Can We Predict the Future of the Climate?

crystal ball
Source: shutterstock

Despite the recent divisiveness of climate change in U.S. politics, it seems no one anticipated the firestorm that erupted from the publication of a New York Times op-ed column last month criticizing the certainty professed by climate change activists. The author, Bret Stephens, received more than 1,800 comments, many expressing anger or concern. The issue was not that Stephens was denying climate change (he specifically stated in the article that he was not). Instead, many accused the author of making factual errors and misrepresenting the current science. For instance, Stephens suggested that “Demanding abrupt and expensive changes in public policy” may not be as merited as climate change proponents claim. One key piece of evidence he uses to support this claim is the “sophisticated but fallible models and simulations by which scientists attempt to peer into the climate future.”


Stephens is not the first member of the popular media to doubt the reliability of climate models. In fact, there has been a great deal of controversy in the media surrounding climate models and how they should be used in public policy. Climate models are our crystal ball into the future, but as many skeptics argue, how can we predict the climate decades from now if weather forecasts for next week are usually wrong?

Furthermore, any attempt toward a reasoned discussion about the limitations and capabilities of climate models can easily slide down a slippery slope toward a debate over the broader issues of climate change. Climate models have been used to support the arguments of parties both for and against a strong national policy response to climate change. Advocates of such policies can easily point to the severe impacts of climate change predicted in some modeling scenarios, while their opponents can just as easily point to imperfections in the models themselves.

On top of all this, one thing is quite certain about climate models: without them, we would have no idea what kind of future climate to prepare for. Observations of surface temperature data show that the globe is overall getting warmer. Some impacts are obvious. Warmer temperatures mean more ice melting at the poles, which leads to sea level rise. Other impacts are much harder to predict because of the incredible complexity of the earth’s systems. For instance, changes in precipitation could affect how and where we are able to grow food. And for all impacts, of course, we want to know who will be affected and how bad it will be. That’s why we need climate models. Climate models give us a range of things we can expect, and improvements in technology are making that range smaller.

Source: IPCC 2013

Climate models do have their limitations. These limitations, though, aren’t cause to completely discard the information the models are giving us, but it is important to understand what these models can and can’t tell us, and to avoid oversimplifying these messages by either overstating their certainty or dismissing their importance to policy decisions.

Climate versus Weather

It is important to note that climate models are designed to predict future climate conditions, which are different from weather conditions. When we talk about weather, we usually mean changes that occur on a daily or monthly basis. This includes things like whether the sun is shining, whether it is raining or snowing, windy or calm, or if today was warmer or colder than yesterday. Climate, on the other hand, is the long-term average of short-term weather patterns. For instance, though the weather on a nice day in Portland, Oregon might be the same as the weather on a relatively cool, rainy day in Phoenix, Arizona, on average, the climate in Oregon is much cooler and rainier than the climate in Arizona, where the climate is often very hot and dry. Because weather can change rapidly and often unpredictably, it can be difficult to predict with any precision. Climate, on the other hand, lends itself better to prediction. For instance, although the tomorrow’s forecast for Portland might be wrong, a general prediction of the climate would say that most days in Portland, especially during certain times of year, are cool and rainy, and that most days in the future will continue to be cool and rainy, even if some days are hot and sunny. If over time, Portland began experiencing more warm sunny days than it did before, then it could be said that the climate at that location is becoming warmer and drier.

For a more in-depth explanation of the differences between weather and climate, see

How Climate Models Work

When we talk about climate models, we are talking mostly about computer programs that simulate conditions on earth. Many models, not just climate models, are constructed this way. For instance, consider making a computer model of a sink. This would be a very simple model with only three parts, the faucet, the sink, and the drain.


The first step in creating a model that works is representing the processes you want to model in mathematical terms so that you can use your model to calculate outcomes. In the sink model, the amount of water in the sink would be equal to the amount of water flowing in from the faucet minus the amount of water flowing out the drain. The amount of water in the sink is, therefore, a variable, which is dependent on two other independent variables, the amount of water flowing in and out. If you had the faucet running full blast but the drain was clogged, this model could tell you how much time you would have before your sink overflowed. Climate models work in much the same way, only they must account for many more variables.

What Do Climate Models Account For?

