Extreme, Cat.-4/-5 hurricanes in U.S.--trends

[Ground Scouring]

Yes, I know that winter is more significant at this time, but in light of debates re: the NAO / blocking and climatology...a case can be made that, in one area at least, extremes have actually decreased. I am not referring to the recent absence of major hurricanes in the United States since 2005...so I would like to keep that debate out of this thread, which is about something quite different. I am referring to the intensity of U.S. major hurricanes and the frequency of Category-4 and -5 landfalls in the United States. The original post on this subject can be located here, but as it then received no attention, I have since decided to make a separate topic.

From 1851-2011, I have looked at historical U.S. landfalls during Atlantic cycles, including inactive (-AMO) cycles. The following compares U.S. landfalls in the period 1995-2011, the latest +AMO phase, to those in historical periods, including both +AMO and -AMO cycles. While the average number of landfalling U.S. majors does not seem to have changed significantly, the intensity of U.S. major hurricanes, as well as the frequency of Category-4/5 hits, appears to have declined during 1995-2012.

While data on active and inactive Atlantic cycles is somewhat unreliable before 1926--and while more reliable intensity data on storms begins from 1900 onward--the numbers from 1851-1900 are still reasonably consistent with data from similar periods after 1900. Of course, some storms may have also been underestimated in intensity as most coastal areas were sparsely populated before 1900. Obviously, most major strikes, and most Category-4/5 hits, tended to come in concentrated spurts or clusters, so I included the median clusters or intervals in which the Cat.-4/5 hits tended to arrive.

1851-1900: 26 major strikes,* avg. intensity 108 kt, Cat. 4/5 once every 9.8 yrs. (median cluster w/in 5 yrs.)

1901-1950: 31 major strikes (19 during active cycle 1926-1950),* avg. intensity 111 kt (avg. intensity 113 kt during active cycle 1926-1950), Cat. 4/5 once every 4.4 yrs. (median cluster w/in 3 yrs.)

1951-1994: 24 major strikes (11 during inactive cycle 1969-1994),* avg. intensity 112 kt (avg. intensity 115 kt during inactive cycle 1969-1994), Cat. 4/5 once every 5.1 yrs. (median cluster w/in 3 yrs.)

1995-2012: 10 major strikes,* avg. intensity 106 kt, Cat. 4/5 once every 9 yrs. (median cluster w/in >9 yrs.)

*Includes major hurricanes that did not make landfall in the United States but produced major-hurricane conditions on a section of the U.S. coastline.

Note: After 1935--the latest year reanalyzed by the Best Track Committee--I usually used the official U.S. hurricane designations and took the lowest possible major (>100 kt) wind speed supportable if HURDAT listed a lower value than 100 kt. If a reanalyzed value was available--i.e., the 105 kt (Strahan et al., 2010) for the 1938 Long Island Express / Great New England hurricane--I used that value instead. After 1943, I utilized the intensities from the 1944-1953 reanalysis of Hagen et al. (2012). Obviously, such values are preliminary and my list will not be exact, but it should give a rough idea as to long-term trends.

Some noted highlights from the findings I made:

• The intensity of major hurricanes and the frequency Cat.-4/5 strikes were actually about the same as in previous periods during the last reliably documented inactive cycle (1969-1994), but the active cycle since 1995 has featured a decline in both the intensity of majors and the frequency of Cat.-4/5 hits relative to those which occurred in previously documented active cycles/periods (1926-1950, 1950-1968).
  
• Gulf Coast Cat.-4/5 hits were much more frequent historically: 5 in 1851-1900, 4 in 1901-1950 (at least 1 in 1926-1950), and at least 3 (perhaps higher if Celia 1970 and Frederic 1979 are reanalyzed as Cat. 4) in 1951-1994 (all in 1951-1969; this may be higher or stay the same if Betsy 1965 is reanalyzed as Cat. 4 in LA). None has occurred in 1995-2012, even though the average return rate in previous active cycles was once per 8.4 yrs. (with a median cluster w/in 6 yrs.).
  

