Poli, C. L. | Search bioRxiv

bioRxiv2025

MONITORING POPULATION EXTINCTION RISK WITH COMMUNITY SCIENCE DATA

AbstractThe robust estimation of local extinction risk is central to inform management and conservation efforts. Still, estimating this key demographic parameter requires standardized monitoring data that are lacking for most species and systems. The analysis of community science data is emerging as a promising alternative. These expansive datasets leverage observations from multiple volunteers that provide higher temporal and spatial resolution. Nevertheless, the proper analysis of community science data is challenging because it requires accounting for additional complexities in the intrinsic ecological and observational processes. To address this issue, we describe and test a quantitative approach that fits continuous state-space models iteratively to eBird data with the ultimate goal of estimating local persistence probability through time. We evaluated model accuracy by comparing estimates and trends from eBird with those from the endangered Everglades snail kite long-term, standardized monitoring project. We also performed two separate sensitivity analyses (temporal and sampling thinning) to assess how robust the persistence estimates are to a reduction in the number of eBird observations available. Our results showed that the temporal trend trajectory of local population persistence estimated from eBird closely matched that from standardized monitoring. Moreover, the trend remained similar even when reducing the amount of eBird data available to 5% of the original data set - a reduction from 258 to 13 weeks or from 7,714 to 385 lists of observations across 5 years of monitoring. Synthesis and applicationsOur modeling framework provides a robust, computationally efficient, and easy-to-apply tool for monitoring local persistence probability that can support global conservation efforts. This will complement the monitoring of species population viability in places where standardized monitoring is still lacking, but community science observations are common.

ecology↗

bioRxiv2020

Identifying latent behavioral states in animal movement with non-parametric Bayesian methods

O_LIUnderstanding animal movement often relies upon telemetry and biologging devices. These data are frequently used to estimate latent behavioral states to help understand why animals move across the landscape. While there are a variety of methods that make behavioral 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 behavioral states. C_LIO_LITo address some of the limitations of existing behavioral state estimation models, we introduce a non-parametric Bayesian framework called the mixed-membership method for movement (M4), which is available within the open-source bayesmove R package. This framework can analyze multiple data streams (e.g., step length, turning angle, acceleration) without relying on parametric distributions, which may capture complex behaviors more successfully than current methods. We tested our Bayesian framework using simulated trajectories and compared model performance against two segmentation methods (behavioral change point analysis (BCPA) and segclust2d), one machine learning method (expectation-maximization 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. C_LIO_LIThe 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 behavior than the other state estimation methods when simulations were generated from less frequently considered distributions (e.g., truncated normal, beta, uniform). Three behavioral states were estimated from snail kite movements, which were labeled as encamped, area-restricted search, and transit. Changes in these behaviors over time were associated with known dispersal events from the nest site, as well as movements to and from possible breeding locations. C_LIO_LIOur non-parametric Bayesian framework estimated behavioral 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 distributions can provide powerful alternatives to address questions in movement ecology. C_LI

ecology↗