How Global Warming Became Scientific Fact

I believe that global warming as a result of human activities is a scientific fact. How do I know this? The reason is this conclusion has resulted from the same process of cultural evolution that has led to every bit of factual knowledge we humans have (and we have a lot of it). This process is one in which status-seeking individuals and groups participate in a competitive process, through which the best ideas (i.e. those that “worked” best and so were copied by others) won out. The cultural evolution process relevant to global warming is the scientific method.

It operates using individual learning combined with direct, prestige and frequency biases. Scientists make observations and formulate conjectures through individual learning and then communicate their findings to other scientists through informal conversations, internal reports, or published papers. Results other scientists find interesting and relevant to their own work are noted (direct bias). Upon further analysis, the original researchers or others produce a testable hypothesis, which may be reported. Predictions from the hypothesis may then be tested against additional observations or experimental results. If the testing is successful and is published in an important journal more people pay attention (prestige bias). Depending on the salience of the result and the prestige of the one reporting it, it may be examined by others. If the finding pans out, it can lead to career success and acquisition of prestige for the initial discoverer, which provides the motivation for hypothesis formulation and verification. If the new knowledge “works” in new hands, the hypothesis becomes a theory. Others will start to use it and it will spread though the community working in this area through frequency bias, until after being used successfully by many people in a variety of situations that provide additional testing, it becomes scientific fact. Scientific fact is not the same as an observational fact such as granite is harder than limestone, but rather confidence in the validity of the predictions obtained from the theory. For example, if the theory predicts the spacecraft will intercept the asteroid at a future time, when that time arrives, the craft will be at the asteroid.

The following sections give a brief account of how the idea that rising levels of CO₂ cause rising average temperature for the Earth evolved via the scientific method from a conjecture to a scientific fact.

The potential for the greenhouse effect was discovered following a number of other scientific discoveries in the 19th Century. Recall that a prism breaks light into a spectrum of different components based on its wavelength, from long (red) to short (violet). The invention of the spectroscope in 1860 allowed scientists to measure the absorption or emission of light energy as a function of wavelength by various materials. Distinctive absorption and emission patterns characteristic of various elements were discovered, which allowed for the identification of their presence in materials through which light passes. For example, French astronomer Pierre Janssen inferred the presence of an unknown element in the sun’s atmosphere (corona) in 1868 based on a bright emission line that did not correspond to any known element.  This hypothetical element was named helium after the sun god Helios. Fourteen years later it was discovered on Earth, confirming Janssen’s discovery.

Over the same period Scottish theoretical physicist James Clerk Maxwell established that visible light was an electromagnetic wave with a wavelength (λ) measured in tenths of a micron. There were many kinds of electromagnetic “light,” or, more accurately, radiation, which was categorized by wavelength. Electromagnetic radiation (EM) with λ longer than 0.7 microns (red) was infrared radiation (IR). Ultraviolet radiation (UV) has λ shorter than 0.4 microns (violet). Instruments like the spectroscope, but which measured absorption of other kinds of EM, were possible once suitable detectors were developed. It was realized that all bodies radiated EM, which at ambient temperatures is all IR; only hot things like fire, molten steel, or the sun, radiate much visible light.

The reason why is that, relative to visible light, IR is lower energy, while UV is higher energy. IR absorption is associated with low-energy vibrational and rotational movements of molecules, whereas visible light, UV are associated with high-energy processes involving electrons changing their orbital positions in atoms. Most locations on Earth are too cold to provide sufficient energy for electrons to shift orbitals, but plenty warm enough to rotate or vibrate molecules. A molecule with more than handful of atoms will have a large number of ways it can vibrate or rotate and so will absorb IR radiation at a great many wavelengths, making it more opaque to IR.

Glass is an example of a material that is transparent to visible light, but not IR. The differential ease of transmission of visible versus IR light in glass is why your car gets so hot when parked in the sun during the summer. Sunlight passes through the transparent glass and gets absorbed by the car interior. The absorbed energy is radiated out as IR, but since glass, like most substances, is not transparent to IR, only a portion of the IR passes through. The rest is absorbed by the glass and radiated back into the interior. The car interior must heat up further to increase its overall IR emissions enough so that the fraction of IR energy making it through the windows matches the energy of the sunlight entering the car. Thus, a car with glass windows, if placed in sunlight, will warm up. This phenomenon is called the greenhouse effect, because a greenhouse is a glass house that becomes warmer than its surroundings by trapping the energy of sunlight.

