Climate models, also known as general circulation models or GCMs, use computers to simulate Earth’s climate. Scientists use climate models to predict weather, but unlike weather forecasts, the timespan is over decades or centuries rather than hours or days. The models use mathematical equations that are based on well-documented physical processes to identify how energy and matter interact within the ocean, atmosphere, and land. Then, variables can be tweaked to simulate responses to global climate changes such as increasing greenhouse gas emissions, increasing pollution, and decreasing forestation.
Modeling a global climate
Climate models can focus on a particular region or the entire planet. According to NASA, the global climate model from the National Center for Atmospheric Research (NCAR) can fill over 18,000 pages of printed text. Hundreds of scientists add hundreds of observations and millions of lines of code to build a model, requiring a supercomputer the size of a tennis court to process the information.
Early climate models simulated only one part of Earth’s climate system, either the atmosphere or the ocean, but not both. These models were still layered with three-dimensional measurements like the height of the atmosphere or depth of the ocean. However, more advanced models started to incorporate many sub-models, linking or coupling multiple systems together.
Over time, models have increased in complexity, adding details such as vegetation growth, algae in the ocean, and dust on snow and sea ice. NCAR scientist David Lawrence said, “The best models can take this information and very accurately reproduce Earth's climate and weather, matching observations reasonably with major precipitation zones, circulation patterns, and modes of variability such as the El Niño Southern Oscillation (ENSO).”
The most recent GCMs incorporate the transfer of chemicals, like the carbon and nitrogen cycles, between living organisms and their environments. If one factor in the climate is tweaked, how does the simulated climate system respond? By fiddling with the variables, scientists can speculate on future changes to rainforests or the Arctic, depending on current environmental conditions.
Researchers know that at least one climate factor is changing—the amount of greenhouse gases in the atmosphere. According to the World Meteorological Organization (WMO), carbon dioxide reached 413.2 parts per million in 2020, equivalent to 149 percent of the pre-industrial level.
Basic physics tells us that increasing greenhouse gases will increase temperatures on average, but where, when, and how exactly the climate will change depends on many other variables, such as atmospheric conditions, ocean circulation, ice loss, and plant growth. NSIDC scientist Julienne Stroeve studies sea ice and climate. She said, “Climate models can tell us about how these components of our planet interact and influence each other, and to understand what happens when we perturb one of these components.”
To test their models and their current understanding of how climate works, climate scientists also run their models through past and current climate scenarios, to determine how well they can replicate the observed climate variability and trends. Stroeve said, “Because the real world provides just one realization of both natural variability and forced climate change, we cannot separate out natural climate change from, say, greenhouse gases in observational data.” Models, she said, allow researchers to run a test many times, and answer questions such as how much added greenhouse gases are changing climate, compared to natural variability.
How a climate model works
A model needs to start somewhere in time. Climate models often rely on a technique called a “spin up,” where once the model starts, the model parameters evolve. Initially, they may be way off, but if the model physics are good, eventually things will evolve to a physically consistent state. For example, you could start a sea ice model with no ice at all in the Arctic Ocean. By setting up a cold atmosphere with below-freezing ocean temperatures, ice will start growing. Initially, it will be very thin, but it will thicken to resemble a physically realistic state.
Another way to start the model is with observations. This is essential for operational models because researchers cannot wait for the ice to reach equilibrium if they want to know how things will look in a day, a week, or a month, or even a year. Passive microwave and other sea ice data can start the model with the observed sea ice conditions. Then, the model can run without the need for further observations.
Once a climate model begins, however, researchers often use a technique called assimilation that combines models and observations throughout the model run. When the model begins in its initial state and then evolves forward one day, observational data can adjust the model fields and “nudge” them to the observations.
One way is to simply replace the model estimates with observations. However, this can mess up the model physics and pull it out of equilibrium. Also, it assumes that the observations are perfect and that is never the case. If there is a disagreement between the observations and the model estimates, how do you know which is right?
