| China is a populous country,food supply and demand is tensive,the improvement of farmland fertility is an important way to promote food production,organic matter and total nitrogen are important parts of soil,which can reflect the level of soil fertility,and closely relate with the yield of crops,therefore,the study on spatial variability of soil organic matter and total nitrogen is the basis of rational fertilization and soil fertility improvement,and it is important to realize the sustainable development of farmland soil.This dissertation uses the theory and method of classical statistics and geostatistics to study the spatial variability of topsoil organic matter of Xinzheng;simultaneously change sampling amplitude and size,and represent the change of spatial scale,further explore the scale effect and nested method of multi-scale information of spatial variability of soil organic matter at county,township and village scale;only change the sampling amplitude,study the spatial variability of soil total nitrogen at township and administrative village scale,and the relationship between total nitrogen variability and sample size;on the basis of the study of spatial variability of soil organic matter and total nitrogen in 2D space,explore the simulation of 3D spatial distribution of soil organic matter and total nitrogen,with the purpose of providing the theoretical and practical basis for the precision management of regional soil nutrients,the design of soil sampling,sustainable utilization of soil resources and sustainable development of food production,using Xinzheng city,a typical agricultural city,as a case study.The main research results are as follows:(1)Spatial variability of soil organic matterThe geostatistics is used to study the spatial structure of soil organic matter,the results show that the topsoil organic matter of Xinzheng has moderate spatial variability(coefficient of variation is 31.71%)and spatial correlation(nugget coefficient is 63.77%),the spatial variability is caused by the structural factor(parent material,topography and climate,etc.)and random factor(tillage,fertilization,irrigation,etc.).Through the analysis of variance,soil type,soil texture and groundwater depth are the important factors influencing the spatial distribution of soil organic matter.(2)Spatial scale effect and multi-scale nesting of soil organic matterThe geostatistics is used to explore the spatial variability structure of soil organic matter at county,township and village scale,the results show that the spatial structure of soil organic matter is dependent on the sampling scale,the single scale can't deeply analyze the spatial variability.Based on the nested structure theory,the nested structure of organic matter of different directions of county and township scales under the anisotropic condition and the spatial structure of organic matter of the village scale under the isotropy condition are nested to construct the multi-scale nested model of organic matter.Through I value test,the spatial interpolation accuracy of multi-scale nested model method is higher than the ordinary kriging method at a single scale(I value 0.48 < 3.31).The multi-scale nested model method is used to estimate the soil organic matter content,the spatial distribution which is fitted by the multi-scale nested model agrees with the spatial distribution of the county scale that is fitted by the ordinary kriging interpolation,but the expression of spatial variability of the township and village scales are more detailed through the multi-scale nested model,which can effectively reveal the spatial variability of soil organic matter.(3)Spatial variability of soil total nitrogen at different scalesThe classical statistics method is used to explore the spatial variability of soil total nitrogen at township and administrative village scale,results show that,the variability of soil total nitrogen at the township scale(coefficient of variation is 13.95% ~ 24.05%)is lower than that at the administrative village scale(coefficient of variation is 2.63% ~ 26.25%),the reliability of sampling according to the township scale is higher than the administrative village scale.To quantify the relationship between total nitrogen variability and sample size by using multiple linear regression models,the results show that the sampling size for the large scale is less than the small scale,from the measurement of soil total nitrogen in county territory,the reasonable sampling number is more than 1000 using the township as the basic statistical unit,and the sampling number is more than 2000 using the administrative village as the basic statistical unit.(4)3D spatial distribution of soil organic matter and total nitrogenThe 3D Kriging interpolation method and inverse distance weighted interpolation method are used to simulate the spatial distribution of soil organic matter and total nitrogen in different depth of soil layer(0 ~ 20 cm,20 ~ 40 cm and 40 ~ 60 cm),3D graphics can intuitively reveal the 3D spatial distribution of soil nutrients,show the relationship between the spatial distribution of soil nutrients and the natural geographical environment,and obtain the spatial distribution information of soil organic matter and total nitrogen in the horizontal and vertical directions,directly show the distribution of the nutrient content in soil arbitrary section.The cross validation method is used to evaluate the prediction accuracy of Kriging and inverse distance weighted interpolation models,the results show that the prediction accuracy of interpolation results of organic matter obtained through 3D Kriging method is higher than that obtained through inverse distance weighting method(RMSE 2.3078 < 2.4720,ME 0.0295 < 0.1273),and the result of total nitrogen is the same,3D Kriging interpolation method can more veritably reflect the 3D spatial distribution of soil nutrients. |
PDF Full Text Request