On the other hand, were the windows made of salt (NaCl) plates, they would be transparent in both directions and there would be no greenhouse warming of the car.  Simple ionic crystals like salt can be relatively transparent to IR radiation. Similarly, monatomic and diatomic molecules like nitrogen, oxygen, and argon (which are major atmospheric constituents) are also transparent to IR; they act like salt plates. Larger molecules present in the atmosphere like CO2, water, and methane have more ways of wiggling and so are not IR-transparent. Their presence will make the atmosphere translucent to IR, causing the surfaces underneath it to warm, like the car with glass windows.

The idea that changing the IR absorption characteristics of the atmosphere can affect global temperatures is something that was possible to anticipate using the physics known by the end of 19th Century. French scientist Joseph Fourier, who in his investigation of heat flow, calculated that the Earth would be much colder than it is if the incoming radiation from the sun were the only warming effect. He conjectured that the Earth's atmosphere acts like an insulator.  Swedish physical chemist Svante Arrhenius hypothesized that this insulating effect arose from the effect on the Earth’s radiation emission by atmospheric components, noting that:

the selective (IR) absorption of the atmosphere is…not exerted by the chief mass of the air, but in high degree by aqueous vapor and carbonic acid (CO₂) which are present in the air in small quantities.

The atomic/molecular phenomenon which I used to explain how materials interact with visible and IR light would not be known for decades. This paper shows how Arrhenius and the other scientists of his day were trying to think through how all this absorption and emission stuff works, and were speculating on potential consequences of these phenomena. Arrhenius, with only fragmentary information, saw how an atmosphere transparent to visible light and translucent to IR would lead to warming, and how small changes in minor atmospheric components like CO₂ (what we call greenhouse gases) could produce large climate effects that might explain things like the ice ages. He then went through a series of calculations, attempting to estimate the size of the effects of changes in greenhouse gases. Here we see the work of a number of scientists producing a hypothesis.

Arrhenius’s paper stimulated follow-up work. In 1900 Knut Ångström published results from laboratory experiments which showed that CO₂ wouldn’t be a very significant greenhouse gas after all.  It seemed that, at the levels the gas was present in the atmosphere, the absorption bands were already saturated and so insensitive to further increase in CO2 levels. This work led most scientists to conclude that Arrhenius’s idea that changing CO₂ level could affect global temperatures was wrong, and most meteorologists to conclude that the greenhouse effect does not happen. This episode shows the action of the scientific method and that the response of the scientific community to the idea that human production of greenhouse gases affects climate was exactly the same as that of the most ardent climate skeptic today. Despite the efforts of researchers like British engineer Guy Stewart Callendar in the 1930’s, who empirically demonstrated a relation between CO₂ and global temperature, the scientific consensus was that the IR-mediated greenhouse effect itself did not happen.

After WWII, Defense Department money became available for the scientific study of atmospheric properties, in support of the supersonic flight program. Already, it was known that the crude experiments of Ångström involving IR absorption by air containing various amounts of CO2 in the laboratory had been misleading. Instead of broad absorption bands, more precise measurements found groups of sharp lines where absorption would occur, with gaps in between where the infrared would get through unhindered.  Was this relevant?

By 1956, further research had established Arrhenius’s qualitative finding of warming with increasing CO₂.  These results were obtained without considering potential effects on cloudiness and humidity resulting from changing levels of CO₂. It was not until 1967 that Manabe and Wetherald provided an account of the impact of changing CO₂ on global temperature that included these factors.  Assuming no change in atmospheric water concentration with changes in CO₂ level, they found a 1.3ºC temperature increase for doubling CO₂ levels, regardless of the presence of clouds. This value can be thought of as the “pure” greenhouse effect of CO2, without considering any other factors known to impact climate. When they allowed water level to change with CO₂ level so as to maintain the same relative humidity, they found values of 2.9º C and 2.3º C for no clouds and average levels of clouds respectively.

