Sustainability & crisis resilience

Our state-of-the-art hydrologic models can forecast riverine floods up to 7 days in advance. Our flash floods model, trained on our new Groundsource dataset, helps to predict flash floods in urban areas up to 48 hours in advance.

We’re leveraging AI and satellite imagery to detect wildfire boundaries. We’re also working to launch more satellites for FireSat, a purpose-built satellite constellation to enable the detection of a garage-sized wildfire anywhere on Earth within 20 minutes.

Our weather models include MetNet for hyperprecise nowcasts and WeatherNext 2 for 15 day forecasts and cyclone predictions, in partnership with Google DeepMind. NeuralGCM is for longer range weather and climate predictions. These power a range of applications for users, developers and weather agencies.

Our large-scale open dataset contains 1.8 billion building detections across Africa, South Asia, South-East Asia, Latin America and the Caribbean. Organizations are using the dataset for post-disaster humanitarian response, urban planning, population mapping and environmental science. Additionally, our new 2.5D Temporal dataset quantifies changes over time in the urban environment.

Green Light uses AI and Google Maps driving trends to build intelligent recommendations for cities to optimize traffic flow. It operates in cities around the world, reducing vehicle emissions at intersections from Boston to Hamburg.

Using satellite imagery and computer vision, we can predict when and where contrails — the heat-trapping condensation trails left behind by airplanes — are likely to form. Pilots from American Airlines and Eurocontrol are using these predictions to adjust flight altitudes in real time, helping mitigate aviation’s climate impact.