基于GWR模型的于田绿洲土壤表层盐分空间分异及其影响因子

发布时间：2018-01-19 06:03

本文关键词： 土壤表层盐分地理加权回归空间分异　出处：《应用生态学报》2016年10期 　论文类型：期刊论文

【摘要】：以于田绿洲为研究靶区,利用24个采样点的土壤表层盐分数据,选取9个与土壤表层盐分密切相关的影响因子,结合空间自相关、传统回归分析和地理加权回归模型,分析表土盐分的空间分布特征及其影响因子的空间分异.结果表明:于田绿洲表土盐分在空间上并非随机分布,而是存在较强的空间依赖关系,空间自相关指数为0.479.地下水矿化度、地下水埋深、高程和温度是影响干旱区平原绿洲表土积盐的主要因子,这些因子具有空间异质性,选取的9个环境变量中除土壤p H值外,其他变量对表土盐分的影响强度均存在显著的空间分异.GWR模型对存在空间非平稳性数据的解释能力和估计精度都优于OLS模型,而且在模型估计参数的可视化上具有明显优势.
[Abstract]:Taking Yutian oasis as the research target, using the soil surface salinity data of 24 sampling sites, 9 factors closely related to soil surface salinity were selected and combined with spatial autocorrelation. Traditional regression analysis and geographical weighted regression model were used to analyze the spatial distribution characteristics of surface soil salt and the spatial differentiation of influencing factors. The results showed that the surface soil salt distribution in Yutian oasis was not random in space. But there is a strong spatial dependence, spatial autocorrelation index is 0.479.groundwater salinity, groundwater depth, height and temperature are the main factors affecting salt accumulation in oasis surface soil in arid area. These factors have spatial heterogeneity, except soil pH value of 9 environmental variables selected. The effect of other variables on the salinity of surface soil has significant spatial differentiation. GWR model has better interpretation ability and estimation accuracy than OLS model for non-stationary spatial data. Moreover, it has obvious advantages in the visualization of model estimation parameters.
【作者单位】：新疆大学绿洲生态教育部重点实验室;新疆大学资源与环境科学学院;新疆大学旅游学院;
【基金】：国家自然科学基金项目(U1138303)资助~~
【分类号】：S156.4
【正文快照】： 本文由国家自然科学基金项目(U1138303)资助This work was sup-ported by the National Natural Science Fundation of China(U1138303).2016-03-14 Received,2016-07-20 Accepted.盐渍化是盐分不断向土壤表层聚积形成盐渍土的自然地质过程,是一定的气候、地形、水文地质和不

【相似文献】