Victor M Eguiluz | IFISC (UIB-CSIC) (original) (raw)

Papers by Victor M Eguiluz

Scientific Reports, 2020

Understanding the response of ecological networks to perturbations and disruptive events is neede... more Understanding the response of ecological networks to perturbations and disruptive events is needed to anticipate the biodiversity loss and extinction cascades. Here, we study how network plasticity reshapes the topology of mutualistic networks in response to species loss. We analyze more than one hundred empirical mutualistic networks and considered random and targeted removal as mechanisms of species extinction. network plasticity is modeled as either random rewiring, as the most parsimonious approach, or resource affinity-driven rewiring, as a proxy for encoding the phylogenetic similarity and functional redundancy among species. this redundancy should be positively correlated with the robustness of an ecosystem, as functions can be taken by other species once one of them is extinct. We show that effective modularity, i.e. the ability of an ecosystem to adapt or restructure, increases with increasing numbers of extinctions, and with decreasing the replacement probability. Importantly, modularity is mostly affected by the extinction rather than by rewiring mechanisms. These changes in community structure are reflected in the robustness and stability due to their positive correlation with modularity. Resource affinity-driven rewiring offers an increase of modularity, robustness, and stability which could be an evolutionary favored mechanism to prevent a cascade of co-extinctions. Network adaptation reflects the ability of a system to respond to external (e.g., environmental) perturbations in terms of establishing dynamic interactions between the remaining elements. This sort of adaptation based on the coevolution of topologies and states appears in many network systems 1,2 , including opinion formation 3,4 , the spread of infectious diseases 5 , online social networks 6 , ecological and biological networks 7-11 , power grids 12 and chemistry 13. An important class of adaptation dynamics occurs as a consequence of extinction events altering network structure. Examples include the disappearance of companies and technologies due to innovations and market competition 14,15 , neurons extinction 16 , gene extinction 17 , extinction of the least populated molecular species 18 local extinction of species in biological networks due to perturbation such as pollution or human disturbance 19,20 , habitat loss 21 , climate change 22,23 or extinction of interaction partners 24,25. In ecosystems, loss or declining abundance of one species can cause the extinction of other dependent species 26-28. Understanding how ecosystems respond to species loss has hence been the subject of numerous theoretical and experimental studies 29,30. While some communities exhibit density compensation 31,32 , a common form of adaptation in food webs is represented by the rewiring of feeding links among members of species in response to variation in resources 33-37. Functional redundancy can help in the process of rewiring, as species have the ability to replace each other. So, for example, pollination networks have been shown to be fairly robust against species extinctions as a result of bipartite network asymmetry (redundancy in the number of floral visitors per plant 25). However, trophic rewiring can lead to overexploitation of resources and aggravate the effects of species loss on food webs 38. From the perspective of the topological structure of the interactions of the species that compose ecosystems , recent studies have revealed key structural features which are believed to influence ecological dynamics 39 : robustness, defined in terms of the largest connected component 40 ; nestedness, where specialists interact with a

arXiv preprint cond-mat/0210173, Oct 8, 2002

Abstract: We study the one-dimensional version of Axelrod's model of cultural transmission f... more Abstract: We study the one-dimensional version of Axelrod's model of cultural transmission from the point of view of optimization dynamics. We show the existence of a Lyapunov potential for the dynamics. The global minimum of the potential, or optimum state, is the monocultural uniform state, which is reached for an initial diversity of the population below a critical value. Above this value, the dynamics settles in a multicultural or polarized state. These multicultural attractors are not local minima of the potential, so that any small ...

Scientific Reports, 2013

Technologically driven transport systems are characterized by a networked structure connecting op... more Technologically driven transport systems are characterized by a networked structure connecting operation centers and by a dynamics ruled by pre-established schedules. Schedules impose serious constraints on the timing of the operations, condition the allocation of resources and define a baseline to assess system performance. Here we study the performance of an air transportation system in terms of delays. Technical, operational or meteorological issues affecting some flights give rise to primary delays. When operations continue, such delays can propagate, magnify and eventually involve a significant part of the network. We define metrics able to quantify the level of network congestion and introduce a model that reproduces the delay propagation patterns observed in the U.S. performance data. Our results indicate that there is a non-negligible risk of systemic instability even under normal operating conditions. We also identify passenger and crew connectivity as the most relevant internal factor contributing to delay spreading.

Journal of Artificial Societies and Social Simulation, 2007

We study the effects of different forms of information feedback associated with mass media on an ... more We study the effects of different forms of information feedback associated with mass media on an agent-agent based model of the dynamics of cultural dissemination. In addition to some processes previously considered, we also examine a model of local mass media influence in cultural dynamics. Two mechanisms of information feedback are investigated: (i) direct mass media influence, where local or global mass media act as an additional element in the network of interactions of each agent, and (ii) indirect mass media influence, where global media acts as a filter of the influence of the existing network of interactions of each agent. Our results generalize previous findings showing that cultural diversity builds-up by increasing the strength of the mass media influence. We find that this occurs independently of the mechanisms of action (direct or indirect) of the mass media message. However, through an analysis of the full range of parameters measuring cultural diversity, we establish that the enhancement of cultural diversity produced by interaction with mass media only occurs for strong enough mass media messages. In comparison with previous studies a main different result is that weak mass media messages, in combination with agent-agent interaction, are efficient in producing cultural homogeneity. Moreover, the homogenizing effect of weak mass media messages are more efficient for direct local mass media messages than for global mass media messages or indirect global mass media influences.

Chaos, 2004

We study the regime of anticipated synchronization recently described on a number of dynamical sy... more We study the regime of anticipated synchronization recently described on a number of dynamical systems including chaotic ones. We use simple linear caricatures to show the minimal setups able to reproduce the basic facts described. The possibility of predicting the behavior of a dynamical system " ''master''… in real time using a similar copy " ''slave''… has been demonstrated theoretically, numerically and experimentally. This surprising result is of general validity, although its main interest concerns those systems, such as chaotic, whose dynamics has an intrinsic degree of unpredictability. The prediction scheme is very simple and relies on the use of time delay lines in the dynamics of the slave system, while the master dynamics is not altered. By focusing on simple linear examples, in this paper we extract and highlight the essential ingredients of this intriguing phenomenon. We also analyze it from the engineering point of view, where the slave is seen as a cascade of control system blocks, an interpretation which might be useful in future applications to system control.

PLoS ONE, 2011

As important as the intrinsic properties of an individual nervous cell stands the network of neur... more As important as the intrinsic properties of an individual nervous cell stands the network of neurons in which it is embedded and by virtue of which it acquires great part of its responsiveness and functionality. In this study we have explored how the topological properties and conduction delays of several classes of neural networks affect the capacity of their constituent cells to establish well-defined temporal relations among firing of their action potentials. This ability of a population of neurons to produce and maintain a millisecond-precise coordinated firing (either evoked by external stimuli or internally generated) is central to neural codes exploiting precise spike timing for the representation and communication of information. Our results, based on extensive simulations of conductance-based type of neurons in an oscillatory regime, indicate that only certain topologies of networks allow for a coordinated firing at a local and long-range scale simultaneously. Besides network architecture, axonal conduction delays are also observed to be another important factor in the generation of coherent spiking. We report that such communication latencies not only set the phase difference between the oscillatory activity of remote neural populations but determine whether the interconnected cells can set in any coherent firing at all. In this context, we have also investigated how the balance between the network synchronizing effects and the dispersive drift caused by inhomogeneities in natural firing frequencies across neurons is resolved. Finally, we show that the observed roles of conduction delays and frequency dispersion are not particular to canonical networks but experimentally measured anatomical networks such as the macaque cortical network can display the same type of behavior.

Models of social diffusion reflect processes of how new products, ideas, or behaviors are adopted... more Models of social diffusion reflect processes of how new products, ideas, or behaviors are adopted in a population. These models typically lead to a continuous or a discontinuous phase transition of the number of adopters as a function of a control parameter. We explore a simple model of social adoption where the agents can be in two states, either adopters or non-adopters, and can switch between these two states interacting with other agents through a network. The probability of an agent to switch from non-adopter to adopter depends on the number of adopters in her network neighborhood, the adoption threshold T and the adoption coefficient a, two parameters defining a Hill function. In contrast, the transition from adopter to non-adopter is spontaneous at a certain rate µ. In a mean-field approach, we derive the governing ordinary differential equations and show that the nature of the transition between the global non-adoption and global adoption regimes depends mostly on the balance between the probability to adopt with one and two adopters. The transition changes from continuous, via a transcritical bifurcation, to discontinuous, via a combination of a saddle-node and a transcritical bifurcation, through a supercritical pitchfork bifurcation. We characterize the full parameter space. Finally, we compare our analytical results with Monte Carlo simulations on annealed and quenched degree regular networks, showing a better agreement for the annealed case. Our results show how a simple model is able to capture two seemingly very different types of transitions, i.e., continuous and discontinuous and thus unifies underlying dynamics for different systems. Furthermore, the form of the adoption probability used here is based on empirical measurements.

The extent of increasing anthropogenic impacts on large marine vertebrates partly depends on the ... more The extent of increasing anthropogenic impacts on large marine vertebrates partly depends on the animals' movement patterns. Effective conservation requires identification of the key drivers of movement including intrinsic properties and extrinsic constraints associated with the dynamic nature of the environments the animals inhabit. However, the relative importance of intrinsic versus extrinsic factors remains elusive. We analyze a global dataset of ∼2.8 million locations from >2,600 tracked individuals across 50 marine vertebrates evolu-tionarily separated by millions of years and using different locomotion modes (fly, swim, walk/paddle). Strikingly, movement patterns show a remarkable convergence, being strongly conserved across species and independent of body length and mass, despite these traits ranging over 10 orders of magnitude among the species studied. This represents a fundamental difference between marine and terrestrial vertebrates not previously identified, likely linked to the reduced costs of locomotion in water. Movement patterns were primarily explained by the interaction between species-specific traits and the habitat(s) they move through, resulting in complex movement patterns when moving close to coasts compared with more predictable patterns when moving in open oceans. This distinct difference may be associated with greater complexity within coastal microhabitats, highlighting a critical role of preferred habitat in shaping marine vertebrate global movements. Efforts to develop understanding of the characteristics of vertebrate movement should consider the habitat(s) through which they move to identify how movement patterns will alter with forecasted severe ocean changes, such as reduced Arctic sea ice cover, sea level rise, and declining oxygen content. global satellite tracking | probability density function | root-mean-square | turning angles | displacements U nifying theoretical frameworks that explain general principles of animal life-history (1), optimal foraging (2, 3), and metabolic scaling in organisms (4, 5) facilitate the interpretation of data and the generation of testable hypotheses. Animal movement accounts for most of the energy budgets of vertebrates because it underpins critical components of their behavior, such as feeding and mating. Following the challenge posed by Aristotle millennia ago in De Motu Animalium (On the Movement of Animals) (6), efforts have been made to develop a unifying framework to study movement (7). [ " Now we must consider in general the common reason for moving with any movement whatever (for some animals move by flying, some by swimming, some by stepping, some in other comparable ways). " (ref. 6, p. 24)] Such efforts have provided clarification that the primary challenge for understanding animal movement lies in the identification of the key external factors, internal states, and the motion and navigation capabilities influencing movement (7). It is also known that animal movement patterns are underpinned by common principles, such as " optimal " resource exploitation by predators (" optimality paradigm " ; refs. 2, 3, 8, and 9) or the use of more efficient search trajectories (" random " paradigm; refs. 10–12). Overall, animal movement patterns have been attributed to extrinsic factors, including the dynamic nature of the environments they inhabit and constrained by intrinsic properties (13–19), including allometric and metabolic scaling with day or home range and locomotion speed, particularly for terrestrial animals (4, 15, 20–23). However, the relative importance of extrinsic versus intrinsic properties in determining the observed patterns of movement of free-ranging animals remains ambiguous. To effectively partition the relative contributions of extrinsic versus intrinsic factors and effectively investigate whether a unifying framework exists irrespective of location, scale of movement, and stage or phase, a large-scale comparison across multiple species is needed (24, 25). Rapid technological developments in animal-attached electronic tags (telemetry/biologging) have generated large tracking datasets across an array of marine vertebrates, now available for Significance Understanding the key drivers of animal movement is crucial to assist in mitigating adverse impacts of anthropogenic activities on marine megafauna. We found that movement patterns of marine megafauna are mostly independent of their evolutionary histories, differing significantly from patterns for terrestrial animals. We detected a remarkable convergence in the distribution of speed and turning angles across organisms ranging from whales to turtles (epitome for the slowest animals on land but not at sea). Marine megafauna show a prevalence of movement patterns dominated by search behavior in coastal habitats compared with more directed, ballistic movement patterns when the animals move across the open ocean. The habitats through which they move will therefore need to be considered for effective conservation.