Climate models aim to account for all the known influences on the earth’s climate and predict, based on those influences, what the climate will be like at a given time and place. There are many climate models in existence today. They are developed and operated by scientists in countries across the world from a wide variety of institutions. There are also different types of climate models, each of which focuses on different sets of components of the earth’s climate. The components from which modelers can choose include the atmosphere, land surface, ocean, and sea ice. Think of each of these components of the earth’s system as “sinks,” except instead of just holding water, they also hold things like heat and carbon, and the arrows that represent the transfer of these things would connect all the sinks to each other. The difficulty is understanding all the variables in each component and how they relate to each other. The diagram below shows a simplified version of what a model of these interactions might look like. Keep in mind, each of the sinks and processes represented here must be quantified mathematically in order to be incorporated into the model. Also, because everything is related to everything else, one faulty equation could lead to the entire model not functioning correctly.







Atmospheric variables that can be included in climate models include the wind (which moves both horizontally and vertically), temperature, pressure, water (in the form of clouds, vapor, rain, snow, etc.), and chemistry. Physical processes that influence these variables include radiation (heat entering the atmosphere from sunlight or from the ground after it has been warmed by the sun), convection (movement of air caused by relatively warm air rising and cooler air sinking), condensation (the process by which clouds form), turbulence (small-scale wind patterns that help mix the air in the atmosphere), gravity wave drag (an effect produced by the earth’s gravity at very high levels in the atmosphere that helps account for the temperature and air flow patterns at those altitudes), and the influence of the land surface. For more information, go to


The ocean is always in motion. Ocean waters can transport things like heat, nutrients, and gases both upward and downward between the shallow and deep ocean waters, as well as around the globe. The figure below shows some of these complex dynamics, not all of which are explained here.




The ocean interacts directly with the atmosphere in many ways, including by the exchange of carbon, water vapor, and temperature. Water holds heat much better than the land. You might have heard of the “moderating” effect of the ocean on climate, which explains why coastal cities have much more consistent (and often rainier) weather than inland cities. Similarly, the ocean helps to moderate global temperature by transporting warm and cool water around the globe on what can be called the “ocean conveyer belt.” For more information on ocean currents and how they influence climate, check out




Land Surface

The land surface plays an important role in climate modeling because it interacts with the atmosphere and oceans in a number of ways. First, the transfer of energy (heat) between the land and the atmosphere must be accounted for. When the sun’s rays hit the earth, some of that heat is absorbed by the surface. Darker objects absorb more of the sun’s heat than lighter ones (ever noticed how hot blacktop pavement gets in the summer?). Lighter or more reflective surfaces (such as snow) might reflect almost all of the sun’s light and send it right back out of the atmosphere. Any time the surface absorbs some heat, though, it will re-emit that heat back into the earth’s atmosphere as an invisible form of energy called infrared rays. Where’s the sun’s light it good at passing straight through the atmosphere, infrared radiation is more likely to get intercepted by gasses in the atmosphere. Gasses are made up of invisible molecules, which you can think of as very tiny particles, just like dust or smoke except that they are made up of just a few atoms. The most common gasses in the air we breathe every day are oxygen and nitrogen, although other gasses that occur in smaller amounts are also important. These include the greenhouse gasses. Greenhouse gasses have been important to life on earth since long before we started thinking about climate change. Simply, without them, we would not be here. Greenhouse gasses are able to intercept the infrared radiation leaving earth and prevent it from escaping uselessly into the atmosphere. Outer space is deathly cold – greenhouse gasses keep our heat in so that we can survive on this planet. This is amazing, considering that greenhouse gasses account for only less than 1% of the gasses in our atmosphere. The concern over climate change arises from the problem of human activities producing excess greenhouse gasses. Less than 1% was plenty enough to keep the planet warm, but any more than that, even fractions of percentages more, and the greenhouse effect becomes stronger and the earth’s atmosphere grows warmer. To accurately model the earth’s climate, then, models must account for how much heat is both absorbed and emitted by the earth, and how much of this gets trapped inside the atmosphere. To do this, models must account for different land surfaces, such as ice, water, vegetation, or barren ground.




The land plays other important roles in climate as well. Water moves continuously between the land, oceans, and the atmosphere. Water vapor in the atmosphere, which develops when water evaporates from land or bodies of water. Once in the atmosphere, water vapor contributes to cloud formation, precipitation, and circulation of air, and water vapor is also a powerful greenhouse gas that helps warm the atmosphere.

Sea ice

Sea ice plays a huge role in earth’s climate system. As mentioned above, ice reflects the sun’s radiation. The light that is reflected (instead of being absorbed by the earth and re-radiated as infrared rays) doesn’t get trapped by the earth’s atmosphere, it just goes right back out into outer space. Ice, therefore, moderates the climate by preventing some of the sun’s heat from actually warming the atmosphere at all.

