With the North Atlantic hurricane season still waiting for a named storm, and the rest of the globe cyclonically challenged during the past several months, it is a good time to catch up on the research end of things. In terms of papers during the past 6-8 months, the amount of tropical cyclone and climate literature has largely dropped off compared to the heyday after Katrina. Regardless, the following is an “omnibus” style blog posting and can serve as the obligatory 2009 Atlantic Hurricanes posting.
First, the forecasts for the upcoming Atlantic hurricane season are decidedly below-normal or below median or less-active depending upon the metric or audience. The UK Met Office, which recently had its climate research budget slashed 25%, has determined the most likely amount of storms is 6, very low compared to what has been seen since the active period in the Atlantic began in 1995. The most likely Accumulated Cyclone Energy (ACE) tally is predicted to be 60, which is in the basement compared to recent seasons.
Here is a helpful comparison of previous seasons of ACE that I plotted up from the HURDAT best-track dataset maintained by the friendly folks at the hurricane center and NOAA.
Figure: The ACE from 1935-2008 for the North Atlantic with the Met Office prediction as the squat, little red bar.
NOAA and Bill Gray’s CSU forecasting outfit are in close agreement:
NOAA: 9-14 named storms (including Baby Whirls), 4-7 hurricanes (1-3 major) and ACE of about 65-125. The ranges are rather large with NOAA forecasts, and the previous 5-6 years’ forecasts have an uncanny resemblance to each other; they are essentially the same forecast year after year.
Gray and Dr. Phil Klotzbach (CSU, June 2 update): 11 named storms, 5 hurricanes (2-major), and an ACE of 85.
Here is a flavoring of the latter’s reasoning:
We expect current neutral ENSO conditions to persist or perhaps transition to weak El Niño conditions by the most active portion of this year’s hurricane season (August-October). If El Niño conditions develop, it would tend to increase the levels of vertical wind shear and decrease the levels of Atlantic hurricane activity. Another reason for our forecast reduction is due to the persistence of anomalously cool sea surface temperatures in the tropical Atlantic. Cooler waters are associated with dynamic and thermodynamic factors that are less conducive for an active Atlantic hurricane season. Another factor in our forecast reduction is the stronger-than-normal Azores High during April-May. Stronger high pressure typically results in stronger trade winds that are commonly associated with less active hurricane seasons.
Secondly, Global tropical cyclone activity (or hurricane, typhoon, cyclone) has continued to plummet to levels now approaching 50-year lows. There is no mystery why this is occurring: it is entirely natural and associated strongly with the fluctuations of large-scale climate including the well-known El Nino – Southern Oscillation. The extended and powerful La Nina period beginning in 2007 precipitated the downturn in global TC activity as measured by ACE. One ENSO index called the MEI highlights the recent strong La Nina and the last few months tick up to more neutral/warm conditions.
Now, when you take a gander at the global ACE using a 24-month running sum, it is rather apparent that there is a strong relationship. Why do I use a 24-month running sum? The time scale of El Nino events is typically 2-7 years, and more importantly, the Southern Hemisphere tropical season peaks in boreal winter. Thus, if you use calendar year January-December metrics, you necessarily chop off or alias the seasonal activity into unrealistic combinations. Thus, the 24-month sums are smoother but no data is removed, it is raw. Additional details are at my FSU website and a recent GRL article I published in March 2009: Northern Hemisphere Tropical Cyclone Activity.
The current global TC ACE number valid for July 1, 2009 summing the previous 24-months is 1083, and the Northern Hemisphere number is 802. For the calendar year 2009, Northern Hemisphere (NH) ACE is at 33 which is at about 50% of the previous 30-year average. This is not unusual as some Western Pacific typhoon and Eastern Pacific hurricane seasons don’t ramp up until mid- to late-July. However, this is little reason to believe above average NH for 2009 is in the cards.
A discerning eye will see that there is no trend in the Northern Hemisphere ACE during the past 50-years even considering that there is a void of Eastern Pacific and Northern Indian Ocean data prior to about 1970. The global data prior to 1970 also suffers from poor intensity estimates in the Southern Hemisphere when/where satellite observations were very limited. Interestingly, one satellite image was from 40-years ago from the Apollo moon-landing (the Lost Tapes have been found!).
Thirdly, the Copenhagen Synthesis Report received laudatory coverage in the media.
The world faces a growing risk of “abrupt and irreversible climatic shifts” as fallout from global warming hits faster than expected, according to research by international scientists released Thursday.
Global surface and ocean temperatures, sea levels, extreme climate events, and the retreat of Arctic sea ice have all significantly picked up more pace than experts predicted only a couple of years ago.
If you sift through the report, you find Figure 6 on page 12 referencing a talk given by Greg Holland on Atlantic hurricane activity.
Figure 6: (A) The numbers of North Atlantic tropical cyclones for each maximum wind speed shown on the horizontal axis. The most intense (Category 5) tropical cyclones have maximum wind speeds of 70 m/s or greater. (B) The proportional increase by cyclone (hurricane) category (1 – least intense; 5 – most intense) arising from increases in maximum wind speeds of 1, 3 and 5 m/s. Note the disproportionately large increase in the most intense tropical cyclones with modest increases in maximum wind speed, compared to the increase in less intense cyclones (23).
After a little digging, reference number 23 points to the talk Dr. Holland gave in Copenhagen back in March, right before Roger Pielke. Here is the extended PDF document.
Figure 1 of the extended abstract shows the maximum intensity distribution of Atlantic tropical cyclones for the period of 1945-2007. He states that theory suggests a potential 1-5 m/s increase in hurricane intensity for a concomitant increase of 1 degree C in ocean temperatures. Quoting: “The impact of a uniform 1, 3 and 5 ms-1 increase in these intensities is shown in Fig. 1b. Quite modest changes of < 20% are indicated for category 1-3 and even category 4 hurricanes. These changes are small compared to interannual variability and would be difficult to detect. But the most intense, category 5 hurricanes could experience an amplified response, doubling in number for a general 5 m/s increase."
Where does Figure 1b come from? It is simply moving the PDF to the right, apparently the tail is a function of the mean. Anyone have an issue with doing that? Just warming up. Figure 2 shows observational proof of this previous hypothesis:
Whoa! Category 5’s have increased by 300-400%! Holland goes on:
Although this example has only considered Atlantic hurricanes, we suggest that the overall result is robust and applicable more widely to weather extremes. Major difficulties arise from the relative rarity and patchy nature of weather extremes and from the inability of current climate models to simulate them.
This seemingly is the converse to the Baby Whirl detection issue brought up here at Climate Audit. The discovery of Category 5 hurricanes in the past is likely a function of our detection capabilities which have markedly improved during the past several decades. Thus, as with Baby Whirls, plotting up Category 5’s as detection has improved apparently will give you hockey stick looking graphs.
Holland points out that changes in the tracks of hurricanes may lead to more Category 5’s, which he alleges are climate signals. Unfortunately, there is no breakdown of natural variability versus climate change / global warming presented. Instead, we are urged to consult the “purely statistical reasoning” in Figure 1, which “implies that other physical factors are contributing to a skewing of the distribution”.
I do not buy the assumption that the PDF of maximum hurricane intensity will simply shift to the right, and I do not find the analysis a “tentative confirmation of (Holland’s) hypothesis.” Furthermore, this analysis is not yet in the peer-reviewed literature.
Lastly, as entertainment value, the Weather Channel on June 30th is going to focus on the “skill” of “hot models” (not Maxim but HWRF) with regards to hurricane forecasting. One such HWRF forecast run of invest 93L in the Caribbean was really out to lunch. Here is the landfall location and an animation to the model run. To be fair, HWRF is “experimental” but it is astounding that it can be so wrong. A trained forecaster would immediately discount this run.
Note: This is NOT a current forecast. The circulation never developed into a tropical storm.
Horribly blown forecast animation for June 27, 2009 00Z forecast.
Couldn’t think of a better selling point for Congressman Alan Grayson (D) and his proposed $50 million U Central Florida / Obama Hurricane Center in Orlando FL, which was added as a Christmas tree ornament in the climate change bill passed by the House of Reps late Friday.
Congressman Grayson said, “With this one move, Central Florida will become a world leader in 21st century meteorology.”
Climate Audit 
238 Comments
I have a suggestion. What if one surveyed all storms that have made landfall in Texas and Florida etc though the years. Is it not reasonable to suggest that such storms would be normally be accounted for? Is it reasonable to suggest that the proportion of storms that make landfalls are constant? If so this would make a better estimate of increase/decrease in storm activity. Of course, one must look at the wind force exclusively and not at the damage.
Mmmm, been there, done that, Knut. The counter argument is that hurricanes only landfall in certain conditions. See Benders excellent analysis of my data here.
.
There’s a Google list of CA pages on landfalling hurricanes here.
.
[Note to self … always search site:climateaudit.org first … lots of info. Why reinvent?]
w.
Re: Willis Eschenbach (#2),
Great note to self and others! Also, excellent presentation: The Thermostat Hypothesis
Re: John Baltutis (#3),
Second to that. Steve, how about inviting Willis to do a guest post along these lines?
TIA & Cheers — Pete Tillman
“My objections to the global warming propaganda are not so much over the technical facts, about which I do not know much, but it’s rather against the way those people behave and the kind of intolerance to criticism that a lot of them have.” — Freeman Dyson, http://www.e360.yale.edu/content/feature.msp?id=2151 (highly recommended)
Having set expectations low (and yet right on the 3-9 average that you’d expect 85% of the time) they can only be pleasantly surprised if it goes over.
By complete coincidence, the Sun continues its completely unpredicted funk and there’s still nothing but solar “Tiny Tims” to relive the monotony.
Ryan, an outstanding piece of scientific compilation. It took me a long time to figure out what you meant, that he was shifting the distribution. He moved the curve untouched to the right, without distortion.
.
The main problem I have with this is his curious idea that this increase could possibly be “uniform”. He says:
When someone says “uni-form” I think “unlikely-form”. Nature doesn’t do uniform. As the most obvious example, in the whole record, there’s not one hurricane with measured wind speed of 85 mps (170 knots). Not one.
.
It appears that here on this non-model planet, with all its natural constraints, it’s never blown a sustained 85 mps. Not once in 63 years of records. There is a natural limit to the strength that hurricanes can attain. This limit is partly due to the fact that the energy in the wind goes up roughly by the cube of the wind speed. At those speeds, we’re talking huge amounts of energy.
.
So when you say “add 5 to the right tip of the distribution”, you are describing a distribution that does not exist in recorded history. But why stop at +5 mps? What about +10? What if tomorrow hurricanes started blowing 90? Run the numbers on that on, it’s huge, add it to Fig. 1(b) up above the +5 line. To go from 80 mps to 90 mps requires 40% more energy, an immense increase over the biggest recorded storm. A bigger hurricane than in recorded history, and the odds on it just went up. Be scared …
.
You are correct, Ryan, good catch. Fig. 1(b) merely reflects exactly what we would expect from that kind of shift. The line for +10 will fit in above it, same shape. Means nothing about the odds of hurricanes blowing 90 mps. That’s simply what happens when you slide an entire distribution to the right. It’s what you get when you assume the impossible. Bad statistician, no cookies.
.
w.
Regarding Figure 2, whose caption says
I call bullshit. The Cat5 data has not been smoothed in the slightest. I just analyzed the data in Fig. 1(a). It shows an average of 1.7 Cat5 storms per year. Note the steps in the Cat5 data. Each step is one Cat5 hurricane. I just plotted it out, the years and the units fit exactly, viz:
Year, Cat5 storms
1932, 1
1933, 2
1934, 2
1935, 1
1936, 2
1937, 2
1938, 1
1939, 1
1940, 1
1941, 0
1942, 0
1943, 0
1944, 0
1945, 1
1946, 1
1947, 1
1948, 2
1949, 3
1950, 2
1951, 2
1952, 2
1953, 2
1954, 1
1955, 1
1956, 1
1957, 1
1958, 1
1959, 2
1960, 2
1961, 2
1962, 2
1963, 1
1964, 0
1965, 0
1966, 0
1967, 1
1968, 1
1969, 1
1970, 1
1971, 1
1972, 0
1973, 0
1974, 0
1975, 0
1976, 0
1977, 1
1978, 2
1979, 2
1980, 2
1981, 2
1982, 1
1983, 0
1984, 0
1985, 0
1986, 1
1987, 2
1988, 2
1989, 2
1990, 3
1991, 2
1992, 1
1993, 1
1994, 1
1995, 0
1996, 2
1997, 4
1998, 4
1999, 4
2000, 4
2001, 3
2002, 2
2003, 6
2004, 6
2005, 8
.
Every year fell exactly on the graphed line, there were no fractions of years. I call bullshit on the claim that it is smoothed. As to whether there were 8 Cat5 storms in 2005, I leave to Ryan.
.
Next, since this is hot-off-the-press stuff, why stop in 2005? There’s been three full Atlantic hurricane seasons since then, why stop there? The paper is dated 2009.
.
Finally (and obviously) if it were actually smoothed with a 5-year filter, the variations wouldn’t be nearly as large … and the graph wouldn’t be as graphic.
.
Since that is unsmoothed data, I doubt it very much. I don’t care what the numbers say. Nature doesn’t do that. Somewhere there’s an error in the count. I can’t say where, but we are not looking at natural observations. He claims that for 67 years there were never more than three Cat5 hurricanes, and then in the following nine years it got up to six and eight, and only fell below four twice?
.
I say bad numbers somewhere to that.
.
w.
Here are two charts I created applicable to this discussion.
US Major Hurricane Index. Only Saffir Scale 3-5 hurricanes are graphed, because no satellites before 1977 or so, and the smaller ones were likely missed. In general there have been less hurricanes the last few decades compared to the 1950’s.
Same thing with Atlantic Hurricane Index. Saffir Scale 3-5 only to take out the little ones. Recent hurricane levels are very similar to the 1950’s. Take out very active seasons of 2005/2006 and we are below average.
Raw data found here.
US Strong/Violent Tornado Index F3-F5.
http://www.nofreewind.blogspot.com
I love this blog. Every day, after WUWT, this is the next one I check. I love it.
OT: I’m a wordsmith, and I’d like to point out the increasing misuse of a common word.
In fact, ‘amount’ refers to any measure but number. The phrase should read: “the most likely number of storms is 6.”
One could refer to “the amount of sand” but “the number of grains of sand”; the “amount of water” or the “number of drops of water”; the “amount of earth” or the “number of mountains”; the “amount of ice” or the “number of ice cubes”. You get the idea.
Using the incorrect term often introduces ambiguity. Quite unscientific.
I hope you can see my gentle smile as I write this. I hope to enhance and clarify the content and process of writing, not obscure or belabour it.
Cheers,
Richard Treadgold,
Editor,
Wordshine.
Ryan, I am sure that you personally know Chris Landsea, (great name for Hurricane scientist). Here is his resignation from the IPCC.
This should be required reading for anyone trying to get to the heart of the “crisis”.
