Spatial Analysis of Weather-induced Annual and Decadal Average Yield Variability as Modeled by EPIC for Rain-fed Wheat in Europe

In our analysis we evaluate the accuracy of near-term (decadal) average crop yield assessments as supported by the biophysical crop growth model EPIC. A spatial assessment of averages and variability has clear practical implications for agricultural producers and investors concerned with an estimation of the basic stochastic characteristics of a crop yield distribution. As a reliable weather projection for a time period of several years will apparently remain a challenge in the near future, we have employed the existing gridded datasets on historical weather as a best proxy for the current climate. Based on different weather inputs to EPIC, we analyzed the model runs (as implemented by IIASA and BOKU) for the rain-fed wheat for 1968-2007 employing AgGRID/GGCMIii) simulations that use harmonized inputs and assumptions (weather datasets: GRASP and Princeton)