Remote Sensing Based Analysis Methods For Heavy-metal Stress Levels In Rice Responses To Physiological Function Variations

Rapid and accurate analysis of heavy-metal contamination in crops in a large area is important with regard to food security and farmland protection,which has been made possible by the development of remote sensing technology.Most studies on monitoring heavy metal stress in crops using remote sensing technology have mainly focused on establishing empiricalmodels between physiological element characteristics(or heavy metal concentrations)and spectral indices.However,the variations of physiological functions have rarely been considered.Moreover,visible/near-infrared(VNIR)range has been widely adopted,while few studies have considered thermal infrared(TIR)characteristics of heavy metal stress.The study area was located in the county of Zhuzhou,Zhuzhou City,Hunan Province,China,which is a primary grain producing area.The data sets collected included heavy metal concentration,leaf area index(LAI),canopy temperature,meteorological data and remote sensing data,including HJ-1 CCD,Landsat 8 OLI/TIRS and GF-1 image data.This study focused on analyzing the methods for monitoring heavy metal stress levels using remote sensing technology with consideration of mechanism and theoretical modelregarding to the variations of physiological functions.The most important findings and conclusions drawn from this study include:(1)The changes in two important physiological functions,i.e.,the photosynthesis rate and dry-matter formationefficiency were chosen to represent the heavy-metal stress levels.Two stress indices which correspond to the daily total CO2 assimilation and dry-matter conversion coefficient,were incorporated into the WOrld FOod STudies(WOFOST)crop growth model.The two stress indices were calculated based on assimilation of remotely sensed LAI and simulated LAI by WOFOST model.The contribution rates ofthe two indices for stress in the WOFOST model were calculated,and based on which,a synthetic stress index was built.The discrimination rule for two stress levels was built using a small amount of sample points,which then was applied to each rice pixel in the study area,enabling the continuous spatial evaluation of heavy-metal stress levels with a correct discrimination rate higher than 90%.(2)Thechanges of rice canopy temperature under heavy metal stress were analyzed,as well as the extreme states of rice growth and the influence mechanism of canopy temperature.A general Canopy-Air Temperature Difference / Leaf Area Index Triangle Theoretical Model was developed based on physical energy balanceequationsto assess heavy metal stress levels in rice.A normalized heavy metal stress index(HMSI)is proposed based on the theoretical model.The model was verified at the field scale,and then was applied in reginal level based on VNIR and TIR remotely sensed data for validation.(3)The dry weight of rice roots and the synthetic stress index that can reflect the variations of two important physiological functions in plant were selected as the stress signs for rice heavy metal contamination.The optimal spatial resolution of image for monitoring the heavy metal stress was identified using the fractal analysis method.Stronger spatial differentiation of the two variables will lead to a higher fractal dimension(FD)value,which can better characterize the subtle changes of stress signs under stress.Therefore,the resolution corresponding to the highest FD value is the optimal spatial resolution for monitoring heavy metal stress.The results showed that the data with resolutions between 4 m and 32 m are appropriate for heavy metal monitoring,and 4 m is the optimal one.