On a slightly related topic, are still having the 2009 hurricane contest? I thought I read earlier that we were going to start in July this year rather than June, thus I had a little more time. If so, my entry for ACE is below normal. I can’t remember what all the categories were. Does anyone have the data? Where is David Smith when you need him.
Steve, Is it normal in the climate field to release studies before they have been peer reviewed?
Re: AndrewPrice (#11), Steve can attest to the use of non-published or peer-reviewed literature making its way into the IPCC or other governmental reports. There has also be a spate of “to-be submitted” papers in the scientific discourse.
Of course, these days you can write up any paper you want, get it published, and then go to town with the press release. That seems to be the place where the societal/public impact is made.
Hehe…Hot models…Ryan is a funny guy!
Nice post. Will take some digesting…
What I found for Cat45 hurricane counts and reported previously here at CA, using the IBTracs data series, was that the trends for the annual occurrence of Cat45 hurricanes globally had a statistically significant trend from 1981-2007, but when coming forward with the starting date towards the present time that the statistical significance was lost even though the trend remained positive.
I also looked at Cat5 hurricanes alone and found a significant positive trend from 1981-2007 that was entirely due to increases in Cat5 hurricanes in the NATL basin. With the NATL counts removed there was no longer a trend. I think these types of sensitivity tests are needed and particularly so when they get left out of some of these studies that are making bold claims. I will be lazy and bypass Image Shack and put a table directly into my post. The first column is year, the second is global Cat5 counts without those from the NATL and the third is the global Cat5 counts that includes those from the NATL.
Year TotCat5W/ONATL Total Cat5
1981 1 1
1982 2 2
1983 3 3
1984 1 1
1985 2 2
1986 2 2
1987 2 2
1988 0 1
1989 1 2
1990 1 1
1991 1 1
1992 0 1
1993 0 0
1994 4 4
1995 0 0
1996 0 0
1997 3 3
1998 3 4
1999 2 2
2000 0 0
2001 0 0
2002 4 4
2003 3 4
2004 3 4
2005 4 8
2006 2 2
2007 2 4
Re: Kenneth Fritsch (#14), thank you so much for your numbers, they have been immensely helpful. I have finally figured out what Holland has done. I reverse engineered his “smoothed” series to calculate the data he used.
First, the numbers. I have multiplied his smoothed series by five to make the difference clearer. They are the sum of the 5 years centered on the year of interest.
Year, Holland Smoothed, Fritsch, Holland data
Y1983, 0, 0, 0
Y1984, 0, 0, 0
Y1985, 0, 0, 0
Y1986, 1, 0, 0
Y1987, 2, 0, 0
Y1988, 2, 1, 1
Y1989, 2, 1, 1
Y1990, 3, 0, 0
Y1991, 2, 0, 0
Y1992, 1, 1, 1
Y1993, 1, 0, 0
Y1994, 1, 0, 0
Y1995, 0, 0, 0
Y1996, 2, 0, 0
Y1997, 4, 0, 0
Y1998, 4, 1, 2
Y1999, 4, 0, 2
Y2000, 4, 0, 0
Y2001, 3, 0, 0
Y2002, 2, 0, 0
Y2003, 6, 1, 1
Y2004, 6, 1, 1
Y2005, 8, 4, 4
Y2006, N, 0, 0
Y2007, N, 2, 2
With the help of your data, I finally unravelled his puzzle. Unwinding his five-year smoothed data revealed what he had done.
But it also revealed that he is using different data for 1998 and 1999. He says two Cat5 storms in each year. Not a large change, but it wipes out four years of no Cat5 storms very nicely. Might be an interpretation error, or transcription, or a different dataset … but it’s odd that all the rest are identical.
He also perseveres with using uniform measurements. He’s measuring the difference in percent change … but you can’t do that. Whether you smooth it or not.
Truly, I take my hat off, we’re in the presence of a master. The concept of using “percent change above/below the mean” on a poisson distribution leaves me in awe.
Look, it’s a poisson distribution, and it probably has a fat tail. Nature likes those. So, four in a year? Not that unlikely. But the whole drama revolves around that one data point. Other than that, it’s a string of zeros, ones, and twos.
I continue to be amazed at what gets passed off as science.
w.
Thanks Ryan. That’s what I was afraid of. I was hoping it was more like medicine where they usually try to get peer review before releasing their findings, but I was deluded.
Re: AndrewPrice (#15), Indeed the press release is not peer-reviewed. It allows the author to frame their work in terms of why it is important, new, or necessitates further inquiry. In doing so, it also allows rife speculation, extrapolation, and plain disinformation to creep in. Since few if any journalists have the patience nor scientific training to even digest the abstract of many journal articles, they are almost completely reliant upon the press releases to generate their news copy.
When the media is compliant and complicit or sympathetic to the authors’ point of view, a positive feedback occurs. Thus, journals as well as authors see how the publicity game works, and eagerly compound the above problems by gearing their research and analysis towards positive result studies. Medicine is well aware of this issue, but climate science is not quite there yet.
Ryan
The linked to article has been refuted as a hoax (Apollo tapes found). There is some good Apollo era imagery of the polar regions but it has nothing to do with lost tapes. The SIM bay and Hasselblad camera film has been in safe storage at JSC for the past 40 years and is now being digitized by Arizona State University.
Ryan, I wanted to make a couple of comments on your thread introduction:
I liked your ENSO difference graph placed on top of the global and NH ACE time series.
Also I appreciate your giving the UK Meteo and Klotzback’s TC forecasts for the NATL. Is not Klotzbach now the lead person on the forecasts with Gray and did he recently obtain a PhD? My first inclination for David Smith’s forecast contest is to use a Meteo/Klotzbach average.
It would appear that we need to get Judith Curry back here to scold you on not moving onto Cat45 hurricane counts and away from ACE indexes. TC counts and ACE are passé.
Re: Kenneth Fritsch (#18),
Tsk, Tsk Kenneth. It’s %Cat 4/5 that has the big trend, i.e. the one that matters.
Re: Kenneth Fritsch (#18), well as ACE goes down the tubes, to the lowest levels in recorded TC best-track history, the number of hurricane days also goes down. That is the denominator of the ratio of Category 4 / Category 1 days. The relevant image:
Notice how the number of hurricane days has completely collapsed, while the number of intense or major days has drifted upwards? Well that is not a global warming signal there at all, but something pretty fundamental. Kim, Webster, and Curry (2009) Science (Thursday) touch on the periphery of the bigger picture in the Atlantic, but go into the weeds a little bit with some very speculative extrapolation of their analysis.
Re: ryanm (#30),
Perhaps improved measurement capability?
Mark
Re: Mark T (#32), ENSO. The tropical Pacific is such a large-scale climate modifier that it is absolutely outrageous that recent TC/climate papers try to “smooth” it away by using 1-3-4-3-1 filters, etc. Or explicitly state that they don’t care.
Re: ryanm (#30),
Ryan, I find you have posted another informative graph in Post #30 above. Perhaps I would offend a statistician but I look at the time series for the intense hurricanes in 3 parts:
The first is 1981-1988 were the trend is flat; the second is an upward trend from 1989-1993: then finally the third, and longest time period, were the trend is flat (with intervening hills and valleys) again from 1994-2008. In effect, I see two platueas with the first occuring being lower than the last and transition period in between with an upward trend. I would think that an attempt would be made to explain the shape of the time series and not simply wave it away with a comment about the noise involved.
The total hurricane days appear to have declined since approximately 1993 to present while the same metric for the Cat345 hurricanes appears to have declined but at a slower rate than the total. The Cat45 hurricane days appear to have a trend that is flat to perhaps slightly declining in the 1993-present time period. My point here is that without further and more detailed explanation what is to say that the relative changes are simply happenstance.
The models that I see for the influence of SST on TC intensity could expalin some small increase in maximum wind speed that would push a TC into a hurricane Cat1 category and Cat2 to a Cat3 and so forth. I do not see Holland’s pdf shifting at 5 m/sec following from that theory.
To be fair, Elsner (2008) found some changes in maximum windspeeds in a quantile analysis of TCs in the global TC basins with larger increases in the upper quantiles of TCs compared to the lower quantiles. As I recall for the period studied (1981-2006) and from reanalysis data, the significant changes in quantile levels were confined to the NATL.
If the trend in counts of Cat45 hurricane days can remain flat to perhaps a slight decline, while overall hurricane days decrease at a faster rate, what value, for the purposes of safety and economic losses, would a %Cat45 metric have?
Re: Kenneth Fritsch (#40), here’s a great way to spend an afternoon: Summit on Hurricanes and Climate powerpoints and podcasts from Corfu Greece from this June.
To answer your last question, you have to include more than just frequency of Category 4/5 or the ratio of 4+5/1. There is considerable year-to-year variability in the tracks of intense tropical cyclones, so you have to include the duration component as well. A normalized ACE per storm is the way to go, and a lot falls out from that metric in terms of understanding the role of hurricanes in climate. This way, you don’t throw out 95% of the hurricane season.
Re: ryanm (#41),
Ryan, your point is well taken. I would like to play with the data used to construct the duration/hurricane category graph that you presented in Post #30. Do you have a link to the data used? Also I want to better understand the data behind the graph. Is the duration the time that a given level hurricane spent at that duration time period at the level given or is the duration for the time taken for a given level of hurricane to build up and die out?
mike B:
I assume by your comment you kenneth (#18) you believe that the trend in %cat 4/5 is real, meaningful and tied to Co2?
If this is the case can you make a decent statistical guess as to when we should expect to get a new high in this percentage given that we know that Co2 is going up continuously? If you had some probabilities associated with the guess that would be great like .20 1 year, .35 2 years …. .85 10 years etc.
thanks
david
Re: david_a (#20), It’s called sarcasm but let’s cut you a break since you don’t know the history here. 🙂
Basically, some researchers (including occasional CA commenter Judith Curry from Georgia Tech) argue that the trend effected by AGW is the ratio of category 4’s to category 5’s…to my knowledge this isn’t something Mike B has advanced so he won’t be able to answer your questions. Th Hypothesis’s proponents could perhaps answer that.
Re: david_a (#20),
Actually, david_a, I was scolded by Judith Curry for using the Cat45 counts and not the percentage that Cat45 represents of the total hurricanes. I found that the percentage basis trend holds up longer than the Cat45 counts but it too peters out to not a significant trend around 1991 – as I recall. Ryan M was scolded for using ACE as a metric and not %Cat45.
I also found that the Cat45 counts from 1983-2007 fit well to a Poisson distribution for all the TC basins of the globe. This would indicate, in my mind, that Cat45 hurricane production are subject primarily to a random gathering of the “right” conditions.
Hurricane data from 2 reanalysis data series showed no trend of Cat45 counts globally over the period 1981-2007. I also showed that a regression of Cat45 counts from IBTracs versus SST showed no statistically significant trend over that period.
Judith Curry pointed to my 1981-2007 positive trend for Cat45 hurricane counts and said in effect thank you very much that proves my point (Curry’s point that is) and did not reply to my other evidence. When I commented that she was acting like an advererial lawyer and only dealing the evidence that helped her case a poster here indicated that we should realize that Judith Curry would have already considered all those counter points and discounted them.
So next time Judith is in town you can ask her your questions. I would add one query and that would be to present a model for hurricane counts and hurricane intensities that would allow a different trend result for using straight Cat45 counts and % Cat45 counts to total hurricanes, i.e. how do you increase Cat45 counts without increasing total hurricane counts.
Re: david_a (#20),
You assume wrong. I was being facetious.
%cat 4/5 is a silly metric.
Kenneth, yes the number of tropical cyclones and ACE are passe. the relevant things to look at are number of hurricanes (gets rid of tiny tim and related issues), % cat45, and integrated kinetic energy (rather than ACE). integrated kinetic energy is basically the radial integral of ACE over the horizontal size of the hurricane (we are just about to submit a paper on this). I know, i always talk about papers in progress. However, we do have a Science paper on hurricanes that will hit the press this Thursday, stay tuned.
Re: Judith Curry (#23), “the radial integral of ACE over the horizontal size of the hurricane,” which assumes a constant vertical depth, of course.
Re: Judith Curry (#23), I wouldn’t be so quick to throw ACE or Kerry’s Power Dissipation metric overboard quite yet. Don’t you love looking at that downward trend into the abyss 🙂 As he states in his 2005 Nature paper, since there is little correlation between size and maximum intensity, that component of the power calculation would not contribute meaningfully to seasonal statistics. However, on a storm by storm basis, Ron McTaggert-Cowan and Lance Bosart in a 2007 paper on the lifecycle of Hurricane Katrina show very nicely how the power dissipation changes on a 6-hourly basis.
If we are ditching ACE as so passe, then can we also ditch the Category 4/5 ratio metric into the same bin as Holland’s category 5 explanation and Holland and Webster 2007 Atlantic frequency paper? If you tread through my little tropical webpage, you can figure out what the problem is with that ratio metric by cross-referencing a few figures. It is all there — and I show that in an upcoming paper. I am still amazed that after all of the correlation games with SST and reanalysis datasets that the most fundamental mechanisms relating tropical cyclones to climate are largely ignored.
To understand hurricanes in climate, we have to look at all of them and not only the most intense 5% or the tails of the distribution. You throw the baby out with the bath water. All that information in the ACE metric has to be used correctly, and not bits and pieces.
It is rather ironic that your upcoming Science article (Thursday) makes my case for the Northern Hemisphere and Globe as a whole as ACE/Power reaches 50-year lows, and further buttresses Emanuel (2005) as the paper still standing after 4-yrs.
Kerry Emanuel (Dean et al. 2009) also has a nice GRL paper in press that uses Kossin’s size database along with the Extended best-track dataset to show that the distribution of hurricane sizes in the North Atlantic is log-normal. I really enjoyed reading this paper, and it is definitely a must read.
I am quite interested in your upcoming paper on Integrated Kinetic Energy and your definition of the horizontal scale of a storm (Dean et al. 2009 have a very novel size model). In my PhD dissertation, I devote some page space to a global examination of TC Power Dissipation as explained in my AMS-New Orleans 2008 Poster Presentation. This largely came out of the comment process Bob Hart and I went through with Sriver and Huber at Purdue in the reanalysis calculations of Power Dissipation. But I am so ready to graduate…
Don’t know much about extreme climate events, but there is a graph show twice in your intro with number on the Y-axis and maximum wind speed on the X. There is a downturn between wind speeds of 25 and 20 m/sec. Simply curious, would have expected the trend to continue up left until off chart. Is there a problem with measurement or counting or definitions?
Hmmm…
So 2005 brought us a large number of NA hurricanes, and 2006 brought us a lot of scientific papers linking number of named storms to AGW. Then 2007 brought us strong Hurricanes Dean and Felix, and the 2008 theme for climate publications? Why of course, cats 4 and 5 are the thing.
Then 2008 brought us Hurricane Ike. Well…got to link size of hurricanes to AGW!
