Centerfield

A couple weeks ago I ran across a couple articles that gave me some doubt that the current climate changes are all caused by man.

Here's a recent paper about the role of the sun from 1900-2000.

"We estimate that the sun contributed as much as 45–50% of the 1900–2000 global warming, and 25–35% of the 1980–2000 global warming. These results, while confirming that anthropogenic-added climate forcing might have progressively played a dominant role in climate change during the last century, also suggest that the solar impact on climate change during the same period is significantly stronger than what some theoretical models have predicted."

Given the above finding, there's also this to consider: The Sun is More Active Now than Over the Last 8000 Years

IMO, we have overestimated the influence we have on our climate.

We should consider all angles of a problem before taking action.

Posted by: Cervus at April 1, 2006 06:21 PM

I'm very suspicious of the initial graph showing increasing numbers of hurricanes. While global shipping has allowed some tracking of tropical storms and hurricanes for a very long time, the most recent data comes from satellites, which allow tracking of even the smallest tropical storms which are never seen by any ship. Thus, some part of the apparent increase may simply because we are better at spotting them usin today's technology.

The Science article, which appears to be quite thorough, does not an increase in the number of tropical storms and hurricanes, but only in the North Atlantic. Other areas of the globe, it says, do not show a statistically significant increase over the past 35 years.

As for global warming, here's the conclusion from the Science article:

In summary, careful analysis of global hurricane data shows that, against a background of increasing SST, no global trend has yet emerged in the number of tropical storms and hurricanes. Only one region, the North Atlantic, shows a statistically significant increase, which commenced in 1995. However, a simple attribution of the increase in numbers of storms to a warming SST environment is not supported, because of the lack of a comparable correlation in other ocean basins where SST is also increasing.

The Science article itself looked only at the satellite era, "because of the known biases before this period, which means that a comprehensive analysis of longer-period oscillations and trends has not been attempted."

In conclusion, the Science article noted: "attribution of the 30-year trends to global warming would require a longer global data record and, especially, a deeper understanding of the role of hurricanes in the general circulation of the atmosphere and ocean, even in the present climate state."

Posted by: PatHMV at April 1, 2006 06:27 PM

The sky is falling, the sky is falling!

Bob, please note that "not inconsistent with" does NOT mean "we proved something." It means "we didn't disprove anything." It's a polite way of saying that they really don't have any usable results at all to report. It's a negative report. Negative reports also have their uses, but they don't prove or support anything.

Please also note that the Webster/Holland study has been roundly slammed by the top experts in the field as clearly speculating well beyond the data, and for using a very abbreviated data set. I think some of the criticism is overdone, mostly because Webster/Holland et al were very careful to point out the extensive limitations of their data and note that that all they'd shown was that they hadn't disproved anything. Unlike, for example, you.

Check this table here. This data set has the advantage of being four times the length of the Webster/Holland set, and having been consistently gathered to cover a much longer historical range. It's not perfect, but it clearly points out the problem of trying to measure (and draw sweeping conclusions from) interdecadal variations in storm cycles from short sets. The only clear trend in it is an apparent 60-year cycle (30 up, 30 down) that would indicate we're well int the first decade of an upswing...and that the Webster/Holland data, beginning in 1970, begins in the early part of a cycle trough.

Now, back to your first graph, which claims to be measuring ALL hurricanes and tropical storms per year, going back to 1870. I note the lack of any definition of data, or commentary on how the data was gathered, or what global regions, etc. I'm quite aware that from 1970 on satellite data is included for most of the globe, and that for the 1940's on there is airplane surveillance data for most of the North Atlantic and parts of the NorthWest Pacific. So, how much global satellite and airplane data do we have, and for which regions, for the periods 1870-1910, 1911-1940, etc.? See the problem there? It's an absolutely useless graph. It conveys no usable information. It's merely a naked assertion of fact without any underlying data to support it, at least, any we can access for assessment.

And I would like to note for the record that the top hurricane scientists are predicting that we're in for a rough ride over the next decade, based solely on their assessments of the cyclical historical data. So calling any hurricane that hits for the next several years a "proof" of your thesis (as you apparently seem to be setting up to do) is kinda like claiming that the sun rising in the morning is also a proof.

We should consider all angles of a problem before taking action.

Amen to that, Cervus.

Posted by: Tully at April 1, 2006 06:27 PM

Cervus: The sun spot activity does not (at first blush) seem to significantly correlate with the tropical storm activity, ocean temperature or global warming. Although obviously solar intensity does have an effect, the question is what is the magnitude of the effect relative to the other forces at work. Also what are you using to measure its historical variation?

See this link on solar-variability
And this one on climate forcing

Tully: We've already gone several round on those subject points. I have no wish to make it round three. This was not an attempt to “win you over” I know your views, and the evidence level you require. If you want to cross your arms and wait till it's all a forgone conclusion that's your choice.

I have located interesting information and presented it in a nice graphical form. There are other people who read this blog, they will have to make their own decision. As I said, “I've done my part”, my conscious is clear.

Pat: click the wunderground link and then pick one of the early years like 1888. Those are pretty specific tracks even with out a satellite. Also consider the question of “Would a satellite in 1870 really have doubled the number of storms observed?”.

Posted by: Bob J Young at April 1, 2006 07:27 PM

Bob: Yes. It's one thing to have a specific track of a storm which was tracked (and even then, I don't know whether wunderground's underlying data actually supports the apparent level of detailed tracking). It's entirely something different to know that a ship happened to be passing within viewing distance of every single TROPICAL STORM in the North Atlantic during any given year. And there's a definitional issues involved, also. We have a solid definition of tropical storm today. But if you're a clipper ship or an early steamer going from London to New York, can you really tell the difference between a tropical depression and a tropical storm on the horizon? How many ships a month DURING HURRICANE SEASON made journeys in the Gulf and the southern end of the North Atlantic? In 1890 was there that much traffic between North Africa and Cuba or Florida?

And, as I noted, the Science article YOU cited said precisely the same thing... the early hurricane records are too uncertain to make any determinations.

Posted by: PatHMV at April 1, 2006 09:00 PM

And as I pointed out, Bob, the authors of that article you cited are themselves quite clear about the extensive problems with what they're trying to find, the limited available evidence, their own conclusions, and the nature of them. You wanna step out miles past the experts, have at. But when the experts you're citing specifically disclaim what you're claiming they've shown, I gotta go with the experts.

That difference of opinion between "levels of evidence" isn't about "cross[ing] your arms and wait[ing] till it's all a forgone conclusion..." It's the difference between critically examining reality, and finding confirmation of a dogma in every breath of wind. The dogma may or may not be correct, but if it's an article of faith absolutely everything will confirm it to the believer, relevant or not.

Especially when your standard of "proof" is a prediction of what the experts are saying, who have already predicted same completely WITHOUT the causal link you're claiming as either contributory or causal.

Pat, ships at sea encountering typhoons/cyclones back as far as the 1750's could take pretty darn good measurements of air pressure and wind and current, if only on their own very narrow track of travel. But it's also worth noting that most ships of those eras actually finding themselves IN the more severe cyclonic storms would not be reporting any data at all. They'd be on the bottom of the ocean in pieces, their crews dead. So the recorded sea observations are pretty certain to be understated.

Posted by: Tully at April 1, 2006 11:04 PM

Tully,

If the evidence for global warming, or at least the human role in global warming, is so unclear how do you account for the almost universal consensus of the scientific community on this? Why are so many scientists and governments duped? I'd wouldn't expect such transnational acceptance and conformity on such weak data.

Most other Western governments are in firm support of limiting emissions and surely they too are aware of the effects that that would have on their economies -- why do they go forward?

And when the primary opposition in recognizing global warming comes from Republicans -- who frankly really have shown a consistent willingness to twist the science whenever it suits their needs, why should I listen to them?

[Yes, I know that something like 99 senators voted down Kyoto, but Dems are more likely to make economic arguments to pull China and India into the mix, while Repubs are more likely to just be deniers of the phenomenon in question.]

Posted by: Adam at April 2, 2006 12:20 AM

Adam, at the moment, I (and I'm pretty sure Tully) am only questioning Bob's asserted claim of historically unprecedented levels of hurricane activity and his allegation that such his a phenomenon resulting from global warming. That's quite different from denying global warming as a phenomenon.

