Answers are grounded in this paper and cite their sources.
Tracking study reveals that statistical significance of animal interactions depends heavily on null model choice, highlighting the need for species-specific agent-based models.
Probabilistic space-time prisms (PSTPs) are used to measure probabilities of location for individuals over time. Intersecting PSTPs for different individuals enables probabilities of co-location and proximity to be computed for each voxel, or volume element, that represents a time step for a specific location. This paper extends PSTP analysis by developing an approach to test the statistical significance of the calculated co-location and proximity measures. First, PSTPs were generated for 20 individual Mallard and Muscovy ducks tracked simultaneously. Second, the PSTPs were used to calculate probabilities of co-location and proximity for all 190 possible dyads. Third, species-specific agent-based models (ABMs) and a generic random walk model (RWM) were developed as null models to simulate tracking data over the same tracking duration, with 1000 replicates used to generate confidence intervals for each dyad. Finally, 95% confidence intervals for co-location and proximity obtained from the PSTP analysis were compared to 95% confidence intervals for co-location and proximity derived from the 1000 replicates of ABMs and RWM to evaluate if dyads were interacting more than, less than, or as expected under the null models. The results showed differences in co-location and proximity between dyads, but most importantly, the significance of those interactions was highly dependent on the null model used for comparison. The RWM typically predicted fewer co-locations and larger proximities in comparison to the species-specific ABMs, often yielding different conclusions about the direction and strength of interactions. The findings suggest that the significance of intra-and inter-species interactions should be evaluated in the context of the species' habitat and movement preferences through ABMs.
No takes yet. Share an insight, caveat, or question.
Smith et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: