应用人工神经网络模型研究福建省山仔水库叶绿素a动态

本文应用山仔水库2003-2006年叶绿素a浓度、总磷浓度、总氮浓度、水温、溶解氧浓度、高锰酸盐指数、pH值7个参数监测数据对人工神经网络模型进行训练,在此基础上应用1997-2002年除叶绿素a浓度外其他6个参数监测数据,推算出1997-2002年间缺失的叶绿素a浓度,并对1997-2006年春末夏初的叶绿素a浓度动态进行分析,结果表明:山仔水库1997年建库初期,叶绿素a浓度处于较高水平,2000年以后叶绿素a浓度开始降低,近几年基本保持稳定.2003-2006年叶绿素a浓度呈季节周期性变化,春末经夏季到初秋,叶绿素a浓度持续升高,冬季下降明显,春季又开始回升;说明近几年山仔水库水体春末夏季秋初处于富营养化水平,秋末冬季处于中营养水平.本研究结果将为山仔水库的富营养化防治提供科学依据.

Study on Chlorophyll-a Trend in Shanzi Reservoir of Fujian Province Using Artificial Neural Network Model

This article used CODMn,temperature,pH value,dissolved oxygen,total phosphorus,total nitrogen and the concentration of chlorophyll-a which were measured in Shanzi reservoir from 2003 to 2006 for training the model of Artificial Neural Network,then the parameters except chlorophyll-a were further used to calculate the missing data of chlorophyll-a concentration in the reservoir during 1997 and 2002,in order to study the trend of chlorophyll-a from 1997 to 2006.The result indicated that the concentration of chlorophyll-a stayed at high level in 1997 during the construction period,then after 2000 it began to fall down,keeping steady in recent years.The concentration also showed seasonal variation from 2003 to 2006:it kept rising from late spring to early autumn,but dropped down obviously in winter and then rise again in next spring,which indicated that water body in the reservoir was mesotrophic in late autumn and winter and it was eutrophic from late spring to summer and early autumn in recent years.The results achieved in the study would provide a scientific reference for the prevention and control of eutrophication in the reservoir.