African elephants in Etosha National Park (data from Tsalyuk et al. 2018)

The relationship between resource availability and wildlife movement patterns is pivotal to understanding species behavior and ecology. Movement response to landscape variables occurs at multiple temporal scales, from sub-diurnal to multiannual. Additionally, individuals may respond to both current and past conditions of resource availability. In this paper, we examine the temporal scale and variation of current and past resource variables that affect movement patterns of African elephants (Loxodonta africana) using sub-hourly movement data from GPS-GSM collared elephants in Etosha National Park, Namibia. We created detailed satellite-based spatiotemporal maps of vegetation biomass, as well as distance from surface water, road and fence. We used step selection functions to measure the relative importance of these landscape variables in determining elephants’ local movement patterns. We also examined how elephants respond to information, in locations they have previously visited, on productivity integrated over different temporal scales: from current to historical conditions. Our results demonstrate that elephants choose patches with higher-than average annual productivity and grass biomass, but lower tree biomass. Elephants also prefer to walk close to water, roads, and fences. These preferences vary with time of day and with season, thereby providing insights into diurnal and seasonal behavioral patterns and the ecological importance of the landscape variables examined. We also discovered that elephants respond more strongly to long-term patterns of productivity than to immediate forage conditions, in familiar locations. Our results illustrate how animals with high cognitive capacity and spatial memory integrate long-term information on landscape conditions. We illuminate the importance of long-term high temporal resolution satellite imagery to understanding the relationship between movement patterns and landscape structure.

资源可获得性与野生动物运动模式之间的关联，对于理解物种行为与生态学特征而言至关重要。动物对景观变量的运动响应存在多个时间尺度，跨度从亚昼夜时段直至多年周期。此外，野生动物个体既会对当前的资源可获得性状况做出响应，也会受过往资源条件的影响。本研究依托纳米比亚埃托沙国家公园内佩戴GPS-GSM项圈的非洲象（Loxodonta africana）所采集的亚小时级运动数据，探究影响非洲象运动模式的当前与过往资源变量的时间尺度及其变化特征。我们基于卫星影像绘制了植被生物量、距地表水源、道路与围栏距离的高精度时空分布图。我们采用步选择函数（step selection functions）量化上述各类景观变量在塑造非洲象局地运动模式中的相对重要性。此外，我们还探究了非洲象在其曾造访过的区域中，如何响应整合了不同时间尺度（从当前至历史时段）的生产力信息。研究结果显示，非洲象会选择年均生产力与草本生物量高于平均水平，但乔木生物量较低的栖息地斑块。非洲象同时偏好靠近水源、道路与围栏的行进路线。这些偏好会随昼夜时段与季节发生变化，由此为理解非洲象的昼夜与季节性行为模式，以及所探究的各类景观变量的生态学意义提供了新视角。我们还发现，在熟悉的区域中，非洲象对长期生产力模式的响应强度要高于对即时草料条件的响应。本研究结果阐明了具备高认知能力与空间记忆能力的动物，是如何整合景观条件的长期信息的。本研究同时阐明了高时间分辨率长期卫星影像，在理解运动模式与景观结构之间关联时的重要价值。