An approach for heavy metal pollution detected from spatio-temporal stability of stress in rice using satellite images

Stable stressors on crops (e.g., salts, heavy metals), which are characterized by stable spatial patterns over time, are harmful to agricultural production and food security. Satellite data provide temporally and spatially continuous synoptic observations of stable stress on crops. This study presents a method for identifying rice under stable stress (i.e., Cd stress) and exploring its spatio-temporal characteristics indicators. The study area is a major rice growing region located in Hunan Province, China. Moderate-resolution imaging spectroradiometer (MODIS) and Landsat images from 2008–2017 as well as in situ measurements were collected. The coupling of a leaf canopy radiative transfer model with the World Food Study Model (WOFOST) via a wavelet transform isolated the effects of Cd stress from other abrupt stressors. An area wavelet transform stress signal (AWTS), based on a time-series Enhanced Vegetation Index (EVI), was used to detect rice under Cd stress, and its spatio-temporal variation metrics explored. The results indicate that spatial variation coefficients (SVC) of AWTS in the range of 0–1 ha d a coverage area greater than 70% in each experimental region, regardless of the year. Over ten years, the temporal variation coefficients (TVC) of AWTS in the range of 0–1 occurred frequently (more than 60% of the time). In addition, the Pearson correlation coefficient of AWTS over two consecutive years was usually greater than 0.5. We conclude that a combination of multi-year satellite-derived vegetation index data with a physical model simulation is an effective and novel method for detecting crops under environmental stress. A wavelet transform proved promising in differentiating between the effects of stable stress and abrupt stress on rice and may offer a way forward for diagnosing crop stress at continental and global scales.