深圳流感发病的气象诱因及预测建模研究

马盼; 王馨梓; 张莉; 唐小新; 冉洪雨; 王式功

doi:10.11676/qxxb2022.039

深圳流感发病的气象诱因及预测建模研究

doi: 10.11676/qxxb2022.039

马盼^1,,
王馨梓^{1, 2},
张莉³,
唐小新^3, ,,
冉洪雨¹,
王式功¹

1.
成都信息工程大学环境气象与健康研究院，气象环境与健康联合实验室，成都，610225
2.
天津市津南区气象局，天津，300350
3.
深圳市国家气候观象台，深圳，518000

基金项目: 深圳市气象服务中心“城市安全与民生重点保障服务”项目（20200605）、北京市科技计划课题（Z201100008220002）、甘肃省科技计划项目（21YF5FA169）、商洛气候适应型城市重点实验室基金（SLSYS2019004）、成都信息工程大学大学生创新创业项目（202010621002）

详细信息

作者简介:
马盼，主要从事气象环境与健康研究。E-mail： mapan@cuit.edu.cn

通讯作者:
唐小新, 主要从事大城市气象防灾应用技术研究。E-mail：786578343@qq.com

中图分类号: P49 X503.1
计量
- 文章访问数: 275
- HTML全文浏览量: 27
- PDF下载量: 86
- 被引次数: 0
出版历程
- 收稿日期: 2021-11-17
- 录用日期: 2022-05-11
- 修回日期: 2022-03-25
- 网络出版日期: 2022-04-12

The association between influenza onset and meteorological inducers in Shenzhen and construction of predictive models

MA Pan^1
,,
WANG Xinzi^{1, 2},
ZHANG Li³,
TANG Xiaoxin^{3
, ,},
RAN Hongyu¹,
WANG Shigong¹

1.
Institute of Environmental Meteorology and Health/Joint Laboratory of Meteorological Environmental and Health，Chengdu University of Information Technology，Chengdu 610225，China
2.
Meteorological Bureau of Jinnan District，Tianjin 300350，China
3.
Shenzhen National Climate Observatory，Shenzhen 518000，China

