Study On Spatial Variability Of Soil Physico-chemical Properties In The Upper Reaches Of Heihe Based On 3S Technology

Heihe River basin is the second largest inland river basin in northwest China and the largest of continental river basin, originated in the southern Qilian mountain, ranges between 98 ° to 101 ° 30 ’E, 38 °- 42 ° N. The main ecological environment problems of Heihe River upstream are grassland degradation, glacier area reducing, terminus Swings back. Monitoring spatial and temporal variations on hydrological and soil condition of Heihe River basin can help to provide certain technical reference to improve the local ecological environment. Through the sampling under watershed scale and field scale in the upper reaches of Heihe River, two dimension research in spatial variability of soil physical and chemical properties have finished, using the classical statistical methods, Geo-statistical methods, fractal method to analysis the distribution patterns of soil physical and chemical properties under different scales. Then, it can help to provide reference for regional agricultural production and environmental protection. Use partial least-squares regression method to extract the soil quality evaluation index under the farmland scale can further supplement farmland soil quality evaluation system. In addition, finding optimal modeling methods to establish the optimal spectral prediction model of soil organic matter among different regression methods can offer more convenient method to the research about soil organic matter in large area. Based on the above, main results are as follows:(1) The study of soil organic matter(SOC) and electrical conductivity(EC) in the upper reaches of Heihe River use classical statistical methods and statistical methods, at the same time, Kriging interpolation prediction analysis was done in the upper reaches of Heihe River. Analysis results show that the spatial variation of soil organic matter, electric conductivity in the upper reaches of Heihe River is caused by itself and the human factors. At the same time, they have the same degree of variation. However, EC has a bigger range of semi-variance than SOC, it suggests that SOC in the study area has higher interpolation accuracy than EC. This study can provide reference for SOC prediction research in Heihe River upstream.(2) Considering the distribution of SOC and EC which influenced by different vegetation types, geology and geomorphology factors in the upstream of the Heihe River. Comprehensive analysis of Kriging interpolation prediction map for SOC and EC, the distribution of them are roughly on the contrary, there is an intuitive negative correlation between them. High coverage grassland, forest farm, farming area are the great importance influence factors of vegetation which influenced the SOC and EC value. Except that, rivers and elevation are also the important factors.(3)Classical statistics, Geo-statistics, Multifractal method, and joint multifractal method were employed to research spatially distribution rules of soil saturated hydraulic conductivity(Ks), electrical conductivity(EC), moisture content(θ) and particle composition(SA, CL, SI) under farmland scale in the upper reaches of Heihe River. The results show that: from classical statistics and multifractal analysis, variability of soil properties abided by the following relation: Ks>EC>θ>SA>CL>SI; Clay and water content in the 0-15 cm deep layer and SI content in the 15-30 cm layer are dominant properties at spatial random distribution when applying geostatistics. Their spatial correlation is weak, while other properties behave better; Correlation between EC and particle distribution or water content is inconsistent at different horizons. During cropland soil hydrological analysis is applied, analyzing independently each property and carrying out field sampling is indispensable for precise simulation.(4) Soil water content(θ), electrical conductivity(EC), water drop penetration time(WDPT), saturated hydraulic conductivity(Ks) and clay(CL), silt(SI), sand(SA) mass fraction of 64 soil samples which collected under farmland scale in the upper reaches of Heihe River were analyzed. Every indicator was assumed to be one kind of soil quality evaluation object for partial least squares regression(PLSR) analysis with the residual soil indicators. The results showed that: Compare with the PLSR method and PCA method, the evaluation index of soil quality which built by PLSR method is better suitable; During the process of building evaluation index for soil water-holding capacity and the degree of salt deposition, CL, SA, SI and WDPT are sensitive indexes, among them, we should take WDPT into the evaluation system of soil quality; PLSR analysis between EC and other indicators pointed out that the degree of salt deposition in surface soil is largely influenced by soil water-holding ability, and the degree of salt deposition is not the main factors to decide soil water-holding capacity.(5) A total of 225 soil samples under watershed scale were collected in an extensive region of the upper reaches of Heihe basin. SOC and spectral reflectance were measured. All the samples were divided into 2 subsets- a modeling subset(180 samples) and a validation subset(45 samples). Six indices were obtained through transformation of soil spectral reflectance(R), continuum-removal(CR), reciprocal(REC), logarithm of reciprocal(LR), first-order differential(FDR) and Kubelka-Munck transformation coefficient(K-M). To build the mathematical model of SOC with 12 spectral indices, two methods, i.e., stepwise linear regression and partial least-square regression were used based on the modeling subset, respectively; the validation subset is used for model evaluation. The results indicated that: Through the comparison of related parameters in the process of modeling and verification, in the six kinds of spectral correlation index of SOC when modeling, based on the spectral variables- soil spectral reflectance of inverse logarithms(LR) SOC models are based on the optimal model is set up, can very good prediction heihe river upstream SOC.Further, based on the variable LR, PLSR method is adopted to establish the prediction accuracy of the SLR method to establish model of SOC is better. Therefore, based on PLSR method of SOC- LR model in many model was chosen as the best model of heihe river upstream soil SOC estimation.