And I don't take as any kind of support for anything the fact that France and Germany and other Western European governments were in favor of Kyoto. Just to pick one example, the French must surely be aware that their employment guarantees are stifling their economy and actually increasing unemployment, but nevertheless they riot in favor of the status quo anyway. Europe doesn't depend on automobiles and 18 wheelers as much as the U.S. does because it is old and compact and has very tiny streets. Emission controls won't impact their economy quite as much as they might ours; plus, France in particular is pretty nuclear-power-friendly. Also, the cynic in me suspects that sometimes some Europeans are happy to support very vocally some P.C. thing they don't really want, simply because they feel certain that the U.S. will kill it in the end. But I have no proof of that.

Tully, I agree that those ships had good measurements for the weather they are in. my point was just that if the ship measures 10 mph winds and a certain barometric pressure, it's hard to tell whether she's close in to a mild storm or at the edges of a bad one.

By the way, I'm not real up on my 18th and 19th century trade patterns. Would they try to avoid shipping during hurricane season? That would make a certain amount of sense, at least for nonperishables.

Posted by: PatHMV at April 2, 2006 12:42 AM

Pat,

I know that the topic was hurricanes, but I've seen Tully and Bob's disputes before, so it lead to me ask.

I'm not sure whether Tully disputes global warming or humanity's role in it.

I share your skepticism towards continental Europe, but what about Japan and England?
And what of our own scientific community?

Posted by: Adam at April 2, 2006 12:54 AM

I meant the UK, not England. I know it can be offensive to refer to England when you mean the UK. But I'm an American, so be happy I even know where the UK is! :)

Posted by: Adam at April 2, 2006 12:56 AM

Tully and Pat,

One final note to be clear. I recognize that the particular claim about global warming driving increased hurricanes seems unsubstantiated. It's just that I've noticed a broader trend in Tully's thinking about global warming in general and was asking him about that.

Posted by: Adam at April 2, 2006 01:03 AM

Adam, I don't pretend to be an expert in the area. I did find this site, which appears to debunk some criticisms made by disbelievers in global warming.

Part of the resistance, I think, comes from the unfortunate fact that many prominent scientists, in the past 20 or 30 years, have become more politically active, signing petitions and taking out full page ads on topics which are primarily political and not in their realm of expertise: war, the death penalty, etc. This costs all scientists some credibility when they do try to speak out on a topic properly within their provenance.

We also regularly see reversals and undoings of previously announced "scientific" findings. Fat is bad for you. No, fat is ok, but not transfat. Eat margerine. No, eat butter. Cellphones cause tumors. No, they don't. While correcting its earlier errors is a fundamental strength of science, public overreactions in support of the "latest" findings cost the profession credibility when those latest findings are proven wrong.

Finally, it wasn't but 30 years ago that there was more concern that carbon dioxide would lead to global cooling rather than global warming. Global warming predicters respond that they have better models and are now able to make deinitive predictions of major climatological changes. But who is to say they are right this time?

Posted by: PatHMV at April 2, 2006 01:58 AM

Back to the original graph I'm disturbed by an alarming uptick in the number of storms per year. My only consolation is that the 1890's have long since passed.

Posted by: c3 at April 2, 2006 10:19 AM

Technology has caused a storm inflation. Roughly 1/3 of all storms classified in the last ten years would not have been such in the time period before that. Microwave technology on the latest generation of sattelites have shown what would have been waves 15-20 years ago, no get classified as storms. Not too mention that a lot of the late season and early season hybrid storms were not classified in the past either. Zeta never would have been classified a tropical system ten years ago, for example. Bret, Jose, Lee, Gamma and Delta are all system that are highly questionable if they would have been classified twenty years ago. That is six systems right there.

The first graph also appears to be deceptive in the way it handles the 1925-late 1930's peak. I would venture to say, if we had todays technology back then, those years could easily, especially 1933, could have met or exceeded that last few.

The pattern was just setup for a year like last year. We have better technology to notice these things.

Could global warming be affecting things. It is possible. But the increase we are seeing now are more comparable to the frequency we saw in the earlier part of the 20th century. I think there is an absolute technological inflation in storm numbers, however, that seems to invalidate the storm freqency curve that was posted here.

Posted by: Jim M at April 2, 2006 01:10 PM

I would also like to throw in a little more relevant number then just hurricanes. Here is the chart for landfalling major hurricanes by decade

period Number of major landfalling hurricanes
Cat. 3 Cat. 4 Cat.5 Total

1891 - 1900 5 1 6
1901 - 1910 5 1 6
1911 - 1920 3 2 5
1921 - 1930 3 2 5
1931 - 1940 6 1 1 8
1941 - 1950 9 1 10
1951 - 1960 5 3 8
1961 - 1970 4 1 1 6
1971 - 1980 4 4
1981 - 1990 3 1 4
1991 - 2000 3 1 4
2001 - 2005 4 3 7

from: http://www.euronet.nl/users/e_wesker/atlhur.html

The last two seasons has really kicked up the number for 2001-2005. However, if we get a couple of slower years, these numbers, while probably ending near the high part of the scale, will probably be right around the 1930-1960 area. Note that 1970-2000 was a very low level compared to earlier. It is too early to say tha we are out of a norm. If we have another 7 major landfalling storms over the next five years for a total of 14, I would start to get some concern. There is some technologoical creep in storms, too. The post storm technology for reviewing systems has changed a few numbers. Technically, this chart is in error because NHC has gone back with new technology and reclassified Andrew as a landfalling Cat 5. This chart only has it as a 4. So even post storm reviews have some technology creep now.

Posted by: Jim M at April 2, 2006 01:31 PM

Ok, the chart did not post well. Where there are numbers missing, the last number is total. All decades have Cat 3's. 1971-1980 have no Cat 4 or 5 systems. Only 1931-1940 and 1961-1970 have Cat 5's. That should make a little more sense. 1941-1950 is the only decade with double digit major storms with 10. Everything else is single digits.

Posted by: Jim M at April 2, 2006 01:37 PM

Jim M: The euronet site is interesting.

You seem smart enough to already have considered these points but I'll make them anyway.

All data sets have errors. You can not make a real world measurement without making an error. The real question is what is the magnitude of that error.

Keep in mind that “technological inflation” of the data can only be reasonable since WWII. When the weather in the Atlantic became of great military value and aircraft started crossing it on a regular basis. The technical advances that started after WWII may add a couple of more storms a year, but they don't really affect the the current trend (which started about 1985).

.Dr. Klotzbach's team (formerly led by Dr. Gray) defines the average number of storms per season (1950 to 2000) as 9.6 tropical storms.

From the data set I've graphed I get the following averages:
1951 to 2000 = 9.98
1901 to 1950 = 7.56
1871 to 1900 = 7.07

So even if you normalize the pre 1950 data by adding 3 more storms per year for “technological inflation” the overall trend shown by the graph is not significantly affected.

Posted by: Bob J Young at April 2, 2006 02:47 PM

If the evidence for global warming, or at least the human role in global warming, is so unclear how do you account for the almost universal consensus of the scientific community on this?

Even accepting that were true, which I don't, I don't need to account for it. It's an argumentum ad populum. Scientists are not immune from fantacism, as this report indicates, and as anyone familiar with the history of science is aware. (Geocentrism, anyone?) You can find a reasonably complete summation of the assorted arguments at Wikipedia.

I'm not sure whether Tully disputes global warming or humanity's role in it.

Without "global warming" Earth would be a cold airless rock, and we would't be here. So what's to dispute? Now, when it comes to claims of human agency as the predictable and controllable major factor in ongoing climactic change (particularly as to scope and scale) leading to disasters of Biblical proportions and therefore requiring beggaring the entire world economy and intentionally implementing measures that are likely to kill millions, even billions, of people, and would in practice demand world government of a particularly oppressive nature to actually enforce, there's a lot to dispute.

As the track record of purported "volunatry" compliers with Kyoto shows, it's one thing to say you'll do something, and another entirely to actually do it when real sacrifice is required. Cost/benefit analyses of Kyoto have problems, but all are pretty clear that the costs outweight the benefits. Even Kyoto supporters agree that it's mostly a cosmetic agreement whose only real benefit is to pyschologically condition nations to take even bigger steps, while transferring wealth from industrialized nations to the third world.

