NHC updated "Deadliest, Costliest, Most Intense" document

[gymengineer]

... through 2010: http://www.nhc.noaa.gov/pdf/nws-nhc-6.pdf

This edition has significant changes in the damage estimates of all the post-1994 hurricanes on the "Costliest" list. They updated the methodology for including flood damage in the total damage estimate, resulting in a signficant increase in most of the damaging storms. As the authors point out, Allison went from $5 billion to $9 billion in the non-adjusted list, and Katrina goes over $100 billion now on the non-adjusted list.

[HurricaneJosh]

Yep-- I'd seen.

They also significantly upped the damage estimate for Celia 1970, a cyclone which really devastated Corpus Christi and which I believe was a Cat 4.

Another interesting tidbit: they lowered the Katrina 2005 death toll by a few hundred.

The updates are great, and very interesting-- but I can't wait until they finish reanalysis so that the document is completely up-to-date. The ~1931-1989 period remains loaded with the old, incorrect info-- i.e., Gloria was a Cat 3 in NY, Audrey 1957 was a Cat 4 in LA, etc. etc.

[Srain]

Ike damage was the notably increased as well.

[Ginx snewx]

Interesting change of deaths in 38 to 256 with a footnote that says offshore losses to 600, really weird when the official death toll just for RI is 276.

[Ginx snewx]

Interesting change of deaths in 38 to 256 with a footnote that says offshore losses to 600, really weird when the official death toll just for RI is 276.

136 in my hometown area alone, probably 10 miles of coast

[gymengineer]

Yep-- I'd seen.

They also significantly upped the damage estimate for Celia 1970, a cyclone which really devastated Corpus Christi and which I believe was a Cat 4.

Another interesting tidbit: they lowered the Katrina 2005 death toll by a few hundred.

The updates are great, and very interesting-- but I can't wait until they finish reanalysis so that the document is completely up-to-date. The ~1931-1989 period remains loaded with the old, incorrect info-- i.e., Gloria was a Cat 3 in NY, Audrey 1957 was a Cat 4 in LA, etc. etc.

I know I've asked you this before, but now that a few years have passed with presumably more data to pour over-- do you think Hugo actually produced category 4 sustained winds anywhere on the SC coast?

I have been looking at the surface observation data with the larger storms (hurricane-force winds, say, 70 miles or more from the center), and just find it hard to believe, even in our post-1990 examples, that any of these larger storms translate their reported max sustained winds at landfall down to 10-m height on the actual coastline. Jeanne and Wilma both come to mind as recent examples. Jeanne's TCR emphasizes that the 105 knot estimate is an "over-water" estimate, and then used the phrasing "It is possible that wind speeds of near 105 kt may have affected a small area of coast east of Sebastian." I again point to Powell's dissent in the upgrade of Andrew to a Cat 5 landfall--- with the key part of whether Andrew's Cat 5 winds were over water or made it onshore. It seems like the NHC accepts now that the max sustained wind at landfall does not need to actually ever hit land- It can remain over the water as the center of the eye crosses land. Is that how they always interpreted the idea?

[HurricaneJosh]

I know I've asked you this before, but now that a few years have passed with presumably more data to pour over-- do you think Hugo actually produced category 4 sustained winds anywhere on the SC coast?

I have been looking at the surface observation data with the larger storms (hurricane-force winds, say, 70 miles or more from the center), and just find it hard to believe, even in our post-1990 examples, that any of these larger storms translate their reported max sustained winds at landfall down to 10-m height on the actual coastline. Jeanne and Wilma both come to mind as recent examples. Jeanne's TCR emphasizes that the 105 knot estimate is an "over-water" estimate, and then used the phrasing "It is possible that wind speeds of near 105 kt may have affected a small area of coast east of Sebastian." I again point to Powell's dissent in the upgrade of Andrew to a Cat 5 landfall--- with the key part of whether Andrew's Cat 5 winds were over water or made it onshore. It seems like the NHC accepts now that the max sustained wind at landfall does not need to actually ever hit land- It can remain over the water as the center of the eye crosses land. Is that how they always interpreted the idea?

Dude, welcome to my world!

When I'm focusing on historic cyclones, these are exactly the questions I'm asking myself all the time: How strong was it at landfall really, and how strong were the winds on the coast? Adam (am19psu) and I have been doing a sh*tload of research lately Re: the reanalysis methodology-- sort of testing it against a "control set" of 80 hurricane landfalls from the last 20 years. For each 'cane, there are 5 data points at landfall: central pressure, latitude, intensity trend, RMW, and speed of motion. Based on these data points, the reanalysis folks essentially try to "solve" for the max wind, and we're sort of testing how well that works against the contemporary sample set.

A couple of general obs:

* Over-water wind vs. over-land wind. My gut feeling-- based on reading reanalysis data/minutes and contemporary tropical cyclone reports (TCRs)-- is that the policy Re: this might be a bit fluid, still. For example, as you pointed out, the official Jeanne landfall intensity is 105 kt, even though the report seems to imply there's a very good chance no such winds actually hit the coast. On the other, I've seen cases in reanalysis where the strongest winds were offshore or affecting another territory (Canada or Mexico) at the time of landfall in the USA, and in those cases, the landfall intensity in the USA is the highest winds to affect the USA coast, not the system's highest estimated wind at that time. (I'm too lazy to find the examples right now.) Also, I've noticed recently in border-region landfalls that each territory that's affected gets assigned the category corresponding to the highest wind that's believed to have affected that territory. Examples: Rita 2005 at landfall at the TX/LA border is Cat 3 in LA and Cat 2 in TX; Hazel 1954 at landfall at the SC/NC border is Cat 4 in NC and Cat 3 in SC (based on a recent reanalysis paper that is not "approved" by the best-track committee but that reflects the most-recent methodology). So, bottom line is that I'm not really sure there's a clear policy yet Re: assigning landfall intensity-- whether it's the system's highest wind at the time of landfall or the highest wind affecting the coast. Either way, if you look at the HRD's surface-wind analyses, the highest marine-exposure isotach usually seems to make it to the open coast. It doesn't penetrate inland, but if it affects just the open coast, that's good enough in terms of saying it affected land.

* Official intensity vs. surface obs. I have no expectation that surface obs will perfectly substantiate best-track values. It's naive to expect that, give that the wind max-- even in a large hurricane-- is quite narrow and rarely passes over an official observing station. (You know this-- I'm just clarifying it for anyone else who's reading.) This having been said, with some hurricane landfalls, I find the disconnect between the official landfall intensity and the surface obs a a little disconcerting. With some, the difference is so great, you just start to wonder.

Re: the 'canes you mentioned:

* Wilma 2005 and Jeanne 2004 are similar in some ways-- both were 950 mb at landfall-- but there were differences. Wilma had some things working against it: the RMW was huge-- larger than Jeanne's-- and it was actually weakening a tad at landfall. On the other hand, it was moving fast, and that augments winds. Except for the large RMW, Jeanne was climatologically ordinary. And when I calculate based on these factors-- using what I understand to be the reanalysis methodology-- I get ~100 kt for each, which corresponds pretty well with the official estimate of 105 kt for each. Re: surface obs... Jeanne hit a populated coastline and was fairly well-sampled, and there are no official readings even approaching Cat 3, with the highest being 79 kt from the the Melbourne NWS-- so you have to wonder there. Wilma, on the other hand, was not well-sampled at landfall. It had a very asymmetric wind field, and the wind max passed over the sticks-- Cape Sable and the Everglades-- so it's hard to "penalize" it for not producing Cat-3 surface obs. The wind data and wind damage over all of Metro S FL, far from the landfall point, was impressive-- huge, solid swaths of Cat-1 and Cat-2 readings and damage, even well inland-- so that gives me confidence that the open Gulf coast down in the Everglades actually did have Cat-3 winds. Bottom line is that I have more confidence in Wilma than in Jeanne.

* Re: Hugo 1989... Adam and I took a close look at it and found two things: 1) While the system might have been large, the core wasn't-- the RMW was an average 20 nmi; 2) The system was actually steady state-- not strengthening-- at landfall. The fast motion at landfall-- over 20 kt-- helped it. Plugging in all these factors, as well as the pressure and latitude, I calculate winds of ~125 kt, which is even higher than best track. Re: surface obs... Charleston was in the left eyewall and there were several official readings of sustained Cat-1 winds in and near the city (76 kt downtown, 68 kt airport/WSO (several mi inland), 74 kt Folly Beach, 71 kt Mt. Pleasant). Since the cyclone was moving fast, it's assumed that the wind field was quite asymmetric, and winds in Charleston would be much, much lower than what happened well to the right of the center. The wind max passed over Bulls Bay and was not well-sampled-- however, the surge was a whopping 20 ft and the forest blowdown was massive. Bottom line is that when you model it mathematically, it comes out to a solid Cat 4, and given that no data contradict a Cat-4 verdict, I believe it was.

Wow-- sorry to be so long-winded! But you asked about something that I've been thinking a lot about these days. Comments are welcomed!

