2024 Post-Season Changes

[ElectricStorm]

What storms do you think will (or should) get a post season upgrade/downgrade? This can be for any basin, not just the Atlantic. I'll start:

Atlantic:
Ernesto: Second peak was likely underestimated, was probably a little stronger than its first peak. I'd bump it up to 90-95kts but lack of direct evidence will probably prevent NHC from doing so, which is understandable since looks don't always tell the whole story.

Francine: I know there's been some debate about the 85kt peak but to me it looks good. Strongest winds like usual were likely over a very small area and may have remained offshore. I don't think we'll see any changes.

94L: Was likely a short-lived TD or minimal TS, would like to see them add it post-season.

96L (added 9/22): Also likely a TC, would like to see it added post season.

EPAC:
Gilma: First peak was likely a solid Cat 4, I'd bump it up slightly. Second peak was probably mid-upper Cat 4, I'd go 125-130kts but again this is just based on looks and without any other evidence I doubt NHC changes the 115kt peak estimate, which again is understandable.

WPAC:
Gaemi: Was likely a super typhoon, and the 125kt estimate is a touch too low IMO. I'd go 130-135kts. I've seen some talk it way have been stronger but since the eye never fully cleared I'm hesitant to go any higher. Probably

Shanshan: Also likely underestimated a bit since cloud tops were generally warm which held Dvorak back. I'd go 125-130kts but once again lack of evidence will likely prevent anything above 115kts.

That's what I have so far, may add more when I look a little more in depth.

[Teban54]

While it's fresh in our memories, I think Gordon's upgrade to a TS may be moved to a little bit earlier, though I can see the argument either way. This is obviously not something most people would care about, but I'm just jotting it down here before we forget lol.

I also think 92L also has as convincing of an argument for a weak TD/TS as 94L does.

[Hurricane2022]

Postseason changes I'd like to see, Part I.
NATL.

BERYL: 145 kt ---> 150 kt. I would say that the data collected by the recon on Beryl's intensity peak supports something around 150 kt, (and not to mention that the recon had probably already missed the intensity peak by 3 - 4 hours, based on the data sent by the plane and satellite images, but I imagine that doesn't make much of a difference in the end). That said, I imagine the NHC will likely maintain Beryl at operational peak intensity.
ERNESTO: 80 kt ---> 90 kt (peak 2). It is very likely that Ernesto returned to category 2 in its second peak of intensity and probably surpassed the strength of the first peak in terms of maximum winds. But the NHC may also consider reasons that would make them keep Ernesto in category 1 during the second peak.
---------------
EPAC.

GILMA: Like several major hurricanes that have wandered in the East Pac in recent years, Gilma was also likely underestimated by the NHC.
Peak 1: 110 kt ---> 115 kt.
Peak 2: 115 kt ---> 125 kt. She probably had his strongest peak intensity around 120 - 130 kts, not at 115 kts, as automatic estimates have failed considerably in recent years.
Peak 3: 95 kt ---> 105 kt.
---------------
WPAC.

GAEMI: 125 kt/919 mbar ---> 140 kt/895 mbar. Buoys and Radar data can support winds of a low-end C5, and an incredibly low minimum pressure between 900 - 885 mb, despite an ragged appearance normally seen in low/mid-end C4 cyclones.
SHANSHAN: 115 kt ---> 125 kt. Probably underestimated by satellite estimates due to somewhat warm cloud tops at peak.
YAGI: 130 kt ---> 140 kt (peak 2). JTWC likely underestimated Yagi at its second peak, despite strong support from Dvorak agencies that mostly peaked around T7.0.

[Sciencerocks]

Add 92L as a tropical depression or borderline storm.
90L off Texas was also a tropical storm and more deserving then 92l and 94l based on observations alone.
and I'd also consider 94L but slightly less data compared to 92L at least in ascat terms but ascat mostly just missed it but if it hadn't about equal. lol

[aspen]

Beryl: I’d go with 150 kt and 931 mbar between 8-10z July 2nd. These super high SFMR readings in relatively high-pressure Cat 5s are always a little suspect, but blending the ~165 kt peak FL winds and ~165-170 kt peak SFMR winds would support a conservative estimate of 150 kt. An argument could be made for 155 kt, though there’s no way the NHC would go that high for a >930 mbar system. Really remarkable that a system of that pressure was able to pull off FL/SFMR readings almost on-par with Dorian. 931 mbar is based on the first eye drop of that morning flight; the pressure rose a bit later on, which is why the NHC went with 934 mbar for 8am.

Gilma: First peak was likely 115 kt; second peak was probably around 125-130 kt based on the ~T6.5 IR presentation and clear eye/stadium effect combo only seen in strong Cat 4s and above; third peak was likely 100-105 kt. I do not think Gilma approached Cat 5 intensity because its second peak was quite brief, unlikely enough time for the winds to catch up to Cat 5 intensity. Out of the recent under-estimated EPac Cat 4s (Douglas, Felicia, Darby, Dora, Gilma), this was easily the most egregious and highlighted how much Dvorak (both subjective and ADT) have been struggling with many majors in the 2020s.

I haven’t been looking at the WPac much but Gaemi is an easy Cat 5 based on those surface observations. Yagi was probably at or just below Cat 5 intensity too, and its peak was earlier (yet another instance of the JTWC assessing a storm’s peak after it’s started weakening, see Bualoi 2019).

[Beef Stew]

I feel that an upgrade to 150 kts for Beryl is very likely- if I remember correctly, the NHC even mentioned something along the lines of Beryl peaking in between advisories, and that it would be reflected in the TCR.

Other than that, I agree with others that Ernesto’s second peak was likely higher, and this may be upgraded in the TCR. I definitely think that 92L met the definition of a tropical cyclone (albeit briefly) and should’ve been classified, but I don’t expect there to be a change here.

[Category5Kaiju]

I feel that an upgrade to 150 kts for Beryl is very likely- if I remember correctly, the NHC even mentioned something along the lines of Beryl peaking in between advisories, and that it would be reflected in the TCR.

Other than that, I agree with others that Ernesto’s second peak was likely higher, and this may be upgraded in the TCR. I definitely think that 92L met the definition of a tropical cyclone (albeit briefly) and should’ve been classified, but I don’t expect there to be a change here.

The one thing about Beryl I'm very curious to see is if its minimum central pressure is lowered. Iirc there were talks of recon believing that the pressure may have tanked lower than 934 mbar in between passes? Would be interesting to see, assuming that is the case, how low, considering Category 5 hurricanes typically have minimum pressures in the 920s or lower.

[ElectricStorm]

Added 96L to the list, was likely a sheared TC

[Sciencerocks]

I'd add 96L as a 40 knt tropical storm. There's no question in my mind. this thing in any other season would have had advisories and a name. No question.

[MHC Tracking]

Changes I would agree with in post:

92L:
Add as a short-lived TD. The LLC was confirmed as closed and tight by reconnaissance aircraft, and the continual bursting of convection when inland (as opposed to thunderstorm Danny )lends credence to this being a TC.
Beryl: 145 knots ->150 knots
Blend of 750 mb FL data and SFMRs supports closer to 155, but as SFMRs appeared to have a high bias through much of Beryl's life, would go with 150.
Beryl's Texas landfall: 70 knots->75 knots
Impressive surface observations for a supposedly low-end category 1. May have intensified a little in the first hour inland according to KHGX wind data.
Debby:
Time off US East Coast changed to SS. Little convection after it emerged off Florida and until it made its Carolina landfall, what convection there was, was heavily displaced from the center and poorly-organized.
Ernesto:
Peak 1 downed to 80 knots, peak 2 upped to 85-90. Peak 1 never had SFMRs above 78 knots, despite high FLs. Dry air likely destroyed mixing. As for peak 2, both SMAP and SAR supported winds of 90-95, in addition to ADT values near 100 knots. NHC doesn't use SMAP, but they have been utilizing SAR data as of late.
Gordon:
TS point pushed back by about a day. MW indicated a weak low-level core 24 hours before the system was officially upgraded.
96L:
Add as a 30 knot TD. Persistent, albeit sheared convection around a well-defined LLC. Fits the bill for a TC.
Will post WPAC and EPAC later.

[Teban54]

An upgrade in Helene's landfall intensity is very possible.

On the other hand, it's time to also add some serious upward adjustments to Isaac and Joyce to this list, if no operational upgrades (especially for Isaac) happen soon.

[kevin]

Let's give it a shot, I'll do the Atlantic first.

-- Named storms --

Alberto: TCR already out, 45 kt sounds good. -> no change

Beryl: As NHC already noted, peak intensity was between advisories and will be fixed in the post-season. Recon supports 150 - 155 kt, but 155 might be a bit generous and NHC is often a bit more conservative. So 150 kt sounds like a good compromise. -> upgrade from 934mb/145kt to 931mb/150kt

Chris: TCR already out, the increase to 40 kt in that report looks good to me. -> no change

Debby: Downgrade from TS to STS after landfall. No change to peak intensity. -> no change

Ernesto: First peak downgraded 85 -> 80 kt. However, second peak upgraded 80 -> 90 kt. Dvorak even supports 95-100kt, but NHC seems to have their reasoning for not giving it much weight operationally and will likely do the same in the post-season analysis. So in the end peak intensity upgrade with a lower pressure of 962mb instead of 968mb. -> 2nd peak stronger than 1st peak, peak intensity 968mb/85kt -> 962mb/90kt

Francine: There has been a lot of discussion regarding a downgrade to 80 kt. I saw all of the beautiful analyses in the Francine thread and it really is a toss-up, most estimates are in the 80 - 85 kt range. For now I'll say no change, but if any storm gets downgraded this year it's Francine. -> no change

Gordon: TS time period extended to 6 - 12 hours before it was operationally named. No change to peak intensity. -> no change

Helene: 136 kt FL before landfall translates to 122 - 123 kt on the ground. NHC themselves already said 'might be a bit conservative' regarding the 120 kt landfall intensity in their own discussion. Furthermore, 10-second winds of 137 kt were measured by the NHC drone just above the surface. Conversion factors for 10-sec to 1-minute are about 1.17 based on the Durst curve. But NHC uses slightly lower conversion factors for on-shore winds near the coast. F.e. Durst curve would say 1.19 for 10-min to 1-min, but NHC says 1.11 for on-shore winds near coast for such a conversion. So the true conversion factor they'll use for 10-sec to 1-minute is probably closer to 1.09 instead of 1.17. This would mean that the dropsonde measurement goes down to 126 kt. All data combined an upgrade to 125 kt is reasonable, though not required. -> upgrade from 120 kt to 125 kt, no change to pressure

Isaac: Looks like it will be upgraded to a cat 2 soon, but the cat 2 period should be expanded. Dvorak supports 90 - 95 kt, I'd go with 90 kt for now. -> upgrade to 90 kt cat 2, but that might already happen operationally

Joyce: I expect upgrades later today, but it's stronger than operationally analyzed atm. Most likely 50 - 55 kt instead of 45 kt right now. -> upgrade to 50 - 55 kt, but that may already happen operationally

-- Potential systems --

90L (June): ASCAT showed 30 kt winds when it was east of Florida. Very maybe a TD for a few hours, but not a TS anyways so I expect no change. -> no change
92L (June): upgrade to a 30 kt TD. Recon found a closed center, consistent convection for 6+ hours. 37 kt SFMR and 32 kt FL blends to 33 kt, but I don't expect them to go to 35 kt for such an edge case. -> upgrade to a 30 kt TD
93L (June): not a TC. -> no change
96L (July): remained a wave for its duration, no upgrade. -> no change
99L (September): plenty of wind, but a frontal low with a naked swirl. -> no change
90L (September): ASCAT showed 35 kt winds, ground measurements indicate a closed LLC and there was plenty of convection. However, since the NHC probably sees it as part of the precursor to Francine, I don't expect an upgrade. -> no change
92L (September): ASCAT had 30 kt winds. Sustained consistent convection for 2 days. Should be upgraded to a 30 kt TD imo. -> upgrade to a 30 kt TD
94L (September): pulsing convection, looked like a TD/TS. But lack of good ASCAT passes makes me doubt any upgrade. -> no change
96L (September): sheared, but consistent convection for 2 days. T2.0 - 2.5, 35 kt according to D-PRINT. Was probably a 35 kt TS. Not sure if NHC will upgrade it, but if we're getting another TS in the post-season this should be it imo. -> upgrade to a 35 kt TS

[Teban54]

Milton's peak intensity will fuel months of debate here. Kirk's too, to a lesser extent, although unfortunately it has now been overshadowed by Milton big time.

