Across much of sub-Saharan Africa, reliable rainfall data remains scarce. Conventional rain gauge networks are sparse, and where they do exist, they often lack the spatial coverage needed to capture the high variability that characterises rainfall across the region. This gap has long posed a challenge for farmers, water managers, meteorological agencies, and disaster-response systems that depend on accurate, timely rainfall information.
A partnership operating under the TEMBO Africa research project is now demonstrating how that gap can begin to be addressed by drawing on an unlikely source of rainfall information: the microwave signals travelling between mobile phone towers.
Commercial Microwave Links (CMLs) are the point-to-point transmission systems that telecommunications providers use to carry data between tower infrastructure. When rainfall occurs along the path of a microwave signal, the signal experiences measurable attenuation a weakening caused by the interaction between the signal and falling raindrops. The relationship between signal attenuation and rainfall intensity has been studied for decades, and researchers have developed methods to invert this relationship and retrieve quantitative rainfall estimates from raw link data.
What makes CMLs particularly valuable in data-sparse environments is their geographic distribution. Telecom infrastructure tends to follow population density and economic activity precisely the areas where weather-related risks are highest and where improved rainfall information would have the greatest impact.
In Ghana, a collaboration between Rainbow Sensing, TU Delft University, TAHMO, AT Ghana, and the Ghana Meteorological Agency (GMet) supported by the Netherlands Enterprise Agency’s Partners for Water programme has achieved significant technical milestones in operationalising CML-based rainfall retrieval.
A critical early step was securing structured access to telecom network data. AT Ghana now provides 15-minute link signal data and associated metadata, made available through a dedicated API designed to handle secure, scalable data transfer. This pipeline feeds directly into near-real-time processing infrastructure deployed within a Dockerised environment, enabling consistent, automated retrieval of CML-derived rainfall estimates.
These estimates do not stand alone. The system integrates CML outputs with satellite-derived precipitation products, surface weather-station observations, and machine-learning methods to produce spatially merged rainfall fields with greater coverage and reliability than any single data source could deliver independently. TU Delft leads the scientific development of the rainfall-retrieval algorithms, while TAHMO contributes ground-based observation data that supports evaluation and validation of the outputs. GMet brings national meteorological expertise and institutional knowledge that helps contextualise the results within Ghana’s operational weather-monitoring framework.
Producing research-grade rainfall maps is one milestone. Translating those outputs into decision-relevant information is another. Rainbow Sensing is now working to integrate CML-derived rainfall products into its data-assimilation and weather-information workflows, moving the system from prototype outputs toward practical climate services that can be delivered at scale.
This integration reflects a broader principle that underpins the TEMBO Africa project: that improving climate monitoring across Africa requires not only scientific innovation, but the development of end-to-end systems capable of converting data into actionable information for users on the ground.
The downstream applications are significant. More spatially complete rainfall estimates can improve the agronomic advisories that help farmers make planting and irrigation decisions under climate uncertainty. They can strengthen the hydrological monitoring that water managers depend on for reservoir operations and catchment management. They can enhance the early-warning systems that disaster-response agencies use to anticipate flooding and heavy-rainfall events. And they can support community-level resilience by improving awareness of rainfall variability and dry-spell patterns.
Ghana’s CML pilot is significant not only for what it achieves locally, but for what it demonstrates more broadly. Telecom infrastructure exists across Africa at a scale that conventional meteorological networks do not. The methodology being developed and refined through the TEMBO Africa partnership combining CML signal data, satellite observations, ground measurements, and machine learning offers a replicable framework for expanding rainfall monitoring coverage in other data-sparse environments across the continent.
The partnerships required to make this work are as important as the technology itself. Collaboration between a telecommunications provider, a national meteorological agency, academic researchers, ground-observation networks, and a commercial weather-services company is not straightforward to establish or sustain. The progress achieved in Ghana demonstrates that it is possible and that the results justify the effort.
As the TEMBO Africa project continues, the Ghana CML pilot represents one of the clearest examples of how the project’s core ambition using innovative data sources and scientific methods to strengthen climate monitoring in Africa is being put into practice.
READ MORE
