Impact of key meteorological parameters on the spread of COVID-19 in Mumbai: Correlation and Regression Analysis

Abstract

Purpose: To understand key meteorological parameters that influence the spread of COVID-19 in Mumbai, India (based on data from April 2020 - April 2021). Methods: The meteorological parameters chosen were Temperature, Dew Temperature, Humidity, Pressure, Wind Speed. The underlying basic relationships between meteorological parameters and COVID-19 information for Mumbai was understood using Spearman's rank correlation coefficients. After establishing basic relationships, Linear analysis and Generalized Additive Model's (GAM) were used to figure out statistically significant weather parameters and model them to explain the best possible variance in the pandemic data. Results: A model of temperature and windspeed could explain 17.3% and 8.3% of variance in Daily new cases and Daily recoveries respectively. As for deaths occurring due to the virus, a model comprising of only pressure best explains a variance of 17.3% in the data. Non-Linear modelling based on GAM confirms the findings of linear analysis and establishes certain non-linear relationships as well. Conclusion: SARS-CoV-2 belongs to the class of Human Coronaviruses (HCoV) which show seasonality depending on weather conditions. The above article focuses on understanding the underlying relationships between SARS-CoV-2 and meteorological parameters that would help progress basic research and formulation of policies around the disease for each weather/season.

Temporal relation between human mobility, climate, and COVID-19 disease

“…While almost all recent studies focus on the direct analysis, the indirect effect is certainly relevant and whose importance should be quantified. In this context, it has been suggested that specific meteorological variables may play a role in the spread of the coronavirus in some countries Wang et al (2020b); Alvarez-Ramirez and Meraz (2020); Bannister-Tyrrell et al (2020); Bukhari and Jameel (2020); Ma et al (2020); Sajadi et al (2020); Wang et al (2020a); Araujo and Naimi (2020); Carleton and Meng (2020); Ficetola and Rubolini (2020); Islam et al (2021);Loché Fernández-Ahúja and Fernández Martínez (2021); Chelani and Gautam (2022); Shetty et al (2022); Souza Júnior et al (2022); Briz-Redón and Serrano-Aroca (2020); Ogunjo et al (2022); Damette et al (2021); Wei et al (2022), while others oppose to strong seasonal effects O'Reilly et al (2020); Jamil et al (2020). There are also proposals that several meteorological factors combined could describe the epidemic trend much better than single-factor models Chen et al (2020).…”