Nonlinear effects of noise on outbreaks of mosquito-borne diseases

Abstract

Mosquito-borne diseases are a significant and growing public health burden globally. Predictions about the spread and impact of mosquito-borne disease outbreaks can help inform direct control and prevention measures. However, climate change is expected to increase weather variability, potentially shaping the future of mosquito-borne disease outbreaks globally. In this study, we sought to determine the effects of demographic and environmental noise (stochasticity) on the duration and size of outbreaks predicted by models of mosquito-borne disease. We developed a demographically and environmentally stochastic Ross-Macdonald model to assess how noise affects the probability of an outbreak, the peak number of cases, and the duration of outbreaks at increasing levels of the basic reproduction number (R0) and environmental noise strength. Increasing environmental noise reduces the risk of endemic disease from 100% down to almost 0%, but the largest outbreaks occur at intermediate environmental noise levels. In this case, if an outbreak dies out, it ends quickly. In the presence of noise, R0 alone is insufficient to definitively predict whether an outbreak occurs. Surprisingly, our modelling results suggest that the dramatic effect on mosquito populations from increases in the frequency of extreme environmental conditions could reduce the risk of endemic disease and epidemics in some settings.