Limits to behavioural plasticity

From everyday foraging trips to extraordinary long-distance migrations: animal movements are pivotal for species survival and underpin key ecological processes of importance for ecosystem functioning and biodiversity (Nathan et al., 2008). To ensure survival and reproduction, animals need to move to acquire resources, find mates and avoid risks. By moving, animals connect habitats, ecosystems and metapopulations, enabling the dispersal of seeds and genes, the spread of resources and pathogens and the flow of nutrients and energy. Therefore, their movements facilitate key ecological processes such as predator–prey interactions, herbivory, seed dispersal, nutrient cycling and disease transmission. In ensuring the distribution of genes, individuals and species in space and time, organismal movement plays a critical role in shaping biodiversity (Jeltsch et al., 2013). However, animals' room to roam is increasingly constrained by human activities, currently affecting more than 70% of the global land surface (IPCC, 2019). The increasing footprint of our activities not only leads to the direct loss of habitat and species but impacts how animals move in and make use of human-modified landscapes. For instance, across the globe, movements of terrestrial mammals are shorter in areas with high human impact (Tucker et al., 2018). While this reduction is thought to be due to behavioural changes of individual animals and the exclusion of species with long-range movements from areas with higher human impacts (Tucker et al., 2018), the underlying mechanisms are not well understood. In particular, it is unknown how plastic individuals and populations are in their responses to anthropogenic pressure—and where the limits of plasticity lie. While critical for management and conservation, understanding and predicting animal responses to human activity therefore remains challenging. Here, the research of Mumme et al. (2023) published in this issue features several important innovations: Comparing multiple populations of two deer species across a wide gradient of human disturbance, they reveal a continuum of movement tactics expressed across species, populations and individuals. Their results highlight that responses to human activity are context-dependent and may vary strongly between individuals of the same population, demonstrating a high level of plasticity and challenging traditional species and population mean approaches. Using a continuous measure of movement characterizing the full spectrum of movement behaviours—from range resident to nomadic to migratory behaviour—allowed the authors to move beyond simple categorizations of movement. Their results show that human activity, represented by the Human Footprint Index (HFI), has a marked truncating effect on movements across populations. Indeed, human activity was a stronger driver of observed movement patterns than continental and taxonomic differences or classic environmental factors (topography and forage predictability) that have previously been shown to drive movement responses. Finally, a key innovation of the authors' research lies in the identification of a potential threshold to plasticity in animals' movement responses to human activity. Due to these advances, their study provides an important and timely contribution to understanding plasticity in wildlife behaviour in an increasingly human-dominated world. The study's findings were facilitated by a unique set-up: Specifically, Mumme et al. (2023) tested the influence of human activity, topography and forage predictability on individuals' movements using an impressive set of elk and red deer movement data including 815 individuals, spanning 14 populations in North America and 14 populations in Europe. The two species belong to the Cervus genus, evolved in biogeographic continuity and are alike in anatomical structure, movement capacity and navigation ability. Both red deer and elk are known to be behaviourally plastic, undertaking directed long-distance as well as area-constrained short movements (Peters et al., 2019). The studied populations cover a large latitudinal gradient and occupy habitats ranging from almost undisturbed and fully protected wilderness to suburban landscapes. In their analyses, this disturbance gradient was represented by the HFI. The HFI integrates eight aspects of human activity (e.g. human population density, pasture lands, roads) into one cumulative proxy of human impact on nature. Ranging from ‘untouched’ at 0 to maximum human pressure at 50, it is mapped for almost all land areas across the globe at 1 km2 resolution (Venter et al., 2016). Cervus movements were quantified using the metric Intensity of Use, an area-standardized measure of the directionality of movements, integrating the total distance travelled and the size of the area covered. Low Intensity of Use values characterize directed, straight movements or roaming within a large area, indicative of either long seasonal migrations or far-ranging nomadic movements. High Intensity of Use values describe tortuous movements within a restricted area, typical for range-resident individuals. With an increase in HFI, Mumme et al. (2023) found an increase in the Intensity of Use, that is, a reduction of long-distance or wide-ranging movements. According to their results, human activity is the strongest determinant of this change in movement expression, adding to a growing body of literature demonstrating that human activity fundamentally impacts wildlife movement patterns. Yet, understanding the underlying mechanisms remains difficult. Indeed, changes in movement can represent various behavioural mechanisms, such