Faddish? Why would I think that?
hi judith,
based upon your understanding of the statistics could you give me your best estimation of the probability distribution that I asked of mike b?
thanks
david
There are many interesting points in this thread. Rather than get into those, though, I’d rather journey down a side alley and offer a few thoughts on the issue of comparing the modern cat 4 and 5 storm record with the historical (pre-modern satellite) record. These are simply offered for consideration.
The first is intended as an example. It is the windspeed history of two Cape Verde storms which had similar tracks. These are the 1947 hurricane #4 (which thumped southeast Florida) and 2004’s Hurricane Frances (which also thumped southeast Florida).
Frances’ record shows intensity ups and downs as it traversed the Atlantic. That reflects our modern ability to detect short-term intensity variation.
On the other hand the 1947 storm record shows an unlikely steady intensification for 13 days. Storms rarely behave that way. I suspect that the meteorologists of the 1940s made educated interpolations and extrapolations from widely-spaced and often indirect measurements. I doubt that they intended to convey that their best track record captured all of the intensity wiggles of the storm.
The next two plots are a bit odd and may take a moment to come into focus. They reflect the changes in windspeed of cat 4 and 5 storms from the time they first become a hurricane until the time they reach their peak windspeed. (This choice tends to capture the open-ocean part of storm existence and avoids the decay phase, which is a phase with particular data problems.)
The first covers the cat 4 and 5 storms from 1995-2008, a period which is blessed with good intensity detection –
It is a plausible-looking distribution, with a portion of the distribution showing decreases, as would be expected, and a rather smooth tailing on the positive side. The positive side shows a respectable presence of rapid intensification (say +15 knots over 6 hrs) which, too, is expected.
Here’s the same type of plot except that it covers 1945-1958, an era of very limited means of measurement –
There are very few values showing wind reduction and there are few showing rapid intensification. The windspeed changes are clustered around +5 knots in 6 hours, which is a value I’d probably use a lot if I was doing a lot of interpolating.
To my eyes the modern distribution behaves differently from the historical distribution. Something appears to have changed over the years. I think the change is, to a large and perhaps total extent, due to improvements in measurement capabilities. As a result I would be hesitant to offer a plot which mixes early intensity data with modern data, as Holland offers, unless I also offered some explanation for the distribution shift.
Re: David Smith (#29), hey David, welcome to Hurricane watch 2009! Wonderful comment — this is a great idea to flesh out in more detail. The easiest explanation that I can think of for Great Hurricane of 47 is that the track points were interpolated between Africa and first recon or when it reached the Islands. A speed was calculated for translation and voila!
Amazing that the things that don’t show a beneficial (to the hypothesis) trend are “passe” and those that do are legitimate. We might as well just start drawing lines on a page telling everyone that’s what’s happening. Oh, wait, that’s what they’re doing with the temperatures, right? Sigh…
Mark
Hi Ken
I’ve asked these kinds of questions relating to probability distributions of climate outcomes of every person I can who has a view as to global warming and its likely course and effects. I am particularly interested in people who have opinions that differ with mine as it might be possible to learn something. To date I have never been given a response by anyone who is a proponent of catastrophic CO2 induced global warming in which the timescale of the prediction is such that I might be around to observe it.
This is a curious result but perfectly understandable if you believe as I do that the vast majority of the climate scientists who are strong AGW proponents do not actually believe in what they have signed up for but instead are along for the financial and political power which accrues to those in the forefront of a popular political idea and perhaps more importantly is taken from those that are opposed.
I can make large well regulated bets on the outcome of next years corn crop, the price of short term credit, the price of Co2 vouchers, whether or not California will default on its bond obligations, or the average price of feeder cattle auctions in the US midwest. Many markets exist where one can make extremely arcane bets related to outcomes which have virtually zero effect on the general population, yet here is an arena where the political determination will have an extraordinary effect on people for decades to come and we have no financial marketplace where the idea can be priced.
You would think that there would be one person out there who thinks that the IPCC mean model forecast has enough of a chance of actually occurring that given a little discount would be happy to increase their net worth comfortable with the knowledge that the science is settled and that there is near unanimity in opinion by scientists as recently relayed to me by my senator.
For those of you who are are interested you can see breakdown of the US hurricane landfall data set at:
http://www.climatedata.info/Impacts/impacts.html
and click on tropical cyclone.
You can also download a digested version of the data set at:
http://www.climatedata.info/Downloads/downloads.html
and click in Impacts.
The great thing with this data set is that it goes back to 1850. Two of its three metrics (pressure and sea surge) have been measured accurately for that all time (wind speed is a bit more problematical). Interestingly it suggests that there has been an increase in hurricane activity since the mid 70s – the backward limit of satellite based data.
Flashback to June 1, 2006: Longer Hurricane Season?
While that advice to Jeb Bush was horrible, ranking up there in the Pantheon of great seasonal forecasts like Accuweather’s “barrel of a gun”, it did bring up an interesting question about the length of seasons supposedly increasing, especially in the wake of the never-ending 2005 season.
James Kossin actually did the scientific analysis in a recent GRL paper. Is the North Atlantic Hurricane season getting longer? PDF from Kossin’s Publication Page. Figure 2 is relevant for detection issues of tropical storms.
RyanM,
Is there a plot that shows ACE and/or hurricane counts against ENSO/AMO/PDO, at least one that plots against best track/detection data?
So far as I’m aware, AMO has a positive and negative phase, and PDO has a positive and negative phase. It would be interesting to see a four quadrant graph (+AMO/+PDO, +AMO/-PDO, -AMO/+PDO -AMO/-PDO) with the ACE of storms in each quadrant.
I think it would also be beneficial to see each of the ACE counts in the same graph or separate with El Nino, La Nina, or ENSO neutral conditions. Maybe dividing each quadrant into El Nino in the upper half and La Nina in the bottom half, and then ACE from lowest to highest within in each sub-quadrant.
Do such graphs exist and if so, can you point me to them. Maybe they would tell me something, maybe they wouldn’t, but I’ve just never seen them.
the “passe” variables are ones that are lacking in either observational or physical justification.
Climateaudit has criticized spurious counting of tiny tims, stating that they inflate recent TC counts. this is probably true, but tiny tims are almost exclusively tropical storms (not hurricanes), and there are other dubious issues as well in terms of counting tropical storms, including classification or not as subtropical, not to mentioned potential missed storms in early part of the record. So # of hurricanes is much more robust observationally than # of TCs
With regards to ACE and PDI, their reflection of TC energetics rests on the assumption that the integrated energy is controlled only by the wind speed, with variations in size being random and uncorrelated with wind speed and so do not impact the seasonally integrated energetics. Our recent work (as yet unpublished M.S. thesis shows that this is incorrect).
with regards to % vs N cat45, the argument we (and everyone else) has been making ties global warming and the increase in tropical SST to intensity (NOT number). So it makes no physical sense to tie an intensity metric to the actual number of tropical cyclones; rather the intensity variable should be independent of the total number of TCs (which are remaining constant or even decreasing slightly globally, although they seem to be increasing in the NATL).
So the best variables to use are determined by the limitations of the observations and the physics, not necessarily the “simplest” or “usual” variables.
Re: Judith Curry (#45),
Fair enough. But you’ve been around here enough to know the next question. Will all the data and code be posted? Actually I’m guessing it will, since TCs are (or at least have been) much more objective than say, adjustments to temperature readings or Ice cores.
Re: Judith Curry (#46),
Bad in the sense of many intense ones, I assume, rather than few weak ones? I.e. bad for humans rather than storms?
Re: Dave Dardinger (#47), As a member of the human race, you can’t fault her for being partisan 😉 Clearly she means lots of intense storms.
Re: Judith Curry (#45),
It makes no sense to disassociate the occurrences of hurricanes at the less intense levels from that for the more intense levels. How does a model for hurricane production at the various category levels account for fewer hurricanes at the lower intensities and overall and more at the higher intensities? Using a percentage metric without looking at the counts for all hurricanes at all levels seems to be ignoring some additional data that would be required to validate any hurricane model.
A starting point in my mind for the effect of SST increasing all hurricane level counts and at the same time increasing Cat45 hurricanes would simply involve TCs with marginally lower maximum wind speed than a Cat1 being bumped up to Cat1 and a marginally lower Cat1 to a Cat2 and so on up to a Cat4 to Cat5.
Not looking and explaining beyond Cat45 percentages being related to intensity changes would appear to some as finding a metric that shows an increase over time that could be simply a matter of happenstance and particularly seem so without providing a model for how this can occur given all of the hurricane level counts.
Re: Kenneth Fritsch (#59),
I don’t see a particular problem here. At least for the NA the overall prevailing wind will be fairly steady, and if we use Morris Code with – = Large storm and . as smaller ones and look at a particular length (using a proportional font)
–…–…
..—..—
we’ll see that larger storms will take up more time / resources leaving less for the smaller storms. Converting that to an algorithm or equation for modelling purposes is another story.
Re: Dave Dardinger (#60),
And just when I thought I had cornered the market on simple mindedness.
jonathan, for a quadrant analysis of AMO/PDO, see my ppt presentation at
Click to access ins_tampa_09.pdf
i have been pushing PDO, but most of the other hurricane climate dynamicists use SST and say there is no extra predictive capability by adding PDO (and many think AMO doesn’t show up in the tropical atlantic). personally i think the combo of warm AMO, cool PDO is VERY BAD NEWS for atlantic hurricanes, and things could get worse than we’ve seen in the past decade (even independent of AGW).
Re: Judith Curry (#46),
Dr. Curry, thanks for the link. Those are pretty much the graphs I was looking for. You state that warm AMO and cool PDO means likely > TC’s. But with an El Nino coming on, do you feel like this year will be less likely to see your concerns come to fruition, more likely, or no change.
Thanks
Re: Jonathan Schafer (#52),
I look forward to reading the paper. Thanks again for the information.
Re: Judith Curry (#46),
Judith, you noted. “…and many think AMO doesn’t show up in the tropical atlantic).”
That’s curious if we use the NOAA ESRL method of calculating the AMO, and that is simply to detrend North Atlantic SST anomalies.
Regardless, in agreement with the recent dip in the AMO, tropical North Atlantic SST anomalies also dropped significantly over the past year.
http://bobtisdale.blogspot.com/2009/07/hurricane-breeding-grounds-sst.html
Regards
We will definitely post all of our data. The hurricane community has been very open about data, mostly since the datasets are publicly available
For the “hurricane community”
Scaled to 10 servings
10 oz vodka
2.5 oz grenadine syrup
10 oz gin
10 oz light rum
5 oz Bacardi® 151 rum
10 oz amaretto almond liqueur
10 oz triple sec
grapefruit juice
pineapple juice
Pour all but the juices, in order listed, into a hurricane glass three-quarters filled with ice. Fill with equal parts of grapefruit and pineapple juice, and serve.
Ryanm (and any other cyclone afficianados) I threw a few typhoon (from the Chinese Tai-feng: supreme wind) parties in Taiwan when a super-typhoon blew through, and my wife grew up on Alabat Island, which seems to be a typhoon magnet… but is there a down-side to fewer cyclonic events?
Re: rephelan (#51),
In some places, agriculture is heavily dependent of cyclonic precipitation, so, yes.
Jonathan, our Science paper (Kim, Webster, Curry) will be released today, it is on the topic of el nino. This is an unusual year, the annual AMO index is slightly negative and ENSO is positive (anomalous for the regime we are currently in). but the punchline of our paper is that there are different types of El Nino, and we this year we may be headed for the anomalous type that doesn’t suppress atlantic TCs.
Re: Judith Curry (#53),
What trend have you determined for global cloud volumes, increasing versus decreasing, and their effect upon the frequencies and intensities of the storms?
Re: Judith Curry (#53), when the paper embargo is lifted, if you are willing and have the time, please add a quick summary and inclusion of the Press Release to a comment thread — and I will move it unadulterated to a posting of its own. I do have some material ready to go, but I think it is more appropriate to include it as a comment to a thread rather than “frame” the debate from the outset.
jonathan, one other thing, the amo/pdo indices in the charts are annual values (not smoothed)
BTW, I see wordpress converted my individual dashes to n-dashes and m-dashes. There should be 2 each in the top row and 3 in the bottom row.
Re: Dave Dardinger (#61), The terms are em dash and en dash…
http://en.wikipedia.org/wiki/Dash
Lot’s of interesting papers are in press at GRL at the moment. I have one I want to ask Ryan about-are you familiar with this?
Yu, J.-Y., C. Chou, and P.-G. Chiu (2009), A revised accumulated cyclone energy index,
Geophys. Res. Lett., doi:10.1029/2009GL039254, in press.
http://www.agu.org/journals/gl/papersinpress.shtml
And might it have any impact on your ACE related work?
(Again, lots there which may be of more general interest to CA readers…)
Re: Andrew (#64), yes I am familiar since I presented the very similar work at AMS in 2008 in New Orleans. The PDF is here, as was up in Comment #28.
The idea of a revised ACE or PDI metric is not new, and the authors likely missed some of the literature. From their conclusion:
I agree with their conclusions. There is considerable information in the ACE metric that can be used in addition to size parameters or their Revised-ACE (which is really just the Power Dissipation), on a storm-by-storm or comparison basis. However, with limited size information of quality decreasing back in time, and the perhaps questioned validity of a modified-Rankine vortex in all instances, on a year-to-year basis, ACE or R-ACE or PDI or PD are very highly correlated due to the dominance of the duration parameter in the accumulated statistic. Thus, we must also examine the correlations between size, duration, intensity, and frequency to determine the actual variance, something that the current datasets barely afford the opportunity to do.
kenneth, number and intensity are two different and mostly independent metrics of hurricane activity. The issue is a shift in the intensity distribution. During the previous active period (say in the 1950s), there was a bump in the pdf of intensity around cat 3 (cat 2 if you do the landsea intensity adjustment). the bump in the last decade has been at cat 4. so the shape of the intensity distribution is not constant. the most frustrating thing is that we haven’t sorted out how to treat intensity (esp for major hurricanes) prior to 1970. The hurdat data almost certainly overestimates the intensity (landsea and emanuel both agree on this), but i don’t think the “fix” that they propose is really correct.
Re: Judith Curry (#66),
It is frustrating for me that I have not had or been offered a reasonable explanation for total hurricane counts remaining constant or declining and %Cat45 hurricanes increasing. No metric is reasonable in my mind without some substantial basis for it.
Given the additional evidence of the good Poisson fits for the Cat45 counts over the three most recent decades and the lack of a statistically significant trend for Cat45 counts over the last three decades or Cat45 % in the two most recent decades and combining it with a lack of a good explanation for the changing distribution of hurricane counts at all levels and further that the hurricane models do not predict a % Cat45 change of this magnitude, pardon me for being skeptical of most of the increase in Cat45% being more than happenstance and/or a matter of changing detection capabilities.
Re: Kenneth Fritsch (#68),
The silence is telling. When the metric produces a pattern that fits the preconceived paradigm you don’t question it. That’s exactly how confirmation bias works. And when the pattern finally falls apart, you pick a new index if you want to salvage the hypothesis. A process that never ends, a hypothesis that always evades refutation. Like a zombie that can never be killed.
