High-resolution images taken by compact satellites could be used to estimate agricultural productivity as well as test intervention strategies in regions of the world where data are currently extremely scarce. Image of maize farm plots in Western Kenya were taken by Terra Bella satellites (left) and an agricultural yield map (right) generated from the same image using machine learning algorithms. Courtesy of David Lobell. Stanford researchers have developed this new technique to estimate crop yields from space, with hopes of helping improve agricultural productivity, ultimately reducing hunger and improving quality of life all over the world. To improve productivity, though, it first has to be measured. Earth-observing satellites have been around for over three decades, but most of the imagery they capture has not been high-enough resolution to visualize the very small agricultural fields typical in developing countries. Recently, however, satellites have shrunk in both size and cost while improving in resolution. "You can get lots of them up there, all capturing very small parts of the land surface at very high resolution," said David Lobell, an associate professor in the department of Earth System Science at Stanford's School of Earth, Energy & Environmental Sciences. "Any one satellite doesn't give you very much information, but the constellation of them actually means that you're covering most of the world at very high resolution and at very low cost. That's something we never really had even a few years ago." In the Stanford study, researchers set out to test whether the images from this new wave of satellites are good enough to reliably estimate crop yields. They focused on an area in Western Kenya where there are many smallholder farmers who grow maize, or corn, on small, half-acre or one-acre lots. The scientists compared two different methods for estimating agricultural productivity yields using satellite imagery. The first approach involved “ground truthing” where they conducted ground surveys by talking to farmers and gathering information about individual farms to check the accuracy of yield estimates calculated using the satellite data. The team also tested an alternative "uncalibrated" approach that did not depend on ground survey data to make predictions. Instead, it used a computer model of how crops grow, along with information on local weather conditions, to help interpret the satellite imagery and predict yields. "Just combining the imagery with computer-based crop models allows us to make surprisingly accurate predictions, just based on the imagery alone, of actual productivity on the field," said Marshall Burke, an assistant professor in the department of Earth System Science. Burke and Lobell have plans to scale up their project and test their approach across more of Africa.