The symptoms of climate change are vast, complex and costly. Extreme weather events cause damage to property, businesses and displaces humans in regions around the globe. While flooding has always been a natural occurrence under stable climate conditions, according to a 2021 report from the World Meteorological Organization, flooding and water related disasters have dominated the lists of both cost to human life and economic losses over the past 50 years.
Advertiser Content From
This advertising content was produced in collaboration between Vox Creative and our sponsor, without involvement from Vox Media editorial staff.
Finding Floods with AI
How IBM is training AI models shaping flood mitigation efforts
As the global population is expected to balloon to 10 billion people by 2050, communities around the globe will need to be better prepared to protect their people, so they’ll need more accurate flood detection and modeling. To do exactly that, IBM Research is developing a Global Climate Network, which uses AI-enabled accelerated discovery, to predict floods before they happen.
“Science requires collaboration. There is no progress without science, so collaboration is the key piece.”
What is a Climate Network? This network is a system of computing systems, where highly distributed climate information and services from across the globe become discoverable or accessible. Participants can share information across the network, and this sharing enables seamless federation and AI modeling to accelerate. That collaborative practice can lead to the development of climate adaptation and mitigation solutions. So what does AI modeling have to do with flood prevention?
In 2021, IBM Research introduced a method that could predict daily precipitation of a region up to six months into the future. “The Climate Network is really doing three things: it is accelerating the discovery of climate information. It accelerates the developments of AI models. And then finally, it accelerates the collaboration” says Chief Scientist for Climate and Sustainability Hendrik Hamann. Critical to this effort is the acceleration of collaboration between data, people and technology.
“Data is the key resource for AI. Climate data is mostly geospatial. It has a space and a time component. It is massive, massive amounts of data” says Hamann. The vastness and complexity of data within the Global Climate Network can mean many things, from satellite images, soil moisture readings or altitude of terrain. Typically, data funneled through an AI model is processed in one place or through one process, hampering efficiency. The use of a Global Climate Network, however, allows participants to discover not just their own climate information, but also information across the network. This method helps accelerate the development of the AI models used, saving time and resources. While each piece of data helps to inform the AI model used by the network, data alone isn’t enough.
Technology is one of three pillars of the Global Climate Network. Key technologies include the use of hybrid cloud, cloud computing, foundational AI and federated learning, which happens at the edge, like mobile phones, laptops or private servers. And as important as data and the various technologies are in building an intelligence ecosystem, like IBM’s Global Climate Network, Hamann reiterates how that’s synthesized and made actionable. “Science requires collaboration. There is no progress without science, so collaboration is the key piece.” Equally important are the people behind each enterprise within the network.
People have an important role to play within the network, like locaI and national governments, scientists, physicists, engineers, business enterprises and academia. By leveraging data, technology and different fields of research, these entities can contextualize data into actionable strategies. Valuable insights which may inform public policy or changes to a communities’ infrastructure, all while maintaining the sovereignty of their information. That is the power of an intelligence ecosystem like the Global Climate Network.
If extreme weather events become more frequent or more hazardous, humans across the globe will need powerful and unique solutions, like actioning IBM’s Global Climate Network. Each flood, while damaging and costly, may provide a lesson for other communities around the globe. The value of this kind of predictive power, means that by understanding the impacts of flooding locally, humans can mitigate against floods and flood impacts globally. “It’s a distributed network. We have started building a climate network with a few key partners. We have demonstrated discovery of climate information in these different instances and we are eager for others to join and work with us, and collaborate, to find better solutions to climate change.”
To learn more about the Global Climate Network, visit IBM.com/research today.
