By Christopher Monckton of Brenchley
The prolonged el Niño of 2016-2017, not followed by a la Niña, has put paid to the great Pause of 18 years 9 months in global warming that gave us all such entertainment while it lasted. However, as this annual review of global temperature change will show, the credibility gap between predicted and observed warming remains wide, even after some increasingly desperate and more or less openly prejudiced ever-upward revisions of recent temperatures and ever-downward depressions in the temperatures of the early 20th century in most datasets with the effect of increasing the apparent rate of global warming. For the Pause continues to exert its influence by keeping down the long-run rate of global warming.
Let us begin with IPCC’s global warming predictions. In 2013 it chose four scenarios, one of which, RCP 8.5, was stated by its authors (Riahi et al, 2007; Rao & Riahi, 2006) to be a deliberately extreme scenario and is based upon such absurd population and energy-use criteria that it may safely be ignored.
For the less unreasonable, high-end-of-plausible RCP 6.0 scenario, the 21st-century net anthropogenic radiative forcing is 3.8 Watts per square meter from 2000-2100:
CO2 concentration of 370 ppmv in 2000 was predicted to rise to 700 ppmv in 2100 (AR5, fig. 8.5) on the RCP 6.0 scenario (thus, the centennial predicted CO2 forcing is 4.83 ln(700/370), or 3.1 Watts per square meter, almost five-sixths of total forcing). Predicted centennial reference sensitivity (i.e., warming before accounting for feedback) is the product of 3.8 Watts per square meter and the Planck sensitivity parameter 0.3 Kelvin per Watt per square meter: i.e., 1.15 K.
The CMIP5 models predict 3.37 K midrange equilibrium sensitivity to CO2 doubling (Andrews+ 2012), against 1 K reference sensitivity before accounting for feedback, implying a midrange transfer function 3.37 / 1 = 3.37. The transfer function, the ratio of equilibrium to reference temperature, encompasses by definition the entire operation of feedback on climate.
Therefore, the 21st-century warming that IPCC should be predicting, on the RCP 6.0 scenario and on the basis of its own estimates of CO2 concentration and the models’ estimates of CO2 forcing and Charney sensitivity, is 3.37 x 1.15, or 3.9 K.
Yet IPCC actually predicts only 1.4 to 3.1 K 21st-century warming on the RCP 6.0 scenario, giving a midrange estimate of just 2.2 K warming in the 21st century and implying a transfer function of 2.2 / 1.15 = 1.9, little more than half the midrange transfer function 3.37 implicit in the equilibrium-sensitivity projections of the CMIP5 ensemble.
Note that Fig. 2 disposes of any notion that global warming is “settled science”. IPCC, taking all the scenarios and hedging its bets, is predicting somewhere between 0.2 K cooling and 4.5 K warming by 2100. Its best estimate is its RCP 6.0 midrange estimate 2.2 K.
Effectively, therefore, given 1 K reference sensitivity to doubled CO2, IPCC’s 21st-century warming prediction implies 1.9 K Charney sensitivity (the standard metric for climate-sensitivity studies, which is equilibrium sensitivity to doubled CO2 after all short-acting feedbacks have operated), and not the 3.4 [2.1, 4.7] K imagined by the CMIP5 models.
Since official predictions are thus flagrantly inconsistent with one another, it is difficult to deduce from them a benchmark midrange value for the warming officially predicted for the 21st century. It is somewhere between the 2.2 K that IPCC gives as its RCP 6.0 midrange estimate and the 3.9 K deducible from IPCC’s midrange estimate of 21st-century anthropogenic forcing using the midrange CMIP5 transfer function.
So much for the predictions. But what is actually happening, and does observed warming match prediction? Here are the observed rates of warming in the 40 years 1979-2018. Let us begin with GISS, which suggests that for 40 years the world has warmed at a rate equivalent not to 3.9 C°/century nor even to 2.2 C°/century, but only to 1.7 C°/century.
Next, NCEI. Here, perhaps to make a political point, the dataset is suddenly unavailable:
Next, HadCRUT4, IPCC’s preferred dataset. The University of East Anglia is rather leisurely in updating its information, so the 40-year period runs from December 1978 to November 2018, but the warming rate is identical to that of GISS, at 1.7 C°/century equivalent, below the RCP 6.0 midrange 2.2 C°/century rate.
Next, the satellite lower-troposphere trends, first from RSS. It is noticeable that, ever since RSS, whose chief scientist publicly describes those who disagree with him about the climate as “deniers”, revised its dataset to eradicate the Pause, it has tended to show the fastest apparent rate of global warming, now at 2 C°/century equivalent.
Finally, UAH, which Professor Ole Humlum (climate4you.com) regards as the gold standard for global temperature records. Before UAH altered its dataset, it used to show more warming than the others. Now it shows the least, at 1.3 C°/century equivalent.
How much global warming should have occurred over the 40 years since the satellite record began in 1979? CO2 concentration has risen by 72 ppmv. The period CO2 forcing is thus 0.94 W m–2, implying 0.94 x 6/5 = 1.13 W m–2 net anthropogenic forcing from all sources. Accordingly, period reference sensitivity is 1.13 x 0.3, or 0.34 K, and period equilibrium sensitivity, using the CMIP5 midrange transfer function 3.37, should have been 1.14 K. Yet the observed period warming was 0.8 K (RSS), 0.7 K (GISS & HadCRUT4) or 0.5 K (UAH): a mean observed warming of about 0.7 K.
