Characterizing crop productivity under heat stress using MODIS data

Stress caused by high temperatures is a critical limiting factor of crop growth and development. Although remote sensing has been used to investigate the impacts of high temperatures on crops, its ability to detect heat stress independently of other stressors and assess its effects on gross primary production (GPP) estimation is unclear. This study developed an innovative approach to distinguish crop heat stress periods from normal growth conditions in croplands independent of water stress and light limitation. Multispectral broad bands and spectral vegetation indices (VIs) derived from MODIS for 78 periods of heat stress were used to assess the sensitivity of canopy reflectance to heat stress and its impacts on GPP. Results reveal that heat stress significantly increased the reflectance in the red band. VIs, in general, enhanced the detection of heat-induced spectral variations, and exhibited sufficient skill in distinguishing crops under heat stress and normal conditions. Three visible-based indices (the Visible Atmospherically Resistant Index, the Green Leaf Index, and the Normalized Green–Red Difference Index) exhibited the highest discriminability (p-value < 0.01 in the Mann–Whitney U test), while the Enhanced Vegetation Index displayed the highest accuracy in GPP estimation (R2 = 0.62, RMSE = 5.49, RRMSE = 0.35) under heat conditions. Overall, the isolation of heat stress impact on crop growth has important implications for advancing large-scale crop modeling and climate change studies, for example, incorporating the suggested VIs into temperature response simulations within crop models.