As you'll notice when reading the Wiki summation, "consensus" is not exactly the term to use in regard to human-agency global warming theory, and we're not talking about definitive science at this point, but modelling extrapolations based on short-run data sets using multiple limited and non-congruous models with limited and inconsistent track records. Passed through political filters to boot. I have a lot of experience with modelling, and am painfully aware of the limitations of it. My required level of evidence is indeed high--as it should be when you start arguing for impoverishing and killing off and/or impoverishing billions of human beings based on incomplete and tenuous evidence filtered through limited models.

NOT disputed--atmospheric CO2 levels have risen about 18-20% in the last 50 years. Some of that is certainly the result of human agency, both through the use of fossil fuel and deforestation. Not disputed--we're in a warming trend. VERY disputed--how much of current short-run temperature trends are due to human agency, if any, and how much is natural variation that is unavoidable.

For some long-term historical (and pre-historical!) context on sea levels, global temperatures, ongoing climactic change, CO2 levels, etc., see here. Also note that evidence exists for non-terrestrial factors currently at work in current short-run warming trends, which supports cyclical climate theory and natural variability. Heck, for all we know human agency is what's keeping us from cycling into another period of glaciation. As Pat pointed out, 35 years ago scientists were freaking out over the possibility of global cooling. For all we know, our production of CO2 is what's holding back the next glaciation. The real news is that, with or without us, global climate is not stable and never has been, and short of achieving global weather control, never will be this side of Sol falling out of hydrostatic balance and becoming a red giant, which will incinerate the planet in a few billion years. No planet, no weather.

Posted by: Tully at April 2, 2006 02:52 PM

And once again, Bob, the people you cite as support don't make the claims you do, any of them, and are making their seasonal predictions based on known and existing storm cycles. Could you please link directly to the base data set for your graph?

Per NOAA:

Landfall US hurricanes 1851-1900 [ALL (major)]: 97 (27)

Landfall US hurricanes 1901-1950 [ALL (major)]: 95 (34)

Landfall US hurricanes 1951-2000 [ALL (major)]: 72 (28)

Hmmm, seems to be a disagreement over that data of yours. And looked at by decades rather than half-centuries, there seems to be some consistent cyclical variability as well.

Posted by: Tully at April 2, 2006 03:14 PM

Your comparing average tropical storm data per year, that may never make landfall, to the sum of landfalling hurricanes, of course the numbers don't match.

Posted by: Bob J Young at April 2, 2006 03:28 PM

Don't be dense, Bob. Tully is pointing out that the only measure we can be fairly certain is not subject to miscalculation due to changes in observation technology, hurricanes which make landfall, shows a conclusion which, if anything, is the opposite of your claim. From the NOAA page cited by Tully:

Table 6, which lists hurricanes by decades since 1851, shows that during the forty year period 1961 2000 both the number and intensity of landfalling U.S. hurricanes decreased sharply! Based on 1901 1960 statistics, the expected number of hurricanes and major hurricanes during the period 1961 2000 was 75 and 28, respectively. But, in fact, only 55 (or 74%) of the expected number of hurricanes struck the U.S. with only 20 major hurricanes or 71% of that expected number. Even the very active late 1990s showed below average landfall frequencies. It could be noted that of the most recent four decades, only the 70's and 80's were significantly below normal in terms of overall tropical cyclone activity.

So if your major concern is accurately predicting that people living in coastal areas need to be even more fearful than average of hurricanes, the trend in fact shows the opposite.

Adding 3 storms a year to your data up through 1960 (an arbitrary cut-off date, as Jim pointed out, we're spotting even more of them now than we did in the 60s) does in fact change the trend line significantly. Go back to Excel and redo your chart. Adding a straight linear trend line (as opposed to the 6 point polynomial line it appears you used) to the ajusted data gives a very, very gentle slope from 9 in 1871 to about 13 in 2005. Adding the same linear trend line to the original data shows a much steeper curve from 5 up to 13. Adding in a 10-year moving average trend line also shows that a sustained period of above-average storm numbers is hardly unprecedented.

Posted by: PatHMV at April 2, 2006 05:30 PM

Actually, Bob, I'm comparing them to an asserted data set that I haven't yet seen. Known and verified first-hand observations of over a century and a half versus an unknown quantity and quality of uncertain data offered only by naked assertion. 150 years of landfall observations are a quite reasonable data proxy for total storm activity on sheer base probability and consistency. Note that the base data I offer has not been massaged or approximated or adjusted in any way, but is the actual sum total of first-hand observations encompassing the entire time period.

All data sets have errors. You can not make a real world measurement without making an error. The real question is what is the magnitude of that error.

Couldn't agree more. As I said in an earlier thread, I'm a professional skeptic. Jim's notation of both technical and definitional inflation of the data is relevant. Adjusting data by fiat assumption produces relevant adjustments only by chance. To even begin meaningful data testing for adjustment, we need both data sets if only to check for significant variations. Of course, we can't judge that at all without access to the data sets.

Could you please link directly to the base data sets for your graph?

Posted by: Tully at April 2, 2006 05:39 PM

Pat: Before you start calling people names, go back and read the posts again.
I'm talking tropical storms not landfalling hurricanes.

By removing everything but landfalling hurricanes, all you really have proven is that the path of hurricanes change. Sometimes they hit land, an sometimes they don't. So what?
Its nice for predicting the weather, but useless for measuring the available heat energy in the ocean.

Tully: The data comes for the two links in the original post.

Posted by: Bob J Young at April 2, 2006 06:21 PM

Hurricane alley's site is down here it is in csv format.
1871 to 2005
6,5,5,7,4,3,8,10,8,9,6,3,4,3,8,10,17,9,9,1,11,9,12,
6,6,6,5,9,6,7,10,5,9,5,5,11,4,8,10,4,4,6,4,1,5,14,3,
5,3,4,6,4,7,8,2,11,7,6,3,2,9,11,21,11,6,16,9,8,5,8,
6,10,10,11,11,6,9,9,13,13,10,7,14,11,12,8,8,10,11,7,
11,5,9,12,6,11,8,8,18,10,13,7,8,11,9,10,6,12,9,11,12,
6,4,13,11,6,7,12,11,14,8,7,8,7,19,13,8,14,12,15,17,14,
21,16,26

weather underground and noaa have the same data set but the format isn't easy to import into a spread sheet.

Posted by: Bob J Young at April 2, 2006 06:49 PM

And we have shown pretty conclusively how uncertain the data you used is, Bob. Moreover, as I noted, if your biggest concern is the impact of hurricanes on people, then the hurricanes which actually make landfall is clearly the most important statistic, not the number of tropical storms out in the ocean that never affect anybody.

Posted by: PatHMV at April 2, 2006 07:36 PM

Absolutely not. You injected “impact on people” not me.
I'm talking energy and its impact on climate.

Posted by: Bob J Young at April 2, 2006 08:07 PM

What your purposing is that by judging the weather along the east coast you can tell what is going on for the entire Atlantic and Gulf of Mexico. You dismiss the ships at seas as having inadequate coverage then you embrace an even smaller coverage area. That's silly. Your just rejecting data till you get the answer you want.

Posted by: Bob J Young at April 2, 2006 08:22 PM

What your purposing is that by judging the weather along the east coast you can tell what is going on for the entire Atlantic and Gulf of Mexico.

Absolutely not. The known data include Gulf Coast strikes, and landfalls in the Gulf outside the US are also on record. You're saying, essentially, that nothing was going on in the Atlantic and Gulf for those periods because there were no planes or satellites to observe it--whcih is kinda the point about the reliability of that pre-aviation/satellite data. Landfall data, OTOH, is completely reliable.

Your just rejecting data till you get the answer you want.

Aren't you doing exactly that, Bob? You refuse to acknowledge that pre-aviation/satellite data for non-landfall storms is inherently limited and has severe problems, and insist on using it as-is to lower your projective baselines. But when landfall data shows a consistent 60-year cycle, one used by the experts to predict today that we're hitting a peak period for storms, you reject it as unusable and unreliable, even though it's 100% reliable and verifiable and is the very basis for those forecasts. You're rejecting the applicability of the reliable data AND the possibility of error in the non-reliable data to reach your baselines.