A more realistic picture may be obtained by dating the calculation from 1950, when our influence first became appreciable. Here is the HadCRUT4 record:
The CO2 forcing since 1950 is 4.83 ln(410/310), or 1.5 Watts per square meter, which becomes 1.8 Watts per square meter after allowing for non-CO2 anthropogenic forcings, a value consistent with IPCC (2013, Fig. SPM.5). Therefore, period reference sensitivity from 1950-2018 is 1.8 x 0.3, or 0.54 K, while the equivalent equilibrium sensitivity, using the CMIP5 midrange transfer function 3.37, is 0.54 x 3.37 = 1.8 K, of which only 0.8 K actually occurred. Using the revised transfer function 1.9 derived from the midrange predicted RCP 6.0 predicted warming, the post-1950 warming should have been 0.54 x 1.9 = 1.0 K.
It is also worth showing the Central England Temperature Record for the 40 years 1694-1733, long before SUVs, during which the temperature in most of England rose at a rate equivalent to 4.33 C°/century, compared with just 1.7 C°/century equivalent in the 40 years 1979-2018. Therefore, the current rate of warming is not unprecedented.
It is evident from this record that even the large and naturally-occurring temperature change evident not only in England but worldwide as the Sun recovered following the Maunder minimum is small compared with the large annual fluctuations in global temperature.
The simplest way to illustrate the very large discrepancy between predicted and observed warming over the past 40 years is to show the results on a dial.
Overlapping projections by IPCC (yellow & buff zones) and CMIP5 (Andrews et al. 2012: buff & orange zones) of global warming from 1850-2011 (dark blue scale), 1850 to 2xCO2 (dark red scale) and 1850-2100 (black scale) exceed observed warming of 0.75 K from 1850-2011 (HadCRUT4), which falls between the 0.7 K period reference sensitivity to midrange net anthropogenic forcing in IPCC (2013, fig. SPM.5) (cyan needle) and expected 0.9 K period equilibrium sensitivity to that forcing after adjustment for radiative imbalance (Smith et al. 2015) (blue needle). The CMIP5 models’ midrange projection of 3.4 K Charney sensitivity (red needle) is about thrice the value consistent with observation. The revised interval of global-warming predictions (green zone), correcting an error of physics in models, whose feedbacks do not respond to emission temperature, is visibly close to observed warming.
Footnote: I undertook to report on the progress of my team’s paper explaining climatology’s error of physics in omitting from its feedback calculation the observable fact that the Sun is shining. The paper was initially rejected early last year on the ground that the editor of the top-ten journal to which it was sent could not find anyone competent to review it. We simplified the paper, whereupon it was sent out and, after many months’ delay, only two reviews came back. The first was a review of a supporting document giving results of experiments conducted at a government laboratory, but it was clear that the reviewer had not actually read the laboratory’s report, which answered the question the reviewer had raised. The second was ostensibly a review of the paper, but the reviewer stated that, because he found the paper’s conclusions uncongenial he had not troubled to read the equations that justified those conclusions.
We protested. The editor then obtained a third review. But that, like the previous two reviews, was not a review of the present paper. It was a review of another paper that had been submitted to a different journal the previous year. All of the points raised by that review had long since been comprehensively answered. None of the three reviewers, therefore, had actually read the paper they were ostensibly reviewing.
Nevertheless, the editor saw fit to reject the paper. Next, the journal’s management got in touch to say that it was hoped we were content with the rejection and to invite us to submit further papers in future. I replied that we were not at all satisfied with the rejection, for the obvious reason that none of the reviewers had actually read the paper that the editor had rejected, and that we insisted, therefore, on being given a right of appeal.
The editor agreed to send out the paper for review again, and to choose the reviewers with greater care this time. We suggested, and the editor accepted, that in view of the difficulty the reviewers were having in getting to grips with the point at issue, which was clearly catching them by surprise, we should add to the paper a comprehensive mathematical proof that the transfer function that embodies the entire action of feedback on climate is expressible not only as the ratio of equilibrium sensitivity after feedback to reference sensitivity before feedback but also as the ratio of the entire, absolute equilibrium temperature to the entire, absolute reference temperature.
We said we should explain in more detail that, though the equations for both climatology’s transfer function and ours are valid equations, climatology’s equation is not useful because even small uncertainties in the sensitivities, which are two orders of magnitude smaller than the absolute temperatures, lead to large uncertainty in the value of the transfer function, while even large uncertainties in the absolute temperatures lead to small uncertainty in the transfer function, which can thus be very simply and very reliably derived and constrained without using general-circulation models.
My impression is that the editor has realized we are right. We are waiting for a new section from our professor of control theory on the derivation of the transfer function from the energy-balance equation via a leading-order Taylor-series expansion. That will be with us at the end of the month, and the editor will then send the paper out for review again. I’ll keep you posted. If we’re right, Charney sensitivity (equilibrium sensitivity to doubled CO2) will be 1.2 [1.1, 1.3] C°, far too little to matter, and not, as the models currently imagine, 3.4 [2.1, 4.7] C°, and that, scientifically speaking, will be the end of the climate scam.