Monitoring Rice Heavy Metal Stress Using Phonological Phase Space From High Spatial-temporal Resolution Time Series Data

Heavy metal pollution of farmland has become a major agricultural ecological environment problem,which poses a serious threat to the sustainable development of modern agriculture and social economy,the agricultural ecological environment and the quality and safety of agricultural products.The application of remote sensing technology to accurately monitor the heavy metal stress of a large area crops is of great significance to food security and environmental protection.Most studies on the remote sensing monitoring of rice heavy metals stress mainly focus on the analysis of their physiological functions as well as characteristics of color and morphology,but these methods can not sufficiently describe the effects of heavy metal stress on rice.The development of methods such as spatio-temporal fusion,time series analysis and phenological phase space make it possible to use the information of rice growth process for remote sensing monitoring of heavy metal stress.In this study,rice farmland in Zhuzhou was selected as the experimental field.Seven sites in the study area were selected to field data collection of soil for measuring their heavy metal contents and divide the seven sites into different heavy metal stress levels according to the heavy metal content of their soil.At the same time,the MODIS data and Landsat data available in the rice growing season between late april and late october in 2016 were collected.An Enhanced spatial and temporal adaptive reflectance fusion model(ESTARFM)is used to generate high spatial and temporal resolution time series data.And the two dimensional phenological phase space model constructed by enhanced vegetation index(EVI)and normalized moisture index(NDWI)time series data are use to describe the growth process of rice,and analyze the effects of different heavy metal stress levels on rice.The main conclusions of this paper are following:(1)The high spatial and temporal resolution time series data with 8 days interval and 30 meter spatial resolution were generated based on ESTARFM spatio-temporal fusion algorithm and used for the extraction and analysis of rice growth dynamic information.In a sense,it overcome the obstacle of cloud cover and remote sensing technologies,making it more practical and more reliable in accurate extraction of rice phenological information..(2)The phenological phase space and rice growth curve are able to express the growth process of rice well,and also depict the difference of rice growth under different heavy metal stress levels.The method of phenological phase space provides a new understanding for remote sensing monitoring of heavy metal stress in rice,which refer to excavating heavy metal stress information from the whole process of rice growth.(3)The GRFI ratio extracted from the rice growth curve in the phenological phase space has a significant correlation with heavy metal stress levels.It can be used to indicate the different stress levels of heavy metals in rice.