The rise of the internet coupled with technological innovations such as smartphones have generate... more The rise of the internet coupled with technological innovations such as smartphones have generated massive volumes of geo-referenced data (big data) on human mobility. This has allowed the number of studies of human mobility to rapidly overtake those of animal movement. Today, telemetry studies of animals are also approaching big data status. Here, we review recent advances in studies of human mobility and identify the opportunities they present for advancing our understanding of animal movement. We describe key analytical techniques, potential bottlenecks and a roadmap for progress toward a synthesis of movement patterns of wild animals.

The association between corals and photosynthetic dinoflagellates (Symbiodinium spp.) is the key ... more The association between corals and photosynthetic dinoflagellates (Symbiodinium spp.) is the key to the success of reef ecosystems in highly oligotrophic environments, but it is also their Achilles' heel due to its vulnerability to local stressors and the effects of climate change. Research during the last two decades has shaped a view that coral host–Symbiodinium pairings are diverse, but largely exclusive. Deep sequencing has now revealed the existence of a rare diversity of cryptic Symbiodinium assemblages within the coral holobiont, in addition to one or a few abundant algal members. While the contribution of the most abundant resident Symbiodinium species to coral physiology is widely recognized, the significance of the rare and low abundant background Symbiodinium remains a matter of debate. In this study, we assessed how coral–Symbiodinium communities assemble and how rare and abundant components together constitute the Symbiodi-nium community by analyzing 892 coral samples comprising 4110 000 unique Symbiodinium ITS2 marker gene sequences. Using network modeling, we show that host–Symbiodinium communities assemble in non-random 'clusters' of abundant and rare symbionts. Symbiodinium community structure follows the same principles as bacterial communities, for which the functional significance of rare members (the 'rare bacterial biosphere') has long been recognized. Importantly, the inclusion of rare Symbiodinium taxa in robustness analyses revealed a significant contribution to the stability of the host–symbiont community overall. As such, it highlights the potential functions rare symbionts may provide to environmental resilience of the coral holobiont.

We study the spreading of cooperative infections in an empirical temporal network of contacts bet... more We study the spreading of cooperative infections in an empirical temporal network of contacts between people, including health care workers and patients, in a hospital. The system exhibits a phase transition leading to one or several endemic branches, depending on the connectivity pattern and the temporal correlations. There are two endemic branches in the original setting and the non-cooperative case. However, the cooperative interaction between infections reinforces the upper branch, leading to a smaller epidemic threshold and a higher probability for having a big outbreak. We show the microscopic mechanisms leading to these differences, characterize three different risks, and use the influenza features as an example for this dynamics.

