From Watts Up With That?

By Andy May

The climate sensitivity to CO₂ and other greenhouse gases (GHGs) is arguably the most important number in the climate change debate. AR6^[1] claims the sensitivity, which they call “ECS” or the equilibrium climate sensitivity, is three degrees per doubling of CO₂, or 3°C/2xCO₂ (“/2xCO₂” simply means per doubling of the atmospheric CO₂ concentration). They claim the very likely (10% to 90%) range of possible values is from 2 to 5°C/2xCO₂ and the likely (66%) range has narrowed to 2.5 to 4°C. Since 1979, with the publication of 1979 Charney Report,^[2] the range of possible ECS values has normally been about 1.5 to 4.5°C for a total range of 3°C, how has it now narrowed to 2.5 to 4°C, a full likely uncertainty range of only 1.5°C? It is generally accepted that the direct warming effect of CO₂ and other greenhouse gases is small, only about one degree per doubling of CO₂,^[3] so the debate is all about the feedbacks, especially cloud feedback to the greenhouse gas warming.^[4]

The real-world effect of changing the CO₂ and GHG atmospheric concentration on climate, whether natural or emitted by humans, has never been measured, only modeled. ECS is defined as the ultimate warming due to an instantaneous doubling of the atmospheric CO₂ concentration. The ultimate climate response to that doubling will not occur for hundreds or thousands of years and everything else affecting the climate, like cloudiness, and insolation will not stay static for that long, so it is an artificial quantity that cannot be measured. Importantly, the IPCC estimate of ECS can probably not be falsified through real world measurements, which means it is not a proper scientific hypothesis. Even with a climate model it is difficult, in Sherwood, et al., they write:

“To calculate the ECS in a fully coupled climate model requires very long integrations (>1,000 years).”^[5]

A more relevant measure of climate sensitivity is the transient climate response, or TCR, which is also calculated by the AR6/CMIP6 climate models. This is the climate response to a steady increase in CO₂ concentration of about 1% per year, to the point where CO₂ doubles, roughly 70 years.^[6] Thus, it is a more realistic and, given the short time frame, it is possibly measurable in the real world. Sherwood, et al.,⁶ a source relied upon in AR6 (Chapter 7 mentions the Sherwood paper 43 times), defines a term called “effective climate sensitivity” that is the climate response to an instantaneous doubling of CO₂, or more specifically half of the climate response to an instantaneous quadrupling of CO₂. By adding “effective” to the name rather than “equilibrium” they cut the time frame to 150 years.

AR6 constrains their estimates of TCR and ECS using four lines of evidence: process (mostly feedbacks) understanding, climate model simulations, historical observations, and paleoclimatic observations, plus a fifth category they call a synthesis of evidence, in explaining this new ECS evaluation they write:

“All four lines of evidence rely, to some extent, on climate models, and interpreting the evidence often benefits from model diversity and spread in modelled climate sensitivity. … unlike in previous assessments, climate models are not considered a line of evidence in their own right in the IPCC Sixth Assessment Report.” AR6, page 1024.

As explained above, ECS is not measurable since it is derived from an unreal model scenario. In AR6 ECS and TCR are referred to as “idealized quantities … that can be inferred from [observations] or estimated directly using climate [model] simulations.”^[7] Thus, all attempts to estimate them in nature require some sort of model to transform the measurements to the modeled ECS or TCR scenario given in AR6.

Nic Lewis and Judith Curry try to simplify the conversion from observations of temperature and CO2 by carefully selecting periods of time when natural forces are as comparable as possible. However, they only consider volcanism and major ocean oscillations, like ENSO and the Atlantic Multidecadal Oscillation (AMO). We have more to say about this idea in Part 4.

The world was cooler in the 19^th century and the Little Ice Age was just ending. For this reason, there were fewer El Niños then than now. El Niños occur due to excess heat buildup in the Pacific Ocean that must be expelled to the atmosphere. They warm Earth’s surface for a few years, but longer term they act as a cooling agent.^[8] The number of strong El Niños and their strength increases as warm periods end and the world grows cooler, as happened at the end of Medieval Warm Period when the Earth dipped into the Little Ice Age from ~1050AD to ~1400AD. Once the world became colder, their strength and number reduced, as observed until the late 20^th century.^[9]

While there is considerable debate on the subject, it is likely that solar variability also plays a role in climate change and the Sun was less active in the 19^th century than during the Modern Solar Maximum from ~1935 to ~2005.^[10] While the IPCC believes that solar variability and other natural factors, except for volcanos over short periods, play no role in climate change over the past 270 years,^[11] Javier Vinós and Ronan Connolly, et al.^[12] have presented considerable evidence that this is not the case. Thus, the calculations Lewis and Curry make to convert their measurements to the modeled quantities of ECS and TCR might be contaminated by natural factors that they did not take into account. Even so, their calculations of ECS and TCR are considerably below the IPCC likely lower limits of 2.5°C and 1.4°C, respectively. Tables of estimates of ECS and TCR will be presented in part 3.

