| 题名 | 结合元胞自动机和数据驱动的太湖蓝藻水华模型 |
| 作者 | 张晓晴 |
| 学位类别 | 博士 |
| 答辩日期 | 2014-05 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 北京 |
| 导师 | 陈求稳 |
| 关键词 | 蓝藻水华 预测模型 小波分析 混合进化算法 元胞自动机cyanobacteria blooms predictive models wavelet analysis hybrid evolutionary algorithm cellular automata |
| 其他题名 | A Cyanobacteria Bloom Model for Lake Taihu Combining Cellular Automata and Data-Mining Methods |
| 学位专业 | 生态学 |
| 中文摘要 | 湖库富营养化及其导致的藻类水华改变了水体理化特性，破坏了生态系统完整性，威胁人类身体健康，对社会经济造成严重的损失。因此，蓝藻水华的防治一直是湖泊研究的热点。本论文旨在揭示水环境因子对浮游植物种群结构及蓝藻水华的影响，建立水华动力学预测模型，模拟藻类时空动态，提升蓝藻水华应急管理能力。 论文以太湖、洋河水库和密云水库为研究实例，调查分析了三个湖库水质时空动态变化及浮游植物群落组成特征，采用多元统计典范对应分析揭示了多环境变量与藻类种群结构的定性关系；应用小波相干性方法识别了太湖蓝藻与环境因子的相干关系及滞后效应，并在此基础上，建立了基于单点的预测模型和基于站点聚合的通用预测模型；最后，通过结合元胞自动机模式和遗传编程方法，模拟了太湖蓝藻的时空动态变化。取得的主要成果包括： 1）初步分析了环境因子与浮游植物种群结构的定性相关关系。通过具有温度梯度和营养盐梯度的三个湖库对比分析，揭示了温度和营养盐浓度对藻类种群结构演替的作用。典范对应分析表明，pH、水温和营养盐为影响浮游植物种群分布的重要因子。具有相似营养盐水平的太湖和洋河水库，氮为影响蓝藻生长的限制因子；由于太湖的平均水温比洋河水库高 3~4℃，导致蓝藻数量明显高于洋河水库。对于处于同一气候带的洋河水库和密云水库，由于洋河水库为重度富营养化，而密云水库为中营养化，导致洋河水库的优势种为蓝藻-微囊藻， 密云水库的优势种为硅藻或者绿藻，磷为限制因子。 2）利用小波相干性方法识别了太湖蓝藻与环境因子之间的相干关系和滞后效应。结果表明，蓝藻与 pH、水温和磷酸盐变化同步，无滞后；蓝藻与透明度、氨氮和硝氮的变化呈反相位，滞后时间分别为 2天、3天和 3天。同时，氮为夏秋季节蓝藻水华的主要驱动因子，尤其在梅梁湾氮为蓝藻生长的限制因子。 3）采用混合进化算法不仅较好地预测了蓝藻水华的发生，而且可以得到显式的数学表达式。基于单点的混合进化算法预测模型能较好的预测不同湖区蓝藻水华的时间和峰值，r2在 0.62~0.83之间；基于站点聚合的通用混合进化算法预测模型对梅梁湾、竺山湖和西部沿岸区蓝藻数量较高的区域预测结果较好，r2最高为0.77。敏感性分析发现，水温为区分蓝藻生长季节（夏秋）和非生长季节（冬春）的标志；蓝藻生长与硝态氮和氨氮呈负相关关系，与磷酸盐呈正相关关系，氮为蓝藻生长的限制因子。 4）结合元胞自动机和遗传编程有效地预测蓝藻水华的时空动态。除了蓝藻含量较低的东部湖区以外，模型能较好的预测高含量蓝藻水华的峰值和时间；同时，模型能较好的捕获整个太湖的蓝藻空间动态，从而提升对蓝藻水华空间斑块形成机制的认识以及水华应急管理能力。 |
| 英文摘要 | The eutrophication and cyanobacteria blooms produce taste and odor problems in drinking water supply, damage the ecosystem diversity, threaten the human health,and cause serious economic losses. Therefore, the prevention and control of cyanobacteria blooms have become the focus of lake studies. This study developed a dynamics model to predict the cyanobacteria blooms based on the responses of phytoplankton to environmental factors. Then an integrated model was used to predict the spatio-temporal dynamics of cyanobacteria blooms. The developed model can help to improve the capability of cyanobacteria blooms management. This study selected Lakes Taihu, Yanghe and Miyun as case studies. Field survey was conducted to analyze the dynamic of water quality and phytoplankton composition. The method of Canonical correspondence analysis was applied to determine the qualitatively relationships between phytoplankton and environmental factors. Then the wavelet coherency analysis was used to identify the correlations and time-lagging responses of cyanobacteria to environmental factors. Based on the wavelet analysis, site-specific and generic hybrid evolutionary algorithm (HEA)based models were built to predict the cyanobacteria blooms. Finally, this research developed a spatio-temoral dynamics model of cyanobacteria blooms by combining the cellular automata and genetic programming techniques. The major contributions of the study are as follows. 