摘要

摘要: 深圳市在人口结构和经济社会发展方面很特殊，而对于其热带向副热带过渡的气候特征对流感发病的影响仍缺乏深入研究。本研究收集长序列（2003—2019年）的深圳市流感样病例（Influenza Like Illness，ILI）监测数据，采用分布滞后非线性模型（DLNM）系统分析了ILI与多种气象因子的关联，并分别使用Prophet时间序列和多元逐步回归模型对流感风险进行预报。近17年来深圳ILI发病在2003—2009年增加、2010—2014年平稳、2015—2019年下降，年周期特征凸显；多数年份发病率呈夏季单峰型，与高温、高湿的气候背景高度相符；个别年份在年末出现次高峰，常与大规模暴发疫情有关。DLNM揭示，高温对ILI风险的即时性影响较强，气温达到29.9℃，相对危险度值（RR）可达1.237（95%置信区间（95%CI）：1.203—1.272）；而低温效应在滞后2—3周起主导作用。70%—75%的湿度范围对应ILI高风险段， 70%相对湿度的RR为1.089（95%CI：1.046—1.135）。偏高的湿度与高温共存可诱使ILI最高风险点出现，即二者有协同增强效应，在其长夏短冬气候下尤其需要注意。ILI危险度在气温日较差为4—6℃或>9℃时均有显著增加，即日内温差对流感的活跃程度亦有显著影响；由于深圳的风速整体较小，其影响整体较弱。Prophet时间序列模型和逐步回归模型的回报准确率相近（>86%），而同时考虑了气象因子和前期发病人数的回归模型预测准确率更高（>80%）。简言之，深圳市ILI风险与温、湿度的非线性协同影响关系最为密切，其发病率很大程度上是可预测的。
- 流感样病例 /
- 气温 /
- 相对湿度 /
- 气温日较差 /
- 预测
Abstract: Shenzhen is very unique in population structure as well as social and economic developments, but the relationship between its transitional monsoon climate from the tropics to subtropics and the influenza morbidity has rarely been evaluated. In order to reveal the epidemiologic features of influenza in Shenzhen, long-term surveillance data of influenza like illness (ILI) for the recent 17 years (2003—2019) are collected. The distributed-lag nonlinear model (DLNM) is adopted to investigate the associations between ILI morbidity and climatic variables. Further, the Prophet time-series approach and a stepwise-regression model are used to predict influenza risks, respectively. The results reveal that the number of ILI outpatients increases at the early stage of the study period (2003—2009), then stabilizes between 2010 and 2014, and decreases slightly from 2015 to 2019. Besides, a relatively clear annual cycle of ILI morbidity is discovered, which shows a unimodal pattern in most of the years. The number of ILI cases usually peaks during the summer months, which is highly consistent with the hot and humid climatic background. A second peak that is usually related to large-scaled influenza epidemics occurs occasionally at the end of the year. The DLNM shows that high air temperature has substantial and immediate impact on the ILI risk (T=29.9℃, RR=1.237, 95%CI: 1.203—1.272), while the effect of low temperature is relatively weak and occurs with a significant lag (2—3 weeks). Relative humidity of 70%—75% can induce the highest risk of ILI (e.g., RH=70%, RR=1.089, 95%CI: 1.046—1.135), and it also has synergistic cross-effect with high temperature. Besides, the number of ILI outpatients increases prominently when daily temperature range (DTR) is within in 4—6℃ or >9℃, indicating a nonlinear relation between DTR and the spread of influenza virus. Due to the overall low wind speed in Shenzhen, its effect on ILI is insignificant. The correct rates of prediction by the Prophet time-series approach and multiple-linear regression model are similar to each other (>86%). Nevertheless, the prediction accuracy of the regression model is relatively high (>80%) when considering meteorological factors and previous ILI cases simultaneously. In summary, the ILI morbidity in Shenzhen is significantly and nonlinearly associated with the synergistic effect of air temperature and humidity. Thus, the influenza risk is to some extent predictable.
- Influenza like illness /
- Air temperature /
- Relative humidity /
- Daily temperature range /
- Prediction

HTML全文

图 1 2003—2019年深圳市每周ILI就诊人次（a）与气象要素（b—f）的时间序列（a. 红色虚线为拟合的ILI发病年周期，b. 平均气温，c. 气温日较差，d. 相对湿度，e. 绝对湿度，f. 风速）

Figure 1. Time series of weekly ILI cases （a） and meteorological factors （b—f） in Shenzhen from 2003 to 2019 （a. the red dashed line indicates the fitted annual cycles of ILI cases，b. temperature，c. daily temperature range，d. relative humidity，e. absolute humidity，f. wind speed）

下载: 全尺寸图片幻灯片

图 2 深圳市2003—2019年流感样病例数及就诊率的逐月分布

Figure 2. Monthly distribution of weekly ILI and ILI% in Shenzhen from 2003 to 2019

下载: 全尺寸图片幻灯片

图 3 深圳市主要气象要素的频次分布特征（a. 平均气温，b. 相对湿度，c. 风速，d. 气温日较差）

Figure 3. Frequency distributions of main meteorological factors in Shenzhen （a. air temperature，b. relative humidity，c. wind speed，d. daily temperature range）

下载: 全尺寸图片幻灯片

图 4 气象要素（a. 平均气温，b. 相对湿度，c. 风速，d. 气温日较差）对ILI的累积滞后（3周）危险度及其95%置信区间

Figure 4. Effects of meteorological variables （a. air temperature，b. relative humidity，c. wind speed，d. daily temperature range） and their cumulative RR（95%CI） on the ILI at lag of 0—3 weeks

下载: 全尺寸图片幻灯片

图 5 气温对ILI风险的影响随着滞后时长（a—d. 滞后0—3周）的变化

Figure 5. Effects of weekly mean temperature on the risk of ILI at different lag periods （a—d. lag 0—3 week）

下载: 全尺寸图片幻灯片

图 6 气温分别与（a）相对湿度和（b）气温日较差对ILI风险交叉效应的三维曲面

Figure 6. Three-dimensional curves of cross-effects （a） between temperature and RH，and （b） between temperature and diurnal temperature range

下载: 全尺寸图片幻灯片

图 7 风速（a）和气温日较差（b）在滞后0 （a₁—b₁）和1 （a₂—b₂）周时对ILI相对危险度的影响

Figure 7. Effects of wind speed （a） and daily temperature range （b） on ILI RR at lag of 0 （a₁—b₁） and 1 （a₂—b₂） week

下载: 全尺寸图片幻灯片

图 8 基于（a）多元逐步回归和（b） Prophet时间序列模型的ILI风险预测效果

Figure 8. Predicted results of ILI risk based on （a） the stepwise-regression equations and （b） Prophet time-series model

下载: 全尺寸图片幻灯片

表 1 深圳市2003—2019年每周流感样病例及气象变量的统计特征

Table 1. Statistics of weekly ILI cases and meteorological variables in Shenzhen from 2003 to 2019

	平均值	标准差	最小值	最大值
ILI	925.94	390.1	134	2823
标化ILI （/万人）	578.9	183.9	145.7	1324.1
ILI%（%）	5.79	1.84	1.46	13.24
平均气温（℃）	23.35	5.18	8.39	30.96
日最高气温（℃）	26.97	5.02	10.60	34.90
日最低气温（℃）	20.92	5.30	6.79	28.37
相对湿度（%）	73.39	10.2	34.57	97.0
绝对湿度/比湿（g/kg）	13.97	4.94	3.28	22.80
24 h降水量（mm）	5.00	8.17	0.0	63.50
气温日较差（℃）	6.05	1.14	2.80	10.13
风速（m/s）	2.15	0.46	1.19	3.97
日照时数（h）	5.17	2.39	0.0	12.03
平均气压（hPa）	1006.0	6.23	989.6	1023.6

下载: 导出CSV

表 2 深圳的四季划分及其流感样病例特征

Table 2. The division of seasons and corresponding ILI characteristics in Shenzhen

	春季	夏季	秋季	冬季
开始日期	2月6日	4月21日	11月4日	1月13日
季长（d）	76	196	69	24
气温（℃）	18.2	27.6	18.2	14.8
降水量（mm）	275.4	1562.5	66.0	27.7
平均ILI（人·次）	838.07	1025.49	783.07	802.95
平均ILI%（%）	5.64	6.31	4.74	5.01

下载: 导出CSV

表 3 周ILI人次与气候因子分季节的斯皮尔曼相关

Table 3. Spearman correlations between ILI outpatients and meteorological variables in different seasons

	夏季	春秋季	冬季	全年
平均气温（℃）	0.219^**	0.168^**	−0.011	0.341^**
最高气温（℃）	0.175^*	0.146^**	0.063	0.325^**
最低气温（℃）	0.245^**	0.191^**	−0.059	0.352^**
气温日较差（℃）	−0.143^**	−0.165^**	0.176	−0.198^**
相对湿度（%）	0.037	0.278^**	−0.161	0.192^**
绝对湿度（g/kg）	0.126^**	0.262^**	−0.100	0.340^**
降水量（mm）	0.148^**	0.157^**	−0.145	0.257^**
平均气压（hPa）	−0.102^**	−0.193^**	−0.006	−0.322^**
风速（m/s）	0.099^*	−0.217^**	−0.127	−0.070^*
日照时数（h）	0.040	−0.165^**	0.175	0.051
注：*和分别表示相关系数通过了显著性水平为0.01和0.05 的显著性t检验。

下载: 导出CSV

表 4 各气象要素的典型值对ILI的单周滞后相对危险度及其95%置信区间

Table 4. RR and 95%CI of individual meteorological variables on ILI risk at various lag times

