文献类型：期刊文献

中文题名：基于GIS和统计模型的黄土丘陵沟壑区土壤水分插值方法

英文题名：The Soil Moisture Interpolation Method Based on GIS and Statistical Models in Loess Plateau Region

作者：姚雪玲[1,2] 傅伯杰[1] 吕一河[1] 孙飞翔[1] 郭秀江[3]

第一作者：姚雪玲

机构：[1]中国科学院生态环境研究中心城市与区域生态国家重点实验室;[2]中国林业科学研究院荒漠化研究所;[3]中国石油天然气股份有限公司华北油田分公司计量中心站

年份：2013

卷号：27

期号：6

起止页码：93-96

中文期刊名：水土保持学报

外文期刊名：Journal of Soil and Water Conservation

收录：CSTPCD;;北大核心:【北大核心2011】；CSCD:【CSCD2013_2014】；

基金：林业公益性行业科研专项项目(201004058);国家"十二五"科技支撑项目(2012BAD16B01)

语种：中文

中文关键词：回归克里格;黄土丘陵沟壑区;土壤水分;插值

外文关键词：regression kriging; Loess Plateau; soil moisture;interpolation

分类号：S127;S152.7

摘要：以黄土丘陵沟壑区羊圈沟小流域的土壤水分插值为例,介绍了基于ArcGIS和统计学软件的回归克里格模型插值方法,其中详细介绍了基于虚拟变量技术将分类变量引入模型的过程;同时,将回归克里格模型、线性回归模型和普通克里格模型的插值精度进行对比。结果表明,在环境条件较为复杂,通过采样点难以反映目标变量的整体空间特征,而目标变量与环境因子有显著相关性的情况下,回归克里格具有较高的插值精度,是更为理想的插值模型。
This study introduced the Regression-kriging operation method based on ArcGIS and statistical models through a case study of soil moisture interpolation in Loess Plateau of China. Particularly, a categori- cal variable was introduced to the model basing on dummy variables technique and the operation method was described in detail. Regression-kriging is a hybrid interpolation technique, which combines kriging and linear regression techniques, using not only the target variable spatial autocorrelation but also the numerical or cat- egorical environmental factors as auxiliary variables, significantly improved the interpolation accuracy. In this paper, the performance of Regression-kriging was compared with linear regression and ordinary kriging. The results showed that when spatial structure could not be well captured by point-based observations, or a strong relationship exists between target variable and auxiliary variables, Regression-kriging is superior to the other two methods, yielding more detailed results and higher accuracy of prediction.