
>>> v = 100
>>> oci = 1010
>>> p0 = 1000
>>> c = 7.4
>>> p = 892
>>> l = 24.8
>>> while p0 > p:
v = v + 0.1
p0 = KZC(KZCroci(v, c, roci, l), oci)


>>> print("%3.0f" % v)
167

That 167 kt is more intense than in best track. When rounded to 165 kt, that KZC output would tie Allen '80 for the highest LMI for the NAtl.

This is a very good baseline for establishing intensity, but it can be improved further, when adjusting for other variables. Wouldn't accounting for the intensity trend (rapidly deepening) and RMW (~5 n mi, much smaller than climatology for its location) add at least ten knots to the KZC output? I, personally, would assume that these factors would bump up the baseline from 165 to at least 175 knots, and that the actual winds were likely similar to Patricia's or Haiyan's (185 to 190 knots). Is that a reasonable guess to you?

[1900hurricane]

I happen to look at re-analysis from 1931 to 1943.

Documentation for 1931 to 1943
http://www.aoml.noaa.gov/hrd/Landsea/la ... e-2014.pdf

I noticed the 1935 Labor Day Hurricane had a large radius of outer closed isobars (ROCI). It has a ROCI of 300 nautical miles, which is quite large. If I recall, the hurricane prior to hitting Florida Keys had hurricane force winds extending up to 15 miles.

Just saw this. Here's what I get when I plug that ROCI of 300 nm and OCI of 1010 mb into KZC when coupled with the 892 mb pressure, 7.4 kt forward speed, and 24.8ºN landfall latitude.

>>> from KZCeq import KZC, KZCroci
>>> v = 100
>>> oci = 1010
>>> p0 = 1000
>>> c = 7.4
>>> p = 892
>>> l = 24.8
>>> while p0 > p:
v = v + 0.1
p0 = KZC(KZCroci(v, c, roci, l), oci)


>>> print("%3.0f" % v)
167

That 167 kt is more intense than in best track. When rounded to 165 kt, that KZC output would tie Allen '80 for the highest LMI for the NAtl.

This is a very good baseline for establishing intensity, but it can be improved further, when adjusting for other variables. Wouldn't accounting for the intensity trend (rapidly deepening) and RMW (~5 n mi, much smaller than climatology for its location) add at least ten knots to the KZC output? I, personally, would assume that these factors would bump up the baseline from 165 to at least 175 knots, and that the actual winds were likely similar to Patricia's or Haiyan's (185 to 190 knots). Is that a reasonable guess to you?

This may be a bit anecdotal simply because there is so little direct recon data from such systems, but KZC's shining moment actually came from an exceptionally deep system that also had an analyzed 5 nm radius of maximum winds: Hurricane Patricia from 2015. Here''s the straight KZC rip out of Patricia's b-deck when going from pressure to wind:

Date & Time Lat/Lon Vmax(kt) Speed Mean Analyzed OCI
exp(act) (kt) R34 Pressure
------------------------------------------------------------------------
10/20/2015 12Z: 13.3N 94.2W 27 ( 30), 2, 0, 1006, 1009
10/20/2015 18Z: 13.2N 94.6W 28 ( 30), 4, 0, 1006, 1009
10/21/2015 0Z: 13.1N 95.1W 33 ( 35), 5, 15, 1004, 1009
10/21/2015 6Z: 12.9N 96.2W 36 ( 35), 11, 45, 1004, 1009
10/21/2015 12Z: 12.9N 97.4W 40 ( 40), 12, 52, 1001, 1008
10/21/2015 18Z: 13.1N 98.7W 51 ( 50), 13, 65, 997, 1009
10/22/2015 0Z: 13.4N 100.1W 62 ( 60), 14, 78, 991, 1009
10/22/2015 6Z: 14.0N 101.7W 76 ( 75), 17, 88, 981, 1007
10/22/2015 12Z: 14.6N 103.1W 92 ( 90), 15, 95, 969, 1007
10/22/2015 18Z: 15.2N 104.2W 104 (115), 12, 98, 957, 1007
10/23/2015 0Z: 15.8N 104.9W 143 (150), 9, 105, 920, 1007
10/23/2015 6Z: 16.5N 105.4W 173 (180), 8, 110, 886, 1007
10/23/2015 12Z: 17.3N 105.6W 185 (185), 8, 110, 872, 1007
10/23/2015 18Z: 18.3N 105.3W 180 (180), 10, 115, 878, 1007
10/23/2015 23Z: 19.4N 105.0W 130 (130), 14, 115, 932, 1007
10/24/2015 0Z: 19.6N 104.9W 115 (110), 13, 82, 946, 1007
10/24/2015 6Z: 21.6N 103.8W 66 ( 50), 22, 20, 985, 1007
10/24/2015 12Z: 23.2N 102.3W 40 ( 25), 21, 0, 1000, 1007

