CMIP6 GCMs versus global surface temperatures: ECS discussion

Spread the love

by Nicola Scafetta

Two publications examining the equilibrium climate sensitivity (ECS) have recently been published in Climate Dynamics:

Scafetta, N. (2022a). CMIP6 GCM ensemble members versus global surface temperatures.

Lewis, N. (2022). Objectively combining climate sensitivity evidence.

These two papers are significant because they take different but complimentary approaches and achieve the same result – ECS <3°C. Scafetta (2022a) extends and confirm Scafetta (2022b) previously published in GRL.

Lewis study was discussed in a previous post, let us here briefly present the main findings of Scafetta (2022).

The Coupled Model Intercomparison Project (phase 6) (CMIP6) global circulation (GCM) models project equilibrium climate sensitivity (ECS) values ranging from 1.8 to 5.7°C. To reduce this range, the 38 GCM were divided into low (1.5<ECS<3.0 °C), medium (3.0<ECS<4.5°C), and high (4.5<ECs<6.0°C) ECS subgroups and their accuracy and precision were evaluated in hindcasting the average global surface warming observed from 1980-1990 to 2011-2021. The study used global surface temperature records are ERA5-T2m, HadCRUT5, GISTEMP v4, NOAAGlobTemp v5, and the satellite-based lower troposphere global temperature UAH-MSU lt v6 record was added as well.

The satellite-based record was added since surface-based records are susceptible to many biases, including urban heat, among others (Connolly et al., 2021; Scafetta, 2021a). The validation tests were conducted using 688 GCM member simulations, 143 average GCM ensemble simulations, and Monte Carlo modeling of internal GCM variability in compliance with three alternative model accuracy requirements.

The period from 1980 to 2021 was chosen because it is when the global temperature records are believed to be affected by the least uncertainty. Moreover, the same time period is also covered by satellite measurements that offer an independent estimate.

The paper’s key finding was that the vast majority of the simulations by the medium and high-ECS GCMs run too hot. From 1980–1990 to 2011–2021, only the simulation of the low ECS GCM group seems to have accurately predicted the warming shown by the surface-based records. For instance, while all temperature data show a warming below 0.6 °C, all GCM averages from the medium and high ECS group forecast a warming over 0.6 °C up to 1.3 °C. These are plainly visible in Figures 1 and 2.

Figure 1: GCM global surface temperature ensembles (yellow area, ±1σ) versus HadCRUT5 (infilled data), GISTEMP v4, NOAAGlobTemp v5, and UAH-MSU-lt v6 temperature records (black, 12-month moving average).

Figure 2: Average temperature changes (2011–2021 minus 1980–1990) hindcasted by 38 CMIP6 GCMs mean simulations. The blue vertical lines represent the temperature change measured by HadCRUT5 (infilled data), ERA5-T2m, GISTEMP v4, NOAAGlobTemp v5, and UAH-MSU-lt v6 temperature records.

If internal model variability is also considered, the conclusion remains unchanged because, as the research clearly shows, 95% and 97%, respectively, of the medium and high ECS ensemble member simulations run hotter than all temperature records. These findings are summarized in Figure 3.

Figure 3: Boxplots of the CMIP6 ensemble members for each CMIP6 GCM; # represents the number of the available simulations for each GCM. The horizontal blue lines represent the global surface warming from 1980–1990 to 2011–2021 reported by HadCRUT5 (infilled data), ERA5-T2m, GISTEMP v4, NOAAGlobTemp v5, and UAH-MSU-lt v6 temperature records, respectively.

Figures 1-3 make it abundantly evident that the warming hindcast by the GCM grows as the ECS increases and that only the low-ECS GCM group can be regarded as being consistent with the data. The research also demonstrates that the outcome holds true regardless of how statistically the internal variability of the models is modeled. Moreover, it is statistically insignificant that a small number of simulations of the medium and high ECS GCMs would seem to coincide with the evidence. Therefore, it follows that the actual ECS should be lower than 3 °C, as Lewis (2022) also found.

However, Figures 1-3 also show that if the actual warming from 1980-1990 to 2011-2021 is better represented by the UAH-MSU-lt v6 temperature record, also the low-ECS GCM would be running too hot. In fact, while the various available surface-based temperature records show a warming roughly ranging between 0.5 and 0.6 °C, the UAH-MSU-lt v6 temperature record show a warming of about 0.4 °C while the low-ECS GCMs show a warming of 0.6±0.1°C. It is worth mentioning that according to the GCMs, the troposphere should experience a greater warming trend than the surface (Mitchell et al., 2020) so that the UAH-MSU-lt v6 might even be overestimating the surface warming. The low-ECM GCMs would therefore need to be scaled down by roughly 33%, assuming that the warming of UAH-MSU-lt v6 is accurate and representative of the warming at the surface. This should imply that the actual ECS might likewise be between 1 and 2 °C.

