A deep dive into techniques to combine ground-based data and remotely sensed data for training and validation of machine learning algorithms used in geospatial impact evaluations of climate sensitive agricultural interventions.