Climate Model Bias 2: Modeling Greenhouse Gases

By Andy May

Christian Freuer has translated this post into German here.

Since the late 19^th century, with the work by Svante Arrhenius, climate models have been used to estimate the amount of global warming due to human greenhouse gas emissions.^[1] Due to the complexity of Earth’s weather and climate, the connection between climate change/global warming and greenhouse gases cannot be observed or measured, it can only be estimated with a model. Arrhenius constructed the first such model and speculated that temperature changed linearly with the logarithm of the CO₂ concentration, specifically he estimated that as CO₂ doubled in the atmosphere, surface temperature should increase 4°C.^[2]

In part 1 of this series, I proposed a definition of a scientific model and described the types of models used by the IPCC in their various reports. We saw that as George Box famously said, “All models are wrong.”^[3] He also said that all models reach a point where they cannot be improved, a point where (paraphrasing) attempts to refine them or make them more accurate amount to self-flagellation. We also saw that there are many examples where the sixth generation CMIP models fail to predict climate accurately. If you can’t get it right after six tries, you might have entered the realm of self-flagellation. Finally, part 1 lists the most common types of bias in the CMIP climate models and the IPCC reports.

Arrhenius’ estimate of the climate sensitivity to a doubling of CO₂ of 4°C, falls within the IPCC AR6 estimate of the likely range of this value, which they abbreviate as “ECS” or the equilibrium climate sensitivity. The AR6 range is 2.5 to 4°C/2xCO₂.^[4] Thus according to the IPCC, Arrhenius’ estimate over 100 years ago is possibly still valid. Arrhenius concludes that the changes due to burning fossil fuels are a good thing, and writes:

“By the influence of the increasing percentage of carbonic acid [CO₂] in the atmosphere, we may hope to enjoy ages with more equable and better climates, especially as regards the colder regions of the earth, ages when the earth will bring forth much more abundant crops than at present, for the benefit of rapidly propagating mankind.”^[5]

Arrhenius, 1908, page 63.

Arrhenius’ model was based upon laboratory measurements of the “permeability to heat”^[6] of CO₂ and water vapor. He tells us that the sources of additional CO₂ are volcanos or human fossil fuel consumption and the main sinks for CO₂ are the ocean, the weathering of rocks by carbonic acid, and growing plants.^[7] Plants are more abundant and robust when the atmospheric CO₂ concentration is high and less so when it is lower. He notes that all CO₂ sinks increase in strength as the CO₂ concentration increases.

Arrhenius’ conclusions were mostly accepted as the world warmed from 1908 to 1945, but when the world cooled from 1945 to 1976, while CO₂ was still increasing, the hypothesis fell into disrepute. The pattern of global average surface temperature change, often used as the primary metric of “climate change” is illustrated in figure 1 using the HadCRUT4 global surface temperature anomaly dataset.^[8] His ideas gained favor again after 1976 when the world began to warm again. The instrumental temperature record reveals that while CO₂ concentration may have some influence on climate, as the Arrhenius and IPCC climate models attempt to show, it cannot be the entire story. It is instructive that both Arrhenius and the IPCC assume, in their models, that solar variability plays no role in recent warming and cooling. Both assume that volcanos do play a role, Arrhenius thought volcanos put more CO₂ into the air and caused warming^[9] and the IPCC assumes that volcanic aerosol emissions reflect more sunlight and have a cooling influence.^[10]

In AR6, they feed recent observations of CO₂ and temperature, paleoclimate CO₂ and temperature proxies from the geological record, and modeled constraints on climate sensitivity into a subjective statistical model to compute a likely ECS range of 2.5 to 4°C/2xCO₂.^[11] This is a change from previous reports where general circulation climate models were used to compute the ECS. As they write in AR6:

“… despite decades of model development, increases in model resolution, and advances in parametrization schemes, there has been no systematic convergence in model estimates of ECS. In fact, the overall inter-model spread in ECS for CMIP6 is larger than for CMIP5; …”^[12]

AR6, WGI, page 1008.