		滞后0 周	滞后1 周	滞后2周	滞后3周
日均气温（℃）	12.8	0.854 （0.830—0.878）	1.025 （0.997—1.053）	1.032 （1.004—1.060）^*	1.094 （1.067—1.121）^*
日均气温（℃）	29.9	1.237 （1.203—1.272）^*	1.058 （1.028—1.088）^*	1.094 （1.063—1.125）^*	1.069 （1.040—1.098）^*
相对湿度（%）	70	1.089 （1.046—1.135）^*	1.044 （1.002—1.087）^*	1.070 （1.027—1.114）^*	1.275 （1.225—1.326）^*
相对湿度（%）	91	0.963 （0.918—1.011）	0.981 （0.935—1.029）	0.955 （0.911—1.001）	1.081 （1.032—1.131）^*
风速（m/s）	1	0.944 （0.916—0.972）	1.013 （0.984—1.044）	1.058 （1.027—1.089）^*	0.916 （0.889—0.944）
风速（m/s）	3	1.015 （1.004—1.026）^*	1.024 （1.013—1.035）^*	0.991 （0.980—1.002）	0.988 （0.977—0.999）
气温日较差（℃）	5	1.194 （1.141—1.250）^*	1.051 （1.006—1.099）^*	1.042 （0.996—1.090）	1.049 （1.003—1.096）^*
气温日较差（℃）	10	1.178 （1.110—1.251）^*	1.055 （0.995—1.118）	1.086 （1.025—1.150）^*	1.018 （0.962—1.078）
注：^*表示相对危险度通过了置信度水平0.05的显著性t检验。

下载: 导出CSV

表 5 ILI预报模型优度检验

Table 5. Testing indices for the predictive model of ILI risk

		回代拟合				试预报
		MAE	RMSE	MAPE	P（%）	MAE	RMSE	MAPE	P（%）
Prophet模型		79.47	109.77	0.135	86.49	133.59	165.97	0.363	68.80
逐步回归	全年	66.92	97.56	0.117	88.62	82.21	133.22	0.199	80.80
	冷季	60.65	90.79	0.118	88.62	88.65	162.66	0.194	80.89
	暖季	68.11	99.99	0.110	89.16	75.22	87.87	0.220	80.13

下载: 导出CSV

表 6 深圳市ILI风险的逐步回归模型

Table 6. The stepwise-regression equations of ILI risk in Shenzhen

	回归方程	R²
全年	Y=1291.439+0.807ILI'+9.219（ΔT₁）⁰+3.386（ΔT₂）¹−1.171（p₁）¹	0.706
冷季	Y=2473.517+0.794ILI'−6.482（Δp₁）⁰+4.127（p₁）¹+7.995（ΔΔp）⁰	0.651
暖季	Y=−13.353+0.826ILI'+1.588（RH₁）¹	0.700
注：（X）^a—发病之前a周的气象因子X，即考虑发病前a周的气象条件；（X_b）^a—气象因子 X 在连续b周内进行平均，a同上；ΔX —气象因子X的周平均值相对于上周平均值的差值，即周际变化；ΔΔX—气象因子X的周内标准差，即体现其周内平均变幅。表中各气象因子单位分别为：T （℃），p （hPa），RH （%），V （m/s）；ILI'代表前一周病例数。

下载: 导出CSV

参考文献(42)

