DECEMBER 31, 2022

BY FRANCISCO CEBALLOS, BERBER KRAMER AND BENJAMIN KIVUVA

IFPRI’s picture-based insurance (PBI) initiative, recognized as a CGIAR@50 Innovation and active since 2016, relies on participating farmers to upload smartphone pictures of their fields at intervals throughout the growing season. In the event of bad weather, pests, disease, or other problems that harm the crops, the photos are used to assess the damage and trigger insurance payouts.

More recently, a new service has been added: The photos can serve as the basis for low-cost, interactive picture-based advisories (PBA)—expert advice tailored to individual farmers—enabling unprecedented levels of targeting and timeliness in remote agricultural extension activities.

Coupling advisories with insurance can reduce barriers to technology adoption and lower overall farming risk by encouraging improved agricultural practices. Beginning in 2022, CGIAR’s Standing Panel on Impact Assessment (SPIA) has funded an impact evaluation of PBA in India and Kenya to evaluate its effects on a range of farming and welfare outcomes and shed light on the system’s scalability.

How did we go about setting up a well-functioning system of remote advisories, personalized to each farmer’s situation, with little automation or heavy human involvement? This post describes our journey for designing this system, a description of its day-to-day functioning in Kenya, and the lessons learned throughout this process.

Results in 2022

At the end of the Long Rains 2022 season, we interviewed the two participating KALRO experts, who reported that farmers’ uploaded images were consistently good enough to enable them to identify specific problems in the crops and provide pertinent advisories. For example, while nutrient deficiency, excess water, and certain diseases can all lead to the yellowing of crop leaves, experts were able distinguish these problems from the photos by looking at other elements—such as the condition of the surrounding soil, close-ups showing telltale signs of a given disease, or noting that nitrogen deficiency results in the gradual yellowing of leaves beginning with those closer to the ground.

Other conditions such as drought could be identified by observing the dryness of the soil and an overall dry, dusty environment in the picture. This highlights the value added of looking at ground pictures of the crop vs. other competing approaches, such as satellite-derived vegetation indices that rely on leaf color alone, with little additional information on its underlying causes.

Experts sent close to 5,000 advisory messages throughout the season covering a range of topics (Figure 1), with the majority of advice relating to the presence of weeds in the fields, the presence of pests or diseases, and a lack of water or the need to irrigate the crop.

See more https://www.ifpri.org/blog/developing-and-implementing-picture-based-advisories-pba-farmers-kenya