Effects Of Salinity On The Changes Of Water, Salinity Content In Cotton-Soil System And Its Monitoring Research

Salinity is considered to be one of the major limiting factors for plant growth and agricultural productivity. Cotton is one of the most important economic crops in China, which has been reported to be salt tolerant. With the reduction of field area, more and more cotton was planted in saline soil. Thus, study the effect and monitoring of soil salinity on water and salinity content is one of important research in development of agriculture. The objective of this study was to determine the crop water stress index, cotton water, salinity content, soil electrical conductivity monitoring model based on hyperspectral reflectance, ascertain the suitable frequency for monitoring soil water and salinity content and establish the imaginary part of dielectric constant models through comprehensive use of the infrared temperature measurement technology, spectral radiation technology and electromagnetic spectrum technology, on the basis of multiple pot experiments under varied soil salinity levels with Sumian12(salinity-sensitivity) and CCRI-44(salinity-tolerance) at Pailou experimental station of Nanjing Agricultural University, in order to provide the technical supports for real-time estimation and precision diagnose of plant water and salinity content in cotton under saline conditions.The main results were as follows:1. The construction of the cotton water stress index and changes of related physiological characteristics under salinity conditionSoil salinity significantly reduced the transpiration rate, water content and net photosynthetic rate of cotton functional leaves. The lower equation of the cotton water stress index was set up in1.25dS m-1salinity rate (well watered), and the cotton water stress index based on the above lower equation under different salinity rates was constructed. Comprehensive analysis the relationship between cotton water stress index and leaf water content and net photosynthetic rate revealed that the cotton water stress index is a good indicator to detect cotton water stress in salinity field. 2. Exploring hyperspectral bands and estimation indices for leaf water content of cotton (Gossypium hirsutum L.) in saline soilThe sensitive spectral bands for EWT and RWC occurred mainly within the near infrared (NIR) and short-wave infrared (SWIR) ranges. The best spectral indices for estimating leaf water content (RWC and EWT) in cotton were NDSI1(R1222, R2264), NDSI2(R1347, R2307), RSI1(R2264, R1321), RSI2(R2307, R1347) and1650/2220nm ratio, and the linear regression models based on the above spectral indices were identified as the best equations for the effective estimation of EWT and RWC in cotton. From testing of the derived equations, the model for EWT estimation based on the NDSI2(R1347, R2307) and RSI2(R2307, R1347) gave R2over0.85with more satisfactory performance than the spectral indices1650/2220nm ratio and the RWC models based on NDSI1(R1222, R2264), RSI1(R2264, R1321) in saline soil. The present spectral parameters of NDSI2(R1347, R2307) and RSI2(R2307, R1347) can be used for monitoring plant water stress in cotton cultivated in saline soil.3. Monitoring cotton(Gossypium hirsutum L.) leaf ion content in saline soil with hyperspectral reflectanceThe Na+, Cl-, and SO42-content in functional cotton leaves increased with the soil salinity rates increasing. In contrast, K+and Ca2+decreased at the same growth stage. The best spectral indices for estimating cotton leaf ion content were found to be NDSI (R1340, R2306) RSI(R2306, R1347); NDSI (R,346, R2276), RSI (R2276, R1343); NDSI (R1380, R2307), RSI (R2306, R1350); NDSI (R1200, R2211), RSI (R2202, R1361); NDSI (R1300, R2250), RSI (R2264, R1335); and NDSI (R1154, R2317), RSI (R2317, R1154) for K+, Na+, Ca2+, Mg2+, Cl-, and SO42-, respectively, and the linear, power and exponential regression models based on the above spectral indices were formulated. Among them, the linear equations based on RSI(R2306, R,347)、RSI(R2276, R1343)、RSI(R2306, R1350), NDSI(R1340,R2306)、NDSI(R1346,R2276)、 NDSI(R1380,R2307)、NDSI(R1200,R2211)、NDSI(R1154、R2317), the power equations based on RSI(R2202, R1361)、RSI(R2317, R1154), NDSI(R1300,R2250) and exponential equations based on RSI(R2264, R1335), for K+, Na+, Ca2+, Mg2+, Cl-, and SO42-, respectively, can well estimated the ion content of cotton under different levels of salinity, After testing of the derived equations, the high fit between the measured and estimated values indicate that the present models based on RSI is better than the models based on NDSI, and could be used for the reliable estimation of leaf salinity in cotton plants with R2greater than0.69under different saline conditions.4. Monitoring simulation of soil electrical conductivity based on hyperspectral parameter of cotton(Gossypium hirsutum L.) functional leaves.During near-infrared and middle-infrared spectral bands, with soil salinity rate increased, the spectral reflectance of cotton functional leaves increased, and spectral parameter Normalized difference spectrum index (NDSI) based on1350nm and2307nm correlated to soil electrical conductivity well, soil EC monitoring model was constructed as EC=-42.899NDSI (R1350, R2307)+27.338, with vegetation index NDSI (R1350, R2307) as independent variable. Vegetation index Thematic Mapper5(TM5-SWIR) was most correlation to soil EC during all derivative spectral parameters, so soil EC monitoring model was constructed as EC=0.0574(TM5-SWIR)2-2.5928(TM5-SWIR)+30.021, with vegetation index TM5-SWIR as independent variable. Take2009experiment data to test soil conductivity models, and show that with the predicted values of soil EC by the two models were very consistent with the observed values, with determination coefficient of0.887,0.8136,and root mean square error (RMSE) of1.09ds-m-1,1.29ds-m-1. The experiment shows that soil EC in saline cotton field can be effectively monitored by two spectral parameters of NDSI (R1350, R2307) and TM5-SWIR.5. Monitoring simulation of soil electrical conductivity based on hyperspectral parameter of cotton (Gossypium hirsutum L.) functional leaves.As the soil salinity increasing, the Na+、K+、Ca2+、Mg2+、Cl-、SO42-and HCO3-content increased. At the same salinity levels, the Na、K+、Ca2+、Mg2+、Cl-、SO42-and HCO3-content declined as the postpone of the growth stage. The imaginary part of soil dielectric constant model was developed through the relation of imaginary part of dielectric constant (ε"), soil bulk conductivity, conductivity of soil solution, and soil ion content in mixed-salinity soil to retrieve soil ion content. After testing with the data of2009, the results showed that the real part of dielectric constant model based on Dobson had high values of R2, low values of RMSE, and can be used to retrieve soil water status well. The inversed values of total concentration of salt (Sc), Cl-, Ca2+by using the model were very consistent with the observed values, with high R2values of0.9417,0.6852,0.7965, and root mean square error (RMSE) lower than0.34g/kg,0.09g/kg and0.13g/kg, respectively. However, the RMSE of total concentration of salt (Sc), Cl-, Na+were small at high frequencies (C-band), with R2values of0.8655,0.8408,0.8117,0.8217, and RMSE value of0.25g.kg-1,0.15g.kg-1,0.02g.kg-1,0.21g.kg-1respectively. The high fit between the measured and invered values indicate that the present models could be used for the reliable estimation of soil salinity in cotton field under different saline conditions.