淮河流域上消化道肿瘤与环境污染的模型分析
收稿日期: 2012-05-13
修回日期: 2012-07-12
网络出版日期: 2012-08-22
基金资助
"十一五"科技支撑项目"淮河流域水污染与肿瘤的相关性评估研究"(2006BAI19B03)。
Model Analysis of Upper Digestive Tract Cancer and Environmental Pollution in Huaihe River Watershed
Received date: 2012-05-13
Revised date: 2012-07-12
Online published: 2012-08-22
自20世纪70年代后期以来,淮河流域不断遭受工业点源污染和其他面源污染,媒体也陆续报道了淮河流域"癌症村"的出现。本文探讨了淮河流域14个监测县5810个行政村的消化道肿瘤与环境因子之间的空间分布规律。作者从流域和行政区划等多维空间角度出发,通过全局的最小二乘法线性回归和稳健回归对环境因子进行筛选分析,以局部地理加权回归方法探测各类环境因子,在不同地区对贝叶斯调整的上消化道肿瘤死亡率的影响程度,建立了消化道肿瘤死亡的风险评估模型,其中,包括地表水水质等级、浅层地下水质量分级、河网密度、土壤多环芳烃含量分级、化肥施用量和经济密度等6类环境危险因素。根据局部回归模型中各监测点环境因子的回归系数和统计学检验结果,提取出当地主要的环境影响因素。从14个监测县区总体上看,地表水水质等级和GDP与肿瘤呈负相关,其他环境因子均与肿瘤死亡存在正相关。但从局部角度看,不同地区环境影响因子种类和影响强度有较大差别。其中淮河流域江苏段以化肥施用量、土壤多环芳烃含量、GDP和河网密度为主要影响因子,安徽段以土壤多环芳烃含量和化肥为主,河南段主要是以地下水质量分级、河网密度和化肥为主,同时河南沈丘县地表水水质等级对当地影响较大。山东段虽然也探测出来部分环境危险因子的存在,但没有发现其与肿瘤死亡的关联关系,尚需进一步深化研究。
戚晓鹏, 计伟, 任红艳, 郭岩, 周脉耕, 杨功焕, 庄大方* . 淮河流域上消化道肿瘤与环境污染的模型分析[J]. 地球信息科学学报, 2012 , 14(4) : 432 -441 . DOI: 10.3724/SP.J.1047.2012.00432
The Huaihe River watershed has been suffering from industrial point pollution and other non-point pollution since 1970s. The media has reported the emergence of ‘cancer village'. Spatial distribution pattern on upper digestive tract cancer and environmental factors was studied in 14 pilot counties including 5810 villages in Huaihe River watershed. From the multiple perspectives such as watershed and jurisdictional areas, Ordinary Least Squares (OLS) and Robust regression model were used for environmental factors selection. Robust regression model can detect abnormal value and put the weight for each of them. Surface water, phreatic water, river density, soil PAHs, fertilizer, population density and economical density (GDP) were imported into the model. Geographically Weighted Regression (GWR) model was used to locally detect the impact of different environmental factors on Empirical Bayes smoothed upper digestive tract cancer mortality. Through the correlative analysis of them, the risk assessment model of upper digestive tract cancer mortality was developed. Based on the regression coefficient of each environmental factor and statistical test in each pilot, the local main environmental factors were extracted. Even six factors were imported into the global regression model, the impact of each factor was different in distinct areas. In general, all the selected environmental factors show the positive correlation with the upper digestive tract cancer mortality except the surface water quality level and GDP. However the pollution type was diverse in different area based on the regression coefficient of each factor. The main findings were listed as follow: fertilizer amount, soil PAHs, GDP and river density are the main factors in Jiangsu segment in Huaihe River watershed; main factors in Anhui segment included soil PAHs and fertilizer amount; main factors in Henan segment included phreatic water, river density and fertilizer amount, With Shenqiu County, as one of the pilots in Henan, showed the strong positive correlation between surface water and cancer mortality; some kind of environmental risk factors were also detected in Shandong segment, but the result showed no correlation between these risk factors and cancer mortality, which needed to be studied any further.
[1] 陈竺.全国第三次死因回顾抽样调查报告[M]. 北京: 中国协和医科大学出版社, 2008.
[2] 李璇. 淮河污染及治污困境分析[J]. 湖北社会科学, 2009(5):3.
[3] Goovaerts P. Geostatistical analysis of disease data: estimation of cancer mortality risk from empirical frequencies using Poisson kriging[J]. International Journal of Health Geographics, 2005,4(31):1-33.
[4] 冯昕,杜世宏,舒红. 空间权重矩阵对空间自相关的影响分析——以我国肾综合征出血热疾病为例[J]. 武汉大学学报(信息科学版), 2011,36(12):1410-1413.
[5] 刘旭华,王劲峰. 空间权重矩阵的生成方法分析与实验[J]. 地球信息科学,2002,4(2):7.
[6] 王鑫,梁念,安海蓉. 淮河污染浅析[J]. 中国环境监测, 2006,22(1):3.
[7] 张宗祜,李烈荣.中国地下水资源与环境图集(M).北京:中国地图出版社,2004.
[8] 周洪建,史培军,王静爱,等. 近30年来深圳河网变化及其生态效应分析[J]. 地理学报, 2008,63(9):969-980.
[9] 国家环境保护总局.1997年中国环境状况公报:耕地,2002. http:www.china.com.cn/chinese/zhuanti/hjgb/1007380.htm.
[10] 徐新良,刘纪远,庄大方. 国家尺度土地利用/覆被变化遥感监测方法[J]. 安徽农业科学,2012,40(4):2365-2369.
[11] Yue T X, Wang Y A, Liu J Y, et al. Surface modelling of human population distribution in China[J]. Ecological Modelling, 2005,181(4):461-478.
[12] Glues Forget D, Lebel J. An ecosystem approach to humman health[J]. International Journal of Occupational and Environmental Health, 2001,7(2):1-38.
[13] Tukiendorf A. An ecological analysis of leukemia incidence around the highest 137Cs concentration in Poland[J]. Cancer Causes Control, 2001,12(7):653-659.
[14] Wang J F, Li X H, Christakos G, et al. Geographical detectors-based health risk assessment and its application in the neural tube defects study of the Heshun Region, China[J]. International Journal of Geographical Information Science, 2010, 24(1):107-127.
[15] 孙振球,徐勇勇. 医学统计学[M]. 北京: 人民卫生出版社, 2002.
[16] 冯国双,罗凤基. 医学案例统计分析与SAS应用[M]. 北京: 北京大学医学出版社, 2011.
[17] 程娟,万建平. 空间统计在公共卫生事业中的应用[J]. 统计与决策, 2009,13(13):2.
[18] Fotheringham A S, Brunsdon C M C. Geographically weighted regression:a natural evolution of the expansion method for spatial data analysis[J]. Environment and Planning A, 1998,30(11):1905-1927.
[19] 覃文忠. 地理加权回归基本理论与应用研究. 上海:同济大学, 2007.
[20] 邢可霞,郭怀成,孙延枫,等. 流域非点源污染模拟研究——以滇池流域为例[J]. 地理研究, 2005,24(4):549-558.
[21] 杨功焕.中国人群死亡及其危险因素:流行水平、趋势和分布的主要发现[J]. 医学与哲学,人文社会医学版, 2007,28(6):1-5.
/
〈 | 〉 |