In light of the current climate negotiations in Paris, I thought it might be interesting to provide a (reasonably) simple explanation of how we know global warming is driven by greenhouse gas emissions. The science has been developed for over a hundred years, and is now understood in considerable detail.

I will begin by showing graphs of atmospheric greenhouse gas concentrations and global temperatures, and then proceed to describe how the changes in each are connected. The final plot reveals the differences we would expect if greenhouse gas concentrations were held constant. It is a crucial plot that underscores why scientists are so confident in their assessment of global warming.

Let’s dive in.

Carbon dioxide (CO₂) is one of the greenhouse gases emitted by human activities, in this case the burning of fossil fuels (coal, oil, natural gas). Other greenhouse gases include methane (the primary component of natural gas) and many refrigerants. To keep things simple, we will concentrate on CO₂ alone.

Figure 1 shows atmospheric CO₂ concentrations since the year 1700. The data until 1958 were obtained from bubbles of air trapped in antarctic ice sheets (the deeper you go, the older the bubbles get), and thereafter they are from regular gas measurements at Mauna Loa, Hawaii. The curve shows an exponential increase beginning in the late 1700s coinciding with the industrial revolution. Atmospheric CO₂ concentrations have now exceeded 400 ppm, a 44% increase from the pre-industrial average. That’s an incredible amount of change.

Atmospheric temperatures have also been rising. Figure 2 shows the annual global surface temperature change (“anomaly” from the forty-year average beginning in 1880) measured using thermometers since 1860. The plot shows that the earth has warmed about 0.85℃ since 1920, although pauses between 1940-1980 and 1998-2012 occurred. Not shown on the plot are the record high temperatures of 2014. We are well on our way to another record this year.

Note from Fig. 2 that there are short-term variations on top of the warming trend. This is the so-called “natural variability” of the climate system.

The two plots reveal a correlation between CO₂ concentration and temperature increases. To firmly connect the two we need to understand the science behind the “greenhouse effect”.

The greenhouse mechanism was proposed in 1896 by Svante Arrhenius following advances in basic physics. His idea was very simple. Visible radiation (i.e., “light”) from the sun is absorbed by all things on the earth, which warm in response. If you have experienced the heat of the sun’s rays as it emerges from behind some clouds then you know exactly what I am talking about. All warm things emit infrared radiation. You can “see” this radiation by using an infrared camera.

Now, if earth had no atmosphere then the upward-travelling infrared radiation would be lost to space. However, greenhouse gases in the atmosphere (CO₂) absorb infrared radiation and send a portion of it back to Earth. This double dose of radiation causes the earth to warm further. It is very fortunate for us that it does: Earth would be a very cold place without the greenhouse effect.

But here we come to the crux of the matter: The addition of more greenhouse gas can be expected to make the earth warmer still. We have been adding greenhouse gases through our burning of fossil fuels, and global warming is the expected result.

The full climate system is more complicated than the simple ideas of Arrhenius. Modern climate models (i.e., supercomputer simulations) include the effects of greenhouse gases, clouds, aerosols, solar cycles, oceans and more. Many such models exist around the world – they are all slightly different – and each can be used to produce multiple simulations of the earth’s climate.

Figure 3 shows a comparison between the measurements (in black) and the models. The thin yellow and blue lines show the results of individual simulations from two different groups of models (CMIP3 and CMIP5). The thick red and blue lines show the model averages.

The comparison shows that the models do an excellent job capturing the overall warming trend. The model averages are quite a bit smoother, but this is no surprise. If we could average the climates from many earths then the measurements curve would be just as smooth, and the agreement would probably be even better.

In some cases the models do pick up the finer detail. For example, the sudden temperature drop following the 1991 eruption of Mount Pinatubo is well captured.

Now, the great thing about climate models is that we can run experiments under different conditions. We can, for example, turn off greenhouse gas changes just to see what will happen. The results of such an experiment are given in Fig. 4. Poor agreement between the measurements and models is found after about 1970. In other words, changes in greenhouse gases, and in particular CO₂, are needed to explain global warming over the past 40 years or so.

I can’t emphasize enough how important this point is. The models show that neither natural variability nor solar cycles – both included in the modelling of Fig. 4 – can explain the warming since 1970. Both play a relatively minor role compared to the impact of our CO₂ emissions. This is why the scientific consensus is that

“It is extremely likely that human influence has been the dominant cause of the observed warming since the mid-20th century”.¹

---

1. See IPCC AR5 WGI, pg. 17. For more information on the attribution of climate change, see Chapter 10 in the same volume.↩