Re: Kenneth Fritsch (#68), Re: bender (#69),
Regarding the ratio of Cat 45 / Cat 1+ ratio changing over time due to low frequency SST changes aka global warming:
Disregard 1970-1978 due to data issues, especially in the Southern Hemisphere.
The numerator has been flat and the denominator also flat during the past 30-years, but the ratio has been trending upward. This is a global metric — so you have to do some sensitivity tests to determine which basins or combinations of basins show this upward trend and if it is indeed significant. If you assume the assignment of maximum wind speed of each storm has been perfect during the past 30-years, then the only task left is to determine why certain years see marginal hurricanes (64-113 kts) more readily develop into Category 4+ (114 kts) storms. Then, examine why certain storms fail to reach Category 4, but indeed only reach Category 3. Or Category 2, etc.
As you go through a thorough examination of the different Saffir-Simpson categories during the past 30-years in each basin, you see that the most critical and by-far important factor of why a storm reaches Category 4 intensity is the track of the storm, mainly its genesis and the atmospheric steering currents. A minimum SST is of course required. Thus, one must balance the thermodynamic environment associated with the oceanic conditions, which are dumb luck — the storm cannot turn around or take equatorward turns (this is part of the basic idea behind a SHIPS model) — and the atmospheric conditions.
So far, this isn’t new knowledge. Pretty basic.
It is no secret that the recent 30-years have been associated with the following large-scale climatic characteristics: a persistently positive PDO, more El Nino’s than La Nina’s, average of 0.4C of tropical SST warming, and the change in 1995 in the Atlantic/East Pacific from inactive/active to active/inactive hurricane seasons — inversely so.
Thus, to test the hypothesis that the 0.4C of tropical SST increases is the primary cause of the increase in the ratio of Cat 45/ Cat 1+, you must also understand the impacts, constructive and destructive, of the natural variability which cannot be simply removed/detrended or dismissed away using simple linear, or other low-frequency filters. Furthermore, you must understand why nature has deigned the cyclones to track where they did during the past 30-years. It is unlikely that a given configuration of the large-scale climate affects each basin the same. Thus, it is probably inappropriate to examine this Cat 45/Cat 1+ ratio globally without taking into account the individual contributions of each basin normalized to the prevailing climate conditions and how they affect that basin.
So, are the cyclones themselves tracking over the warmer water? That answer is indeed yes. If you take the recent storm tracks (2004-2008 SSTs) and run them over previous waters (1979-1983 SSTs), as an example, you get an average trend/difference throughout the 30-years of about 0.4C. But are the cyclones preferentially tracking over warmer water — in other words, are the tracks being reconfigured or somehow are the cyclones adapting?
Re: Ryan Maue (#70),
Ryan, thanks for taking the time for providing some insights into analyzing the Cat45% trends and providing an alternative explanation for the occurrence patterns we see for hurricanes for all categories. I would like to digest what you have put forth better and see whether it unhangs my hang-up over increasing or flat Cat 45 % trend with a decreasing trend of annual counts of hurricanes of all categories.
I totally agree (not that a layperson’s view carrys much weight) that each global TC basin should be handled separately. When I looked at the good fits of annual Cat45 counts to Poisson distributions I did the counts by basin.
The global Cat45 counts/percentages using straight counts and the hurricane days (from the data that you provided) shows a global decline in total hurricanes from appropximately 1990 to 2008 while the Cat345 occurrences show a lesser decline and the Cat45 occurrences are nearly flat. I have a breakdown by TC basin (which you provided) that I need to look at again. I know that there are significant differences basin to basin.
In line with RyanM’s post above indicating that analysis of hurricane categories should start with results from the individual TC basins of the globe, I am posting the annual hurricane days for Cat12345, Cat345 and Cat45 hurricanes by TC basin and globally with data provided by RyanM.
There are differences between basins in the relative trends in hurricane counts as classified for these graphs, but I do see a general decreasing trend from approximately 1990 for all hurricane days (IO with relatively low counts excepted) and a lesser decreasing to flat trend for the Cat345 and Cat45 categories.
Re: Kenneth Fritsch (#72), Kenneth, Can you point me to the source of your Data? I am fed up listening to poorly researched presentations from “Carbon Accountants” who point to insurance industry evidence and conclude that there has been a huge increase in storm activity.
Ed
Re: EddieO (#91),
EddieO, RyanM posted the global results at the post linked below. For some reason the graph no longer shows there. I received the basin data from RyanM on request. I would feel better if you emailed RyanM for the data directly. I believe that the data were retrieved originally from a TC series (IBTracs?) that lists the maximum sustained winds over 6 hour periods of time.
Re: ryanm (#30),
EddieO, you might want to follow the current discussion going on here between RyanM and Angela Fritz. The newer metric may be size matters and/or how to determine/estimate size.
Kenneth, no one understands why the total number of global tropical cyclones remains nearly constant. and why is it 90 (rather than 9 or 900). This is a big question that has no simple answer. In my opinion (i have published nothing on this), tropical cyclones form in the summer hemisphere when the pole to equator temperature gradient is weak, as the major form of heat transfer from the tropics to polar regions (second law of thermodynamics in action). In the winter hemisphere, the larger pole to equator temperature circulation drives a larger mean circulation that accomplishes much of the heat transport. It is my guess that if we were looking at integrated kinetic energy (integrated over the size of the tropical cyclone) rather than ACE, we would be able to start to make sense of the interactions between tropical cyclone energetics and planetary scale energetics. I have some ideas on how to approach this from a diagnostic/observational perspective, but the data sets are not yet up to the task (working on it). The only way to test hypotheses about the global tropical cyclone frequency using models is to run a series of simulations using coupled atmosphere ocean models at high resolution (say 20 km) that can resolve to some extend the individual tropical cyclones. This is not currently feasible; the Japanese Earth Simulator has only managed to run a few atmospheric only simulations at this resolution (unless there is something recent i don’t know about).
We DO have a theory for tropical cyclone intensity, albeit a crude one, that includes a relationship with tropical SST. it doesn’t really say anything about the distribution of TC intensities (this is back to the energetics issue raised above). The empirical observation of Webster et al. (2005) of the %NCAT45 increasing is just that, an empirical observation, that can be related to the increase in SST through theories for tropical cyclone intensity.
So in the absence of a satisfactory explanation for all this, should we pretend that the increase in %NCAT45 doesn’t exist? that it isn’t related to an increase in SST? that it isn’t likely to be associated with greenhouse warming through the increase in SST? No. The observation is there, a physical link is hypothesized, and a possible continued increase in %NCAT45 with future increases in SST is a risk that needs to be considered
Re: Judith Curry (#73),
With the total number of TC remaining relatively constant, if the number of CAT45 increases, then the lower strength will decrease. But expect the CAT 45 worldwide to drop when the SST decreases if that is a cause. Is the change in SST caused by greenhouse warming? If the Ocean temps are not increasing, then expect the SST to not increase worldwide. A somewhat related event is that the total number of thunderstorms is also relatively constant. There is a certain amount of energy that these systems generate and release. Bottom line – a lot more study is needed.
Re: Judith Curry (#73), Re: David Smith (#75),
Speaking from a totally ill-informed point of view: Has anyone ever done a statistical analysis on the distribution of start dates (say) of tropical cyclones, over the course of the season? If all dates were equally likely (a dubious hypothesis), they should end up looking like a Poisson distribution; measuring the deviation from Poisson might yield some useful quantitative information on the extent to which the formation of one tropical cyclone inhibits the formation of others (albeit, without giving any information on the mechanism). Such an analysis could shed at least a little light on “why the number of global tropical cyclones is 90, rather than 9 or 900”.
Re: jc-at-play (#77),
You might wish to take a look at comment 271 on an earlier CA thread. There are several graphs showing the distribution of statrting dates for tropical storms in the Atlantic basin. The dates are definitely not uniform in their distribution during the hurricane season.
Re: RomanM (#81),
Quite right; I should have searched CA before submitting my comment. The question I was trying to ask is more in line with Paul Linsay’s analyses in that thread and elsewhere — especially in http://www.climateaudit.org/?p=1068, which shows a suppression of hurricane formation for intervals of less than three days. By itself, that’s doesn’t seem to be enough of an effect to limit the number of tropical storms very much.
Re: Judith Curry (#73),
Thanks, Judith, for an insight into your method of reasoning on this issue of %Cat45. You never appear to include the points that I have attempted to bring to the discussion and from your comment above I guess it does not matter.
I’ll continue to look at more of the detailed data that go into a %Cat45 calculation, as for example the hurricane counts/trends for all categories, and the distribution and the trends over time of these hurricanes.
The ratio is not consistent with MPI theory. It can’t be, and no one has established that it has.
It’s an intriguing question. My favorite idle conjecture is that the answer will be found in the tropical upper troposphere, with tropical cyclones modifying the upper atmosphere in ways that hinder the formation of other tropical cyclones. Basically, a negative feedback mechanism will be found which limits the number of cyclones that form. A crude analogy is that of a tree seedling which grows and chokes off the sunlight and nutrients available to other seedlings, thus limiting the number of trees in the forest.
Certain upper-tropospheric conditions (low wind shear, regions of divergence/diffluence, etc) or combinations of conditions favor the transition of seedlings into cyclones. The aerial extent and “strength” of these upper-troposphere conditions tend to increase over time if no cyclones are present. But, when a cyclone forms, that cyclone modifies the upper troposphere via its exhaust. This modification may occur locally or far downstream. The impact of this modification is that other seedlings are less-likely to transition into cyclones, thus reducing the chances that other seedlings become cyclones and thereby limiting the global rate of cyclone formation.
Of course this is all simply blogosphere conjecture and may be true or wildly wrong.
To determine the average relationship between kinetic energy (KE) and intensity, a best-fit applied to aircraft recon data from 1995-2005 (MACLAY et al. 2008, MWR). 124 storms yielded a total of 1244 flight-level wind field analyses. The best-fit is KE = 3×10^13 (Vmax)^1.872 and variance explained is 82% or R=0.91. Thus, KE increases nearly with the square of the maximum winds (ACE).
“It should be noted that the mean KE-intensity relationship does not describe the evolution of KE for individual storms and there is considerable variability for a given intensity.”
From MACLAY et al. (2008), Figure 2
Re: Ryan Maue (#78), Hi Ryan – I think that you are using Maclay and DeMaria (2008) out of context here. They are calculating ICKE (inner-core kinetic energy) to determine the lifecycle of storms, NOT to determine the energetics of a TC season. The critical radius used in their study is RMW, which, I think we can both agree, has little to do with the total IKE of a storm. The radius that should be used for an energetics calculation is R34. I have done this, and I can show that IKE and ACE are uncorrelated on all time scales. These results will be submitted by the end of the summer, at which point I can go into more detail.
Re: Angela Fritz (#88), thanks for chiming in. No one disputes that the real-2 Dimensional wind field is much better than just maximum sustained wind for calculating integrated metrics such as Power Dissipation, IKE, etc, especially size and maximum intensity are not well-correlated.
See the comment Bob Hart and I published a couple years back regarding Sriver and Huber (2006) usage of reanalysis datasets to calculate the PD. The duration component absolutely dominates the calculation, and that is what ACE shows, nothing more, nothing less. It is elegant in its simplicity just like Emanuel’s PDI.
ACE and IKE are not apples to apples, but near the inner-core, as you suggest, they converge to unity. Thus, you are prescribing the result beforehand by choosing an arbitrary R34 threshold for the size of the storm. While that may be what is reported in the extended best-track, do you really trust that data? I don’t. You can use modified Rankine vortex with varying drag coefficients to calculate the IKE or PDI, and on a storm by storm basin, you will find considerable differences dependent upon the “size” of the storm. Thus, to be able to satisfy a hypothesis that ACE and IKE are uncorrelated on all time scales, you have to show that the individual components of the calculation are independent and robust. Deconvolving the size component of IKE is straightforward, I think.
I would be happy to discuss offline (fire off an email) my limited efforts concerning the calculation of IKE/PDI/ACE using a variety of data sources, some of which are brand new for size estimates. I have no stake in the battle consuming the surge community with regards to IKE vs. other metrics, and have no plans to publish anything related to the IKE/ACE stuff due to uncertainty in the datasets.
Here’s a question with regard to the “trend” in %Cat4+5: Why does all the increase happen during the part of the satellite period before Dvorak IR observations, and they afterward level off?
http://www.heartland.org/bin/media/newyork09/PowerPoint/Stan_Goldenberg.ppt
How can such a trend be attributed to AGW in any case? That does really sound like the same pattern seen in SST!
Here’s a Powerpoint with a nice slide eight, which shows seasonality by basin
Link
Here’s a time series related to global tropical cyclone count. There is a prize of 200 quatloos for anyone who can accurately guess what this shows:
I’ll post the answer tomorrow. I think it’ll be sort of interesting.
Re: David Smith (#84),
Increase in detectability and/or detection effort leading to a positively biased count in later years. AKA ‘Tiny Tim’ hypothesis.
Here’s the time series, this time with a label:
It is a plot of the annual (calendar year) global count of short-duration tropical cyclones. These short-lived systems are often near-shore, entangled in frontal systems, highly-sheared, amd otherwise poorly-defined. They are, in a phrase, junk storms, the international version of Tiny Tim.
This plot suggests the possibility that either (1) more of these junk storms are truly occurring or (2) these junk storms have been there all along but, unlike in the past, are now being counted as tropical cyclones.
The significance of this is that, if one subtracts these “junk storms” from the global database and looks only at “bona-fide” longer-duration storms, then one may find that the global count of tropical cyclones is not constant but rather that it has been slowly declining in recent years. That would be different from the conventional thought about global tropical cyclone count.
(This data comes from an unrelated screening exercise I’m doing using Unisys maps and data and just sort of popped up in the screening. I need to repeat this on a different database to confirm the pattern. But, in my experience with the Unisys data, it can be wrong in spots but it is not wildly wrong over extended periods. We’ll see.)
Here’s the global tropical cyclone count with and without the short-duration storms:
Perhaps the “true” global cyclone count does wiggle and vary over time.
(Another type of “junk storm” for which detection and classification practices may have changed are the ones without a clear warm core, the hybrids where there is no clear core characteristic. Perhaps FSU’s phase-space analysis has become a factor in classification.)
I’ll have to argue with you on the R34 point. I don’t think that R34 is arbitrary at all. I think that the radius of gale force winds distinguishes the storm from the environment. And I have no reason to assume that the R34 in EBT is incorrect, nor Kossin (2007). I guess my point is that I would much rather use EBT or Kossin ’07 to measure energy, which use a number of in situ observations, than an integral of 10 meter 1-minute sustained winds at a single point in the storm every six hours. Physically, I don’t think that variable means very much. What I do think means a lot is that our IKE calculation shows significant variance explained by both size AND wind speed.
What no one has yet to show is how ACE is what it says it is (accumulated cyclone energy). ACE is really a misnomer for TC duration (which I know you don’t argue with). If we are talking about global climate and energetics, a good measure of IKE and PD should be what we strive for.
I will definitely chat with you in email, I’d like to hear what you have done with size.