Then you want to use those self-same forecasts based on the historical trends of landfalls as a basis for your claims, when the experts say it doesn't show what you're claiming and that what you're claiming was NO part of their forecast models, one way or another. And you want to use partial-cycle data from a limited test model as a positive affirmation of a new long-term theoretical trendline, when the people who wrote the study are quite careful to say they haven't shown any such thing, merely failed to exclude it as a possibility.

Oy, veh.

Posted by: Tully at April 2, 2006 09:02 PM

I'm sorry, Bob, but I must have misunderstood when you began your post by saying:

"I didn't warn the public about what was going to happen during the 2005 hurricane season. I had the data set, but I just didn't know how to embed the graph in a web site. Not that anyone would have paid any attention, but it would have salved my conscious."

Or maybe it was the note on which you finished your post:

Note to the general public: I've done my part. Please take appropriate action to prevent your butt from being photographed by the MSM, as you scream for rescue atop of your flooded home.

I guess the concern you expressed was for warning everybody about the general perils of global warming, about which they should take immediate action for their own personal safety. My bad.

Posted by: PatHMV at April 2, 2006 09:19 PM

Tully: That was so convoluted I couldn't make any sense of it. But I'm pretty sure (outside of the direct quotes) that I never said any of it. To the best of my knowledge the data set I'm using does include landfall data. Use the weather underground link to look at the tracks.

Refusing to include the landfall data would be just as silly as refusing to accept the data from ships.

Posted by: Bob J Young at April 2, 2006 09:26 PM

Pat: Ok! You got me on that one.
I shall try and rephrase.
“Absolutely not. You decided to parse out a subset of data that only includes “impact on people” not me. I'm talking energy and its impact on climate.”

Posted by: Bob J Young at April 2, 2006 09:47 PM

All the cyclical data I've seen cited says another busy year - I don't have reason to discount that. S. Louisiana and Mississippi are not ready for another big one, so I'm hoping this consensus is wrong.

And anyone doubting global warming at this point has their head in the sand, or is expecting the Rapture to take them away before it matters. This needs to be discussed from a "reality based" perspective.

Posted by: JP at April 3, 2006 12:44 AM

A couple of comments here as I was studying physical climatology in the late 70's(and still have the $75.00 Mikhail Budyko classic "Climate and Life" to show for it).

While global cooling was a possible concern so was global warming was equally so at the time. In fact the two were about equal. A lot had to be considered. The heat island effect of large metropolitan centers, aerosol deposition into the atmosphere from natural and anthropogenic sources, the effects on albedo - or surface reflectivity - from deforestation and desertification and on and on and on. Esoteric things like the effect of releasing krypton gas due to the production of nuclear energy and the resultant changes in atmospheric buoyancy and air circulation and the effects on cloud formation and precipitation. I think I still ahve that article somehwere... Literally tons of factors.

At that time we just didn't have the computing power to really get a handle on what was going on with atmospheric heat balance, the global circulation models were flawed and primitive, and historical records that could be used to check the models were kind of spotty to say the least. So I would say that given what was a rather primitive state of the art at the time, any declaration in the 70's has to be treated with the considerable uncertainty that even scientists then admitted to.

Saying that, just peeking into my old copy of "Ecoscience"(Holdren and Erlich) from about 30 odd years ago I see this on p687: "If there is any consensus on the overall direction... it is a shaky one to the effect that the dominant influence is or soon will be in the direction of global warming"...It goes further to say, and correctly I think, "That changes in patterns of circulation and the accompanying changes in precipitation and extremes of temperature are more important than changes in mean global temperature cannot be emphasized enough." How true.

In one of the classes I took from Dr. Orman Granger*, we discussed a lot about changes in meridional (north south) circulation and how with warming due to antrhopogenic additions of grenhouse gasses. In a nutshell that's the jet streams' having a more vertical component and thereby causing large changes in climate patterns with just minor shifts. Think of an oscilloscope with a sine wave. Now using the horizontal control just nudge that thing over a c-hair and you'll see what I mean. Little movement, big changes. Something normally mundane becomes more violent, or vice versa.
Another effect, not discussed much but one to be aware of is how greenhouse gasses can affect the vertical temperature profile. These are minor changes but they are global in scope so the cumulative impact on global circulation is quite significant.

One of the reasons why we started taking human impacts so seriously back then was that the magnitude of human impact was no longer in the minor leagues but in the big leagues. An example from the 70's;
Natural emissions or formations of hydrogen sulfides were estimated to be 130 - 200 million MT. Anthropogenic emmisions of sulfur dioxide at the time in the same range, 130-200 million MT.
And as has been noted, the release of CO2 into that atmosphere has been largely anthropogenic, due to fossil fuel burning, deforestation, slash and burn agriculture in 3rd world countries, etc.
That's the problem. We're outdoing Ma Nature.

Hence the consensus these days that global warming is here, that human activities do play a significant role, that human activites can initiate some changes and that when there are natural changes due to events such as volcanoes, that human activities keep the engine running so to speak- (recent article in Science but can't find it now)

Philisophically you can ask yourself this: "Do you like being your own guinea pig? And as your own guinea pig, in this case alteration of the climate due to man-made activities, are you prepared to aaccept the consequences or are you just going to scream out as if in some cheap sci fi film "But we didn't know!!! We didn't know!!!!!" ???

*At the time I knew Orman we were in the middle of a major drought. Using a rather clever econometric model(and the old CDC6400 in the bowels of Evans Hall) linking rainfall patterns in the Mediterranean to those in California he accurately predicted the break in the drought and his precipitation forecast was within half an inch of the rainfall totals for that year.

Posted by: Marcus at April 3, 2006 04:17 AM

The actual points show something, but the line graph portion is obvious based on assumptions pulled from someone's rectal cavity.

Here's a question: How come the data point for, say 1934 is 21 or 22, but the line graph goes through ___8____? But then in 2005 the data point is 26 and the line graph goes through 24?

Answer? Because the 1934 line graph area gets averaged out by what happened in subsequent years. Here's the thing: we really don't know the extent to which current storm levels are a cyclical peak versus being an enduring trend. We can only guess. And this graph's prediction that the last few years are the beginning of an endurring upward trend is just that, a guess.

You could have made the same sort of hysterical upward trend guess if you had "modeled things in 1935 or 1936.

So let's kep an eye on this graph, and then we can make a new guess about where we are headed after a few more years data. The line graph suggests we are headed for numbers like 29, 33, 35, 29, 31, 25, 33, etc. But if instead the nbumbers for the next 5 years are 17, 21, 23, 16, 15, 18, then the line graph will curve back down.

Bob, I can't believe how willfully opaque you are about all the points Tully has, made, especially calling his last argument "convoluted" when it was really a very clear and precise demolishing of everything you've marketed here.

I especially don't see how you call landfall data unreliable when it's given as half-century tallies. And I'm still waiting for you to address the point Tully has brought up several times now, that you are actually claiming that your preferred data shows something that even its collector/authors do not claim to be true. If you can't address this point, I can't consider you an honest broker.

Marcus, as it relates to being a guinea pig, I have a question. In what sense would, say, enacting and following the Kyoto protocol NOT make me and the rest of us guinea pigs, albeit for a different hypothesis?

Posted by: bk at April 3, 2006 10:10 AM

The trend line is a sixth order polynomial fit of the data performed by Microsoft Excel.

Posted by: Bob J Young at April 3, 2006 10:29 AM

Here is an article on how it works.

Posted by: Bob J Young at April 3, 2006 10:43 AM

As for a link between climate change and hurricanes here is an article that can explain it better than I can.

Posted by: Bob J Young at April 3, 2006 11:03 AM

From the RealClimate article you cited, Bob:

Hurricane forecast models (the same ones that were used to predict Katrina's path) indicate a tendency for more intense (but not overall more frequent) hurricanes when they are run for climate change scenarios (Fig. 1).

Once again, an article you cite specifically denies the correlation you tried to make, which was an increase in the number of tropical storms and hurricanes. And again, here:

But what about the past? What do the observations of the last century actually show? Some past studies (e.g. Goldenberg et al, 2001) assert that there is no evidence of any long-term increase in statistical measures of tropical Atlantic hurricane activity, despite the ongoing global warming. These studies, however, have focused on the frequency of all tropical storms and hurricanes (lumping the weak ones in with the strong ones) rather than a measure of changes in the intensity of the storms.

