LAUSR

Accurately assessing the heavy-metal contamination in crops is crucial to food security. This study provides a method to distinguish heavy-metal stress levels in rice using the variations of two physiological functions as discrimination indices, which are obtained by assimilation of remotely sensed data with a crop growth model. Two stress indices, which correspond to daily total CO2 assimilation and dry-matter conversion coefficient were incorporated into the World Food Study (WOFOST) crop growth model and calculated by assimilating the model with leaf area index (LAI), which was derived from time-series HJ1-CCD data. The stress levels are not constant with rice growth; thus, to improve the reliability, the two stress indices were obtained at both the first and the latter half periods of rice growth. To compare the stress indices of different stress levels, a synthetic stress index was established by combining the two indices; then, three types of stress index discriminant spaces based on the synthetic index of different growth periods were constructed, in which the two-dimensional discriminant space based on two growth periods showed the highest accuracy, with a misjudgment rate of 4.5%. When the discrimination rules were applied at a regional scale, the average correct discrimination rate was 95.0%.