October 26, 2020
by Pat Frank
This paper extends the previously published assessment of CMIP5 climate models to the predictive and physical reliability of global average CMIP6 air temperature projections.
Before we move on, a big thank you to Anthony and Charles, the moderators, for providing such an excellent forum for openly communicating ideas and for making my work public. Having a voice is so very important. Especially these days when so many are working to silence it.
I have already reported on the predictive reliability of climate models in Watts Up With That (WUWT) here, here, here and here. Those who prefer a video presentation of their work can find it here. To ensure full transparency, the video review published here by Dr. Patrick Brown (now Prof. Brown at San Jose State University), who was refuted in the comments section below this video from here on.
Those who read these reviews will find that Dr. Brown shows no obvious training in physical failure analysis. He made the same novice-level mistakes that climate modellers often make, which are discussed in detail here and here.
In our debate, Dr. Brown very polite and polite. He looked nice and well meaning. However, when his teachers and mentors did not allow him to evaluate the accuracy and quality of the data, they betrayed him.
Lack of training in data quality assessment appears to be an educational gap for most, if not all, AGW consensus climate scientists. They find no meaning in the critically central distinction between precision and accuracy. There can be no possible advance in science at all if workers are not trained to be critical of the quality of their own data.
The best overall description of climate model errors is still Willie Soon et al., 2001 Modeling the Climatic Impacts of Anthropogenic Carbon Dioxide Emissions: Unknowns and Uncertainties. Almost all of the simulation errors and deficiencies described remain to this day.
Jerry Browning recently published a rigorous mathematical physics study that uses the simulation errors described by Willie et al. Reveals at its source. He showed that the incorrectly formulated physical theory in climate models generates discontinuous heating / cooling terms that lead to a reduction in the simulation accuracy by orders of magnitude.
These discontinuities would cause climate simulations to diverge quickly, except that climate modelers suppress them with a hyperviscous (molasses) atmosphere. Jerry's paper offers the way out. Nevertheless, discontinuities and molasses atmospheres remain features of the new improved CMIP6 models.
In the Fifth Assessment Report 2013 (5AR), the IPCC used CMIP5 models to predict the future of global air temperatures. The upcoming 6AR will use the improved CMIP6 models to predict the thermal future that awaits us if we continue to use fossil fuels.
CMIP6 cloud error and detection limits: Figure 1 compares the CMIP6-simulated global average annual cloud percentage with the measured cloud percentage and shows the difference between 65 degrees north and south latitude. The mean annual error of the cloud fraction in the root mean square (effective value) is ± 7.0%.
This error calibrates the average accuracy of CMIP6 models against a known observable cloud fraction. The mean annual rms error of the CMIP5 cloud fraction over the same latitude is ± 9.6%, which indicates an improvement in CMIP6 of 27%. Nevertheless, CMIP6 models still make significant simulation errors in the global cloud fraction.
Figure 1 lines: red, MODIS + ISCCP2 annual average measured cloud fraction; blue, CMIP6 simulation (9 model average); green, (measured minus CMIP6) annual average calibration error (width rms error = ± 7.0%).
The following analysis is a simple extension of the earlier error propagation on CMIP6 models that was applied to the air temperature projections of CMIP5 climate models.
Errors in the simulation of the global cloud fraction lead to downstream errors in the long-wave cloud forcing (LWCF) of the simulated climate. LWCF is a source of thermal energy flow in the troposphere.
The tropospheric heat energy flow is the determinant of the tropospheric air temperature. Simulation errors in LWCF lead to uncertainties in the heat flow of the simulated troposphere. These in turn bring uncertainty into the projected air temperatures.
For more information, see here – Figure 2 and surrounding text. The dissemination of the above linked error paper also provides a detailed discussion of this point.
The global annual average of the longwave LWCF root mean square calibration error of the top of the atmosphere (TOA) of CMIP6 models is ± 2.7 Wm² (28 model average from Figure 18 here).
I was able to check the validity of this number as the same source also provided the average annual LWCF error for the 27 CMIP5 models evaluated by Lauer and Hamilton. The Lauer and Hamilton CMIP5 annual average LWCF error is ± 4 Wm². Independent redetermination resulted in ± 3.9 Wm²; the same within the rounding error.
The little question of resolution: Compared to the CMIP6-LWCF calibration error (± 2.7 Wm²), the average annual increase in CO2 forcing between 1979 and 2015 is 0.025 Wm² according to the EPA. The annual average increase in the sum of all drives for all major greenhouse gases in the period 1979-2015 is 0.035 Wm².
The annual average CMIP6-LWCF calibration error (± 2.7 Wm²) is therefore ± 108 times greater than the annual average increase in propulsion due to CO2 emissions alone and ± 77 times greater than the annual average increase in propulsion from all greenhouse gas emissions .
