The climate crisis is arguably the single biggest threat to humanity. But science provides a powerful weapon, and research communities around the world are racing to find ways to predict, and safeguard, the future of our planet. This Earth Day, we’re highlighting five projects at The Alan Turing Institute that are contributing to this effort, involving researchers across a range of disciplines and collaborations. Environmental challenges are global, and the sheer volume of data that needs to be incorporated into decision-making, and into computational models of our climate, means that data science and AI will be a crucial part of the response.
Using AI to predict Arctic sea ice loss
Algorithms are essential for extracting useful intelligence and insights from large datasets. In this area, the Turing is working with the British Antarctic Survey AI Lab to use machine learning techniques to make seasonal forecasts of Arctic sea ice cover.
Researchers hope that being able to provide a reliable forecast of Arctic sea ice cover will help improve regional weather model predictions, and could provide an early warning system to mitigate the risk to wildlife and indigenous communities who depend on the ice. To meet this aim, our researchers have teamed up with the World Wide Fund for Nature (WWF) to develop forecasts of the sea ice freeze-up and break-up days in Hudson Bay in north-east Canada to support polar bear conservation efforts there.
Supporting Pangeo – the community-driven platform for big data geoscience
In order for environmental scientists to store, analyse and process data, they need a flexible, reliable and scalable digital infrastructure. Since 2020, the Turing has been working alongside the Met Office and Microsoft to support Pangeo – a project that is helping the Earth science community to analyse data in the cloud. The platform allows for interactive and scalable computing and can be used by ocean, atmosphere, land and climate scientists.
The Turing's Research Engineering (REG) team is working closely with researchers in the Met Office’s Informatics Lab and Microsoft’s AI for Earth programme to improve Pangeo’s ability to handle petabyte-scale datasets on the Microsoft Azure platform. So far, the researchers have developed and deployed a platform that allows easy access to the Pangeo suite of tools on Microsoft Azure’s cloud infrastructure.
Mapping the UK’s solar panels
There are no comprehensive records of the location of the UK’s solar panels, which means that precisely how much solar energy is being pumped into the UK’s electricity grid at any time is not well known, even by the National Grid. In the absence of accurate measurements and predictions of solar energy input, fossil fuels are burnt unnecessarily to keep generators running so they can take the strain when the network is underpowered. Our project, in collaboration with Open Climate Fix and OpenStreetMap, aims to resolve that by using a combination of AI (machine vision), open data and short-term forecasting.
Through a crowdsourcing exercise, volunteers tagged the locations of solar panels on OpenStreetMap, mapping one-quarter of all the solar panels in the UK. Using this data, the researchers created an open dataset of solar panel locations that will help provide a short-term forecast of how much solar power is being fed into the National Grid, in turn helping to cut carbon emissions. The team is also working on machine learning methods to detect solar panels from satellite images, which would automate the process and fill in some of the map's gaps, further improving solar power forecasting.
Understanding the risks of climate change to human and national security
Recent developments in data science, statistical modelling and machine learning can help researchers to better model, predict and forecast the impacts of climate change on populations, livelihoods and natural resources, and so gain a better understanding of how conflict, political stability and migration might be affected. A project being undertaken by the Turing’s defence and security programme is exploring the implications and risks of climate change to human and national security.
Researchers have so far analysed the interactions between the societal, economic and environmental factors that can affect instability and conflict resulting from climate change. This gives a better understanding of the modelling and data requirements needed to effectively build a data model(s) that can be used in various scenarios to show future potential climate security tipping points. A recent report details this initial research.
Developing a digital twin for the world’s first underground farm
Beneath the busy streets of Clapham, London, lies an underground farm that is growing subterranean salad leaves. Based in a former World War II air raid shelter, the aim of Growing Underground is to grow crops all year round in optimum conditions, while saving on valuable land resources.
To support the farm, researchers at the Turing and the University of Cambridge have developed a ‘digital twin’ – a computer model of the farm that uses sensors to help the growers optimise crop growth and quality while minimising energy usage. For instance, the growers can use the model to predict when they might need to alter the light or heat levels, and then find out how effective those adjustments were.
Climate action at the Turing
For more information on the Turing’s response to climate change, head to our dedicated climate action page. We also have a new Environment and Sustainability Interest Group, which aims to build a cross-disciplinary community to explore the role of data science and AI in addressing key environmental challenges.