Identifying crop heat stress with MODIS and FLUXNET data

Abstract

Stress caused by high temperatures is a critical limiting factor of crop growth and development. Accurately identifying heat stress is crucial to assess and mitigate the negative impact of high temperatures on crop growth. However, isolating the independent effects of heat stress from other factors, such as moisture stress, poses a challenge in field conditions. This study developed an innovative approach to distinguish crop heat stress periods from normal growth conditions, disentangling them from moisture stress and light limitation in croplands. Utilizing FLUXNET data, including air temperature, gross primary productivity, soil water content, and shortwave radiation observations, we identified 78 heat periods and corresponding normal growth conditions. The identified heat and normal periods were further related with remote sensing to extend the identification process to a large scale. Single bands and spectral vegetation indices (VIs) derived from MODIS were employed to evaluate the capability of multispectral data in detecting heat stressed crops from healthy crops. The analysis revealed a significant increase in the reflectance of red band during heat stress. VIs, in general, enhanced the visibility of heat-induced spectral variations and exhibited sufficient capability in distinguishing crops at heat and normal conditions. Visible bands-based indices (EVRI, GLI, and NGRDI) exhibited the highest distinguishability (p-value < 0.01 in the Mann–Whitney U test). These findings underscore the significance of visible bands, especially the red band, in advancing large-scale crop heat stress detection, agricultural monitoring, and crop modeling considering heat stress.