典型岩溶山地土壤厚度的空间插值模型应用研究

发布时间：2018-05-20 22:32

本文选题：岩溶山地 + 土壤厚度　；参考：《昆明理工大学》2015年硕士论文

【摘要】：我国是世界上遭受岩溶石漠化危害最严重的国家之一,云南省是我国遭受石漠化危害的八个省其中之一。云南省岩溶面积约为11.09万平方公里,为全省土地面积的28.93%。岩溶石漠化导致严重的水土流失,植被覆盖度降低,耕地面积不断减少等生态环境问题,严重威胁我国西南岩溶山地的生态环境安全。土壤厚度直接制约岩溶山地植被的生长状况,直观地反映岩溶山地的水土流失情况。通过人工方式可以改善、调整土壤质量,而土壤厚度的减少却不能在短时间内恢复。因此,对岩溶山地土壤厚度空间分布特征的探测与模型模拟研究是岩溶山地石漠化综合治理工程能否合理布局的重要依据与基础研究。论文以典型岩溶山地云南省文山州八宝镇为研究区。在分析研究区土地利用状况的基础上,确定研究对象为集中在研究区内主要分布的耕地、林地和裸地三大地类的土壤厚度特征。依据研究区各地类的分布情况、地类特点,分别选取3个具有不同主体类型代表性的样地进行研究。具体研究中,首先经野外土壤厚度调查,测得各样地的土壤厚度数据,并利用直方图和正态QQ图对各样地的土壤厚度数据进行正态分布检验。在此基础上,对各样地进行土壤厚度异质性分析,包括描述性统计分析、空间变异性和趋势分析。论文研究工作内容及主要成果包括：1)利用描述性统计分析方法,实现了对各样地的土壤厚度数据进行总体的统计分布分析；2)运用空间变异性分析,对各样地土壤厚度进行K-S非参数检验,在确定数据为正态分布的情况下,完成了对各样地进行土壤厚度数据的空间结构分析。3)综合运用趋势分析方法,主要揭示了研究区各样地土壤厚度空间分布的总体规律,反映其在空间区域上变化的主要特征,为在之后的空间插值建模中可以方便的剔除全局趋势。论文研究工作的重点与创新主要是体现在上述分析基础上,分别采用了反距离加权插值、局部多项式插值、全局多项式插值和径向基函数插值4种确定性插值方法,以及普通克里格插值、简单克里格插值、泛克里格插值和协同克里格4种地统计插值方法,较为深化的对各样地土壤厚度数据进行空间插值建模与模型模拟分析,并采用交叉检验和验证数据集检验对各样地土壤厚度空间插值结果进行精度检验及对比分析,得到能反映研究区中典型岩溶山特征的三种典型地类的最优空间插值方法。泛克里格插值为耕地土壤厚度插值模拟的最优方法,泛克里格插值方法为林地土壤厚度插值模拟的最优方法。径向基函数(平面样条函数)为裸地土壤厚度插值模拟的最优方法。
[Abstract]:China is one of the countries most seriously affected by karst rocky desertification in the world, and Yunnan Province is one of the eight provinces in China. The karst area of Yunnan Province is about 110900 square kilometers, which is 28.93% of the province's land area. Karst rocky desertification leads to serious soil and water loss, low vegetation coverage, decreasing cultivated land area, and other ecological environmental problems, which seriously threaten the ecological environment safety of karst mountains in southwest China. Soil thickness directly restricts the growth of vegetation in karst mountainous areas and directly reflects the soil and water loss in karst mountainous areas. The soil quality can be improved and adjusted by artificial means, but the reduction of soil thickness can not be restored in a short time. Therefore, the study on the spatial distribution of soil thickness in karst mountainous area is an important basis and basic research for the rational layout of the comprehensive control project of rocky desertification in karst mountainous area. This paper takes the typical karst mountain area of Wenshan, Yunnan Province as the research area. Based on the analysis of the land use status in the study area, the soil thickness characteristics of the three major land types, namely, cultivated land, woodland and bare land, which are mainly distributed in the study area, are determined. According to the distribution of the classes and the characteristics of the ground in the study area, three representative plots with different main types were selected for the study. In the specific study, the soil thickness data of each land were measured by field soil thickness investigation, and the normal distribution of soil thickness data was tested by histogram and normal QQ map. On this basis, soil thickness heterogeneity analysis, including descriptive statistical analysis, spatial variability and trend analysis were carried out. The contents and main results of this paper include: (1) the use of descriptive statistical analysis method to realize the overall statistical distribution analysis of soil thickness data in various lands. (2) the use of spatial variability analysis. K-S nonparametric test was carried out on the soil thickness of various plots. Under the condition that the data were normal distribution, the spatial structure analysis of soil thickness data was completed. 3) the comprehensive trend analysis method was used. This paper mainly reveals the general law of the spatial distribution of soil thickness in the study area, and reflects the main characteristics of its variation in the spatial region, which can conveniently eliminate the global trend in the later spatial interpolation modeling. The emphasis and innovation of this paper are mainly embodied in the above analysis. Four deterministic interpolation methods, namely inverse distance weighted interpolation, local polynomial interpolation, global polynomial interpolation and radial basis function interpolation, are adopted respectively. And four geostatistical interpolation methods, such as ordinary Kriging interpolation, simple Kriging interpolation, Pan Kriging interpolation and Cooperative Kriging interpolation, are used to model and simulate the soil thickness data. By using cross-test and validation data set test, the accuracy test and comparative analysis of spatial interpolation results of soil thickness in various plots were carried out, and three optimal spatial interpolation methods for three typical land types were obtained, which can reflect the characteristics of typical karst mountains in the study area. Pan Kriging interpolation is the best method to simulate the soil thickness of cultivated land, and the pan Kriging interpolation method is the best method to simulate the soil thickness of woodland. The radial basis function (plane spline function) is the best method for soil thickness interpolation.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S157

【参考文献】