Future warming would be moderate and not particularly concerning if the actual ECS was between 1.5 and 3.0°C. The IPCC’s predictions of future climate catastrophes if CO2 emissions are not severely cut to essentially zero would be unfounded if the actual ECS is considerably lower, which is 1-2°C. As a result, it’s critical to assess if a warming bias, as multiple studies have already revealed, may be affecting surface-based temperature data.

To check the last point, the work adds an extended section where the observed and GCM modelled warming over the land and over the ocean are compared. As a result, the land has warmed 2.0–2.3 times faster than the ocean according to surface-based temperature records, 1.5 times faster according to satellite-based temperature records, and somewhere in the middle according to the GCMs: 1.75±0.20. In addition, the surface-based temperature records over land show a warming that is around 0.4°C more than the satellite readings, whereas the surface-based temperature records over the ocean show a warming that is just slightly larger (up to 0.1°C) than the satellite observations. These results suggest that the warming reported by the surface-based temperature records, especially over land, is too large and incompatible both with the satellite measurements and the land/ocean ratio prediction of the models. These findings imply that the warming indicated by surface-based temperature records, particularly over land, is excessive and inconsistent with both satellite observations and theoretical model predictions of the land/ocean ratio.

Based on the aforementioned findings, it was determined that the surface-based temperature records may be at least 10% off when it comes to actual warming. By reducing the ECS of the low-ECS GCMs by 10%, the ECS range changes from 1.8-3.0°C to 1.6-2.7°C, which is in good agreement with Lewis’s conclusion (1.7-2.7 °C).

However, if the real warming is closer to that indicated by the UAH-MSU-lt v6 temperature record, or if the climate system is controlled by multidecadal and millennial natural oscillations that the GCMs are unable to replicate, it is possible that the ECS may be much lower (for example, 1-2°C). For example, Scafetta (2013, 2021b) deduced an ECS between 1.0 and 2.3 °C by assuming (solar-astronomically induced) natural climatic oscillations of quasi 20, 60, 115 and 1000 years, which are observed in many climatic data throughout the Holocene but not reproduced by the GCMs. The same result is obtain using solar records showing a large secular variability, while the GCMs use solar forcings taken from the solar proxy reconstructions that show the least secular variability (Connolly et al., 2021).

Because they imply that anthropogenic global warming for the upcoming decades will inevitably be moderate, the results by Scafetta (2022a) and Lewis (2022)  cast serious doubts on climatic alarmism.

Bibliography

Connolly R, Soon W, Connolly M et al. (2021). How much has the Sun influenced Northern hemisphere temperature trends? An ongoing debate. Res Astron Astrophys 21:131. https://doi.org/10.1088/1674-4527/21/6/131

Lewis, N. (2022). Objectively combining climate sensitivity evidence. Climate Dynamics https://doi.org/10.1007/s00382-022-06468-x

Mitchell DM, Lo YTE, Seviour WJM, Haimberger L, Polvani LM (2020). The vertical profile of recent tropical temperature trends: persistent model biases in the context of internal variability. Environ Res Lett 15:1040b4. https://doi.org/10.1088/1748-9326/ab9af7

Scafetta N (2013). Discussion on climate oscillations: CMIP5 general circulation models versus a semiempirical harmonic model based on astronomical cycles. Earth Sci Rev 126:321–357. https://doi.org/10.1016/j.earscirev.2013.08.008

Scafetta N (2021a). Detection of non-climatic biases in land surface temperature records by comparing climatic data and their model simulations. Clim Dyn 56:2959–2982. https://doi.org/10.1007/s00382-021-05626-x

Scafetta N (2021b). Reconstruction of the interannual to millennial scale patterns of the global surface temperature. Atmosphere 12:147. https://doi.org/10.3390/atmos12020147

Scafetta, N. (2022a). CMIP6 GCM ensemble members versus global surface temperatures. Climate Dynamics. https://doi.org/10.1007/s00382-022-06493-w

Scafetta N (2022b). Advanced testing of low, medium, and high ECS CMIP6 GCM simulations versus ERA5-T2m. Geophys Res Lett 49: e2022GL097716. https://doi.org/10.1029/2022GL097716

via Climate Etc.

September 25, 2022

CMIP6 GCMs versus global surface temperatures: ECS discussion | Climate Etc. (judithcurry.com)