Arguments
Patrick Michaels: Cato's Climate Expert Has History Of Getting It Wrong
Posted on 13 July 2013 by dana1981
This is a re-post of a devastating debunking of Patrick Michaels' history of wrong climate predictions by Shauna Theel at Media Matters.
A review of claims made by the Cato Institute's Patrick Michaels over the last quarter century shows that he has repeatedly been proven wrong over time. Michaels is one of a few contrarian climate scientists who is often featured in the media without disclosure of his funding from the fossil fuel industry.
Patrick Michaels' Losing Bets
On Temperature Trends
Michaels "Bet" In 1999 There Would Be A "Statistically Significant Cooling Trend" From 1998 To 2008. In a Cato post that was later published as a Washington Times op-ed, according to Nexis, Patrick J. Michaels wrote that he would place a "bet" that "the 10 years ending on December 31, 2007, will show a statistically significant global cooling trend in temperatures measured by satellite":
I'm willing to wager two things. First, I'll bet that anyone who said global warming is an overblown bunch of hooey had a terrible time at this year's holiday cocktail parties. Second, I'll take even money that the 10 years ending on December 31, 2007, will show a statistically significant global cooling trend in temperatures measured by satellite.
[...]
Last year was so warm that it induces a statistically significant warming trend in the satellite data. Thus the second bet: Starting with 1998, there will almost certainly be a statistically significant cooling trend in the decade ending in 2007. [Cato, 1/18/99]
Satellite Records For That Decade Showed No Statistically Significant Trend. From 1998 to 2008, the University of Alabama in Huntsville satellite record shows a warming trend that is not statistically significant at the 95 percent level (a warming of 0.074°C per decade plus or minus 0.439°C). The Remote Sensing Systems satellite record shows a cooling trend that is not statistically significant at the 95 percent level (a change in temperature of -0.053°C per decade plus or minus 0.425 °C). The three surface temperature records showed a "warming trend" for that time period according to a Skeptical Science report on a 2008 paper by the Australian Bureau of Meteorology. [Calculated using Skeptical Science's Temperature trend calculator, 7/1/13] [Skeptical Science, 1/10/13]
Michaels' New Bet: "We Are Going To Go Nearly A Quarter Of A Century Without Warming." In a Washington Times op-ed in January 2013, Michaels stated "it's a pretty good bet that we are going to go nearly a quarter of a century without warming." [The Washington Times, 1/17/13]
On Ice Ages
Michaels Makes An "Easy Prediction" That By 2000 The "Vogue Environmental Calamity Will Be An Ice Age." Michaels wrote in a 1992 Washington Times op-ed:
About 15 years ago it was all the rage in the climate business to proclaim the coming ice age.
[...]
Here's an easy prediction: By the year 2000, plus or minus a few, the vogue environmental calamity will be an ice age. And this nouvelle apocalypse, revised version, will predict that global warming will cause sea level to fall, exposing Bangladesh to wrenching cultural changes, and therefore we should give more money to the Third World. [The Washington Times, 2/11/92, via Nexis]
Michaels' Failed Prediction Based On Misrepresentation Of Scientific History. There has been a persistent increase in temperatures and continuing warnings about the impacts of climate change from scientists and advocates, contrary to Michaels' prediction.
[World Meteorological Association, 2013, via The Washington Post]
Furthermore, as a review published by the American Meteorological Society concluded, there was no consensus about a "coming ice age" in the scientific literature in the 1970s:
There was no scientific consensus in the 1970s that the Earth was headed into an imminent ice age. Indeed, the possibility of anthropogenic warming dominated the peer-reviewed literature even then.
[...]
When the myth of the 1970s global cooling scare arises in contemporary discussion over climate change, it is most often in the form of citations not to the scientific literature, but to news media coverage.
[...]
Even cursory review of the news media coverage of the issue reveals that, just as there was no consensus at the time among scientists, so was there also no consensus among journalists. For example, these are titles from two New York Times articles: "Scientists ask why world climate is changing; major cooling may be ahead" (Sullivan 1975a) and "Warming trend seen in climate; two articles counter view that cold period is due" (Sullivan 1975b). Equally juxtaposed were The Cooling (Ponte 1976), which was published the year after Hothouse Earth (Wilcox 1975).However, the news coverage of the time does reflect what New York Times science writer Andrew Revkin calls "the tyranny of the news peg," based on the idea that reporters need a "peg" on which to hang a story. [Bulletin of the American Meteorological Society, 2/8/08]
On Green Technology
Michaels In 2001: Prius Will "Never" Deliver Profit For Toyota. In a Washington Times op-ed Michaels declared that the Prius would "never" make a profit for Toyota and suggested that demand would always be "weak" because "no one except diehard technophiles and hyper-greens are willing to shell out several thousand extra for a hybrid":
Last year, the Big Three participants -- Ford, GM and Daimler-Chrysler -- got around to building their first hybrid gas-electric prototypes. By then it was obvious that the technology would not work in anything approaching a cost-effective fashion. Why did they continue?
Two nonparticipants -- Honda (Insight) and Toyota (Prius) -- had already demonstrated the futility of trying to produce the impossible cheaply. And both were in the process of finding out that gas is so inexpensive in this country (despite its 40 cents per gallon tax) that no one except diehard technophiles and hyper-greens are willing to shell out several thousand extra for a hybrid.[...]
Honda's prototype, the JV-X, was completed in 1997. Toyota's was even earlier, as the Japanese Prius was first sold in 1997. Years before the Big Three bothered to finish their prototypes, both Toyota and Honda let it be known they were losing big bucks on these cars, and neither company, when pressed, would say if or when they would make a profit on them. That's because the answers are no and never.
Amazingly, and despite largely glowing reviews and favorable articles by our green press, demand has been weak. [The Washington Times, 8/20/01]
Four Months Later, Toyota Announces It Is Making A Profit On Prius. The Los Angeles Times reported at the time in an article titled "Toyota Says It's Now Turning A Profit On The Hybrid Prius":
Toyota Motor Corp. said it is starting to make a profit from its Prius gasoline-electric hybrid car, four years after introducing the low-pollution vehicle.