Infectious diseases have been modelled on networks that summarise physical contacts or close prox... more Infectious diseases have been modelled on networks that summarise physical contacts or close proximity of individuals. These networks are known to be complex in both their structure and how they change over time. We present an overview of recent progress in numerically determining the epidemic threshold in temporally-switching networks, and illustrate that slower switching of snapshots relative to epidemic dynamics lowers the epidemic threshold. Therefore, ignoring the temporally-varying nature of networks may underestimate endemicity. We also identify a predictor for the magnitude of this shift which is
based on the commutator norm of snapshot adjacency matrices.

Scientific Reports, 2017

Coextinction models are useful to understand community robustness to species loss and resilience ... more Coextinction models are useful to understand community robustness to species loss and resilience to disturbances. We simulated pollinator extinctions in pollination networks by using a hybrid model that combined a recently developed stochastic coextinction model (SCM) for plant extinctions and a topological model (TCM) for animal extinctions. Our model accounted for variation in interaction strengths and included empirical estimates of plant dependence on pollinators to set seeds. The stochastic nature of such model allowed us determining plant survival to single (and multiple) extinction events, and identifying which pollinators (keystone species) were more likely to trigger secondary extinctions. Consistently across three different pollinator removal sequences, plant robustness was lower than in a pure TCM, and plant survival was more determined by dependence on the mutualism than by interaction strength. As expected, highly connected and dependent plants were the most sensitive to pollinator loss and collapsed faster in extinction cascades. We predict that the relationship between dependence and plant connectivity is crucial to determine network robustness to interaction loss. Finally, we showed that honeybees and several beetles were keystone species in our communities. This information is of great value to foresee consequences of pollinator losses facing current global change and to identify target species for effective conservation. Anthropogenic disturbances, such as habitat transformation, climate change or biological invasions, are at present main drivers of species loss and disruption of ecological interactions 1–3. Ecological interactions among species are an important component of biodiversity because they provide relevant functions for populations, communities and ecosystems, such as pollination services 4–6. For that reason, predicting responses of ecosystems and tolerance of species to disturbances is a key issue in ecology. The extinction of a species inevitably causes the extinction of interactions, which in turn can trigger additional extinctions of species. Models with ecological interaction networks constitute a useful tool to simulate coextinction cascades, with the ultimate goal of understanding community robustness to species loss and resilience 7–10. Topological coextinction models (TCMs) represented the first attempt to explore patterns of extinction in plant-pollinator networks 7. Nevertheless, these models are mainly based on static network structure and have several important constraints. For instance, they assume that a species can only become extinct when all its interacting partners are lost; however, the primary loss of a pollinator species in real plant-pollinator networks may cause the coextinction of a plant, leading in turn to the coextinction of other pollinators that strongly depended on that plant, and even to the indirect coextinction of other plants which rely on those pollinators. TCMs also neglect interaction strength heterogeneity, assuming that after the loss of a species, all its partners are equally likely to coextinction; nevertheless, pollination interactions differ in terms of quantity and quality effects 11–13 and thus the coextinction probability of a species partner varies depending on the ecological effect of the lost

Antibiotic-resistant organisms, an increasing source of morbidity and mortality, have a natural r... more Antibiotic-resistant organisms, an increasing source of morbidity and mortality, have a natural reservoir in hospitals, and recent estimates suggest that almost 2 million people develop hospital-acquired infections each year in the US alone. We investigate the temporal network of transfers of Medicare patients across US hospitals over a 2-year period to learn about the possible role of hospital-to-hospital transfers of patients in the spread of infections. We analyze temporal, geographical, and topological properties of the transfer network and show that this network may serve as a substrate for the spread of infections. Finally, we study different strategies for the early detection of incipient epidemics on the temporal transfer network as a function of activation time of a subset of sensor hospitals. We find that using approximately 2% of hospitals as sensors, chosen based on their network in-degree, with an activation time of 7 days results in optimal performance for this early warning system, enabling the early detection of 80% of the C. difficile. cases with the hospitals in the sensor set activated for only a fraction of 40% of the time.

