
LeiGhaThis product shows the distribution of rainfed and irrigated agricultural land across Kenya. This dataset has been produced using a large training dataset that contains a total area of 71,900 km² in training samples. The algorithm in the product used three years of Sentinel-2 data (2021-2023). Harmonics of NDVI were used to overcome the limitations of data gaps due to clouds and noise in the time series. The harmonics of the NDVI were then used along with averaged evapotranspiration (ET) and land surface temperature data from MODIS to feed a random forest classifier to produce the final product, along with several post-processing steps to smooth the dataset. The product can provide meaningful insights into areas in need of intervention due to climatic stress and areas within proximity to alternative water sources to assist in irrigation decision making.