LB 5-41 - Pairing automated mark–recapture and social network models to explore the effects of landscape configuration on hummingbird foraging patterns

Background/Question/Methods:

Landscape changes can alter pollinator movements and foraging patterns which can influence demographic processes of both pollinator and plant populations. Unfortunately, approaches for studying pollinator movement directly have been limited due to small body sizes. We leveraged social network models and fixed arrays of hummingbird feeders equipped with radio frequency identification (RFID) data loggers to study rufous hummingbird (Selasphorus rufus) foraging patterns in a heterogeneous landscape. Using a space-for-time approach, we asked whether forest encroachment on alpine meadows is likely to restrict hummingbird foraging movements and impede resource discovery. We implanted 163 rufous hummingbirds with passive integrated transponders (PIT tags) and maintained four arrays of five RFID-equipped feeders over four summers. We fit sender-receiver models developed for analyzing social network data in a generalized regression framework to data on 2,221 movements among feeders made by 29 hummingbirds and included information on the local vegetative cover (forested/non-forested) and characteristics of the intervening landscape as explanatory variables.

Results/Conclusions:

Movements were made primarily by females, likely due to male territoriality and early migration dates. Distance was the driving factor in determining the rate of movement among feeders. The estimated probability of at least one movement per bird per week between two locations approached zero at distances greater than two kilometers. The posterior mean effects of local forest cover and intervening forest on movement were estimated to be negative, but with considerable uncertainty. Isolating two meadows by increasing the amount of forest in the intervening landscape from 0% to 100% is expected to reduce the background movement rate between them by 45%. Similarly, movement between two food sources that are overgrown by woody plants is expected to be 69% less than if the two food sources are in the open. Finally, we found strong reciprocity in hummingbird movements (i.e., the rate of movement from feeder i → j ≈ j → i). Together, these findings suggest that reciprocal movements by female hummingbirds could help maintain bidirectional gene flow among nearby subpopulations of ornithophilous plants; however, if the distance among meadows increases with further forest encroachment, this may limit foraging among progressively isolated meadows.