[1]	郭貔,李克. 2011. 香港1997—2008年甲型流感季节性波动与气候因素的相关性分析. 中华流行病学杂志,32(10):1005-1008 doi: 10.3760/cma.j.issn.0254-6450.2011.10.012 Guo P,Li K. 2011. A retrospective analysis on the association between seasonality of influenza A and climate factors in human from 1997 to 2008,in Hong Kong. Chinese J Epidemiol,32(10):1005-1008 (in Chinese) doi: 10.3760/cma.j.issn.0254-6450.2011.10.012
[2]	刘刚,张敏,廖异等. 2021. 深圳市社区医生流感疫苗接种意愿及因素分析. 中国公共卫生,37(3):401-404 doi: 10.11847/zgggws1130217 Liu G,Zhang M,Liao Y,et al. 2021. Willingness to have influenza vaccine vaccination and its associates among community health workers in Shenzhen city:A cross-sectional survey. Chinese J Public Health,37(3):401-404 (in Chinese) doi: 10.11847/zgggws1130217
[3]	逯建华,程小雯,房师松等. 2009. 深圳市2005—2008年流感样病例暴发疫情特征分析. 中国热带医学,9(11):2102-2103,2111 Lu J H,Cheng X W,Fang S S,et al. 2009. Features of outbreak of influenza-like infection in 2005-2008 in Shenzhen. China Trop Med,9(11):2102-2103,2111 (in Chinese)
[4]	马盼,王式功,尚可政等. 2018. 气象舒适条件对呼吸系统疾病的影响. 中国环境科学,38(1):374-382 doi: 10.3969/j.issn.1000-6923.2018.01.042 Ma P,Wang S G,Shang K Z,et al. 2018. The impact of meteorological comfort conditions on respiratory disease. China Environ Sci,38(1):374-382 (in Chinese) doi: 10.3969/j.issn.1000-6923.2018.01.042
[5]	彭质斌,王大燕,杨娟等. 2018. 中国流感疫苗应用现状及促进预防接种的政策探讨. 中华流行病学杂志,39(8):1045-1050 doi: 10.3760/cma.j.issn.0254-6450.2018.08.007 Peng Z B,Wang D Y,Yang J,et al. 2018. Current situation and related policies on the implementation and promotion of influenza vaccination,in China. Chinese J Epidemiol,39(8):1045-1050 (in Chinese) doi: 10.3760/cma.j.issn.0254-6450.2018.08.007
[6]	王伟. 2016. 2009—2014年山东省流感样病例流行特征及其与气象因素关系研究[D]. 济南: 山东大学. Wang W. 2016. Study on epidemic characteristics of influenza-like illness and relationship with meteoro-logical factors in Shandong province, 2009-2014 [D]. Jinan: Shandong University （in Chinese）
[7]	翟红楠,张莉,孙石阳等. 2009a. 深圳市流感高峰发生的气象要素临界值研究及其预报方程的建立. 数理医药学杂志,22(2):188-192 Zhai H N,Zhang L,Sun S Y,et al. 2009a. Critical range of Meteorological factors influence on the influenza peak in Shenzhen and the establishment of influenza forecasting equation. J Mathem Med,22(2):188-192 (in Chinese)
[8]	翟红楠,张莉,孙石阳等. 2009b. 深圳市流感就诊率季节特征及夏季流感就诊率气象预报模型. 气象科技,37(6):709-712 Zhai H N,Zhang L,Sun S Y,et al. 2009b. Seasonal characteristics of clinical rate of influenza-like illness in Shenzhen and its meteorological forecast model. Meteor Sci Technol,37(6):709-712 (in Chinese)
[9]	张静,杨维中,郭元吉等. 2004. 中国2001—2003年流行性感冒流行特征分析. 中华流行病学杂志,25(6):461-465 doi: 10.3760/j.issn:0254-6450.2004.06.001 Zhang J,Yang W Z,Guo Y J,et al. 2004. Epidemiologic characteristics of influenza in China,from 2001 to 2003. Chinese J Epidemiol,25(6):461-465 (in Chinese) doi: 10.3760/j.issn:0254-6450.2004.06.001
[10]	中国国家流感中心. 2017. 全国流感监测技术指南（2017年版）. Chinese National Influenza Center. 2017. Technical Guidelines for National Influenza Surveillance （in Chinese）
[11]	周艳丽,张海艳,王珺等. 2021. 北京市东城区流感样病例与气象因素的相关性分析. 中国公共卫生管理,37(5):650-653 Zhou Y L,Zhang H Y,Wang J,et al. 2021. Analysis of the relationship between influenza-like illness and meteorological factors in Dongcheng district,Beijing dity. Chinese J Public Health Manage,37(5):650-653 (in Chinese)
