Improving COVID-19 forecasts by accounting for seasonality and environmental responses

Policy-makers urgently need medium- and long-term forecasts for SARS-CoV-2, but we are currently ignorant of the virus' seasonal dynamics. In the absence of data, forecasters have been forced to assume that SARS-CoV-2 will have identically strong environmental and seasonal responses to other coronaviruses[1]. We will address this important source of uncertainty by measuring changes in transmission across climatic gradients (e.g., temperature and humidity) to forecast seasonal responses. We will do so by leveraging classical ecological theory related to niche modelling (the measurement of species' environmental tolerances) and space-for-time substitution (using variation across space to predict variation through time). We will integrate this approach into existing forecasting models developed by the Imperial College London COVID-19 Response Team, which have informed the public-health response worldwide. This will allow mitigation strategy to account for climate-driven spatial changes in transmission, highlighting where and when stronger interventions are needed. As SARS-CoV-2 is evolving, and so potentially adapting, as it spreads, we will also conduct preliminary work to assess the extent to which the virus is adapting in situ to environmental conditions. This project will provide new insight into the relationships between SARS-CoV-2 transmission rates, seasonality, and environmental factors, which will inform the development of targeted optimal intervention strategies against COVID-19. We have prioritised our efforts around delivering forecasts before autumn and winter, leveraging an inter-disciplinary team of epidemiologists, quantitative ecologists, evolutionary biologists, and bioinformaticians. [1] Kissler et al. (2020) Science eabb5793.