As the statistics suggest we should have seen two or more Cat.-4/5 hits in the U.S. since 1995, I would very much appreciate some discussion as to why major landfalls seem to be getting less intense, and Cat.-4/5 hits less frequent, than the historical record would suggest. Pro-met responses would be especially appreciated re: possible factors that may be causing a less-than-favorable environment for such hits than in comparable periods in the past.

[andyhb]

Less sexycanes for Josh = the world will suffer.

BTW, is Charley not considered a Gulf Coast hit?

[Ground Scouring]

Less sexycanes for Josh = the world will suffer.

BTW, is Charley not considered a Gulf Coast hit?

In my post, Gulf Coast refers to Big Bend of FL WESTWARD, because the climatology for SW FL is different than for areas farther NW. So I did not include Charley in the Gulf-Coast total, for the upper-level pattern that caused Charley to recurve was different from that of most landfalls from the FL Panhandle to TX.

[andyhb]

In my post, Gulf Coast refers to Big Bend of FL WESTWARD, because the climatology for SW FL is different than for areas farther NW.

Ok.

[skierinvermont]

Seems to me that the sample sizes are too small to conclude much. A lot of short-term luck comes into play within the long-term factors.

AGW is supposed to slightly increase the frequency and intensity of the strongest storms, but obviously that has not been borne out yet.

When specifically referring to the Gulf Coast west of the bend in the panhandle of FL (which apparently is how you defined the Gulf Coast) it is possible AGW would decrease the frequency of major hurricane landfalls because of a predicted drying of the western and central U.S. which would lead to more dry air over the northern and western Gulf.

Any major approaching the northern/western Gulf this year would likely have had serious dry air entranement due to the low soil moisture and humidity over the U.S. The frequency of such heat and drought is forecasted to increase with AGW (and has already been observed to do so).

By the way, somebody should notify some folks in the climate forum of this thread. Many appear to be under the impression that AGW has caused a huge increase in major hurricanes.

[aslkahuna]

By the way, somebody should notify some folks in the climate forum of this thread. Many appear to be under the impression that AGW has caused a huge increase in major hurricanes.

Wouldn't do any good-their minds are already made up so why confuse them with facts?

Steve

[Nikolai]

Measurements before the 1960s were often extremely inaccurate... I think many of the 'major' landfalling canes prior to modern standards were likely not actually major hurricanes. One of the best examples is the 'New England Express' of 1938, which was probably the same strength as Sandy.

[Amped]

Measurements before the 1960s were often extremely inaccurate... I think many of the 'major' landfalling canes prior to modern standards were likely not actually major hurricanes. One of the best examples is the 'New England Express' of 1938, which was probably the same strength as Sandy.

Nice try but no.

[HurricaneJosh]

Measurements before the 1960s were often extremely inaccurate... I think many of the 'major' landfalling canes prior to modern standards were likely not actually major hurricanes. One of the best examples is the 'New England Express' of 1938, which was probably the same strength as Sandy.

I agree with your post except the last sentence, which is completely incorrect. The 1938 hurricane was recently reanalyzed by the top hurricane scientists, and they estimate it hit Long Island with winds of 105 kt and CT/RI with winds of 100 kt. The wind and storm surge damage in these areas was extreme and substantiates these estimates. Sandy is fresh in everyone's minds and so it feels so "big", but it does not hold a candle to 1938 by any yardstick, except that its gale radius was larger-- which makes perfect sense, given that Sandy was a nor'easter, not a tropical cyclone.

[HurricaneJosh]

I'm wondering if, despite reanalysis, the historic storms are still being overestimated.

