Tracking Data Reveal How Each Animal Carves Its Own Ecological Niche

Every animal of the same species does not live the same life. Some individuals of a species consistently seek out wet soils rich in earthworms, while others tolerate busy farmland or crowded human landscapes. Ecologists have long known that such between-individual differences in resource use, responses to environmental gradients, and movement decisions can shape population dynamics, competition, and even species coexistence. What has been missing is a practical, comparable way to translate the flood of modern GPS tracking data into operational measures of individual specialization in habitat use. A new study published in Individual-based Ecology now offers exactly that: a complete statistical workflow that quantifies what its authors call the individualized niche, implemented entirely in the R programming language and demonstrated with publicly available tracking data from northern lapwings across Western Europe.

The ecological niche is one of the foundational concepts in ecology, but its definition has been contested for a century. Charles Elton in 1927 framed the niche as a species’ functional role within a community, emphasizing interactions with other species. Joseph Grinnell in 1928 instead defined it as the physical space in which a species lives. The most widely used formulation came from G. Evelyn Hutchinson in 1957, who described the niche as a hypervolume in an n-dimensional environmental space, roughly spanning biotic and abiotic factors. Hurlbert later compiled more than twenty different definitions of the niche, and the debate continues. Crucially, almost all of these definitions operate at the species level or above, treating variation among individuals as noise around an ecological optimum.

That species-level focus is now recognized as a serious limitation. Foundational work over recent decades has shown that many species contain distinct phenotypes linked to age, sex, morphology, or behavior, and that consistent individual differences appear in cognition, antipredator behavior, breeding behavior, and responses to chemical cues. The Niche Variation Hypothesis, proposed by Van Valen in 1965, holds that the width of a species’ niche is driven by among-individual variation, a relationship demonstrated empirically across multiple taxa. Individual-level differences therefore matter for population-level dynamics, species coexistence, and the indefinite persistence of species. The term individualized niche emerged in the early 2010s to describe individual niche specialization, and it inherits the plasticity of the broader niche concept, which is precisely why its quantification has been considered a laborious task.

The new workflow, presented by researchers building on a formal definition published by Takola and Schielzeth, distinguishes between realized and potential individualized niches, following Hutchinson’s classic separation of the fundamental and realized niches. The realized individualized niche represents the environmental space where an individual actually is found, while the potential niche represents the space where it could be found. The workflow proceeds in three stages: data preparation, analysis, and output generation. Recorded GPS locations correspond to used locations, and available locations are generated randomly within a radius set by the organism’s dispersal rates. These points are then overlaid with environmental variables, which can be measured, inferred, or calculated with the aid of remote-sensing tools and grouped by resource type such as food, soil, and weather.

The analytical core is a mixed-effects resource selection function, or RSF, fitted for each resource-type group using a binomial model with a logit link. The response variable is binary, distinguishing used from available locations. Each model contains the relevant environmental predictors along with random intercepts and random slopes for each individual. The random intercepts capture baseline differences among individuals, whereas the random slopes, also known as reaction norms, estimate individual variation in the strength of selection along each environmental gradient. The model outputs represent population-average effects that describe selection and individualized deviations that quantify consistent among-individual differences along specific environmental axes. Fitted predictions are converted from the log-odds scale back to probabilities of use, and these predicted values are then scaled into relative selection probabilities, or RSPs, which set each individual’s maximum predicted selection value to one and express all other locations relative to that maximum.

These RSPs serve as observation-level weights in the construction of individualized niche hypervolumes. For each resource type, the workflow transforms the ecological group into a standardized multidimensional environmental space and estimates a weighted hypervolume separately for each individual using Gaussian kernel density estimation with Silverman’s bandwidth estimator. Because density-estimated hypervolumes have continuous boundaries, the workflow converts them to hard boundaries by selecting the region containing 95 percent of the probability estimates. Niche breadth is then defined as the volume of each individual’s 95 percent hypervolume, roughly corresponding to the within-individual variance of the individualized niche. Pairwise niche overlap between individuals is quantified with Jaccard and Sørensen similarity indices, which provide insights into intraspecific competition and niche partitioning along different environmental axes. Notably, the breadth of a potential niche is not necessarily larger than that of the realized niche, contrary to the traditional species-level definition, a phenomenon related to the so-called zero-niche paradox in which animals survive under conditions outside their species’ fundamental niche.

