Spatial Variability And Uncertainty Assessment Of Soil Properties In Longtan Watershed In Shucheng County Of Anhui Province

The soil is the basic information for people's agricultural production activities; it also is a natural history of complex whose form and evolution are very complex. The study of the spatial variability of soil properties can not only reveal the distribution of soil properties and their variability, but also better to explain the impact of natural and human activities on spatial variability of soil properties, so as to provide a scientific basis for guiding human agricultural production, improving soil nutrient availability, precision fertilization as well as the implementation of regional management.The Longtan watershed in Shucheng county of Anhui province is taken as the study area in this paper. 117 surface soil samples are collected and 5 indexes of soil physical and chemical properties are analysed in laboratory such as soil particles, organic matter(OM), total nutrients(TN), total phosphorus(TP) and available phosphorus(AP). Then the spatial variability properties of surface soil are analysed and the spatial distribution pattern of soil properties are revealed by using the geostatistics, GIS spatial analysis and sequential Gaussian simulation method. At last the paper evaluates the uncertainty of property index which exceeds a given threshold. The results show that:1. The variation coefficients of soil properties varies from 23.45% to 85.92%, of which the highest coefficient of variation of available phosphorus is 85.92%, relatively strong variability; the lowest coefficient of variation of sand is only 23.45%, indicating relatively low varibaility of sand. The indicators of soil physical and chemical properties of the coefficient of variation is in the following order: soil sand, soil silt, soil clay, organic matter, total nutrients, total phosphorus and available phosphorus, and the variance is in the rang of 10% to 100%, both belongs to moderate variability.2. The spatial variability characteristic of several physical and chemical properties of soil in the basin shows a good spatial structure, random variation in their systems is the proportion of 10.34% to 24.80% in total variance, of which only the soil clay of the C0/(C0+C) is greater than 25%, a medium degree of spatial autocorrelation, and the remaining indicators of physical and chemical proporition of the C0/(C0+C) is less thane 25%, belongs to the strong spatial autocorrelation. Clay, TP, OM and AP have a large-scale spatial autocorrelation and their autocorrelation distances are all over 690m, silt, sand and TN have a relatively small-scale spatial autocorrelation and their autocorrelation distances are all below 550m, of which the autocorrelation distances of TN is the smallest one, only 438.87m, the autocorrelation distances of clay is the biggest, 836.16m. The autocorrelation distances of soil physical and chemical properties is in the following order: TN, silt, sand, OM, TP, AP and clay. The isotropic of all soil physical and chemical properties in different direction is not obvious, only in some individual properties within a limited rang of steps can show demonstrate isotropic.3. Taking advantage of ordinary kriging interpolation to interpolate soil properties, the maps of their spatial distribution pattern are obtained. The interpolation result shows more intuitive illustration of the properities of the reaction the content of the distribution of indicators, also can see that soil physical and chemical properties index in the study area have a high degree of spatial heterogeneity. Among them, the spatial distribution pattern of clay and silt is similar, the contents of them are both found that lower in central, higher in north, east and southwest. While the spatial distribution pattern of sand is opposite to clay and silt, its content is higher in central, lower in the north and southwest. TN and AP have a similar spatial distribution, their contents are both higher in the south of the terrain relative to the horizon, lower in northern and western regions with relatively high terriain. The content of OM concentrates between 6 g/kg and 55 g/kg, with an area of 7435100 m~2, and that of AP concentrates between 11 mg/kg and 140 mg/kg, with an area of 6186300 m~2.4. Using sequential Gaussian simulation to evaluate the uncertainty of clay, silt, sand, TN, TP, OM and AP with 6%, 40%, 50%, 1.3 g/kg, 0.5 g/kg, 45 g/kg and 100mg/kg as the threshold, by 1000 times consistent Gaussian simulation, the purple white great-probability area of all properties which exceed the threshold are significantly smaller than the area in the simulation map. The probability of these high-probability areas exceed the threshold is great, basicly more than 80%, indicating that these high-probability area is more easy to determine the rang of the area which exceed the threshold value. Other low-probability because its probability which exceed the threshold is low, can not determine the rang of the area which exceed the given threshold value.