Quantifying the response of wheat yields to heat stress: The role of the experimental setup

Previous studies suggested a wide range of sensitivities of wheat yields to heat stress around anthesis. The aim of this study was to improve the understanding of the reasons of the disagreement by testing the response of wheat yield and yield components to differences in the method of heating, the temperature measurement point and soil substrate under sole heat and combined heat and drought stress around anthesis. Growth chamber experiments performed at different sites showed that increasing of the ambient air temperature at anthesis corresponding to a temperature sum of 12000 °C min above 31 °C resulted in a significant yield reduction of −24% for plants grown on sandy soil substrate but not for those grown on a soil with high soil water holding capacity. The grain yield of wheat also declined by −16% for sandy soil substrate but at a much lower level of heat stress when the temperature of the ears was increased by infrared heaters (a temperature sum of 1900 °C min above 31 °C). The yield reduction increased significantly under combined heat and drought compared to sole heat stress. Grain number significantly declined in all experiments with heat stress and combined heat and drought stress at anthesis. Single grain weight increased with heat stress around anthesis and partly compensated for lower grain numbers of pots containing a soil with high soil water holding capacity but not in experiments with sandy soil substrate. We demonstrate, based on data from previous heat stress studies, that statistical relationships between crop heat stress and yield loss become stronger when separating the data according to the soil used in the experiments. Our results suggest that the differences in the yield response to heat may be caused by additional drought stress which is difficult to avoid in heat stress experiments using sandy soil substrate. We conclude that differences in the experimental setup of heat stress experiments substantially influence the crop response to heat stress and need to be considered when using the data to calibrate crop models applied for climate change impact assessments.