There are a couple of points (18Z October 22 and 00Z October 23) where the winds outperformed expectations, but most points, especially those near peak intensity, are in very good agreement with KZC expectations (the 06Z October 23 point was 46 minutes prior to an 879 mb pass, 1 mb off from the 878 mb expected from the analyzed 180 kt).

Now, this does come from using r34 to estimate S (storm size parameter), as outlined in Courtney & Knaff. I have considerably more confidence in this method than the ROCI method used by Knapp et al. The R34 method feels much more though to me, requiring several equations to derive S. The ROCI method only uses one. Both methods are below, beginning with the R34 method and then the ROCI method.

Code: Select all

V500 = r34 / 9 - 3
Rmax = 66.785 - 0.09102 * v + 1.0619 * (lat - 25)
X = 0.1147 + 0.0055 * v - 0.001 * (lat - 25)
V500c = v * (Rmax / 500) ** X
S = V500 / V500c

Code: Select all

S = (roci / 60) / 8 + 0.1

[tatertawt24]

Earlier I took a look at majors up to 1978 (although some may warrant further revision). Now here is my thought for the major hurricanes (either in the best track or that I believe were such) from 1979 to 1994 (i.e. before the most recent era):

1979 - David (BT 150 kt, no change), Frederic (BT 115 kt, no change)
1980 - Allen (BT 165 kt, no change), Frances (BT 100 kt, increase to 125 kt), Ivan (BT 90 kt, increase to 100 kt)
1981 - Harvey (BT 115 kt, no change), Irene (BT 105 kt, no change); REMOVED: Floyd (BT 100 kt, decrease to 85 kt)
1982 - no majors; REMOVED: Debby (BT 115 kt, decrease to 95 kt)
1983 - no majors; REMOVED: Alicia (BT 100 kt, decrease to 90 kt)
1984 - Diana (BT 115 kt, no change)
1985 - Elena (BT 110 kt, no change), Gloria (BT 125 kt, increase to 135 kt), Kate (BT 105 kt, decrease to 100 kt)
1986 - no majors
1987 - Emily (BT 110 kt, increase to 120 kt)
1988 - Gilbert (BT 160 kt, increase to 170 kt), Helene (BT 125 kt, increase to 140 kt), Joan (BT 125 kt, no change)
1989 - Gabrielle (BT 125 kt, no change), Hugo (BT 140 kt, increase to 150 kt)
1990 - Gustav (BT 105 kt, no change)
1991 - Bob (BT 100 kt, increase to 110 kt)
1992 - Andrew (BT 150 kt, increase to 160 kt)
1993 - Emily (BT 100 kt, increase to 110 kt)
1994 - no majors

The biggest increase is with Frances 1980 (best analysis is T6.5, before even reaching 30W longitude!) while Debby is the biggest decrease (no evidence it was higher than T5.0 looking at satellite).

