PARENT SESSION
Wednesday, August 9, 1:30-5:00 pm
COS 73 - Ecological modeling III
L-11, Lobby Level, Cook Convention Center
Presiders: M Kenney and S Hoeppner

Simple stochastic time series models in ecology and infectious diseases.

Alonso, David^*,1, Pascual, Mercedes¹, Smith, David², ¹ University of Michigan, Ann Arbor, MI² National Institute of Health, Washington, DC

ABSTRACT- Stochastic time series models with explicit description of nonlinear ecological interactions are an important yet challenging approach to obtain parameters, test hypotheses via model selection, and evaluate predictability in population dynamics. Recent work has proposed one such model (known as TSIR) for Susceptible-Infectious-Recovered dynamics in an application to measles. We present here a general framework to build and test mechanistic time series models with demographic noise. We briefly illustrate the approach with the case of simple density-dependent population growth, and show how it provides analytical basis for the distributions arising in the TSIR model. We further illustrate the application to infectious disease dynamics with two models of increasing complexity for malaria, used to test different hypotheses on the processes underlying the recent exacerbation of the disease in African highlands.

Key words: stochastic time series, model selection, TSIR, TSIC, malaria