Re: Angela Fritz (#90),
R34 is a the gale force radius, which is Beaufort number 8 on the Beaufort scale. This scale could be likened to the Saffir Simpson (or Fujita) scale, with categories developed out of non-continuous wind speed data. To adequately separate the storm from the environment, you need a representation of the environment, which is obviously not included in the extended best-track (EBT) data.
Some of the scatterometers and radiance instruments indirectly measure wind speed and direction, but the coverage is probably insufficient prior to 1999. Reanalysis data is problematic because the tropical cyclone depiction is almost always unrealistic or plain crappy. But, if this is the only data available, you can perform a TC removal technique which is sort of the opposite as the TC bogus insertion idea in NOGAPS or the GFDL model for instance. It may be that the average tangential wind speed is 34 knots exactly, but I have not seen evidence of that.
In the Sriver and Huber (2006) GRL paper, they prescribed a storm-footprint of 7×7 degrees (a square) and calculated the average wind speed inside with reanalysis data to come up with the Power Dissipation. Thus, the area-integral even taking account of the change of x with cosine latitude, is largely a constant and can be taken outside the double integrals. However, the R34 does change from storm to storm, and indeed I know that is your point in the argument between ACE and IKE. There can no be no argument to that. That’s why I worked a bit to see if PDI is actually an accurate estimate of the full PD. But, from a climate point of view, year-to-year, the duration component completely dominated the calculation, as it should. The correlations and covariances were large.
I can make a case why the duration component is physically meaningful. I will happily provide that if you can provide the physical reasoning about Category 45/1 ratios being a climate signal 😉
With the end of July, time to update a few tallies on the Northern Hemisphere tropical cyclone activity or lack of it. Indeed, July mustered a record low, blowing away the previous tallies. Indeed, the ACE of 15 during July 2009 was 10% of the most active months of July 1971, 1972 and way behind the previous 30-year average of 68. The hemisphere is so far behind normal, it is becoming worrisome.
I updated my website FSU Tropical with some graphics and the new numbers including ugly looking Tropical Storm Lana.
Scraping bottom just like our current economy. Perhaps we need a Cash for Clunkers program for tropical cyclones?
Here’s a plot of the date on which the first Atlantic storm forms and the final storm count for that season –
It gives a visual impression of the lateness of the 2009 season. Interestingly, it also shows that even a late-starting season can be active, though the odds drop. There’s a hint, based on scant data and perhaps vivid imagination, that the late-starting seasons may fall into two groups, one of which is a definite dud while the other, though late, comes on with a bang.
And here is the same type of plot, with seasonal ACE substituted for seasonal storm count:
Based on history and the current lack of any healthy seedling, the chances of above-average ACE in 2009 are now low.
Re: David Smith (#95), top notch stuff David. However, interpreting the season “start” date physically is not intuitive. Could you instead take the first Category 1 hurricane of the season and see if that shows the same signal as the tropical storms?
If the North Atlantic somehow became hyperactive in the next 2-3 months, it would be the most colossal seasonal forecast blunder in history.
Re #96 Will do.
By the way, I noticed this headline from NOAA
Global Surface Temperature warmest on Record for June . The same report may be made for July. At the same time, though, global TC activity is weak. That doesn’t seem to square with the idea that rising SST leads to stronger TC activity.
It is a weather-not-climate situation, of course, but fun to note.
David and Ryan, you must get off the counts and ACE kick and onto the %Cat45 bandwagon. With the global TC activity counts (and Cat45 counts)and ACE indexes showing no positive trends over the past couple of decades and with a current season showing a dearth of activity, we must divert our attenion to the all important (if not fully explained or understood) %Cat45 index. You simply cannot ignore this metric and its possible connections to upward trending SST when viewed, not over the past couple of decades, but over the past 4 decades (assuming measurement changes notwithstanding).
I look at the current season as one where we may have low TC activity and even low Cat45 activity but the potential for a few Cat45 events in this situation could produce a high %Cat45 index for the season and you know what that means – don’t you?
Here’s a plot of the date of the first hurricane of a season and the total count of hurricanes for that season –
Hurricanes tend to concentrate later in the season due to lower shear, eastward expansion of the warm SSTs, etc. The typical first Atlantic hurricane occurs about August 14. While the 2009 season is “late” in terms of first cyclone, history suggests that even late-starting seasons can have an average count of hurricanes.
Re: David Smith (#99),
David, could you do the same first hurricane of the season plot using %Cat45 on the y axis. I think your post above may be indicating that the later in the season we see much hurricane activity the greater chance of seeing a higher %Cat45. And that means …..
Re 3100 OMG………..
Will do
I just noticed the confusing labeling on my earlier plots. The x-axis shouldn’t show a year (“09”), just the month and day. Sorry.
Ken, here’s the start date versus % cat 4 and 5
I believe this reflects the fact that hurricanes in June and July tend to be weaker than those happening later in the season. Thus the early-start seasons have a bit of dilution from those weaker early storms.
Tropical Storm Ana formed last evening. This August 15’th start to the season is the eighth-latest in the last sixty years.
The other late-starts averaged less than seven storms over their season (an average season has ten) with an ACE of 61 (an average season is about 100).
Re: David Smith (#103), … and you know what someone is going to attribute that to.
…But right now…14:15 UTC, “the area of low pressure” latest NHC
DESCRIPTION…looks like a giant circulation compared to little “Ana”
Very good example for beginners:”Which is the TS and which is the ALP?”[ESE of “Ana”…]
Mann, et al have a new hurricane paper out. It’s being discussed at Roger Pielke, Jr.’s blog. There appear to be smoothing issues.
Now that Bill seems to have peaked, the media is through fulminating about having to watch all those tropical storm systems over the last week and half (so many crises—so little time).
So how about an early season game?
Here’s the Atlantic storms to-date.
http://www.wunderground.com/hurricane/at2009.asp
Based on time, location, and strength, do you think that in 1930:
a) Ana would have been named?
b) Bill would have been called a cat 4 (or would have gone in the record books at 140 mph)?
c) Claudette would have been named?
How about 1950 or 1970?
My answers for 1930: no, no, yes. Same for 1950. 1970? Bill cat 4—maybe.
Re: John M (#107),
Make that 135 mph (hell, let’s just say “greater than 130 mph”).
Re#107
John M, Tropical Storm Danny should be added to the list of dubious 2009 storms. I’m not convinced it has yet been a tropical cyclone or would have been so classified in prior decades.
Maximum winds are far from the center, the core has been mostly void of rain and the central pressure has not been been particularly low. When it initially formed the analysis indicated it was a coin toss as to whether it was tropical or subtropical.
We’ll see once all the evidence is compiled at the end of the season.
Re: David Smith (#109),
If the lack of Atlantic activity continues there is probably going to be a scramble to name anything with rain and wind as a TS.
Re: BarryW (#110), Yep. This year a big-enough school of splashing dolphins might get named.
When reading this:
“The UK Met Office, which recently had its climate research budget slashed 25%, has determined the most likely amount of storms is 6, very low compared to what has been seen since the active period in the Atlantic began in 1995. The most likely Accumulated Cyclone Energy (ACE) tally is predicted to be 60, which is in the basement compared to recent seasons.”
I can’t help but think of:
“There are major differences between real science and bureaucratic science (BS). Real science involves living with the prospect of failure. In BS, failure is not allowed. The whole project is mapped out beforehand in forms such as Gantt charts. There are deliverables that have to be delivered on the due date. With the exception of really big physics, real science is carried out by small groups. It is the same with BS, except that there are about five managers for every researcher. Above all the expected result must be delivered on time. Those who desire further patronage never report a negative result or, indeed, a result at variance with the expectations of the sponsors.”
http://www.numberwatch.co.uk/lying.htm
Of course, perhaps the Met Office budget cuts have nothing to do with the current lack of new hurricane research available to support government propaganda.
OK, question about Erika now.
It hung around quite a while and eventually dropped a lot of rain on Puerto Rico.
Anyone want to venture a guess as to how it would have gone in the record books 60 or 80 years ago? I suppose a “dumb ship” could have easily wandered into it.
BTW, I guess I didn’t realize I had asked a trick question in 107. When did they start naming hurricanes?
John M, the naming of Atlantic storms began in the early 1950s.
Standard military names (Able, Baker, Charlie, etc) were used for several years starting in 1950. Then, in 1953, US female names were used. The name list became gender- and nationality-neutral in the 1970s.
An interesting website is this one , where you can find the estimated wind fields of the 2009 storms. Most of the earlier seasons can also be found via the pull-down menu.
Here’s the wind field for Erika at its max:
The system had a small but respectable field of 35+ winds over open ocean. This was short-lived, though, and the field seems to have collapsed within 24 hours.
Would a ship have encountered this? Possibly, but not likely, in my opinion. Would the islands have seen storm force winds? No. Would the barometric pressure in the windy region been low enough to lead one to suspect a storm? No. Would the time of lowest pressure been roughly coincident with the time of greatest windiness? Probably not, unless a ship was traveling northeast across the center. Were there definite surface westerly winds to indicate a cyclone? Yes, but they were weak.
My guess is that this system would probably have been assigned to the Monthly Weather Review’s list of “North Atlantic Gales” for the month and never be elevated to a suspected tropical storm. My sense is that, prior to about 1940, the agency looked for certain triggers in ship reports (low pressure, shifting winds, strength of winds, more than one report, etc) before elevating a “Gale” into a tropical cyclone. Erika probably would not have triggered the elevation.
http://www.latimes.com/news/nationworld/world/la-fg-mexico-drought7-2009sep07,0,6988447.story
here is an interesting take on the “Warmer World is a Calamitous Event” (it is Hurricane relevant)
regards
tty
Your comment on the Sea Ice thread:-
Today’s weather in Algeria is advecting from the south towards southern slopes of the Atlas Mountains. Since the start of September, as this convecting air mass moved across the central Sahara, the convection has been creating dust storms north of Tamanrasset (e.g. EUMETSAT FRAME_OIS_RGB-dust-westernAfrica_0909022000.jpg)
Monsoon anyone?
With the naming of Fred this week, and its eventual rapid intensification to major hurricane strength way out in the Eastern Atlantic, the Hurricane season has seen its 6th name storm. Fred was the replacement name for Fabian back in 2003 which ravaged Bermuda.
So, with the 6th storm named, regardless of the intensity and longevity of each: Baby Whirl vs. Major Hurricane… we are at the peak of the Atlantic hurricane season, so it is likely and expected that an additional 4-5 storms will at least develop through the rest of September and October bringing the yearly total to 10-12, which is just about average. The ACE will likely remain near-normal, probably 75-90.
In terms of Northern Hemisphere overall activity as measured by ACE, 2009 continues the spate of depressed tropical cyclone activity seen since 2007. The climatological ACE for the NH (thru Sept 7, based upon 1979-2008 averages) is 282. Currently the NH ACE is 188. Here is a text list of the previous September 7 to-date totals for NH ACE. LIST
Not much else to update except the status quo. The Western Pacific is very very quiet.
Re: Ryan N. Maue (#117), For the Atlantic, this is very “average” for this point in the season in terms of named storms:
But Hurricanes are still a little behind (probably due to the season’s slow start):
Sorry to be off topic, but there isn’t a Saharan rainfall thread here.
Here is a link to the TRMM record showing the last 7 days satellite estimate of the rainfall amounts in Africa .
Note the rainfall totals for Saharan Algeria, south of the Atlas mountains (date of posting 9th September) and in particular the rainfall in northern Mali and last week’s flood disaster in Burkina Faso.
Re: PhilipM (#119), why do we need a Saharan rainfall thread…not sure anyone else is clamoring for it. It floods a lot in the tropics, please tell us why this is important for Hurricanes in the Atlantic aside from the well-known Sahel connections to tropical waves and the tropical easterly jet.
Re: ryanm (#120),
ryan
The weather has to go somewhere. More rainfall in the Sahara, more dust, more atmospheric damping of the eastern Atlantic, less hurricane activity.
Re: PhilipM (#121), Please tell me how wet sand gets blown up into the Atlantic. That’s nonsensical.
Re: Andrew (#122),
Andrew
In the Sahara not all precipitation from convective cells reaches the ground, most of the falling rain evaporates in the hot lower tropospheric desert air. That is why I made the comment that the precipitation amounts measured by TRMM are an only an estimate of rainfall. All satellite analyses, including rainfall estimates, needs to be ground truthed against surface measurements. Not an easy task in the uninhabited Great Western Erg.
What is clear however is that the cold down-draft associated with the evaporating precipitation continues down to the ground surface where it spreads out in an arc picking up dust and forming a haboob .
Plenty of examples of haboobs in the EUMETSAT images. Of particular interest is the way major dust clouds form the site of new convection cells in a self-seeding process of repeated convection, that lasts over a number of days.
Re: PhilipM (#123),
I had noticed that most of the few weather stations in Western Sahara have been reporting rain the last several days, in some cases in large amounts:
http://www.uni-koeln.de/math-nat-fak/geomet/meteo/winfos/wetterk_africa_world.html
I think much of the area is probably too wet for dust storms right now.
Re: tty (#134),
tty
Thanks for the link. Very useful.
So, if you were watching the post-Obamacare speech on Fox News last night, you would have seen Accuweather’s Joe Bastardi pontificating about the California wildfires and discussing global warming. He then went on to present “Accuweather’s” version of my NH 24-month ACE plot as depicted above, which was published in Maue (2009) and is seen on my website. The graphic and the underlying idea of plotting it (24-month running sums) is clearly my intellectual property and someone from Accuweather (Joe Bastardi) had better stop using my work without proper citation or in this case — PAYMENT.
Re: Ryan Maue (#124),
You have one year from first public disclosure to file for a utility patent on this. The claim would be something like “a computer means for generating a representation of hurricane data for display” as well as a another claim for “transmitting the image for reception on at least one of a a television set and a compute monitor”.
Haboobs also are seen in the US southwest desert areas.
Re: Fred Harwood (#126),
I see a lot of them at the beach.
Oh, wait.
Never mind!
Re: Jeff Alberts (#127),
No Jeff, arcs not curves.
Re: Fred Harwood (#126),
Re: Jeff Alberts (#127),
Re: PhilipM (#128),
Seriously, haboobs are EXTREMELY common in Phoenix, where they’re just called ‘duststorms’. In fact, they can occur most every evening during Phoenix’s “monsoon season” [roughly mid-July to early September.]
Re: jc-at-play (#129),
jc
Thanks for the tip. Much appreciated!
Arizona Dust Storm and Monsoon Pictures
Meanwhile Hurricane Fred north west of the Cape Verde Islands becomes a Tropical Storm and is forecast to weaken further.
So, Anthony Watts helpfully cross-posted Bastardi’s appearance from the O’Reilly Factor, and I helpfully providing a screen shot. I never received attribution from Bastardi for using my research property meaning the ACE graphic :in the format: that I present it. When Dr. Steve Lyons at the Weather Channel used it a month or so back, he asked for and received permission. Look for yourself: Same time period, same y-axis labels. Looks like they just used a tracer to dump my graphic into their system.