[12]	Alonso W J,Viboud C,Simonsen L,et al. 2007. Seasonality of influenza in Brazil:A traveling wave from the Amazon to the subtropics. Amer J Epidemiol,165(12):1434-1442 doi: 10.1093/aje/kwm012
[13]	Azziz Baumgartner E,Dao C N,Nasreen S,et al. 2012. Seasonality,timing,and climate drivers of influenza activity worldwide. J Infect Dis,206(6):838-846 doi: 10.1093/infdis/jis467
[14]	Barreca A I, Shimshack J P. 2012. Absolute humidity, temperature, and influenza mortality: 30 years of county-level evidence from the United States. Am J Epidemiol, 176 Suppl 7: S114-S122
[15]	Cheng X W,Tan Y,He M L,et al. 2013. Epidemiological dynamics and phylogeography of influenza virus in southern China. J Infect Dis,207(1):106-114 doi: 10.1093/infdis/jis526
[16]	Cheng Y H,Wang C H,You S H,et al. 2016. Assessing coughing-induced influenza droplet transmission and implications for infection risk control. Epidemiol Infect,144(2):333-345 doi: 10.1017/S0950268815001739
[17]	Chong K C,Goggins W,Zee B C Y,et al. 2015. Identifying meteorological drivers for the seasonal variations of influenza infections in a subtropical city—Hong Kong. Int J Environ Res Public Health,12(2):1560-1576 doi: 10.3390/ijerph120201560
[18]	Dai Q G,Ma W,Huang H D,et al. 2018. The effect of ambient temperature on the activity of influenza and influenza like illness in Jiangsu province,China. Sci Total Environ,645:684-691 doi: 10.1016/j.scitotenv.2018.07.065
[19]	Finkelman B S,Viboud C,Koelle K,et al. 2017. Global patterns in seasonal activity of influenza A/H3N2,A/H1N1,and B from 1997 to 2005:Viral coexistence and latitudinal gradients. PLoS One,2(12):e1296
[20]	Gasparrini A,Armstrong B. 2010. Time series analysis on the health effects of temperature:Advancements and limitations. Environ Res,110(6):633-638 doi: 10.1016/j.envres.2010.06.005
[21]	Gasparrini A,Leone M. 2014. Attributable risk from distributed lag models. BMC Med Res Methodol,14:55 doi: 10.1186/1471-2288-14-55
[22]	Gomez-Barroso D,León-Gómez I,Delgado-Sanz C,et al. 2017. Climatic factors and influenza transmission,Spain,2010-2015. Int J Environ Res Public Health,14(12):1469 doi: 10.3390/ijerph14121469
[23]	Huang X D,Mengersen K,Milinovich G,et al. 2017. Effect of weather variability on seasonal influenza among different age groups in Queensland,Australia:A Bayesian spatiotemporal analysis. J Infect Dis,215(11):1695-1701 doi: 10.1093/infdis/jix181
[24]	Liu T,Kang M,Zhang B,et al. 2018. Independent and interactive effects of ambient temperature and absolute humidity on the risks of avian influenza A (H7N9) infection in China. Sci Total Environ,619-620:1358-1365 doi: 10.1016/j.scitotenv.2017.11.226
[25]	Liu X X,Li Y H,Zhu Y B,et al. 2017. Seasonal pattern of influenza activity in a subtropical city,China,2010-2015. Sci Rep,7(1):17534 doi: 10.1038/s41598-017-17806-z
[26]	Lofgren E,Fefferman N H,Naumov Y N,et al. 2007. Influenza seasonality:Underlying causes and modeling theories. J Virol,81(11):5429-5436 doi: 10.1128/JVI.01680-06
[27]	Lowen A C,Steel J,Mubareka S,et al. 2008. High temperature (30°C) blocks aerosol but not contact transmission of influenza virus. J Virol,82(11):5650-5652 doi: 10.1128/JVI.00325-08
