论文部分内容阅读
许多水文、生态环境和土地管理规划的决策当中,都迫切需要空间连续的土壤性质信息来提高其建模和决策的精确性和可靠性。依靠传统土壤调查技术获得的土壤图则无法满足这种需求。本研究以浙江省龙游县研究区为例,利用第二次土壤普查的土壤性质数据,生成土壤性质-环境因子空间数据库。运用决策树建模方法将土壤性质含量与一些易于广泛观测的景观属性,包括地形、地质、土地利用和遥感影像建立联系,从而将有关土壤性质含量分布的知识转入一种清楚的、定量的、与环境因子相关联的规则系统中,并以此来预测研究区土壤性质的连续空间分布。研究结果表明,所建立的决策树模型可以解释75～81%的土壤性质空间变异。
In many hydrological, ecological and land-management planning decisions, there is an urgent need for spatial continuous information on soil properties to improve the accuracy and reliability of their modeling and decision-making. Soil maps based on traditional soil investigation techniques do not meet this need. In this study, Longyou County, Zhejiang Province as an example, the use of the second soil survey soil properties data to generate soil properties - spatial database of environmental factors. Using decision tree modeling methods to relate soil properties to landscape attributes that are readily and widely observed, including topography, geology, land use, and remote sensing images, the knowledge about the distribution of soil properties is transferred to a clear, quantitative , And the environmental factors associated with the rules of the system, and in order to predict the continuous spatial distribution of soil properties in the study area. The results show that the established decision tree model can explain 75 ~ 81% of the spatial variability of soil properties.