Combining animal movements and behavioural data to detect behavioural states

Animal movement paths show variation in space caused by qualitative shifts in behaviours. I present a method that (1) uses both movement path data and ancillary sensor data to detect natural breakpoints in animal behaviour and (2) groups these segments into different behavioural states. The method can also combine analyses of different path segments or paths from different individuals. It does not assume any underlying movement mechanism. I give an example with simulated data. I also show the effects of random variation, # of states and # of segments on this method. I present a case study of a fisher movement path spanning 8 days, which shows four distinct behavioural states divided into 28 path segments when only turning angles and speed were considered. When accelerometer data were added, the analysis shows seven distinct behavioural states divided into 41 path segments.     

Making sense of ultrahigh‐resolution movement data: A new algorithm for inferring sites of interest

Decomposing the life track of an animal into behavioral segments is a fundamental challenge for movement ecology. The proliferation of high‐resolution data, often collected many times per second, offers much opportunity for understanding animal movement. However, the sheer size of modern data sets means there is an increasing need for rapid, novel computational techniques to make sense of these data. Most existing methods were designed with smaller data sets in mind and can thus be prohibitively slow. Here, we introduce a method for segmenting high‐resolution movement trajectories into sites of interest and transitions between these sites. This builds on a previous algorithm of Benhamou and Riotte‐Lambert (2012). Adapting it for use with high‐resolution data. The data’s resolution removed the need to interpolate between successive locations, allowing us to increase the algorithm’s speed by approximately two orders of magnitude with essentially no drop in accuracy. Furthermore, we incorporate a color scheme for testing the level of confidence in the algorithm's inference (high = green, medium = amber, low = red). We demonstrate the speed and accuracy of our algorithm with application to both simulated and real data (Alpine cattle at 1 Hz resolution). On simulated data, our algorithm correctly identified the sites of interest for 99% of “high confidence” paths. For the cattle data, the algorithm identified the two known sites of interest: a watering hole and a milking station. It also identified several other sites which can be related to hypothesized environmental drivers (e.g., food). Our algorithm gives an efficient method for turning a long, high‐resolution movement path into a schematic representation of broadscale decisions, allowing a direct link to existing point‐to‐point analysis techniques such as optimal foraging theory. It is encoded into an R package called SitesInterest , so should serve as a valuable tool for making sense of these increasingly large data streams.     

Using spatially explicit models to characterize foraging performance in heterogeneous landscapes

ABSTRACT The success of most foragers is constrained by limits to their sensory perception, memory, and locomotion. However, a general and quantitative understanding of how these constraints affect foraging benefits, and the trade-offs they imply for foraging strategies, is difficult to achieve. This article develops foraging performance statistics to assess constraints and define trade-offs for foragers using biased random walk behaviors, a widespread class of foraging strategies that includes area-restricted searches, kineses, and taxes. The statistics are expected payoff and expected travel time and assess two components of foraging performance: how effectively foragers distinguish between resource-poor and resourcerich parts of their environments and how quickly foragers in poor parts of the environment locate resource concentrations. These statistics provide a link between mechanistic models of individuals' movement and functional responses, population-level models of forager distributions in space and time, and foraging theory predictions of optimal forager distributions and criteria for abandoning resource patches. Application of the analysis to area-restricted search in coccinellid beetles suggests that the most essential aspect of these predators's foraging strategy is the "turning threshold," the prey density at which ladybirds switch from slow to rapid turning. This threshold effectively determines whether a forager exploits or abandons a resource concentration. Foraging is most effective when the threshold is tuned to match physiological or energetic requirements. These performance statistics also help anticipate and interpret the dynamics of complex spatially and temporally varying forager-resource systems.     