As for your chart, if you do a 2nd order polynomial fitting, the graph looks nowhere near as scary. And if you do a 10-year moving average, it looks less alarming still. And of course there are all the problems with that historical data already noted.

Seriously, Bob, every article that YOU have cited denies a significant statistical trend upwards in the number of tropical storms and hurricanes. Is there ANY scientific article that supports your claim?

Posted by: PatHMV at April 3, 2006 12:18 PM

I have to support the argument that pre-space age counts of tropical cyclones were subject observation capability bias. However, even if global warming has no effect on the number of storms, which scientists on RealClimate.org have agreed with, it is possible that global warming can make the ones that form more intense than they would otherwise be. A quantitative measure of the strengthening of storms is the Power Dissipation Index (PDI).

I recall reading on RealClimate.org some time ago that the shifting of tropical cyclones to higher intensity has been a phenomenon across multiple hurrican basins. It would take some time for me to find that article.

Posted by: Scott Smith at April 3, 2006 12:21 PM

We also regularly see reversals and undoings of previously announced "scientific" findings. Fat is bad for you. No, fat is ok, but not transfat. Eat margerine. No, eat butter.

I'm not aware of dieticians having ever reached the level of consensus that climate scientists have reached on global warming.

Posted by: Scott Smith at April 3, 2006 12:28 PM

Big effing whoop that it's a sixth order polynomial model. I know how it works. It works like a a sixth order polynomial guess pulled out of one's rectum. Is there any inherent reason to believe that climate data can be predictably modeled by a 6th order polynomial? Or a 5th, or a 4th or a 3rd, or a 7th, or an 11th?

I dare you to do 6th order polynomial model of the data up until 1935, and tweak it for the best bfit of THAT data, and then see how well it predicts the next 10 or 20 years. You can jerk round with polynomials until you get an equation that seems to mimic your data. But all you did was a clever trick. It's bass-ackwards. There's no real reason to think that a 6th order polynomial models storm data especially well. It's simply the case that if you ALREADY have variate up/down data, you can FIND a polynomial equation that sort of matches it. You just need to pick a polynomial that has an order as high as the number of dips and turns, and then jerk around with it.

And you still haven't addressed the point that you are making claims about this data that even your authors contradict. At my count, that's at least 3 times you've spit the bit on this. If you are going to change the minds of people, you can't have us thinking you're not an honest broker. And that's precisely what I think right now.

Posted by: bk at April 3, 2006 12:58 PM

As I have learned at my cost, never argue math and statistics with Brian!

Posted by: PatHMV at April 3, 2006 01:21 PM

Scott, I wasn't trying to equate the two (cardiologists and meteorologists), but simply to point out that the general public hears a lot of "science" day in and day out, some of it later thoroughly debunked. That hurts the credibility of the entire industry. It's mostly the media's fault for not being more careful in what they choose to make HUGE stories, but it's a fact all the same.

Posted by: PatHMV at April 3, 2006 01:23 PM

Brian,

You can accomplish the same thing by saying that a sixth-order polynomial isn't particularly parsimonious.

Posted by: Scott Smith at April 3, 2006 01:24 PM

Brian--rusty as I may be, I seem to recall there's a real good reason not to trust higher-order polynomial smoothing when plotting curves to limited and highly variable data-point sets, especially cyclical ones. Runge?

Bob, your graph suggests a rapid two- to four-fold or more increase in total number of storms of ALL sizes, purportedly resulting from an atmospheric increase in CO2 of approx. 33% from baseline--WAY out of line with the models you're basing the causal link claim on. But we don't even need to go into the causal problems there, because the graph's an observational stochastic analysis and thus needs no theory, only examination of the data and projections for technique and reliability.

Now, assuming a steady-state cycle and knowing the technical problems with the historical non-landfall storm data, we can confidently predict an upward-sloping curve simply from improved data reception over time. The only reliable data we have for retroactive comparisons is the landfall-storm data, going back 155 years--and it confirms the cycle, but NOT an upward-sloping trendline across cycles.

None of which either proves or disproves the theories offered by the climatologists, who are still arguing with each other about both theory and specifics. We're at or near peak of the 60 year cycle, so we're more likely to have more and nastier storms, and NOAA has so predicted same, based on that. We won't have good data about that peak until we're heading back into the trough. And even then they'll be arguing over variability versus human agency.

Posted by: Tully at April 3, 2006 02:27 PM

bk: I haven't addressed it because I don't think it matters. The data can speak for itself and people can draw their own conclusions.

The only people I have linked to that really has a problem with linking climate change and hurricanes are Dr. Gray and Dr. Philip J. Klotzbach from the Colorado State University. An the only think I used from them, is the average tropical storms per year. How does that impinge on my honesty? Especially if I supplied the link in the first place.

Everyone else uses weasel words like, "is not inconsistent with", to insulate themselves from criticism. If I was trying to get something published in a peer reviewed journal I would use the same phrase.

As for the polynomial, all it's doing is showing a trend line. If you do a running average you get about the same thing. I tried several approaches and this seems to fit the data the best. If you have a better curve fit or analysis then post it or supply a link to it.

Posted by: Bob J Young at April 3, 2006 02:28 PM

tully:

"We won't have good data about that peak until we're heading back into the trough. And even then they'll be arguing over variability versus human agency. "

Which brings us to the heart of the problem.
If you hear a train approaching do you step off the tracks now, or wait till you can see the whites of the engineer's eyes? Especially when it's not obvious if you can actually move at all, when it comes time to jump.

My personality does not lend itself to playing chicken. Particularly when there are other good reasons to reduce CO2 levels (high gas prices, dependence on foreign oil, air pollution, baby caribou.....)

Posted by: Bob J Young at April 3, 2006 02:49 PM

Translation: Your conviction is an emotional one, so it's justified to rationalize the evidence to say what you want it to.

Gotcha.

Posted by: Tully at April 3, 2006 03:41 PM

Funny!!
Especially since I was trying to restrain myself from saying the same thing about your conviction.

It's a reasonably good data set, with an incredibly ordinary trend line, that looks a lot like the CO2 curve.

Your response of shooting the messenger seems like an extremely blatant case of "protesting too much".

Posted by: Bob J Young at April 3, 2006 03:52 PM

A better analogy might be: Your Porche gets stuck on the train tracks. You hear a horn which might be the train, or might just be an 18-wheeler passing safely by on the adjacent interstate. Do you: (a) blow up your car (the economy) immediately, because the train must surely be coming and you don't want to cause a derailment; or (b) wait a little while longer until you can tell for sure whether it's a train or not?

Like all analogies, this one's not perfect. But it does at least acknowledge that there is a MAJOR cost to acting on the worst-case scenarios on the basis of uncertain evidence. There's plenty of ground to debate just how certain the evidence is. But that's very different from saying, "sure, we don't know, but if our fears are right (Malthus), then it'll be really bad, so let's not examine whether we're right or not."

Posted by: PatHMV at April 3, 2006 03:54 PM

I have no idea what you guys are talking about. (-:. The science is beyond me. So I won't discuss that. But what I will say is that Tully has a good point about the tendentiousness of some scientific analysis. In the early 70s, we had the Club of Rome predictiong imminent disaster from overpopulation (which obviously has not happened). And you can, I'm sure, find a lot of examples of that kind of thinking. And, there is little doubt that scientists are as subject to political bias as anyone else. Plus, sometimes they can just be wrong.

Still, I don't think the fact that scientists may have a political agenda necessarily makes their scientific conclusions incorrect. And I am struck by the consensus (albeit obviously not unanimity) of the scientific community on global warming. It's hard for me not to credit at least some of their conclusions; I just don't believe every scientist is skewing his or her data to support a political position.

Posted by: Marc at April 3, 2006 04:08 PM

Pat: Which is exactly why I am in favor of solutions that will solve multiple problems at once.

Victory gardens, wind power, Nuclear power, electric light rail for commuting, reorienting city design to reduced automobile use, encouraging local produce consumption, bio-fuels, a moratorium on coastal building, relocating people off the coast lines as their homes are destroyed.......

There is much that can be done without blowing up your Porche, and at the same time solves other pressing problems.

