Study On Spatial Variation And Establishment Model Of Soil Properties In Regional

Spatial variability of soil property is one of the main research directions in soil. Based on geostatistics and semi-variogram,the analysis of spatial structural of soil properties was common.This study had selected a typical plot of HongHuaTao Township as research object,to study the spatial variability of soil properties.Meanwhile,several soil chemical parameters on sampling points were chose,in order to discuss the spatial distribution,using the regression analysis and kriging-interpolating.For the forecast of soil property in this area,the regional soil models were established.The achievements of this study can be summarized as follows:(1).Explanation of distributional characteristics for soil properties in study areas. The range of soil pH value was 4.0 to 7.9,and changes has mini,which distributions relatively stable in the research area.The result of test of normal frequency distribution, showed available P(AP) and available K(AK) were log-normal distribution,the others were normal distribution.In addition,the coefficient of variability of alkali-hydrolyzable N(AN),total organic matter content(OMC),pH value and cation exchange capacity (CEC) were lowest.Consequently,available P(AP) and available K(AK) had great variability,while rest had moderate variability.But pH value and cation exchange capacity(CEC) had low variability with CVs was 0.13,0.15,respectively.(2).Research and evaluation of spatial autocorrelation of regional soil properties. The semi-variograms of alkali-hydrolyzable N(AN) and total organic matter content (OMC) were described by exponential model.The best of definitive coefficients(R²) was 0.945;belong to pH value,which was described by sphere model.The other variables had lineal model.The value of nugget(C₀)/sill(C₀+C) for AP,AK were higher,and had weak spatial variability,compared to OMC and AN was decided by random factors and structural factors with moderate spatial variability.Soil pH value had great spatial variability with decided by structural factors and CEC had pure nugget effect model.(3).Analysis in anisotropic of semivariance for soil attributes.The semi-variograms of AN and CEC in each directions had lower different,were neatly same,it showed their weak anisotropic,actual was isotropic.The semi-variograms of soil pH value had different between 90°and 45°directions,while AK with though the log-transformation only had different between 0°and 45°directions.So this different was omitted and AK showed isotropic.AP had different between 90°and 45°directions,yet OMC had different between 0°and 90°.(4).Revelation of correlationship in soil attributes indexes and establishment regression model.The correlation matrix of soil properties displayed a significant correlations with AN and OMC,meanwhile AN had negative significant correlation to CEC with the coefficient correlation were 0.798 and -0.140,respectively.AP had significant correlations with AK,pH value and CEC.AK had significant correlations with AP,additionally,had significant correlations with pH value,OMC and CEC.In addition, pH value had no correlation with OMC,but had correlation with CEC.Based on these correlations,the regional regression analysis model and kirging-interpolating model were established.