Forecasting Infectious Diseases both with and without Climate Forcing

The spread and proliferation of infectious diseases within a host population proceeds as a function of both intrinsic system dynamics (e.g. pathogen replication rates, host interactions, host-vector interactions, etc.) and environmental conditions. Models depicting and forecasting infectious disease outcomes often represent the former as an initial value problem and the latter as a boundary value problem. Here we show that long-lead predictions (>3 months) for a number of diseases are possible using climate information and seasonal climate forecasts as boundary forcing; however, for more precise, shorter lead forecasting (<3 months), models that depict the intrinsic nonlinear dynamics of disease transmission are often needed. These shorter lead forecasts can sometimes be improved through the additional incorporation of weather and climate forcing into these model structures. Comprehensive early warning systems should utilize both long-lead and short-lead forecasting approaches in order to provide predictions over a range of times scales.