Posted by: Bob J Young at April 3, 2006 04:14 PM

Bob, then you should talk about THAT rather than using BS statistics not supported by the very scientists upon whom you purport to rely. Rather than try to club people over the head with scare tactics about hurricanes to support a dogmatic point (global warming), just make the case for "let's adopt windpower", at a very small cost. Or describe who will bear the costs of prohibiting further development in coastal communities, and why they should. If the costs are really not that great, and they provide other benefits anyway, what do you need the bogeyman of hurricanes and global warming for?

Posted by: PatHMV at April 3, 2006 04:54 PM

I don't need a bogeyman. The data existed and needed to be presented. I'm not going to hide data just because it's politically unpopular.
I was just as surprised as everyone else first time I saw the graph.

Posted by: Bob J Young at April 3, 2006 05:06 PM

It's a reasonable good data set, with an incredibly ordinary trend line, that looks a lot like the CO2 curve.

And that's a blatant lie, Bob. You're taking the least reliable data set, dismissing the most reliable data set, and eyeballing your preferred trend line to fit (the one that gives you the "answer" you want). And even then it doesn't look like the CO2 line, other than being an upward-trending line. Which by the nature of the flaws in the data, it must be. All before we even come close to that causation/correlation problem. Then you say you're letting the data "speak for itself."

Bull. You wrote all the lines and tried to muzzle the reliable data that disagrees. "Look, a train's coming!" We look at the most reliable data (that you're doing your best to dismiss in favor of your less-reliable data) and ask "What train?" and your response is "If we don't act NOW, we're DOOMED!" We ask "Doomed by what?" and you say "Look, a train!" And we look at the best data and ask "What train?" And around we go.

Marc, it's not about global warming, which is a seperate subject. To get that far you have to clear several high hurdles on proofs. Bob hasn't cleared the first hurdle yet, and the folks in the biz say he's trying to jump the wrong hurdle and has several more to go to reach his allegation with any justification.

He's saying that his flawed and subjective data are more relevant than the certain objective data of known storm activity. He admitted he fitted the trend line to the bad data by eye, just diddling with fits until it looked right to him. Science that ain't--and he knows it. Or should.

BTW, Bob, higher-order polynomials increase function-fitting errors. Runge's phenomonon. The higher the order, the more variable the data, the greater the error.

Posted by: Tully at April 3, 2006 05:11 PM

The annoying thing is that I agree we should be doing some of that stuff anyway, or at least figuring out if it can be efficiently and economically implemented.

Posted by: Tully at April 3, 2006 05:12 PM

Bob, one more time, presented Garrett Morris style for the hard of hearing. THE DATA YOU PRESENTED IS UNRELIABLE. IT DOES NOT SUPPORT THE CLAIM YOU MAKE. THE SCIENTISTS YOU CITE SPECIFICALLY SAY THAT THE CLAIM YOU MAKE IS NOT SUPPORTABLE BASED ON THE AVAILABLE EVIDENCE.

Posted by: PatHMV at April 3, 2006 05:26 PM

Pat: I do not concur with your assessment.
I've seen nothing so far that supports your statements.
If you have meaningful evidence supply it.
Otherwise, stop yelling.

Posted by: Bob J Young at April 3, 2006 07:52 PM

By the way, you sited the realclimate article:

“Some past studies (e.g. Goldenberg et al, 2001) assert that there is no evidence of any long-term increase in statistical measures of tropical Atlantic hurricane activity”

There wasn't any evidence in 2001 when Goldenberg et al did their study. It's now 2006 and more data is available.

Posted by: Bob J Young at April 3, 2006 08:13 PM

Amazingly enough, that's not right either, Bob. The Goldberg article (which was cited by the RealClimate post YOU first linked to) says that no prediction can be made because the pre-1944 data is not reliable, and it really wasn't until the 1960s that they got really reliable. From the Goldberg article (sorry, no link, because I reached it through my university account):

The Atlantic tropical cyclone record, which (except for U.S. landfall data) is not considered reliable before 1944 (33), shows less than one complete cycle of the multidecadal signal.

And from footnote 33:

Although these data are available since 1851, only the data for the years since 1944, when routine aircraft reconnaissance of Atlantic tropical cyclones began, are considered very reliable. The greatest reliability starts around the mid-1960s when operational satellite detection of Atlantic tropical cyclones began (50). Before satellite coverage, a portion of the lifetimes of many systems had probably been missed.

While the Goldberg article does predict above average hurricane seasons in the near future, it very clearly concludes that it is not possible to say whether this can be laid at the feet of global warming:

The possibility exists that the unprecedented activity since 1995 is the result of a combination of the multidecadal-scale changes in Atlantic SSTs (and vertical shear) along with the additional increase in SSTs resulting from the long-term warming trend. It is, however, equally possible that the current active period (1995-2000) only appears more active than the previous active period (1926-1970) due to the better observational network now in place. During the previous active period, only 1966-1970 had continual satellite coverage (33,50). Further study is essential to separate any actual increase from an apparent one due to more complete observations.

And even Goldberg's prediction of increased activity is very uncertain as to the duration of such increases:

This means that during the next 10 to 40 years or so, most of the Atlantic hurricane seasons are likely to have above average activity, with many hyperactive, some around average, and only a few below average.

10 to 40 years. That's some range, and emphasizes how much we don't know. Of course, anyone living on the Gulf coast should be prepared for hurricanes, prepared to evacuate, prepared to live for several days or more without any outside help. But "doom is coming, and it's all the fault of global warming" is not supported by the evidence, period.

Posted by: PatHMV at April 3, 2006 10:16 PM

The text of the Goldenberg article is available here, though free registration may be required. Its own graph of number of major hurricanes per year begins with 1944, and the caption reads:"Fig. 1. Number of major hurricanes from 1944 through 2000 (32). Less reliable data before routine aircraft reconnaissance dictate caution in the use of these data before 1944 (33)." Also, it doesn't show a made-up polynomial trend, but a more straight-forward 5-year running mean.

Posted by: PatHMV at April 3, 2006 10:31 PM

bk,
the difference is that this grand experiment of global warming can cause trillions in economic costs. Just look at the cost of one hurricane and one tsunami last year. Think of the value of American coastal real estate that is up to 20 feet above sea level. Let's also add the large cost of rebuilding port facilities and ancillary transportation infrastructure. Port of Oakland could very well be at my front door. We're already spending 140 million or so up in Alaska to move a town because the tundra has softened up too much and there's not enough sea ice to protect that area during storm season.

What you'll get out of cutting greenhouse emissions will be more business for companies that will feed that new market. You may have to drive no faster than 55 or 60. big deal. Imagine, some poor arab oil magnate having to reduce his harem from 140 women to a paltry 135 women.
So instead of using 300 million disposable coffee cups a day, we bring our own. Gasp, such hardship.
Okay so we4 lose a few paper mill jobs. You may see the disappearance of some butt ugly big cars, to be replaced by butt ugly small cars.
I call that reduction in negative visual impact.
More motorcycles. Cool, actually the new crop of bikes has me itching right now to spend that I do niot have........You may actually have to plant some trees along your street. For carbon sequestration and for temperature moderation so you use your AC less. You'll have to look at windmills instead of smoke stacks and maybe a few piles of dead birds here and there. You may have to insulate your house. Use compact flourescent lights. Put up a solar panel or 10. Germany, BTW, is mandating the use of solar panels on new and old construction. Not only to comply with Kyoto accords but to reduce the need to import oil and gas.

See if we make choices now we may avoid having to make harder choices in the future. If you want the current pre-eminent example, look no further than the free-spending drunken sailors that control the House, the Senate and the White House and our lovely budget quagmire.

Posted by: Marcus at April 4, 2006 03:01 AM

bk,
the difference is that this grand experiment of global warming can cause trillions in economic costs.

The boldfacing of the single most crucial word in your sentence is mine, Marcus. It's all about the "can," which is the weasel word that attempts to negotiate the unknown path. Ultimately, we need to be far more certain about the likelihood of the most negative impacts before we undertake economically draconian measures. BTW I am very pro on windmills, motorcycles, and smaller and hybrid cars (though I haven't started enjoying my own farts yet Tully LOL). I'm still agnostic on solar, not on the continued investigation of it, but on mandating it.

