Improving the Crop Germination Cover for Smallholder Agriculture in East Africa Using Multi-Sensor Data

Smallholder farmers in Africa are vulnerable to climate risks. Spring drought, in particular, is a common threat to farmers rely on the start of the rainy season to germinate their seeds. If it doesn't rain, they may lose a substantial fraction of their investment and be left without the money to buy more seeds for replanting. An emerging business model is thus to provide hybrid seed with a germination cover, mostly based on comprehensive weather-index. However, due to the sparse weather station and limited weather forecast capacity in these underdeveloped regions, weather-based index insurance can hardly resolve individual fields and hence the actual farm-level loss. In this study, we will demonstration how well can high-resolution satellite data detect crop progress in early growing season, especially crop germination at field level. Our objectives are 1) to assess Sentinel-1 and 2 data fusion results over this smallholder farming system; 2) to map early season crop phenology at 10m resolution using fused Sentinel data; and 3) to identify the relationships between remotely sensed crop phenology metrics and the crop progress data at district level derived from thousands of farmers. We expect this early-stage efforts can inform the design and implementation of germination insurance program.