The world is a complex place. The natural environment consists of many different components that can interact with each other, sometimes predictably but at other times with unpredictable outcomes. Our climate and weather are examples of such complex systems. Dr Ken Amor explains how new developments in Artificial Intelligence can help us understand these systems, and to find new ways of tackling climate change.
The accuracy of the weather forecast is often limited to one or two days in the future, beyond which details become sketchy. Our climate does have some predictable patterns and we know that summers are warmer than winters for example, but trying to predict an accurate forecast for the weather in six months’ time is impossible at present. British winters can be mild, wet and windy, or cold and frosty, depending upon whether low or high pressure dominates the season.
This vagueness is because there are many different components and mechanisms that ‘react’ with each other in different ways depending upon the state of other variables. We do not fully understand how these systems work and this can lead to uncertainty in our predictions. For example, our weather in Britain is controlled by the jet stream, a current of air that circles the globe in the upper atmosphere; the Gulf Stream, a swift ocean current in the Atlantic that brings warm water to Western Europe; the distribution of atmospheric high- and low-pressure systems; wind direction; sea and air temperature; humidity; cloud cover; and many other factors. Furthermore, the condition of each of these components is determined by what happens in other parts of the world and their position and intensity are constantly changing with time.
Trying to predict the outcomes of all these interactions requires super-computers. This is an area where AI can and does help because it can process vast amounts of data from satellite and ground-based observations. It is also good at recognizing patterns because its way of ‘thinking’ attempts to mimic human thought and we are very good at distinguishing patterns in the information we receive. Although we can’t make accurate weather forecasts six months in the future at the present time, climate modelling using AI is being used to make long-range forecasts such as whether next winter will be mild and wet or cold and dry.
These same climate models can help predict the outcome and consequence of increasing air and sea temperatures resulting from rising levels of carbon dioxide in the atmosphere. However, if we can’t accurately predict the weather even a few weeks in advance, how can we be certain about predictions for human-induced climate change decades into the future? Different computer models produce different climate scenarios, but they all agree that temperature will continue to rise. AI can be used to fine-tune our projections of global climate change, by analysing and comparing the very large datasets of historical and modern data. This will give us a better understanding of the ways different parts of the world respond to climate change, for example the probability of extreme weather events (floods, drought, severe storms, and so on), changes in rainfall amounts, and the different rates at which regions are warming (the polar regions are predicted to see the fastest warming).
Another way that AI is being used to tackle climate change is predicting extreme weather-related hazards such as flooding and landslides. A project is underway whereby an intelligent machine is learning to recognize landslides from Google Earth images. The next step is to associate these events with rainfall amounts and other local events such as changes in land use, deforestation, and so on. The goal is to be able to predict what happens next, identifying the trigger and to prevent these geohazards from endangering human life and property. AI can also be used for improved local weather forecasting and this will give better predictions of power output from renewable sources such as wind and solar. It isn’t just on the global and national scale where AI systems are being used to tackle climate change. AI can improve energy efficiency in buildings and homes by controlling automated systems to adjust heating and lighting depending upon the number of people in a room, the time of day, how much light is being received through the windows and switching on your washing machine to run during off-peak electricity times while taking advantage of a cheaper energy tariff. Human behaviour is an example of a complex system – each of us is an individual with our own habits, which is why intelligent control systems could support the energy usage in our own homes.
Dr Ken Amor is the Access & Outreach STEM Associate at St John’s College, Oxford
If you enjoyed this article…
You might want to explore our courses in:
Geography | St John’s College, Oxford
Computer Science & Joint Schools | St John’s College, Oxford
Orwell Youth Prize – Writing the Climate Crisis
The Orwell Youth Prize is an annual writing prize for students in Years 8–13, encouraging young people to use the writing of George Orwell as a starting point to inspire them to write about their own ideas and experiences. The theme of this year’s prize is “Coming up for Air: Writing the Climate Crisis”.
The Prize invites you to think and write critically and creatively about how the climate crisis affects the people, places, and things around you. How will our impact on the environment affect the places you know? How can we confront this crisis in a way which is fair to everyone? What is missing from the conversation?
And if you had the power, what changes would you make?
More details about the 2022 competition, and how to enter, can be found on the Orwell Youth Prize website!