Efficient or Inaccurate? Analytical and Numerical Modelling of Random Search Strategies

A large number of observational and theoretical studies have investigated animal movement strategies for finding randomly located food items. Many of these studies have claimed that a particular strategy is advantageous over other strategies or that the spatial distribution of the food items affects the search efficiency. Here, we study a deliberately idealised problem, in which a blind forager searches for re-visitable food items. We show analytically that the forager’s efficiency is completely independent of both its movement strategy and the spatial pattern of the food items and depends only on the density of food in the environment. However, in some cases, apparent optima in search strategies can arise as artefacts of inappropriate and inaccurate numerical simulations. We discuss modifications to the idealised foraging problem that can confer an advantage on certain strategies, including when the forager has some memory or knowledge of the environment; when the food items are non-revisitable; and when the problem is viewed in an evolutionary context.     

Animal movement in dynamic landscapes: interaction between behavioural strategies and resource distributions

Most ecological and evolutionary processes are thought to critically depend on dispersal and individual movement but there is little empirical information on the movement strategies used by animals to find resources. In particular, it is unclear whether behavioural variation exists at all scales, or whether behavioural decisions are primarily made at small spatial scales and thus broad-scale patterns of movement simply reflect underlying resource distributions. We evaluated animal movement responses to variable resource distributions using the grey teal (Anas gracilis) in agricultural and desert landscapes in Australia as a model system. Birds in the two landscapes differed in the fractal dimension of their movement paths, with teal in the desert landscape moving less tortuously overall than their counterparts in the agricultural landscape. However, the most striking result was the high levels of individual variability in movement strategies, with different animals exhibiting different responses to the same resources. Teal in the agricultural basin moved with both high and low tortuosity, while teal in the desert basin primarily moved using low levels of tortuosity. These results call into question the idea that broad-scale movement patterns simply reflect underlying resource distributions, and suggest that movement responses in some animals may be behaviourally complex regardless of the spatial scale over which movement occurs.     

The influence of turning angles on the success of non-oriented animal searches

Animal searches cover a full range of possibilities from highly deterministic to apparently completely random behaviors. However, even those stochastic components of animal movement can be adaptive, since not all random distributions lead to similar success in finding targets. Here we address the general problem of optimizing encounter rates in non-deterministic, non-oriented searches, both in homogeneous and patchy target landscapes. Specifically, we investigate how two different features related to turning angle distributions influence encounter success: (i) the shape (relative kurtosis) of the angular distribution and (ii) the correlations between successive relative orientations (directional memory). Such influence is analyzed in correlated random walk models using a proper choice of representative turning angle distributions of the recently proposed Jones and Pewsey class. We consider the cases of distributions with nearly the same shape but considerably distinct correlation lengths, and distributions with same correlation but with contrasting relative kurtosis. In homogeneous landscapes, we find that the correlation length has a large influence in the search efficiency. Moreover, similar search efficiencies can be reached by means of distinctly shaped turning angle distributions, provided that the resulting correlation length is the same. In contrast, in patchy landscapes the particular shape of the distribution also becomes relevant for the search efficiency, specially at high target densities. Excessively sharp distributions generate very inefficient searches in landscapes where local target density fluctuations are large. These results are of evolutionary interest. On the one hand, it is shown that equally successful directional memory can arise from contrasting turning behaviors, therefore increasing the likelihood of robust adaptive stochastic behavior. On the other hand, when target landscape is patchy, adequate tumbling may help to explore better local scale heterogeneities, being some details of the shape of the distribution also potentially adaptive.     

Behavioural responses of the lizard Pedioplanis l. lineoocellata to overgrazing

We studied the effects of overgrazing on the foraging behaviour of the lizard Pedioplanis l. lineoocellata (Spotted Sand Lizard), a sit-and-wait forager, in habitats of differing vegetation states to determine the effects of habitat degradation on this species. At high grazing intensity where vegetation cover and diversity is low, the lizard P. lineoocellata moves more frequently, spends more time moving and covers larger distances than in habitats where vegetation cover and diversity is high. These behavioural changes in movement patterns can be explained by less abundant prey in habitats with low vegetation cover and diversity. Although morphology, phylogeny and physiology of P. lineoocellata should constrain the change in foraging behaviour, the species has modified its foraging strategy from sit-and-wait to actively foraging. We assume that this behavioural flexibility of P. lineoocellata is a buffer mechanism enabling the species to use and survive in degraded (unfavourable) habitats.     