But I am also very pro-disposaable cups, plates, and so on, although I could see putting deposits of some of that stuff. IMO recycling makes a ton of sense as long as we can do it the right way. The current model where each individual has to spend time sorting their own tuff by hand strikes me as very inefficient. I think we should move towards standardized containers and shapes and sizes to allow mechanized collection and sorting. the only reason we're doing it the way we are now is because the cost of each person's labor to sort their own stuff is only a cost for you and me, not a cost to the system. But recycling is another time vampire of 21st century modern culture.

BTW, Pat and Tully, I don't want to misrepresent myself as an extremely advanced mathematician. I don't even have a math degree, although I studied research methods and testing in college and graduate school. I'm just decent at the things that I do know and happen to have studied. The problem that I have with polynomial curve fitting is the problem I have with curve fitting in general. You can always find a graph that makes a nice curve that mimics data somewhat reasonably. The question is whether that graph has predictive value for the data yet to come.

Like Marcus trumpeted, past results are no guarantee of future peeformance. The model Bob uses doesn't need to be fitted for future fluctuations because they don't exist yet. So all it does is follow the existing data and then at the end, it may well assume that the final trend following the last big turn will persist. Which would be a stupid assumption.

I haven't addressed it because I don't think it matters. The data can speak for itself and people can draw their own conclusions.

So your thesis here is that the opinion of the authors is not relevant, even though they are the ones most familiar with their formulation of data? I'm going to remind you that you made that argument some day when it's inconvenient for you, Bob, and you are not going to like it. Does this mean that in the future when you cite data, you'll refrain from quoting the authers when they ahppen to agree with you because you believ that data should always "speak for itself?" :-) LOL

But hey, why not take it on to the carpet here? Do you really believe that the curved line you're showing is making an appropriate forecast, or do you just like that it mimics what has already happened? How about we take your model, and see what it predicts for 2006, 2007, 2008, 2009, and 2010. And then let's see if you are willing to make a wager based on those numbers. If you come shopping very high 20s and on into the 30s, I'll place $100 on the under.

Posted by: bk at April 4, 2006 10:22 AM

BK: I'll pass on "taking it to the carpet". I've lost track of how much I've bet on this site, and I never gamble with the rent money.

My "thesis" is that if you give people information they can make decisions for themselves. If I provide a link to the experts, I've done my due diligence. People can click the link and read it themselves.

I told you my guess of 23 storms back in December of 2005. In our previous discussion I believe I stated that there has to be some kind of physical upper limit to the number of storms. At some point new storms would not be able to form regardless of how much energy is available. They will bump into the existing ones. My guess is you will have an increase till some peak in yearly totals and durations (upper limit as yet undetermined) with the hurricane season starting sooner and lasting longer. We're screwing with a nonlinear system guessing beyond a year would not be smart.

And by the way, Excel calls it a trend line , I never used the word "model". It shows current trends. I would never consider using it for long term predictions.

Posted by: Bob J Young at April 4, 2006 01:34 PM

And the current peak is exactly what is being predicted by NOAA without any factoring at all of human agency. Zero, zip, nada.

The problem with your graph is you're running a trend line against a cyclical oscillation function with incomplete data on the front end. Because of the incomplete data, the results of that are fore-ordained by the contruction. The slope will always be elevated in the direction of the more complete data. Always.

Then you throw in the apocalyptic wolf-crying on the backside, and say you're just "letting the data speak for itself." While misrepresenting the data, and insisting that the "check" data (landfall data) that IS complete is not applicable.

My "thesis" is that if you give people information they can make decisions for themselves.

And when you present incomplete information labelled as definitive, misrepresent the information they are getting, and combine it with information that hasn't been shown to be relevant, Bob, your "thesis" is cheerleading for dishonest polemic. Science it ain't, no matter how it's packaged to look like it.

Just from eyeballing the graph and knowing how it was constructed, I'll take the under at 25 for next season at even money. :-)

Posted by: Tully at April 4, 2006 02:15 PM

Piling on:

Experts forecast another active Atlantic hurricane season

"No credible observational evidence is available or likely will be available in the next few decades which will directly associate global surface temperature change to changes in global frequency and intensity."

And that's before you make that next causal-chain leap to human agency.

Posted by: Tully at April 4, 2006 02:25 PM

The experts quoted in the article are from "The Tropical Meteorology Project", lead by Klotzbach's and Gray, long term climate change dissenters . As of last December here is their past accuracy:.

The Tropical Meteorology Project forecast
2005 forecast for 2006 season Named Storms = 17, actual ?
2004 forecast for 2005 season Named Storms = 11, actual 26
2003 forecast for 2004 season Named Storms = 13, actual 16
2002 forecast for 2003 season Named Storms = 12, actual 21
2001 forecast for 2002 season Named Storms = 13, actual 14
2000 forecast for 2001 season Named Storms = 09 actual 17

This comment was updated on 04/06/2006: I accidentally included tropical depressions when looking at the accuracy of the TMP forecasts for 2001 thru 2004 and left out two storms in 2005. The correct list for tropical storm formation and prediction is listed below. (Assuming I did the math right)

The Tropical Meteorology Project forecast
2005 forecast for 2006 season Named Storms = 17, actual ?
2004 forecast for 2005 season Named Storms = 11, actual 28
2003 forecast for 2004 season Named Storms = 13, actual 15
2002 forecast for 2003 season Named Storms = 12, actual 16
2001 forecast for 2002 season Named Storms = 13, actual 13
2000 forecast for 2001 season Named Storms = 09 actual 16

Posted by: Bob J Young at April 4, 2006 02:52 PM

Pig pile!!

Call it whatever you want Bob. It's not a very GOOD trend line, and you could admit that any time now. It sure looks like the trend line does little to account for the sort of fluctuation the data have shown in the period prior to, oh, let's say 2000.

You're trumpeting 6th order polynomials. meanwhile, a 3rd grader could have drawn a line from (1991, 8) to (2001, 17) and gotten the same results. And the method would in this case be just about as valid.

See, I have no big problem with the actual points. Those actually ara data, for whatever they are worth. But the trend line ISN'T data. The trend line does is absolutely NOT "the data speaking for itself." The trend line is the curve-fitting software speaking for the data. And not speaking very well. As you yourself have implicitly conceded by declining the bet and saying you'd never use it for prediction.

I mean come ON! Seriously. Look at the trend line. Ask yourself why it ignores the really high and really low numbers from all the way from 1870 to 1995, and then starting in 1996 goes really close to the higher numbers. Admit it, the last 3 columns of the trend line are a bit of a joke.

Posted by: bk at April 4, 2006 02:55 PM

The trend line gives great weight to the last years of the chart, simply because it has no data from future years to, perhaps, pull it down. If you cut off the chart at the peak year of 1933, then the 6th order polynomial trend peaks at well over 15, some 8 or 9 points above where it shows up in Bob's trend line. Cutting the chart off in 1933 makes it look every bit as scary as Bob tried so hard to make his chart look for the current year. But because Bob's chart is not accurate as a prediction model, any predictions made on the basis of a scary graph in 1933 would be completely wrong.

Posted by: PatHMV at April 4, 2006 03:10 PM

Cannonball from CNN

The 2006 hurricane season will not be as ferocious as last year when Hurricane Katrina devastated New Orleans, but will still be unusually busy, a noted forecasting team said Tuesday.

The Colorado State University team led by Dr. William Gray, a pioneer in forecasting storm probabilities, said it expected 17 named storms to form in the Atlantic basin during the six-month season starting in June.

Nine of the storms will likely strengthen into hurricanes....
...

But they also said there was less chance of major storms making landfall in the United States compared with 2005's record-breaking hurricane season.

"Even though we expect to see the current active period of Atlantic major hurricane activity to continue for another 15-20 years, it is statistically unlikely that the coming 2006-2007 hurricane season, or the seasons that follow, will have the number of major hurricane U.S. landfall events as we have seen in 2004-2005," Gray said in a statement.

I wonder what the 6th order polynomial trend line looks like if you add (2006, 17), or (2006, 16) or (2006, 15) or god forbid (2006, 14). Jeepers, pretty soon we'd need a [gasp!!] 7th order polynomial...