While reading this quote from AR6, we need to remember Box’s cautions about “overelaboration and overparameterization.” It is also worrying that even after refining the IPCC/CMIP models for over 30 years, the spread in model results is not converging, it is increasing. This is a sign that the model premises^[13] and/or design are flawed. It is also a clue that the conceptual IPCC model that human activities are controlling climate change,^[14] has reached the end of its useful life.

The likely lower limit of 2.5°C/2xCO₂ was raised from the AR5 lower limit of 1.5°C/2xCO₂.^[15] However, their subjective statistical model is flawed and internally inconsistent with the confidence intervals of each input variable. When Nic Lewis redid their work using an objective statistical model, with improved and more recent data, he produced an ECS range of 1.75-2.7°C/2xCO₂ (5-95%ile).^[16] Like the IPCC, Lewis assumes that increasing CO₂ has caused nearly all the warming since the 19^th century. If that assumption is incorrect and there was some natural warming due to the Modern Solar Maximum^[17] then even Lewis’ calculated ECS values are too high and are maximum, not likely values.^[18] A large body of evidence for an ECS less than 2°C/2xCO₂ exists,^[19] which supports Lewis’ view that the AR6 subjectively modeled ECS is biased high.

AR6 does mention that observation-based studies exist that produce estimates of ECS below 2°C/2xCO₂ but dismisses them because they believe that the feedback to increasing CO₂ and warming increases with warming.^[20] This concept of climate feedback changing as the world warms to increase the warming rate is a perfect example of Box’s “overelaboration.” They write that since observation-based estimates assume that the net feedback remains constant over the period being studied (usually 1870 to 2020), the estimates are incorrect. All the authors of the Clintel AR6 assessment dispute the idea that CO₂ warming feedback changed significantly from 1870 to 2020.^[21]

When a favored model, such as the IPCC model that CO₂ is the “control knob”^[22] for climate reaches its end, it is common to reach for any idea that can save it from extinction (overelaboration). However, the idea that CO₂-caused warming feedback changes radically over geologically short periods of 150 years is a very desperate reach.

Laboratory measurements show that infrared active gases, like CO₂, water vapor, and some hydrocarbon gases selectively absorb some of the infrared radiation emitted by Earth’s surface.^[23] When they do this, the gas molecules become excited and vibrate faster for periods of a few milliseconds to a few tenths of a second and in this excited state, they collide more often with the gas molecules around them, warming them.^[24] Then they emit the remainder of the radiation they absorbed. In the denser and more humid atmosphere near Earth’s surface, this has a warming effect because almost all the emitted surface radiation joins the neighborhood thermal energy pool.^[25] The higher in the atmosphere the surface radiation is absorbed, the less energy is retained in the immediate neighborhood and the more energy is emitted by the greenhouse gases as radiation, some of which goes to outer space and escapes Earth’s climate system. The altitude where more of the radiation goes to space than to the surrounding air or back to the surface is traditionally called the “top of the atmosphere” or TOA.

Summary

In this part we examined the IPCC “conceptual model” or hypothesis that attributes modern warming to human influence through our land use and greenhouse gas emissions,^[26] and the idea’s 100-year plus history. While adding greenhouse gases to the atmosphere probably causes some warming there are numerous studies that suggest some of the warming is natural and unrelated to human activities. The proportion of the warming that is due to human activities is unknown, but as we have seen in this post, it is very unlikely to be either zero or 100%, it falls in the middle somewhere, and may be small.

Modern general circulation climate models (called “Earth System Models” or “ESMs” in AR6) are all based upon the idea that most or all the warming in the past 150-170 years is due to human activities, with minimal impact from volcanism or solar activity changes. However, the models do not match observations in numerous critical areas.^[27] As noted above, in some cases the CMIP6 models are worse than the previous generation CMIP5 models. Since these mismatches have persisted through six iterations of the models, it is reasonable to assume the flaw is in the assumptions, that it is in the hypothesis itself. In other words, it is likely the conceptual model should be scrapped, and a new one with different assumptions constructed.

