Identifying stationary phases in multivariate time series for highlighting behavioural modes and home range settlements

Abstract

1. Recent advances in biologging open promising perspectives in the study of animal movements at numerous scales. It is now possible to record time series of animal locations and ancillary data (e.g. activity level derived from on-board accelerometers) over extended areas and long durations with a high spatial and temporal resolution. Such time series are often piecewise stationary, as the animal may alternate between different stationary phases (i.e. characterized by a specific mean and variance of some key parameter for limited periods). Identifying when these phases start and end is a critical first step to understand the dynamics of the underlying movement processes. 2. We introduce a new segmentation-clustering method we called segclust2d (available as a r package at cran.r-proje ct.org/packa ge=segcl ust2d). It can segment bivariate (or more generally multivariate) time series and possibly cluster the various segments obtained, corresponding to different phases assumed to be stationary. This method is easy to use, as it only requires specifying a minimum segment length (to prevent over-segmentation), based on biological rather than statistical considerations. 3. This method can be applied to bivariate piecewise time series of any nature. We focus here on two types of time series related to animal movement, corresponding to (a) at large scale, series of bivariate coordinates of relocations, to highlight temporary home ranges, and (b) at smaller scale, bivariate series derived from relocations data, such as speed and turning angle, to highlight different behavioural modes such as transit, feeding and resting. 4. Using computer simulations, we show that segclust2d can rival and even outperform previous, more complex methods, which were specifically developed to highlight changes of movement modes or home range shifts (based on hidden Markov and Ornstein-Uhlenbeck modelling), which, contrary to our method, usually require the user to provide relevant initial guesses to be efficient. Furthermore, we demonstrate it on actual examples involving a zebra's small-scale movements and an elephant's large-scale movements, to illustrate how various movement modes and home range shifts, respectively, can be identified.

Identifying latent behavioral states in animal movement with M4, a non-parametric Bayesian method

“…Unlike existing methods that assign a single behavior to observations or track segments, LDA classifies each track segment as a combination of multiple behaviors (Blei et al 2003; Valle et al 2014). This conceptually fits with our notion that track segments of animal movement may not be entirely comprised of a single behavior (Pohle et al 2017; Patin et al 2020). However, LDA can also classify track segments as belonging to a single behavior if that is supported by the data (Valle et al 2014).…”

High fidelity of sea turtles to their foraging grounds revealed by satellite tracking and capture-mark-recapture: New insights for the establishment of key marine conservation areas

“…For each of the 24 satellite-tracked turtles that remained in the area for at least 40 days, we checked the stationarity of the distribution of locations using a new segmentation method (Patin et al, 2020; see Supplementary method S1). The corresponding Utilisation Distributions (UD) were estimated using the Kernel Density Estimation (KDE) method as implemented in the adehabitatHR R package (Calenge, 2006) with an ad hoc smoothing parameter (Kie, 2013).…”

Biologging Special Feature

“…Animal movements are fundamentally characterized by facultative switches between distinct movement modes (Fryxell et al, 2008) and many methods have been developed to identify and segment movement paths into different behavioural sections (Barraquand & Benhamou, 2008;Beyer, Morales, Murray, & Fortin, 2013;Edelhoff, Signer, & Balkenhol, 2016;Gurarie et al, 2015;Leos-Barajas et al, 2017;Michelot & Blackwell, 2019;Wang, 2019), where issues of scale and the difference between stationary and non-stationary movements are of particular importance (Benhamou, 2014). Here, Patin et al (2020) contribute to this growing literature by extending the K-segmentation approach of Lavielle (2005) to identify breakpoints in time-series of biologging data (or more generally any multivariate time-series) and potentially categorize resulting segments into common groups based on similarities in data characteristics.…”

“…These GPS-based papers investigate predator-prey spatiotemporal interactions among elk Cervus canadensis and wolf Canis lupus (Cusack et al, 2020), quantify foraging niche overlap between sympatric seabird species (Dehnhard et al, 2020), or assess effects of personality on the consistency and repeatability of foraging trips in black-legged kittiwakes Rissa tridactyla (Harris et al, 2020). Other contributions present novel statistical methods to estimate individual variation in habitat selection (Muff, Signer, & Fieberg, 2020) or to identify different movement modes in movement tracks (Patin, Etienne, Lebarbier, Chamaillé-Jammes, & Benhamou, 2020), whereas other studies use fine-scale movement data to quantify the impact of wind turbines on functional habitat loss of a soaring terrestrial bird, the black kite Milvus migrans (Marques et al, 2020), or identify mating tactics of male African elephants Loxodonta africana (Taylor et al, 2020).…”