Update: Sept 21 — all is well, Joe Bastardi and I had a wonderful conversation & are moving forward after this misunderstanding.
The latest from Judith:
http://www.gatech.edu/newsroom/release.html?id=3262
Comments?
Re: Andrew (#133),
Do climatologists get paid for each use of the word ‘unprecedented?’
Re: David Cauthen (#137), Wouldn’t know, but sometimes it seems that way, doesn’t it?
Re: Andrew (#133),
Typical imo. Here research is constantly changing…counts, then ace, then the next metric, then tornadoes with a constant refrain of “the she feels that AGW is likely true”. If only she could settle on a course of research.
“I bless the rains down in Africa…” – Toto
Rich.
Economists make predictions about unemployment, consumer confidence, GDP. The media with the Associated Press the biggest offender, always describes the news/numbers as “unexpected”. All economics to the liberal media is “unexpected”.
Re: Ryan Maue (#139), You might be interested in this:
http://www.climateaudit.org/phpBB3/viewtopic.php?f=3&t=781
I haven’t gotten to the details yet, but I’m planning on seeing what happens to North Atlantic ACE if you exclude the weakest, most short-lived storms. It would be nice if you could direct me to more accurate sources, to. Wiki so far is all I can come up with, and I’m not clear how exactly they got their numbers (I presently doing a sanity check against the seasonal totals-could take a while).
Re: Ryan Maue (#139),
Old joke: “The nice thing about having amnesia is that you are always meeting new people and learning new things.”
I have a file with all the global storms with a zillion statistics about each including lifetime, hurricane days, ACE, PDI, etc. I will dig it up and try and provide some documentation.
Wiki isn’t to be trusted in general.
Re: Ryan Maue (#141), Thanks, let me know! 🙂
An odd aspect of the 2009 season is that, so far, four storms have dissipated at sea, over tropical (26C+) water. That’s a very high number and may be approaching a “record”.
Re: David Smith (#143),
Anecdotally, it seems as if ‘shear’ has been killing off the storms for the last two years. Are upper level winds stronger?
Re: David Smith (#143), wow David — you have a very good paper idea (please no one else read this). But, if you could look at the dissipation characteristics of hurricanes or storms as a function of intensity over > 26C water over the past 30-years in the North Atlantic, you should see the following:
The effects of large-scale climate on the hurricane tracks themselves as well as their proximity to the SST gradients. Probably the location and strength of the Bermuda (subtropical) high would be the dominant mechanism controlling the magnitude of the shear. With archived satellite imagery over the past 30-years every 3-hours for the globe + reanalysis datasets of dubious quality, you could still say quite a bit about the behavior of cyclones before dissipation.
Re: David Smith (#143), Maybe I misunderstood it, but per Vecchi and Knutson (2008) aren’t cyclones which dissipate over the ocean precisely the sort likely undercounted in the ship based observation era? Having never quite made it to shore, their chances of getting noticed way back when would have been poor, right?
I suspect that wind shear and, to a lesser extent, dry air have been snuffing out the storms this year.
Interestingly, one of this year’s storms (Erika?) clearly suffered from shearing winds below the standard 200mb level, perhaps at 300-400mb, which tilted the system and eventually killed it. The standard shear analysis (200mb-850mb) probably would not have detected that. I sometimes wonder if mid-level (say 300-400mb) shear is one of the poorly-recognized factors in how a storm, or season, behave.
Andrew, I agree that the storms which died at-sea and/or were short-lived and/or were in remote regions were the ones which often went undetected in prior to modern detection, esp satellite. Several of this years storms would likely have been recorded as a “North Atlantic gale”, if at all.
I noticed that Belanger (2009) has a hockey stick:
Interesting. The rest of my September is busy but I think I’ll visit the paper in October.
Re: David Smith (#150),
http://en.wikipedia.org/wiki/Hurricane_Beulah
Tornadoes: 115 (in 1967)
Not broken until Ivan:
http://en.wikipedia.org/wiki/Hurricane_Ivan
http://en.wikipedia.org/wiki/Hurricane_Ivan_tornado_outbreak
Tornadoes: 119 (2004).
Now what did they say in their press release?
Three things strike me immediately- First, I can’t find where Beulah in their chart, either “observed” or “modeled”-damned smoothing!. Second, Beulah could well have had more tornadoes than Ivan. Third, their model uses intensity and other measures to predict past tornadoes-but these records aren’t homogeneous either!
Not to mention that their calibration period is entirely within the positive phase of the “AMO”.
Re: Andrew (#151), please send me an email rmaue at met.fsu.edu & I will point u to a per-storm information dataset for global TCs from 1979-2008…
Re: ryanm (#152), Will do…As soon as I’m at my home computer that is. Which should be in a couple of hours. Thanks! 🙂
Re: ryanm (#152), Ryan, Your latest update seems to have a typo in your graphs:
http://www.coaps.fsu.edu/~maue/tropical/
The labeling of the two images is identical:
They both say >64 knots. That’s wrong, right?
Re: Andrew (#154), one is 24-month running sum, the other is 12-month. With the Southern Hemisphere tropical cyclone season occurring mainly from December – March, when the Northern Hemisphere is asleep, I find it useless to use calendar years — especially with ongoing or terminating La Nina or El Nino events affecting the frequency of TCs in the Southern Hemisphere.
Just wondering if Accuweather’s resident Mr. Hurricane will steal these images too.
Ryan have been following your research with a whole lot of interest. Damn impressive.
You gotta great mind and a whole career worth of scientific advancement for the human race ahead of ya. Godspeed.
Hope you get an explanation and apology from Bastardi for using unattributed research. I don’t blame you for being ticked.
Ironically, he endorses you on his blog as if this did not happen:
He owes you an explanation….even if he erred unknowingly.
Chris
Norfolk, VA, USA
Re: Christopher Shark (#156), Thank you for the nice comments. Back in March 2009, I gave permission to Jesse Farrell to use the graphic (which appeared here on Climate Audit) in his blog postings. He agreed in an email to both provide a proper citation and a link to my tropical website. That’s all that I ask. However, when the face of the company (Mr. Hurricane) goes onto a national cable network (Fox News) and passes off my work as “Accuweather’s” with their logo, it is troubling. From the y-axis units it is 100% clear that the ACE quantity is the 24-month running sum. Looks like they simply traced my plot into their graphics system.
I do not have a monopoly on the ACE metric and publicly available hurricane data, but I do have intellectual property rights to “value-added” products which I produce for my research and publish. Unless I receive an email today from someone at Accuweather, I am passing off this issue to the University lawyers.
Ryan-I’ve been looking over those files you pointed me to. It’s interesting…there seems to be a strong trend in the annual ACE in the Northern Hemisphere for storms with ACE less than 1. What would you make of that? I have some thoughts but actually I’m surprised such an effect would show up even since 1979.
Re: Andrew (#158), I have not read anything in the literature that would explain an AGW induced increase in very weak, short-lived tropical cyclones. However, I haven’t looked at the track locations. It would be my first guess that increased detection capabilities associated with improved monitoring mainly by satellites would be the reason why.
Even though the number of global tropical cyclones undergoes large interrannual variability, there is no discernible trend during the past 30-years. Removing all storms that fail to reach hurricane strength gives a similar picture.
Re #158, #159
Regarding tracks and locations, here are a couple of plots I made some months ago on the origin points of short-lived (24 hrs or less) tropical cyclones.
The first shows pre-1946 –
.
.
.
The second shows 1946-2007 –
.
.
.
The remarkable concentration along the Gulf Coast is unphysical. It hints of coastal buoys, radar and oil rigs.
Re: David Smith (#160), Very interesting, although my finding more directly reminds me of your post here:
http://www.climateaudit.org/?p=6383#comment-349210
Because of the larger geographic scope and the shorter time period.
I think that Landsea, Vecchi, Bengtsson, and Knutson (2009) showed rather convincingly that the really short lived storms have an impossibly large effect on the trends. Ryan has mentioned it before, but your post just now reminds me of it and just in case you hadn’t checked it out (it’s fantastic):
ftp://ftp.gfdl.noaa.gov/pub/gav/PAPERS/LVBK_08_SHORTSTORMS.submitted.pdf
Re: Andrew (#161),
“This is a disappointing paper in many respects, particularly given the involvement of certain researchers (Vecchi, Bengtsson, Knutson) whose work has generally been far more careful and solid than this.”
— Anonymous reviewer of a paper that was published in JClimate by Landsea, Vecchi, Bengtsson, and Knutson
Re: Ryan N. Maue (#162), I can find no specific fault in it, so I’m curious what “carefulness” is believed to have been abandoned here. Were they not sufficiently careful to avoid suggesting that AGW won’t lead to more tropical storms? If so, why should they be careful about implying something that their study suggests? I’m confused to say the least. I smell some serious tribalism here and I suspect someone is upset at a perceived “defection”. Sigh.
Creating a new post for an update to the Northern Hemisphere Tropical Cyclone year of 2009: which is “circling the drain”.
As a preview I posted this to an email list: and it is also on my tropical webpage (which I am converting to a blog).
As far as I can tell using the best-tracks, the last day in September without Accumulated Cyclone Energy (ACE, Bell et al. 2000) being registered was September 24, 2003. During the past 20-years, *15* September days have not had an active
tropical cyclone with *5* of those days occurring at the end of September 1999.
Through September 21, comparing 2009 to the previous 20 Septembers for NH ACE, the current total from the operational advisories of about 71 is one standard deviation below the 20-year mean (mean=117,sigma=36).
These links are for two plots that show the yearly Northern Hemisphere ACE for September and the average ACE per day during the month.
Since daily ACE represents a 4-times daily sum of wind speed squared, an “average” September 21st could see one of the following (among other combos):
One TC at 125 knots
Two TCs at 90 knots
Three TCs at 70 knots or
Six TCs at 50 knots
Current TCs = 0
Anthony Watts has helpfully copied my website particulars to his awesome blog — please take a look. Watts Up With That.
There is a decent chance of a tropical depression or storm forming in the Eastern Pacific during the next day or so, but nothing looks remotely close to brewing in the North Atlantic — is the season done?
What are the odds that another hurricane will NOT form in the Atlantic in 2009 ?
Re: Ryan Maue (#165),
Let’s ask Judith!
In 15 of the years since 1945 no hurricanes formed after September 30th in the Atlantic basin. Although 4 of those years were in the last 7. 2002, 2003, 2004, 2006.
Quite a difference depending on whether you play the short term or long term odds.
Re: Bob Koss (#167),
Gosh, sounds like a trend to me. Now what do you suppose we can attribute that to? Hmmmm, what seems to cause every unusual (and maybe not so unusual) event these days?
Maybe we’ll see several papers on this next year?
Re: Bob Koss (#167),
http://en.wikipedia.org/wiki/Gambler%27s_fallacy
Here’s an interesting plot:
This shows the annual count of tropical cyclones which formed east of 80W (Miami) and then made landfall on the US Gulf Coast. These tend to be longer-lived than storms which formed in the Gulf or Western Caribbean and include some monster Cape Verde storms. That longer duration probably leads to greater average breadth and, arguably, greater average strength.
Eyeballing the (imperfect) data gives the impression of concentrated activity in the 1930s and since the late 1990s. That probably conforms to AMO activity.
Hurricane breadth (aerial coverage) is apparently a factor in tornado generation. Given that, it seems like the active 1930s would have been a good choice for comparing with the active 1995-2008 period. However, if I recall correctly, Belanger (2009) used 1948-1964 (red line) for the comparison, a period that was rather lacking in the eastern-born (likely longer-duration) storms. So, for tornado comparisons, I wonder if the 1930s would have been a better choice than 1948-1964.
Note – the pre-WW2 weather map data may be of such poor quality that it’s of little use for the tornado work. Also, the eastern portions of hurricane tracks were probably soft on accuracy due to limited sampling. So, beware the quality trap.
October it’s not over.
Tropical Storm Grace
Tropical Storm Henri
NHC Henri Advisory No1
Henri is a baby whirl that may not last too long but gets the same weighting as a “storm” as Katrina. With Grace up by Ireland now, wonder if these later season developments are evidence of some sort of climate signal? (uh, no.)
Re: Ryan Maue (#172),
Ryan, I noticed that Jeff Masters was making a big deal about Grace being the second most northerly forming TC since satellite record keeping began in 1960. Now, he didn’t say anything about a climate signal, but any enterprising alarmist writer, blogger, etc. could easily seize on that comment and attempt to make something out of it.
Re: Mike B (#173), astronomers will announce in Nature that Saturn actually has an enormous radius but microdense ring that surrounds the planet — that was up till now undetected. As new satellites and instruments are launched, we begin to see even more exciting things about our solar system. Seemingly in 1960, the detection of a very small tropical storm so far north would have been difficult to say the least — especially if it was only known as a pure tropical system for 12 hours.
Re: Ryan Maue (#172), On a positive note, it does NOT get the same weight in ACE, since it should be 1. Short lived and 2. Weak
Re: Andrew (#175),
Please pardon the OT. Andrew might be interested in the latest Tsonis & Swanson offering. It is about circulatory modes as unstable eigenthingies (for lack of a better term). Of course, it is relevant to hurricane climatology, but it is also somewhat tangential.
Re: bender (#176), Definitely worth a look. Thanks.
And of course the Atlantic and Pacific eigenthingies have often been said to be associated with hurricane activity in there respective basins. Not endorsing that, but it’s been said.
Re: Andrew (#177),
Yes, that is why it could be relevant. A theory of “punctuated activity” – like the current “warming pause”. It could be Ryan’s next big paper.
They seem to be calling the 2009 hurricane season before it’s officially over: link
I noticed they were blaming the lack of storms on shear due to el nino.
Re: BarryW (#179),
It ain’t over till the GT lady sings.
Re: sky (#180),
Indeed. “Different” sort of El Nino, and all, wot?
http://www.climateaudit.org/?p=6383#comment-347631
http://www.climateaudit.org/?p=6464#comment-347701
That latter comment was based on a “Modoki” El Nino, with Nino 4 anomaly greater than 1 deg and Nino 3 less than 1 deg.
So we have to wait for the “late season” to run its course.
As of Oct 26 Nino 4 was 1.4 deg, Nino 3 was 0.8 deg
Re: John M (#183),
Well now it looks like both 3 and 4 are over 1 degree.
So does that mean this is no longer a Modoki El Nino?
Does this remove the impetus for an active late hurricaine season?
If instead, we get a blurp of late season activity like this guy says might happen, what then?
Will the argument be that most of the season was influenced by a Modoki El Nino?
Join us later this month for the next episode of…
As the World Churns.
The 2009 season may well set, or tie, a couple of minor records.
In number of hurricanes, 2009 is tied with 1982 for the fewest (just two) of the last sixty years –
And a more-obscure one is the farthest-west longitude of any hurricane during the season. The 2009 season has seen no hurricane winds west of 68W –
Speaking of hurricanes, Eric Berger has an article this morning quoting Ryan Maue and others on the matter of global storms and global warming.
