Models Of Estimating Nutrient Elements In Leaves And Canopy Of Zizyphus Jujuba Using Hyperspectral Data

Ziziphus jujuba is native to China.In recent years,with the development of forestry industry in Xinjiang,jujube has become the main economic tree species in Xinjiang,and has also become the main mode for farmers to adjust the structure of agricultural industry and increase economic efficiency.Due to the lack of timely,timely,and non-destructive timely monitoring techniques for nutrient content in jujuba,it is not conducive to the monitoring of growth and nutritional status of large-scale planting dates.Our study focus on Ziziphus jujuba planted in a large scale at Aksu red-flag farm.By combining outdoor observation measurement and indoor chemical analysis,we analyzed the changes at different Ziziphus jujuba growth time with spectrum reflectivity(R_λ)and first order differential spectrum reflectivity(R’_λ)in leaf and canopy scale,respectively.So we got the dynamic changes of the nutrient elements,such as,N,P,K,Ca,Mg,Fe,Mn,Cu and Zn,in different period.And we also tried to find the relationship between different period’s leaf and canopy spectrums and the nutrient elements.Finally,we successfully constructed a hyperspectral spectrum estimation model to estimate the content of nutrient elements.Our main results as followings:As suggested in the followings’analysis about leaf and canopy spectrum features in different period,the difference of the periods’’leaf spectrum features is very small,this is because leaf spectrum features main influenced by the content of chlorophyll and inner structure.But a large change happened at canopys spectrum.This large and obvious change may be due to some factors such as the canopy structure,tree’s growth condition,or water enough or not,besides the chlorophyll content and leaf inner structure.Then there exists some difference of the content of nutrient element in different growth period,such as the changes of P and Cu were relatively stable,relatively small for Mn,Zn,Ca and Mg,but relatively large for N,K,Fe element.These differences are main because the tree’s different needs of nutrient elements during different growth period.By analyzing the relationships between leaf and canopy spectrum and nutrient elements content,we can see clearly that first order differential spectrum analaysis can significantly enhance the relationships between spectrum and element content.The sensitive wave range of leaf and canopy in different period have some difference,for leaves,the range main at visible light band and near infrared wave band.But for canopy,it’s sensitive wave range distribute almost the whole light wave band,due to the influence of canopy structure,chlorophyll content and leaf inner structure,etc.Finally,by using leaf and canopy reflectivity’s sensitive wave bands as independent variables,leaf and canopy nutrient elements content as dependent variables.We constructed a different periods suitable nutrient elements content inversion model for Ziziphus jujuba.Unambiguously,this model finally can estimate the content of N,P,K element very well using sensitive wave band as independent variables,and also well estimation results for Ca,Mg,Fe,Mn,Cu,Zn by using first order differential double sensitive wave bands as dependent variables.But,the fitting degree of canopy is lower than leaves within this model,so how to improve the fitting accuracy of this model needs some more deep research.