Impact of Weather on Incidence and Mortality of COVID-19: An Application of the ARIMAX Model

SARS-CoV-2 is primarily transmitted by direct contact between infected individuals, but other factors, such as meteorology, can affect mortality rates and the incidence of this disease. The purpose of this study was to examine the impact of meteorological factors on COVID-19 incidence and mortality in a center of Iran. In fact, this study sought to pursue two main goals: first, to find climate and air pollutant risk factors that seem to be related to people's respiratory conditions, and their effect on the number of daily cases and deaths caused by COVID-19, and the second one was to use the time series regression model as the appropriate model for such data instead of one-variable models.

Data collected over time can be modeled and forecasted using time series methods. It is common for time series models to be based on a single response variable, such as the Autoregressive Integrated Moving Average (ARIMA) model. In addition to the number of deaths and confirmed cases of COVID-19 as the response variable, we have also considered meteorological indices as independent variables. ARIMAX time series method was applied in this case.

The ARIMAX model was fitted in five lags (lag time in days). It was found that the average daily temperature in lag 10 and relative humidity in lag 7 were related to the mortality caused by COVID-19. The average visibility also had a significant and inverse relationship with the number of deaths in lag 14 and 7; this relationship was also observed with the number of confirmed cases, so in lag 3, as average visibility decreased, the number of cases increased.

It seems that some factors, such as temperature and severe storms, can affect the severity of the disease and should be considered in such conditions, especially for heart and respiratory patients. Thus, the necessary measures should be taken to reduce the severity of the infection with COVID-19 and the deaths caused by it.