Modeling Canopy Temperature As a Driver of Heat Stress in Wheat.

The temperature of a crop canopy (T_can) often differs substantially from the reference air temperature (T_air) that commonly is assumed to drive processes in crop simulation models. To improve our understanding of how heat stress affects wheat (Triticum aestivum), the AgMIP wheat team compared nine models that estimate canopy temperature and found that an energy balance correcting for atmospheric stability conditions was required to obtain accurate values of T_can. We describe efforts to better understand sources of variability in comparisons of measured and simulated T_can with emphasis on models that correct for atmospheric stability. Factors considered included use of daily maximum T_air as a proxy for T_air at time of measurement, differences in simulated vs. actual crop growth and water use, and possible associations between specific weather variables and residual error. Considering T_air at time of measurement substantially improved accuracy of T_can, emphasizing the need for weather and T_can data with higher temporal resolution. Large effects of simulated evapotranspiration and crop growth confirmed that errors simulating in other crop processes propagate to T_can. Weak associations of individual weather variables with residual error suggest that fundamental errors or bias in modeling the energy balance are currently less important than temporal resolution or overall accuracy of simulated crop water use and growth. Our results raise questions about possible limits inherent in modeling crop responses with daily weather.