By Andy May
Christian Freuer has translated this post into German here.
The previous parts of this series investigated model bias in the CMIP6 models and in their interpretation in AR6 WGI. This part looks at model bias in AR6 WGII, Climate Change 2022: Impacts, Adaptation, and Vulnerability.[1] The IPCC WGII report uses the possible future climate projections from the WGI report to project the future impact of climate change on society. It uses socio-economic models to accomplish this. As we saw in the previous parts of this series, the WGI report is biased and ignores possible natural contributions to recent observed global warming from changes in the Sun, cloud cover, and the meridional transport of energy.
The WGI/CMIP6 models, rather arbitrarily, assign all warming since 1750 to human influences, particularly CO2 emissions.[2] WGII accepts this controversial conclusion. It uses projected CO2 emissions combined with the WGI/CMIP6 models to predict future temperature and projected knock-on effects to other climate components, like precipitation, to model the future impact on human civilization.
WGII states that:
“Human-induced climate change, including more frequent and intense extreme events, has caused widespread adverse impacts and related losses and damages to nature and people, beyond natural climate variability.”[3]
AR6 WGII, page 9
This is only true if we accept their assumption about the range of natural climate variability, but as we saw in the previous parts of this series, their assumptions about natural warming, especially the impact of solar variability, are very controversial. Further, whether climate change is natural or human-caused, someone, somewhere, is nearly always going to be adversely affected by a change in climate, while others will benefit from the same change. How widespread is “widespread?”
WGII liberally discusses the potential negative impact of climate change,[4] and they discuss the potential benefits of their recommended adaptation and mitigation policies, but the report rarely mentions the well documented potential benefits of global warming and additional atmospheric CO2. [5] The fact that WGII only considers the problems of climate change and not the benefits, reveals their bias and invalidates their analysis. Even when mentioning a benefit, they find something negative in it. For example, they mention that elevated CO2 benefits woody plants, but that woody plants can cause an increase in atmospheric carbon.[6]
As Brian O’Neill writes, while many studies anticipate problems in the future, they also predict a future where humanity is better educated, better fed, longer lived, healthier, with less poverty, and less conflict. This is simply continuing a trend that has been underway for many decades.[7] O’Neill reports that currently there are 700-800 million people at risk of hunger globally. By 2050, even including the possible effects of 2°C of warming, that number will fall to 250 million.[8]
Currently the world’s economy is growing between 2 and 3% per year[9] and this is not expected to change much in the future. Looking ahead at a possible 2.5°C of warming in the next century or so, economists anticipate between a positive net climate change impact of about 2% and negative net impact of about 2.5% on global GDP. It is significant that the sign of the net economic impact due to climate change is not known. The average impact for 2.5°C of warming is a negative 1.3% for the average person.[10] In the next 80 years global GDP would be expected to grow between 487% and 1,000%, so a negative 1.3% due to climate change is unlikely to be noticed. Richard Tol writes that the uncertainty in the estimates of the impact of climate change on total economic welfare is very large and if we take this uncertainty into account, the impact of climate change does not significantly deviate from zero until 3.5°C of warming.[11]
Emissions and impact scenarios
The future cannot be predicted. So, the concept of “scenarios” was developed in the 1960s by Herman Kahn, a military strategist with the RAND Corporation.[12] The idea is to develop a “business as usual” forecast that assumes no unusual events occur over the planning period. Then you vary something and compute an alternative forecast that shows the difference between the baseline, business-as-usual, forecast and your model. It is just a learning tool and like all models, used to investigate the possible impact of policy changes, regulations, or tactical decisions in wars or battles. We are not supposed to believe any of the forecasts, it is just the relative values between various assumptions that are important. Scenario analysis is widely used to do cost-benefit analysis. However, since WGII only incorporates the costs and leaves out the benefits, their cost-benefit analysis is invalid.
It is very important to remember that the projections used in WGII assume that there will be no natural warming or cooling between now and 2100. If there are natural forces acting on climate, then the greenhouse gas-based projections they rely upon will be wrong and their projected impacts on human civilization must be wrong as well. The AR6 scenarios of temperature change relative to 1850 to 1900 are shown in figure 1.
Hausfather and Peters[13] have called the higher scenarios, SSP3-7.0 and SSP5-8.5 (as well as their AR5 equivalent RCP8.5) unlikely, but since this view is contested,[14] AR6 WGII takes no position on which of the scenarios in figure 1 is most likely.[15] This is unfortunate since the difference in the scenarios in 2100, only 76 years from today, is over three degrees. The combination of the uncertainty in the projected warming and in the potential impact of the warming is extremely large.