Furthermore, sea ice helps drive ocean circulation which, as mentioned above, helps distribute temperatures around the globe. Ice cools ocean water, and cool water sinks. When it does, warmer water flows in to replace it, but then it too cools and sinks. The process is repeated, and a continuous circulation cycle is created.


Modeling sea ice is difficult because, in modern times, the amount of ice is shrinking. That means that the reflectiveness of the earth’s surface is changing, as is the circulation patterns of the oceans.


Newer models have also been including the carbon cycle and other biogeochemical processes in their calculations. This includes how carbon moves between the land, atmosphere, and ocean. These processes add another significant level of complexity to climate models and are seen as a major improvement in technology.

How do climate models predict the future?

Climate model predictions simply take known information about the earth’s climate and try to use it to predict the future climate. They do this by looking at information from the past climate and trying to make their computer models reproduce the conditions we have observed. If a model successfully reproduces past climates, then that means that the conditions on which the models are built are realistic and that it can be used to predict future changes.

Challenges of climate modeling

Climate modeling is very complex. The factors outlined above are just a glimpse into all the processes that must be accounted for in a model. Recall as well that everything must be described in mathematical equations, which requires very specific knowledge of the process. Scientists haven’t yet learned everything there is to learn about the earth’s processes so knowledge can be a limiting factor on the accuracy of models. Some have argued that the sheer complexity of the earth’s climate makes it impossible to model. Fortunately, this isn’t quite true. The earth’s climate is not a chaotic system – to the contrary, it is quite predictable if we have enough information. Modeling is an opportunity to test how well we really know the climate system and to improve our knowledge of it if predictions don’t quite fit observations.

Furthermore, scientists may be limited by funding and computing power. The amount of data that must be run by a computer in a complex climate model is massive. It requires what we call “supercomputers,” which are expensive to own and operate. As a result, not all climate models include all aspects of the climate. Depending on what kind of questions a study wants to answer, they might not need to include chemical processes, only physical ones such as temperature exchanges.

Dealing with complexity and uncertainty

It’s true, models are experimental. Scientists are very aware that the predictions their models make might not come to be in exactly the way that they predicted. There are a few ways, though, that climate modelers can reduce the uncertainty of their predictions. One major way that modelers double check their predictions is by comparing their results with other models. The chart below shows a complete list of all the models used in the most recent report released by the Intergovernmental Panel on Climate Change (those included in the CMIP5 program) compared to those used in the previous report, the CMIP3 program.


Source: IPCC 2013


Notice that some models specialize in different components of the model. Because each model element is so complex, sometimes it is useful to focus on the greater accuracy of fewer components. Results between models focusing on different elements can then be compared before conclusions are drawn.

Scientists are very aware of uncertainties in their climate models. Simply, we don’t know everything there is to know about how the climate works. Critics of climate models often claim that this alone is enough reason to dismiss model predictions. They point to graphs showing that the observed global temperature has in recent years been lower than many model predictions. This doesn’t mean the models are wrong, but it is useful feedback for climate scientists. Noting shortcomings of models is the first step to improving them. Models have improved in many ways in recent years. Not only are they beginning to include more factors and processes than before, but scientists’ understanding of all processes is improving.

Why understanding climate models is important

Climate models aren’t magic. They can’t tell us what exactly will happen in our climate. They also can’t predict the future. These facts are thoroughly noted by critics of climate models, but it is also important to understand that climate modeling isn’t guesswork. Its complexity should be reassuring, not discouraging. Models are based on known information about a multitude of factors that affect the earth’s climate. The effects of these factors aren’t guesswork either, it is quantified by scientists working independently across the globe whose findings are constantly shared and compared to achieve the best accuracy. This article presents just a small glimpse into the work that these scientists do. Entire careers could be devoted to just understanding small fragments of climate models.

In this rapidly changing world, we should be welcoming any information that might help us better prepare for our future. Climate models provide us with some of the best information to this end that we could ask for.

Author’s note: To read a reflection on this article written by the author, click here.


Additional sources:

Flato, G., J. Marotzke, B. Abiodun, P. Braconnot, S.C. Chou, W. Collins, P. Cox, F. Driouech, S. Emori, V. Eyring, C. Forest, P. Gleckler, E. Guilyardi, C. Jakob, V. Kattsov, C. Reason and M. Rummukainen, 2013: Evaluation of Climate Models. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.