东北典型黑土区土壤养分空间分异影响因素分析

Spatial Variability Characteristics of Soil Nutrient Contents in the Black Soil Region of Northeastern China

结合常规统计方法与GIS空间分析方法,采用经典回归模型和空间滞后模型,定量研究了榆树市土壤养分空间变异的主要影响因素。研究表明,土壤中碱解氮元素的空间变异主要受河流、居民点、小路3个因子影响;土壤中速效磷空间变异的影响因子是到居民点的距离、到小路的距离,说明速效磷的空间变异主要受人文因素的影响;土壤中速效钾空间变异影响因素相对复杂,水文因素、人文因素、地貌因素对速效钾的空间分布均有影响,其中地貌因素的影响最为显著;土壤中有机质空间变异的影响因子包括湖泊、河流、大路、居民点、坡度,和速效钾空间变异的影响因子相同,但各个因素的影响相对均衡。比较两种模型发现,在土壤养分存在明显自相关的情况下,空间滞后模型的模拟效果明显优于经典回归模型。

Combining the methods of Statistical Analysis and GIS spatial analysis, a quantitative study on the spatial variability characteristics of soil nutrient contents are made by taking the standard muhiple line ar regression model and spatial log regression model in Yushu city. The results showed that the spatial variability characteristics of nitrogen was mainly influenced by the distance of river, settlements and path. The influencing factors of the spatial variability characteristics of phosphorus was the distance of settlements and path, it is suggested that the spatial variability characteristics of phosphorus was mainly influenced by the humanity factors; the spatial variability characteristics of potassium is a very complicated process and might be influenced by multiple factors such as hydrological effects, humanity factors and to pography, among these factors topography is the most striking; the influencing factors of the spatial variability characteristics of soil organic matter is same as potassium, but the influence of the factors is relatively balanced. Compared with the models, we can see that the spatial log regression model is superior to the classical regression model when the soil nutrients have remarkably spatial autocorrelation.