基于HEA模型的达里诺尔叶绿素a浓度模拟与分析

Simulation analysis of chlorophyll-a concentration in Lake Dali by Hybrid Evolutionary Algorithm

针对达里诺尔湖富营养化问题,通过2000—2020年达里诺尔水质监测数据,运用混合演化算法(HEA)对达里诺尔西部、中部和东部叶绿素a浓度的变化规律进行模拟和预测,并揭示其三个采样点的变化规律。结合阈值数据和敏感性分析量化叶绿素a浓度与环境因子的关系。研究结果表明:(1)达里诺尔叶绿素a浓度的季节变化呈显著性差异,表现为夏季和秋季显著高于春季,但三个采样点之间不具有显著性差异;(2)达里诺尔湖西部、中部和东部HEA模型模拟值与实际监测值之间相关系数(R^(2))分别为0.50、0.62和0.75,均方根误差(RMSE)分别为2.08、1.62和1.09 mg·m^(-3);(3)在一定的变化范围和阈值条件下,叶绿素a浓度与高锰酸盐指数(COD_(Mn))、生化需氧量(BOD_(5))、铵态氮(NH_(4)_N)和水温(WT)呈较强的正相关关系,与水深(WD)、透明度(SD)呈较强的负相关关系,与pH相关性不明显。研究显示,HEA算法较好地解释并模拟了叶绿素a的含量变化,模型总体精确度较高,可为达里诺尔藻华早期的预测预警提供一定的数据基础。

In order to solve the eutrophication problem in Lake Dali,the variation of chlorophyll-a concentration in the west,center and east of Lake Dali was simulated and predicted by Hybrid Evolutionary Algorithm(HEA)based on the monitoring points data of water quality in Dali from 2000 to 2020,and the three monitoring points of chlorophyll-a concentration were revealed.The threshold and sensitivity analysis were combined to quantify the relationship between chlorophyll-a concentration and environmental factors.The results showed that:(1)the concentration of chlorophyll-a showed significant seasonal variations,which was significantly higher in summer and autumn than spring,but there was no significant difference in three monitoring points;(2)the correlation coefficients(R^(2))between simulated values and actual monitoring values in the western,central and eastern parts of Lake Dali were 0.500.62 and 0.75,respectively.And the root mean square error(RMSE)were 2.08,1.62 and 1.09 mg·m^(-3),respectively;(3)under certain range and threshold conditions,chlorophyll-a concentration had a positive correlation with permanganate index,biochemical oxygen demand,ammonia nitrogen and water temperature,and a negative correlation with water depth and Secchi disk depth,but no obvious correlation with pH.The results showed that the HEA could explain and simulate the changes of chlorophyll-a,and the accuracy of the model was high,which provided a certain data basis for the early prediction and early warning of lake Dali bloom.