This means that a lower limit of the CMIP6 resolution is ± 77 times greater than the disturbance to be detected. This is a small improvement over CMIP5 models, which had a lower limit resolution ± 114 times too large.
For analytical accuracy, the detection limit (resolution) of the instrument must generally be ten times smaller than the expected measurand. In order to fully record a signal of CO2 or GHG emissions, current climate models have to improve their resolution by almost 1000 times.
Another possibility is that CMIP6 climate models may not be able to capture the effects of CO2 emissions or greenhouse gas emissions on the earth's climate or global air temperature.
This fact is destined to be ignored in the consensus climatologist community.
Emulation validity: Papalexiou et al., 2020, found that "the credibility of climate projections is typically defined by how accurately climate models represent historical variability and trends". Figure 2 shows how well the linear equation previously used to emulate CMIP5 air temperature projections reproduces GISS Temp anomalies.
Figure 2 lines: blue, GISS Temp 1880-2019 land plus SST air temperature anomalies; red, Emulation only using the Meinshausen RCP forces for CO2 + N2O + CH4 + volcanic eruptions.
Emulation is running through the middle of the trend and is especially good in the post-1950 region where air temperatures are supposedly determined by greenhouse gas (GHG) emissions. The nonlinear temperature drops due to volcanic aerosols are successfully reproduced in 1902 (Mt. Pelée), 1963 (Mt. Agung), 1982 (El Chichón), and 1991 (Mt. Pinatubo). We can proceed after proving the credibility of the published standard.
CMIP6 World: The new CMIP6 projections have new scenarios, the Shared Socioeconomic Pathways (SSPs).
These scenarios combine the representative concentration paths (RCPs) of the 5AR with “quantitative and qualitative elements, based on worlds with different challenges in mitigation and adaptation (with) new scenario storylines (including) quantifications of the associated population and income development … For the Research community on climate change. "
You can find increasingly developed descriptions of these storylines here, here and here.
The emulation of the CMIP6 air temperature projections follows the identical method described in the propagation of the error paper linked above.
The analysis here focuses on projections created with the CMIP6 IMAGE 3.0 earth system model. IMAGE 3.0 was designed in such a way that all extended information is contained in the new SSPs. The IMAGE 3.0 simulations were chosen for practical reasons only. The project published in 2020 by van Vuulen et al. Published paper conveniently included both the SSP forces and the resulting air temperature projections in Figure 11. The published data was converted to points using DigitizeIt, a tool that served me well.
Here is a short, descriptive quote for IMAGE 3.0: “IMAGE is an integrated assessment model that simulates global and regional environmental impacts of changes in human activities. The model is a simulation model; H. Changes in model variables are calculated based on the information from the previous time step.
“(IMAGE simulations are based on) two main systems: 1) the human or socio-economic system, which describes the long-term development of human activities relevant to sustainable development; and 2) the Earth system, which describes changes in natural systems such as the carbon and water cycles and climate. The two systems are linked through emissions, land use, climate feedback and possible human-political reactions. (my bold print) "
Regarding incorrect iterations: The sentence in bold above describes the step-by-step simulation of a climate, in which each previously simulated climate condition in the iterative calculation supplies the initial conditions for the subsequent climate condition simulation up to the final simulated condition. Simulation as a step-by-step iteration is standard.
If the physical theory used in the simulation is incorrect or incomplete, each new iterative initial state carries its error into the subsequent state. Each subsequent state is then additionally subject to a further induced error from the operation of the wrong physical theory on the error-prone initial state.
Critically and as a consequence of the step-by-step iteration, systematic errors in each intermediate climatic condition are transferred to each subsequent climatic condition. The uncertainties from systematic errors then spread through the simulation as the sum of the roots (rss).
Fittingly, Jerry Browning's article has analytically and rigorously shown that climate models apply a wrong physical theory. Figure 1 above shows that one of the consequences is an error in the simulated cloud fraction.
In a projection of future climatic conditions, the physical simulation errors are unknown because future observables are not available for comparison.
The RSS propagation of a known model calibration error through the iterated steps, however, generates a reliability statistic on the basis of which the simulation can be evaluated.
The above summarizes the method used to evaluate the projection reliability in the propagation paper and here: first calibrate the model against known targets and then propagate the calibration error through the iterative steps of a projection as the root sum squared uncertainty. Repeat this process until the last step, which describes the predicted final future state.
The final uncertainty of the root sum square (rss) indicates the physical reliability of the final result, since the physically true error cannot be seen in a future forecast.
This method is standard in the natural sciences when determining the reliability of a calculated or predictive result.
Emulation and Uncertainty: One of the key demonstrations in the paper on error propagation was that advanced climate models only project air temperature as a linear extrapolation of greenhouse gas forcing.