Quantitative analysis of bottlenose dolphin movement patterns and their relationship with foraging

1. Broad-scale telemetry studies have greatly improved our understanding of the ranging patterns and habitat-use of many large vertebrates. However, there often remains considerable uncertainty over the function of different areas or the factors influencing habitat selection. Further insights into these processes can be obtained through analyses of finer scale movement patterns. For example, search behaviour may be modified in response to prey distribution and abundance. 2. In this study, quantitative analysis techniques are applied to the movements of bottlenose dolphins, recorded from land using a theodolite, to increase our understanding of their foraging strategies. Movements were modelled as a correlated random walk (CRW) and a biased random walk (BRW) to identify movement types and using a first-passage time (FPT) approach, which quantifies the time allocated to different areas and identifies the location and spatial scale of intensive search effort. 3. Only a quarter of the tracks were classed as CRW movement. Turning angle and directionality appeared to be key factors in determining the type of movement adopted. A high degree of overlap in search effort between separate movement paths indicated that there were small key sites (0.3 km radius) within the study area (4 km(2)). Foraging behaviour occurred mainly within these intensive search areas, indicating that they were feeding sites. 4. This approach provides a quantitative method of identifying important foraging areas and their spatial scale. Such techniques could be applied to movement paths for a variety of species derived from telemetry studies and increase our understanding of their foraging strategies.     

Exploitation of distant Antarctic waters and close neritic waters by short‐tailed shearwaters breeding in South Australia

Identifying the primary foraging grounds of abundant top predators is of importance in marine management to identify areas of high biological significance, and to assess the extent of competition with fisheries. We studied the search effort and habitat selection of the highly abundant short‐tailed shearwater Puffinus tenuirostris to assess the search strategies employed by this wide‐ranging seabird. During the chick‐rearing period 52 individuals were tracked performing 39 short foraging trips (1–2 days), and 13 long trips (11–32 days). First‐passage time analysis revealed that 46% of birds performing short trips employed area‐restricted searches, concentrating search effort at an average scale of 14 ± 5 km. Foraging searches were more continuous for the other 54%, who travelled faster to cover greater distances, with little evidence of area‐restricted searches. The prey returned indicated that continuous searchers consumed similar prey mass, but greater prey diversity than area‐restricted search birds. On long trips 23% of birds travelled 500–1000 km to neritic (continental shelf) habitats, showing weak evidence of preference for areas of higher chlorophyll a concentration, and foraged at a similar spatial scale to short trips. The other 76% performed rapid outbound flights of 1000–3600 km across oceanic habitats commuting to regions with higher chlorophyll a. The spatial scale of search effort in oceanic habitat varied widely with some performing broad‐scale searches (260–560 km) followed by finer‐scale nested searches (16–170 km). This study demonstrates that a range of search strategies are employed when exploiting prey across ocean basins. The trade‐offs between different search strategies are discussed to identify the value of these contrasting behaviours to wide‐ranging seabirds.     

Scavengers on the Move: Behavioural Changes in Foraging Search Patterns during the Annual Cycle

Background

Optimal foraging theory predicts that animals will tend to maximize foraging success by optimizing search strategies. However, how organisms detect sparsely distributed food resources remains an open question. When targets are sparse and unpredictably distributed, a Lévy strategy should maximize foraging success. By contrast, when resources are abundant and regularly distributed, simple Brownian random movement should be sufficient. Although very different groups of organisms exhibit Lévy motion, the shift from a Lévy to a Brownian search strategy has been suggested to depend on internal and external factors such as sex, prey density, or environmental context. However, animal response at the individual level has received little attention.

