How do we predict the consequences of climate change?


The Intergovernmental Panel on Climate Change, or IPCC, has not only projected rising temperatures and sea level; it is also made projections about rainfall, growing zones, changes to ecosystems, and more. How can scientists do that?

First, note that they use the word projection rather than prediction. A projection is an educated guess about the future based on what we know now. It’s widely understood that what we know is not perfect, so a responsible projection will offer a range of outcomes, not a specific prediction.

Even if what we know is not perfect, we do know a lot. We know the basic physics of the atmosphere. We know how carbon dioxide, or CO2, and other greenhouse gases trap heat. We know how changes in the brightness of the surface of the Earth — for example, when ice melts to expose rock or water — affect how much sunlight the Earth absorbs. We know how temperature differences between different regions of the Earth force air masses to move around . . . and so on.

We also know, from air bubbles trapped in ancient ice, how CO2 has affected the Earth’s temperature in the past. We know how high sea level was when the Earth was warm enough to melt all the ice on the planet, and how low it was in the depths of the ice ages. We know a lot about the distribution of vegetation during these very different climate regimes. And we know much more accurately than ever what conditions are like today almost everywhere on the planet — temperatures, ice cover, greenhouse gas concentrations, sea level, and more.

When scientists want to project what will happen by 2100, for example, they take all of this information and feed it into global climate models. These are complex computer programs that simulate major parts of the climate system, including the oceans, atmosphere, and land processes. Scientists then test the ability of these models to reproduce actual observations of the present and reconstructed information about the past. Once they see that the models can match observed and historical data, climate scientists can confidently use them to analyze future scenarios. They can “add” greenhouse gases virtually, in different amounts, and project the effects on rainfall, oceans, and more 50 or 100 years from now.

Scientists also test the models against nature’s own climate experiments. One major test involved the Philippine volcano Mount Pinatubo. It erupted in 1991, pouring huge amounts of particles into the atmosphere. The physics incorporated into the models said such dust should block the Sun and temporarily slow global warming — and that’s just what happened.

In sum, climate models are not perfect representations of the world. However, internationally recognized climate models are the best available substitute. As climate modelers gain more experience, and have the benefit of better data, their models keep getting more sophisticated, and their projections of future climate become more and more reliable.