A case against anthropogenic climate change Part 3

Written by Mentemalleo on March 4, 2023. Posted in Current News

The IPCC’s AR5 “Detection and Attribution of Climate Change” report ^[1] concluded that “It is extremely likely that human activities caused more than half of the observed increase in GMST [Global Mean Surface Temperature] from 1951 to 2010”.

Part 1 in this series however demonstrated that the Arctic temperature anomalies are not caused by IPCC-modeled climate forcings, such Anthropogenic Forcing or Polar Amplification.

Part 2 documented that the Arctic anomalies are very likely caused by an IPCC-ignored climate forcing, Geothermal Forcing.

This post conclusively demonstrates that the 1880-2010 global temperature variations were largely caused by changes in Geothermal Forcing (GF) and not by human activities.

Covariation versus Causation

Three conditions are necessary to prove causation.

1)         The cause must precede the effect

2)         The cause and effect must covary

3)         No plausible alternative cause exists

Covariation is therefore an essential step in proving causation – cause and effect must vary in sync with each other – but not the definitive one, as spurious relationships can exist between events.

A spurious relationship (or spurious correlation) is one where events covary even though they are not causally related (for example Fig. 1). Some covarying events are linked via a “confounding” or “lurking” variable.

For example, ice cream sales may covary with swimming pool drownings, whereby the lurking variable is outdoor temperature. In this instance changes in sales and drownings are said to be dependent (dependent variables) on changes in temperature. Temperature is the independent variable: sales and drownings depend on the outdoor temperature and not vice versa.

Covariation between two independent variables indicates either a cause-effect relationship or fortuitous covariation. An example of the latter is presented in Fig. 1, which illustrates the need to investigate alternative plausible causes before rushing to conclusions (and asking Mr. Cage to stop making movies).

The analyzed time series should always be expanded to include the maximum amount of relevant data, for example to the period before Mr. Cage’s acting career began.

Sherlock Holmes famously stated “When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth.”. A Holmesian fallacy occurs however when all plausible alternatives have not been ruled out.

The IPCC climate change study

The main focus of the IPCC AR5 climate change attribution study ^[1] was to determine the relative contributions of Natural versus Anthropogenic causes to the GMST changes, with special attention dedicated to the observed increase of 0.6 °C in the 1951-2010 period.

Their attribution analysis used climate models to simulate the changes in temperatures caused by the changes in Earth’s physical processes, termed “climate forcings”. Two Natural Forcings (NAT: Solar Irradiation, Volcanic) and two Anthropogenic Forcings (ANT: Radiative Forcing due to ‘Greenhouse Gasses’ (RF, GHG), Other Anthropogenic (OA)) were modeled.

Their analysis (Fig. 2) indicated that the 1951-2010 global warming was caused by changes in RF: the variance of the GMST data could only be explained by the changes in RF (GHG; Fig. 2). The study ignored changes in Geothermal Forcing, thereby unintentionally committing a Holmesian fallacy.

Fig. 3 summarizes the time series used in this analysis, which as much as possible were obtained from the same data sources used in the IPCC AR5 study.

whereby C₀ is a reference concentration (in ppm) – here taken to be the 1951-1980 average of 322.4 ppm – and ΔC is the concentration change in ppm. Note that RF=0 for 1967, the year atmospheric CO₂ exceeded the reference concentration, and that therefore RF < 0 between 1880-1966 and RF > 0 between 1968-present

Covariation between two variables is commonly investigated via scatterplots. A scatterplot – also termed cross-plot, XY-plot, etc. – plots the value of the (presumed) independent variable along the horizontal axis and the value of the (presumed) dependent variable along the vertical.

A linear regression model, such as the Ordinary Least Squares (OLS) regression models in Fig. 4, tests whether the dependent variable varies linearly as a function of the independent variable.

The model’s “goodness of fit”, that is how well the linear regression model’s values compare to the dependent variable data, is often indicated by its coefficient of determination, denoted r², which represents the percentage of dependent variable variance that is predictable from knowledge of the independent variable.

Figure 4 shows the scatterplots and OLS regression models for GF and RF versus GMST anomalies. Both regression models were forced through (0,0) as this is expected (no forcing results in no anomaly), which did not significantly alter either their regression equation or their r² values.

Both show good, significant (p-values < 0.0001) correlations, though the RF model (r²=0.9) possibly explains more of the GMST anomaly variance than the GF model (r²=0.78). Fig. 4 illustrates a problem: the two independent variables RF and GF cannot both explain a majority of the GMST anomaly variance. At least one of the correlations must therefore be spurious.

