Forcing Versus Feedback: Epidemic Malaria and Monsoon Rains in Northwest India

Karina Laneri, Anindya Bhadra, Edward L. Ionides, Menno Bouma, Ramesh C. Dhiman, Rajpal S. Yadav, Mercedes Pascual

Abstract: Author Summary Malaria epidemics can exhibit pronounced variation from year to year that can be driven by external forcings, such as climate, or can be generated instead by dynamic feedbacks within the disease system itself. For example, levels of immunity in the population (or control efforts) can rise and fall as the result of past levels of infection. This type of feedback is found in the dynamics of all (nonlinear) biological systems. Feedbacks can interact in complex ways with external drivers, for example by creating refractory periods. It remains a challenge to identify internal feedbacks vs. external forcings from available temporal records of aggregated reported cases and forcing variables. We propose a quantitative approach that can statistically compare the hypotheses of feedbacks vs. forcings (epidemiological vs. climate) based on dynamical and mechanistic models. Our approach is computational, based on a large number of computer simulations of the different models. We illustrate and apply the approach to the analysis of extensive monthly records for malaria incidence in desert regions of India that span two decades. Our analyses confirm the strong role of rainfall, and quantify this effect with transmission model(s) for malaria that include rainfall and are shown to exhibit a remarkable prediction skill.