This document discusses an empirical Earth Observation (EO) based approach to wheat yield forecasting and its adaptation within the GEOGLAM framework. GEOGLAM is an international initiative that aims to inform agricultural decisions through coordinated Earth observations. The approach uses multi-temporal satellite data like NDVI to develop quantitative wheat yield forecasts at regional/national scales. A key challenge is obtaining annual crop type maps for spatially explicit time series. Aggregating high-resolution wheat masks to coarser resolutions can mitigate effects of crop rotations over time.