Higher production volume of the Prius, introduced in Japan in 1997 and in the U.S. last year, and technological gains are helping the costs of its advanced battery and electrical components, said Hiroyuki Watanabe, Toyota's senior managing director for hybrid and fuel-cell systems. [Los Angeles Times, 12/19/01]
Prius Now World's Third Best-Selling Car. Bloomberg reported:
Toyota Motor Corp. (7203)'s Prius, a niche oddity when it went on sale 15 years ago, jumped to the world's third best-selling car line in the first quarter as U.S. demand and incentives in Japan turned the hybrid into a mainstream hit. [Bloomberg, 5/29/12, via Media Matters]
Michaels' Climate Excuses Didn't Pan Out
On The Hemispheres
Michaels Mocked IPCC's "Scary" Assertion That Northern Hemisphere Will Warm Up Faster. From a 1992 op-ed by Michaels in The Washington Times on the Intergovernmental Panel on Climate Change's (IPCC) assertions in 1990:
IPCC Assertion: "The Northern Hemisphere [where everybody lives and almost all of the world's food is produced: scary] will warm faster than the Southern. Fact: According to the last 30 years' data from any source - satellites, ships, or Pop's thermometer - the opposite has been occurring. Our figure subtracts Southern Hemisphere temperatures from Northern ones since 1950, and if the United Nations is right, the trend line should be going up. Instead, it is pointed significantly downward. The United Nations is using a forecast that has the world turned upside down as a basis to induce major controls on the energy economy. A few pages later, the Policy-makers Summary asserts that "the temperature rise has been broadly similar in both hemispheres." [The Washington Times, 4/8/92, via Nexis]
Northern Hemisphere Heated Up More Than Southern Hemisphere. Climate Central reported on the results of two studies:
If global warming were a race, the Northern Hemisphere would be winning. It is warming faster than the Southern Hemisphere, with some of the most rapid warming rates on Earth located in the Arctic, where sea and land ice is shrinking and thinning. Not only is the North winning now, but projections show that, largely due to the influence of manmade greenhouse gas emissions, it is likely to widen its lead in the coming decades.
[Climate Central, 4/9/13]
This graph, created using NASA's Global Maps tool, shows temperature anomalies in both hemispheres from 2000 to 2012, as compared to 1951 to 1980:
[NASA's Global Maps tool, accessed 7/1/3]
On The Urban Heat Island Effect
Michaels In 1989: Warming Over Last Century May Simply Be "Artificial." Michaels pushed the idea that "there may have been no global warming to speak of during the last century" and that this was instead "artificial" warming from the urban heat island effect:
Twentieth-century U.S. temperature data, which formed a part of NASA's congressional testimony last year, hide a drastic warm-measurement bias. NOAA scientist Tom Karl, who arguably knows more about regional climate variation than anyone in the world, has calculated that NASA's record over the United States has warmed up nearly a degree during this century mainly because cities tend to grow up around their weather stations, not because of the greenhouse effect.
If the effect of urbanization ("artificial" warming) on the temperature record averages the same over the rest of the world (and there's no reason to believe it doesn't), then there may have been no global warming to speak of during the last century. Karl's finding surprised none of us who daily toil with the data. But it should be a major shock to those who are using those figures for policy purposes. [The Washington Post, 1/8/89, via Nexis]
A Peer-Reviewed Study Found "No Evidence" That Temperature Trends Are "Inflated Due To Poor Station Siting." A 2010 peer-reviewed study by NOAA scientists and published in the Journal of Geophysical Research addressed the issue of temperature stations sited near heat sources and found "no evidence" that the temperature trends "are inflated due to poor station siting." Peter Thorne of NOAA's National Climatic Data Center referenced the study, and told The New York Times' Andrew Revkin that "if anything, we are under-estimating the real world warming trends for the contiguous United States." [Journal of Geophysical Research, 6/8/10] [The New York Times, 7/30/12]
On Nighttime Warming
Michaels: Warming Is Only During Nighttime, Which Is Beneficial. From a 1991 article by The Oregonian:
Michaels, in a paper presented Monday at the annual meeting of the American Association for the Advancement of Science, said evidence now suggests that most warming would likely occur at night in high latitudes. He added that that could be beneficial because nighttime warming may mean fewer droughts, a longer growing season and better crop yields.
If global warming occurs primarily at night, polar temperatures would remain far enough below freezing to minimize the melting of polar ice and a rise in sea levels.
Night temperatures have been rising in the Northern Hemisphere since 1950, Michaels said, although overall temperatures haven't changed significantly.
[...]
"The critical scientific question on global environmental change is not how much will the globe warm but rather how will it warm?'' said Michaels, who is Virginia's state climatologist. [The Oregonian, 2/19/91, via Nexis]
Experts: Nighttime Heat Is Especially Dangerous For People. The Associated Press reported:
Excessive heat is the No. 1 weather killer in the United States and it's at its most dangerous when it doesn't cool down at night.
[...]
Q: What's so disturbing about this current heat wave?
A: It's unrelenting stubbornness. There is no relief at night. Phoenix set a record for highest nighttime temperature: 91. Las Vegas has gone three days without getting below 90, according to readings at the airport.
"Nighttime heat is especially bad," said Eli Jacks, chief of fire and public weather services at the National Weather Service. "Not to get below 90 is crazy."
Q: What's so dangerous about that?
If you aren't in an air-conditioned place, "your body never has a chance to recover" at night, Jacks said. Normally the "feels-like" index -- which factors in temperature and humidity -- has to get to 80 degrees or below for your body to recover from the daytime heat, Jacks said.
The lack of nighttime cooling is more dangerous than the 117 degree all-time record in Las Vegas, experts said.
Q: How do heat waves compare to other weather killers?
A: In recent years, heat has been more deadly than other weather extremes in the United States. [The Associated Press, 7/1/13, via The Washington Post]
Daily Record Highs Have Also Increased. A 2009 study by scientists at the National Center for Atmospheric Research found that daily record highs occurred twice as often as record lows over the last decade. Although the press release noted that this was "could be attributed more to a comparatively small number of record lows than to a large number of record highs," indicating "much of the nation's warming is occurring at night," record highs also increased:
[National Center for Atmospheric Research, 11/12/09]
On Satellites
Michaels: Satellites, Which Almost "Cannot Be Wrong," Show No Warming. From a 1998 Washington Times op-ed by Michaels:
The three measures [according to a paper by the National Climatic Data Center] were land surface temperatures, which by definition are hardly global; sea surface temperatures taken from ships; and data from a network of buoys whose deployment was begun in the mid-1980s. The last two measurements are very different from the first, and in order to create the desired fruit salad, NCDC adjusted the sea surface temperature data up by 25 percent after 1982. That certainly might make things appear to be a bit warmer in recent years!
In point of fact, the sea surface temperature data are increasingly at odds with air temperatures taken over the ocean. No one knows the reason for this, but the air temperatures just happen to match up perfectly with those recorded by NASA's satellites, which happen to match up perfectly with the Weather Bureau's (what it was called before it became a "service") weather balloons. None of those records shows a lick of global warming in the last 20 years.