Houston Chronicle; Global storms and warming
The Modoki paper was based upon a couple extraordinary years. If 2009 is designated as such, then the exceptional weak Atlantic season pretty much wipes away that hypothesis.
Re: Ryan N. Maue (#185),
Pardon my cynicism, Ryan, but the execptionally weak hurricane season means there is little chance that this year will be classified as Modoki.
Summer 2009 is of the mixed type (both El Nino and Modoki), although this winter could be heading for Modoki
Re: Judith Curry (#186), Have there been previous “mixes”?
Might be yet another variable for consideration-what happens to hurricanes during a “mix”.
Of course, then again it becomes a little cumbersome and inelegant a theory.
Here’s and idea. The proposed mechanism behind conventional El Nino effects is wind shear. Why devote so much effort to El Nino, then, and it’s curiosities, and not the wind shear?
Just as there may be seasonal “flavors” of ENSO (modiki, conventional El Nino, La Nina, neutral) there may also be seasonal “flavors” of wind shear.
I’m increasingly inclined to think that, during a season, the conventional 850/200 mb shear might behave differently than the 850/300mb or 850/400mb shear, affecting that seasons storm count and intensities.
And, this seasonal lower-level wind shear behavior may be driven by currently-unrecognized seasonal weather modes.
Re: David Smith (#188),
David, when can I make my 2009 NATL hurricane, TC and ACE predictions for the contest. The predictions were based a lot on wind shear analysis and a little bit on Modoki – although he was not a big help. I also threw in percent Cat 45 hurricanes, average hurricane size and number of tornadoes generated by hurricanes for good measure.
Re: Kenneth Fritsch (#189), Ken, we’ll open the 2009 contest in a few weeks, around US Thanksgiving Day. I feel super-lucky this year.
Re: David Smith (#191),
David,
I thought we did have a contest already this year? See Re: Jonathan Schafer (#10). I was down for below normal ACE. I want another certificate of achievement, having correctly determined the season two of the past three years. 🙂
New form of El Nino may increase Atlantic hurricanes
From USA Today Article July 2, 2009
Re: Ryan N. Maue (#190),
Would that predicting be from empirical data or a computer model?
.
Did not we analyze a paper at CA not long ago on partitioning an index to supposedly improve TC/hurricane forecasts?
I take back my criticism. The inclusion of Hurricane Ida is well predicted by the Modoki El Nino partition. One data point is enough for me especially when only a few years are considered.
So what’s the shortest hurricane on record?
Re: John M (#196), poor observational ability in the early Atlantic record shows 14 hurricanes with only one observation. Three hurricanes since the air-recon era in the mid-1940s have 1.75 days of observations. ANNA in 1956 being the most recent.
Since 1950, I have a list of the North Atlantic storms, their ACE, maximum intensity, and number of hurricane days, etc. From my list, there have been some similar storms to Ida that have a maximum intensity of 65 kts and for only 6 hours. These include:
Interestingly, the bulk are in the past 6 years. Hmmm… These storms have low ACE as they are short-lived in general. Ida hasn’t completely given up and may reintensify during the next few days as it enters the Gulf of Mexico.
Re: Ryan N. Maue (#198),
C’mon Ryan, you’re not trying.
Gotta find those pseudo-superlatives wherever we can.
Re: John M (#199),
Aha! Something has changed – and it cannot be good.
.
My latest 2009 prediction did have a November hurricane making landfall, although I might have been off by a day or two.
Re: Ryan N. Maue (#198), Ryan, here’s a plot of NATL windspeed observations since 1978:
The frequency of 65 kt (minimum hurricane) observations stand out. I wonder if there’s some kind of measurement or interpretative bias which nudges the number upwards.
Re: David Smith (#200), first appearance of an eye perhaps would herald a 65 kt reading.
Re: Ryan N. Maue (#198),
Erika, Gaston, Cindy, Vince and Ernesto are a rag-tag assortment of storms.
Erika, Gaston and Cindy were classified as hurricanes after their seasons ended, during reanalysis
Vince and Ernesto had no surface or aircraft readings of hurricane-force winds but satellite and other bits of data allowed them to be classified as brief hurricanes.
My belief is that, in earlier times, analysts had neither the instruments nor the inclination to look for brief intensity bursts. Thus, this is mainly a modern phenomena.
That unnamed 1959 hurricane was really moving along. Traveled almost 2500 nm. in 4 days. They likely left it unnamed until they decided after the season that it qualified as a tropical storm.
Looks like Ida is bone fida (poetic license, eh)
So in the continuing saga, will Ida now be the first November storm in an odd-numbered El Nino year making landfall as a hurricane twice?
Is she the daughter of the son of Modoki?
The sister of El Nino’s half-brother?
This must mean something profound.
Anyway, hope no serious damage results.
Oops.
Wrote that before I read the news from El Salvador.
Make that, hope no more serious damage results.
Ida may be a harbinger of a dangerous trend, consistent with increasing CO2 and AGW:
From 1950 to 1984 there were none. Then, consistent with the rise in CO2, we had “Juan” in 1985. Now, in 2009, we have “Ida”. Looks like an expotential trend to me, which Excel kindly plotted.
By 2050 we’ll likely have double-digit double-voweled November Gulf storms.
Re: David Smith (#207),
In that case it seems to me there is a cheap and easy fix for the problem. Just stop using two vowelled names and the problem will flatline …
😉
Re: David Smith (#207),
Hmphh. Reasoning from two data points. Of course, that’s one more than most theorists use.
Re: NW (#210),
Of course, one point defines a curve, two points defines a family of curves!
Ida has been downgraded to a tropical storm according to my local, Florida news report. I doubt it’ll get back to hurricane strength as close to land as it is, but who knows.
Re: David Smith (#207),
David, did not naming start in 1941? What happened from 1941 to 1949 with regards to two voweled November hurricanes or are we waiting for re-analysis? Perhaps, if we went back and named storms back to the late 1880’s you would have better statistics – although I doubt it would change that trend much.
Re #207 Oops. “Expotential” is, of course, a misspelling. And, perhaps less obviously, it should have been written as “5’th order polynomial”. Speed kills.
Roger K. Smith, M. T. Montgomery and S. Vogl recently published “A critique of Emanuel’s hurricane model and potential intensity theory” (Quaterly Journal of the Royal Meteorological Society 134: 551-561, 2008). In this paper the basic foundations of the potential intensity (PI) theory are criticized that were laid down as early as in a 1986 paper of K. Emanuel.
The emphasis is that of the authors.
In my view, this publication might be interesting for a general CA reader in several aspects. First, a serious drawback in a mainstream, widely cited and widely applied model is claimed to have been spotted that has escaped critical attention from 1986 to 2008, i.e., for 22 years. Second, it will be informative to see how this criticism will be handled by the criticized party (the critique appeared on-line 14 May 2008, i.e. a year and a half ago) as an illustration of how progress is made in modern climate science. Finally, the main question is strictly scientific — what is the physical effect that determines the radial pressure gradient in the hurricane independently of the stipulations of Emanuel’s model?
Re: Anastassia Makarieva (#213), this is not the first critique of Emanuel Potential Intensity theory and won’t be the last. In my opinion, Smith et al. (2009) attempts to improve the theory not discard it. Thus, this is actually probably not of interest to the general reader since it requires a few years of in depth study of tropical cyclone dynamics and thermodynamics beyond what most PhD students even deal with. When you can put the abstract of the paper in layman’s terms, let us know.
Re: Ryan N. Maue (#214), In my opinion, a “general reader of CA” is different from a general “general reader” in being critically-minded, receptive to common sense and capable of learning. Plus a little bit of physical training — and one would be able to occupy oneself with the papers of K. Emanuel and that of Smith et al. and the like not less meaningfully than, e.g., with the reading of Mann et al., Briffa et al., etc., and their critique.
The hurricane problem at large can be seen as the physical task to explain what dictates the observed wind velocities up to 90 m/s that are well above the global mean wind speed of a few m/s. In the dynamics context, there are four variables involved: three velocity components (radial, tangential and vertical), air pressure, and distance to the center. There are three conservation laws: angular momentum, mass and energy. But three equations are not enough to solve the five-variables dynamic problem. An equation relating air pressure to distance must be found. The fame of the PI theory of K. Emanuel has to do with the fact that it claims to derive the missing relationship of air pressure on radius from thermodynamic considerations, i.e. viewing the hurricane as a heat engine. (BTW, there is a brief overview of this theory for physicists in Physics Today (K. Emanuel, Hurricanes: Tempests in a Greenhouse, August 2006. p. 74).)
As I understand, Smith et al. namely criticize this derivation, which is fundamental to the PI theory overall consistency. They claim that the dependence of air pressure on distance is made under an unphysical assumption of gradient wind balance at the surface. In simple words, gradient wind balance means that for any air volume the inward radial acceleration is balanced by the outward acceleration of the centrifugal force. (E.g., a satellite rotating around the Earth at a constant orbit is an example of such a balance, with the gravity force “pulling” it inward, and the centrifugal force outward. In the case of air rotation the inward-pulling force is that of pressure gradient, with air pressure diminishing towards the hurricane center.) Absence of radial acceleration results in a zero radial velocity. Smith et al. numerically illustrate the obvious implausibility of such an assumption for the hurricane. The overall conclusion is that the pressure gradient [above the boundary layer] must be prescribed or determined independently. This essentially returns one where one was from the very beginning: facing the need of an extra equation for the relation between pressure and distance.
Therefore, I would disagree that Smith et al. attempt to improve the model. Rather, they spot a critical inconsistency in the physical foundations of the model. In their own words, they show that
Re: Anastassia Makarieva (#216),
Well in the dynamical case there is actually one variable more (if one forgets the multiphase problem due to the presence of water vapour) : the density .
There is also the detail that momentum conservation (Navier Stokes) is a vectorial equation so it actually gives 3 scalar equations .
All that makes 5 equations , 5 unknown scalar functions (3 velocity components , pressure , density) and 4 variables (3 space and 1 time) .
I am not familiar with what Emanuel did but there is no missing equation for the purely dynamical problem – one has the right number of equations (5) for the number of unknown functions (5) .
If heat transfer is involved (latent heat exchanges , convection) then it gets significantly more complicated and 2 more unknown functions must be added – the temperature and the humidity .
Re: TomVonk (#217), The main physical problem is that in the hurricane the air accelerates. If one recalls the appearance of Navier-Stokes equations, pressure gradient is located in their right-hand side together with any external force (e.g., friction, gravity, etc.) acting on the considered air volume. When the horizontal air motion is considered, there is no external force that would lead to accumulation of kinetic energy except for the radial pressure gradient itself. Therefore, if one does not specify the pressure gradient force and has no clue for its physical cause, the only solution one gets from the full set of hydrodynamics equations is a trivial one with radially constant pressure and zero velocity.
This is simple but fundamental in importance: if one does not encode a physical driver of air motion into the Navier-Stokes equations, no non-trivial solution can be obtained. For a layman, the equations themselves do not know whether the air moves or not.
Thus, the formal calculations of the unknown functions and variables in hydrodynamics equations as in #217 are not directly relevant to the above problem of running away form a trivial solution. Yet for the sake of being precise, for an incompressible fluid (and air is considered at constant density in the boundary layer by both Emanuel and Smith et al.) there are three scalar Navier-Stokes equations plus mass conservation (continuity equation), i.e. four equations for four unknown functions (three velocity components and pressure) as dependent on spatial coordinates and time. Energy conservation and angular momentum conservation are comprised by Navier-Stokes equations; they cannot be summed up in a line with the former as independent stipulations.
Re: Anastassia Makarieva (#218),
This is not correct .
You will have surely noticed that I didn’t speculate about the form of the solution for example an uniformly vanishing velocity field . I merely pointed out that the system was closed , e.g that there were enough equations to solve for the unknown functions .
To recapitulate :
6 unknown functions – density , pressure , 3 velocity components and temperature . Of a particular note is that temperature appears only in the energy equation .
.
6 independent equations :
1 continuity equation (mass conservation)
3 scalar Navier Stokes equations (momentum conservation)
1 energy conservation equation (featuring T)
1 equation of state p = f(T,rho) . This equation couples the energy conservation equation to the N-S equations .
.
If the hypothesis of an incompressible flow is made , we have
div (V) = 0 (btw this doesn’t mean that density is constant only that the material derivative Drho.Dt is 0 . Another way to say that is that the fluid is homogenous) .
As this assumption adds one equation featuring only velocity components , it effectively decouples the energy conservation from N-S and the velocity field can be solved without using the energy conservation .
As for the angular momentum conservation this is a consequence of the fact that the stress tensor is symmetrical . That’s why it is (almost) never used and instead it is implicitely assumed that the stress tensor is symmetrical .
.
Of course I agree that in order to fully define the system it is necessary to specify both the stress tensor (6 components depending on the characteristics of the fluid) and the external forces .
In addition comes that if the flow is compressible , it is necessary to use the energy conservation equation (as well as the state equation) what leads to the necessity to specify the radiative properties of the fluid (emissivity etc) .
This case is so untractable that in studies of atmospheric systems the incompressibility hypothesis is always taken and as the Mach numbers are always low , this hypothesis is also correct .
.
What I like about your approach is that you use the analytical method instead of overparametrised numerical models which would have made me immediately skeptical .
If I understand well , what you do is to specify the external forces (basically gravity), derive pressure profiles from the energy conservation (assuming water phase changes) and state equations and then inject this in N-S to find the velocity field .
I have not read the papers so won’t comment on the detail but the method is certainly commendable .
What I would have to check is how you treat the stress tensor because THIS ONE is of paramount importance for the turbulent energy dissipation .
And the rhyme for November is?
November Remember!!
My friend tells me that one of his favourite books in Jamaica as a kid was “Hornblower in the West Indies”… Hornblower was sailing home from the end of his tour in November, when he was ship wrecked by a hurricane.
I for one appreciate the efforts to explain the hurricane models and the weaknesses thereof.
Re: Kenneth Fritsch (#219), Re: jeff id (#220), Many thanks for your interest and kind words.
Actually in the paper Jeff mentions we arrived at the same conclusion as Smith et al. did although we pursued a very distinct formal route. Regarding the proposed physical mechanism for hurricanes and tornadoes, in my opinion it is so transparent that it should be clear to most educated people.
Air pressure at the Earth’s surface is approximately equal to air weight in the atmospheric column. Weight is proportional to the number of gas molecules in the column. Condensation removes water vapor from the gas phase, reduces the weight of air column and, hence, air pressure at the surface. In the result, air starts to stream from the neighborhood to the area where condensation takes place. It brings together more water vapor, which sustains condensation and low surface pressure. The condensation-induced wind persists as long as there is enough water vapor in the inflowing air. Whether it will be a mild wind, a hurricane or a compact tornado, will largely depend on the horizontal size of the area where condensation takes place, and on surface roughness.
In simple words, mass non-conservation of a gaseous component of the atmosphere leads to a release of potential energy, which is converted to kinetic energy of wind.
