典型云南高原湖泊叶绿素a与影响因子的定量关系及对比分析

湖泊富营养化的影响因子涉及水文、物化、生物等多方面,具有复合性和非线性特征,定量化其与影响因素间的相关关系有助于识别影响湖泊营养状态的关键因子,可以用较低的成本、较短的时间达到理想的控制效果.云南高原湖泊具有易发生富营养化的自然和气候特征,对其富营养化发生条件及影响因子的分析可为科学的控制决策提供参考.本文选取云南滇池、程海、抚仙湖和异龙湖4个高原湖泊,比较湖泊自然特征与流域社会经济条件的异同;构建包括绝对主成分多元线性回归分析(APCS-MLR)、结构方程模型(SEM)及人工神经网络模型(ANN)的综合分析方法,重点关注并确定浮游初级生产力的代表指标(叶绿素a,Chla)与相关影响因子间的定量相关关系.研究发现:14个湖泊中,对Chla浓度变化影响最大的均为理化因子,但在各湖中该影响的正、负性及不同理化因子的贡献权重有较大差异;2流域污染源构成的不同在一定程度上影响了入湖的氮、磷负荷,使4个湖泊表现出不同的营养盐限制性特征;3流域面积、湖泊形态及湖体水动力条件影响着营养盐在湖体中的迁移转换,造成4个湖泊富营养化的差异性特征;4对Chla与影响因子间因果关联的识别须结合深入的机理过程分析.

Quantitative relationship between chlorophyll a and key controlling factors in Four Plateau Lakes in Yunnan Province, China

Most lakes in China are impaired by eutrophication, therefore it is necessary to take effective measures to restore the water quality. Plateau lakes in Yunnan Province, China are unique due to their distinct natural characteristics. Different dynamic processes contribute to the eutrophication of each lake, making it impossible to implement the same management method to all the lakes. Therefore, it is essential to identify the key driving factors of lake eutrophication to support more reliable and effective decision making. Previously, many water quality management and restoration projects have been implemented to address the water quality problems in lakes in Yunnan Province; however, the achieved water quality improvement is far from expected environmental objectives. Therefore, conducting comparative research among the plateau lakes in Yunnan Province can not only identify the difference in the driving factors and dynamic processes between these plateau lakes, but also find some similar key processes in eutrophic mechanism. In this study, four typical plateau lakes were selected, including Lake Dianchi, Lake Chenghai, Lake Fuxian and Lake Yilong. An integrated approach of absolute principle components score-multivariate linear regression (APCS-MLR) and structural equation modeling (SEM) method were developed in this study to understand the influence of water chemistry variables on chlorophyll a (Chla). The SEM results were further validated with the artificial neural networks (ANN). The model results demonstrated that among all factors, the physical and chemical conditions in the lakes have the greatest influence on Chla. However, different physical and chemical factors make different contribution to the Chla between the four lakes. The comparative analysis also showed that lake morphology, watershed population and industrial composition, and aquatic ecosystem variations would have great influence on lake eutrophication.