Parenthetically, we might note that recent reports about the satellite data being in error are themselves in error. Annual temperature averages taken by weather balloons look exactly like those measured by the satellites. So the satellite cannot be wrong unless, somehow, thermometers in the 1,125,000 weather balloons launched over the last 20 years have been making exactly the same mistakes in temperature measurement as the satellites. [The Washington Times, 8/31/98, via Nexis, emphasis added]
After Errors Were Fixed, Satellites Showed Same Amount Of Warming As Surface Temperatures. The New York Times reported in 2005:
Some scientists who question whether human-caused global warming poses a threat have long pointed to records that showed the atmosphere's lowest layer, the troposphere, had not warmed over the last two decades and had cooled in the tropics.
Now two independent studies have found errors in the complicated calculations used to generate the old temperature records, which involved stitching together data from thousands of weather balloons lofted around the world and a series of short-lived weather satellites.
A third study shows that when the errors are taken into account, the troposphere actually got warmer. Moreover, that warming trend largely agrees with the warmer surface temperatures that have been recorded and conforms to predictions in recent computer models.
The three papers were published yesterday in the online edition of the journal Science. [The New York Times, 8/12/05]
Who Is Patrick Michaels, Cato's Contrarian Climate Scientist?
Despite Being One Of A Few Contrarians, Michaels Is Prominent Media Figure. Michaels made about 49 major media appearances from 2007 to July 2011 (13 of them on Fox News), according to a Nexis search. In addition, The Washington Times and Forbes often publish opinion pieces by Michaels and do not disclose his funding from the fossil fuel industry. Yet during that time, Michaels only published four peer-reviewed climate articles. In comparison, 97 percent of the most actively publishing climate researchers agree that "most" of recent warming is manmade, and 84 percent of climate scientists say the public should be told to be worried or "very worried" about climate change. Despite this, USA TODAY, The Washington Post, and CNN all hosted or quoted Patrick Michaels in 2012.
[Media Matters, 7/7/11] [Media Matters, 6/25/13] [Skeptical Science, accessed 7/9/13] [Cato, accessed 7/9/13] [Nexis search, 7/10/13]
Michaels Estimated That 40 Percent Of His Funding Comes From Fossil Fuel Industries. In 2010, Patrick Michaels estimated that about 40 percent of his funding comes from fossil fuel industries:
FAREED ZAKARIA, CNN HOST: Can I ask you what percentage of your work is funded by the petroleum industry?
PATRICK MICHAELS: I don't know. Forty percent? I don't know. [CNN, Fareed Zakaria GPS, 8/15/10, via Think Progress]
Michaels Initially Did Not Disclose His Publication Was Funded By Coal Industry Association. The Society of Environmental Journalists reported in 2007 that Michaels initially did not disclose that World Climate Report, published by Michaels' PR firm New Hope Environmental Services, was partially funded by the Western Fuels Association, an association of coal mining companies and coal-fired utilities:
Michaels' web publication, World Climate Report, and its skeptical predecessors have been heavily funded by coal and electric utility industries with a large financial stake in preventing regulation of greenhouse emissions. In the 1990s, he published World Climate Review without clearly disclosing in the publication itself that it was funded by the Western Fuels Association - until after journalist Bud Ward brought this to light in the Environment Writer newsletter. [Society of Environmental Journalists, 9/19/07] [Western Fuels Association, accessed 7/8/13]
Patrick Michaels later refused to disclose the donors for New Hope Environmental Services in a trial where he was retained by automobile manufacturers and dealers attempting to prevent Vermont from regulating greenhouse gases, as he stated in an affidavit:
[O]n or around April 7, 2007, I informed plaintiffs' counsel that I would not testify at trial. My sole reason in doing so as concern that my trial testimony would result in the loss of confidentiality for the New Hope information.
[...]
Large companies are understandably adverse to negative publicity. Thus, the global warming controversy has created an environment in which companies who wish to support New Hope's research and advocacy about global warming science are increasingly willing to do so only if their support remains confidential.
Public disclosure of a company's funding of New Hope and its employees has already caused considerable financial loss to New Hope. For example, in 2006 Tri-State Generation & Transmission Association, Inc., an electric utility, had requested that its support of $50,000 to New Hope be held confidential. After this support was inadvertently made public by another New Hope client, Tri-State informed me that it would no longer support New Hope because of adverse publicity. Also, in 2006, when a $100,000 contract between New Hope and electric utility Intermountain Rural Electric Association to synthesize and research new findings on global warming became public knowledge, a public campaign was initiated to change the composition of the board of directors so that there would be no additional funding. That campaign was successful, as Intermountain has not provided further funding. [Affidavit of Patrick Michaels, 7/06/07, via SourceWatch]
Yet Michaels Accused Scientists Of Adjusting Data For Money. Michaels wrote in a 2008 Wall Street Journal op-ed "If global warming isn't a threat, who needs all that funding?":
This was not the first time Michaels had accused scientists of stoking fears of climate change for money, writing in 1992: "if you think that scientists are going to jump up and say, well, maybe my cash cow (global warming, global cooling, acid rain, the ozone hole, air pollution, water pollution, AIDS, deforestation, biodiversity, population, etc . . . ) isn't the end of the world after all, and please pass the funding somewhere else or just save it, you probably believe that men get pregnant." [The Washington Times, 6/29/92, via Nexis]There have been six major revisions in the warming figures in recent years, all in the same direction. So it's like flipping a coin six times and getting tails each time. The chance of that occurring is 0.016, or less than one in 50. That doesn't mean that these revisions are all hooey, but the probability that they would all go in one direction on the merits is pretty darned small.
[...]
This prompts the ultimate question: Why is the news on global warming always bad? Perhaps because there's little incentive to look at things the other way. If you do, you're liable to be pilloried by your colleagues. If global warming isn't such a threat, who needs all that funding? Who needs the army of policy wonks crawling around the world with bold plans to stop climate change? [The Wall Street Journal, 4/18/08, via Factiva]
This is too timely and too good not to share.
Thanks.
Patrick Michaels - renowned AGW contrarian - a closer look
http://whatsupwiththatwatts.blogspot.com/2013/07/patrick-michaels-renowned-agw.html
An interesting article. In the interests of equity can you find and post an article showing how many predictions made by the IPCC have not come to pass?
[DB] In the interests of equity, you are welcome to post such a link to a reputable source where this is discussed, provided it is evidenced-based and references the primary literature.
Ray, do you know of any?
There are several articles here identifying IPCC predictions (search on IPCC) which are likely to be incorrect (e.g. sea level rise is clearly going to be higher than past projections, arctic sea ice will certainly melt out far sooner than predicted, Antarctic ice mass loss will be much more rapid than predicted, et cetera), but I'm not sure that any of these qualify as already having "not come to pass".
As a follow up to CBD's comment, I notice that all the IPCC projections he mentioned are in error on the conservative side. Michaels claims that the IPCC overestimates the problem. Ray should compare when the IPCC overestimates the problem. I also note that the IPCC would be expected to have more missed projections because it has projected many more future issues than MIchaels.
