基于分布滞后非线性模型的淡色库蚊种群密度动态与气象因素关系研究

STUDY ON THE RELATIONSHIP OF DYNAMICS OF CULEX PIPIENS PALLEN POPULATION DENSITY TO METEOROLOGICAL FACTORS BASED ON DISTRIBUTED LAG NON-LINEAR MODEL

为建立基于分布滞后非线性模型(Distributed Lag Non-linear Model, DLNM)的CO_(2)诱蚊灯监测的蚊密度模型,探索不同气候条件下上海市淡色库蚊密度变化的特征,在上海市15个区,于4-11月每旬共设置229个CO_(2)诱蚊灯,监测淡色库蚊密度。以2018-2020年的监测数据建立基于DLNM的蚊密度与气象因素的方程。以改进的赤池信息准则(QAIC)对各个气象因素、连接函数和参数组合进行评价,最终选择QAIC最小的模型。最终建模结果取类泊松模型为DLNM的连接函数,进入模型的气象因素包括每旬的日平均最高气温、平均日降雨量和平均相对湿度。结果显示,旬平均最高气温与淡色库蚊密度呈倒“U”型关系,在平均最高气温大于10℃后逐步升高,在30℃达到密度峰值,其后逐步降低。平均日降雨量与淡色库蚊密度呈“J”型关系,在日平均降雨量大于6 mm后逐步升高。旬平均相对湿度与淡色库蚊密度呈倒“U”型关系,在平均相对湿度为70%时达密度峰值。结果表明,DLNM可以解释气象因子变量对蚊虫种群密度的影响,从而预测未来蚊虫密度的变化,为开展蚊虫控制措施提供科学依据。

To establish a CO_(2)mosquito trap monitoring model based on the distributed lag non-linear model(DLNM), and to explore the characteristics of density changes of Culex pipiens pallens in Shanghai. A total of 229 CO_(2)mosquito traps were set up in 15 districts of Shanghai from April to November to monitor the density of Culex pipiens pallens. With the monitoring data from 2018 to 2020, the equation between mosquito density and meteorological factors based on DLNM was established. The QAIC model was used to evaluate the meteorological factors and parameters, and the model with the lowest QAIC was selected to analyze. The Poisson like model was taken as the connection function of DLNM. The meteorological factors included the daily average maximum temperature, average daily rainfall and average relative humidity in every ten days. The results showed the ten day average maximum temperature and the density of Culex pipiens pallens showed an inverted "U" shape, gradually increased after the average maximum temperature was greater than 10 ℃, reached the peak density at 30 ℃, and then gradually decreased. The average daily rainfall and the density of Culex pipiens showed a "J" shape, and gradually increased after the average daily rainfall was greater than 6 mm. The ten day average relative humidity and the density of Culex pipiens pallens showed an inverted "U" shape, and reached the peak when the average relative humidity reached 70%. The DLNM can explain the influence of meteorological factors on mosquito population density, so as to predict the change of mosquito density in the future and provide a scientific basis for the design of mosquito control measures.