Study On Soil Salinity Factor Spectrum Diagnosis Of The Ebinur Lake Wetland Natural Protection Area

Ebinur Lake Wetland National Nature Reserve is an important part of China's biodiversity and a natural gene pool.It is rich in natural resources and unique in wetland ecological environment.However,with the unrestrained development of wastelands in Xinjiang and the abuse of water resources,the area of soil salinization is increasing,and the problem of salinization in Lake Aibi has become more and more serious,which in turn has eroded farmland land resources,directly affecting Ebinur Lake biological and ecological environment system.Remote sensing technology can monitor the dynamic changes of soil salinity in real time in a short period of time,and can accurately and scientifically obtain information on the degree of salinization,spatial distribution characteristics,water and salt transport laws,and other aspects of soil salinity treatment in the Ebinur Lake region.Utilization and overall planning are of great significance.In this paper,the soil of Ebinur Lake Nature Reserve is taken as the research target,and the characteristics of soil salinity and seven salt factors are analyzed.The ground hyperspectral and Landsat8 image data are used as the spectral data,and the mathematical combination algorithm is used to establish the single-band and multi-band combination indices.Soil salinity factor inversion model.The main conclusions are as follows:1.According to the distribution of salt factors in different soil layers,salt factors mainly accumulated in the soil surface,and the salt content gradually decreased with increasing depth.The content of HCO3-is basically the same in different soil layers;the content of Cl-is slightly decreased with increasing soil depth.The Ca2+ content gradually decreased with the increase of soil depth;Mg2+ content did not change significantly in different soil layers;Na+ content decreased with increasing soil depth.The coefficient of variation of SO42-and Cl-of more than 100%is a strong spatial variability.The maximum coefficient of variation of Cl-in the 0-10 soil layer is 340%,which is highly spatial variability.From the analysis of the soil profile,we can see that:A is divided into three stages,and the water and salt conditions are different in each stage and there is a certain regularity;in B area,the overall surface soil salinity is the highest,followed by the 10-20cm soil change is weak;In the upper part of the Ebinur Lake,the surface layer in the C area and the salt content in the 10-20cm soil are not much different;the salt content in the 10-20cm and 20-40cm layers of the soil around the D area is higher than that in the surface layer;the salt content in the surface layer of the E area is higher than 10-20cm And 20-40cm soil layer.2.The analysis of spectral curves with different salt content shows that the spectral characteristics of the soil are basically the same in terms of form and change trend.Analysis of Na+ positively correlated with soil spectrum at 600-1400 nm,1450-1750 nm,2000-2100 nm,and 2200-2250 nm.Based on the spectral index modeling results of the band combination,the R2 values of the logarithm and power function regression models of Na+,Cl-,and SO42-are 0.7198,0.5365,and 0.7129,respectively.The R2 values of the linear regression and power function regression models of Cl-,SO42-and Ca2-are 0.9479,0.6438 and 0.5696,respectively.The best model validation of the three exponential spectral parameters(RSI,NDSI,and DI)is that the SO42-fitting coefficients are R2=0.6461,R2=0.5631,and R2=0.5384,respectively.3.Analyzing the correlation between the multi-spectral bands and the salinity factors shows that the correlation between SO42-and Ca2+ in Bandl-Band7 bands is decreasing.The verification results show that the accuracy R2 of the SO42-and Bandl,Band2,Band3,Band4 and Band5 band models are:0.6777,0.6744,0.5506,0.6296 and 0.5295,respectively,and the verification results are good.Secondly,the correlation between the multispectral band combination index and the salinity factor is established,and the effect is not good.This article will not repeat here.4.In this paper,an index is used to fuse multispectral and hyperspectral data together to establish a comprehensive index of inversion salt factors,and to analyze the correlation between the composite index and salinity factor,and finally to establish a comprehensive index model.The results show that the NDSI model validation results are good,in which the Ca2+ verification accuracy is the best,R2 = 0.8596,followed by Mg2+ verification accuracy R2 = 0.4629.The model established by the NDD index,in which the model established by SO42-has the best verification effect,with an accuracy of R2=0.6519.SI composite spectral index and SO42-verification result R2=0.7648,Ca2+ model verification result R2=0.457.