基于空间地理加权回归分析的山东省居民饮用水水碘的影响因素分析  

Analysis of influencing factors on water iodine in Shandong Province based on geographically weighted regression

作　　者：高红旭[1] 张本政 尹玉岩[1] 蒋雯[2] 王晓明[2] 高杰 Gao Hongxu;Zhang Benzheng;Yin Yuyan;Jiang Wen;Wang Xiaoming;Gao Jie(Department of Endemic Fluorosis and Kaschin-Beck Disease,Shandong Institute for Prevention and Control of Endemic Disease,Jinan 250014,China;Department of Iodine Deficiency Disorders,Shandong Institute for Prevention and Control of Endemic Disease,Jinan 250014,China;Maternal and Child Health Development Research Center,Shandang Maternal and Child Health Hospital,Jinan 250014,China)

机构地区：[1]山东省地方病防治研究所地方性氟中毒与大骨节病防治研究科,济南250014 [2]山东省地方病防治研究所碘缺乏痛科,济南250014 [3]山东省妇幼保健院妇幼健康发展研究中心,济南250014

出　　处：《中华地方病学杂志》2018年第12期948-953,共6页Chinese Journal of Endemiology

基　　金：山东省医药卫生科技发展计划项目(2013WS0177).

摘　　要：目的探讨山东省居民饮用水水碘的影响因素。方法以山东省为研究现场,各县(市)水碘含量数据来源于山东省地方病防治研究所2008—2010年水碘普查资料。以水碘含量作为因变量,土壤类型、水文地质类型、地形地貌和距黄河主干道的距离为自变量进行回归分析。因变量的正态性检验和一般线性回归分析采用SAS 9.3软件;地理加权回归(GWR)分析采用SAM V4.0软件;相关电子地图的绘制采用ArcGIS 9.3软件。结果共收集108164份水碘含量数据。一般线性回归分析显示,构建的回归模型具有统计学意义(F=16.29,P<0.01),土壤类型被纳入模型,决定系数(R^2)=0.51。GWR分析显示,R^2=0.63,调整决定系数(R^2adj)=0.59,在考虑变量空间自相关性的前提下,模型的拟合优度优于传统一般线性回归模型。GWR分析显示,土壤类型和距黄河主干道的距离是影响山东省居民饮用水水碘分布的主要因素。土壤类型与水碘的空间变异性在山东省绝大部分地区内呈负相关,由西向东逐渐减弱,呈明显的地理梯度变异性;距黄河主干道的距离与水碘含量的空间变异性在山东省部分地区呈现负相关,而在一部分地区呈现出正相关,由西南向西北呈现出明显的地理梯度变异性。结论土壤类型和距黄河主干道的距离是山东省水碘含量空间分布的重要影响因素。Objective To analyze the influencing factors of water iodine in Shandong Province.Methods The county-based study set Shandong Province as a research site.The water iodine data of county (city)from 2008 to 2010 were obtained from Shandong Institute for Prevention and Control of Endemic Disease.Water iodine content was used as a dependent variable,and soil type,hydrogeological type,topography and distance to the Yellow River were analyzed as independent variables.Normality test and general linear regression analysis of the dependent variables were performed using SAS 9.3 software;geographically weighted regression (GWR)analysis was performed using SAM V4.0 software;related electronic maps were drawn using ArcGIS 9.3 software.Results A total of 108164 water iodine content data were collected.General linear regression analysis showed that the constructed regression model was statistically significant (F=16.29,P <0.01),and the soil type was included in the model with a determination coefficient (R^2)=0.51.GWR analysis showed that R^2=0.63 and the adjustive determination coefficient (R^2adj)=0.59.Considering the autocorrelation of the variable space,the model's goodness of fit was better than that of the traditional general linear regression model.GWR analysis showed that soil type and distance to the Yellow River were major factors related to water iodine in Shandong Province.There was a negative correlation between soil type and spatial variability of water iodine in most areas of Shandong Province,the correlation coefficients weakened gradually from west to east,indicating a geographic gradient variability.The correlation coefficients of distance to the Yellow River and spatial variability of water iodine were negative in some areas,while they were positive in other areas,indicating a clear geographical variability from southwest to northwest. Conclusion The soil type and the distance to Yellow River are important factors affecting the spatial distribution of water iodine in Shandong Province.

关 键 词：饮用水 碘 因素分析 统计学 地理加权回归 

分 类 号：R123.1[医药卫生—环境卫生学]