Modeling Disease Dynamics in Italian Livestock: A Multilayer Analysis of Cattle and Buffalo Movement Network

Epidemics are a significant issue that affects various aspects of our lives, both in terms of public health and the global economy. For this reason, several mathematical models have been developed, such as compartmental models, agent-based models, and network-based models, to study the spread of diseases and the role that individual or community behaviors play in them. This thesis primarily focuses on the Italian cattle and buffalo movement network and the associated risk of disease spread. When applying the model mentioned above to this context, several challenges immediately arise. Indeed, the Italian farms exhibit certain characteristics that are often overlooked and understudied. Pasture movements are the seasonal displacement of cattle to shared mountain pastures during the summer months. While movements between herds must be recorded, this is not the case for movements between two pastures, complicating both the analysis and the tracking of contacts. On the other hand, several farms raise both buffalo and cattle in the same spaces, facilitating, through proximity, disease transmission between two different species. The presence of buffalo is often overlooked in studies of the Italian network. Still, they could have a significant role, acting as a reservoir for pathogens that can also infect cattle. Understanding these unique characteristics is crucial to improving disease modeling and implementing effective containment or treatment measures. In this thesis, we will first aim to observe the impact of cattle movements to pastures on the spread of pathogens and explore ways to improve the available data, highlighting existing issues. Subsequently, we will investigate whether the Italian buffalo network, which has yet to be studied, could sustain the spread of a pathogen. Finally, we will assess whether ignoring the potential cattle-buffalo interaction might lead to an underestimation of epidemic risk. To achieve this, we will use existing animal movement data to create a contact network between different farms. These farms will be the epidemiological units of interest instead of individual animals. Using a combination of the three models mentioned above, we will simulate the spread of a pathogen to test our hypothesis. Additionally, for the cattle-buffalo interaction, we will adapt the concept of a multilayer network, where each layer will represent the movements of a single species, and the connections between layers will describe the interactions occurring within a single farm. Our findings emphasize the importance of data quality in modeling disease spread in animal populations. We conclude that both movements to pastures and the dynamics of farms raising both buffalo and cattle play a significant role that should be better considered and studied by improving data collection processes beforehand. Furthermore, a comprehensive approach appears crucial for disease control, as focusing on a single species for certain pathogens could lead to underestimations of the actual risk. Future work could incorporate various environmental factors or wildlife interactions to provide an even more complete representation