Session summary
This GeoField 2026 session presents three applied use cases showing how Earth observation can complement field experiments, surveys, and program data in agricultural impact evaluation. The presentations focus on irrigation efficiency, fertilizer management, and climate risk management for smallholder farmers across Bangladesh, India, and Nigeria.
Moderator: Matt Hallas
Alternate Wetting & Drying for Irrigated Rice in Bangladesh
Kunwar Singh and Reeya Shrestha present a study of alternate wetting and drying, a climate-smart irrigation practice that allows rice fields to dry before irrigation is applied again. Building on prior randomized trials, the use case asks whether reductions in water use and changes in irrigation efficiency can be detected from space. Using Sentinel-2 imagery, weather data, and evapotranspiration modeling in Google Earth Engine, the study examines whether treated areas show evidence of more productive water use over time.
Scaling Leaf Color Charts Among Cotton Farmers in India
Tomoko Harigaya and Yuntian Bi present research on leaf color charts, a low-cost tool that helps farmers calibrate nitrogen fertilizer use based on crop leaf color. The study examines whether incentive structures encourage farmers to adopt and use the tool, and uses satellite imagery to enrich baseline measurement and explore treatment effects across the crop health distribution. The presentation also highlights practical constraints, including cloud cover during the monsoon season and the difficulty of linking optical imagery to in-season crop management.
Mutual Crop Insurance in Nigeria
Peter Njiforti and Kendra Walker present a study of mutual weather index crop insurance for smallholder farmers in northern Nigeria. The project tests a Sharia-compliant, takaful-style insurance model designed for farmers facing drought, dry spells, and rainfall variability in millet and sorghum systems. The presentation discusses survey-based evidence on willingness to participate in future insurance and explores how satellite-derived vegetation indices could help estimate crop condition, validate reported damage, and reduce reliance on costly field visits.
Together, the session illustrates the value of combining Earth observation with randomized trials, household surveys, field measurements, and administrative data. The presentations show how satellite data can extend learning across time and space, while also underscoring common challenges such as cloud contamination, field boundary uncertainty, ground-data linkages, and the need to distinguish crop recovery from other vegetation growth.