GARBAL: An EO-powered call centre for farmers and herders in the Sahel

GARBAL is a specialized call center service that bridges the digital divide for farmers and pastoralists across Mali, Niger, and Burkina Faso by transforming complex Earth Observation (EO) data into actionable advice. By integrating satellite imagery on vegetation and water, weather forecasts, and real-time ground data on market prices and cattle concentrations the service provides localized insights via a simple phone call. Operating in 9 local languages, GARBAL allows users without smartphones or internet access to receive precise 'scripts' from call center agents regarding pasture availability, water points, and weather forecasts, effectively turning sophisticated space-based data into a practical tool for daily decision-making in the Sahel.

“If you can predict when the rainy season will start,” opined Amadou, “it will help us decide our planting schedule.” He was one of a group of 30 farmers and livestock herders who had crowded under a shade tree in Houndé, Burkina Faso, to attend our focus group on using remote sensing data for local agriculture. “And if a satellite can show me where last month’s rain fell,” mused a herder, “I will know where to pasture my sheep.” Incongruously, given the myriad of other issues facing the community, all anyone wanted to talk about was satellites.

Similar to many drylands across the world, transhumant movements in the West African Sahel are determined by a diverse array of factors including pasture availability, rainfall, access to water, market conditions and security. Timely and reliable information on these factors is critical for pastoralists for making decisions on mobility. Endogenous information systems exist within pastoral communities to relay much of this information, however they are often reliant on the physical presence of an interlocutor such as a scout who seeks out pasture conditions, which brings risks associated with timeliness and insecurity. Regular data on natural conditions like pasture and water have become increasingly available to the spatial community through the growth of open access satellite imagery and open source GIS software. However, few channels exist to transmit this information to pastoral communities.

This focus group was one of dozens conducted over the past decade for GARBAL, a call centre operating in Mali, Niger and Burkina Faso since 2015 that uses EO data to provide decision insights to farmers and herders in the area. These discussions were how we defined our indicators and prioritised datastreams. Thanks to GARBAL, farmers and herders can access the full benefits of EO imagery with a simple phone call, no computer or smartphone needed. Often fielding more than one million calls per year, the service provides live information on pasture availability, surface water and rainfall to any caller via an interface accessed by a call centre agent. The information is provided in 9 languages and in many places GARBAL is the only source of localized weather forecasts and agrometeorological data.

GARBAL is run through a consortium coordinated by the NGO SNV, whereas the call centre is managed by the telecom providers Orange in Mali and Burkina Faso and Airtel in Niger. Field data on livestock and markets are collected by producer organisations (part of the RBM Network) and static datasets on pastoral infrastructure are provided by government agencies in Mali and Burkina Faso. The data platform is run by Hoefsloot Spatial Solutions, a Dutch geospatial company.

GARBAL offers callers a variety of data from Earth observation and field sources, blending satellite imagery with traditional knowledge systems. Field data is collected and treated through pastoral networks, relying on traditional information systems for assessing pasture quality and herd concentration. Information from dozens of sites on market prices, cattle concentration and pasture quality is collected on a weekly basis, processed and displayed in an interactive dashboard for the call centre agent.

Earth observation imagery is collected and processed into a number of indices, notably the Normalised Difference Vegetation Index (NDVI), Dry Matter Productivity (DMP) and Small Water Bodies (SWB), from a number of different satellites including Sentinel-2 and PROBA-V. These images, collected on a dekadal basis, allow for a near-real-time picture of pasture and water availability, which is likewise displayed alongside the field data in the dashboard. The dashboard offers both a map and a textual output, or “script”, providing an analysis of the latest data (biomass conditions, distance to the nearest water point) listed in the table below to help guide the call centre agent during their discussion with the caller. Information from government databases is also included, allowing the caller to know where to find veterinary services and ground water sources. Testing with call centre operators has shown that the large volumes of calls received often prevent agents from being able to carefully analyse maps or graphical outputs. As a result, the textual outputs of the service, which provide basic analysis of the data in a “dialog box”, tend to be used more often than maps by the call centre agents.