Statistical analysis shows pattern consistent with pre-industrial temperature swings, study concludes
From McGill University’s Shaun Lovejoy
Statistical analysis of average global temperatures between 1998 and 2013 shows that the slowdown in global warming during this period is consistent with natural variations in temperature, according to research by McGill University physics professor Shaun Lovejoy.
In a paper published this month in Geophysical Research Letters, Lovejoy concludes that a natural cooling fluctuation during this period largely masked the warming effects of a continued increase in man-made emissions of carbon dioxide and other greenhouse gases.
The new study applies a statistical methodology developed by the McGill researcher in a previous paper, published in April in the journal Climate Dynamics. The earlier study — which used pre-industrial temperature proxies to analyze historical climate patterns — ruled out, with more than 99% certainty, the possibility that global warming in the industrial era is just a natural fluctuation in the earth’s climate.
In his new paper, Lovejoy applies the same approach to the 15-year period after 1998, during which globally averaged temperatures remained high by historical standards, but were somewhat below most predictions generated by the complex computer models used by scientists to estimate the effects of greenhouse-gas emissions.
The deceleration in rising temperatures during this 15-year period is sometimes referred to as a “pause” or “hiatus” in global warming, and has raised questions about why the rate of surface warming on Earth has been markedly slower than in previous decades. Since levels of greenhouse gases have continued to rise throughout the period, some skeptics have argued that the recent pattern undercuts the theory that global warming in the industrial era has been caused largely by man-made emissions from the burning of fossil fuels.
Lovejoy’s new study concludes that there has been a natural cooling fluctuation of about 0.28 to 0.37 degrees Celsius since 1998 — a pattern that is in line with variations that occur historically every 20 to 50 years, according to the analysis. “We find many examples of these variations in pre-industrial temperature reconstructions” based on proxies such as tree rings, ice cores, and lake sediment, Lovejoy says. “Being based on climate records, this approach avoids any biases that might affect the sophisticated computer models that are commonly used for understanding global warming.”
What’s more, the cooling effect observed between 1998 and 2013 “exactly follows a slightly larger pre-pause warming event, from 1992 to 1998,” so that the natural cooling during the “pause” is no more than a return to the longer term natural variability, Lovejoy concludes. “The pause thus has a convincing statistical explanation.”
The methodology developed in Lovejoy’s two recent papers could also be used by researchers to help analyze precipitation trends and regional climate variability and to develop new stochastic methods of climate forecasting, he adds.
—————————-
The paper:
“Return periods of global climate fluctuations and the pause”, Shaun Lovejoy, Geophysical Research Letters, published online July 14, 2014. DOI: 10.1002/2014GL060478
Abstract
An approach complementary to General Circulation Models (GCMs), using the anthropogenic CO2 radiative forcing as a linear surrogate for all anthropogenic forcings [Lovejoy, 2014], was recently developed for quantifying human impacts. Using preindustrial multiproxy series and scaling arguments, the probabilities of natural fluctuations at time lags up to 125 years were determined. The hypothesis that the industrial epoch warming was a giant natural fluctuation was rejected with 99.9% confidence. In this paper, this method is extended to the determination of event return times. Over the period 1880–2013, the largest 32 year event is expected to be 0.47 K, effectively explaining the postwar cooling (amplitude 0.42–0.47 K). Similarly, the “pause” since 1998 (0.28–0.37 K) has a return period of 20–50 years (not so unusual). It is nearly cancelled by the pre-pause warming event (1992–1998, return period 30–40 years); the pause is no more than natural variability.
Preprint paper here:
dbstealey says:
July 24, 2014 at 11:10 am
“Global warming stopped many years ago, but the alarmist crowd, exemplified by hanzo, routinely disregards that fact. Because if they admitted it, their entire argument would be debunked, and any rationale for continued spending on ‘climate studies’ would be highly questionable
That may be the case, but its not over yet. As many here will be aware, H2O is the key player in all of this. Roy Spencer has some interesting graphs here
http://www.drroyspencer.com/2014/07/june-2014-update-of-ssmi-ocean-products/
that show the changes in total H2O in the atmospheric column. Now which came first, chicken or egg is another matter, but the graphs seem to indicate more warming to come if the lag to the ocean heating follows normal behaviour.
Terry,
Dr Spencer says that there is no verified ‘fingerprint’ of greenhouse gases, therefore AGW remains a conjecture. If there is more warming, which is entirely possible, it is natural. The planet is still emerging from the LIA, but there is no evidence that CO2 has anything to do with it.
Re: the ‘chicken and egg’, T is the chicken, and CO2 is the egg. There is plenty of evidence that ∆T causes ∆CO2. But there is no evidence that changes in CO2 cause subsequent changes in temperature. Thus, AGW remains a conjecture [and CAGW is not even worth discussing; it is nonsense].