I am not aware of any IPCC projections that are in error and overestimate the problem. If Ray would list several such examples it would be a learning experience.
Ray @ 2,
From the first IPCC report in 1990:
Warming has occurred faster over land, with minimal warming in Antarctica.
Warming has been nearly 3 times as much at the north Pole than the global mean.
(Don't know if that has occured or not)
This occurred a year later with the eruption on Pinatubo, and the influence on short-term temps was well-predicted by Hansen's 1988 model.
They are now.
Current rise is 3cm per decade, and acceleration may have been occurring (the trend for last century is 1.7cm per decade).
Specific humidity has increased since the 1970s.
Excepting the South Pole, that is the case, but this report also predicted minimal warming over Antarctica.
Global precipitation has increased over the last century.
The FAR (First Assessment Report) pointed out that there will be significant regional variation for various indices, which has also happened.
I picked the FAR, not just because it was the first report, but also because there was not comprehensive indices at that time for many components of the climate system. In the 22 years since then, much data has been collated, verifying the sign of almost all predicted trends.
There are a few unresolved predictions.
1) Tropical tropospheric temps should warm faster than the rest of the atmsphere.
This is not a GHG prediction, but a prediction that the tropicqal tropospheric temps should warm faster no matter what the cause. Observations are not uniform in confirming this, and it remains a very active area of research.
2) Antarctic sea ice should decline.
Over the periof of satellite measurements, it has increased, but if the record is longer extended, it has declined, with caveats.
Regarding #2 Antarctic sea ice should decline.
"surface area" may be increasing... but just as with the Arctic, the thinning of ice says more than the "extent" of short-term winter ice.
====================================
http://www.scar.org/news/antarctic/ - 28 June 2013
More than half of the melting of Antarctica's ice occurs at just ten small ice shelves.
The results, which appear in Science, suggest that warm ocean currents are melting ice shelves predominantly at certain locations around the continent, to an extent that accounts for 55% of the annual meltwater. The findings will help scientists to tackle larger questions about how the Antarctic ice sheet might change in future and its contribution to global sea-level rise.
For more information, please read the Nature News item or read the full paper:
Rignot, E., Jacobs, S., Mouginot, J. & Scheuchl, B. Science DOI: 10.1126/science.1235798
Oh and about that surface snow/ice increases, doesn't that have a lot to do with more moisture in the air, warmer ocean temps, more snow fall...
and such dynamics that have been directly influenced by global warming?
citizenschallenge,
I'm not aware of any studies of Antarctic sea ice thickness. Latest results on the increase in extent point to changing wind patterns, but no one really knows why it happening.
Also from FAR;
Confirmed over the satellite record (from 1979).
You can find other predictions at the source (Chapter 5).
Several comments from Dr Pauchauri such as that there has been no warming for 17 years and that discussion of the science is essential suggest that all the forecasts from the IPCC may not be set in stone. For example sea level rises from IPCC are 3.1mm/year; data from NOAA for 2005-2012 is 1.1-1.3 mm/year which is a bit less than half. NOAA values from satellite altimentry and ARGO. Latter values actually about 0.3mm/year but a "correction factor" by NOAA of 0.9mm/year increases ARGO values. With error bars min and max values from NOAA are 0.2 -2.2mm/year quite different from IPCC. Is NOAA wrong? And why, if global temperature is increasing and ice melt is increasing, why are current increases in sea level rises not increasing also?
[DB] "Several comments from Dr Pauchauri such as that there has been no warming for 17 years"
Please furnish a linked to a reputable source directly quoting Dr Pachauri on this. An approved transcript and/or unedited video will suffice. Until then this claim is unsupported and off-topic in this thread.
Ray, you state, "Several comments from Dr Pauchauri such as that there has been no warming for 17 years..."
Please show the credible source of that alleged remark. I'll wager it doesn't exist. Secondly, sea levels do continue to rise. :Please link to the credible source that supports your assertion that they do not.
Concerning Michaels' claim to be "state climatologist of Virginia" there is no "State Climatologist" of Virginia.
Ray @9,
1) The Australian newspaper says that Pauchauri agrees there has been no temperature increase over the last few years. That is dubious in that it is not true, and it is not what Pauchauri typically says on the issue. The Australian will not release transricpts of the interview to verify its claim, so it is likely that they have misrepresented some related comment by Pauchauri. The most typical misrepresentation in this sort of case is misrepresenting "the rise in temperature has not been statistically significant" as "there has been no rise in temperature".
2) NOAA claims a sea level rise of 2.9 mm/ year, with the most recent period being greater than that:
NOAA notes that adding the global isostatic adjustment would raise that rate of sea level rise by between 0.2 and 0.5 mm/year.
I note, however, that you are prepared to cherry pick data to create a misleading impression.
BTW, GISS L-OTI anomaly has been updated: .68. That's the second warmest June in the instrumental period, after 1998 (during the peak of the big 97/98 El Nino). No El Nino to help out. Seven of the eighteen warmest months in the instrumenal period that have exceeded June 2013 occurred in El Nino conditions. The first six months of 2013 have all been above .5. That's happened four other times: 1998 (El Nino), 2005, 2007, and 2010 (El Nino). ENSO MEI for June is neutral at best.
I emailed Mr. Michaels, and managed to get him to agree to a bet. But he won't use GISTEMP. He wants to use CRU. Any advice on the bet? I used to win a lot at Intrade on betting GISS and ICE, so I'm hoping to structure an even money bet on a reasonably marker of what he calls ""...a pretty good bet that we are going to go nearly a quarter of a century without warming."
Any suggestions are welcome, in order to help me resume my denier's tax.
@DSL, yeah, the Intrade markets were set up at .45, .55, .65, and .75 markets. I went long on them all, and did well, but I would have raked it in last month because I'd have been increasing bets every month on the higher ends until they paid off. You could get .65s for around a dollar a share usually.
There is an unfortunate tendency for using global average surface air temperatures as THE measure of global temperature and it's not confined to laypersons. ie people who should know better say things like "The world warmed by x degrees" or "The world has experienced no significant warming for x years" (suggesting air temperatures near the surface is equivalent to 'the world', when if fact most of the heat, whether during the periods when "global temperatures" are experiencing there normal (plus a bit) ups or normal (plus a bit) downs, goes into the oceans. And this increase in heat content hasn't slowed or stopped.
Warming of surface air temperatures is not the same as the world warming because air only holds a few percent of the heat the world has gained.
In my opinion climate science communicators who conciously or unconciously use air temperatures as the definitive measure of change to our climate system - or just as shorthand - are contributing to ongoing misunderstandings. If Pauchauri has been one of them I'm disappointed.
Back to scrutinising the 3% - Michaels' actual knowledge of climate makes him more effective at pushing the denialist/obstructionist line. Certainly he would know that the world continued to warm during the past 17 years and that is shown clearly in heat content with most of it in going into rising ocean heat, so therefore, to push his biased conclusion that preceded his formulation of his arguments, he would know to distract attention from heat content and keep the focus strictly on surface air temperatures. When the next el Nino sends that 'stalled' warming back into record territory, he will suddenly rediscover some of the natural processes that make that variability - processes that he currently avoids mention of in discussing why we can have a 17 year 'pause' without global warming having stopped or slowed.
Supak @14:
First, you should insist on using HadCRUT4 rather than HadCRUT3.
Second, find out the start year for Michael's predicted 25 years with no warming.
Third, from existing HadCRUT4 data, find the standard error on 25 year trends.
Fourth, bet that on the 25th year, the warming trend will be greater than two times the standard error for HadCRUT4 trends.
Alternatively, find the AR4 multimodel mean prediction with a HadCRUT mask (either version). Bet that the GMST as measured by HadCRUT will be within 0.2 degrees of that plus an adjustment factor based on ENSO as per Foster and Rahmstorf. His position will be that it is within 0.2 C of the 1996-2012 mean. Each year, there is a standard payout from the person the other person to the person where the temperature falls within "their" predicted range. Obviously if it falls in both, there is no payout.
I'm guessing that the focus on surface air temperatures is because 1), it is only recently that ocean heat content indicea have been collated, 2) impact of global warming (for humans anyway) will be experienced mostly at the surface.
IPCC reports have a strong (but not exclusive) emphasis on surface temperatures. AR4 Summary for Policy Makers has a few short paragraphs on ocean temperatures, for example, but much more on surface data. There are no ocean heat content projections in AR4, so it's no wonder people from all points of the debate have focussed on surface data.
There were plenty of predictions made in FAR (mostly Chapter 5) with far less climate data than today. A balanced scorecard, pointing out what they got right and wrong could make a very strong point in the debate. Worth an article at SkS?
(I know there have been articles on FAR surface temp projections, but there were many other predictions made, some of which I mentioned in my first post above)
Tom Curtis @12
Your comment
"The most typical misrepresentation in this sort of case is misrepresenting "the rise in temperature has not been statistically significant" as "there has been no rise in temperature".
If the rise is not statistically significant is it really a rise? If it is does this make statistical evaluation redundant?
Ray, if you pick short enough periods in a time series with a low signal to noise ratio, you can get many or even all those periods to lack a statistically significant rise. But longer periods that contain those short periods do have a statistically significant rise. That apparent logical incompatibility is only apparent, when you understand what "statistically significant" really means.
Ray
You are missing the point of Tom's comment. If we are talking about the 'statistical significance' of something, we are discussing how likely that something was to have occurred due to random chance or not. Not whether that 'something' happened at all. Tom is referring to the fact that the concept from statistics - 'statistically significant' - is misleadingly equated with the colloquial English word 'significant' when actually the two terms have very different meanings.
However, if one seeks to mislead people, not pointing out the distinction is all that is needed to con a reader who is unfamiliar with the meaning of the term from statistics.
Lets take Phil Jones' oft mis-represented comment about there hasn't been a statistically significant warming since x (I can't remember the year) He actually said that it wasn't statistically significant at the 95% confidence interval which is the usual standard measure. He also said that it was statistically significant at the 90% confidence interval.
His meaning, in plain English - 'There has been warming; the chances that this observed warming is not random are not greater than 95%; but they chances that the observed warming is not random are greater than 90%'
Thank you for the comments in posts 21 and 22. In my own field of science which is biochemistry, Student's t-test is routinely used as a measure of significance. If I get a t value of p<0.5 then I can't claim the results didn't happen by chance alone. Certainly no journal in my field would accept a claim on my part that my finding was unequivocally real. Similarly no journal would accept my use of a 90% confidence interval to validate my findings. Interestingly there has been an evaluation of global temperatures from 6 major data sets (GISS, Hadcrut 3 and 4, RSS, UAH, NOAA) using data retrieved from Skeptical Science and the SkS Temperature Trend Calculator. Statistically significant warming at the 95% confidence limit has not been shown in any of the data sets for periods ranging from 18-23 years.
[DB] "Interestingly there has been an evaluation of global temperatures from 6 major data sets (GISS, Hadcrut 3 and 4, RSS, UAH, NOAA) using data retrieved from Skeptical Science and the SkS Temperature Trend Calculator. Statistically significant warming at the 95% confidence limit has not been shown in any of the data sets for periods ranging from 18-23 years."
You'll need to provide a proof link for this assertion. As longtime commenter KR has observed many times in this venue:
Therefore, in the absence of a change in trend, the previous warming trend is therefore still in place.
Ray @20, first I'll note that Glenn Tamblyn's response is excellent. If you think carefully about what he wrote I need add nothing (though I'll add it anyway). In particular, he pointed out that for any period in the last forty years, even though the linear trend over that period is positive and significant (0.167 +/- 0.037 C per decade), there is a shorter period in which the trend is not statistically significant. If you interpret "the trend is positive but not statistically significant" as meaning there is no trend, then for each of those periods there both was a positive trend (because it is part of an extended period with a statistically significant rise), and no trend (because it is part of a period without a statistically significant trend). Thus, the common misinterpretation of lack of statistical significance is directly inconsistent. Logically, it cannot be true.
In fact, statistical significance speaks not to whether the observed trend is rising or falling, but rather how confident we can be about that claim. If the trend is rising but not statistically significant, that means there is:
That is all it means.
If we know that the trend is positive at the 90% confidence level but not at the 95% confidence level, then:
The reason it is a 95% chance rather than a 90% chance is because the confidence interval is symetrical, so there is a 5% chance that the trend will be positive, but larger than the upper bound of the 90% confidence interval, which is, of course, still positive.
Of course, those probability estimates are based only on the information in the temperature data alone (plus some well characterized statistical assumptions). If we include other data, such as the known relationship between ENSO oscillations and temperature, and the known ENSO fluctuations over the period, the probability that the underlying trend is positive becomes much greater that indicated the temperature data alone. Indeed, including the information about ENSO and other known short term influences shows the underlying trend since 1997 to by 0.209 +/- 0.085 C per decade (GISTEMP adjusted).
In any event, turning the claim that the trend is positive but not statistically significant, ie, the claim that:
into the claim that there is not positive trend can at best be construed as woefull ignorance, but in many cases is more likely to be deliberate misrepresentation.
Ray @23, you now show evidence of some statistical sophistication. Given that, I wonder how you can justify parsing:
"The hypothesis of no warming has not been excluded"
as:
"There has been no warming".
To my mind the later is a positive claim, in effect stating that the hypothesis of warming has been excluded by the data. And that is, of course false. In science, no hypothesis gets prefferential treatment. Not even the nul hypothesis, and failing to disprove the null hypothesis is not the same as proving the null hypothesis, no matter how many times deniers pretend to the contrary.