Besides TCR and ECS, the classical climate sensitivity quantity, which we simply call “climate sensitivity to CO₂,” is totally evidence based and determined from observations. The classical quantity is best defined as the surface air temperature sensitivity (SATS) to an increase in CO2.^[13] The units used for SATS are degrees C per Watts per square meter of forcing. By assuming all the forcing is due to CO2, the value can be converted into °C/2xCO₂. The further conversion of this value to ECS or TCR requires making assumptions about the time required for Earth’s surface (mainly the oceans) to come into equilibrium from the change in forcing inclusive of any feedbacks to the CO₂-caused warming. In the tables shown in part 3, we list many of these observation-based estimates of climate sensitivity. Some of them, including Lewis and Curry, use simple models to translate the measurements into pseudo ECS and TCR. When this is done, typically the same assumptions made by the IPCC are used for the conversion model. Other estimates of the classical climate sensitivity just present the measurements, but they assume a forcing for the CO₂ changes observed.

Mauna Loa^[14] measured atmospheric CO₂ is increasing at about 2 ppm (0.5%) per year and is not far off from the TCR defined 1% per year. Since the preindustrial era, or the Little Ice Age, CO₂ has increased about 50% (one-half of a doubling), thus we are in a time when TCR is relevant.

As mentioned above, AR6 does not use models to directly compute ECS and TCR as they did in the past. Instead, they use five lines of evidence to constrain the final ECS and TCR model calculations.^[15] This process is laid out in considerable detail in Sherwood, et al.^[16] and in AR6 section 7.5. Sherwood’s analysis tries to show that all values used in the current calculation of ECS are narrowly constrained except for the cloud feedback to surface warming, and in particular, the feedback due to lower-level clouds. It is important to understand that all the methods that AR6 uses to constrain their estimates of ECS and TCR rely, to some extent, on climate models. We show the cloud feedback relationship to ECS in part 2.

The statistical analysis methods that Sherwood, et al. use to integrate various estimates of climate sensitivity into one range are subjective and seriously flawed, as shown by Nic Lewis.^[17] Lewis corrected their errors and lowered their estimate of climate sensitivity from 3.23 to 2.16K, about 33%. Nic Lewis points out that “Climate sensitivity has been estimated from various types of evidence, but none of these has narrowly constrained its value.”

AR6 isolates the processes that they think contribute to warming and constrain them with observations, this is a reasonable approach if the full range of possible processes affecting warming are considered. The authors of Connolly, et al. ^[18] believe that AR6 have not properly considered the potential influence of solar variability. Connolly, et al. demonstrate that insolation and other solar variability may be much more important than assumed by the IPCC. As shown in figure 1, the AR6 estimate of natural warming (volcanism and solar variability) is zero, or slightly negative.

Figure 1. AR6 estimated temperature change contributions from 1750 to 2019, with uncertainties. The assumed natural contribution is zero, or slightly negative, plus or minus a small amount. Source: AR6, chapter 7, page 961.

Global surface warming from 1971 through 2018 is about 0.85°C, according to HadCRUT4.^[19] According to AR6^[20] this corresponds to a top of the atmosphere (TOA) energy imbalance of +0.57 W/m² (+0.2% of the incoming ~340 W/m² from the Sun) for the same period. For Earth’s surface to warm, it must retain more thermal energy than it emits to space. When this surface energy imbalance, which is measured in Watts per square meter of surface (W/m²), causes warming, it is positive by convention. Some of the excess energy warms the surface, and the warmer surface and lower atmosphere emit more energy to space, resulting in the positive 0.57 W/m² increase in emissions at the TOA.

Thus, if we were to assume all the surface warming is due to increasing CO₂ and other greenhouse gases (GHGs), the surface air temperature sensitivity (SATS) to GHGs is about 1.6°C/W/m², if nothing else changes. This is an extraordinarily large number. The classical values, based on observations,^[21] are typically between 0.1°C/W/m² to 0.5°C/W/m². This suggests that all recent warming is not entirely due to GHGs.

The oceans are not really a factor in the short term, since IR (Infrared Radiation) emitted by CO₂ and other GHGs cannot penetrate far below the ocean surface, like sunlight does. Most incident IR is absorbed in the first millimeter of the ocean and re-emitted or evaporated away shortly after. IR does warm the sea surface and some of this heat will go into the deeper ocean through conduction and turbulent mixing, but IR is not as transmissible to the deep ocean as visible light, especially blue light.^[22]

If the AR6 estimate of the radiative imbalance is correct, the 48-year period from 1971 to 2018 had an average imbalance of 0.01 W/m² per year. This is tiny and far below what we can measure today. The accuracy of our satellite measurements of Earth’s incoming and outgoing radiation is no better than ±2 W/m².^[23] Besides the contribution of GHGs, there are other natural factors, such as changes in cloud cover and type, that can play a large role in either increasing or decreasing the radiative imbalance at Earth’s surface.

