Analysis Of Cadmium Stress In Rice Using Phenological Characteristics From Remote Sensing Data

Cadmium?Cd?contamination in soil has become an increasingly serious concern because of the rapid development of industry in China.Cd not only destroys the function of soil and inhibits crop growth but also poses a severe threat to human health through the food chain,even when Cd concentrations in the soil are low.Phenology is a sensitive indicator of environmental change,and soil pollution by heavy metals will change the phenological characteristics of above-ground crops.The phenological metrics of crops can be obtained by remote sensing method,the remote sensing phenological information provides a new idea for monitoring soil Cd stress.Based on remote sensing phenological characteristics,this study integrated two phenological metrics to develop a Stress Index?SI?to evaluate Cd stress levels of rice,which can realize rapid and accurate evaluation of Cd stress degree of rice.This research selected four experimental areas located in Zhuzhou City,Hunan Province,China.Soil samples were collected from the rice experimental areas and heavy metal concentrations were measured.The heavy metal concentrations were divided into mild stress and severe stress levels,each with two rice experimental areas.Eight cloud-free quality Sentinel-2 images were acquired and processed from July to October 2017.Firstly,the red-edge chlorophyll index?CIre?were calculated from sentinel-2 data,and the CIre time-series curve was obtained by using the Double-logistic function fitting method,and the tillering period?T?,the heading period?H?and the maturity period?M?of rice were extracted by using the first derivative analysis method.Then,phenological metrics related to Cd stress of rice were obtained and related phenological metrics were analyzed by Fisher discriminant analysis.The phenological metrics with the highest discrimination accuracy were selected to develop the SI.Finally,the SI calculation was carried out for rice fields in a large area of xiangjiang river basin,and the spatial distribution of Cd stress level was obtained.The main conclusions were as follows:?1?By analyzing the difference of CIre and NDVI in the mild and heavy Cd stress levels of sentinel-2 images,it was found that the difference of CIre in the two stress levels were greater than that of NDVI,which proved that the CIre was more sensitive to the health status of rice than NDVI,and the CIre had better ability to distinguish the Cd stress of rice.?2?The phenological metrics under mild Cd stress were generally higher than those under severe Cd stress.These phenological metrics including the time interval of the tillering to maturity phase(DOY_T-M),the daily CIre and their corresponding first derivative value?FD?,the CIre value of the tillering period?CIre_T?,the maximum value of CIre(CIre_MAX),the CIre value of the maturity period?CIre_M?,Skew,CIre_MAX/DOY_T-M,and the change rate of CIre?CRCIre?and the time-integrated CIre?TICIre?,which were used to characterize the photosynthetic rate and biomass,respectively.?3?Fisher discriminant analysis was used to analyze TICIre_T-H and TICIre_H-M?CIre_MAX and DOY_T-M?as well as CRCIre and TICIre at the three growth phase.The SI constructed based on CRCIre and TICIre of the tillering to heading phase?the T-H phase?can effectively evaluate the stress levels of Cd.When SI?0,rice was under mild Cd stress and when SI<0,rice was under severe Cd stress.The overall accuracy was 86.02%,and the accuracy in identifying severe stress was 97.30%.?4?Through the SI calculation of rice pixels in the study area,it was found that the rice pixels under mild Cd stress accounted for 65.70%of the total number of rice pixels.The rice pixels under severe Cd stress accounted for 34.30%of the total number of rice pixels,which proved that the rice Cd stress in the study area was very serious.Severe Cd stress is mainly distributed near the vicinity of Xiangjiang river and residential areas.As the distance away from Xiangjiang river increases,away from residential areas,Cd stress levels of rice gradually changed from severe stress to mild stress.