A key innovation is the use of repeatability estimates to address the problem of resource availability, which strongly influences individual variation in use. Repeatability quantifies the proportion of variance attributable to between-individual differences. The workflow calculates repeatability separately for realized niches, based on the environmental conditions individuals actually use, and for potential niches, based on the conditions available to them. Higher realized than potential repeatability indicates that individuals differ in use beyond what would be expected from their opportunity sets alone, supporting an interpretation of genuine individualized selection or specialization. Similar values suggest that differences in use largely reflect differences in availability, while higher potential repeatability points to spatial or home-range segregation. The workflow implements these tests with 1,000 bootstrap replicates and 1,000 permutations to obtain empirical 95 percent confidence intervals.

The case study showcases the workflow with northern lapwing tracking data publicly available through the Movebank database, covering the Netherlands, Belgium, England, France, and Spain from April 2021 to March 2023, with between 28 and 1,266 locations per individual. The environmental variables integrated into the lapwing niches were remarkably rich: temperature, precipitation, wind speed, the normalized difference vegetation index, crop type, cropping intensity, soil water index, soil organic carbon and its stock, soil nitrogen, pesticide application rates, population density, human footprint, and earthworm abundance. The mixed-effects RSFs showed that lapwings avoided densely populated areas and frequently cropped fields but preferred certain crop types, water, and high earthworm abundance, confirming what is already known about the species’ ecology while the random slopes revealed how each of the 13 tracked birds responded differently to each environmental variable.

The lapwing results illustrate the diagnostic power of the paired niche components. In many cases, for example in the selection of soil and human-presence variables, realized niche breadth was smaller than potential niche breadth, providing evidence of stronger individual specialization. High overlap among potential niches for soil variables suggested limited potential for individual differentiation in a highly suitable environment, while high overlap among realized niches for food resources, including the soil water index, earthworm abundance, and their interaction, indicated that individuals tended to use the same resources. Repeatability values ranged from 0 to 0.55 for potential niches and from 0 to 0.83 for realized niches, and in most cases realized repeatability exceeded potential repeatability, consistent with the hypothesis that northern lapwings exhibit individual variation in space use.

The authors emphasize that the workflow rests on explicit assumptions: individuals move in landscapes that on average match their phenotype, higher relative selection probabilities correspond to conditions providing greater expected lifetime reproductive success, and individualization is supported when realized niches show greater among-individual structure than potential niches. The framework is transferable across taxa of mobile organisms and responds to a decade of calls for incorporating individual-level variation into ecology and conservation. By quantifying individual niche variation, filtering for areas of high selection probability, and aggregating resource-use data into meaningful indices of intraspecific variation, the workflow promises more accurate and mechanistic individual-based models of wild population dynamics. Future research can focus on linking individualized niches to population growth rates, and because climatic niche breadth has been shown to predict recent range shifts across moths, butterflies, and birds, accounting for intraspecific variation may improve forecasts of persistence and redistribution under global change. As the individualized niche concept becomes established, quantifying it from movement data may prove an essential tool for behavioral ecologists seeking to draw population-level inferences from individual-level data.

Subject of Research: Individual specialization in habitat use quantified from animal movement data

Article Title: The individualized niche in motion: Quantifying individual specialisation with movement data

Article References: Takola, E. (2026). The individualized niche in motion: Quantifying individual specialisation with movement data. Individual-based Ecology, 2, Article e203247. https://doi.org/10.3897/ibe.2.203247

Image Credits: AI Generated

DOI: 10.3897/ibe.2.203247

Keywords: ecology, movement ecology, individualized niche, resource selection functions, hypervolumes, repeatability, northern lapwing, GPS tracking, niche breadth, niche overlap, individualized, niche