Deep Learning for Vegetation Health Forecasting: A Case Study in Kenya

East Africa has experienced a number of devastating droughts in recent decades, including the 2010/2011 drought. The National Drought Management Authority Kenya relies on real-time information from MODIS satellites to monitor and respond emerging drought conditions arid semi-arid lands Kenya. Providing accurate timely vegetation health—and its probable near-term future evolution—is essential for minimising risk evolving into disasters as country’s herders directly rely grasslands. Methods field machine learning are increasingly being used hydrology, meteorology, climatology. One particular method that shown promise rainfall-runoff modelling is Long Short Term Memory (LSTM) network. In this study, we seek test two LSTM architectures health forecasting. We find these models provide sufficiently forecasts be useful monitoring forecasting purposes, showing competitive performances with lower resolution ensemble methods improved over shallow neural network persistence baseline.