Methodology/Principal Findings

We used GPS satellite-telemetry data of Egyptian vultures Neophron percnopterus to examine movement patterns at the individual level during consecutive years, with particular interest in the variations in foraging search patterns during the different periods of the annual cycle (i.e. breeding vs. non-breeding). Our results show that vultures followed a Brownian search strategy in their wintering sojourn in Africa, whereas they exhibited a more complex foraging search pattern at breeding grounds in Europe, including Lévy motion. Interestingly, our results showed that individuals shifted between search strategies within the same period of the annual cycle in successive years.

Conclusions/Significance

Results could be primarily explained by the different environmental conditions in which foraging activities occur. However, the high degree of behavioural flexibility exhibited during the breeding period in contrast to the non-breeding period is challenging, suggesting that not only environmental conditions explain individuals'' behaviour but also individuals'' cognitive abilities (e.g., memory effects) could play an important role. Our results support the growing awareness about the role of behavioural flexibility at the individual level, adding new empirical evidence about how animals in general, and particularly scavengers, solve the problem of efficiently finding food resources.     

State‐space modeling reveals habitat perception of a small terrestrial mammal in a fragmented landscape

相似文献   

Adaptive Lévy Processes and Area-Restricted Search in Human Foraging

A considerable amount of research has claimed that animals’ foraging behaviors display movement lengths with power-law distributed tails, characteristic of Lévy flights and Lévy walks. Though these claims have recently come into question, the proposal that many animals forage using Lévy processes nonetheless remains. A Lévy process does not consider when or where resources are encountered, and samples movement lengths independently of past experience. However, Lévy processes too have come into question based on the observation that in patchy resource environments resource-sensitive foraging strategies, like area-restricted search, perform better than Lévy flights yet can still generate heavy-tailed distributions of movement lengths. To investigate these questions further, we tracked humans as they searched for hidden resources in an open-field virtual environment, with either patchy or dispersed resource distributions. Supporting previous research, for both conditions logarithmic binning methods were consistent with Lévy flights and rank-frequency methods–comparing alternative distributions using maximum likelihood methods–showed the strongest support for bounded power-law distributions (truncated Lévy flights). However, goodness-of-fit tests found that even bounded power-law distributions only accurately characterized movement behavior for 4 (out of 32) participants. Moreover, paths in the patchy environment (but not the dispersed environment) showed a transition to intensive search following resource encounters, characteristic of area-restricted search. Transferring paths between environments revealed that paths generated in the patchy environment were adapted to that environment. Our results suggest that though power-law distributions do not accurately reflect human search, Lévy processes may still describe movement in dispersed environments, but not in patchy environments–where search was area-restricted. Furthermore, our results indicate that search strategies cannot be inferred without knowing how organisms respond to resources–as both patched and dispersed conditions led to similar Lévy-like movement distributions.     

Analysis of foraging movements of Atlantic bluefin tuna (<Emphasis Type="Italic">Thunnus thynnus</Emphasis>): individuals switch between two modes of search behaviour

We investigate the application of quantitative techniques for distinguishing adaptive search behaviour in Atlantic bluefin tuna (Thunnus thynnus). The analysis demonstrates the application of a novel spectral analysis technique for resolving and measuring periodicity in animal behaviour patterns. Two different search strategies are identified that include regulation of turning (klinokinesis) and speed (orthokinesis). Our results provide evidence that bluefin tuna attempt to optimize their searching efficiency through adjustments in the duration and timing of switching between these two searching strategies. Repetitive, diurnal deep dives were also found to coincide with switching of search behaviour. Additional tracking experiments with larger sample sizes are needed to better identify how individuals switch between the two search strategies and how such decisions may collectively improve the searching and foraging efficiency of their schools (synchrokinesis, social taxis) in response to changes in the size or composition of prey aggregations.     

