The next chapter in flood resilience: Open sourcing Google’s hydrology framework

Google has open-sourced its flood forecasting framework, which replicates operational FloodHub model training settings and reflects methodology described in a 2024 Nature paper for global ungauged flood prediction. The framework combines physical hydrology with machine learning to provide an open-source tool for enhancing flood forecasting, particularly in data-scarce regions. This release enables broader adoption of Google's hydrology and flood resilience technology across research and operational communities.

Source: Google Research | 2026-06-03