摘要：Climate extremes have profound impacts on key socio-economic sectors such as agriculture. In a changing climate context, characterised by an intensification of these extremes and where the population is expected to grow, exposure and vulnerability must be accurately assessed. However, most risk assessments analyse extremes independently, thus potentially being overconfident in the resilience of the socio-economic sectors. Here, we propose a novel approach to defining and characterising concurrent climate extremes (i.e. extremes occurring within a specific temporal lag), which is able to identify spatio-temporal dependences without making any strict assumptions. The method is applied to large-scale heat stress and drought events in the key wheat producing regions of the world, as these extremes can cause serious yield losses and thus trigger market shocks. Wheat regions likely to have concurrent extremes (heat stress and drought events) are identified, as well as regions independent of each other or inhibiting each other in terms of these extreme events. This tool may be integrated in all risk assessments but could also be used to explore global climate teleconnections.