This immediately responds to the main question — what should be the magnitude of pressure drop in wind circulation patterns? Well, of course, the scale will be the partial pressure of atmospheric water vapor pv (approx. 40 mb at 30 deg C, 20 mb at 20 deg C), which shows how much vapor can condense at maximum. To make an appeal to one’s everyday life, if one cares to notice pressure changes as the cyclones come and go, they are always of the order of a few dozens millibars. This coincidence with the characteristic atmospheric water content is not incidental! For hurricanes eye formation leads to an additional drop of air pressure totaling 2.5pv, yielding a maximum drop down to about 900 mb, in excellent agreement with observations for most severe hurricanes. Some details could be found here and references therein.
I am now shamelessly using Jeff’s direct and Kenneth’s indirect encouragement to advertise this work (and I do hope to find more opportunities to speak of that in this exciting place), because, though of course prejudiced as an author, I find this explanation physically elegant and a sort of a triumph of fundamental physics against the alleged unintelligibility, complexity and ultra-sophistication of the climate problem.
I do not aim to fully replicate climate in a model, but I do aim to understand all the basic physical parameters of it in quantitative terms from fundamental physical laws, like the hydrostatic equation allows me to understand the scale height of terrestrial atmosphere. The wind-driving pressure gradient is one of the most important such parameters. I am very much looking forward to find interested people among physicists and climate scientists as well who would be willing to look into the new approach.
Re: Anastassia Makarieva (#221), I haven’t read Smith but your paper made so much sense it really hit home. It’s an excitingly simple physical explanation, as a non-expert I won’t think of hurricanes or tornadoes the same again. The condensation aspect has to factor into the rotational pressure gradient discussed above as well.
Re: Anastassia Makarieva (#221),
Your paper makes passing references to agreement with empirical evidence, but I would suppose the next steps are to test your model against empirical data and other models.
Re: Kenneth Fritsch (#223), In the paper you are referring to the “missing” equation (cf. #216) between pressure and radial distance is derived (Equation 4) from the consideration of the effect condensation has on air pressure. Briefly, as the air spirals in, it rises and some part of water vapor condenses as the rising air cools, producing a pressure drop. The closer to the center, the more vapor condenses, so the pressure drop becomes deeper. This relationship between vertical and radial velocities and condensation magnitude yields Equation 4.
This sums up to an internally coherent physical description of a stationary radially symmetric wind circulation pattern. This picture quantitatively reproduces all the major characteristic magnitudes of hurricanes (and tornadoes) like wind velocity magnitudes and pressure drop from a very limited number of parameters — saturated vapor pressure at a given temperature, hurricane linear size and surface roughness.
To all my knowledge, there is no competing model/theory that would achieve the same, remaining physically coherent and quantitatively realistic. The proposed approach is not a numerical model with numerous empirical parameterizations, but a theoretical description of a previously unknown physical effect derived from a few basic physical laws. (E.g., Navier-Stokes equations, energy conservation law, gravity law, etc., are not models, and so theories derived directly from them are not models either.) My view is that the mechanism of condensation-induced dynamic air flow will ultimately form the basis of the atmospheric circulation theory.
The next steps depend on where one wants to get. The work is in the public domain; every physicist capable of appreciating the idea can further develop and test it. We have solved the equations analytically in a pole approximation; there are several theoretical tasks to be solved towards a more exact description of the hurricane eye (a complex task) and a more exact incorporation of friction. Regarding climate scientists, so far quite a few meteorologists have been skeptical about a possible paradigm shift. So my current steps are still directed to better explain the physical bases of the new approach to climate scientists. For example, some people tend to think that water vapor condensation increases air pressure. This is incorrect.
If hurricanes are condensation-driven, while modern models ignore the effect, I expect there would be many people who would prefer that the hurricane forecasts affecting their well-being be made on the basis of a physically sound theory. That is why I believe exposing our work to the attention of educated public is worthwhile. The growing potential of informed public pressure as illustrated by the Climate Audit is very impressive.
Re: Anastassia Makarieva (#224),
Is there some physical basis for the misunderstanding? It seems obvious that increasing the density of the water would reduce the density of the surrounding gas so I must be missing something.
Re: jeff id (#225), Well, our main point is that condensation reduces air pressure. Pressure and density are different things, as there is also temperature and molecular mass involved, like this:
p = Rho R T/M
where p is pressure (N/m^2), Rho is density (kg/m^3), R is the universal gas constant (8.3 J/mol/K), M is molecular mass (kg/mol).
Suppose you have an air column in hydrostatic equilibrium, then pressure at the surface p_s
p_s = Rho g h, h = RT/Mg is the atmospheric scale height, g is acceleration of gravity.
and mtot = Rho h (kg/m^2) is the total amount of gas per square meter in the column. Now suppose we reduce temperature T of our column, so some water vapor condenses and precipitates. The total amount of gas in the column mtot diminishes and so does the surface pressure p = mtot g. This is our point — less gas, less weight, less pressure.
The opponents’ point about condensation increasing air pressure is related to the release of latent heat that accompanies water vapor condensation. As is well-known, this makes the temperature decrease more slowly for a given amount of energy taken from the gas. So in the atmospheric column where moist air parcels rise and condensation occurs, temperature drops with height more slowly than in the region where dry air would descend. In the result, at the same height air temperature can be higher (if surface temperature is the same) where condensation occurs than where the air is dry. From this (I presume) some people conclude that as far as temperature is higher, the pressure in the moist column is higher as well. However, air pressure approximately changes with height z as p(z) = p_s exp(-z/h), where p_s is surface pressure. So where p_s is lower, air pressure in the lower atmosphere will be lower too irrespective of temperature. This is the feature of gas in hydrostatic equilibrium: its surface pressure ONLY depends on the amount of gas in the column, not on its temperature. As I understand (and some opponents admit), it is this point that is missed in their critique.
Re: Anastassia Makarieva (#226),
.
Anastassia, I selfishly hope you are able to provoke some responses here at CA from the physicists and climate scientists who post here both infrequently and frequently as I enjoy these types of discussions. If you have had any responses to your non-empirical model that you can disclose without violating any confidentiality considerations, that always makes for interesting posts.
.
It will take me awhile to comprehend the full import of your paper, but I must say that you have put much effort into providing an explanation for the relatively well-informed layperson (who may or may not be me). As an aside and a chemist from a long time ago, I believe you are referring to the ideal gas law in your reply to Jeff. Is not the van der Waals equation of state more appropriate for the real gases of the atmosphere?
Re: Kenneth Fritsch (#227), Atmospheric pressure is low enough for air to be well described by the ideal gas law. The only condensable gas is water vapor; to account for the temperature-related changes of its saturated concentration one uses the Clausius-Clapeyron equation.
I would also be very much interested to receive some feedback.
Re: Anastassia Makarieva (#228),
Hi Anastassia.
I am an old chemist too, and my recollection of physical chemistry was that air shouldn’t be considered to be a mixture of ideal gases (if for no other reason than the fact there are water aerosols and various kinds of solid particle aerosols suspended in it). There are too many intermolecular and intramolecular interactions going on such that even at normal atmospheric pressure there are measurably significant deviations from the ideal gas laws. One could, however, consider mixtures of purified, bottled nitrogen 79%, oxygen 20%, and argon 1% in the lab as such a mixture of ideal gases for many (if not most) kinds of analytical purposes.
Re: Kenneth Fritsch (#227),
Hi Ken.
I think that we are of similar age at least with respect to the physical chemistry of air. Yours is a recollection that mirrors my recollection. However, I have not kept up with the chemical literature since my university days; since my professional life was in the medical field (where one becomes an applied bayesian statistician 🙂 if one plans on having any chance of survival).
Re: cdquarles (#229), Hi Charles. Actually representing atmospheric air as an ideal gas is a very good approximation for all large-scale atmospheric problems. The inaccuracy involved is vanishingly small compared to the various uncertainties and unknown parameters of atmospheric circulation. Of course, microphysics studies focusing on surface processes on suspended particles would take a different approach.
Re: TomVonk (#230), I do not think there can be much conceptual discrepancy around the Navier-Stokes equations. In the general case which you are talking about of course one needs to involve an additional equation for viscous dissipation and heat transfer, which is derived from the consideration of energy conservation for a unit volume of fluid. In the case of incompressible fluid I considered, as you agree, energy conservation is not added to the Navier-Stokes equations as an independent equation to fully solve the problem.
In the hurricane case, writing the equations for the approximately horizontal streamline near the surface, alike Bernouilli’s equation, one gets a transparent interpretation of what happens – potential energy of the pressure gradient is converted to the kinetic energy of the mainstream flow (i.e., leads to the increment of tangential, radial and vertical velocities) minus energy losses to friction. This is what I meant by energy conservation comprised within the hydrodynamics equations for that particular case.
Leaving aside the (otherwise important indeed) question about how to represent turbulent friction (actually we devoted an earlier paper to discussing this issue in greater detail), one can get a first-order approximation of the hurricane just considering the hurricane as a dynamic machine yielding kinetic energy of wind that is then transported away and further dissipates somewhere outside the hurricane. Before kinetic energy to dissipate, it must originate from somewhere. So for the hurricane turbulent dissipation is not a major physical issue. I repeat, the main issue is how the air accelerates to hurricane velocities and where the potential energy necessary for that comes from.
Our contribution is Equation 4 here, which shows how pressure drops towards the center as dependent on the amount of water vapor that condenses as the air rises. In simple words what happens – as the surface air moves inward, it also rises which causes water vapor to condense. This diminishes the amount of gas in the column and, hence, reduces pressure in the surface layer. There is this positive feedback — pressure drop in the air that has already traveled a bit inward makes the outward air stream inward too. This continues as long as there is enough water vapor outside the hurricane.
Taking Equation 4 one can attempt to solve the problem in greater details than we did and relax some of our approximations. There is nothing particularly great we did when formally solving the equations. What we consider to be our achievement is pinpointing and quantifying the physical process responsible for the release of potential energy for the hurricane wind formation.
As we show, hurricane can be consistently described as a purely dynamic phenomenon, which arises when the previously accumulated potential energy of water vapor is released in a relatively short burst. It can be meaningfully compared to an avalanche, for example.
Re: Anastassia Makarieva (#231),
I’m not a fan of weather engineering but it seems that there may be a possibility to seriously disturb a hurricanes symmetry at least by seeding the condensation. If the condensation happened before the area of updraft or perhaps at a low level or heavily on one side the buildup of wind speed might be affected.
If it worked, it could do some damage to a tornado possibly as well. Dust missiles anyone?
When Anastassia first started posting on the other thread I clicked on her name and found a page full of references to papers. She’s got review comments and everything on some. The one that caught my attention was: On the validity of representing hurricanes as Carnot heat engine
Which is familiar talk for an aero engineer. I found the potential explanation of the energy driving the hurricane very interesting and almost left a comment on that other thread but decided it was too far o/t at the time. It was very worth reading for me. So much to learn so little time.
A second time the discussion above has been of excellent quality so thanks again.
A hurricane would require a substantial amount of dust – a cargo ship or maybe several, but if it worked it could be a worthwhile effort. Without good symmetry wind speeds drop fast.
Tom Vonk (#230),
You might want to read the manuscript by Andrew Majda and coauthors on hurricanes. They use the reduced system suggested by Kreiss and me that is applicable to tropical dynamics and midlatitude mesoscale (and smaller) dynamics. The reduced system is essentially just a level by level shallow water system with the forcing of the horizontal divergence given by the total heating at the level and is easy to understand. Majda used different prescribed forcings to show how hurricane could be initiated.
Jerry
Have you a link Jerry ?
.
Anastassia Makarieva .
It all depends what you call a major issue .
I consider turbulent viscous dissipation a major issue because the scale and structure of the hurricane is precisely dictated by turbulent dissipation .
In another context (climate model) Jerry Browning had precisely shown what havoc is happening with unphysical viscous dissipation when one makes the hydrostatic assumption .
But I will have a look at the paper you linked .
Re: TomVonk (#235), Well, to clarify in what physical context dissipation is not a major issue — turbulent dissipation arises when there is kinetic energy to dissipate, not vice versa. The hurricane does not arise because of turbulent dissipation like the car does not move because of friction losses, but because it has a motor.
Likewise in the hurricane the huge kinetic power that ultimately dissipates originates due to the radial pressure gradient (with a total pressure drop reaching up to 100 mb over a distance of a few hundred kilometers). In our approach this gradient is not prescribed from observations but is derived from the consideration of potential energy release during water vapor condensation.
But it is impossible to disagree that once the pressure gradient is found, the more accurately one takes turbulent dissipation into account, the more precise picture of the hurricane velocity field one will obtain.
Re: Jeff Id (#232), In my opinion, large-scale weather engineering is simply dangerous. But since there are anyway projected attempts to control hurricanes and tornadoes, a correct understanding of the background physics is absolutely indispensable. If one thinks (incorrectly, from the above viewpoint) that the hurricane arises due to heat input from the ocean, then one could conceive, for example, as Bill Gates recently did, cooling the coastal waters by mixing to scare the hurricanes away. However, since, as we argue, hurricanes are driven by water vapor store previously accumulated in the atmosphere, then creating a cold zone might enhance condensation and lure the hurricane towards the coast, i.e. produce an opposite effect. And yes, as hurricanes and tornadoes can only move until there is water vapor condensation, creating an area of artificially enhanced condensation might indeed change the direction of their movement towards that area.
Re: Anastassia Makarieva (#237),
Since we are sneaking in some statements that could be construed as relating to general policy issues here, I would state that not considering human intervention into climate is short sited thinking. I do think that events such as tornadoes and hurricanes are where we would initially consider intervention. AGW at any level is due to human influence on the climate and any attempts at mitigation would have to be in a form of human intervention.
.
Jim Hansen has stated that we have probably avoided the next ice age (an estimated 10 to 30K years away without the predicted effects of AGW) with the cumulative effects of AGW to date, or if we haven’t, by his calculations, we could avoid the ice age by a controlled release of the proper GHGs from a single facility producing it.
.
When it comes to ice ages, I consider Mother Nature a cold-hearted bitch. When it comes to making a case for the effects of a few degrees C increase in temperature over centennial time periods, notwithstanding Jim Hansen’s pronouncements, I do not think we have made a case for the detrimental effects and/or that the net effect is a not a benefit.
.
So if we understand how a hurricane is produced and furthermore we understand how to mitigate its forces and any side effects of the mitigation, why not intervene? What bothers me about the much of the current science efforts on climate is that it appears aimed at showing the bad effects of any claimed AGW without regard to any benefits of human intervention whether that intervention be intentional or not – except, of course, by the single minded government decreed lowering of GHGs. There is no doubt in my mind that the current thinking on climate is primarily closed-minded and probably no doubt in an attempt to protect the efforts of those with the consensus frame of mind.
Tom,
The reference is
(with Y. Xing, M. Mohammadian) “Moist Multi-Scale Models for the Hurricane Embryo,” J.F.M, submitted, March 2009 (PDF)
and a PDF version can be seen on Andy Majda’s home page
http://www.math.nyu.edu/faculty/majda/
Jerry