For smallholder farmers across sub-Saharan Africa, the difference between a successful planting season and a total crop failure often comes down to the first three weeks of rainfall after sowing. Whether seeds germinate depends on whether the soil receives enough moisture in that narrow window. Until recently, tracking that moisture accurately at the farm level was simply not possible for most of rural Africa. TEMBO Africa, working with TU Delft, has changed that using data from the Sentinel-1 satellite orbiting 693km above Earth.
THE CHALLENGE: RAINFALL DATA AT THE WRONG SCALE
Africa’s ground-based rainfall monitoring network is critically sparse. In Northern Ghana, the density of automatic weather stations is far below what is needed to capture the spatial variability of rainfall across agricultural landscapes. Rainfall is inherently patchy — convective storms can drench one community while leaving a neighbouring farm completely dry.
Existing satellite rainfall products fill some of this gap, but typically operate at resolutions of 10 kilometres or more per pixel. At that scale, a single data pixel covers hundreds of farms. When deciding whether a germination insurance payout is triggered, that resolution is not adequate — it tells you what happened over a large area, not what happened on the specific field where the farmer planted.
SENTINEL-1 AND THE FIELD-LEVEL SOLUTION
The European Space Agency’s Sentinel-1 constellation provides synthetic aperture radar (SAR) imagery at 10 metres by 10 metres — a scale equivalent to the size of a single crop row. Unlike optical satellites, SAR works through clouds and at night, making it reliable in tropical climates where cloud cover is frequent.
TU Delft researchers within TEMBO Africa developed a method to detect rainfall occurrence from Sentinel-1 data by analysing short-term changes in the radar backscatter signal from agricultural fields. The concept: within manually selected agricultural fields around TAHMO weather stations, a climatology of the Sentinel-1 backscatter was created at the individual pixel level. This gives a baseline of how wet or dry each pixel should be at each time of year. If the relative wetness increases sharply between two satellite acquisitions typically 6 to 12 days apart this indicates that rainfall occurred during that window.
This data-driven statistical approach is designed specifically for African agricultural contexts, where traditional forward models linking satellite data to soil moisture have been shown to perform poorly. The approach shows particular promise during the transition from dry to wet season, when the first rains occur exactly the window that matters most for germination insurance.
FROM SATELLITE TO INSURANCE PAYOUT
The Sentinel-1 rainfall and soil moisture detection method feeds directly into TEMBO Africa’s Germination Index Insurance (GII) product, developed with the Ghana Agricultural Insurance Pool (GAIP). In the GII model, a payout is triggered if cumulative rainfall in the 21-day window following a farmer’s planting date falls below the threshold required for seed germination.
Using Sentinel-1 data, TEMBO can assess this threshold not based on the nearest weather station which may be tens of kilometres away but based on what the satellite detected over the farmer’s
actual field. A farmer in a rain shadow who did not receive adequate moisture will now be correctly identified by the system, even if a nearby weather station recorded sufficient rainfall.
In the 2025 TEMBO pilot covering 160 enrolled farmers in Northern Ghana, the rainfall index showed over 95% correlation with actual germination outcomes observed in farmers’ fields. A peer-reviewed paper on the Sentinel-1 methodology has been accepted for publication in the Journal of Hydrometeorology.
OUTCOMES
WHY IT MATTERS
The combination of freely available Sentinel-1 data, TAHMO ground-truth validation, and TEMBO’s data-driven algorithmic pipeline represents a scalable, affordable model for field-level rainfall monitoring across Africa. As the Germination Index Insurance product scales to more districts, more crops, and more countries, Sentinel-1 will remain at its core enabling fairer, faster, and more accurate payouts for farmers who need support most.
