Projecting the impacts of climate change on wheat yields in the UK and Germany and assessing modeling uncertainties

Climate change and variability are projected to negatively affect wheat production in Europe. The impacts of climate change are typically projected using global and regional climate models (GCMs and RCMs) and impact assessment tools such as crop models. However, this impact simulation chain can propagate uncertainty, as errors are introduced by GCMs, RCMs, and crop models. There are also many intermediate steps and decisions in the impact simulation process that are influenced by the different communities of practice that utilize climate and crop models. These differences in methods and approaches can also influence the range of future yield projections. Yield projections are thus considered inherently uncertain because of this cascade of uncertainty.
This interdisciplinary study projects the impacts of climate change on wheat yields in the UK and Germany, two key wheat-growing countries. Added value is found when using RCMs for downscaling temperature and precipitation simulations for the impact assessment. However, these GCM-RCM simulations are shown to have significant errors relative to observations, necessitating a bias correction (BC) step. Different BC methods are shown to be effective in improving simulations. Two BC calibration approaches, one that corrects RCM-only error and the other GCM-RCM error, are used to examine how different GCM-RCM combinations can affect projected changes in climate. Future climate projections are used in a multi-method crop modeling approach, and the uncertainty in the resulting yield projections is analyzed. Key findings are that wheat yields in the UK and Germany will be affected by changes in temperature and precipitation. However, these impacts are shown to be region-dependent and vary based on the crop modeling method, making the choice of crop modeling method a major contributor to uncertainty.