Edited better:
In hypothesis testing, it is good practice to formulate the hypothesis precisely and to design hypothesis tests that reduce the probability of type I and II errors.
A type I error occurs when the null hypothesis is true, but is rejected. A type II error occurs when the null hypothesis is false, but erroneously fails to be rejected.
A formulation of the AGW null hypothesis states that there is at most a minor relationship between atmospheric CO2 concentration and global mean surface air temperature (global mean SAT) relative to natural variation.
INITIAL FORMULATION: A null hypothesis TEST was proposed that measures the mean SAT over a selected 17 year period during which a coincident 25% increase in atmospheric CO2 occurred. RESULT: A SAT plateau was observed during the test period, which implied there was no correlation between CO2 and temperatures; so the null hypothesis was not rejected.
In devising the hypothesis test, the prudent scientist must consider how likely it would be for a set of observations to occur if the null hypothesis were true but erroneously rejected (type I error). Likewise, one must consider how likely it would be for a set of observations to occur if the null hypothesis were false, but erroneously not rejected (type II error). This analysis is important as it informs selection of sample size and design of of controls.
ERROR. The 17-year hypothesis test was flawed since there existed evidence that suggests a high probability of a type II error as demonstrated by the following paradoxical observations.
During the time between 1965-2013, there were eight 15-year periods where mean global SAT had plateaued according to a review of HADcrut3v data.
http://www.skepticalscience.com/Has-Global-Warming-Stopped.htmlhttp://www.skepticalscience.com/pics/SkepticFrame.jpg
Thus, the null hypothesis could be ACCEPTED multiple times depending on the ~15 year trends selected. However, the global mean SAT trend is positive and statistically significant over >30 year trends. Paradoxically, the null hypothesis could also be REJECTED based on any observed > 30 yr trends or any observed 30 year period during which a coincident increase in atmospheric CO2 occurs. RESULT: A statistically significant positive SAT trend was detected, showing an increase of 0.6 deg C/ century, which is 10x faster than any pre-industrial global mean proxy temp increase as far back as 800,000 years. This data is consistent with the correlation between CO2 and global mean temperatures; so the null hypothesis is rejected.
BETTER FORMULATION: A null hypothesis TEST is proposed that measures the mean SAT over a selected 99%).
http://www.physics.mcgill.ca/~gang/eprints/eprintLovejoy/neweprint/Anthro.climate.dynamics.13.3.14.pdf
Hanzo challenges anyone at WUWT to refute any or all aspects of the above analysis without conspiracy theory or ad hominem speculation.
(Please disregard previous post as use of greater/less than symbols results in omission of whole sentences)
In hypothesis testing, it is good practice to formulate the hypothesis precisely and to design hypothesis tests that reduce the probability of type I and II errors. A type I error occurs when the null hypothesis is true, but is rejected. A type II error occurs when the null hypothesis is false, but erroneously fails to be rejected.
THE FLAWED 17-YEAR TEST: A null hypothesis TEST was proposed that measures the mean SAT over a selected 17 year period during which a coincident 25% increase in atmospheric CO2 occurred. RESULT: A SAT plateau was observed during the test period, which implied there was no correlation between CO2 and temperatures; so the null hypothesis was not rejected. The 17-year hypothesis test was flawed since there existed evidence that suggests a high probability of a type II error as demonstrated by the following paradoxical observations.
PARADOX. During 1965-2013, there were eight 15-year periods where mean global SAT had plateaued according to a review of HADcrut3v data.
http://www.skepticalscience.com/Has-Global-Warming-Stopped.html
http://www.skepticalscience.com/pics/SkepticFrame.jpg
Thus, the null hypothesis could be apparently ACCEPTED multiple times depending on the particular ~15 year trends selected. Or paradoxically, it could be REJECTED based on any observed greater than 30 yr trends or based on any observed lesser than 30 year trend where known sources of natural variation were removed (volcanic aerosols, ENSO, AMO, PDO). The null hypothesis cannot be both accepted AND rejected. Thus, trend duration AND known sources to natural variation (noise subtraction) MUST be considered in test as it is a scientific expectation to reduce the probability of either type I or II errors.
BETTER FORMULATION: A hypothesis TEST is proposed that measures the mean SAT over a selected greater than 30 year period during which a coincident increase in atmospheric CO2 occurs, or measured over a selected less than 30 year period where known short term natural variations that augment or dampen the mean global SAT are removed.