Over this same period, accurate measurements of atmospheric CO₂ level had become technically feasible. Continuous monitoring of CO₂ began in 1958, and within just a few years it was clear that levels were rising. Figure 1 shows the rise in CO₂ measured since 1959 and a smoothed temperature trend since 1945. Global temperature reconstructions up through the late 1960’s had shown that temperature had risen during the first four decades of the 20th Century, but then had been roughly flat or slightly down since 1940.

Figure 1. CO₂ and temperature  rise since WW II.
Temperature decadally smoothed using 21 point binomial filter. M&W refers to Manabe and Wetherald. The model is a logarithmic extrapolation from 1967 based on their 2.3 degree value for doubling of CO₂.

A decade of measurements had shown CO₂ levels rising 2.7% which, if extrapolated to cover the three decades since 1940, amounted to a 10% increase. Such a rise should have given a temperature increase of 0.3º C since 1940 according to Manabe and Wetherald’s model―yet none had been seen. That rising CO₂ would produce warming temperature through its impact on IR absorption was a result obtained by application of well-established theories of radiation absorption and could not be easily discarded. Given the results of the radiation calculations, it was further hypothesized that the greenhouse effect was being offset by other, cooling factors that were not yet understood. Researchers noted that their model employed a very simple approach to the issue of humidity and a physically unrealistic accounting for the effects of clouds. And there could still be other factors not yet considered. Although greenhouse warming could well become a problem in the future, at the present it was not.

The situation changed when warming resumed in the mid-1970’s. Later research revealed that the injection of aerosols into the atmosphere from industrial emissions exerted a cooling effect like that produced by volcanic eruptions. The introduction of widespread pollution controls reduced this impact allowing the underlying warming due to the greenhouse effect to manifest. Since 1975, Figure 1 shows a fairly linear rise in temperature of 0.18 degrees per decade. The model of Manabe and Wetherald also shows a fairly linear rise with a rate of 0.16 degrees per decade. In other words, once warming resumed, it has proceeded in accordance with the predictions of what is, in today’s terms, a very simple model, for nearly half a century.

Figure 1 shows the rising trend in CO₂ level on a logarithmic scale, showing a fairly linear trend. Logarithmic plots are used to represent exponential processes in a linear form. Since CO₂ production is directly related to industrial economic growth, which is usually described as exponential, it makes sense for CO₂ rise to be exponential and the log plot to be linear. Greenhouse warming predicted by models like that of Manabe and Wetherald, can be well approximated as a logarithmic function of CO₂ level. The logarithm of an exponential function is a linear function, so it is expected that the temperature rise would be linear given an exponential CO₂ trend, which is the case, as shown in Figure 1. Assuming these trends remain in place to the end of the century, CO₂ levels should reach around 600 ppm, with temperatures 1.4ºC warmer than today (2.4ºC warmer than before industrialization). Projections for CO₂ levels in 2100 range are high as 970 ppm.  For this level, the projected global temperature would be around 4.1ºC higher than before industrialization. This simple projection of 2.4° to 4.1°C temperature increase compares well with the 2-4°C consensus of modern sophisticated models.  Uncertainty over how much greenhouse gas humans will choose to put into the atmosphere is the largest factor affecting how high temperatures will rise in this century.

If one installs a light-colored screen just behind the windshield of a car parked in the hot sun, it will not get as hot as it does without the screen. If this wasn’t true nobody would do it. The fact that this works is because the greenhouse effect is a real thing one can experience for oneself. Given this fact, the consequences shown by Manabe and Wetherald’s model are unavoidable, unless some other factor intervenes to interfere with it, as pollution did in the mid-20th Century, or the windshield screen does in your car. But once that factor is no longer active, the planet will warm just as your car gets hot without the windshield screen.

The account given above shows the evolution of the present scientific consensus about global warming from 19th century speculations to hypothesis in 1896 and to a theory in 1967, which was confirmed as established scientific knowledge in subsequent decades by the correctness of its predictions as shown in Figure 1 and subsequent discoveries about climate dynamics (e.g. discovery of predicted stratospheric cooling and elucidation of net cooling effects of aerosols) made possible by the greenhouse theory. That is, the greenhouse effect has become scientific fact.