非点源污染负荷预测研究现状及发展趋势

Prediction for non-point source pollution load current status and future progress

简要回顾了非点源污染负荷预测的发展历程,对国内外在该领域的研究现状及发展趋势作了较全面的分析与展望.介绍了4种非点源污染负荷预测新方法:人工神经网络预测法、偏最小二乘回归预测法、支持向量机预测法、自记忆原理预测法.指出了现状研究的不足之处,今后研究的重点包括非点源污染负荷迁移转化规律及其机理研究、基础数据库的建立、控制管理措施的优化、不确定性分析方法的建立、预测模拟新方法的研究.

The paper reviews briefly the developmental history of non-point source pollution load prediction,and summarizes comprehensively its current research status and priority research areas inland and abroad.Four new methods for prediction of non-point source pollution load are presented：artificial neural network prediction method,partial least square regressive prediction model,support vector machines prediction method and self-memory prediction method.Research indicates shortage of current situation,and emphases include transference mechanism,date-base foundation,control measurement optimization,uncertainty analysis methods foundation,and prediction of new methods study for non-point source pollution load in the future.