It is often helpful for climate scientists to think of the Earth as a system, composed of several components, or smaller systems: the atmosphere, the oceans, land, the biosphere, and the cryosphere (glaciers and ice). The interactions between these components can be very complex, and this complexity makes it extremely difficult to know how and why climate will respond to changes. Computer models allow us to conduct simulations of climate – including such things as motion in the atmosphere and oceans, and exchanges of gases, heat, and water between each of the components of the system. 

Computer models are the only means we have of testing the sensitivity of the climate system to various factors that might cause changes. These may include anthropogenic (human-caused) factors, such as an increase in carbon dioxide or other greenhouse gases, or changes in the land surface, or factors that are completely natural. These would include small changes in Earth’s orbit, changes in the intensity of the Sun, volcanic eruptions, changes in the positions of continents, or even natural fluctuations in greenhouse gases.

Climate models in general range in complexity, from very simple models, that only examine the effects of one factor (such as a change in the sun) on the global average temperature, to very complex models that simulate all of the interactions between the various components of the Earth system. Any change in one component of the climate system could affect things that are happening in every other component.

Simulations of Past Climate

Complex climate models are especially useful for studying past climates, because they allow us a means to try to recreate the world as it once existed. Generally, models can help us in two specific tasks: (1) help us explain the geologic data that might indicate a particular type of climate by allowing us to experiment with the various factors that could have been affecting climate at a particular time period, and (2) models can be used to fill in the gaps in the geologic data, and provide field researchers with clues as to where they might find additional data.

Simulations of Eocene Climate

To reconstruct the Eocene climate, we used a global climate model developed at the National Center for Atmospheric Research. This model is most commonly used to simulate present day and future climate. We moved the continents to their approximate locations in the early Eocene, 50 million years ago (see figure). The climate model uses a gridded format, so the continents lose their smoothed out features and become "blocky" (see figures for comparison). The challenging part is how to best set up the simulation to depict the conditions during the Eocene. The geologic record suggests that it was very warm in the early Eocene – even at high latitudes. But what factors would make it warm at high latitudes?

One of the most likely possibilities is that there was a high level of greenhouse gases in the atmosphere at the time. This is not something we know for certain. This is something we must infer based on the geologic record. Because we don’t know for certain, we must run several test simulations – each with a different level of greenhouse gases in the atmosphere. If one of the test simulations comes close to reproducing the environment depicted in the geologic record, we might conclude that one simulation is more realistic for the early Eocene than the other. However, it’s not always that simple. Sometimes a simulation will accurately depict the geologic data in one part of the world, but not in another part. So, we must continue refining our simulations to better reproduce the climate, and, hopefully, better understand the mechanisms driving climate changes at the time.

The model simulations presented in this activity are two test simulations of Eocene climate. In one simulation, carbon dioxide level in the atmosphere is set to 560 ppm (the present day level is 375 ppm), and in another, the carbon dioxide level is set to 2240 ppm. Everything else about the two simulations is identical. The interface here will allow you to compare the climatology in the two simulations, along with the climatology from a simulation of present day climate.