自组织特征映射神经网络识别珠江口夏季水质空间格局  

Assessment of spatial trend of water quality in the Pearl River estuary by the self-organizing map

作　　者：杨志浩 吴梅林[2] 吴鹏[2,3] 

机构地区：[1]广东核力工程勘察院,广州510800 [2]中国科学院南海海洋研究所热带海洋环境国家重点实验室,广州510301 [3]国家海洋局南海环境监测中心,广州510300

出　　处：《生态科学》2015年第3期20-25,共6页Ecological Science

基　　金：国家自然科学基金项目(31270528;41206082);热带海洋环境国家重点实验室(中国科学院南海海洋研究所)开放课题(LTO1408);国家海洋局近岸海域生态环境重点实验室资助项目(201508);广东省渔业生态环境重点实验室开放基金(LFE-2010-14);广东省自然科学基金(2014A030310495)

摘　　要：通过建立珠江口2009年夏季水质综合评价的自组织特征映射网络模型,探索了珠江口不同河段水质状况。结果表明,利用自组织特征映射网络人工神经网络可以直观清晰地对珠江口海域水质空间进行分类。珠江口海域水质可分为三大类,第一类为受到人类活动影响显著的广州河段区域;第二类为内伶仃洋海域,该区域主要受到咸淡水混合的影响;第三类为主要受到外海水交换影响的外伶仃洋海域。结果阐明人工神经网络模型能为珠江口环境保护与生物资源可持续利用提供科学的决策依据。The self-organizing map has been used to identify the spatial trend of water quality in the Pearl River estuary collected in summer 2009. The results indicated that the self-organizing map could clearly classify the water quality in the Pearl River estuary into three distinct groups. Water quality in the Guangzhou reach of the Pearl River was mainly influenced by human activities. However, water quality in the middle reaches of the Pearl River estuary（the inner Lingding Bay） was mainly controlled by the mixture of fresh water and sea water. The invasion of seawater was the main factor to shape the water quality in the outer Lingding Bay. The information extracted by the self-organizing map would be very useful for regional administration in developing strategies to carry out scientific resource-utilization plans based on marine system functions.

关 键 词：自组织特征映射网络 珠江口 水质 空间分布 

分 类 号：X832[环境科学与工程—环境工程]