Network analysis of dairy cattle movements in Ontario to support livestock disease simulation modelling

Date

Authors

Dubé, Caroline

Journal Title

Journal ISSN

Volume Title

Publisher

University of Guelph

Abstract

This thesis was an investigation of the movement of dairy cattle in Ontario in order to improve the realism of simulated epidemics of foot-and-mouth disease (FMD). Canada does not have a livestock movement database to develop parameters for simulations. We evaluated the Dairy Herd Improvement (DHI) database, which stores cow lactation records, as a source of movement information for adult milking cows in Ontario. We used social network analysis techniques to study the structure of the networks formed among DHI farms in Ontario resulting from the movements of adult milking cows in 2004-2006. We showed how the topology of such networks influenced how diseases could spread among farms. We used existing network analysis measures to estimate potential epidemic size at first detection of FMD and compared these estimates to those provided by the infection chain. The infection chain is a modified breadth-first search that we consider the most biologically sensible measure for estimating potential epidemic size in both connected and fragmented networks of direct animal movements. The DHI network information was then used to validate the contact structure represented in the North American Animal Disease Spread Model (NAADSM). Our findings showed that the current parsimonious approach in NAADSM would represent up to 99% of monthly farm-to-farm networks. However, the current approach does not represent extreme situations in which some farms might have a high number of contacts. We proposed alternatives to consider for future versions of NAADSM. We also studied the movements of dairy cattle through two large markets in Ontario in the year 2004. Markets played a central role in the dairy industry as they are the main point for selling slaughter cows and calves. Our study demonstrated the need for quickly closing market sales at first detection of FMD in the Province. In addition, we proposed that tracing efforts be focussed upon determining if infected animals have gone through markets. This thesis shows the importance of obtaining and maintaining livestock movement information and the benefits of using network analysis. However, the rules that define certain network analysis measures must be clearly understood to be useful.

Description

Keywords

network analysis, dairy cattle, movement, Ontario, livestock disease, simulation, modelling, foot-and-mouth disease

Citation