Identifying latent behavioural states in animal movement with M4, a nonparametric Bayesian method

Abstract

1. Understanding animal movement often relies upon telemetry and biologging devices. These data are frequently used to estimate latent behavioural states to help understand why animals move across the landscape. While there are a variety of methods that make behavioural inferences from biotelemetry data, some features of these methods (e.g. analysis of a single data stream, use of parametric distributions) may limit their generality to reliably discriminate among behavioural states. 2. To address some of the limitations of existing behavioural state estimation models, we introduce a nonparametric Bayesian framework called the mixed-membership method for movement (M4), which is available within the open-source bayesmove R package. This framework can analyse multiple data streams (e.g. step length, turning angle, acceleration) without relying on parametric distributions, which may capture complex behaviours more successfully than current methods.We tested our Bayesian framework using simulated trajectories and compared model performance against two segmentation methods (behavioural change point analysis (BCPA) and segclust2d), one machine learning method [expectationmaximization binary clustering (EMbC)] and one type of state-space model [hidden Markov model (HMM)]. We also illustrated this Bayesian framework using movements of juvenile snail kites Rostrhamus sociabilis in Florida, USA.3. The Bayesian framework estimated breakpoints more accurately than the other segmentation methods for tracks of different lengths. Likewise, the Bayesian framework provided more accurate estimates of behaviour than the other state estimation methods when simulations were generated from less frequently considered distributions (e.g. truncated normal, beta, uniform). Three behavioural states were estimated from snail kite movements, which were labelled as 'encamped', 'area-restricted search' and 'transit'. Changes in these behaviours over time were associated with known dispersal events from the nest site, as well as movements to and from possible breeding locations. 4. Our nonparametric Bayesian framework estimated behavioural states with comparable or superior accuracy compared to the other methods when step lengths and turning angles of simulations were generated from less frequently considered distributions. Since the most appropriate parametric distributions may not be obvious a priori, methods (such as M4) that are agnostic to the underlying | 433Methods in Ecology and Evoluঞon CULLEN Et aL.

Biologging as an important tool to uncover behaviors of cryptic species: an analysis of giant armadillos (Priodontes maximus)

“…We relied on activity count, speed (as calculated above), and turning angle for the estimation of latent behavioral states using the non-parametric Bayesian mixture model for movement (M3; Valle et al, 2022 ). Each variable was discretized into bins as required by the model, where the number and width of bins were selected to characterize the continuous distributions in as few bins as necessary ( Cullen et al, 2022 ). Activity count ranged from 1 to 300 and was discretized into six bins of equal width.…”

“…The three discretized variables ( i.e., activity count, speed, and turning angle) were analyzed using M3 via the package ‘bayesmove’ v0.2.0 ( Cullen et al, 2022 ) in the R statistical software (v4.0.2; R Core Team, 2020 ). This model clusters observations (pooled across all individual tracks) into an unknown number of discrete latent behavioral states.…”

Burmese pythons in Florida: A synthesis of biology, impacts, and management tools

“…For example, accelerometers can be paired with GPS tags to combine data collection on fine-scale behavior with large-scale movements (Whitney et al 2021). Advanced behavioral modeling algorithms are also emerging (e.g., Jeantet et al 2020;Cullen et al 2022), which could reduce the need to simultaneously observe individual behaviors in animals equipped with biologging devices. However, novel statistical analyses are needed to extract patterns from data streams coming from multiple sensors with differing temporal resolutions (McClintock et al 2017) before simultaneous observations are not needed.…”

abmR: An R package for agent‐based model analysis of large‐scale movements across taxa

“…movements in predicted future land-use changes) while studying the emerging properties of the system (Yin et al, 2022). Finally, abmR functionalities can be integrated with other approaches, such as a step-selection function (Thurfjell et al, 2014) and/or Bayesian statistics to estimate movement parameters and explore migration patterns or habitat occupancy of a particular set of agents (Cullen et al, 2022;Joo et al, 2020).…”