STAC Overflow: Map Floodwater from Radar Imagery Hosted By Microsoft AI for Earth


About the project

This challenge uses Sentinel-1 global SAR imagery to detect the presence of floodwater on the surface of the Earth. Traditional overheard electro-optical (EO) imagery relies on passive, visible or infrared sensors to detect features in a scene by measuring reflected solar light. SAR uses an active microwave sensor to produce its own pulses and then records the portion of the outgoing signal reflected back to it.

Electromagnetic Spectrum

SAR operates in the microwave band of the electromagnetic spectrum.

SAR has several unique advantages for flood detection. It can capture images at night as well as through clouds, precipitation, smoke, smog, and vegetation. The Sentinel-1 mission performs C-band imaging, which operates at a frequency of 4-8 GHz and is particularly useful for change detection in areas with low vegetation. Sentinel-1 transmits and receives energy that is both horizontally and vertically polarized, providing valuable information about the structure of imaged objects.

The interpretation of SAR images, however, is far from straightforward. Geometric distortion is an inherent characteristic of its unintuitive, side-looking geometry. As a general rule of thumb, regions of calm water tend to appear dark as the radar reflects away from the spacecraft, while rough wet surfaces tend to reflect radar in all directions and often appear brighter. Wind-roughened water can backscatter particularly brightly when its waves are close in size to the incident radar's wavelength, while a hill or mountain may appear bright on one side and dim on the other.

SAR Polarization

Readings from different polarizations carry information about structure based on three types of scattering.
Image Credit: NASA SAR Handbook.

With the increasing prevalence of extreme weather events due to climate change, we believe that remote sensing and specifically C-band SAR has the potential to strengthen the speed and accuracy of flood mapping, inform disaster risk management, and ultimately, save lives.

About the project team

Microsoft AI for Earth empowers organizations and individuals working to solve global environmental challenges. The program drives innovation in environmental sustainability by developing technology to accelerate environmental science and by providing grants to organizations addressing problems in biodiversity, climate change, water, and agriculture. Since 2017, AI for Earth has awarded grants to research teams across more than 60 different countries.

Microsoft’s new Planetary Computer hosts the Sentinel-1 imagery, NASADEM digital elevation data, and JRC permanent water data used in this competition. The Planetary Computer puts global-scale environmental monitoring capabilities in the hands of scientists, developers, and policy makers. It combines a multi-petabyte catalog of analysis-ready environmental data with intuitive APIs, a flexible development environment, and applications to put actionable information in the hands of conservation stakeholders.

Note on available resources: The Planetary Computer Hub provides a convenient way to compute on data from the Planetary Computer. If you're interested in using the Planetary Computer Hub to develop your model, you can request access here. Make sure to include "DrivenData" in your area of study on the account request form. To be approved more quickly, you can also email after submitting the form and say that you are competing in this competition.

Cloud to Street is an organization committed to using flood mapping and monitoring to protect the world’s most climate-vulnerable communities from natural disasters. By harnessing a global network of optical and radar satellites, Cloud to Street tracks worldwide floods in near real-time and uses artificial intelligence to remotely analyze local flood exposure. Its innovative platform has been used by governments, insurers, and disaster response agencies across 15 countries to understand flood risk, insure populations, and prepare for flood events.

You can learn more here: