National-Scale High-Resolution Crop Condition Maps: Assessing Drought Impact on Croplands in Kenya Using Sentinel-2

The crop condition mapping product monitors cropland pixels affected by drought using Vegetation Indices (VIs), such as the Normalized Difference Vegetation Index (NDVI), Normalized Difference Moisture Index (NDMI), Normalized Difference Red Edge (NDRE), and Green Normalized Difference Vegetation Index (GNDVI), generated from Sentinel-2 images. A binary classification is performed to map drought-affected and unaffected croplands. A random forest model is trained using VI time series data from both drought and non-drought years for each Agro-Ecological Zone (AEZ). The outputs display the spatial characteristics of drought impacts on croplands at a national scale. The dataset includes seasonal crop condition maps for 2016-2022 at a 20m spatial resolution, classifying pixels as 0: non-croplands, 1: unaffected pixels, and 2: drought-affected pixels. Two maps per year are provided for the long rains (season 1) and short rains (season 2). The output is validated through comparison with other datasets such as the Global Drought Observatory, FAO Agriculture Stress Index System (ASIS), East Africa Drought Watch, and reports from the National Drought Management Authority (NDMA). Additionally, user validation has been conducted through engagement with relevant stakeholders, ensuring the outputs align with ground realities and user needs. Each map is accompanied with quality maps based on number of available clear sky observations and classification probability.