Roger Pielke Jr. and Justin Ritchie tell us that the ancestor of the SSP5-8.5 scenario in figure 1 originated in the first IPCC report in 1990. In 1990, with what was known then, it was a reasonable “business-as-usual” scenario. It predicted a large increase in coal consumption and a CO2 concentration of 1,200 PPM in 2100. Today that emissions scenario is reached in SSP5-8.5, but with what we know today it is not “business-as-usual,” in fact it is an implausible future, that is becoming more impossible with each passing year.[16] To be fair, the IPCC does not call SSP5-8.5 business-as-usual, that label is used by others, presumably because that is what it is called in the first report in 1990.[17]
Marcel Crok reports in the book that he and I edited, The Frozen Climate Views of the IPCC, that the unlikely, and now implausible, SSP5-8.5 and its predecessor RCP8.5 are mentioned in AR6 41.5% of the time according to Roger Pielke Jr., much more than the more likely SSP2-4.5 or RCP4.5 scenarios (mentioned 17% of the time). The latter two scenarios more closely match recent observations.[18] Thus, WGII often uses the biased and too hot WGI models as input to maximal and implausible emissions scenarios to do their modeled climate impact projections.
Ignoring the Good News
While using implausible scenarios and biased climate model results in assessing the impacts of climate change is unwise, ignoring the positive impacts of climate change and focusing only on the bad may well be worse. The whole idea of using scenarios is to investigate the full range of possible outcomes, not cherry-pick the model input to manufacture a desired outcome, a problem often called reporting bias. It is this part of the WGII procedure that cost them credibility.
Marcel Crok shows us that U.S. major and all landfalling hurricanes have been declining since 1900.[19] Globally, there is no trend in cyclones and hurricanes.[20] There is also no trend in accumulated global cyclone energy.[21] AR6 WGI finds that since 1950 there has been an increase in the number of hot days and heatwaves,[22] but as figure 1 in part 2 shows the world was cooling in 1950. At least in the United States, records show that peak hot days and heatwaves were in the 1930s.[23] AR6 WGI also finds that there is “low confidence in general statements to attribute changes in flood events to anthropogenic climate change.”[24] The idea that extreme weather is increasing globally is very controversial.
It is worth noting that AR6 WGII states that they have high confidence that some extreme weather is increasing as a result of climate change, including extreme rainfall events, more frequent and stronger cyclones/hurricanes, and that recent devastating floods were made more likely due to climate change.[25] This appears to be directly contradicted by what is stated in AR6 WGI, but WGII cleverly sidesteps the contradiction by specifying “Some extreme weather…” and “devastating floods in western Europe…” Thus, to make their point, they cherry pick locations and events and avoid discussing global impacts that have not changed or are decreasing.[26] In any given year, extreme weather events are increasing somewhere, that is the nature of weather. Their assertion is contradicted by the work of Zhongwei Yan, Philip Jones, and Anders Moberg already mentioned in part 5.[27]
Finally, both WGI and WGII completely ignore evidence that global warming and additional CO2 have many benefits. Bjorn Lomborg reports that human welfare will likely increase 450% in the 21st century and damages due to climate change might reduce this to 434%,[28] which will be hard for most people to detect. Lomborg also finds that non-climate-related deaths, due to earthquakes, tsunamis, volcanoes, etc. have fallen only slightly in the past 100 years, but climate-related deaths have fallen a staggering 99%. Part of this is that cold-related deaths are much more common than heat-related deaths, and as the world warms, cold-related deaths fall more than heat-related deaths increase.[29]
Cherry picking
The authors of AR6 WGII were particularly guilty of selecting papers to discuss that supported their assumptions and ignoring papers that refuted or disagreed with them. In a classic case they discussed Grinsted, et al.,[30] which claims to be able to attribute some U.S. hurricane losses to human-caused global warming. Grinsted is the only paper, out of many[31] that was able to attribute hurricane losses to human-caused or human-enhanced hurricane activity. However, Roger Pielke Jr. has found that the paper is flawed and has requested that it be retracted.[32]
Even though the paper is likely flawed and is contradicted by many other studies, it is used to support the idea that some U.S. hurricane losses can be “partly attributed to anthropogenic climate change” in AR6 WGII.[33] To be fair, they do mention one of the many studies that disagree with Grinsted. However, they also mention one other paper, Estrada et al.,[34] that they imply supports attribution to human-caused climate change, but the paper does not say that. Estrada, et al. say that their results are ambiguous, and that in 2005 2-12% of normalized losses “could be attributable to climate change.” So, they chose one year, and only considered the United States, and maybe 2-12% of the damage was due to climate change. In Estrada’s conclusions they note:
“Increases in wealth and population alone cannot account for the observed trend in hurricane losses. The remaining trend in itself does not prove the existence of a climate change signal, as it could be due to causes not considered here.”