Re: the reanalysis work... In the absence of reliable, representative wind data from the core of a landfalling hurricane (a very rare treat, even today), landfall intensities of historical storms are assigned using formulas that combine central pressure (observed or calculated), RMW, forward speed, latitude, and intensity trend. When applying these same methodologies to contemporary storms, I find that they often (but not always) yield wind speeds that are higher than what we're seeing lately via reliable dropsonde and SFMR data.

So, like I said, I'm just wondering if the historical storms are still being overestimated, despite reanalysis. It just seems like storms now-- for example, Isaac-- are held to a much more stringent verification standard for assigning max winds.

[Dark Energy]

I'm wondering if, despite reanalysis, the historic storms are still being overestimated.

Re: the reanalysis work... In the absence of reliable, representative wind data from the core of a landfalling hurricane (a very rare treat, even today), landfall intensities of historical storms are assigned using formulas that combine central pressure (observed or calculated), RMW, forward speed, latitude, and intensity trend. When applying these same methodologies to contemporary storms, I find that they often (but not always) yield wind speeds that are higher than what we're seeing lately via reliable dropsonde and SFMR data.

So, like I said, I'm just wondering if the historical storms are still being overestimated, despite reanalysis. It just seems like storms now-- for example, Isaac-- are held to a much more stringent verification standard for assigning max winds.

Are you leaning twds the point where the NHC was so anal retentive with the clasification of recent storms this year and going by the book? Is that kind of thinking going twds the reanalysis these days?

[HurricaneJosh]

Are you leaning twds the point where the NHC was so anal retentive with the clasification of recent storms this year and going by the book?

Yes, I'm alluding to that-- and I particularly noticed it with Isaac this year. I'm not going to go as far as to say I think they're being "anal"-- I think they're just getting more and more precise as the science improves, and that's a good thing.

However... It makes historical comparisons hard, as I think the reanalysis methodologies (that I describe above) yield values that are more generous than what you're going to get from verifiable SFMR/dropsonde data.

Is that kind of thinking going twds the reanalysis these days?

As I mentioned above, reanalysis is relaying on formulas using central pressure, RMW, forward speed, latitude, and intensity trend, whereas with contemporary storms like Isaac, they're relying on observed winds. That's going to skew things and make the older storms look stronger, I think.

[Dark Energy]

Do you think that because of the change in NHC directors that this may be the cause of change? Would Sheets have done the same thing?

[HurricaneJosh]

Do you think that because of the change in NHC directors that this may be the cause of change? Would Sheets have done the same thing?

I honestly don't think it's political. I think it's just a reflection of the improving science-- they have more tools now for measuring and verifying winds.

[Dark Energy]

I know it took a long time to upgrade Andrew to a cat 5.

[Dark Energy]

I agree with this 100%. You see the same thing with TOR verifications now then in the past. Bigger network, more technology, more reports.

[HurricaneJosh]

I know it took a long time to upgrade Andrew to a cat 5.

But that upgrade was also a product of improved science. They upgraded it because over that decade (1992-2002), they learned a lot more about how flight-level winds reduce to the surface.

Maybe it's because I'm so wonky about this stuff, I just don't factor in politics.

[Dark Energy]

And that's a good thing. You have to trust the forecasters there to give you a great report and know that it is top notch. I'm sure they use your data to help them in their studies.

[Dark Energy]

Btw, do they ever ref you in their reports??

[HurricaneJosh]

And that's a good thing. You have to trust the forecasters there to give you a great report and know that it is top notch. I'm sure they use your data to help them in their studies.

Btw, do they ever ref you in their reports??

Yes, they do credit me when they use my data-- for example, in the reports for Karl 2010, Jova 2011, and Rina 2011. I expect they'll reference my data in the Ernesto report, since I got a reliable, controlled pressure reading within a couple of miles of the cyclone's center in a jungly, remote part of the Yucatan with no observing stations. Given this, I think my pressure reading is a very important clue Re: how strong it was at landfall.

