Study On Spectral Monitoring Cotton And Soil Water Salt Change In Jiangsu Coastal Saline

Saline soil was an important limiting factor for restricting land productivity in coastal areas,which restricted regional agricultural development in crop growth and production.Cotton was as the pioneer crops to saline alkali land.The abilityin monitoring changes of cotton salt-water and changes ofsoil water-salt were very important for cotton yield and quality.To construct hyperspectral monitoring model for cotton leaf water-salt and hyperspectral monitoring and soil water-salt in cotton field,We researched of cotton leaf spectra and surface soil spectral inquiry and determined the close relationship between cotton leaf water-salt content and cotton leaves spectrum,between soil water-salt content and soil surface spectrum.Based on the coastal saline alkali land area of Jiangsu Province,this study selected 5 representative ecological experiment sites of salinity gradient.We obtained cotton leaf light spectrum of cotton growth period and cotton soil surface spectral indoor.The study focused on:the water-salt monitoring precise model inhyperspectral parameters of cotton functional leaves;the hyperspectral remote sensing monitoring model in coastal saline soil moisture;the hyperspectral remote sensing monitoring model in coastal saline soil salt.The main results were as follows:1 Study on Water and Salt Monitoring Model Based on Hyperspectral Parameters of Cotton Functional Leaves in Coastal Saline Alkali LandOur study selected 5 representative ecological experiment sites of salinity gradient in coastal saline alkali soil of Jiangsu Provincein 2014-2015,and made clear the relationship between cotton leaf spectral reflectance and leaf water-salt content.Besides,we corrected model parameters based on existing leaf water salt spectral monitoring model and built up accurate saline cotton salt water spectral monitoring models.The results showed that when the agronomic parameters were water,the larger band about RWC,EWT and leaf spectral reflectance of 1390-2430nm,1390-2430nm(r>0.6),respectively,showed a negative correlation and positive correlation,the spectral parameter correction model accuracy was better than previous model;when agronomic parameters were salt,the larger bands about Cl-,Na+ content and leaf spectral reflectance of 1390-2450nm,1390-2440nm(r>0.6),showed a positive correlation and the correction of spectral parameters of model accuracy was better than previous models.Therefore,the new constructed models land about cotton leaf RWC,EWT,Cl-,Na+ content and leaf spectral reflectance in coastal saline alkali could be used to better monitoring the state of water and salt of cotton leaves in the coastal saline soil.2 Study on Hyperspectral Parameters and Prediction Model of Soil Moisture in Coastal SalineBased on analyzing the data of soil moisture content and soil surface spectral reflectance indoor from five sampling points of coastal saline soil and the relationship between soil moisture content and soil spectrum in wavelength 350-2500nm;we determined spectral parameters under Ratio Spectral Index(RSI),Normalized Difference Spectral Index(NDSI)and Difference Spectral Index(DI);established quantitative model of soil moisture content.It showed significant negative correlation between spectral reflectance and soil moisture content,and the maximum negative correlation was near 1930nm(r=0.86).By comparion of the regression equation of RSI,NDSI and DI,the result showed that the regression equation of exponential function built by soil moisture content with being based on RSI(R1407,R1459)presented the maximum 82(0.780)and the minimum SE(0.016),regression equation:y=0.00001e9 72053x.It could be better monitored saline soil moisture conditions.The established the model for estimating soil moisture content based on RSI(R1407,R1459)could be used to monitor soil moisture content accurately in Jiangsu coastal saline soils.3 Study on Hyperspectral Monitoring of Soil Electrical Conductivity in Coastal Saline Alkali SoilThe experiment based on coastal saline soil of different salinization of typical test points,through the indoor access to surface soil hyperspectral reflectance,a comprehensive analysis of the relationship between soil electrical conductivity and soil reflectance spectra in different forms of transformation,determined the construction model of sensitive bands,through different forms of transformation stepwise regression equation monitoring model comparison and establish the best coastal saline alkali land soil salt content spectral monitoring model.Results showed that there was a significant negative relationship between soil electrical conductivity and soil spectral reflectance,the regression equation based on the logarithm of the inverse transform to 760nm and 1052nm as the independent variables has obtained good monitoring effect,test index decided coefficient(R2),the root mean variance(RMSE)minimum and slope(Slope)closest to 1,the model equation:y=88.77158-44.1497x1+90.0946x2.Conclusion:the constructed estimation model of the soil electrical conductivity based on the logarithm of the reciprocal transformation could realize the accurate detection of the soil salinity in the coastal saline alkali soil of Jiangsu.In coastal saline alkali land,the soil water and salt movement affected the water and salt content in cotton leaves,and the spectrum directly reflected the change process.Study found that they were dependence relation and sensitive bands were 1390-2430nm.Scholars constructed the estimation models of cotton leaf water-salt content based on soil salinity simulation,but it need to improve accuracy according to field experiment.The research under the actual situation of coastal saline alkali soil modified existing estimation model.It could play an important role for model practical application.In addition,correlational dependence is between soil spectra and soil water-salt content,and the increase of soil spectra caused by the increase of soil moisture and soil conductivity,and thus established the quantitative model.Methods about spectral pre-processing method and the spectral index makes the model more accurate,but it needed field experiment to do further research because of the limit of spectrum in the room.