In the next part we will examine the IPCC/CMIP assumption that changes in solar activity over the past 170 years have been too small to make any difference in Earth’s climate. In part 4 we will examine the assumption that changes in atmospheric and oceanic circulation make no net difference in the climate over time, since all these changes supposedly do not change the total thermal energy (heat) stored in the climate system, they just move it around.

Download the bibliography here.

1. 
   

   
   (Arrhenius S. , 1896) ↑
   

   
   
2. 
   

   
   (Arrhenius S. , 1908, p. 53) ↑
   

   
   
3. 
   

   
   (Box, 1976) ↑
   

   
   
4. 
   

   
   We abbreviate the temperature increase due to a doubling of CO₂ as “2xCO₂.” The IPCC AR6 estimate can be found here: (IPCC, 2021, p. 926). ↑
   

   
   
5. 
   

   
   (Arrhenius S. , 1908, p. 63) ↑
   

   
   
6. 
   

   
   (Arrhenius S. , 1908, p. 52) ↑
   

   
   
7. 
   

   
   (Arrhenius S. , 1908, pp. 54-55) ↑
   

   
   
8. 
   

   
   (Morice, Kennedy, Rayner, & Jones, 2012) and (Morice C. , 2017) ↑
   

   
   
9. 
   

   
   (Arrhenius S. , 1908, p. 55) ↑
   

   
   
10. 
    

    
    (IPCC, 2021, pp. 961-962) ↑
    

    
    
11. 
    

    
    (IPCC, 2021, p. 926 & 1007), (Sherwood, et al., 2020), and (Lewis, 2023) ↑
    

    
    
12. 
    

    
    (IPCC, 2021, p. 1008), AR5 is the fifth major WGI IPCC report: (IPCC, 2013). CMIP5 are the models used in AR5. ↑
    

    
    
13. 
    

    
    Model premises are all the assumptions and data used to build it. ↑
    

    
    
14. 
    

    
    “Taken together with numerous formal attribution studies across an even broader range of indicators and theoretical understanding, this underpins the unequivocal attribution of observed warming of the atmosphere, ocean, and land to human influence.” AR6 WG1 TS, page 63. ↑
    

    
    
15. 
    

    
    (IPCC, 2013, p. 81) ↑
    

    
    
16. 
    

    
    (Lewis, 2023) ↑
    

    
    
17. 
    

    
    The Modern Solar Maximum is from 1940 to 2000. See (Usoskin, Solanki, & Kovaltsov, 2007). ↑
    

    
    
18. 
    

    
    (May, The Mysterious AR6 ECS, Part 4, converting observations to ECS, 2023b) and (Crok & May, 2023, p. 103) ↑
    

    
    
19. 
    

    
    (May, The Mysterious AR6 ECS, Part 3, What is Climate Sensitivity?, 2023) ↑
    

    
    
20. 
    

    
    (IPCC, 2021, p. 996) ↑
    

    
    
21. 
    

    
    (Crok & May, 2023, p. 104) ↑
    

    
    
22. 
    

    
    (Lacis, Hansen, Russell, Oinas, & Jonas, 2013), (Lacis, Schmidt, Rind, & Ruedy, 2010), and (IPCC, 2021, p. 179) ↑
    

    
    
23. 
    

    
    (Tyndall, 1861) and (Wijngaarden & Happer, 2020) ↑
    

    
    
24. 
    

    
    (Pierrehumbert, 2011) ↑
    

    
    
25. 
    

    
    (Wijngaarden & Happer, 2020) and (Pierrehumbert, 2011) ↑
    

    
    
26. 
    

    
    AR6 WG1 TS, page 63 ↑
    

    
    
27. 
    

    
    (McKitrick & Christy, A Test of the Tropical 200- to 300-hPa Warming Rate in Climate Models, Earth and Space Science, 2018), (McKitrick & Christy, 2020), (Lewis, 2023), (IPCC, 2021, p. 990) ↑