Advances in modern climate science study the dynamics of weather patterns and provide insights into how global warming will affect the world. Such forecasts have been made since the 1960s by using climate models. Let’s find out how they work and whether they can always be trusted.
What are climate models?
Climate models are computer simulations used to analyse past and present changes in the planet’s climate and to make predictions. They consist of three macroblocks
- A thermodynamic system that exchanges energy and matter both within itself and with outer space. Its state is governed by physical and chemical laws expressed in the form of differential equations.
- Climatically relevant chemical processes. For example, the carbon cycle in the ocean, the ozone-oxygen cycle in the stratosphere and so on.
- Climate-forming processes — the activities of all living organisms on the planet and their impact on weather patterns.
The first climate models were attempted back in the 19th century. But there were no computers back then. However, at least a mathematical apparatus appeared in the early 19th century: the Fourier transform, an algorithm that can represent an arbitrarily complex function as a sum of simple functions (ideally, an infinite number of simple functions) was created. This algorithm was immediately useful for predicting tides, and it was the first climate model. Scientists created an analogue computer, that is, a mechanical device where the movement of gears performed the same role as the processor in modern computers. Such devices were incredibly sophisticated, and by the end of the 19th century they were also being tried to predict changes in the atmosphere. However it was not until the 1950s, with the invention of the first electronic computers, that real models came into use.
Climate models make it possible to study the sensitivity of the climate to various factors by means of experiments that cannot be performed in reality. For example, what would happen if the global population doubled? What if all the factories in the world became carbon neutral? The results help to create strategies that can slow the rate of global warming and avert climate catastrophe.
Global forecasting requires impressive computing power, so modern climate models run on massive supercomputers. For example, the Met Office Hadley Centre, one of the UK’s leading centres for climate change science, has three Cray XC40 supercomputers capable of performing around 14,000 trillion calculations per second.
A model of aerosol motion created at NASA on a supercomputer
Photo: NASA
Types of climate models
The precursors of climate models are energy balance models (EBM). They do not predict the climate, but the change in the temperature of the planet’s surface. EBMs take into account the difference between the solar energy entering the Earth’s lower atmosphere and the heat released back into space. The simplest EBMs can be run in a spreadsheet and contain just a few lines of computer code.
The most sophisticated climate models available are called the General Circulation Model (GCM), or Global Climate Model. They consist of about 500,000 lines of code and are mostly written in Fortran, a scientific programming language. GCMs include several separate ecosystem models: atmosphere, ocean, land and sea ice, and others. Each of these models is developed separately and then combined with the others, taking into account their interchange of heat, moisture and electromagnetic pulses.
Nobel Prize for Climate Modelling
Only 30 years ago, scientists regarded climate models with great scepticism. Now there is no doubt about their accuracy, as predictions made between 1970 and 2007 have already come true. A team of scientists at the University of California, Berkeley, has reviewed 17 global temperature models from this period and found that 14 of them are as accurate as possible. This is impressive, because in the 1970s scientists had fairly little evidence of rising global temperatures.
In 2021, climate forecasters Shukuro Manabe and Klaus Hasselmann won the Nobel Prize for their contribution to protecting the future of the planet. Princeton University senior climatologist Shukuro Manabe was the first person to investigate the interaction between the vertical movement of air masses and the radiative balance of the Earth’s surface. His work laid the foundation for modern climate systems. For his part, Klaus Hasselmann, a professor at the Max Planck Institute for Meteorology in Hamburg, proved that we can trust climate models, despite the variability of the weather. He has also found a link between global warming and greenhouse gas emissions.
Image by: Johan Jarnestad / The Royal Swedish Academy of Sciences“
Climate models are changing — the threat of global warming remains
Climate scientist and professor at Texas Tech University, Catherine Hayhoe, noted that climate models have become more sophisticated and accurate in recent decades. But throughout the modernisation process, they have continued to point to the threat of global warming, once again confirming its reality. Moreover, some catastrophic changes may come sooner than previously thought.
The Arctic, for example, has been found to be heating up faster than anywhere else on the planet. Over the past decade, the average temperature in the region has risen by 1 oC. If emissions of greenhouse gases are not reduced, the Arctic will be 4 oC warmer all year round by the middle of the 21st century. At that time, the global temperature will have risen by 2 oC, a level often referred to as the “point of no return”. Given that a climate catastrophe will occur in the Arctic much sooner than elsewhere, scientists are paying maximum attention to observing the region.
Artificial intelligence predicts the rate at which Arctic ice is melting
Due to global warming, Arctic sea ice has begun to melt at an alarming rate. Over the past 40 years, its area has shrunk by half, an area lost that is 25 times the size of Britain.
Arctic ice is a critical component of the global ecosystem. Its loss could lead to a decrease in the polar bear population, as well as an increase in the intensity of poisonous algal blooms. If the concentration of toxins they release becomes too high, the fish on which Arctic indigenous peoples depend for their food security will die. Due to the movement of air masses, Arctic sea ice also affects the planet’s climate. For example, their intense melting can make European summers wetter and Northern Hemisphere winters extremely cold.
To predict changes in sea ice, an international team of scientists led by the British Antarctic Survey and the Alan Turing Institute has created an artificial intelligence-based system called IceNet. It allows you to find out how the state of the Arctic ice cover will change in the coming season, and to take measures in time to mitigate the effects of these changes.
One of the authors of the project, Tom Andersson, noted that IceNet works thousands of times faster than traditional forecasting methods, and the inaccuracy of the system is no more than 5%. It combines data from meteorological satellites and information obtained from climate models. The next goal is to create an improved version of the system that works in real time. This will make it possible to learn as early as possible about the risks associated with the rapid melting of the Arctic ice.
Author: Vera Zhikhareva
Cover photo: Felipe Dana / AP
Comments