Study On Spatial Distribution Of Soil Nutrients In The Peri-urban Zone

The spatial variability of soil characteristics is widespread.The regional distribution with the continuous data of soil nutrient content is the basis of the space analysis.Due to restrictions of the variety conditions in practice,we can only laid sampling points in accordance with the representation and economic principles,leading to a high discreteness of the sampling data.Interpolating the sampling points is the effective means to obtain relatively continuous nutrient data in the study area.As for the region with the complex land-use pattern and the higher degree of fragmentation,if directly used Kriging interpolation, resulted in their relatively low accuracy,did not match with the actual situation.Therefore,this article put the Peri-urban of Shijiazhuang as the research object,on the basis of using the BP neural network model to forecast soil nutrients in the spatial interpolation,combined with the GIS technology and the geostatistics principle,studied the spatial variability of soil nutrients,which contained the organic matter,the total nitrogen,the available phosphorus and the available potassium in the soil surface layer(0-20cm) of the Peri-urban of Shijiazhuang,explored a new statistical analysis method in a regional with relatively small the number of sampling points,complex land-use type,high landscape heterogeneity.The purpose is to understand the spatial distribution of soil nutrients in the entire study area more accurately and intuitively.The results showed that:BP neural network models combined with statistical method was applied to analyze the spatial features of soil nutrient in the areas of Shijiazhuang Peri-urban zone,which has been divided.The results of our description prove that BP modeling can not only enable us to grasp intelligently the mapping relation between spatial position and soil nutrients,but can also help to predict robustly the soil nutrient contents in every spatial interpolating dot.Soil nutrient data in this study included organic matter,total N,available P and K.The results showed that the semi-variograms of soil nutrients were best described by spherical model.Shijiazhuang city marginal zone of soil organic matter,total nitrogen,phosphorus, potassium middle variation coefficients of variation were 18.24%,15.94%,22.14%,available K is a weak variation of coefficient of variation was 7.42%.After the forecast of soil organic matter,total nitrogen,available P and K variogram of the theoretical models are in line with the spherical model,and its coefficient of determination R~2 were 0.967,0.792,0.927,0.942. Organic matter,available P and K has a strong spatial correlation;block-based ratios were 0.12%,15.10%and 9.63%,total soil nitrogen associated with the medium.Block-based ratio of 25.64%,from farming,fertilizer,irrigation,field management level of human factors to some extent undermined the relevance of nitrogen space.Four kinds of available nutrient have a strong pattern of spatial autocorrelation,but their spatial autocorrelation changes with significant differences in scale.Organic matter and nitrogen-way closer to the change were 502.6m and 458.6m,available P and K variation smaller scale,were 277.0m and 385.1m.The conclusions were obtained from the analysis:the four soil nutrients showed the patchy distribution obviously,the various high-value of the soil nutrient distribution is the west and south of Shijiazhuang City.The content of the four soil nutrient showed that the decreasing trended from west to east,from south to north.Through the establishment of regression model,the article introduced two indexes,which are landscape patterns of land and the comprehensive scores of soil nutrients,and studied the relations between the contents of soil nutrients and land use pattern.The results showed that:the fitting model between land landscape pattern index and the comprehensive scores ofsoil nutrients is y=1.014×exp(x/1.794)+2152,coefficient of determination R~2=0.657.There is an obvious trend of the exponential growth.The spatial variation of soil nutrient is closely related to the landscape patterns of land use.The higher Landscape Pattern index,the more complex types of land use is,correspondingly,the content of soil nutrient is high,and vice versa.