1951-2022 correlations

Figure 5 presents the 1951-2022 period crossplots between GMST Anomaly (dependent variable) versus four independent variables: Nicolas Cage’s age (born in 1964; negative ages pre-1964), the population of Texas (Source: ^[9]), RF, and the decrease in Total Geomagnetic Field Strength (GFS) near Winnipeg, Canada (Lat=50, Long=100, reference date: 1967; Source: ^[10]), a location near the geomagnetic field maximum over North America.

Changes in North American GFS are a proxy for GF for the 1951-2010 period (Part 2). All linear regression models for the correlations show a good fit, though clearly at least 3 of the four correlations must be spurious.

The Nicolas Cage age, Texas Population, and RF plots are remarkably similar to each other and to the GMST Anomaly vs Year plot (Fig. 3) for the 1951-2022 period. These good correlations – and their similarities – are therefore the result of a “lurking” variable, namely time, indicating the relationships are potentially spurious.

Any linearly (Nicolas Cage’s age) or slightly-exponentially (Texas Population, RF) increasing time series will show a good correlation, as all can be well-approximated by a linear function that varies with Year.

The high r² values of all correlations implies that testing causation hypotheses such as “Nicolas Cage’s increasing age is causing global warming” must fully rely on demonstrating there are no plausible alternatives, that is demonstrating that no other physically plausible planetary processes exist.

Demonstrating that something does not exist – Bigfoot, Aliens, etc. – is often a tough challenge, as it can be very hard to prove a negative.

Physical processes however always involve energy exchanges, so any “plausible alternative” planetary process must involve changes in planetary energy exchanges on the order of 0.1 W/m^{2 [11]}, which effectively rules out Nicolas Cage Forcing as well as Texas Population Forcing: both correlations are spurious.

Estimates of RF (Fig. 3) indicate its changes over the 1951-2022 period are of the right magnitude, though its similarity to the other spurious correlations, as well as its spurious correlation to the Arctic temperature anomalies (Part 1 of the series), warrants further investigation of its plausibility (below).

Part 2 of this series demonstrated that changes in GF have resulted in increases of ocean floor heat flux from ~60 mW/m² (average ocean floor) to ~160 mW/m², that is increases on the order of 0.1 W/m², as well as increased submarine volcanics and hydrothermal venting, which locally increase power densities by multiple orders of W/m².

Changes in RF and GF are therefore both plausible causes for the recent global GMST increases.

The 1880-2022 GMST-RF correlation is spurious

The 1880-2022 GMST data ^[3] can be divided into a number of distinct periods of change:

• 1900-1910 strong decrease of roughly 0.35 °C
• 1910-1942 increase of roughly 0.5 °C
• 1942-1950 decrease of roughly 0.2 °C
• 1950-1976 period of “muddle”, whereby GMST anomalies varied irregularly between -0.2 °C and 0.15 °C
• 1976-2005 strong increase of 0.6 °C
• 2005-2013 period of “muddle”, whereby GMST anomalies remained relatively constant around 0.65 °C
• 2013-2018 strong increase of roughly 0.2 °C
• 2018-2022 small and possibly insignificant decrease of roughly 0.03 °C

whereby the quoted changes represent the LOESS-averaged changes from Fig. 3.

Fig. 6 demonstrates the poor covariation between GMST Anomalies and RF for RF <0: the OLS model r² drops to 0.33 for the pre-1967 values, indicating this essential condition of causation has not been met for the pre-1967 period.

The OLS relationship of Fig. 4 does a very poor job both mathematically and visually of predicting either increases or decreases prior to 1967. For example, during the 1910-1942 period the OLS model predicts a 0.12 °C GMST increase for the measured 0.185 RF increase, though in reality a 0.5 °C increase was measured, indicating RF was not the dominant climate forcing during this period.

In addition, because RF has been monotonously increasing since 1750 ^[5], it cannot be responsible for either of the two periods of temperature decrease, indicating a different process, e.g. heat radiation to space, dominated during the 1900-1910 and 1942-1950 periods.

A possible explanation might be Volcanic Forcing, which typically limits solar irradiation for 2-3 years following eruptions ^[1]. However, most of the large volcanic eruptions between 1880-1967 have occurred after a GMST decrease had already started, or during periods of GMST increase.

