Climate models are the main tool climate scientists use to predict how Earth will respond to more heat-trapping pollutants in the atmosphere.
But what exactly is a climate model? Let’s start off easy by breaking down the phrase “climate model.” The “climate” is simply the weather averaged over a long period of time. A “model” in this case is a physical approximation of a complex system. So a climate model is an approximation of the Earth’s weather over a long period of time.
Since their debut in the 1960s, scientists have been improving and increasing the complexity of climate models (check out my History of Climate Models blog), and my colleagues and I at UCS continue to use them today.
General circulation models
When climate scientists reference a climate model, they are generally referring to a general circulation model (GCM), which is the main tool climate scientists use to simulate and understand how the Earth’s oceans, land, atmosphere, and cryosphere (a word to describe the planet’s sea and land ice) respond to changes in both its own internal dynamics as well as changes in heat-trapping pollutants.
Just by looking at the name, you can see that a GCM is a model that simulates the circulation of Earth’s different physical systems like the atmosphere and ocean. What causes a circulation? In my blog on the potential collapse of the Atlantic Meridional Overturning Circulation (AMOC), which is the conveyor belt of water moving in the Atlantic Ocean, I discussed how regions around the equator are warmer than the poles due to different amounts of incoming solar radiation, that is, energy from the sun.
The Earth’s climate system doesn’t like imbalances in heat given the difference in density: Earth will do everything in its power to mix the cold poles and the hot tropics. The Earth’s atmosphere and oceans create circulations in order to mix temperature differences between regions; GCMs, or climate models, simulate these circulations quite well.
How exactly do GCMs simulate circulations? In order to model the climate system, a GCM uses a set of equations that explains how energy, momentum (e.g., moving air), and water interact and change within the atmosphere and oceans. GCMs simulate the Earth as a giant three-dimensional grid and calculate how different variables (e.g., temperature, rainfall, etc.) change at each grid point. The models further simulate how heat and other climate variables travel to and influence values in other grid points.
A climate model is made up of many models
In my blog on the history of climate models, I discussed how the first climate model back in the mid 20th century was actually just a single model of the atmosphere, which is just one part of the climate system. We know that there are other components of Earth’s climate besides the atmosphere, for example, the ocean, the land, and ice. Today’s climate models are so complex because they are made up of all of these components: atmosphere, land, ocean, and ice. We also have scientists who specialize in each component, allowing for further complexity and improvement in prediction of the Earth’s climate system. Today, a climate model is made up of smaller, component models of the atmosphere, ocean, land, and cryosphere.
How exactly do all these different components of Earth’s climate system communicate with each other while a climate simulation is running? Through something called a coupler, which connects the different model components so that data can easily flow between the different sub-models.
Why do we need so many different models? Each model simulates something specific in its respective system. An ocean model calculates ocean circulation (like the AMOC) as well as ocean biogeochemistry, which is the science of how different molecules, such as carbon or nitrogen, cycle through the ocean. A land model will simulate:
- vegetation
- snow cover
- soil moisture
- evapotranspiration (process by which water moves from the land surface or vegetation to the atmosphere)
- river flow
- and carbon storage
A sea-ice model will calculate
- reflection of incoming sunlight
- air-sea heat exchange
- and moisture interaction between ice and water
An atmospheric model calculates changes in
- atmospheric circulation
- radiation
- clouds
- and aerosols
Model parameterizations
You might be thinking, how could we possibly simulate clouds if they’re created from many tiny water droplets and ice crystals? If we were to simulate a cloud and all of its tiny droplets, our three-dimensional grid would have to be extremely detailed. Unfortunately, we don’t have the computer power to perform these kinds of detailed calculations (we also don’t fully understand the dazzling complexity of all the physics involved), so scientists developed something called a parameterization. A parameterization can be thought of as a model within a model.
Let’s say there’s a cloud in the eastern Tropical Pacific Ocean near the Galápagos Islands. This cloud exists under certain atmospheric conditions (temperature, moisture, wind) that support its existence.
If we were to simulate this cloud in a GCM, these atmospheric conditions would first be reported to the cloud parameterization scheme from the main atmospheric model. The parameterization then calculates certain properties of the cloud, like how much sunlight the cloud reflects or how much cloud coverage there is in the cloud’s surroundings. The parameterization then reports back its findings to the main atmospheric model, which allows for continuous communication between the main atmospheric model and the parameterization to follow the cloud through its lifecycle.
Many small-scale processes are parameterized in GCMs. Beyond clouds, air quality and turbulence are also parameterized. Turbulence is just the word for abrupt, small-scale changes in wind (think of being in a plane and suddenly experiencing a bump, or playing frisbee in a park and the frisbee changes direction or elevation as it suddenly experiences a gust of wind).
What are climate models used for?
The obvious use for climate models is to predict how the Earth’s climate may change given a “forcing” applied to Earth’s atmosphere. A forcing is typically a change in the composition of Earth’s gases in the atmosphere or a change in incoming solar radiation that leads to a radiative imbalance.
What do I mean by this? A key feature of the Earth’s climate system is that it is always trying to maintain equilibrium—that is, the energy coming into the planet must always equal the energy leaving the planet. Why? Because the whole of the Earth’s climate system is subject to the laws of thermodynamics: energy in = energy out. But if the composition of gases in the atmosphere changes, then this can affect the energy balance.
When CO2 is added to the atmosphere, an energy imbalance is established, and the only way to reach energy equilibrium again is for the planet to warm up. This is why the Earth is warming in response to added CO2 in the atmosphere.
In the 1960s, it started to become clear, with the help of climate models and theory, that fossil fuel use would warm the planet. The National Academy of Sciences released The Charney Report in 1979, which used climate models to predict, and warn the U.S. government, that the planet would warm due to fossil fuel emissions (though the U.S. government was warned about global warming as early as 1965). The authors estimated that the world would warm 3°C (5.4°F) given a doubling of atmospheric CO2 based on their climate model simulations in the 1970s.
But this is just one example. You could use a climate model to ask any question that would affect the climate system: “What would happen if the Yellowstone supervolcano erupted?” “What if the sun disappeared for five days?” “What if all atmospheric nitrogen was removed?” You can also construct a climate model with any arrangement of continents—for example, a climate model to represent Pangea Earth or a “Waterworld” planet with no continents at all. Some scientists even built a climate model to simulate the climate of Westeros from the Game of Thrones TV show.
Today, climate models are so complex that we can study how climate may be changing on a more regional level. In my research, I’ve run climate models to study how drought in the U.S. Northeast is changing with climate change, how the Earth may start to rapidly warm in the near-future given a change in oceanic warming, and how precipitation patterns might shift in the Southwestern U.S.
Climate models will continue to become more complex and more accurate
GCMs are complex, made up of multiple sub-models, and have a few parameterizations. They have been improved on for decades and are the combined work of climate scientists, physicists, mathematicians, and computer scientists. They’re also incredibly accurate—model simulations run in the 1990s predicted how much the Earth would warm by 2025, which matches our current observations.
In the future, climate models will become even more complex, perhaps resolving small-scale features, like clouds, rather than parameterizing them. We need these improved climate models to better predict and reduce uncertainty of regional climate change. The more scientists can equip society and decision makers with the best available climate science, the more we can sufficiently respond, adapt, and prepare for the changes underway.