A multilayer perspective for inferring spatial and social functioning in animal movement networks

Abstract

211. Animal movement patterns are increasingly analysed as spatial networks. Currently, structures 22 of complex movements are typically represented as a single-layer (or monoplex) network. 23However, aggregating individual movements, to generate population-level inferences, 24 considerably reduces information on how individual or species variability influences spatial 25 connectivity and thus identifying the mechanisms driving network structure remains difficult. 26 2. Here, we propose incorporating the recent conceptual advances in multilayer network analyses 27 with the existing movement network approach to improve our understanding of the complex 28 interaction between spatial and/or social drivers of animal movement patterns. 29 3. Specifically, we explore the application and interpretation of this framework using an empirical 30 example of shark movement data gathered using passive remote sensors in a coral reef 31 ecosystem. We first show how aggregating individual movement networks can lead to the loss 32 of information, potentially misleading our interpretation of movement patterns. We then apply 33 multilayer network analyses linking individual movement networks (i.e. layers) to the 34 probabilities of social contact between individuals (i.e. interlayer edges) in order to explore the 35 functional significance of different locations to an animal's ecology. 36 4. This approach provides a novel and holistic framework incorporating individual variability in 37 behaviour and inter-individual interactions. We discuss how this approach can be used in 38 applied ecology and conservation to better assess the ecological significance of variable space 39 use by mobile animals within a population. Further, we argue that the uptake of multilayer 40 networks will significantly broaden our understanding of long-term ecological and evolutionary 41 processes, particularly in the context of information or disease transfer between individuals. 42 43 Keywords 44 animal tracking; connectivity; graph theory; multilayer; spatial network; spreading processes; 45 telemetry 46 47

Multiple spatial behaviours govern social network positions in a wild ungulate

“…Several frameworks have been proposed to facilitate the untangling of spatial and social 49 processes in wild animals (Jacoby & Freeman 2016;Silk et al 2018Silk et al , 2019Webber & 50 Vander Wal 2018; Mourier et al 2019). To date, associated statistical methodology focusses 51 on incorporating spatial activity into the node-and-edge structure of network data, including 52 e.g.…”

“…To date, associated statistical methodology focusses 51 on incorporating spatial activity into the node-and-edge structure of network data, including 52 e.g. null network permutations (Firth & Sheldon 2016), spatially embedded networks 53 (Daraganova et al 2012), and nested super-networks composed of movement trajectories 54 (Mourier et al 2019). Many such analyses involve reducing movement patterns into some 55 form of spatial network based on home range overlap or spatial proximity between dyads 56 (Mourier et al 2019).…”

“…null network permutations (Firth & Sheldon 2016), spatially embedded networks 53 (Daraganova et al 2012), and nested super-networks composed of movement trajectories 54 (Mourier et al 2019). Many such analyses involve reducing movement patterns into some 55 form of spatial network based on home range overlap or spatial proximity between dyads 56 (Mourier et al 2019). For example, statistical models named "animal models" can examine 57 spatial variation by fitting such matrices as variance components, potentially alongside other 58 dyadic similarity matrices, to quantify genetic and non-genetic contributions to individuals' 59 phenotypes (Kruuk 2004;Stopher et al 2012b; Regan et al 2016; Thomson et al 2018;60 Webber & Vander Wal 2018).…”

“…For example, statistical models named "animal models" can examine 57 spatial variation by fitting such matrices as variance components, potentially alongside other 58 dyadic similarity matrices, to quantify genetic and non-genetic contributions to individuals' 59 phenotypes (Kruuk 2004;Stopher et al 2012b; Regan et al 2016; Thomson et al 2018;60 Webber & Vander Wal 2018). However, movement paths are highly multivariate, and dyadic 61 home range overlaps represent just one aspect of spatial behaviour (Mourier et al 2019; Pasquaretta et al 2020). As such, controlling for structuring using space sharing alone 63 potentially risks missing important variation.…”

A multilevel statistical toolkit to study animal social networks: the Animal Network Toolkit Software (ANTs) R package

“…In recent years the use of network theory has developed considerably in animal research ( Fig. 1), facilitating the study of relationships between individuals and in a variety of research fields, including genetic 1 , spatial movements 2 , animal groups and species assemblages 3 , epidemiology 4 , information transmission 5 , ethology, sociobiology 6 , evolution 7 , conservation 8 and animal management 9 . Animal social networks are also dynamic as they change through time based on group composition and social choices/assemblages of individuals [10][11][12][13] .…”