What happens more in practice is that there is a give and take between observations and the model. As the model goes forward, researchers compare the observations and the observational errors. With that information, they know how far off the model is and how much it needs to be adjusted. Sometimes a weighted average between the observation and the original model estimate can provide an improved model estimate.
Of course, researchers cannot assimilate future observations into the model, but past data can inform a model into a more accurate position. And as the model moves forward in time, researchers can continually add data and adjust the model.
An example of climate model in action
Adding in detailed factors like plant growth makes climate models more complicated, but they also help scientists better understand what influences that ecosystem and how that ecosystem influences global climate.
On the Arctic tundra, the number of shrubs is increasing as the air and land get warmer. Shrubs provide more shading than grassy tundra, leading some scientists to argue that shrub expansion could slow down permafrost thaw because shading the ground below cools it. But the shrubs themselves have additional effects on the climate.
NCAR scientist David Lawrence has been researching how changing plant life in the Arctic could affect climate. Lawrence said, “Shrubs are darker than the surrounding tundra, so they tend to absorb more solar radiation, thereby warming the nearby air.” In 2011, Lawrence published a paper that used a climate model to study how an expected increase in shrubs will alter permafrost vulnerability to climate change. He found that more shrubs are likely to increase permafrost vulnerability to climate change, in contrast to previous studies that suggested that shrubs would have the opposite effect.
A study from 2020 adds more complexity to the shrub expansion under a warming climate. Plants in many parts of the Arctic are withering. As the Arctic warms two to three times the rate of the rest of the planet, summers are getting too hot and too dry for growing shrubs. Meanwhile, Arctic sea ice is also shrinking in all seasons, with summers being most pronounced. Could there be a connection between sea ice loss and the browning of shrubs?
This study compared Arctic sea ice decline with shrub growth and resilience, analyzing tiny growth rings in the wood trunks of willows, birch, and other plants from around the Arctic. Although 57 percent of shrubs have boomed as sea ice has retreated since the 1990s, 39 percent have been stunted. The disparity directly correlates to the availability of water.
The Arctic landscape is not uniform. In most places, shrinking sea ice triggers higher air temperatures, increasing precipitation. In other areas, like parts of western Greenland, where shrubs grow in drier or poorly drained soils, the higher temperatures push the vegetation into drought conditions. Intricate climate models can reveal connections and direct consequences of one changing factor on another.
Climate models have improved since Norman Phillips, an American meteorologist, first developed a true general circulation model (GCM) in 1956. With visionary pioneers like Syukuro “Suki” Manabe, who won the Nobel Prize in physics in 2021 for his 1960s research into climate modeling, the mechanisms behind climate science have become more accessible and tackle to everyone. While global climate models do a good job of simulating Earth’s climate, they are not perfect and still have room for improvement. Perhaps this is a never-ending quest for perfection, but the closer observations match climate models, the more accurate their future predictions will be.
References
Buchwal, A., P. F. Sullivan, M. Macias-Fauria, E. Post, et al. 2020. Divergence of Arctic shrub growth associated with sea ice decline. Proceedings of the National Academy of Sciences, 117 (52). doi:10.1073/pnas.2013311117
Docquier, D., and T. Koenigk. 2021. Observation-based selection of climate models projects Arctic ice-free summers around 2035. Commun Earth Environ 2, 144. doi:10.1038/s43247-021-00214-7
Lawrence, D. M. and S. Swenson. 2011. Permafrost response to increasing Arctic shrub abundance depends on the relative influence of shrubs on local soil cooling versus large-scale climate warming. Environmental Research Letters 6 045504. doi:10.1088/1748-9326/6/4/045504
Stroeve, J. C., A. P. Barrett. Assessment of Arctic Sea Ice in the CMIP5 Climate Models. American Geophysical Union 2011 Fall Meeting. C21D-04 Community Earth System Model Project Web site: http://www.cesm.ucar.edu/