Accounting for Detection Heterogeneity and Host Movements in a House Finch-Mycoplasma gallisepticum Disease System

Abstract

During the course of my dissertation research, I made use of capture-recapture methods to investigate local house finch (Carpodacus mexicanus) demography and movements in the context of understanding seasonal Mycoplasma gallisepticum (MG) infection dynamics. Capture-recapture design, estimation, and modeling explicitly accounts for variable detectability of individuals and provides a framework for making multi-model inference, thereby incorporating inherent model selection uncertainty (via Akaike information criterion) into the inferential process. The biological focus throughout my research has generally been centered on the relationship between local spatial scale host population structure, movements, and host-pathogen dynamics. Broadly, my work illustrates the importance of accounting for animal detection probabilities when estimating epidemiological statistics and parameters. I also highlight the importance of considering different forms of animal movements (either biologically induced or as a consequence of sampling design) with respect to understanding dynamics in the finch-MG system (specifically), but also applicable to other host-pathogen systems (generally). I estimate host transient movements, completely observable within-study area movements, proportional recruitment, and temporary movements from the study area (representing partially observable movements); all of which are very important elements to consider for understanding the dynamics of highly mobile animal populations (especially in the presence of a virulent pathogen). My research, conducted at a local spatial scale in Ithaca, NY complements analyses using House Finch Disease Survey data (Dhondt et al. 1998) at a broader spatial scale, and provides a point of entry for understanding the critical linkage of scale dependent processes influencing finch-MG dynamics. Throughout this dissertation, I have sought to characterize the structure of this local finch population, and establish how both host population structure and movements lead to a better overall understanding of MG infection dynamics. As such, the complete body of work produced here represents the most comprehensive investigation of wildlife disease dynamics to date, which has incorporated and accounted for sampling and biologically driven heterogeneity in host encounter probabilities. Beyond the proximate benefits that this research contributes to understanding of the finch-MG system, my hope is that this work will in part serve as a precedent for future empirical investigations of wildlife-pathogen dynamics.