The Great Climate Change Debate: William Happer v. David Karoly, Part B

By Andy May

In Part A of the Great Debate series (see here) we discussed Dr. David Karoly’s and Dr. William Happer’s arguments regarding how unusual the recent global warming is and how we know the recent observed increase in CO2 is due to human activities. In Part B we will examine their thoughts on questions three and four.

3. How do we know that the increase in CO2 and other greenhouse gases in the atmosphere have caused most of the recent global warming?

4. Climate models have been used to compute the amount of warming caused by human activities, how accurate are they?

For an introduction to the debate and links to the original documents see Part A.

3. How do we know that the increase in CO2 and other greenhouse gases in the atmosphere have caused most of the recent global warming?

Karoly’s first argument is an appeal to the “consensus.”

“Nevertheless, an overwhelming consensus of climate scientists agree that human-caused climate change is happening, and that global warming will continue throughout the current century, with many adverse impacts on human and natural systems.” From the Karoly statement.

Happer’s detailed response contains his view of “consensus science.”

“Truth has never been determined by “an overwhelming consensus,” and in fact, consensuses have often been completely wrong.” From Happer’s detailed response.

In Karoly’s statement, he goes straight into a discussion of how we know much of the increase in CO2 concentration in the atmosphere is from humans (see previous question in Part A, note Happer agrees the increase in CO2 is due to human activities) but provides no more evidence that CO2 and other greenhouse gases have caused most of the recent warming. For the calculation of the greenhouse gas contribution, he turns to climate models.

How computer models are used to separate natural warming from human-caused warming:

“… climate model simulations have been used to assess the relative importance of different forcing factors on the climate system and how well they explain the observed global warming. The simulations are driven by natural forcing factors, such as changes in solar radiation and volcanic aerosols, as well as human-caused changes in greenhouse gases and human activity-related climate forcing factors, including industrial aerosols and land use change.” From the Karoly statement.

Karoly refers to a whisker plot from the IPCC AR5 report (Figure 3) which shows the impact of human and natural “climate forcings” as computed by the climate models from 1951-2010.

Figure 3. The assumed temperature changes due to GHG (greenhouse gases), ANT (anthropogenic), OA (other anthropogenic), NAT (natural forces), and INT (internal climate variablity) from 1951 to 2010. Source Figure 10.5 from (IPCC 2013) page 884.


Karoly also provides the following illustration of the reasons we “know” humans are causing global warming (Figure 4).

Figure 4. Reasons that we know humans are causing [some] global warming. Source


Most of the reasons given are related to human CO2 emissions, which both Happer and Karoly agree are the main cause of the rising CO2 concentration in the atmosphere and oceans. The principal disagreement between them is how much warming is caused by CO2. The tropopause rises as the surface warms, this is not controversial. Karoly believes that since winter is warming more than summer and nights more than days, this excludes solar variability as a cause of the warming, perhaps this is so. But, none of the reasons given exclude internal variability (mainly ocean cycles).

The cooling upper atmosphere (stratosphere) could be due to more CO2, since CO2 radiates a lot of infrared radiation (IR) to space from the stratosphere. It could also be due to increased shielding of thermal radiation from below due to additional IR absorption in the troposphere. But, regardless of the cause, there is a noticeable correlation between the observed tropospheric warming and the stratospheric cooling in the atmosphere. For a discussion of the significance of the correlation, displays of the data and additional references see Munshi, 2018 (Munshi 2018). Thus, the essence of the debate is not if CO2 warms the lower atmosphere, the argument is about how much warming is due to CO2 and other greenhouse gases. More detail on Figure 4 and references can be seen at this link by John Cook.

Happer points out at the beginning of his statement that CO2 levels today are much lower than during almost any time in the history of life on Earth. Figure 5 shows the history of CO2 levels for the past 550 million years.

Figure 5. History of CO2 levels for the past 550 million years. Source: (Berner and Kothavala 2001).


Happer writes this about Figure 5:

“The important message of Fig. [5] is that CO2 concentrations have been much higher than present values over most of the history of life. Even though CO2 concentrations were measured in thousands of parts per million by volume (ppm) over most of the Phanerozoic, not the few hundred ppm of today, life flourished in the oceans and on the land. Average pH values in the ocean surface were as low as pH = 7.7, a bit lower than the pH = 8.1 today. But this was still far from acidic, pH < 7, because of the enormous natural alkalinity of seawater. The mean global temperature was sometimes higher and sometimes lower than today’s. But the temperature did not correlate very well with CO2 levels. For example, there were ice ages in the Ordovician, some 450 million years ago, when the CO2 levels were several thousand ppm.” (Berner and Kothavala 2001) and (Quinton and MacLeod 2014) From Happer’s statement.

Tamblyn writes the following about the same graph and a separate, but similar one in his detailed response:

“Lest we wonder how the Ordovician ice age could have occurred … GEOCARB III (which Professor Happer also references) is a geochemical model, which estimates past CO2 levels from the chemistry of rocks. Its calculations are run over steps of 10 million years and averaged over 50 million years. It is not sensitive enough to detect shorter-term changes. Direct geological evidence shows that CO2 levels fell sharply during that period, in 1 to 2 million years or less — too small for GEOCARB III to capture. A higher-resolution geochemical model applied just to this period suggests a decline of CO2 levels from ~5000 to 3000 ppm. With differences in solar output, 3000 ppm then is equal to 500 ppm today. Climate models applied to late Ordovician conditions predict icehouse conditions at CO2 levels below about 2240 to 3920 ppm.

