蓝藻水华预报模型及基于遗传算法的参数优化

蓝藻水华预报是应对水危机,保障水资源供给的一项重要工作。以太湖北部三湾(竺山湖、梅梁湾、贡湖)为研究对象,采用动态空间环境建模技术,构建了蓝藻水华预报模型,并通过实地观测建立了模拟的初始参数集。利用2008年04-09月太湖水环境、气象等实测数据,采用遗传算法优化叶绿素a浓度预报模型中敏感度较高的4个参数。研究结果表明,该模型在蓝藻水华空间分布的预报上达到了一定的精度;采用遗传算法能全面、高效地进行参数优化,降低了模拟结果的相对残差,提高了模型预报精度。

Cyanobacteria bloom prediction model and parameters optimization based on genetic algorithm

Cyanobacteria bloom prediction is very important for water crisis and water resource security. Technique of dynamic spatial environmental modelling is used to develop cyanobacteria bloom prediction model used in three bays (Meiliang Bay, Zhushan Bay, Gong Bay) of northern Taihu Lake. The initial model parameters are obtained from field observation. The four parameters highly sensitive in chlorophyll-a concentration prediction are determined using Genetic Algorithm optimization technique. The observed field data of water environment and meteorological conditions in Taihu Lake from April to September 2008 are used for this purpose. The results showed that, Genetic Algorithm is comprehensive and efficient in optimizing model parameters, thus effective in improving prediction accuracy of the model and the relative residual decreases.