Scientific Reports, 2017

Antibiotic-resistant bacterial infections are a substantial source of morbidity and mortality and... more Antibiotic-resistant bacterial infections are a substantial source of morbidity and mortality and have a common reservoir in inpatient settings. Transferring patients between facilities could be a mechanism for the spread of these infections. We wanted to assess whether a network of hospitals, linked by inpatient transfers, contributes to the spread of nosocomial infections and investigate how network structure may be leveraged to design efficient surveillance systems. We construct a network defined by the transfer of Medicare patients across US inpatient facilities using a 100% sample of inpatient discharge claims from 2006–2007. We show the association between network structure and C. difficile incidence, with a 1% increase in a facility's C. difficile incidence being associated with a 0.53% increase in C. difficile incidence of neighboring facilities. Finally, we used network science methods to determine the facilities to monitor to maximize surveillance efficiency. An optimal surveillance strategy for selecting " sensor " hospitals, based on their network position, detects 80% of the C. difficile infections using only 2% of hospitals as sensors. Selecting a small fraction of facilities as " sensors " could be a cost-effective mechanism to monitor emerging nosocomial infections. Healthcare-associated infections are a significant source of morbidity and mortality, imposing substantial clinical and financial costs to the US health care system 1–6. Many infections have a common reservoir in inpatient settings such as hospitals and rehabilitation facilities, but they are difficult to monitor on a national scale. A 2013 Centers for Disease Control and Prevention (CDC) report on antibiotic-resistant bacteria identified the lack of infrastructure to detect and respond to emerging resistant infections as a pressing gap 2. While patient transfers could plausibly act as a mechanism for epidemiologic spread from facility to facility, only a few studies have investigated the possible role of transfers for the spread of infections at the country level, which constitutes arguably the biggest scale for these kind of systems. Some studies have focused on the structure of the nationwide critical care transfer network 7–10 , while others have had a more restricted scope, limited to geographical units such as counties or states 11–14. In this study, we consider nationwide transfers of Medicare patients 65 or older, who constitute about 15% of the US population 15 , and about 37% of all hospital admissions 16. This population is also arguably at highest risk for morbidity and mortality from health care associated infections. As a case study of nosocomial infections we use data on Clostridium difficile [C. difficile], which is an anaer-obic, gram-positive, spore-forming bacteria that occurs frequently in health care settings. It is found in >20% of patients who have been hospitalized for more than one week. The disease is spread by ingestion of C. difficile spores, which are very hardy and can persist on environmental surfaces for months without proper hygiene 17. C. difficile associated infections reached half a million in the United States only, with 29,000 patients deaths, 15,000 of which were estimated to be directly caused by C. difficile infections (80% of patients 65 or older). Furthermore, approximately two thirds of the C. difficile infections are associated with a stay in an inpatient facility 18 .

The exploitation of high volume of geolocalized data from social sport tracking applications of o... more The exploitation of high volume of geolocalized data from social sport tracking applications of outdoor activities can be useful for natural resource planning and to understand the human mobility patterns during leisure activities. This geolocalized data represents the selection of hike activities according to subjective and objective factors such as personal goals, personal abilities, trail conditions or weather conditions. In our approach, human mobility patterns are analysed from trajectories which are generated by hikers. We propose the generation of the trail network identifying special points in the overlap of trajectories. Trail crossings and trailheads define our network and shape topological features. We analyse the trail network of Balearic Islands, as a case of study, using complex weighted network theory. The analysis is divided into the four seasons of the year to observe the impact of weather conditions on the network topology. The number of visited places does not decrease despite the large difference in the number of samples of the two seasons with larger and lower activity. It is in summer season where it is produced the most significant variation in the frequency and localization of activities from inland regions to coastal areas. Finally, we compare our model with other related studies where the network possesses a different purpose. One finding of our approach is the detection of regions with relevant importance where landscape interventions can be applied in function of the communities.