Here's my barogram. It was a very small cyclone with an extremely steep pressure gradient near the center, and damaging wind confined very close to the center:

[wxmx]

I'm wondering if, despite reanalysis, the historic storms are still being overestimated.

Re: the reanalysis work... In the absence of reliable, representative wind data from the core of a landfalling hurricane (a very rare treat, even today), landfall intensities of historical storms are assigned using formulas that combine central pressure (observed or calculated), RMW, forward speed, latitude, and intensity trend. When applying these same methodologies to contemporary storms, I find that they often (but not always) yield wind speeds that are higher than what we're seeing lately via reliable dropsonde and SFMR data.

So, like I said, I'm just wondering if the historical storms are still being overestimated, despite reanalysis. It just seems like storms now-- for example, Isaac-- are held to a much more stringent verification standard for assigning max winds.

We've had several loosely wound systems with relatively low ambient pressures these past few years. It goes the other way too, high ambient pressures, relatively high central pressures, but stronger than expected surface winds. Science will eventually find out, but I still have confidence in the work of NOAA's scientists, and I think their estimates have no to little bias toward stronger landfalling hurricanes.

[Nikolai]

I agree with your post except the last sentence, which is completely incorrect. The 1938 hurricane was recently reanalyzed by the top hurricane scientists, and they estimate it hit Long Island with winds of 105 kt and CT/RI with winds of 100 kt. The wind and storm surge damage in these areas was extreme ans substantiates these estimates. Sandy is fresh in everyone's minds and so it feels so "big", but it does not hold a candle to 1938 by any yardstick, except that its gale radius was larger-- which makes perfect sense, given that Sandy was a nor'easter, not a tropical cyclone.

But they were the same strength. They just had different wind speeds. Sandy and 1938 both made landfall as 940-945mb storms. I think the major difference was the forward speed in 1938, which (obviously) led to the 100kt winds on the eastern side of the storm. Sandy was moving significantly slower--if it had been roaring along at 60MPH, winds on the eastern flank would have approached or exceeded 100 knots as well.

I think the main difference between Sandy and 1938 comes down to the forward speed of each storm, with the faster speed of '38 the only reason winds so strong were recorded. I think even the Cat 3 estimates are dubious, but the forward speed of '38 does support gusts to 110MPH+ for sure.

[HurricaneJosh]

But they were the same strength. They just had different wind speeds. Sandy and 1938 both made landfall as 940-945mb storms. I think the major difference was the forward speed in 1938, which (obviously) led to the 100kt winds on the eastern side of the storm. Sandy was moving significantly slower--if it had been roaring along at 60MPH, winds on the eastern flank would have approached or exceeded 100 knots as well.

I think the main difference between Sandy and 1938 comes down to the forward speed of each storm, with the faster speed of '38 the only reason winds so strong were recorded. I think even the Cat 3 estimates are dubious, but the forward speed of '38 does support gusts to 110MPH+ for sure.

No, they weren't the same strength.

The NHC stopped using central pressure to rate intensity twenty years ago-- so the fact that they had similar pressures doesn't mean anything. (And the fact that you're using central pressure to compare intensities of two tropical-origin systems tells me you're not up on the latest science Re: this topic.) They were structurally completely different systems: the 1938 cyclone was still a (mostly) topical cyclone with a core of very high winds; Sandy was a nor'easter. A tropical cyclone can generate the high winds that a nor'easter can't because the structure is entirely different.

Also, you are incorrect that 1938's high winds were entirely the result of the forward speed. No. If that were the case and it were as simple as adding the forward speed to winds on the right and subtracting them from winds on the left-- and if the 1938 storm really was the same strength as Sandy-- then NJ, NYC, and W Long Island would have had winds of about 10 kt in the 1938 storm.

Silly comparison.

[HurricaneJosh]

We've had several loosely wound systems with relatively low ambient pressures these past few years. It goes the other way too, high ambient pressures, relatively high central pressures, but stronger than expected surface winds. Science will eventually find out, but I still have confidence in the work of NOAA's scientists, and I think their estimates have no to little bias toward stronger landfalling hurricanes.