RESULT: A statistically significant positive SAT trend was detected, showing an increase of + 0.6 deg C/ century for greater than 30 year datasets (10x faster than any pre-industrial global mean proxy temp increase as far back as 800,000 years). For a less than 30 year dataset, a positive SAT trend was also detected and showed an increase of +0.16 deg C per decade. This data is consistent with the correlation between CO2 and global mean temperatures; so the null hypothesis is rejected.
Hanzo challenges anyone at WUWT to refute any or all aspects of the above analysis without conspiracy theory or ad hominem speculation.
hanzo says:
Hanzo challenges anyone at WUWT to refute any or all aspects of the above analysis without conspiracy theory or ad hominem speculation.
I challenge ‘hanzo’ to admit that global warming stopped many years ago.
But he won’t. hanzo preposterously believes that global warming is continuing. He has no understanding of the null hypothesis, which, as Dr Roy Spencer states, “…has never been falsified.”
Who to believe? Dr Spencer? Or the anonymous ‘hanzo’, who obvioulsy feeds at the taxpayer trough.
And hanzo might start to get some credibility if he stops linking to that incredible blog, SS.
So there is my challenge: admit that global warming has stopped. Claiming that global warming is continuing as usual brands ‘hanzo’ as a lunatic.
katatetorihanzo:
At July 26, 2014 at 9:48 am you say
That is no “challenge” because it is so very, very easy to do.
Your “analysis” is complete nonsense because it is based on your total lack of any understanding of the scientific method; e.g. your “analysis” confuses hypothesis testing and the Null Hypothesis.
See how easy that was? All aspects of your so-called “analysis” were accurately and factually refuted by a single, true sentence which contained no conspiracy theory, no ad homs., and no speculation.
Richard
R. Courtney said: “your “analysis” confuses hypothesis testing and the Null Hypothesis.”
Please elaborate. Clear evidence in the 130 year SAT dataset compels a test design that reduces probability of Type II error. Need greater than 30 yr global SAT data to allow known non-trending natural fluctuations to cancel out or apparent accelerated or dampened temperature trend may give misleading results. Alternatively, known short-term natural variations must be subtracted from observed trend to reveal magnitude of unnatural or unknown component.
http://en.m.wikipedia.org/wiki/Trend_estimation#Data_as_trend_plus_noise
Please refute any or all aspects of the above analysis without conspiracy theory or ad hominem speculation.
Until hanzo admits that global warming is currently halted, he has no credibility.
“Until hanzo admits that global warming is currently halted, he has no credibility.”
The current evidence doesn’t support such a claim any more than it did at earlier plateau periods. We still have a measurable radiative imbalance at the top if the atmosphere and the 1st law of thermodynamics predicts planetary heat build up. With global ice mass and arctic sea albedo diminishing, ocean heat content and sea level rising without pause, it’s just too premature to assume that what started 133 years ago has suddenly stopped 17 years ago with no reason put forward. Feel free to suggest an alternative scientific explanation why anyone should not be concerned.
“Until hanzo admits that global warming is currently halted, he has no credibility.”
The current evidence doesn’t support such a claim any more than it did at earlier plateau periods. We still have a measurable radiative imbalance at the top if the atmosphere and the 1st law of thermodynamics predicts planetary heat build up. With global ice mass and arctic sea albedo diminishing, ocean heat content and sea level rising without pause, it’s just too premature to assume that what started 133 years ago has suddenly stopped 17 years ago with no reason put forward. Feel free to suggest an alternative scientific explanation why anyone should not be concerned.
Why did temperatures in the 1930’s go up before co2 levels really began to rise?
katatetorihanzo:
At July 26, 2014 at 2:12 pm you ask me
I DID in my post at July 26, 2014 at 10:12 am. The refutation says
And your response to it demonstrates the truth of that refutation.
Richard
http://www.oism.org/pproject/s33p36.htm
http://www.oism.org/pproject/Slides/Presentation/Slide3.png
http://www.oism.org/pproject/Slides/Presentation/Slide3.png
Everyone refutes hanzo’s wrong-headed belief that global warming is continuing. It is not. As usual, hanzo keeps bringing up irrelevant arguments. Why? To muddy the waters. “Say anything” is an alarmist tactic, used when they lack honest, testable evidence to support their anti-science.
Satellite data — the most accurate there is — shows conclusively that global warming stopped many years ago. But rent-seekers like hanzo don’t want to accept reality. His income depends on his promotion of the lie that global warming is continuing. Despicable.
=========================
Russ Browne,
Thanks for the links. They easily refute hanzo’s nonsense. The OISM site makes monkeys out of those promoting climate alarmism, for which there is zero scientific evidence. Alarmists are self-serving purveyors of pseudo science. They are sounding a false alarm, for which there is no measurable, testable scientific evidence. They are lying for money. Shame on them.