[Hurricane2022]

Post-season changes I'd like to see, Part II.

ISAAC: 90 kt --> 95 kt. Isaac's best IR Dvorak symmetry probably supported winds somewhere around 95 kts. There are those who argue that 100 kts could also be considered but the cold waters that Isaac crossed at peak probably affected the way the strongest winds reached the surface. For this same reason and the lack of support from Dvorak agencies it is possible that the NHC mantains the operational estimate. Oh, I also would push Isaac TCG by 1 day earlier than the actual estimate.

JOYCE: 45 kt --> 50 kt. SAR passes and microwave images easily supported winds somewhere around 50 - 55 kts but the NHC preferred to keep Joyce at 45 kts. I would also move the TCG by around 1 day earlier than previously analyzed.

KIRK: 125 kt --> 135 kt. Dvorak supports 140 kts but as only ADT did so, I can't see any other more concrete evidence that supports an upgrade to Cat5 in the TCR. In this case, an upgrade to 135 kts seems quite reasonable.

LESLIE: 80 kt --> 85 kt (Peak 1). Some SAR passes could support 100 - 105 kts but the NHC still has very little confidence in this technology, unlike the JTWC. But it's probably proof that Leslie may have been a little stronger than they had thought. So I assume that 85 kts is also a pretty reasonable estimate.


I also considered expressing some opinions about Helene and Milton but I think the NHC worked very well with the data that Recon collected on these two. Maybe I can see the NHC lowering Milton's min. pressure to 896/895 mbar but I don't think it's necessary, as it was also noted in an excellent analysis by our ProMet friend CrazyC83 at the end of Milton's thread.

[ljmac75]

My uninformed opinions:

Changes I think might happen:

Helene landfall intensity: 120 kts, 938 mbar ---> 120 kts, 940 mbar
I think 120 kts will remain as the landfall intensity, but I think the estimated central pressure was slightly too low. The lowest recorded pressure on land was 943 mbar near the landfall point, so 940 might still be too low. However, another station recorded 947 about an hour and 40 minutes later. I might be wrong, but only 4 mbar of filling in that time seems a little low, so I am assuming the landfall pressure was somewhat below 943 (I might be talking nonsense here).

Beryl peak intensity: 145 kts, 934 mbar ---> 150 kts, 931 mbar
Satellite.

Milton landfall intensity: 105 kts, 954 mbar ---> 100 kts, 954 mbar
Maybe I'm being generous, but I think they'll keep it as a major hurricane.

Changes that I don't think will happen but are possible:

Milton peak intensity: 155 kts, 897 mbar ---> 160 kts, 895 mbar
The recon data didn't really seem too different from other 185 mph storms, although I'm drawing from a sample size of two here, and it's conceivable that the storm was slightly stronger before recon entered after it exploded in intensity. Also this is nonsensical reasoning but every cat 5 in the gulf since 2005 (Katrina, Rita, Michael, Ian) has had its winds upped by 5 kts in post-analysis and I like to think the pattern will continue.

[ljmac75]

Oscar Landfall intensity: 70 kts,986 mbar ---> 80 kts, 980 mbar
80 kts might still be too low but the NHC doesn't really like to change intensities by that much I've noticed.
Kristy Initial peak: 135 kts, 928 mbar ---> 140 kts, 925 mbar
I don't know if they'll change it since the peak may have been non-synoptic.

[kevin]

Let's give it a shot, I'll do the Atlantic first.