Estrada, Botzen, and Tol, Nature Geoscience, 2015
In other words, they detect a trend in normalized hurricane damage that cannot be fully explained by increasing wealth and population and it is possible that this excess is due to climate change. Estrada, et al. explain that prominent ocean oscillations, such as the Atlantic Multidecadal Oscillation (AMO) can account for some of the excess hurricane damage observed. Also, data problems prior to 1940 could produce a spurious upward trend in damage. So, Estrada, et al.’s analysis uncovered a small excess trend in damage that might be explainable by climate change but could also be caused by other factors. Not very convincing.
AR6 WGII leaves the reader with the idea it is two against one, when actually one of the pro-attribution studies is inconclusive and they ignored a large number of studies that found no connection between hurricane damage and climate change. WGII does make the following statement, which partially absolves them:
“Climate change explains a portion of long-term increases in economic damages of hurricanes (limited evidence, low agreement).”[35]
IPCC AR6 WGII, page 1978
They are saved by the “limited evidence, low agreement” bit, but somehow that part is always left out of the press releases and news media.
WGII Model Bias, Summary
Just as WGI ignored the potential impact of solar variability and changes in meridional transport, WGII ignored the potential benefits of warming and additional atmospheric CO2. This invalidates the report. By ignoring the well-documented benefits of global warming and additional CO2, they clearly cannot assess the impact of climate change or our vulnerability to climate changes. It makes their report useless for policy making or cost-benefit analysis.
It is hard to decide exactly how to characterize this problem in AR6 WGII, it could be described as reporting bias, since they ignored so many studies that report warming and CO2 benefits. It could also be described as confirmation bias given their stated assumption that warming and additional CO2 is a bad thing. But, either way, they failed to honestly report the current state of the existing literature on the subject.
Next, we look at model bias in WGIII.
Download the bibliography here.
(IPCC, 2022) ↑
(IPCC, 2021, p. 67) ↑
(IPCC, 2022, p. 9) ↑
(IPCC, 2022, pp. 44-70) ↑
(May, Are fossil-fuel CO2 emissions good or bad?, 2022g), (Idso, 2013), (Zhu, Piao, & Myneni, 2016), (Tol R. S., 2018) , (Tol R. , Correction and Update: The Economic Effects of Climate Change, 2014b), and (O’Neill, 2023) ↑
(IPCC, 2022, p. 264) ↑
(O’Neill, 2023) ↑
(O’Neill, 2023) ↑
(International Monetary Fund, 2022) ↑
(Tol R. S., 2018) ↑
(Tol R. S., 2018) ↑
(Pielke & Ritchie, 2021) ↑
(Hausfather & Peters, 2020) ↑
(IPCC, 2022, p. 136) ↑
(Pielke & Ritchie, 2021) ↑
(Pielke & Ritchie, 2021) and (Hausfather & Peters, 2020) ↑
(IPCC, 1990, pp. 55-56) ↑
(Crok & May, 2023, pp. 122-126), (Hausfather & Peters, 2020), and (Pielke Jr, Burgess, & Ritchie, 2021) ↑
(Crok & May, 2023, p. 142) ↑
(Weinkle, Maue, & Pielke Jr., 2012) and see Dr. Maue’s site https://climatlas.com/tropical/ ↑
(Crok & May, 2023, p. 147), also see Dr. Maue’s site https://climatlas.com/tropical/ ↑
(IPCC, 2021, p. 82) ↑
(Crok & May, 2023, p. 146) ↑
(IPCC, 2021, p. 1569) ↑
(IPCC, 2022, p. 588) ↑
(Lomborg, Welfare in the 21st century: Increasing development, reducing inequality, the impact of climate change, and the cost of climate policies,, 2020), (Lomborg, We’re Safer From Climate Disasters Than Ever Before, 2021), and (Pielke Jr., 2021) ↑
(Yan, et al., 2001) ↑
(Lomborg, Welfare in the 21st century: Increasing development, reducing inequality, the impact of climate change, and the cost of climate policies,, 2020) ↑
(Dixon, et al., 2005) ↑
(Grinsted, Ditlevsen, & Christensen, 2019) ↑
For a list see: (Crok & May, 2023, p. 153) ↑
(Pielke Jr., Apples, Oranges, and Normalized Hurricane Damage, 2024) ↑
(IPCC, 2022, p. 1978) ↑
(Estrada, Botzen, & Tol, 2015) ↑
(IPCC, 2022, p. 1978) ↑
9 thoughts on “Climate Model Bias 6: WGII”