Speaker
Grey Nearing, Research Scientist on the Google Flood Forecasting team, Google, Zürich, Switzerland

The seminar is open to the public.
To join online, please contact seminars@cluttereawag.ch for access details.

Abstract
Over the past eight years, Google Research developed a state-of-the-art global flood forecast system called FloodHub. This started as a research project but has become part of a core Google service that provides multi-hazard disaster warnings to a significant fraction of the global population. This talk will trace this flood forecasting project from early research through operationalization to open questions about the role of AI and of the private sector to assist within the global Early Warnings for All (EW4A) ecosystem.

Perhaps for the first time, state-of-the-art forecast systems and global-scale information dissemination services exist within the same organization. This presents both an opportunity to help address the global EW4A challenge, and a challenge in that these tools must be leveraged to support, rather than supplant, national authoritative agencies. The open question is about what combination of technology development and capacity development is needed to fill the gap. AI/ML forecast systems offer the promise of lowering barriers to access – for example, lower computational cost opens potential for local data assimilation without heavy HPC resources – but the technologies and global partnerships necessary to realize this promise are still in the process of being developed.