Terrain AI
Figuring out ’where’ and ‘how’ of tackling greenhouse gas emissions
‘Think global, act local’ is one of the enduring slogans of the environmental movement. It also raises important questions - where exactly should we target measures to tackle the climate emergency? And how do we measure those emissions to ensure the interventions are working?
SFI research is looking at how to capture and use information smartly from different land types in Ireland, in a bid to better measure and ultimately manage carbon management and greenhouse gas emissions.
The need is urgent, according to Professor Rowan Fealy, because even though we know we need to reduce global greenhouse gas emissions, it is not happening in practice.
"There is a significant gap between the ambition to reduce greenhouse gas emissions and the realisation", he says. “We think that could be improved if we have better ways of measuring and reporting emissions from different land types, rather than measuring and reporting the overall emissions from sectors, such as agriculture or transport, which is what we do now.”
Professor Fealy is a co-leader of the Terrain AI project, which is focusing on more than 20 benchmark sites around Ireland, including agricultural grassland, forests, peatlands, wetlands and urban sites.
Jointly funded by Microsoft Ireland and Science Foundation Ireland, Terrain AI is a collaboration with Teagasc, Trinity College Dublin, University College Dublin, Dublin City University, and University of Limerick, and the project is particularly interested in how human activities at different land types impact emissions.
“You need to understand the geography of where the emissions are coming from, and the potential sequestration of carbon,” says Professor Fealy, who is based at Maynooth University’s Geography Department. “Then you understand where to intervene to achieve the most efficient and effective reductions. It makes difficult decisions simpler.”
Using technology such as drones, air quality sensors and instruments to measure biological activity, the research is gathering measurements from soil to sky at the benchmark sites around Ireland. But it’s not just about the technology, a major aspect of the project is to bring together the people and organisations involved in land use - such as Teagasc, EPA, National Parks and Wildlife Services and Bord na Mona - and in measurement, analysis and policy development, explains Professor Tim McCarthy, who co-leads Terrain AI with Professor Fealy.
"We are bringing those communities together, and that is really important," says Professor McCarthy, who a Professor at Maynooth University’s Department of Computer Science.
"Up to now, you have had an individual focused on, say, a sensor taking one type of measurement from a grass field, and that is all they are interested in. But with these communities working together, we can bring things to a higher level, we have comparability, the readings are standardised and we can do something much more with the cumulative dataset than we could with the information from individual sensors. It makes the information much stronger.”
The benchmark sites were chosen carefully, to develop a better understanding of a wide array of land types or terrains. But we can’t measure every field, city and forest in Ireland, so Terrain AI is developing computer models to make sense of the data in a national and international context, and offer support to policymakers tasked with regulating and protecting the environment.
“Once we have got a really good comprehension of the linkages between human activity and the dynamic, environmental response, we can take that information and apply it to a model context,” says Dr Fealy. “It is like draping the circus tent over the posts, where the posts are the benchmark sites. These draped models allow us to get an insight for locations where we are not directly monitoring, but we want to know the likely impact of human activity and its impact on whether carbon is being taken up from the atmosphere or released to the atmosphere.”
In 2021, Terrain AI won the AI Award for Best Application of AI to Achieve Social Good at the AI Awards, which seek to recognise how AI can be ethically used to help solve some of the biggest business, academic, environmental, and health challenges facing our society today.
Professors Fealy and McCarthy believe that Terrain AI painting the bigger picture of land use and emissions will help to ensure that interventions looking to reduce emissions are applied where they will have more effect, and that we will be able to measure and report the changes in emissions as a result of those interventions.
“Climate change is a car crash is slow motion,” says Professor McCarthy. “To tackle it, we need more effective and credible interventions, and to do that we have to have more scientific accuracy, more robust measurements and models that give us the answers that best reflect what is going on.”