In this case, however, the null hypothesis is only disconfirmed if you cherry pick your start points to limit the data available; and if you exclude other known relevant data (such as ENSO states). I wonder how you would mark a student who, having tested a hypothesis and excluded the null, dropped half of the data and redid the test, then reported only the data used in the second test along with a confident claim that the failure of that data to exclude the null hypothesis showed that the hypothesis being tested was false (ie, "There has been no warming.")
The link is http://en.wordpress.com/tag/werner-brozek/. In the article the author comments on the slope of the lines. As I stated I'm a biochemist not a climate scientist and I acknowledge that comments such as those by Tom Curtis and others are from those with more understanding of the science than I have. That said, the scientically trained layman can see that "the science is settled mantra" may be premature from some the findings reported.
Ray makes two statements (in two different comments):
The unstated question is: not statistically significant compared to what? If you are as familiar with a Student's t test as you claim, then you will remember that an important step in the calculations is where you determine "observed - expected". In the case of a temperature trend, "observed" is the slope from the regression, and "expected" is the value from your null hypothesis. But what is your null hypothesis?
- if you want to see if the slope is non-zero, then your null hypotheis is zero and your expected value is zero. This is very common.
- if you want to see if the past warming trend has "stalled", or "stopped", etc., then your expected value is not zero, as your null hypotheis is that the trend has not changed. To use zero as the expected value is wrong - you should use the previous trend value as your expected value. This mistake is also very common.
For the case where a calculated trend is neither significantly different from zero, nor significantly different from the previous trend, then the data that you have is insufficient to distiguish between those two null hypotheses. Which is where Glenn and Tom's comments about short time periods become relevant.
If you are unaware that a t-test can use an expected value other than zero, then you've got some learning to do. If you are aware that an expected value can be non-zero, then you've got some explaining to do.
Ray,
I am still waiting for you to provide examples of the IPCC exaggerating the AGW problem. You suggested at 2 above the IPCC exaggerates the AGW problem. SInce you referenced WUWT above I presume you have not seen any posts there that document the IPCC exaggerates the AGW problem. If you provide no examples that would indicate that your suggestion of exaggeration was incorrect.
Please provide a link to a Skeptical Science thread that claims "the science is settled". There is much to learn about AGW. On the other hand, even the skeptics now admit that the temperature is increasing. The greenhouse effect and the contribution of CO2 is also no longer debated by scientists. If you want to suggest that we have to convince WUWT you will have to be more specific about what you claim is not settled. What exactly are you claiming is not settled? Please provide specific examples.
It is easy to make absurd claims if you are not required to back them up with specific examples.
And finally, Ray, if you come up with a period of statistically insignificant warming, or even a period with a negative trend, so what? What does it mean? Is it "aha! gotcha!"?
David Rose of the Daily Mail thinks that it means that "global warming has stopped," and he's quite willing to use his power to spread that message far and wide. Do you think a sixteen year statistically insignificant warming trend in global average surface temp means that global warming has stopped? If so, what does a sixteen year trend of .286C per decade mean (1992-2007)? Do you think it's good methodology to use the surface trend to claim a theory has been demolished, and to do so without performing a component analysis, and without accounting for ocean heat content and the energy used in (accelerating) global ice mass loss.
Aha! Gotcha! Oh ho! Uh huh!
No. The theory of anthropogenic global warming is not based on the global average surface temp trend. The theory did not start with "hey. it's warming. i wonder what's causing it." It started with "hey, per Tyndall, CO2 is a greenhouse gas. I wonder what would happen if humans increased atmospheric CO2."
@Glenn Tamblyn, Tom Curtis, Bob Loblaw:
It would be extremely beneficial to all of our readers if you guys were to meld your responses about statitiscal significance into a blog post article.
Michael Sweet Himalayan glaciers vanishing by 2035. I'm fairly that's not an "absurd claim" Sorry but I don't know if there is a link to "The science is settled" at SkepticalScience but here's one to Real Climate in a piece written by Gavin Schmidt. http://www.realclimate.org/index.php/archives/2009/12/unsettled-science I thought this comment of his was very relevant: "The reason why no scientist has said this is because they know full well that knowledge about science is not binary – science isn’t either settled or not settled". Too true
Bob Loblaw you appear to be under the impression that I did the statistical analyses of the temperature data. I didn't You'd have to ask those that did to find out what their null hypothesis was. Hope that is explanation enough but I don't recall seeing it specifically stated
[TD] The Himalayan glaciers error is covered here.
Not an exaggeration, an error, which IPCC acknowledged when it was pointed out. AR4 has a few other errors in it.
You're better off choosing the first or second assessment reports (FAR 1990, SAR 1995), Ray, because projections from the TAR (2001) or AR4 (2007) generally can't be tested owing to the short time span since.
If you want to focus on surface data, 20 years is a fairly good minimum time period, and a bit more for satellite data - if atmospheric temperatures are the metric you prefer.
Skeptical Science has a handy tool for that - here. The trend results come with uncertainty estimates. You can get an idea of how much data is needed to achieve statistical significance by changing the time period length. You might want to test claims you've read about by plugging in the time periods, and seeing if the trends are greater than the uncertainty. It's a useful bookmark for the climate debates.
A warming climate does not preclude weather events, and at short time periods the 'noise' can obscure the signal.
There is no statistically significant warming for the last 3 years. Do you think that represents a trend in climate?
You need a premise with which to begin assessing climatic changes rather than the variance within the system. One of the first things to ask is, "how long is enough determine global climate change?" Obviously a few days or months or a handful of years isn't enough. So how long do you need before the short-term effects roughly balance and a climate signal emerges?
Robert Grumbine attempted a mathematical estimate some time ago.
http://moregrumbinescience.blogspot.com.au/2009/01/results-on-deciding-trends.html
The following is worth reading, too.
http://tamino.wordpress.com/2012/07/06/how-long/
(In the 7 to 8 years I've been browsing the climate debates, not once have I come across a 'skeptic' doing analyses like these. It is obvious why that is so)
Ray:
No, I was not under the impression that you did the statistical analysis. I was hoping that you understood the analysis well enough to know what it really meant. That hope is currently dwindling.
I am still at an impass, with insufficent observations to distinguish between two hypotheses. Those were listed in my last paragraph of the comment above, but I will reword them here, perhaps a little more directly (or bluntly):
1) are you unfamiliar with the proper way of determining a null hypothesis and incoporating that into the "expected value" of a statistical significance test (e.g., t-test)? If this is the case, then are you willing to learn about it before you make any further uninformed comments about significance testing?
