Knowing where water is and where floods might happen is really important for managing water, preparing for floods, and protecting communities. But many parts of the world don’t have access to this kind of up-to-date-information. Thanks to recent advances in cloud-computing and machine-learning, it’s now possible to use satellite-data to create detailed-maps of surface-water and make short-term flood forecasts.

In this project, we use satellite-images from Sentinel-1, which can “see” through clouds and take pictures even during storms. These images help us make water-maps every six-days, with details as small as 10-meters across. We built a system that shows where floods are happening, how rivers are changing, and what areas might flood soon. The system works with SWORD-database and lets people explore maps and forecasts through an easy-to-use online-app.

In addition to flood forecasting, we also use the same data and methods to help farmers. We created tools that show the health of crops, estimate crop-yields, and even forecast future harvests. Right now, our system focuses on rice fields/(paddy) and will soon include other crops in the U.S. It’s built so landowners can check their farm's condition and get yield forecasts right on their phones—helping them make better decisions/improve profits.