My own tidbits, to be taken for what they are worth:

*Some of the systems listed here may be underestimate due to small size. For instance, I would not necessarily keep Joan (1988) at *only* 125 knots. Joan was a very compact, rapidly deepening system as it neared landfall in Nicaragua, much like Ethel (1971), which, despite *only* attaining a minimum recon-derived pressure of 943 mb (which likely was not the lowest pressure, as landfall occurred several hours later, and the storm was rapidly intensifying at the time of the last mission), was likely fully deserving of its Cat-5 designation in HURDAT as it made landfall near Cabo Gracias a Dios, on the Honduras/Nicaragua border. Both Joan and Ethel featured pinhole-type, well-defined eyes on aircraft radar, but the satellite estimates likely did not register the small inner cores. A similar situation actually existed with Felix (2007), whose satellite presentation just before landfall in Nicaragua did *not* support Cat-5 intensity—only a fortuitously timed recon mission hours earlier found that winds in the storm were higher (high-end Cat-4) than satellite indicated, and based on the continued improvement in presentation afterward, the NHC went with 140 knots for the landfall. In all three cases, the highest winds on land affected sparsely populated regions and thus went unrecorded. Based on available data and pre-landfall intensity trends, all three storms likely featured pressures in the low to mid 930s at their peaks/landfalls and similar winds, in the 135-to-145-knot range. My guesses: 135 knots for Joan, 145 knots for Ethel and Felix.

The 1935 Labor Day hurricane, Ethel, Joan, Andrew (1992), Iris (2001), Charley (2004), and Felix, among others, all belong in the same class: tiny, explosively deepening cyclones whose actual intensities were far higher than were/would have been discernible on satellite. Case in point for Joan: the last recon mission into Joan found a pressure of 933 mb hours before landfall over Central America, along with a very small eye. If anything, Joan was probably closer to 135 knots than 125 knots at its peak intensity. As for Andrew, the flight-level winds, pressures, and radar data from both its initial and secondary peaks strongly suggests that 160 knots is probably too low for such a small, explosively deepening, rapidly translating, westward-moving system situated beneath a very strong subtropical ridge and embedded in high environmental pressures. A more realistic estimate for Andrew, in my view, would be 170 knots (with well-documented 922-mb pressures) for both peaks, to the east of the Bahamas and at landfall on South FL. Despite studies, it is by no means conclusive that the extreme wind damage in Naranja Lakes et al., assumed to be from mesovortices in Andrew's eyewall over Dade County, was *entirely* due to mesovortices rather than the representative winds in a very extreme, compact, Cat-5 hurricane. The 1935 hurricane, being even smaller and more intense (892 mb, if not lower) than Andrew, is a very good candidate for Patricia- or Haiyan-league winds—185 to 190 knots at landfall on the FL Keys.

One key note about Andrew: the extreme wind-caused damage swaths were co-located with strong, N-eyewall (hence RMW) convection, suggestive of *potentially* representative winds.

Re: David and Hugo, I concur with your estimate for Hugo, but perhaps not for David. Even visible satellite imagery reveals that David had a very bright, white CDO, extremely impressive, circular outflow, and stadium-effect eye at the time(s) of its likely peak intensity over the northeastern Caribbean Sea. Aircraft-derived surface estimates and measured flight-level winds, while less reliable in the late 1970s vs. today, suggested winds higher than the pressure-wind relationship indicated. David was not a midget-sized storm at the time, but it certainly wasn't a large one, in terms of wind radii. It was pretty similar to Inez (1966) in size. Interestingly, aircraft infamously reported 700-mb winds of 170 knots in the northern eyewall of Inez, concurrent with rapid intensification and a minimum (observed, likely not the deepest) pressure of 927 mb. David and Inez were both similar in terms of overall size, RMW, observed pressure(s), and intensity trend(s) around peak(s); therefore, one may logically deduce that both were similarly intense, with minimum pressures likely in the low 920s and peak sustained winds of 155 to 160 knots. In other words, Irma- and Maria-type winds, but higher pressures because smaller RMWs and wind radii. Another convincing piece of evidence in favour: Inez and David (like Felix at its peak!) were embedded in high environmental pressures. ...

