Predictability of Atmospheric Conditions Potentially Favorable for COVID-19 Transmissibility

Abstract:

Despite predictions of epidemiological models for southern shift of the epicenter of COVID-19 from Wuhan to Bangkok, the disease extended longitudinally in a narrow latitudinal band. This band was characterized by surface temperature and humidity within a well defined range. Based on this observation we suggested the hypothesis that atmospheric conditions may provide an epidemiological background where COVID-19 has the potential to proliferate (Sajadi et al., Jama Network Open 2020). Specific society based control measures e.g., stay at home orders, can be very effective in decreasing the number of new cases, even in areas with potentially favorable atmospheric conditions for COVID-19 spread. However, the decision making process for imposing such drastic measures is complex. Factors ranging from the job loss, mental health, and hospital capacity have to be taken into account. Knowledge of the environmental potential for proliferation of COVID-19 may thus help this decision making process.

In this paper we discuss methodologies for monitoring and forecasting atmospheric conditions that are potentially favorable for COVID-19. We first focus on quantifying the information content added by the use of forecast models versus use of climatological seasonality. We then explore the predictability of the combined temperature and humidity conditions that are favorable for COVID-19 for several forecast lead times. We conclude the presentation by introducing an experimental real-time forecasting system and discuss the next research and development steps for improving this system.