Posted by: bk at April 4, 2006 04:13 PM

BK: So what's your suggestion?
It's not linear, a power function, or exponential, and a 10 point running average looks a lot like the polynomial (just bumpier). Some kind of modified sinusoidal might work but that's not available in my version of Excel, and my minds to rusty to work it out long hand.

If I adjust the number of point in the running average I can arbitrarily emphasize or dampen what is obviously a new trend that started in the last 6 years.

Average for the entire data set is 8.76~
Standard deviation for the data (calculated using the "unbiased" or "n-1" method) is 4.21~
That means three standard deviation is 12.63~ (call it 13)
Now you may have a better grasp of statistics than I do, and a more appropriate distribution, but for my work the normal distribution is usually a good enough approximation.

So for the last 6 consecutive years, the number of storms per year, has been outside 3 standard deviations (15, 17, 14, 21, 16, and 26).

So how would you fit an unbiased trend line?

Posted by: Bob J Young at April 4, 2006 04:20 PM

This is what we get for trying to focus individually on the various specific problems rather than constantly all of them... most especially the fact that pre-1944 data is extremely unreliable (see Goldenberg), and not 100% reliable until satellite coverage began in the 1960s. The standard deviation calculation will be wrong if the raw data is wrong. Plus, if the data is known to be cyclic, the standard deviation won't be as meaningful, since there are known peaks and troughs.

Posted by: PatHMV at April 4, 2006 04:37 PM

And if there is an objection to the pre 1944 data then we can use only data from 1944 forward:

Average for the entire data set is 10.72~
Standard deviation for the data (calculated using the "unbiased" or "n-1" method) is 3.996~
That means three standard deviation is 11.989~ (call it 12)

Again, for the last 6 consecutive years, the number of storms per year, has been outside 3 standard deviations (15, 17, 14, 21, 16, and 26).

Posted by: Bob J Young at April 4, 2006 04:44 PM

Ok.
Now I am going to have to eat crow.
It should be average plus three standard deviations.
2 stdev =18.71~ and three stdev is =22.72~
For post1944 data.

Posted by: Bob J Young at April 4, 2006 04:55 PM

The experts quoted in the article are from "The Tropical Meteorology Project", lead by Klotzbach's and Gray, long term climate change dissenters.

Lie, as in bald-faced. Love that dismissive ad hominem, don't we? William Gray is a dissenter from the "human-agency climate-change severe-and-right-now" school. He's not at all a dissenter on the occurence of climate change. In fact, he's a top expert on it. He's been studying it for fifty years.

Gray also is the world's top expert on hurricanes, and climate change proponents aren't. He founded the science of hurricane forecasting. When I want to know about hurricanes, Gray gets my respectful attention. That's also why the networks go to him, and not the climate change proponents. As Gray himself put it:

"There’s almost an equation you can write--the degree to which you believe global warming is causing major hurricanes to increase is inversely proportional to your knowledge about these storms."

Now, why does that sound familiar?

The problem with using your post-1944 data is that it doesn't really cover a cycle, much less multiple cycles. Which is why you should be using the 150+ years of landfall data as a "check" set, which you keep notably refusing to do. Probably because you've noticed it highlights the shortcomings of the data you're using. Why don't you plot out that landfall data and see what it tells you, Bob? C'mon, it's not that tough. You only need to do two columns, just the total storms and major storms....

Posted by: Tully at April 4, 2006 05:46 PM

And you could throw in the STDEV and AVG just for giggles. C'mon, you can do it!

Posted by: Tully at April 4, 2006 05:52 PM

Tully: Nice attempt to duck, that his accuracy sucks.

I did plot them. Congratulations, you just proved that hurricanes don't always hit land. Sometime weather patterns keep them at sea and sometimes they hit land. So what?

However as a consolation prize you win a hurricane information link: The Bermuda High

And here is a new toy to keep you occupied: The Hurricane Track Simulator

Posted by: Bob J Young at April 4, 2006 06:26 PM

Who's ducking? I was gonna be nice, Bob, but since you insist--you've significantly misstated Gray's accuracy stats. I checked. In the world of real science that's called a "lie." Or maybe you simply relied on a third source, rather than cross-checking the actual forecasts.

You've also failed to mention ANY relevant information about real-world forecasting, such as using real methodolgy instead of divine revelation, and accepting that even the best mehtodology leads to errors which help refine the methodology. As compared to doom-saying by fudging statistics from bad data and misdirecting the audience to promote a mathematically and scientifically unjustified assertion. You want to bullshit in some apocalysm with invalid statistical applications and bad data and faux causality, but those who use the tools available correctly and scientifically are all suckage, eh? Only you have the divine revelation....

I did plot them. Congratulations, you just proved that hurricanes don't always hit land. Sometime weather patterns keep them at sea and sometimes they hit land. So what?

So you plotted them, but the most reliable hurricane data available, over a century and half of solid and verified reports, wasn't worthy of a graph or any statistical analysis, eh? Why not? Because it didn't show what you wanted it to show? Why is it that you refuse to utilize the most reliable data tool available to check the reliability of your own shaky methodology?

Seems to be a pattern developing here. Maybe we should graph it.

Posted by: Tully at April 4, 2006 07:15 PM

Damn! I did make a mistake!
And It's a big enough one that its going to take weeks to sort out.
I guess thats why they have peer review.

Hurricane alley is including tropical depressions in their count!!!
How did I miss that?

The base data set I'm using isn't tropical storm and hurricanes. It's tropicals storms, hurricanes and tropical depressions.

Now I have to go check everything again from the start.

Anyone know the rule of for how to handle this? I'm inclined to delete the entire thread.

Posted by: Bob J Young at April 4, 2006 08:19 PM

Ain't no rule, so your call.

Posted by: Tully at April 4, 2006 09:08 PM

But you'll still have the same structural problem with your graph, only more so. The smaller storms were even more likely to be missed in the reporting the farther back you go.

Posted by: Tully at April 4, 2006 09:12 PM

Bob, my suggestion is to NOT fit an equation to the data, just show the plotted data points. Or to have a button to click through various speculative fits while stressing that they're speculative.

I think fitting lines can be prone to leading and misleading.

I think that what's defensible is to look at these data and say, "OK I see a slight upward trend, but I also know that such data is cyclicial, and I also know that #1 we don't know HOW cyclicially, and our data set is, geologically speaking, small.

What really bugs me when I play the role of skeptic on things like global warming is being mischaracterized as an apologist or a corporate tool, or to have my points misstated back to me as something like "how can you deny that the world is warming?"

I've never denied it. I've simply pointed out that the data is skethcy, and so it makes sense to be cautious in drawing conclusions. As Ron Bailey recently pointed out, we have at this point in the debate reached a point where a lot of careful collecting and parsing of vast sets of data suggest that Earth is in fact warming. But it's most likely warming quite gradually. We still don't know how much, but it might be warming up by something like a half a degree fahrenheit per decade.

We really don't know how that will affect our climate, or whtehr the changes will all be bad. We also don't know the extent to which the changes are due to human agency, so we don't know if the trend is durable.

That's just the way it is. And different people ar going to have very different opinions about how worried we should be. The thing is that those levels of worried are mostly based very different reactions to the same or similar stimuli. In other words, even though the dbate is about a sceince subject, the science is inadequate to answert the questions we really need answered, so the debate continues to unfold on moral and political and economic policy grounds, not scientific grounds.

I don't know how we go about determining whether the trend is durable or the extent to which it is being caused by human agency. And given that IMO we can't be at all certain that the sum total of the outcomes of the trend will be bad, it doesn't make sense to me that we frighten ourselves into drastic changes.

Posted by: bk at April 5, 2006 11:30 AM

I would also add to Brian's excellent post that we don't know what feedback mechanisms in the earth's climate might be triggered by the gradual temperature increase. For example, hurricanes generally impart a cooling influence on SST (sea surface temperature, for those joining the debate late). If higher SSTs do in fact increase hurricane formation, or intensity, it may be that the hurricanes then also provide more cooling to the SSTs, moderating or nullifying the increase in SST after a few seasons of higher
than-average hurricanes. Who knows what other self-moderating feed-back loops there may be.

Posted by: PatHMV at April 5, 2006 02:31 PM

Thank you. http://certificate-deposit-rate.blogspot.com

Posted by: certificate-deposit-rate at April 10, 2006 05:18 PM