On another note, Hugo is an interesting case to reanalyse, not just for its (probably 150-knot) Cat-5 peak over the MDR, but also for its Cat-4 landfall in SC. Radar data from Charleston, showing a continued tightening of the eye and intensification of lightning/convective activity, suggest that Hugo continued to deepen after the last aircraft mission, while satellite data from the time of landfall (and before) clearly suggested a higher intensity than 120 knots. Even aircraft-derived winds suggest that 120 knots was a cautious estimate. Given evidence of mesovortices in the eye of Hugo, the minimum pressure of 934 mb at landfall is probably a bit high. Finally, while this is inconclusive, the wind damage in the Bulls Bay–McClellanville area was consistent with a much stronger hurricane. All in all, a solid case can be made that Hugo's true intensity at landfall in SC was in the 130-to-135-knot range, with a pressure of 930-932 mb.

Oh, and I also concur with Chris90 re: Lenny being a solid candidate for Cat-5 status, however briefly it attained that plateau.

Terrific post.

1.) Wasn't Naranja Lakes situated in the southern eyewall of Andrew? Hence the mesovortex explanation for the extreme damage.

2.) Do you have examples of the stronger wind damage from Hugo? Not that I disagree, I'd just really love to see them. And by evidence of mesovortices in the eye, do you mean the satellite imagery?

[Shell Mound]

Earlier I took a look at majors up to 1978 (although some may warrant further revision). Now here is my thought for the major hurricanes (either in the best track or that I believe were such) from 1979 to 1994 (i.e. before the most recent era):

1979 - David (BT 150 kt, no change), Frederic (BT 115 kt, no change)
1980 - Allen (BT 165 kt, no change), Frances (BT 100 kt, increase to 125 kt), Ivan (BT 90 kt, increase to 100 kt)
1981 - Harvey (BT 115 kt, no change), Irene (BT 105 kt, no change); REMOVED: Floyd (BT 100 kt, decrease to 85 kt)
1982 - no majors; REMOVED: Debby (BT 115 kt, decrease to 95 kt)
1983 - no majors; REMOVED: Alicia (BT 100 kt, decrease to 90 kt)
1984 - Diana (BT 115 kt, no change)
1985 - Elena (BT 110 kt, no change), Gloria (BT 125 kt, increase to 135 kt), Kate (BT 105 kt, decrease to 100 kt)
1986 - no majors
1987 - Emily (BT 110 kt, increase to 120 kt)
1988 - Gilbert (BT 160 kt, increase to 170 kt), Helene (BT 125 kt, increase to 140 kt), Joan (BT 125 kt, no change)
1989 - Gabrielle (BT 125 kt, no change), Hugo (BT 140 kt, increase to 150 kt)
1990 - Gustav (BT 105 kt, no change)
1991 - Bob (BT 100 kt, increase to 110 kt)
1992 - Andrew (BT 150 kt, increase to 160 kt)
1993 - Emily (BT 100 kt, increase to 110 kt)
1994 - no majors

The biggest increase is with Frances 1980 (best analysis is T6.5, before even reaching 30W longitude!) while Debby is the biggest decrease (no evidence it was higher than T5.0 looking at satellite).

My own tidbits, to be taken for what they are worth:

*Some of the systems listed here may be underestimate due to small size. For instance, I would not necessarily keep Joan (1988) at *only* 125 knots. Joan was a very compact, rapidly deepening system as it neared landfall in Nicaragua, much like Ethel (1971), which, despite *only* attaining a minimum recon-derived pressure of 943 mb (which likely was not the lowest pressure, as landfall occurred several hours later, and the storm was rapidly intensifying at the time of the last mission), was likely fully deserving of its Cat-5 designation in HURDAT as it made landfall near Cabo Gracias a Dios, on the Honduras/Nicaragua border. Both Joan and Ethel featured pinhole-type, well-defined eyes on aircraft radar, but the satellite estimates likely did not register the small inner cores. A similar situation actually existed with Felix (2007), whose satellite presentation just before landfall in Nicaragua did *not* support Cat-5 intensity—only a fortuitously timed recon mission hours earlier found that winds in the storm were higher (high-end Cat-4) than satellite indicated, and based on the continued improvement in presentation afterward, the NHC went with 140 knots for the landfall. In all three cases, the highest winds on land affected sparsely populated regions and thus went unrecorded. Based on available data and pre-landfall intensity trends, all three storms likely featured pressures in the low to mid 930s at their peaks/landfalls and similar winds, in the 135-to-145-knot range. My guesses: 135 knots for Joan, 145 knots for Ethel and Felix.

The 1935 Labor Day hurricane, Ethel, Joan, Andrew (1992), Iris (2001), Charley (2004), and Felix, among others, all belong in the same class: tiny, explosively deepening cyclones whose actual intensities were far higher than were/would have been discernible on satellite. Case in point for Joan: the last recon mission into Joan found a pressure of 933 mb hours before landfall over Central America, along with a very small eye. If anything, Joan was probably closer to 135 knots than 125 knots at its peak intensity. As for Andrew, the flight-level winds, pressures, and radar data from both its initial and secondary peaks strongly suggests that 160 knots is probably too low for such a small, explosively deepening, rapidly translating, westward-moving system situated beneath a very strong subtropical ridge and embedded in high environmental pressures. A more realistic estimate for Andrew, in my view, would be 170 knots (with well-documented 922-mb pressures) for both peaks, to the east of the Bahamas and at landfall on South FL. Despite studies, it is by no means conclusive that the extreme wind damage in Naranja Lakes et al., assumed to be from mesovortices in Andrew's eyewall over Dade County, was *entirely* due to mesovortices rather than the representative winds in a very extreme, compact, Cat-5 hurricane. The 1935 hurricane, being even smaller and more intense (892 mb, if not lower) than Andrew, is a very good candidate for Patricia- or Haiyan-league winds—185 to 190 knots at landfall on the FL Keys.

One key note about Andrew: the extreme wind-caused damage swaths were co-located with strong, N-eyewall (hence RMW) convection, suggestive of *potentially* representative winds.

Re: David and Hugo, I concur with your estimate for Hugo, but perhaps not for David. Even visible satellite imagery reveals that David had a very bright, white CDO, extremely impressive, circular outflow, and stadium-effect eye at the time(s) of its likely peak intensity over the northeastern Caribbean Sea. Aircraft-derived surface estimates and measured flight-level winds, while less reliable in the late 1970s vs. today, suggested winds higher than the pressure-wind relationship indicated. David was not a midget-sized storm at the time, but it certainly wasn't a large one, in terms of wind radii. It was pretty similar to Inez (1966) in size. Interestingly, aircraft infamously reported 700-mb winds of 170 knots in the northern eyewall of Inez, concurrent with rapid intensification and a minimum (observed, likely not the deepest) pressure of 927 mb. David and Inez were both similar in terms of overall size, RMW, observed pressure(s), and intensity trend(s) around peak(s); therefore, one may logically deduce that both were similarly intense, with minimum pressures likely in the low 920s and peak sustained winds of 155 to 160 knots. In other words, Irma- and Maria-type winds, but higher pressures because smaller RMWs and wind radii. Another convincing piece of evidence in favour: Inez and David (like Felix at its peak!) were embedded in high environmental pressures. ...

On another note, Hugo is an interesting case to reanalyse, not just for its (probably 150-knot) Cat-5 peak over the MDR, but also for its Cat-4 landfall in SC. Radar data from Charleston, showing a continued tightening of the eye and intensification of lightning/convective activity, suggest that Hugo continued to deepen after the last aircraft mission, while satellite data from the time of landfall (and before) clearly suggested a higher intensity than 120 knots. Even aircraft-derived winds suggest that 120 knots was a cautious estimate. Given evidence of mesovortices in the eye of Hugo, the minimum pressure of 934 mb at landfall is probably a bit high. Finally, while this is inconclusive, the wind damage in the Bulls Bay–McClellanville area was consistent with a much stronger hurricane. All in all, a solid case can be made that Hugo's true intensity at landfall in SC was in the 130-to-135-knot range, with a pressure of 930-932 mb.

Oh, and I also concur with Chris90 re: Lenny being a solid candidate for Cat-5 status, however briefly it attained that plateau.

Terrific post.

1.) Wasn't Naranja Lakes situated in the southern eyewall of Andrew? Hence the mesovortex explanation for the extreme damage.

2.) Do you have examples of the stronger wind damage from Hugo? Not that I disagree, I'd just really love to see them. And by evidence of mesovortices in the eye, do you mean the satellite imagery?

1) For Hugo's landfall in SC, here are several images, including scenes from Bulls Bay and the Francis Marion National Forest (the latter two locations being inside the right-front RMW). Note, in particular, the image of the smashed car; virtually all the pine trees in the background are snapped and twisted eight to ten feet (???) above ground level. Pine trees, especially in the deep South, are very durable, mature trees being notably sturdy. The same species of pine grew (and still grows) in the areas that Hugo and Andrew devastated. The images from the RMW of Hugo in SC are identical to those from Andrew's RMW in Dade County. The same severity and type of damage to identical vegetation (pines) is evident. More information on Hugo may be found here.

2) As far as Naranja Lakes is concerned, it is directly W of (inland from) Andrew's landfall location in mainland South FL, Fender Point. Andrew was moving quickly westward as it struck Elliott Key, just offshore of Dade County, and then made its second landfall near Homestead. So the very centre of the eye appears to have passed right over Naranja Lakes. Indeed, the minimum pressure on land in South FL, 922 mb, was measured not too far from Naranja Lakes. Thus, Naranja Lakes appears to have experienced the W and E eyewalls, not the N and S ones; however, convection in the northern eyewall rotated SW into Naranja Lakes. Dr. Theodore "Ted" Fujita conducted a study on this and identified this N-eyewall convection as the apparent source of the mesovortices that he deemed responsible for the extreme damage swaths.

[1900hurricane]

Regarding Joan '88, I created a plot from the 20th Century Reanalysis v2 dataset to estimate OCI and ROCI to use with KZC. For OCI, I have 1008 mb, and for ROCI, I have 285 nm. For storm speed, I took the distance between the 00Z and 06Z best track points and divided by six for the six hour average speed, which ended up being 6.83 kt. L is 11.9ºN from best track, and P is 932 mb also from best track. When all plugged into KZC, expected Vmax is a cool 130 kt even.

>>> from KZCeq import KZC, KZCroci
>>> p = 932
>>> c = 41 / 6
>>> c
6.833333333333333
>>> 9.5 * 30
285.0
>>> roci = 285
>>> l = 11.9
>>> oci = 1008
>>> p0 = 1000
>>> v = 100
>>> while p0 >= p:
v = v + 0.1
p0 = KZC(KZCroci(v, c, roci, l), oci)


>>> print('%3.0f' % v)
130

I've been playing with the ROCI version of KZC a little bit, and for the most part it appears to match up very well with the TS wind radius version that is operationally used. It does diverge a little bit towards the 'windier' out beyond 155 kt or so, but not exceptionally so, and I actually prefer the ROCI version when it comes to weaker storms that sometimes have inconsistent wind radii. Naturally, I then pulled up some old surface charts of past storms and computed away. Here's a few that I did.

1900hurricane: p = 936, c = 13.8, roci = 250, l = 29.1, oci = 1009, v = 121 kt
(p = 931) v = 127 kt
Andrew '92: p = 922, c = 17.3, roci = 113, l = 25.5, oci = 1013, v = 148 kt
Celia '70: p = 945, c = 13.3, roci = 120, l = 27.6, oci = 1009, v = 114 kt
Camille '69: p = 900, c = 14.0, roci = 120, l = 30.3, oci = 1003, v = 156 kt
Carla '61: p = 931, c = 4.33, roci = 375, l = 28.0, oci = 1004, v = 114 kt

[CrazyC83]

The posts in this forum are NOT official forecast and should not be used as such. They are just the opinion of the poster and may or may not be backed by sound meteorological data. They are NOT endorsed by any professional institution or STORM2K. For official information, please refer to products from the National Hurricane Center and National Weather Service.

As for Gilbert, if I reanalyzed it based on data I have seen, here's what it would look like:

AL081988, GILBERT, 49,
19880908, 1800, , TD, 12.0N, 54.0W, 25, 1008, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880909, 0000, , TD, 12.7N, 55.6W, 25, 1007, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880909, 0600, , TD, 13.3N, 57.1W, 30, 1006, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880909, 1200, , TS, 14.0N, 58.6W, 35, 1004, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880909, 1800, , TS, 14.5N, 60.1W, 40, 1001, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880910, 0000, , TS, 14.8N, 61.5W, 45, 998, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880910, 0600, , TS, 15.0N, 62.8W, 55, 994, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880910, 1200, , TS, 15.3N, 64.1W, 60, 992, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880910, 1800, , HU, 15.7N, 65.4W, 70, 987, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880911, 0000, , HU, 15.9N, 66.8W, 85, 979, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880911, 0600, , HU, 16.2N, 68.0W, 90, 976, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880911, 1200, , HU, 16.1N, 69.5W, 100, 975, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880911, 1800, , HU, 16.2N, 70.7W, 100, 970, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880912, 0000, , HU, 16.8N, 72.0W, 105, 964, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880912, 0600, , HU, 17.3N, 73.7W, 110, 962, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880912, 1200, , HU, 17.6N, 75.3W, 115, 959, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880912, 1700, L, HU, 17.8N, 76.8W, 120, 957, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880912, 1800, , HU, 17.9N, 76.9W, 115, 958, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880913, 0000, , HU, 18.2N, 78.5W, 110, 962, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880913, 0600, , HU, 18.5N, 79.7W, 120, 952, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880913, 1200, , HU, 18.8N, 81.1W, 140, 934, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880913, 1800, , HU, 19.4N, 82.5W, 165, 905, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880914, 0000, , HU, 19.7N, 83.8W, 170, 884, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880914, 0600, , HU, 19.9N, 85.3W, 160, 889, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880914, 1200, , HU, 20.4N, 86.5W, 155, 892, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880914, 1500, L, HU, 20.7N, 87.0W, 150, 900, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880914, 1800, , HU, 20.9N, 87.8W, 120, 925, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880915, 0000, , HU, 21.3N, 89.5W, 85, 944, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880915, 0600, , HU, 21.6N, 90.7W, 85, 949, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880915, 1200, , HU, 21.9N, 91.7W, 85, 950, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880915, 1800, , HU, 22.1N, 92.8W, 90, 950, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880916, 0000, , HU, 22.5N, 93.8W, 95, 948, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880916, 0600, , HU, 22.9N, 94.8W, 100, 944, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880916, 1200, , HU, 23.7N, 95.9W, 105, 943, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880916, 1800, , HU, 23.9N, 97.0W, 100, 947, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880916, 2200, L, HU, 24.2N, 97.8W, 90, 955, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880917, 0000, , HU, 24.4N, 98.2W, 75, 964, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880917, 0600, , TS, 24.8N, 99.3W, 50, 988, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880917, 1200, , TS, 25.0N, 100.5W, 35, 996, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880917, 1800, , TD, 25.4N, 101.9W, 30, 1000, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880918, 0000, , TD, 26.0N, 103.2W, 30, 1002, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880918, 0600, , TD, 27.6N, 103.7W, 30, 1004, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880918, 1200, , TD, 29.3N, 102.6W, 25, 1003, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880918, 1800, , TD, 31.5N, 101.3W, 25, 1003, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880919, 0000, , TD, 33.2N, 99.7W, 25, 1002, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880919, 0600, , TD, 35.8N, 97.7W, 25, 1001, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880919, 1200, , TD, 37.7N, 93.2W, 25, 999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880919, 1800, , EX, 40.2N, 89.9W, 25, 998, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,
19880920, 0000, , EX, 43.4N, 86.5W, 25, 995, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999, -999,

Some reasoning:

* Limited Recon data suggests that Gilbert intensified a bit more earlier in its life due to lower pressure measurements. Hence the intensities on the 10th and 11th were largely increased.
* The intensity at Jamaican landfall is estimated at 120 kt. That is based on surface data of 100-105 kt sustained winds, which may have very well been 10-min winds, and the limited spatial observations. That is despite a fairly high pressure (analyzed at 957 mb). Slight weakening occurred thereafter, probably due to an ERC, as the pressure filled a bit per Recon.
* The lowest pressure was 888 mb at 2130Z on September 13. The next Recon penetration was not until about 0500Z on September 14, when 894 mb was measured, and then the following pass measured 890 mb. Due to the difficulty of measuring low pressures in extreme conditions, 889 was maintained at 0600Z. The minimum pressure was hence analyzed at 884 mb at 0000Z, assuming additional deepening after the 888 report.
* There is also uncertainty in the peak intensity - I went with 170 kt, although a case for an even higher intensity could be made. That is based on the extremely low pressure and the 173 kt measured at 700 mb, which in a rapidly deepening storm the 90% rule may be very conservative. Those winds were also not particularly reliable in 1988. I could have gone with 175 or even 180 kt, knowing that Dvorak was an easy T8.0, which is unique in the record.
* At Yucatan landfall, I analyzed an intensity of 150 kt. That is higher than HURDAT, based on flight level winds around 155 kt and the 900 mb pressure.
* In the Gulf, uncertainty is quite high. Flight level winds were all over the place - ranging from 85 kt to 115 kt. Yet the pressure only dropped a bit to about 943 at its minimum, and then rose to 955 at landfall. Dvorak largely ran at T5.0 or T5.5. There is nothing to suggest it was a cat 4 (as with the BT), hence the intensities come down. I have it emerging from the Yucatan with 85 kt winds, increasing to a final peak of 105 kt, then down to 90 kt at final landfall. Those are very uncertain estimates though.

[1900hurricane]

Did you know the 888 mb in Gilbert wasn't actually measured? It's actually an estimate based on the 700 mb extrapolated pressure which was then bias corrected. The paper by Willoughby, Masters, and Landsea goes into detail how the value was derived. It's quite an interesting read.

[CrazyC83]

Did you know the 888 mb in Gilbert wasn't actually measured? It's actually an estimate based on the 700 mb extrapolated pressure which was then bias corrected. The paper by Willoughby, Masters, and Landsea goes into detail how the value was derived. It's quite an interesting read.

Interesting findings. I didn't realize there was a pass at 23Z or so, but that decrease may be conservative if it missed the center. That would go well with an estimate of 884 at 00Z (I know in Patricia, they had winds as strong as 60 kt in the passes establishing the minimum pressure, and that would translate down to 885-886).

[1900hurricane]

For me personally, I'd probably keep Gilbert '88's 888 mb as is, although it isn't impossible that it may have been a couple millibars lower (I wouldn't go lower than 886 mb though). Using 20th Century Reanalysis, I was able to derive an outermost closed isobar of 1005 mb with an average radius of about 330 nm. CSU's ebtk lists the OCI at 1008 mb with an r of 510 nm, but that looks too large to me (even large WPac super typhoons rarely get that big), so I stuck with the data I derived myself via 20c v2. Using a version of KZC that derives the S parameter from the ROCI (which I have tweaked/bias corrected so it behaves better at extreme intensities), I end up getting an expected Vmax of 166 kt. If Patricia '15's 700 mb to surface wind reduction of about 0.95 near peak intensity is used with Gilbert's 173 kt flight level wind, almost the same expected Vmax is derived (about 164 kt). Based on this, I'd set Gilbert's peak intensity to 165 kt/888 mb.

>>> oci = 1005
>>> roci = 330
>>> p = 888
>>> l = 19.7
>>> c = 76 / 6
>>> v = 100
>>> p0 = 1000
>>> while p0 > p:
... v = v + 0.1
... p0 = KZC(KZCroci(v, c, roci, l), oci)
...
>>> print('%3.0f' % v)
166
>>> 173 * 0.95
164.35