Machine Learning to Detect Irrigation

Using satellite, climate and soil data to predict water use for agriculture


Interactive, zoomable maps of worldwide irrigation for agriculture


Measure and protect fresh water, ensure food for people


Satellite data and machine learning models to predict extent

Farms use a lot of water. Two-thirds of it.

Yes, most of the fresh water we draw goes into growing food. Did you know we need 1 to 3 tons of water to grow a kilo of cereal? Speaking of cereal, we have to double what we grew in 2005 to feed everyone in 2050!

Yet we're not managing our water. We don't know where in the world we rely on irrigation.

We should start by finding out how much of our cropland is irrigated. We should also know how it's changing over time. We can then detect unsustainable farming, prevent over-drawing of water, and ensure we can grow enough food.

Technology can help. Satellite data and machine learning can provide us insights.

We've used pictures taken from satellites and ground-level climate and soil data to detect irrigated croplands. We've developed a machine learning model for this.


Explore our maps!

The Maps

How we built our product

About Us

See the faces behind it

About Us