The effect of temporal resolution and contact duration on Real-Time location system-based contact networks for confined feedlot cattle

Abstract

Empirical data on livestock contact networks are scarce but digital technologies are increasingly used to characterize animal behavior and describe the dynamics of contact networks. The objective of this study was to use contact network analysis to quantify contacts within three pens of feedlot cattle across three consecutive years at varying temporal resolutions to better inform the construction of network-based disease transmission models for cattle within confined-housing systems. We also aimed to describe the influence of the variation in Real-Time Location System (RTLS) average tag read rates and the effect of increasing minimum contact duration (MCD) on the contact networks of feedlot cattle. Three pens of feedlot steers were tagged with RTLS, one pen in each of three consecutive summers from 2017 to 2019. Contacts were defined with a spatial threshold of 0.71 m and an MCD of either 10, 30, or 60 seconds. Static, undirected, weighted contact networks were created for the full study duration and then split into daily (24-h), 6-h period, and hourly networks to better assess network heterogeneity. For the full study duration time scale, all three networks were found to be densely connected. The networks showed more heterogeneity in network density and clustering coefficient when smaller time scales (6-h period and hourly) were applied. When contacts were defined with a MCD of 30 or 60 seconds, the total number of contacts seen in each network decreased, indicating that most of the contacts observed in our networks may have been transient passing contacts. For example, the total defined contacts for the 2017 native read rate network were 930,843 at 10 s MCD, 95,570 at 30 s MCD, and 19,135 at 60 s MCD. Though the same system was used for all three years, variation in average tag read rate was observed (range: 2.2 readings/min (2018) to 7.4 readings/min (2017)). When the networks were down-scaled from higher average tag read rates to match lower tag read rates, the full study networks maintained similar network density and clustering, though the average edge weight between pairs decreased. Overall, the networks created here from high-resolution spatial and temporal contact observation data provide estimates for a contact network within a commercial US feedlot pen.