For example, the only significant (though still rather unimportant) volcanic eruptions during the 1942-1950 and 1950-1967 periods of temperature decrease and stability (resp.) were the rather smallish Bezymianny, Kamchatka (1955) and Mount Agung, Bali (1963) VEI=5 eruptions, both much smaller than the Quizapú, Chile VEI=6 eruption that occurred in 1932 during a period of GMST increase.

Changes in RF were minimal in 1750 ^[7], so Natural Forcings must have dominated climate change for a certain period after. According to the IPCC, at some point between 1750 and 1951, a changeover occurred whereby Anthropogenic Forcing – RF – became the dominant climate forcing.

Fig. 6 demonstrates that such a change definitely did not occur prior to 1950, the end of the last major recent global temperature decrease. Natural Forcings therefore caused the 1910-1942 increase of 0.5 °C, dominated the pre-1950 period, and likely dominated the period afterwards.

Part 1 demonstrated that changes in RF did not cause the current Arctic temperature anomaly, Earth’s largest temperature anomaly, nor can RF explain why both the 1910-1942 and current warming phases were largely confined to North America’s periphery:

“the most pronounced warming [occurred] in the Arctic during the cold season, followed by North America during the warm season, the North Atlantic Ocean and the tropics.” (IPCC; ^[1]).

The RF-GMST correlation is very likely spurious.

Geothermal Forcing was the dominant climate forcing between 1880-2022

Figure 7 demonstrates that GF -in contrast to RF – does covary with GMST between 1880-2022: periods of synchronous increase (red), decrease (blue), and “muddle” (yellow) occur over most of the period.

Figure 5 also highlights that when GFS is low (~0) that geothermal heating is dominated by a cooling process – heat radiation to space – which cools the Earth. In other words, similar to RF, GF is only a heating and not a cooling process between 1880-2022.

The Earth only cools when the supplied geothermal heat (GFS <= 0) is insufficient to compensate Earth’s heat radiation to space, which happened between 1968-1974. In contrast, Earth can never cool if RF is the dominant climate forcing and RF is increasing.

The OLS model in Fig. 7 (red line) is not the optimal predictive model, as it only predicts heating of the North American SWIC circuit periphery.

As such, it needs to be included into a global climate model, whereby this North American heat is distributed via weather, ocean currents etc., in order to obtain a better match to the global measurements. Note that GMST and GF are both decreasing in sync after 2018.

The continuation of this trend – i.e. post-2018 global cooling – will further disprove ‘anthropogenic climate change’.

Summary

Part 1 in this series demonstrated that the Arctic temperature anomalies are not caused by IPCC-modeled climate forcings, such Anthropogenic Forcing or Polar Amplification.

Part 2 documented that these anomalies are caused by an IPCC-ignored climate forcing: Geothermal Forcing.

Part 3 demonstrated that these Arctic learnings can be expanded to the rest of the Earth: Anthropogenic Forcings are not dominating climate change.

References

[1] Bindoff, N.L., et al., 2013: Detection and Attribution of Climate Change: from Global to Regional. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.

[2] https://tylervigen.com/spurious-correlations

[3] https://data.giss.nasa.gov/gistemp/tabledata_v4/GLB.Ts+dSST.txt

[4] https://data.giss.nasa.gov/modelforce/ghgases/CMIP5/CO2_OBS_1850-2005.lpl

[5] https://gml.noaa.gov/ccgg/trends/data.html

[6] Mason, P., 1998, Energy Futures and Future Transport. ACE 614. Doi: 10.13140/RG.2.1.4085.9768

[7] https://gml.noaa.gov/aggi/

[8] Myhre, G., Highwood, E.J., Shine, K.P., Stordal, F., 1998, New estimates of radiative forcing due to well mixed greenhouse gases. Geophysical Research Letters. 25: 2715–8. doi:10.1029/98GL01908.

[9] https://www.macrotrends.net/states/texas/population

[10] https://www.ngdc.noaa.gov/geomag/calculators/magcalc.shtml#igrfwmm

[11] Loeb, N.; Johnson, G.; Thorsen, T.; Lyman, Jo.; et al., 2021, Satellite and Ocean Data Reveal Marked Increase in Earth’s Heating Rate. Geophysical Research Letters. 48. doi:10.1029/2021GL093047

Header image: Alamy

Please Donate Below To Support Our Ongoing Work To Defend The Scientific Method

PRINCIPIA SCIENTIFIC INTERNATIONAL, legally registered in the UK as a company incorporated for charitable purposes. Head Office: 27 Old Gloucester Street, London WC1N 3AX. 

Trackback from your site.