as fleeing or avoiding humans/predators, travelling longer or shorter distances to find resources/mates, constrained movements due to physical barriers, or altered efficiency of moving through modified habitats. As Mumme et al. (2023) discuss, human activity may not only decrease animals' ability to move but also their need to do so if resource availability is locally increased (for instance through irrigated agricultural fields or supplemental feeding). To further complicate things, both mechanisms can be at play simultaneously: As, for example, Jerina (2012) showed for red deer in Slovenia, both supplemental feeding and roads decreased home range sizes. Given that the different mechanisms have different management implications, disentangling their influence constitutes an important priority for future research. Current limitations in our ability to do so also highlight shortcomings of the multi-proxy HFI: Due to its cumulative nature, it is impossible to determine which influence factor (e.g. agricultural land use or roads) dominates animal responses, especially in areas with intermediate HFI levels. This emphasizes the need for better global data sets quantifying various aspects of human activity in detail, including temporally varying human mobility (Ellis-Soto et al., 2023). Importantly, the authors' results indicate that Intensity of Use only increased until a HFI threshold of around 6.6. Beyond this threshold, Intensity of Use remained constant—suggesting that there may be a limit to Cervus movement plasticity in response to human disturbance. Interestingly, this threshold only became evident because the investigated populations inhabit the full HFI range. Just comparing different populations at a continental level would have yielded different—even reversed—results, highlighting the importance of assessing animal responses across the entire range of experienced environmental gradients. The identification of such thresholds can provide valuable insights for management and conservation efforts. Accordingly, the fact that Mumme et al. (2023) were able to find this pattern is highly relevant. Notably, the identified threshold of 6.6 is fairly low: for instance, the Bavarian Forest National Park in Germany, one of the study sites and by definition an area designated for the conservation of wild nature, has a mean HFI of 7.7. This suggests a high sensitivity to human impact even in a genus for which behavioural plasticity is known to be high (Peters et al., 2019). As mentioned above, however, explaining the threshold is challenging due to the ambiguity in interpreting HFI levels, and further complicated by the HFI relying on an interval scale, not a ratio scale (Kennedy et al., 2018), which means that doubling HFI values do not necessarily imply a doubling of human disturbance. Future work could focus on identifying the drivers of the indicated plasticity threshold and on investigating whether similar thresholds exist for other species. Depending on the underlying mechanism, an increase in the Intensity of Use can have diverging implications for individual fitness, and therefore population persistence. Moving forward, an important next step would be to link the observed continuum of expressed movement tactics to individual fitness. This would be critical to better understand the consequences of impeded movements at the population level. Irrespective of the underlying mechanism, however, a decline in the extent and directionality of animal movement will have consequences for movement-driven ecosystem processes. Ecological permeability, the unimpeded movement of species and flow of natural processes, hinges to a large degree on animal movements—especially movements of large-bodied, far-ranging species such as red deer and elk. To facilitate large-scale ecological and evolutionary processes and species range shifts in response to climate change, maintaining ecological permeability is critical. Quantifying disturbance thresholds that truncate animal movements is therefore pivotal to prevent further loss of permeability and/or to guide restoration efforts. The fact that the identified plasticity threshold only emerged in models combining populations from both continents, covering the full HFI range, showcases the importance of tracking populations across gradients, that is, from wild to urban environments. Achieving this requires coordination and collaboration between research groups. The work of Mumme et al. (2023) thus provides an important example for comparative studies in movement ecology and highlights the great potential of data sharing. This has been an early theme in the field of movement ecology, supported by data repositories such as Movebank and further driven by various initiatives such as EuroMammal, Afrimove, MegaMove, the Biologging Covid-19 Initiative, and the Global Initiative for Ungulate Migration. Their work towards interoperable databases is critical in facilitating novel and urgently needed insights into wildlife responses to global change. Studies like Mumme et al. (2023) illustrate that these collaborations afford inference that otherwise would not be obtainable. Providing new evidence of how radically our activities impact the natural world, the results of Mumme et al. (2023) should give us pause. Their work, however, delivers critical insights for designing management measures safeguarding animals' room to roam. Mitigating the constraining impacts of human activity on animal movements will be key for biodiversity conservation in the advancing Anthropocene. The authors declare that they have no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analysed for the current article.