The most powerful computers on Earth are used to run climate models. Scientists use these models to understand how Earth’s climate works and to make predictions about how it might change in the future.
Climate models have successfully helped reconstruct climates of the distant and recent past, answering important questions, such as “what caused the last ice age?”. The proven ability of climate models to describe Earth’s past and present climates gives us confidence that they can simulate the planet’s future climate, too.
There are many research centres and thousands of climate scientists creating and fine-tuning computerized climate models worldwide. For example, the Met Office Hadley Centre for Climate Science and Services in England is one of the world’s leading climate research institutions. It has over 200 staff dedicated to climate research and uses clusters of supercomputers to create some of the most effective climate models in the world. Here in Canada, the federal Canadian Centre for Climate Modelling and Analysis (CCCma) has been creating excellent climate models since the early 1980s.
Climate scientists at institutions like the Hadley Centre and the CCCma combine scientific research and advanced computing to predict what the climate will look like in the decades to come.
General Circulation Models
The first computer models describing global warming were created in the 1960s. They calculated changes in the temperature of the planet by modelling the balance between the energy coming in from the sun and the energy escaping Earth’s atmosphere back into space. The scientific community has been improving on these first models for over 50 years now.
Modern climate models are called "General Circulation Models" or “Earth Systems Models.” They address much more than the sun-Earth energy balance. Working from the foundations of physics and chemistry, they take thousands of factors into account to model the entire climate system, including solar radiation, greenhouse gas emissions, volcanic eruptions, cloud formation, ocean currents, chemical reactions in the atmosphere, land use changes, and much more.
New climate models are constantly improving our understanding of the climate system and our improving understanding of the climate system is improving the models. The overall conclusions of the models have not changed—human greenhouse gas emissions are driving global warming.
How do we know climate models work?
Models allow scientists to test their understanding of how the climate works and how it might change in the future. But do the models do a good job of simulating the real world?
Climate models are extensively tested by “hindcasting”, which means modelling the climate of the past. Climate models are considered successful only if they can recreate to a high degree the averages, extremes, and seasonal patterns that match up with observed climate.
There is one major difference between predicting climate and predicting weather. A weather forecast attempts to create very precise hour-by-hour predictions on a very small scale. Climate models, on the other hand, effectively combine many simulations of possible weather to produce the accumulated story of what average and extreme conditions might happen. In other words, the goal of a climate model is to project changes in climates over years, decades and longer, whereas weather predictions are interested in what is expected to happen in the coming days.
Climate models do create simulations of day-to-day weather over many years, but in no way are they meant to be used to state what the weather is expected to be in the future, on any particular day or sequence of days. Instead, these daily simulations are interpreted statistically, resulting in statements about the probability of particular weather conditions being observed in the future. They do not produce weather forecasts; they produce climate projections.
And importantly, scientists keep track of how well models perform as the years pass. Since 1990, for example, the observed rate of global warming is well within the original range projected by climate models.
Why do we use so many climate models?
Canada is one of dozens of countries that has independently produced its own climate models. International organizations help coordinate all of these modelling experiments and gather them into an “ensemble”, or collection of many different models.
Why do this? Although each model is carefully designed to be consistent and plausible, working with an ensemble of many models lets us look at a range of future projections instead of just one. Working with ensembles lets us do a better job of taking natural climate variability into account, helps eliminate the effects of modelling uncertainty, and means that our conclusions are not biased by the weaknesses or strengths of any one model on its own.
Comparing the results of many different models makes one thing very clear: all climate models clearly indicate that temperatures will continue to rise as greenhouse gas emissions accumulate in the atmosphere.
- U.S. Global Change Research Program. "Chapter 4: Climate Models, Scenarios, and Projections"
- IPCC Fifth Assessment Report (2013)Working Group 1: The Physical Science Basis
- NASA Science Briefs : The Physics of Climate Modeling
- Skeptical Science. FAQ on climate models: Part I and Part II.
- Steve Easterbrook. Timeline of Climate Modeling.
- Intergovernmental Panel on Climate Change. “What is a GCM?”
- “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.
- NOAA. “Breakthroughs: The First Climate Model.”
- Carbon Brief. “How do Climate Models Work?”