The subtropical ocean gyres are recognized as great marine accummulation zones of floating plasti... more The subtropical ocean gyres are recognized as great marine accummulation zones of floating plastic debris; however, the possibility of plastic accumulation at polar latitudes has been overlooked because of the lack of nearby pollution sources. In the present study, the Arctic Ocean was extensively sampled for floating plastic debris from the Tara Oceans circumpolar expedition. Although plastic debris was scarce or absent in most of the Arctic waters, it reached high concentrations (hundreds of thousands of pieces per square kilometer) in the northernmost and easternmost areas of the Greenland and Barents seas. The fragmentation and typology of the plastic suggested an abundant presence of aged debris that originated from distant sources. This hypothesis was corroborated by the relatively high ratios of marine surface plastic to local pollution sources. Surface circulation models and field data showed that the poleward branch of the Thermohaline Circulation transfers floating debris from the North Atlantic to the Greenland and Barents seas, which would be a dead end for this plastic conveyor belt. Given the limited surface transport of the plastic that accumulated here and the mechanisms acting for the downward transport, the seafloor beneath this Arctic sector is hypothesized as an important sink of plastic debris.

Recent developments of the multilayer paradigm include efforts to understand the role played by t... more Recent developments of the multilayer paradigm include efforts to understand the role played by the presence of several layers on the dynamics of processes running on the networks. The possible existence of new phenomena associated to the richer multilayer topology has been discussed and examples of these differences have been systematically searched for. Here, we show that the interconnectivity of the layers may have an important impact on the speed of the dynamics run in the network and that microscopic changes such as the addition of one single inter-layer link can notably affect the arrival at a global stationary state. As a practical testbed, these results obtained with spectral techniques are confirmed with a Kuramoto dynamics for which the synchronization consistently accelerates after the addition of single inter-layer links.

The growing number of large databases of animal tracking provides an opportunity for analyses of ... more The growing number of large databases of animal tracking provides an opportunity for analyses of movement patterns at the scales of populations and even species. We used analytical approaches, developed to cope with " big data " , that require no 'a priori' assumptions about the behaviour of the target agents, to analyse a pooled tracking dataset of 272 elephant seals (Mirounga leonina) in the Southern Ocean, that was comprised of >500,000 location estimates collected over more than a decade. Our analyses showed that the displacements of these seals were described by a truncated power law distribution across several spatial and temporal scales, with a clear signature of directed movement. This pattern was evident when analysing the aggregated tracks despite a wide diversity of individual trajectories. We also identified marine provinces that described the migratory and foraging habitats of these seals. Our analysis provides evidence for the presence of intrinsic drivers of movement, such as memory, that cannot be detected using common models of movement behaviour. These results highlight the potential for " big data " techniques to provide new insights into movement behaviour when applied to large datasets of animal tracking. Movement is a fundamental aspect of animal behaviour 1. The need to search for food, mates and shelter shapes many aspects of animal ecology and is central to developing conservation and management strategies for any species 2, 3. Studies of animal movement were catalysed by the introduction of satellite-linked tags and the Argos satellite system in the late 1970's 4, 5 , which for the first time allowed animals to be tracked in a near-real time across habitats such as the forests, skies and open oceans that had previously been largely inaccessible to researchers. Observations describing horizontal displacements have been the most common product of satellite-linked tags. Analysis of these tracks can reveal the processes that underlie the movement strategies of the target species 6 and have mostly focused on the role of prey distribution in determining movement patterns 7–9. However, movement patterns are unlikely to be solely a response to the spatial and temporal distribution of food 10, 11. Animals have the capacity to learn and react to important aspects of their environment for many reasons, such as reproduction and anti-predator behaviour 12 , or even fear 13, 14. Some movement behaviours may even be genetically programmed 15, 16. Examination of these ideas has been limited in the past by the small sample sizes of most tracking studies due to the expense of satellite tags. Such low replication led to mostly individual-based analysis rendering any intrinsic (learning, genetic) component of movement behaviour difficult to detect. In recent years,