-- Named storms --

Alberto: TCR already out, 45 kt sounds good. -> no change

Beryl: As NHC already noted, peak intensity was between advisories and will be fixed in the post-season. Recon supports 150 - 155 kt, but 155 might be a bit generous and NHC is often a bit more conservative. So 150 kt sounds like a good compromise. -> upgrade from 934mb/145kt to 931mb/150kt

Chris: TCR already out, the increase to 40 kt in that report looks good to me. -> no change

Debby: Downgrade from TS to STS after landfall. No change to peak intensity. -> no change

Ernesto: First peak downgraded 85 -> 80 kt. However, second peak upgraded 80 -> 90 kt. Dvorak even supports 95-100kt, but NHC seems to have their reasoning for not giving it much weight operationally and will likely do the same in the post-season analysis. So in the end peak intensity upgrade with a lower pressure of 962mb instead of 968mb. -> 2nd peak stronger than 1st peak, peak intensity 968mb/85kt -> 962mb/90kt

Francine: There has been a lot of discussion regarding a downgrade to 80 kt. I saw all of the beautiful analyses in the Francine thread and it really is a toss-up, most estimates are in the 80 - 85 kt range. For now I'll say no change, but if any storm gets downgraded this year it's Francine. -> no change

Gordon: TS time period extended to 6 - 12 hours before it was operationally named. No change to peak intensity. -> no change

Helene: 136 kt FL before landfall translates to 122 - 123 kt on the ground. NHC themselves already said 'might be a bit conservative' regarding the 120 kt landfall intensity in their own discussion. Furthermore, 10-second winds of 137 kt were measured by the NHC drone just above the surface. Conversion factors for 10-sec to 1-minute are about 1.17 based on the Durst curve. But NHC uses slightly lower conversion factors for on-shore winds near the coast. F.e. Durst curve would say 1.19 for 10-min to 1-min, but NHC says 1.11 for on-shore winds near coast for such a conversion. So the true conversion factor they'll use for 10-sec to 1-minute is probably closer to 1.09 instead of 1.17. This would mean that the dropsonde measurement goes down to 126 kt. All data combined an upgrade to 125 kt is reasonable, though not required. -> upgrade from 120 kt to 125 kt, no change to pressure

-- Potential systems --

90L (June): ASCAT showed 30 kt winds when it was east of Florida. Very maybe a TD for a few hours, but not a TS anyways so I expect no change. -> no change
92L (June): upgrade to a 30 kt TD. Recon found a closed center, consistent convection for 6+ hours. 37 kt SFMR and 32 kt FL blends to 33 kt, but I don't expect them to go to 35 kt for such an edge case. -> upgrade to a 30 kt TD
93L (June): not a TC. -> no change
96L (July): remained a wave for its duration, no upgrade. -> no change
99L (September): plenty of wind, but a frontal low with a naked swirl. -> no change
90L (September): ASCAT showed 35 kt winds, ground measurements indicate a closed LLC and there was plenty of convection. However, since the NHC probably sees it as part of the precursor to Francine, I don't expect an upgrade. -> no change
92L (September): ASCAT had 30 kt winds. Sustained consistent convection for 2 days. Should be upgraded to a 30 kt TD imo. -> upgrade to a 30 kt TD
94L (September): pulsing convection, looked like a TD/TS. But lack of good ASCAT passes makes me doubt any upgrade. -> no change
96L (September): sheared, but consistent convection for 2 days. T2.0 - 2.5, 35 kt according to D-PRINT. Was probably a 35 kt TS. Not sure if NHC will upgrade it, but if we're getting another TS in the post-season this should be it imo. -> upgrade to a 35 kt TS

Updated with Atlantic systems after Helene since Isaac and Joyce were still active during my last post.

Isaac: A case could be made for 95 kt, but I think NHC will keep it at 90 kt. Isaac's time period as a TC will be extended. -> no change

Joyce: There was quite some SAR and Dvorak clearly supported 50 - 55 kt, so I was a bit surprised the NHC kept it at 45 kt. I do expect a 5 kt upgrade in the post-season analysis. -> upgrade from 1001mb/45kt to 997mb/50kt

Kirk: Now this is a difficult one. Dvorak clearly supported 140 kt, but I don't expect a 15 kt upgrade by the NHC. 130 - 135 kt is more realistic, but if we ever get a 15 kt TCR upgrade this is the one. -> upgrade from 934mb/125kt to 924mb/135kt

