基于时间序列模型的海南省白纹伊蚊种群动态与气象影响因素研究  

Research on the population dynamics and the meteorological influencing factors of Aedes albopictus in Hainan Province based on time series models

作　　者：李明发 刘莹[2] 刘璞瑜 吴群 曾雪霞[2] 孙定炜[2] 杨国静 LI Mingfa;LIU Ying;LIU Puyu;WU Qun;ZENG Xuexia;SUN Dingwei;YANG Guojing(NHC Key Laboratory of Tropical Disease Control,School of Tropical Medicine,International School of Public Health and One Health,Hainan Medical University,Haikou,Hainan,571199,China;Institute for Prevention and Control of Tropical and Chronic Diseases,Hainan Center for Disease Control and Prevention,Haikou,Hainan 570203,China)

机构地区：[1]海南医学院国家卫生健康委员会热带病防治重点实验室,热带医学院,公共卫生与全健康国际学院,海南海口571199 [2]海南省疾病预防控制中心,热带病与慢性病预防控制所,海南海口570203

出　　处：《中国热带医学》2024年第3期282-286,314,共6页China Tropical Medicine

基　　金：海南省自然科学基金项目(No.821CXTD440,No.821MS0844,No.821QN420);海南省科技专项资助(No.ZDYF2023SHFZ134,No.ZDKJ2021035)。

摘　　要：目的探讨基于气象因子的时间序列模型在海南省白纹伊蚊种群密度中的应用,为海南省登革热防控提供科学依据。方法采用双层叠帐法和布雷图指数法,对2017—2022年海南省18个市县不同生境的白纹伊蚊逐月进行监测,记录帐诱指数和布雷图指数。采用Mann-Kendall趋势性检验分析2个密度指数的时间变化趋势;采用Spearman相关分别分析各气象因子与两个指数的关联强度,并剔除无关变量,进一步通过全子集回归法筛选出最终变量;基于时间序列分析对2个密度指数分别构建3个时间序列模型,并以均方根误差、平均绝对误差等为精度评价指标确定最优模型;对2023—2024年的密度指数进行预测。统计分析均在R(4.3.1)中进行。结果帐诱指数和布雷图指数整体呈逐年下降趋势(Mann-Kendall趋势检验Z分别为-6.15和-4.03,P均<0.05);与帐诱指数关联性最强的气象因子是月平均最低气温;与布雷图指数关联性最强的气象因子是月平均最低气温和月平均相对湿度。综合各项评估指标,多元时间序列模型精度最高。2023年和2024年帐诱指数和布雷图指数有1~2个峰值,且峰期分别在6—9月和5—9月。2023年预测值与实测值特征基本一致,预测效果显著。结论本研究在季节性时间序列的基础上加入了气象因子,更准确地预测了2023年和2024年海南省伊蚊密度,可为海南省登革热防控提供参考依据。Objective To explore the application of time series models based on meteorological factors in the population density of Aedes albopictus in Hainan Province,and to provide a scientific basis for the prevention and control of dengue fever in Hainan Province.Methods The density of Aedes albopictus in different habitats in 18 cities and counties of Hainan Province from 2017 to 2022 was monitored monthly using the double-mosquito net trapping index and the Breteau index.Mann-Kendall trend test was used to analyze the temporal trends of the two density indices;Spearman's correlation analysis was employed to assess the correlation strength between each meteorological factor and the two indexes,eliminating unrelated variables,and further selecting the final variables through the full-subset regression method.Three time-series models were constructed for the two density indices,with root mean square error(RMSE),mean absolute error(MAE),and other accuracy metrics used to determine the optimal model;predictions for the density indices for 2023-2024 were made.All statistical analyses were performed in R(4.3.1).Results The net trapping and Breteau indices showed an overall decrease over the years(Z-values of Mann-Kendall trend test were-6.15 and-4.03,respectively,and P<0.05).The meteorological factor most strongly associated with the trap index was the monthly average minimum temperature;monthly mean minimum temperature and monthly mean relative humidity were strongly correlated with the Breteau index.Based on various evaluation indicators,the multivariate time series model demonstrated the highest accuracy.The study predicts one to two peaks in both the trap index and Breteau Index for the years 2023 and 2024,with peak periods occurring between June to September and May to September,respectively.The predicted value for 2023 aligns with the measured value,demonstrating outstanding predictive accuracy.Conclusions This study has introduced meteorological factors into the seasonal time series model,allowing for more accurate pre

关 键 词：白纹伊蚊 媒介监测 消长趋势 影响因素 多元时间序列 

分 类 号：R384.1[医药卫生—医学寄生虫学]