The role of aquatic reservoir fluctuations in long-term cholera patterns.

We propose and analyze an important extension of standard cholera epidemiological models, explicitly accounting for fluctuations of water availability to the human community under study. The seasonality of water input in the reservoir drives the variation of concentration of Vibrio cholerae. Two compartments are added to the Susceptible-Infected-Bacteria model. First, the recovered individuals, which, over many seasons, lose their immunity to the disease and replenish the Susceptible group. Second, the water volume of the reservoir, which determines bacterial dilution and, consequently, the probability of contracting cholera by ingesting contaminated water. By forcing the model with a seasonally varying hydrologic input, we obtain simulations that can be compared to available data for various regions of the World characterized by different hydrological and epidemiological regimes. The model is shown to satisfactorily reproduce important characteristics of disease insurgence and long-term persistence. Using bifurcation analysis of nonlinear systems, we also explore how different degrees of seasonality and values of the basic reproductive number can change the expected long-term epidemiological time series. We find that there exist parametric conditions where the model shows chaotic patterns - i.e. high unpredictability especially in the amplitude of prevalence peaks - which very much resemble actual data on long-term cholera insurgence.