Data from: From fine-scale foraging to home ranges: a semi-variance approach to identifying movement modes across spatiotemporal scales

Understanding animal movement is a key challenge in ecology and conservation biology. Relocation data often represent a complex mixture of different movement behaviors, and reliably decomposing this mix into its component parts is an unresolved problem in movement ecology. Traditional approaches, such as composite random walk models, require that the timescales characterizing the movement are all similar to the usually arbitrary data-sampling rate. Movement behaviors such as long-distance searching and fine-scale foraging, however, are often intermixed but operate on vastly different spatial and temporal scales. An approach that integrates the full sweep of movement behaviors across scales is currently lacking. Here we show how the semivariance function (SVF) of a stochastic movement process can both identify multiple movement modes and solve the sampling rate problem. We express a broad range of continuous-space, continuous-time stochastic movement models in terms of their SVFs, connect them to relocation data via variogram regression, and compare them using standard model selection techniques. We illustrate our approach using Mongolian gazelle relocation data and show that gazelle movement is characterized by ballistic foraging movements on a 6-h timescale, fast diffusive searching with a 10-week timescale, and asymptotic diffusion over longer timescales.

解析动物运动模式是生态学与保护生物学领域的核心难题之一。动物重定位数据往往混杂着多种不同的运动行为模式，而将这类混合行为可靠地拆解为各自独立的组分，仍是运动生态学中尚未解决的关键问题。传统研究方法例如复合随机游走模型（composite random walk models），要求表征运动的时间尺度均需与通常为任意设定的数据采样速率相匹配。然而诸如长距离搜索与精细尺度觅食这类运动行为，往往相互交织，却在截然不同的空间与时间尺度上展开。目前仍缺乏能够整合全尺度范围内所有运动行为的研究方法。本研究阐明，随机运动过程的半方差函数（semivariance function, SVF）既可识别多种运动模式，又能解决采样率适配难题。我们通过半方差函数表征了涵盖连续空间、连续时间的各类随机运动模型，并通过变异函数回归将这些模型与重定位数据关联，最终借助标准模型选择技术完成模型对比。我们以蒙古原羚的重定位数据为例演示了该方法，结果表明原羚的运动模式可分为三类：6小时尺度下的弹道式觅食运动、10周尺度下的快速扩散搜索运动，以及更长时间尺度下的渐近扩散运动。