[MJW155]

Wilma, on the other hand, was not well-sampled at landfall. It had a very asymmetric wind field, and the wind max passed over the sticks-- Cape Sable and the Everglades-- so it's hard to "penalize" it for not producing Cat-3 surface obs. The wind data and wind damage over all of Metro S FL, far from the landfall point, was impressive-- huge, solid swaths of Cat-1 and Cat-2 readings and damage, even well inland-- so that gives me confidence that the open Gulf coast down in the Everglades actually did have Cat-3 winds. Bottom line is that I have more confidence in Wilma than in Jeanne.

Josh,

Not sure if you know this, but the dyke on the southern part of Lake Okeechobee was damaged during Wilma. As far as I know, they are still reparing the dyke. I went backpacking last winter (Feb of 2010) and I had to detour about 5-10 miles because the Florida Trail was closed. (the trail around the lake is the top of the dkye.)

Also, my sister lives in Coral Springs and she had a ton of damage. Every roof in the neighborhood was damaged. She told me there were areas of Broward County that lost power for 3 weeks. The eye passed over and she said everything went nuts during the backside of the storm. LOL.

[gymengineer]

Dude, welcome to my world!

When I'm focusing on historic cyclones, these are exactly the questions I'm asking myself all the time: How strong was it at landfall really, and how strong were the winds on the coast? Adam (am19psu) and I have been doing a sh*tload of research lately Re: the reanalysis methodology-- sort of testing it against a "control set" of 80 hurricane landfalls from the last 20 years. For each 'cane, there are 5 data points at landfall: central pressure, latitude, intensity trend, RMW, and speed of motion. Based on these data points, the reanalysis folks essentially try to "solve" for the max wind, and we're sort of testing how well that works against the contemporary sample set.

A couple of general obs:

* Over-water wind vs. over-land wind. My gut feeling-- based on reading reanalysis data/minutes and contemporary tropical cyclone reports (TCRs)-- is that the policy Re: this might be a bit fluid, still. For example, as you pointed out, the official Jeanne landfall intensity is 105 kt, even though the report seems to imply there's a very good chance no such winds actually hit the coast. On the other, I've seen cases in reanalysis where the strongest winds were offshore or affecting another territory (Canada or Mexico) at the time of landfall in the USA, and in those cases, the landfall intensity in the USA is the highest winds to affect the USA coast, not the system's highest estimated wind at that time. (I'm too lazy to find the examples right now.) Also, I've noticed recently in border-region landfalls that each territory that's affected gets assigned the category corresponding to the highest wind that's believed to have affected that territory. Examples: Rita 2005 at landfall at the TX/LA border is Cat 3 in LA and Cat 2 in TX; Hazel 1954 at landfall at the SC/NC border is Cat 4 in NC and Cat 3 in SC (based on a recent reanalysis paper that is not "approved" by the best-track committee but that reflects the most-recent methodology). So, bottom line is that I'm not really sure there's a clear policy yet Re: assigning landfall intensity-- whether it's the system's highest wind at the time of landfall or the highest wind affecting the coast. Either way, if you look at the HRD's surface-wind analyses, the highest marine-exposure isotach usually seems to make it to the open coast. It doesn't penetrate inland, but if it affects just the open coast, that's good enough in terms of saying it affected land.

* Official intensity vs. surface obs. I have no expectation that surface obs will perfectly substantiate best-track values. It's naive to expect that, give that the wind max-- even in a large hurricane-- is quite narrow and rarely passes over an official observing station. (You know this-- I'm just clarifying it for anyone else who's reading.) This having been said, with some hurricane landfalls, I find the disconnect between the official landfall intensity and the surface obs a a little disconcerting. With some, the difference is so great, you just start to wonder.

Well, I think more data than what we have now (airplane, radar, and ground observations) would help. Maybe more consistent post-storm damage surveys, like after tornadoes? Of course estimating winds from damage to vegetation and structures has its own huge set of issues and lack of reliability, and you would have to be in an area where the surge didn't wipe out the structures anyway. But if used as a corroboration, I think damage surveys would help with the confidence of the more intense landfalls. What I mean is if category 4 winds did actually hit land, there would have to be wind damage of at least a certain degree (to trees even, if there aren't many structures around). If anything, damage usually ends up more spectacular than winds would suggest. I know surveys were completed after Charley, but I would like to see them incorporated into the NHC's TCR.

ETA: And there is certainly precedence for the NHC to include damage survey results into the TCR's. Claudette is a key example-- the NHC needed to "throw out" 2 observations suggesting a Category 2 landfall. They didn't just rely on aircraft or radar data but instead felt the need to introduce damage survey results from NWS Corpus Christi, NWS Houston, and Haag Engineering to strengthen their argument.