2) If you are familiar with how to appropriately choose an "expected value", then why are you making statements that are not supported by the statistical tests that other have done? - e.g., the incorrect idea that a non-significant difference from zero means that the trend is zero? If you do know the meaning of such tests, and are deliberately twisting them into something that they do not actually mean, then that raises questions as to your sincerity.
There may be another explanation that I am not considering, so feel free to provide alternatives. For now, if I had to bet, I'd go with option 1, given that you do not seem to know what null hypotheses others have used in their trend calculations.
To try to keep at least a bit of this on-topic, Pat Michaels seems to often fit into option 2: I think he knows better, but he realizes that many casual readers/observers do not. Consequently, he twists the science into something it is not, makes grandiose statements that lead people to where he wants them to be, and to hell with what the evidence really says. I would not buy a used car from him.
Supak @14, 15: Any bet that I made, I would want to bet on different outcomes for different emissions scenarios, so that I wouldn't end up having a financial stake in "rooting" for the human race to continue to drive itself into devastation. But I'm not a climate expert - maybe 25-year predictions don't depend much on which emissions scenario happens during those 25 years, in which case I applaud your denier tax.
On the other hand, is Michaels making enough of these bets that he is liable to go bankrupt and leave you unable to collect?
Ray: You comments are based on a simple but common misunderstanting of statistical significance. We have been meaning to do an article on this, but there never seems to be time. See Misuse of Statistics for an overview.
A statistical significance test on a trend can only disprove the null hypothesis, it cannot prove it. Therefore the fact that the trend on some arbitrary short period can never falsify the hypothesis of continuing warming. That's a simple statistical fallacy.
The confusion arises from the fact that the test is not symmetric. It can falsify the hypothesis of 'no trend', but it cannot falsify the hypothesis of 'continuing trend'. For that you need to change the test.
That may seem counterintuitive, but it's really very simple. If a trend is not statistically significant, it's telling you one of two things. Either
But it doesn't tell you which, so you have no more information than when you started. On the other hand if the trend is statistically significant, then you know that both
(subject to the given confidence intervals).
Thus, whenever you hear someone quoting the lack of statistical significance of a trend as evidence for a hypothesis, you can immediately conclude that they don't know what they are talking about.
Ray,
If the best you can come up with for the IPCC exaggerating AGW is an obvious typographical error I take that to mean you agree that the IPCC has not exaggerated AGW at all. A non-skeptic scientist found that typo and the IPCC has publicly acknowledged that they copied an error from another publication.
I dare say that all of the regular posters here agree with Gavin on the "science is settled" statement. Your post here implies that at Skeptical Science we claim the science is settled. Perhaps you need to write your posts more carefully so that you are not misunderstood.
[TD] Ray, the IPCC mistake about Himalayan glaciers is covered here.
Tom Curtis @ 17
While some of what you said went over my head, I appreciate your input and will definitely try to hunt down some of the information you offered.
jdixon1980 @ 36
"so that I wouldn't end up having a financial stake in "rooting" for the human race to continue to drive itself into devastation."
That's funny, and what my wife often laments about my propensity to want to gamble on the future of the human race. I don't have these qualms because I don't have to root for a bad outcome. It's already going to happen.
My guess is that even if he has millions in the bank, he has no plans to pay out on any such bet, so I'm probably wasting my time. But I am keeping what he's saying and plan to make a blog post out of it at some point. I'm just looking to get some input from you guys so I don't look like a fool.
See, while I understand a lot of this stuff, I'm really just a gambler who is always looking for a sure thing. I cannot think of anything much more sure than the fact that we are going to see significant warming of surface temps in the next 25 years.
I do, however, like the comment from Ken in Oz @ 16 that going by surface temps alone is a fool's errand. We shoud be sure to always mention that the vast majority of the warming has gone to the oceans. If only there were some accurate measure of ocean warmth I could get Micheals to agree on, then I'd include that as a seperate bet.
With regard to post no. 34, there are more fundamental problems with current models in making projections because they are largely unverifiable.
[PW] Discussion of 'models are wrong' are veering a bit off-topic for this thread: further discussion should be posted here.
http://skepticalscience.com/climate-models.htm
Barry @18 -
This is true, however most of the discussion about statiscally significant warming and the false climate denier conclusion that warming has slowed or stopped for 17years that has followed our comments looks kind of pointless in light of the clear and measurable global warming - 4 Hiroshima bombs per second worth - that has been occurring during this period of "no significant warming". Even if temperature records for oceans don't go back as far as surface air temperatures (because existing meteorological records was where much initial effort was spent looking for climate change signal) they are adequate now to measure a fundamental change in the climate system's energy balance - which SAT's do not.
Of course this is just one way to frame the discussion - alternatives include focus on the known natural oscillations, cycles and underlying climate processes. In order for the deniers and obstructors to maximise the spin value (spun falsewards) it is necessary for them to falsely imply that if SAT's don't rise consistently year to year, decade to decade then something besides GHG's are responsible - and other things besides are indeed responsible for most of any short term up or down but the deniers and obstructors assiduously avoid examining them closely. Flattening of SAT's presents a window of opportunity to mislead the public and the public's political representatives. As I said before, when el Nino hits, "global" temperatures hit new highs, they will rediscover selected natural phenomena like ENSO that, when applied to the current period, reveal the lack of foundation for their 'warming has stopped, climate scientists are wrong' arguments.
Foster and Rahmstorf showed clearly that when you take the known natural influences out of the SAT records much of the internal variation disappears and a clearer look at the underlying trend is revealled. ie there is no 17 years of no warming. It's entirely an artifact of an artificial argument, intended to obfuscate the seriousness of a global problem of unparalleled extent.
But they overlay a strong clear warming trend with ups and downs. Well, ups and levelling offs because of that underlying warming. That's worth pointing out; if there was no underlying global warming those would indeed be ups and downs - 'statistically significant' periods of SAT's going unmistakeably down. Short lived periods, of course, because the bulk of those variations are due to oscillations like ENSO.
If global heat content were THE measure of change to our climate system all these arguments about statistically significant trends over 'x' number of years become more clearly what they are - time wasting distractions.
glhoffman,
With regard to post no. 34, there are more fundamental problems with current models in making projections because they are largely unverifiable.
That's a bold claim, supported by the following:
Decadal predictions have a high profile in the climate science community and beyond, yet very little is known about their skill.
Emphasis mine.
Can you see the difference?
In the context of repeated statements about the fact that the signal:noise ratio is too high over short time periods to even be able to confirm global warming is still happening over those short periods — purely because of that signal:noise ratio, not because it isn't still happening! — what makes you think that saying it's difficult to verify model predictions over those same time scales is saying something profoundly important about their skill?
Of course models are verifiable, and there have been plenty of model predictions that have been verified, including multiple cases where the data apparently falsified the models at the time and then it was discovered that the data was wrong! The trick is that we just have to be patient purely and simply because the signal:noise ratio does not allow us to verify these models until enough time has passed for the short-term weather noise to cancel out revealing the underlying climate change. If you do that, then even quite simple models do a decent job of hindcasting the last hundred years or so.