Using a Partial Sum Method and GPS Tracking Data to Identify Area Restricted Search by Artisanal Fishers at Moored Fish Aggregating Devices in the Commonwealth of Dominica

Foragers must often travel from a central place to exploit aggregations of prey. These patches can be identified behaviorally when a forager shifts from travel to area restricted search, identified by a decrease in speed and an increase in sinuosity of movement. Faster, more directed movement is associated with travel. Differentiating foraging behavior at patches from travel to patches is important for a variety of research questions and has now been made easier by the advent of small, GPS devices that can track forager movement with high resolution. In the summer and fall of 2012, movement data were collected from GPS devices placed on foraging trips originating in the artisanal fishing village of Desa Ikan (pseudonym), on the east coast of the Caribbean island nation of the Commonwealth Dominica. Moored FADs are human-made structures anchored to the ocean floor with fish attraction material on or near the surface designed to effectively create a resource patch. The ultimate goal of the research is to understand how property rights are emerging after the introduction of fish aggregating device (FAD) technology at the site in 1999. This paper reports on research to identify area-restricted search foraging behavior at FAD patches. For 22 foraging trips simultaneous behavioral observations were made to ground-truth the GPS movement data. Using a cumulative sum method, area restricted search was identified as negative deviations from the mean travel speed and the method was able to correctly identify FAD patches in every case.     

Large‐scale identification of genetic design strategies using local search

In the past decade, computational methods have been shown to be well suited to unraveling the complex web of metabolic reactions in biological systems. Methods based on flux–balance analysis (FBA) and bi‐level optimization have been used to great effect in aiding metabolic engineering. These methods predict the result of genetic manipulations and allow for the best set of manipulations to be found computationally. Bi‐level FBA is, however, limited in applicability because the required computational time and resources scale poorly as the size of the metabolic system and the number of genetic manipulations increase. To overcome these limitations, we have developed Genetic Design through Local Search (GDLS), a scalable, heuristic, algorithmic method that employs an approach based on local search with multiple search paths, which results in effective, low‐complexity search of the space of genetic manipulations. Thus, GDLS is able to find genetic designs with greater in silico production of desired metabolites than can feasibly be found using a globally optimal search and performs favorably in comparison with heuristic searches based on evolutionary algorithms and simulated annealing.     

From random walks to informed movement

The analysis of animal movement is a large and continuously growing field of research. Detailed knowledge about movement strategies is of crucial importance for understanding eco‐evolutionary dynamics at all scales – from individuals to (meta‐)populations. This and the availability of detailed movement and dispersal data motivated Nathan and colleagues to published their much appreciated call to base movement ecology on a more thorough mechanistic basis. So far, most movement models are based on random walks. However, even if a random walk might describe real movement patterns acceptably well, there is no reason to assume that animals move randomly. Therefore, mechanistic models of foraging strategies should be based on information use and memory in order to increase our understanding of the processes that lead to animal movement decisions. We present a mechanistic movement model of an animal with a limited perceptual range and basic information storage capacities. This ‘spatially informed forager’ constructs an internal map of its environment by using perception, memory and learned or evolutionarily acquired assumptions about landscape attributes. We analyse resulting movement patterns and search efficiencies and compare them to area restricted search strategies (ARS) and biased correlated random walks (BCRW) of omniscient individuals. We show that, in spite of their limited perceptual range, spatially informed individuals boost their foraging success and may perform much better than the best ARS. The construction of an internal map and the use of spatial information results in the emergence of a highly correlated walk between patches and a rather systematic search within resource clusters. Furthermore, the resulting movement patterns may include foray search behaviour. Our work highlights the strength of mechanistic modelling approaches and sets the stage for the development of more sophisticated models of memory use for movement decisions and dispersal.     