1）The qualitatively relationships between phytoplankton and environmental factors were obtained. This study revealed the impacts of several key environmental factors on cyanobacteria growth in limnological data of three lakes categorized by climate and trophic state. Results clearly showed that pH, water temperature and nutrients were the principle factors that affected the phytoplankton distribution. Nitrogen was the limiting factor in Lakes Taihu and Yanghe, which locate in different climate but similar trophic state. The higher temperature (3~4℃) in Lake Taihu contributed to the higher abundance of Microcystis compared to Lake Yanghe. By contrast, Lake Miyun locates in different trophic state but similar climate with Lake Yanghe. The Lake Yanghe was in hyper-eutrophication state and the dominant species was cyanobacteria. The Lake Miyun was in mesotrophication state and the dominant species was diatom or chlorophyta, which was supported by findings of PO4-P limitation. 2）The wavelet analysis was applied to identify the correlations and time-lagging responses of cyanobacteria to environmental factors. Results revealed that the time series between cyanobacteria and pH, water temperature and phosphate were in phase with no delay; the time series between cyanobacteria and Secchi depth、NH4-N and NO3-N were out of phase with a short lag times of 2 days, 3 days and 3 days respectively. Nitrogen were also identified as key driving forces for cyanobacteria growth in summer and autumn because of seasonal N-limitation in the Meiliang Bay. 3）The HEA model could not only well predict the outbreak of cyanobacteria blooms, but also present an explicit function to describe the influence of different environmental factors on the growth of cyanobacteria. The resulting site-specific models well matched the timing and magnitude of the observed cyanobacteria dynamics for all areas, which was reflected by coefficients of determination (r2 ) of 0.62~0.83. The generic model achieved good predictive results in Meiliang Bay, Zhushan Bay and Western Lake with high cyanobacteria population, with the highest r2 value of 0.77. The sensitivity analyses revealed the water temperature was the threshold to distinguish between the slow growing season (winter to spring) and the fast growing season (summer to autumn) of cyanobacteria. The cyanobacteria had inhibitory relationships with nitrate and excitatory relationships with phosphate, which suggested N-limitation of the lake existed in summer and autumn. 4） The spatio-temporal dynamics model of cyanobacteria blooms was successfully developed by integrating cellular automata and genetic programming techniques. The model well predicted the magnitude and timing of cyanobacteria blooms for all areas except the eastern lake with lower cyanobacteria population. Simultaneously, the model could well capture the spatial dynamics of cyanobacteria, intitively enhance our cognition on space plaque formation mechanism of cyanobacteria blooms and help to improve the capability of cyanobacteria blooms management. |
| 公开日期 | 2015-07-08 |
| 内容类型 | 学位论文 |
| 源URL | [http://ir.rcees.ac.cn/handle/311016/15701]  |
| 专题 | 生态环境研究中心_环境水质学国家重点实验室 |
| 推荐引用方式 GB/T 7714 | 张晓晴. 结合元胞自动机和数据驱动的太湖蓝藻水华模型[D]. 北京. 中国科学院研究生院. 2014. |
| 个性服务 |
| 查看访问统计 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论