I have a couple of problems with this article, although I am by no means a fan of Pat Michaels.
"Yet during that time [2007 to July 2011], Michaels only published four peer-reviewed climate articles. In comparison, 97 percent of the most actively publishing climate researchers agree that "most" of recent warming is manmade, and 84 percent of climate scientists say the public should be told to be worried or "very worried" about climate change. Despite this, USA TODAY, The Washington Post, and CNN all hosted or quoted Patrick Michaels in 2012."
I have no data to go on (although John Cook may well have, having just done a paper ) but based on no more than the SkS policy of citing learned articles as (Whozitz, 2013) and not (Whozitz, 2013a, b, and c) that one peer-reviewed article per year would put him in at least the upper quartile of those publishing, not the lower quartile as this paragraph rhetorically implies. I would be much more interested in how he got the sort of nonsense he persists in spouting in interviews past peer review.
Yes, he does get 40% of his funding from the fossil fuel industry, and yes, the Cato Institute was the Charles Koch Foundation for the first two years of its existence. You justifiably heap scorn on his claim that scientists support AGW because of "funding" (not money ... I think there is a difference) because that is an extraordinary claim for which he provides no evidence. But where is your evidence that Pat Michaels would be saying anything different if he did not receive any money from the fossil fuel industry? While not directly stated, the imputation is of venality on his part, and while I can believe there are some people who would be willing to prostitute themselves into saying anything in their client's interest, most of them go into law rather than science. Corelation is not evidence of causation.
Today is the 68th anniversary of the first field test of the biggest science project in history, in Los Alamos, New Mexico. I wish you all the best of luck keeping politics out of your science project ,,, at least much more than those scientists had.
Best wishes,
Mole
Old Mole @43, I believe Anderegg et al, 2010 is the more appropriate study for comparison purposes. They show that mean expertise of climate scientists unconvinced by the evidence of AGW (UC) as measured by lifetime peer reviewed publications is 60, about half of that of those convinced by the evidence (CE) at 119 publications. Michaels appears to fall just below the mean for those who are UC, with around 50 peer reviewed publications over his lifetime, although he has far more non-peer reviewed publications for Cato or Heartland. Restricted to publications from 2007 forward, my search only reveals three publications. If he has indeed published four over that period, and a similar level of inaccuracy applies to the whole search it is certainly possible that he has more than fifty lifetime publications, and may have as many as one hundred. For comparison, here is a graph of the number of publications by researchers analyzed by Anderegg. Michaels falls in the second smalles category:
Tom Curtis @ 17
<<
First, you should insist on using HadCRUT4 rather than HadCRUT3.
Second, find out the start year for Michael's predicted 25 years with no warming.
Third, from existing HadCRUT4 data, find the standard error on 25 year trends.
Fourth, bet that on the 25th year, the warming trend will be greater than two times the standard error for HadCRUT4 trends.
>>
Here's his offer:
<<
statistically significant (p= .05) warming trend for 25 years based on annual data beginning in 1997, using HadCRU4
>>
Supak @45, notice how he offers a cowards bet by choosing the largest (or second largest depending on index) El Nino of the twentieth century as a start year. He also reinforces that by using the thermometer based index which is most effected by ENSO fluctuations due to lack of coverage in the Arctic and North Africa. Nevertheless the bet is safe for you, IMO, provided you have a clause calling the bet of if there is a major volcanic eruption (equivalent to Pinatubo or larger) in the last five years of the bet period. You may also want a clause calling the bet of if carbon sequestration schemes reduce CO2 concentration below 400 ppmv by the 25th year, but that is sufficiently unlikely to not be a necessary clause.
More technically, the current 25 year HadCRUT4 trend is 0.147 +/- 0.69 C. The current trend since 1997 is 0.045 +/- 0.12 C. The error for twenty five year trends tends to be just below 0.7 C, so you are betting that the trend will increase by more than 0.35 C to about half the current 25 year trend and to about two thirds of the IPCC predicted trend over that period. Failing a major volcano, if you do not win the bet the IPCC have got it wrong in a big way.
Finally, you may want a clause switching the bet from HadCRUT4 to later HadCRU products which supercede it if they have more extensive coverage, of if the HadCRUT4 product is no longer produced.
Supak,
So if there is a 93% chance of a temperature rise Michaels wins and can claim there is no temperature rise? Perhaps if p<.05 you win .20>p>.05 no winner and p>.2 Michaels wins.
If Michaels wants to win on a 90% probability of temperature rising (with a cherry picked start date), he certainly thinks the temperature will go up. The question is how much will it go up.
michael sweet @47,
First, I have just noticed that I accidentally dropped a decimal place from the standard errors of the 25 year trends in my preceding post, which should have been 0.069 C and 0.07 C.
Second, here are the AR4 short term predictions:
You will notice that the "commitment" prediction matches the 1998 observations, but that all other predictions are for temperatures 0.3 C above 1998 values, or higher. If the IPCC is anywhere near accurate, 1998 values should be routinely exceeded by 2020, and I suspect will be routinely eceeded from 2015 forward. By 2023, the termination of the bet, even very strong La Ninas will be hotter than the 1998 El Nino.
Put another, way, if the IPCC is correct, then the trend to 2023 has only a 2.5% chance of being less than 0.13 C per decade. Using the 1997 El Nino as a start point increases that chance, but not enough to bring the lower 95% confidence interval below 0.07 C per decade. Michaels bet accurately represents the claim of 25 years without warming (if we ignore the cherry picked start point). Not taking it if you are prone to bet would reflect a lack of confidence in the IPCC predictions.
Tom,
I am not much of a betting man. I see your point that the IPCC projections are that you should measure a temperature increase at the 95% level after 25 years. If you like to bet, it would be a good bet to make.
On the other hand, Michaels publicly claims that the temperature rise has stopped. In this bet he is claiming the temperature has not risen when there is a 90% or greater (as long as it is less than 95%) probability that the temperature has risen. If you just look at the data the slope is always positive. In the peer reviewed data scientists prefer a 95% confidence to make a claim. In common speach when there is a 90% chance of something not being chance you would say it is happening. Michaels is making the opposite claim, ie that if it is not 95% significant there is no rise. That is a false claim.
To me, if Michaels is only willing to bet on a cherry picked date and he wins at 93% confidence, that means he is convinced the temperature is rising and will continue to rise.
OK, I've managed to get Michaels to publicly bet with me. I basically had to call him out in the comments at Roy Spencer's blog to do it, but...
http://www.drroyspencer.com/2013/09/pat-michaels-bets-on-25-years-of-no-warming
So, Pat and I have a bet. Details here.