基于BP神经网络与GIS的巢湖水华危险性评价  

Hazard Evaluation of Chaohu Lake Water Bloom Based on BP and GIS

作　　者：何君[1] 金菊良[1] 黄诗峰[2] 许浒[3] 罗慧娟[2] 

机构地区：[1]合肥工业大学土木与水利工程学院,安徽合肥230009 [2]中国水利水电科学研究院遥感技术应用中心,北京100048 [3]安徽省水利科学研究院安徽省水利水资源重点实验室,安徽蚌埠233000

出　　处：《水电能源科学》2011年第5期34-36,151,共4页Water Resources and Power

基　　金：国家水体污染控制与治理科技重大专项基金资助项目(2009ZX07106-001);水利部公益性行业科研专项基金资助项目(200901077);国家自然科学基金资助项目(51079037);国家高技术研究发展计划基金资助项目(2009AA12Z141463)

摘　　要：选择溶解氧、化学需氧量、总磷、总氮、叶绿素浓度作为评价指标，建立了巢湖水华危险性评价分级标准。采用BP神经网络模型对指标进行了评价，得出巢湖12个水质监测站点的水华危险性等级，并借助GIS的空间信息展布功能，获得了巢湖2000-2007年的水华危险性区域分布及年际变化。结果表明，巢湖西部湖区的水华危险性程度最高，暴发水华的可能性最大，中部湖区的水华危险性程度较高，东部湖区的水华危险性程度一般；BP神经网络模型与GIS相结合的方法直观、简便，具有推广应用价值。DO, COD, TP, TN and chlorophyll concentration are chosen as evaluation indexes and the grading criterion of hazard evaluation in Chaohu Lake is established. BP neural network model is adopted to evaluate index and it obtains the hazard level of water bloom for 12 water quality monitoring stations distributed in Chaohu Lake. The regional distribution and annual change of water bloom hazard from 2000 to 2007 is gained with the aid of GIS spatial information analysis. The results show that the risk of water bloom in western area of Chaohu Lake is the most serious and it is most likely to cause water bloom, followed by middle part and eastern part in Chaohu Lake. The combination method of BP and GIS is simple and visualized, which can widely be applied to other disaster risk evaluation.

关 键 词：湖泊富营养化 水华 危险性评价 BP神经网络 GIS 地统计方法 巢湖 

分 类 号：X824[环境科学与工程—环境工程] X524