[28]	Moura F E A,Perdigão A C B,Siqueira M M. 2009. Seasonality of influenza in the tropics:A distinct pattern in Northeastern Brazil. Amer J Trop Med Hyg,81(1):180-183 doi: 10.4269/ajtmh.2009.81.180
[29]	Nimbalkar P M,Tripathi N K. 2016. Space-time epidemiology and effect of meteorological parameters on inﬂuenza-like illness in Phitsanulok,A northern province in Thailand. Geospat Health,11(3):447
[30]	Price R H M,Graham C,Ramalingam S. 2019. Association between viral seasonality and meteorological factors. Sci Rep,9(1):929 doi: 10.1038/s41598-018-37481-y
[31]	Shaman J,Goldstein E,Lipsitch M. 2011. Absolute humidity and pandemic versus epidemic influenza. Amer J Epidemiol,173(2):127-135 doi: 10.1093/aje/kwq347
[32]	Soebiyanto R P,Adimi F,Kiang R K. 2010. Modeling and predicting seasonal influenza transmission in warm regions using climatological parameters. PLoS One,5(3):e9450 doi: 10.1371/journal.pone.0009450
[33]	Soebiyanto R P,Clara W,Jara J,et al. 2014. The role of temperature and humidity on seasonal influenza in tropical areas:Guatemala,El Salvador and Panama,2008-2013. PLoS One,9(6):e100659 doi: 10.1371/journal.pone.0100659
[34]	Soebiyanto R P,Gross D,Jorgensen P,et al. 2015. Associations between meteorological parameters and influenza activity in Berlin (Germany),Ljubljana (Slovenia),Castile and León (Spain) and Israeli districts. PLoS One,10(8):e0134701 doi: 10.1371/journal.pone.0134701
[35]	Tamerius J D,Shaman J,Alonso W J,et al. 2019. Environmental predictors of seasonal influenza epidemics across temperate and tropical climates. PLoS Pathog,9(3):e1003194
[36]	Tang X J,Fang S S,Chiu A P Y,et al. 2018. Unsynchronized influenza epidemics in two neighboring subtropical cities. Int J Infec Dis,69:85-87 doi: 10.1016/j.ijid.2018.02.019
[37]	Taylor S J,Letham B. 2018. Forecasting at scale. Amer Stat,72(1):37-45 doi: 10.1080/00031305.2017.1380080
[38]	Wang X L,Yang L,He D H,et al. 2017. Different responses of influenza epidemic to weather factors among Shanghai,Hong Kong,and British Columbia. Int J Biometeorol,61(6):1043-1053 doi: 10.1007/s00484-016-1284-y
[39]	World Health Organization. 2017. Up to 650 000 people die of respiratory diseases linked to seasonal flu each year. Geneva: World Health Organization. https://www.who.int/news/item/13-12-2017-up-to-650-000-people-die-of-respiratory-diseases-linked-to-seasonal-flu-each-year
[40]	Ye C C,Zhu W P,Yu J X,et al. 2019. Understanding the complex seasonality of seasonal influenza A and B virus transmission:Evidence from six years of surveillance data in Shanghai,China. Int J Infect Dis,81:57-65 doi: 10.1016/j.ijid.2019.01.027
[41]	Yu H J,Alonso W J,Feng L Z,et al. 2013. Characterization of regional influenza seasonality patterns in China and implications for vaccination strategies:Spatio-temporal modeling of surveillance data. PLoS Med,10(11):e1001552 doi: 10.1371/journal.pmed.1001552
[42]	Zhang Y Z,Ye C C,Yu J X,et al. 2020. The complex associations of climate variability with seasonal influenza A and B virus transmission in subtropical Shanghai,China. Sci Total Environ,701:134607 doi: 10.1016/j.scitotenv.2019.134607