Analysis And Prediction Of Influencing Factors Of Norovirus Based On Generalized Additive Model

Objective: This paper investigates the short-term effect of meteorological factors on norovirus load in sewage,predict the future norovirus load,and provide scientific reference for the prevention and early warning of norovirus infection in population.Methods: This paper based on the monthly monitoring data of norovirus load in sewage in Guangzhou from 2013 to 2018 and factors icalmeteorolog data in the same period.The generalized additive model was used to evaluate the nonlinearity and lag of meteorological factors on norovirus load in sewage.The single-factor and multi-factor model analysis was conducted in sequence,and the seasonal analysis was divided into different seasons.The ARIMA model was established to analyze the trend change of norovirus and predict the change content in the next year.Results: From January 2013 to December 2018,the norovirus load in sewage in guangzhou showed a seasonal trend,and the average monthly norovirus load was 13104736copies/ml(range 307480copies/ml to 131481040copies/ml).Norovirus load in sewage was significantly correlated with average temperature,average air pressure,relative humidity,average wind speed,rainfall and sunshine duration(p<0.05).Meteorological factors have different lag effects on norovirus load in sewage.Average temperature,rainfall lag 2months and the model is good.The ER value corresponding to temperature was-0.1407(-0.2657,0.0056),indicating the norovirus load increased by 1.15% for every 1℃ decrease in temperature.The ER value corresponding to rainfall was-0.9054(-0.9756,-0.6341),indicating that the norovirus load increased 2.47% for every 1 unit decrease in rainfall.Cold season and warm season are divided,and the ER value in cold season is-0.164(-0.3037,0.0031),and that in warm season is-0.19(-0.2870,-0.0809).The ER value of low rainfall is-0.394(-0.2035,0.0010),and that of high rainfall is 13.40(3.6458,89.3802).According to principal component regression,rainfall,average atmospheric pressure,average wind speed,average temperature,average vapor pressure,relative humidity and sunshine duration are main meteorological factors affecting norovirus load.Conclusions: 1.There is a close correlation between meteorological factors and norovirus load in sewage in Guangzhou,and the influence of average temperature and rainfall is more significant.2.The changes of norovirus load caused by different meteorological factors were different in the same month,1 month and 2 months,The virus changes the most when average temperature and rainfall lag by 2 months,respectively 3.The fitting effect of ARIMA model is ideal and has high reliability.Results show that the norovirus load in Guangzhou in 2019 is slightly lower than that in 2018.