The most comprehensive live agricultural dataset in East Africa

Cropland NDVI minus each unit's own seasonal baseline: a 10 m composite calibrated onto an eight-year 300 m climatology. 243 units: 47 Kenyan counties, 31 Tanzanian regions, 135 Ugandan districts, 30 Rwandan districts. Snapshot 2026-10-19.

No unit on this map is compared with its neighbour. Each is compared with itself: this dekad's cropland vegetation against what the same weeks looked like over the 8 years before. A county that is always greener than everyone around it only counts as above normal once it beats its own history.

At this snapshot 12 of the 243 units read at or above normal, and most of the region was still climbing out of a dry July. The story is the same across 3 independent sources: the July early-warning bulletin named Kakamega, Bungoma, Uasin Gishu, Elgeyo Marakwet and Trans Nzoia as a hotspot; the rainfall anomaly of 26 August put Bungoma at 26 % and Kakamega at 39 % of normal for the long rains; the press signals a week later named the same 5 counties. 4 datasets, 1 shared county vocabulary, 1 answer.

The fresh reading is a 10 m optical product, cloud-masked and cropland-masked, calibrated per unit against an 8-year 300 m climatology. Where the 2 disagree past a fixed tolerance, the platform falls back to the lagged product and says so in the response rather than guessing.

Demand and crop stress