Abstract: Testing our understanding of crop yield responses to weather fluctuations at global scale is notoriously hampered by limited information about underlying management conditions, such as cultivar selection or fertilizer application. Here, we demonstrate that accounting for observed spatial variations in growing seasons increases the variance in reported national maize and wheat yield anomalies that can be explained by process-based model simulations from 34 to 58% and 47 to 54% across the 10 mo…