Data from: Estimating interactions between individuals from concurrent animal movements

1. Animal movements arise from complex interactions of individuals with
their environment, including both conspecific and heterospecific
individuals. Animals may be attracted to each other for mating, social
foraging, or information gain, or may keep at a distance from others to
avoid aggressive encounters related to, e.g., interference competition,
territoriality, or predation. With modern tracking technology, more data
sets are emerging that allow to investigate fine-scale interactions
between free-ranging individuals from movement data, however, few methods
exist to disentangle fine-scale behavioural responses of interacting
individuals when these are highly individual-specific. 2. In a framework
of step-selection functions, we related movements decisions of individuals
to dynamic occurrence distributions of other individuals obtained through
kriging of their movement paths. Using simulated data, we tested the
method’s ability to identify various combinations of attraction,
avoidance, and neutrality between individuals, including asymmetric (i.e.
non-mutual) behaviours. Additionally, we analysed radio-telemetry data
from concurrently tracked small rodents (bank vole, Myodes glareolus) to
test whether our method could detect biologically plausible behaviours. 3.
We found that our method was able to successfully detect and distinguish
between fine-scale interactions (attraction, avoidance, neutrality), even
when these were asymmetric between individuals. The method worked best
when confounding factors were taken into account in the step-selection
function. However, even when failing to do so (e.g. due to missing
information), interactions could be reasonably identified. In bank voles,
responses depended strongly on the sexes of the involved individuals and
matched expectations. 4. Our approach can be combined with conventional
uses of step-selection functions to tease apart the various drivers of
movement, e.g. the influence of the physical and the social environment.
In addition, the method is particularly useful in studying interactions
when responses are highly individual-specific, i.e. vary between and
towards different individuals, making our method suitable for both
single-species and multi-species analyses (e.g. in the context of
predation or competition).