Using animal movement paths to measure response to spatial scale

Animals live in an environment that is patchy and hierarchical. I present a method of detecting the scales at which animals perceive their world. The hierarchical nature of habitat causes movement path structure to vary with spatial scale, and the patchy nature of habitat causes movement path structure to vary throughout space. These responses can be measured by a combination of path tortuousity (measured with fractal dimension) versus spatial scale, the variation in tortuousity of small path segments along the movement path, and the correlation between tortuousities of adjacent path segments. These statistics were tested using simulated animal movements. When movement paths contained no spatial heterogeneity, then fractal D and variance continuously increased with scale, and correlation was zero at all scales. When movement paths contained spatial heterogeneity, then fractal D sometimes showed a discontinuity at transitions between domains of scale, variation showed peaks at transitions, and correlations showed a statistically significant positive value at scales smaller than patch size, decreasing to below zero at scales greater than patch size. I illustrated these techniques with movement paths from deer mice and red-backed voles. These new analyses should help understand how animals perceive and react to their landscape structure at various spatial scales, and to answer questions about how habitat structure affects animal movement patterns.     

Flying bats use serial sampling to locate odour sources

Olfactory tracking generally sacrifices speed for sensitivity, but some fast-moving animals appear surprisingly efficient at foraging by smell. Here, we analysed the olfactory tracking strategies of flying bats foraging for fruit. Fruit- and nectar-feeding bats use odour cues to find food despite the sensory challenges derived from fast flight speeds and echolocation. We trained Jamaican fruit-eating bats (Artibeus jamaicensis) to locate an odour reward and reconstructed their flight paths in three-dimensional space. Results confirmed that bats relied upon olfactory cues to locate a reward. Flight paths revealed a combination of odour- and memory-guided search strategies. During ‘inspection flights’, bats significantly reduced flight speeds and flew within approximately 6 cm of possible targets to evaluate the presence or absence of the odour cue. This behaviour combined with echolocation explains how bats maximize foraging efficiency while compensating for trade-offs associated with olfactory detection and locomotion.     

trajr: An R package for characterisation of animal trajectories

Quantitative characterisation of the trajectories of moving animals is an important component of many behavioural and ecological studies, however methods are complicated and varied, and sometimes require well‐developed programming skills to implement. Here, we introduce trajr, an R package that serves to analyse animal paths, from unicellular organisms, through insects to whales. It makes a variety of statistical characterisations of trajectories, such as tortuosity, speed and changes in direction, available to biologists who may not have a background in programming. We discuss a range of indices that have been used by researchers, describe the package in detail, then use movement observations of whales and clearwing moths to demonstrate some of the capabilities of trajr. As an open‐source R package, trajr encourages open and reproducible research. It supports the implementation of additional methods by providing access to trajectory analysis “building blocks,” allows the full suite of R statistical analysis tools to be applied to trajectory analysis, and the source code can be independently validated.     

Is Foraging Time Limited During Winter? – A Feeding Experiment with Tree Sparrows Under Different Predation Risk

Small passerines are faced with a trade‐off when foraging during winter. Increasing energy reserves makes them more vulnerable to predators, while a low level of reserves exposes them to a high risk of starvation. Whether small birds under these circumstances are allowed to reduce their foraging activity under increased predation risk, for example in feeding sites more exposed to predators, remains controversial in former behavioural and ecological researches. In this study, we investigated the foraging activity of free‐living Tree Sparrow Passer montanus flocks feeding on an artificial feeding platform. The predation risk perceived by the sparrows was manipulated by placing the platform either close to or far from a bushy shelter. Foraging activity, assessed as cumulative activity of sparrows per unit time on the platform, did not differ between the low‐risk and the high‐risk conditions and did not significantly change during the day. Feeding efficiency, assessed as pecking rate, was not either reduced under the high‐risk condition. Our results suggest that sparrows were forced to feed almost continuously during the day in order to maintain their preferred level of energy reserves. However, several behavioural changes helped sparrows to adopt a safer foraging policy when feeding far from cover, as we found in another study. Altogether, sparrow flocks feeding far from cover decreased the overall foraging time (the time when any sparrow stayed on the platform) by approximately 20% as compared to the near cover condition. A possible way to maintain the same level of foraging activity despite of the reduction in overall foraging time is discussed.