In part 2 of this series, we will examine the largest uncertainty in the AR6 ECS estimate, cloud feedback. In part 3 of the series, we will compare the AR6 ECS and TCR estimates to observation-based estimates. Some of the observation-based estimates are considered by AR6, and some are not. We will see that many observation-based estimates of climate sensitivity are considerably lower than the AR6 likely lower bound of 2.5°C.

Finally, in part 4 we examine how Lewis and Curry^[24] convert their selected observations into a value that can be compared to the totally model-based value called “ECS.” It is unusual to convert measurements to model values, usually it is done the other way around, but is their conversion valid? What assumptions do they make? The Lewis and Curry ECS is significantly lower than the AR6 likely lower bound, how do we interpret that difference?

The bibliography can be downloaded here.

1. (IPCC, 2021) or AR6. ↑
2. Charney, J., Arakawa, A., Baker, D., Bolin, B., Dickinson, R., Goody, R., . . . Wunsch, C. (1979). Carbon Dioxide and Climate: A Scientific Assessment. National Research Council. Washington DC: National Academies Press. doi:https://doi.org/10.17226/12181 ↑
3. (Charney, et al., 1979, p. 8) ↑
4. Dessler, A. E. (2013). Observations of Climate Feedbacks over 2000-10 and Comparisions to Climate Models. J of Climate, 333-342. ↑
5. Sherwood, S. C., Webb, M. J., Annan, J. D., Armour, K. C., J., P. M., Hargreaves, C., . . . Knutti, R. (2020, July 22). An Assessment of Earth’s Climate Sensitivity Using Multiple Lines of Evidence. Reviews of Geophysics, 58. ↑
6. AR6, p 992 ↑
7. AR6, p 992 ↑
8. Vinós, J. (2022). Climate of the Past, Present and Future, A Scientific Debate. Spain: Critical Science Press. Pages 53-54. Link. ↑
9. (Moy, Seltzer, & Rodbell, 2002) ↑
10. Vinós, J. (2022). Climate of the Past, Present and Future, A Scientific Debate. Spain: Critical Science Press. Page 192 and Connolly et al., R. (2021). How much has the Sun influenced Northern Hemisphere temperature trends? Research in Astronomy and Astrophysics, 21(6). Link. ↑
11. AR6, page 961. ↑
12. Connolly et al., R. (2021). How much has the Sun influenced Northern Hemisphere temperature trends? Research in Astronomy and Astrophysics, 21(6). ↑
13. Newell, R., & Dopplick, T. (1979). Questions Concerning the Possible Influence of Anthropogenic CO2 on Atmospheric Temperature. J. Applied Meterology, 18, 822-825 and (Idso S. , 1998). ↑
14. Global Monitoring Laboratory – Carbon Cycle Greenhouse Gases (noaa.gov) ↑
15. AR6, page 993 ↑
16. Sherwood, S. C., Webb, M. J., Annan, J. D., Armour, K. C., J., P. M., Hargreaves, C., . . . Knutti, R. (2020, July 22). An Assessment of Earth’s Climate Sensitivity Using Multiple Lines of Evidence. Reviews of Geophysics, 58. doi:https://doi.org/10.1029/2019RG000678 ↑
17. Lewis, N. (2022). Objectively combining climate sensitivity evidence. Climate Dynamics. ↑
18. Connolly et al., R. (2021). How much has the Sun influenced Northern Hemisphere temperature trends? Research in Astronomy and Astrophysics, 21(6). ↑
19. (Met Office Hadley Centre, 2017) ↑
20. AR6 p 937 ↑
21. Newell, R., & Dopplick, T. (1979). Questions Concerning the Possible Influence of Anthropogenic CO2 on Atmospheric Temperature. J. Applied Meterology, 18, 822-825. and Idso, S. (1998). CO2-induced global warming: a skeptic’s view of potential climate change. Climate Research, 10(1), 69-82. ↑
22. Homewood, P. (2015, May 28). Yes, The Ocean Has Warmed; No, It’s Not Global Warming. Retrieved from Not a Lot of People Know That. Also see Britannica here. ↑
23. Loeb, N. G., Doelling, D., Wang, H., Su, W., Nguyen, C., Corbett, J., & Liang, L. (2018). Clouds and the Earth’s Radiant Energy System (CERES) Energy Balanced and Filled (EBAF) Top-of-Atmosphere (TOA) Edition-4.0 Data Product. Journal of Climate, 31(2). ↑
24. (Lewis & Curry, The impact of recent forcing and ocean heat uptake data on estimates of climate sensitivity, 2018)