As I've said several times above, I have faith in them, too. But that doesn't mean that the science won't continue to evolve. For example, even after the initial round of reanalysis was completed, they went through and lowered the intensities of several storms by 5 kt-- for example, Indianola 1886, Galveston 1900, Tampa Bay 1921, etc.

[Ground Scouring]

As I've said several times above, I have faith in them, too. But that doesn't mean that the science won't continue to evolve. For example, even after the initial round of reanalysis was completed, they went through and lowered the intensities of several storms by 5 kt-- for example, Indianola 1886, Galveston 1900, Tampa Bay 1921, etc.

But at least the first two, and most other cyclones originally assessed as being Cat. 4, retained their Cat.-4 status while only some storms (like the 1909 and 1915 LA hurricanes) were downgraded from Cat. 4 to Cat. 3 at landfall. Also, RMW size, movement, observed surges, etc. are not the only factors that go into reanalysis--usually, in most pre-1920s and particularly pre-1900 U.S. landfalls, data near the center were usually very scarce, and the nearest wind or pressure observations were often 40 n mi, or much farther, from the eye. So in many cases the size of the storm was calculated by using observations of pressures from the nearest land or ship stations. The pressure measurement at the estimated outer closed isobar (OCI)--the outermost edge of the circulation--was used to calculate the likely location of and the pressure at the center of the eye. Such data, which are usually a bit more reliable than RMW data for very old storms, can be located in the official list of reanalyzed U.S. hurricanes.

Note that OCI data before 1898 are lacking due to scarce observations; the RMW sizes for pre-1898 cases were usually based upon an observation or two of wind pattern or storm surge at a point closest to the eye. The data do show that many Cat.-4 storms from 1898-1935 had average or slightly larger-than-average OCIs, but more data nearer the eye became available with which to estimate RMWs better during this time, which counteracts the likelihood that the storms would have been overestimated by more than 5 kt. I am assuming here that a climatological mean can be deduced from this period, so similar patterns may have occurred before 1898 as well. So I would tend to agree with Jorge that the reanalyzed data, while not perfect, are generally accurate for older storms.

With that said, are you willing to comment on the suggestion that we may be seeing weaker landfalls / less frequent Cat.-4/5 hits in the U.S.?

[HurricaneJosh]

But at least the first two, and most other cyclones originally assessed as being Cat. 4, retained their Cat.-4 status while only some storms (like the 1909 and 1915 LA hurricanes) were downgraded from Cat. 4 to Cat. 3 at landfall. Also, RMW size, movement, observed surges, etc. are not the only factors that go into reanalysis--usually, in most pre-1920s and particularly pre-1900 U.S. landfalls, data near the center were usually very scarce, and the nearest wind or pressure observations were often 40 n mi, or much farther, from the eye. So in many cases the size of the storm was calculated by using observations of pressures from the nearest land or ship stations. The pressure measurement at the estimated outer closed isobar (OCI)--the outermost edge of the circulation--was used to calculate the likely location of and the pressure at the center of the eye.

Yes. As I wrote above, they use central pressure, either observed or calculated.

Such data, which are usually a bit more reliable than RMW data for very old storms, can be located in the official list of reanalyzed U.S. hurricanes.

snowflake, my friend-- knowing what a big reanalysis nerd I am, did you think I wasn't aware of this chart? C'mon.

Note that OCI data before 1898 are lacking due to scarce observations; the RMW sizes for pre-1898 cases were usually based upon an observation or two of wind pattern or storm surge at a point closest to the eye. The data do show that many Cat.-4 storms from 1898-1935 had average or slightly larger-than-average OCIs, but more data nearer the eye became available with which to estimate RMWs better during this time, which counteracts the likelihood that the storms would have been overestimated by more than 5 kt. I am assuming here that a climatological mean can be deduced from this period, so similar patterns may have occurred before 1898 as well. So I would tend to agree with Jorge that the reanalyzed data, while not perfect, are generally accurate for older storms.

