COS 246-4 - Human disturbance compresses the spatiotemporal niche

Background/Question/Methods

Human disturbance alters the environment and composition of animal communities, but little is known about the implications of human disturbance for species interactions. Our objective was to determine how human disturbance alters spatiotemporal cooccurrence of animals, since species must cooccur in space and time to interact. We hypothesized that human disturbance alternatively 1) compresses the space and time to be partitioned by communities, leading to greater spatiotemporal cooccurrence or 2) expands the space and time to be partitioned, leading to less spatiotemporal cooccurrence. To address these hypotheses, we used 4 years of data on 18 animal species (16 mammals, 2 terrestrial birds) from a network of >2,000 cameras maintained by community scientists in Wisconsin, USA. We used the time between detections of species pairs (n = 74) from cameras as an index of their spatiotemporal cooccurrence and created species networks for each camera based on pair cooccurrence. In addition, we ranked the antagonism level (low, medium, high) of each pair based off relative body size, since disturbance may have variable effects on different interactions. Finally, we used Bayesian hierarchical models to relate spatiotemporal cooccurrence and network statistics (the response variables) to human disturbance in the landscape surrounding each camera.

Results/Conclusions

Averaged across pairs, we found evidence of greater spatiotemporal cooccurrence in disturbed landscapes. Compared to the least-disturbed landscapes, cameras in disturbed landscapes showed 1.4 ± 0.4, 1.9 ± 0.4, and 2.1 ± 0.6 (mean ± standard deviation of model predictions) fewer days between successive detections for low-, medium-, and high-antagonism pairs, respectively. Increasing antagonism was associated with reduced spatiotemporal cooccurrence; at average values of human disturbance, high-antagonism pairs showed 0.30 ± 0.24 and 1.76 ± 0.36 days longer between detections than medium- and low-antagonism pairs, respectively, while medium-antagonism pairs showed 1.46 ± 0.32 days longer between detections than low-antagonism pairs. In support of the compression hypothesis, species networks in disturbed landscapes were denser (i.e., greater proportions of links observed between nodes) than those in less-disturbed landscapes, a trend that was particularly apparent in the summer. In addition, networks in disturbed landscapes were comprised of a greater proportion of low-antagonism pairs and smaller proportions of medium- and high-antagonism pairs than networks in less-disturbed landscapes. These results indicate that human disturbance generally compresses the spatiotemporal niche, suggesting that human disturbance restructures species interactions and may thus have far-reaching ecological effects.