Leslie: ADT supports 90 kt for the secondary peak, so I don't expect any changes there. In a similar fashion, ADT supports 85 kt for the first peak, which might thus mean a 5 kt upgrade during that time period. -> no change to peak intensity

Milton: Had pretty great recon coverage, but both peak intensities were measured at the time when subsequent passes showed an increasing pressure. Based on some intuition and a ton of analyses in the Milton thread, I expect a very slight upgrade to the lowest pressure but no change to peak winds. Landfall intensity might be reduced from 105 kt to 100 kt. -> upgrade from 897mb/155kt to 895mb/155kt

Nadine: Seems reasonable to me. ADT might indicate a lower landfalling pressure, but there are no surface observations I found that support that so I expect this one to remain the same. -> no change

Oscar: Time period as a TC is moved forward with more gradual intensification, because a random hurricane popping up out of nowhere doesn't make sense. Peak intensity should be increased to an 85 kt cat 2 imo. A case could be made for an even higher intensity. -> upgrade from 986mb/80kt to 978mb/85kt

[Teban54]

I also think it's well within the realm of possibility that Oscar's genesis will be moved much earlier, when it was Invest 94L near Cabo Verde, before it degenerated and moved further west as a remnant low.

At that time, it had 35 kt ASCAT and a closed circulation, and many people suspected the main reason for the lack of upgrade was due to its short life span. But since the same system eventually became Oscar anyway, designating this first phase as a TC wouldn't be a "name waster".

[sasha_B]

With the caveat that I'm a total amateur and I doubt the NHC would ever go for a post-season correction this dramatic, I think Oscar has a case for 90-100 knots and 975-965 hPa. Some here are saying that 85 knots is a generous estimate, but between SAR, Dvorak, and surface radar, it's hard for me to believe that it was anything less than a solid Category 2 at landfall. In addition to individual indirect observations (satellite, radar, etc.) we can also look at the rate of change in its presentation on those instruments, which was quite rapid and pronounced in the hours leading up to landfall, with intense lightning around the center + indications of a closed eyewall before it fell apart on landfall; those sorts of improvements are generally indicative of RI, not minor fluctuations in intensity. In a word, I think the NHC would be entirely justified in going 85 knots at least.
As for the other storms for which intensity was in question this season, I don't think Kirk will (or should) be upgraded to a SSHWS Cat 5, but 135 knots and a much lower MSLP minimum (something like 930-924 hPa) would make sense; even if there's insufficient evidence for C5 sustained winds, what we saw on satellite was indicative of a deeper system than the provisional best track shows.
I very much doubt they'll drop Milton to Category 2 on landfall; surface pressure and the direct measurements from the final recon mission into the storm fully support 100-105 knots. Sampling of wind measurements after landfall is always distorted - just like with Helene, the fact that there wasn't an anemometer on the ground picking up C3 sustained winds is not sufficient to throw out the recon measurements from 'the moment before' landfall. Changes to Milton's peak will probably be minimal, because differences in intensity are marginal at that point; dropping the MSLP minimum to 893-895 hPa seems possible, with an increase to max winds less likely. (The same goes for Beryl, which could have briefly reached 931-928 hPa and 150 knots after recon left. It may have been weaker than operationally assessed on landfall in the Yucatan but perhaps not sufficiently for a post-season adjustment.)
For Helene, no change to central pressure at landfall + an increase to 125 knots for max winds seems to have already been telegraphed by the NHC when they acknowledged that their operational wind estimate was "possibly conservative" - this is the change I'm most confident will actually be made in the TCR, as there's much more uncertainty in the other cases.

[Travorum]

Regarding Milton's central pressure, I went ahead and modelled a polynomial regression using the recon data from just before and just after the peak. The one change I made was changing the 899mb VDM to the 897mb that was given in the next advisory. I tried to minimize the residual of the 897mb data point as I figured that would give the best estimate of the central pressure just prior to that point, and a 6th order polynomial regression was the lowest order that managed to produce a reasonably low residual at that point (any lower order overshot 897mb and wouldn't get down below 900mb at any point in the curve). I also plotted an 8th order polynomial, both are shown below:

6th order polynomial regression	8th order polynomial regression

So, based on a high order polynomial regression the minimum central pressure could have been as low as 893-890mb, at roughly 2000-2030z on October 7th.

---

However, high order polynomial regression is not always desirable and has a tendency to overshoot the ground truth data (Runge's phenomenon), especially in larger gaps between data points like the gap between recon flights. To sanity check this, I also calculated a cubic spline interpolation, which uses a piecewise spline approximation of third order polynomials, thus avoiding potential overshoot. Credit to Timo Denk (https://tools.timodenk.com/cubic-spline-interpolation) for making an online calculator and saving me an hour writing an implementation in MATLAB. (Note that this is interpolation of the points and not a regression model).

cubic spline interpolation

Using this method results in a minimum central pressure of 896mb at roughly 2200z, which is less than 30 minutes prior to recon arriving and measuring a sub-900mb pressure. If this method is to be trusted it may indicate that a high order polynomial approximation is over(under?)shooting the minimum pressure by a few mb, and that recon really did measure right near the peak on that first pass.

---

Now, a sanity check and a thought experiment. When recon left, the last VDM recorded 912mb and had been dropping at an average of 8mb/hr over the past three hours (and 9.5mb/hr between the last two passes). When the next recon arrived, pressure was rising at a rate of 5.5mb/hr over the next 3 hours. This to me suggests that the peak was more likely to be towards the middle of the recon gap rather than right prior to the later recon flight arriving. I wanted to simulate a near-best case scenario for Milton, so I assumed a 9.5mb/hr rate of deepening continued for the next 1.5 hours after recon left Milton (this is perhaps optimistic but it gives a data point of 898mb which is about as best case as I can go without explicitly assuming Milton exceeded its operational minimum pressure, and 1.5 hours was conveniently about the frequency of recon fixes at that point). I redid both methods of regression with this simulated recon fix added into the dataset:

6th order polynomial regression	8th order polynomial regression

cubic spline interpolation

So, based on this best case simulation, using the polynomial regression results in an estimated minimum central pressure of 890-891mb and the cubic spline interpolation gives an estimated minimum central pressure of 893mb, both at around 2030z Oct 7th.

---

Now what is my gut feeling about all this? Above all else my gut says that relying on regression models as simple as these to model dynamics as complicated as a hurricane won't give a super accurate estimate of the minimum central pressure, especially an edge case like Milton of a sub-900mb Cat 5 with a 5nm pinhole eye. The pressure rate of change in such a small intense eye cannot be assumed to follow a polynomial curve between points in the 5 hour recon gap when the instantaneous rate of change on recon departure was 9.5mb/hr, which is nearly unprecedented in the Atlantic basin. It is possible that rate of change continued and Milton bottomed out below 890mb, and it is also possible that the pressure quickly plateaued and ended up around 895mb. Still, I do think that the simplified regression and interpolation models are useful enough to provide a broad range and given the 9.5mb/hr rate of deepening on recon departure and the 5.5mb/hr rate of filling on return, I am more inclined to believe that the pressure followed a steeper curve between recon. Therefore, my gut feeling looking at the data is that Milton's lowest central pressure was likely between 890mb-895mb, and probably towards the upper bound of that range.

An entirely different question is what the NHC will place Milton's pressure at in the TCR. In this regard Rita's TCR provides a remarkably good analogue. With Rita, recon measured via dropsonde 899mb w/32kt winds at 2309z 21 September, almost exactly similar to Milton's peak. There was then a six hour gap in recon before the next flight found 898mb at 0538z and 899mb at 0715z 22 September. Based on this Rita's intensity was estimated as 895mb at the non-synoptic point of 03z 22 September. Rita has a bit stronger of a claim to a peak lower than recon measured given that sub-900mb pressures were measured on both sides of a recon gap, but I would think that given Milton's rate of pressure change on either end of its recon gap will be sufficient for the NHC to lower Milton's central pressure a few mb in the TCR. Given that the NHC is unlikely to break Rita's record of lowest Gulf pressure without more conclusive evidence than Rita itself had, my guess is that the NHC puts Milton's pressure at 895mb in the TCR.