The explanation is that changes in the greenhouse effect over time, due to changes primarily in CO2 and methane levels, explain much of the observed climate history. They aren’t the only factors. Orbital cycles influence how much sunlight the Earth receives. The extent of snow and ice changes how reflective the Earth is. The position of the continents matters (land reflects more sunlight than the oceans), and continents shape the flow of ocean currents, influencing heat transport. For around 30 million years, Antarctica has been isolated and surrounded by a huge current, the Antarctic Circumpolar Current, promoting cooling. For many, hundreds of millions of years before, this wasn’t so.” Tamblyn detailed response.

So, 3000 ppm in the Ordovician ice age is equal to 500 ppm today, yet we are currently at 400 ppm? Tamblyn’s point about resolution is valid but does not prove his case and his numbers are inconsistent with his thesis. He lists many valid influences on climate but fails to make the case that CO2 is the dominant influence. The quoted passage appears to weaken Tamblyn’s case and strengthen Happer’s.

4. Climate models have been used to compute the amount of warming caused by human activities, how accurate are they?

Karoly explains that natural forces are separated from human influences on climate with more than 30 computer models that compare well to observations since 1860 and writes:

“The observed significant cooling for one to two years after major volcanic eruptions —Santa Maria (1903), Agung (1963), El Chicon (1982), and Pinatubo (1991) —is simulated very well. The observed global mean temperature variations throughout the whole period lie within the range of all the model simulations with combined forcings, indicating the models simulate well the chaotic interannual variability of global mean temperature. There is very good agreement between the observed long-term global warming since the late nineteenth century and the average global warming across all the model simulations for combined natural and anthropogenic forcing.” From the Karoly statement.

He shows the following illustrations (Figure 6) from IPCC AR5 (IPCC 2013):

Figure 6. (a) The ensemble mean of numerous CMIP3 and CMIP5 climate models are shown in blue and red and compared to observations shown in black. The range of modeled results is shown with yellow and gray shading. The minimum warming in (a) from the models can be almost a degree C lower than the maximum model value for any given date and the error increases with time. This display models both human-caused and natural warming. (b) Same as (a) but only natural warming is simulated. (c) Same as (a) but only greenhouse gas warming is simulated. (c) suggests that net forcing (except for greenhouse gases) is negative, since natural forcings are assumed to be zero by the IPCC, they claim the negative forcings are “other anthropogenic.” The temperature anomalies are computed relative to the mean from 1880-1919. Source IPCC AR5, figure 10.1 (IPCC 2013) page 879.

Thus, the evidence for the magnitude of the human impact on climate is entirely model-based and direct measurements of the impact have not been made. Critics also point out that the warming from 1910 to 1945 and 2000 to 2012 are poorly matched by the models. Karoly responds:

“Of course, a small number of scientists say that the climate models are tuned to simulate the recent observed warming but are unreliable for projecting future warming trends. Others say that they show too much global warming, because the observed warming from 1998 to 2010 was very small, while the simulated warming continued, if you consider the average across all the climate model simulations. As shown already when considering the observed global mean temperature variations, there is large natural variability in global mean temperature in the observations and the models. The observed departure in 2010 from the multi-model mean is no larger than in 1910 or in 1940 and is well within the envelope of all the model simulations.” From the Karoly statement.

Thus, the poor reproduction of 1910-1945 and 2000-2012 is acknowledged, but in Karoly’s opinion the mismatch is acceptable. The error from 1910 to 1945 ranges from -0.2° to +0.4°C, approximately equal to the warming from 1975 to 2010, but human emissions of CO2 were not considered (or modeled) to be significant before 1951. How interesting that both sides use the poor model match from 1910-1945 in their arguments, but with opposite intent.

Happer has the following to say about using climate models to “prove” global warming since 1951 is primarily due to human activity:

“I disagree. This statement is based on excessive faith in computer models. The wide availability of computers and powerful software to make color displays has been a serious problem, since it has blurred the lines between reality and virtual reality. These are not the same. In my Statement and Interview I tried to stick to real satellite pictures of visible and thermal radiation from the Earth, real measurements of ocean pH, real records of tornados, hurricanes, floods, droughts, etc. Essentially all of Dr. Karoly’s claims of warming from greenhouse gases come from computer models, with lurid, threatening reds to represent the supposedly harmful effects of the demon gas, CO2.” Happer’s detailed response.

What is coming next?

The answers the scientists give to these questions highlight a significant and key difference in interpretation. Karoly is comfortable with the results from the climate models and Happer is not. As a result, Karoly is comfortable with the model-based conclusion that CO2 and other greenhouse gases are responsible for most of recent warming. Happer is not convinced the models are accurate and provides his own calculations and observations that show far less warming from CO2. Part C of this series will discuss their thoughts on the accuracy of future temperature estimates. It examines questions five and six. These are:

  • How do we know global warming and more CO2 will have substantial adverse impacts on humans and the planet?
  • Should anything be done to combat global warming?

Equilibrium Climate Sensitivity (ECS) to a doubling of the CO2 concentration has not been measured and is poorly estimated, the IPCC range of possible values is 1.5 to 4.5 and it has stayed at this range since 1979. How much difference does this range make?

Works Cited

Berner, Robert, and Zavareth Kothavala. 2001. “Geocarb III: A Revised Model of Atmospheric CO2 over Phanerozoic Time.” American Journal of Science 301 (2): 182-204.

IPCC. 2013. In Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, by T. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley. Cambridge: Cambridge University Press.

Munshi, Jamal. 2018. “Climate Change, Tropospheric Warming, and Stratospheric Cooling.” SSRN (SSRN).

Quinton, P., and K. MacLeod. 2014. “Oxygen isotopes from conodont apatite of the midcontinent US: Implications for Late Ordovician climate evolution.” Palaeogeography, Palaeoclimatology, Palaeoecology 404: 57-66.