Remember, I am the biggest proponent of the reanalyzed data on this board. I'm constantly citing those data-- so obviously I respect the work that's been done. But as I said above, the science is evolving, and we've already seen some of the reanalyzed storms receive additional adjustments. The reanalysis team will be the first ones to tell you this is a fluid, living data set.

The 1947 Fort Lauderdale 'cane is a good example a storm that I think might get downgraded as the science evolves. Why? Because, from what we've seen in the last twenty years, I'm not sure I believe that a large-ish, slow-moving, weakening storm had winds of 115 kt with a pressure in the low 940s at that latitude. I'm not saying it's impossible-- I just have my doubts. Let's see what the best-track committee says.

With that said, are you willing to comment on the suggestion that we may be seeing weaker landfalls / less frequent Cat.-4/5 hits in the U.S.?

I think it has more to do with the different standards of rating older versus modern storms (see above), as opposed to a grand climatological shift.

[weatherwiz]

Measurements before the 1960s were often extremely inaccurate... I think many of the 'major' landfalling canes prior to modern standards were likely not actually major hurricanes. One of the best examples is the 'New England Express' of 1938, which was probably the same strength as Sandy.

Completely not true at all.

[CT Rain]

But they were the same strength. They just had different wind speeds. Sandy and 1938 both made landfall as 940-945mb storms. I think the major difference was the forward speed in 1938, which (obviously) led to the 100kt winds on the eastern side of the storm. Sandy was moving significantly slower--if it had been roaring along at 60MPH, winds on the eastern flank would have approached or exceeded 100 knots as well.

I think the main difference between Sandy and 1938 comes down to the forward speed of each storm, with the faster speed of '38 the only reason winds so strong were recorded. I think even the Cat 3 estimates are dubious, but the forward speed of '38 does support gusts to 110MPH+ for sure.

Yeah you're totally wrong here dude. Josh is right.

1938 flattened entire sections forests with widespread structural damage from the wind. The damage was consistent with low end cat 3 damage in southern CT, Long Island, and S RI. It's pretty well documented by photographs and accounts. Not really much of a mystery here.

While horrific, Sandy's surge was pretty much what you'd expect to see given a storm arriving at high tide moving at that direction. New York City and the western part of Long Island Sound is particularly vulnerable to storm surge given the geography of the coast. If anything we've been lucky in the past. I really wouldn't say Sandy was anywhere near a worst case scenario to be honest. The impact of the 1938 storm on the region would be hard to fathom IMO.

[weatherwiz]

Yeah you're totally wrong here dude. Josh is right.

1938 flattened entire sections forests with widespread structural damage from the wind. The damage was consistent with low end cat 3 damage in southern CT, Long Island, and S RI. It's pretty well documented by photographs and accounts. Not really much of a mystery here.

While horrific, Sandy's surge was pretty much what you'd expect to see given a storm arriving at high tide moving at that direction. New York City and the western part of Long Island Sound is particularly vulnerable to storm surge given the geography of the coast. If anything we've been lucky in the past. I really wouldn't say Sandy was anywhere near a worst case scenario to be honest. The impact of the 1938 storm on the region would be hard to fathom IMO.

Completely agree. The thought of something similar to 1938 occurring again is completely frightening. We'd be looking at millions without power for several weeks and maybe even longer. It's a highly scary thought and I don't even know how you'd explain that to people.

[andyhb]

Completely agree. The thought of something similar to 1938 occurring again is completely frightening. We'd be looking at millions without power for several weeks and maybe even longer. It's a highly scary thought and I don't even know how you'd explain that to people.

Or worse, a storm of the intensity of the Long Island Express combined with the size of Sandy and/or the tidal element.