Figure 3, panel a: Points are the air temperature projection of IMAGE 3.0 from, blueScenario SSP1; and red, Scenario SSP3. Solid lines are the emulations of the IMAGE 3.0 projections: blue, SSP1 projection and red, SSP3 projection, created using the linear emulation equation described in the published Analysis of CMIP5 Models. Panel b is as in panel a, but also shows the expanding 1s-root-sum-square uncertainty envelopes produced when ± 2.7 Wm2 of annual average LWCF calibration error is propagated through the SSP projections.
In Figure 3a above, the dots show the air temperature projections of the SSP1 and SSP3 storylines that were created with the IMAGE 3.0 climate model. The lines in Figure 3a show the emulations of the IMAGE 3.0 projections created using the linear emulation equation that is fully described in the error propagation paper (also in a 2008 article in Skeptic Magazine). The emulations are 0.997 (SSP1) or 0.999 (SSP3), which are correlated to the IMAGE 3.0 projections.
Figure 3b shows what happens when ± 2.7 Wm² of the annual average LWCF calibration error is transmitted via the global air temperature projections IMAGE 3.0 SSP1 and SSP3.
The uncertainty envelopes are so large that the two SSP scenarios cannot be statistically differentiated. It would be impossible to choose either a projection or an SSP air temperature projection as more representative of the evolving air temperature, since every possible change in the physically real air temperature is immersed in all projection uncertainty envelopes.
An interlude – There are dragons: I will take a break here to forestall an earlier misunderstanding that was hotly, persistently and repeatedly alleged. These uncertainty envelopes in Figure 3b are not physically real air temperatures. Don't entertain this bogus idea for a second. Get it out of your head. Crush his movements without mercy.
These uncertainty bars do not imply any future climate conditions with 15 ° C warmer or 10 ° C cooler. Uncertainty bars describe a breadth in which ignorance prevails. Your message is that the projected future air temperatures are somewhere within the uncertainty range. But nobody knows the place. CMIP6 models cannot say anything in particular.
Terra Incognita is in this uncertainty bar. There are dragons.
For those who insist that the uncertainty bars imply actual real physical air temperatures, consider how this idea works against the need that a physically real ± C uncertainty requires the simultaneity of hot and cold states.
Uncertainty bars are strictly axial. They are plus and minus on either side of a single data point. The assumption of two simultaneous, equally large but oppositely polarized physical temperatures, which stand on a single point of the simulated climate, means to assume a physical impossibility.
The idea does not possibly require that the earth occupy the global climatic states of hot house and ice house at the same time. Please, for the few who have entertained the idea, put it firmly behind you. Close your eyes. Never lift again.
And now back to our functional presentation: The following table contains selected projection anomalies for IMAGE 3.0 SSP1 and SSP3 scenarios and their corresponding uncertainties.
Table: FIGURE 3.0 Projected air temperatures and uncertainties for selected simulation years
|action||1 year (C)||10 years (C)||50 years (C)||90 years (C)|
|SSP1||1.0 ± 1.8||1.2 ± 4.2||2.2 ± 9.0||3.0 ± 12.1|
|SSP3||1.0 ± 1.2||1.2 ± 4.1||2.5 ± 8.9||3.9 ± 11.9|
Neither of these projected temperatures is different from physically meaningless. None of them say anything physically real about possible future air temperatures.
Several conclusions follow.
First, like their predecessors, CMIP6 models project air temperatures as a linear extrapolation of the drive.
Second, like their predecessors, CMIP6 climate models cause large simulation errors in the cloud fraction.
Third, like their predecessors, CMIP6 climate models produce LWCF errors that are vastly larger than the tiny annual increase in tropospheric forcing caused by greenhouse gas emissions.
Fourth, CMIP6 climate models, like their predecessors, generate uncertainties that are so large and immediate that the air temperatures cannot be reliably projected even after a year.
Fifthly, like their predecessors, CMIP6 climate models must have a resolution that is around 1000 times better in order to reliably record a CO2 signal.
Sixth, CMIP6 climate models, like their predecessors, generate physically meaningless air temperature projections.
Seventh, like their predecessors, CMIP6 climate models have no predictive value.
As before, the inevitable conclusion is that an anthropogenic air temperature signal in climate observables could neither be detected nor can it be detected at present.
I will end with an observation that we have already made before: We now know for sure that the whole intoxication about CO₂ and the climate was in vain.
All tortured adults; all the desperate young people; All high school students were afraid of tears and accusations through lessons about the coming fate, death and destruction. all the social unrest and dislocations. Everything was in vain.
All of the blame, all of the character assassinations, all of the damaged careers, all of the excessive winter deaths from fuel starvation, all of the men, women, and children who continue to live with smoke indoors, all of the enormous sums of money that have been distracted, everyone the tainted landscapes, all the chopped and burned birds and the destroyed bats, all the huge funds that the middle class transferred to rich subsidized farmers:
All for free.
Finally, a page from Willis Eschenbach's book (Willis always gets to the heart of the problem). If you have any issues with this work in the comments, please quote my actual words.