Abstract
Across landscapes, shifts in species composition often co-occur with shifts in structural or abiotic habitat features, making it difficult to disentangle the role of competitors and environment on assessments of patch quality. Using 22 years of rodent community data from a long-term experiment, we show that a small, ubiquitous granivore (Chaetodipus penicillatus) shifted its use of different experimental treatments with the invasion of a novel competitor, C. baileyi. Changes in population metrics such as residency, probability of movement between patches, and the arrival of new individuals in patches resulted in changes in which treatment supported the highest abundances of C. penicillatus. Our results suggest that the invasion of a new species worsened the quality of the originally preferred treatment, probably through its impacts on resource availability. Paradoxically, the invader also increased C. penicillatus’ use of the less preferred treatment, potentially through shifts in the competitive network on those plots.
Footnotes
We made the decision to reframe the manuscript, shifting away from metacommunity paradigms and placing our work in more of a habitat- or patch-selection context. As such, we were able to make stronger connections to key papers on habitat selection, especially those using rodent experiments similar to the Portal Project. To clarifying our thoughts on some of the potential incongruencies in the results and make stronger connections to theory, we have added a more detailed explanation in the discussion pulling from coexistence theory in competitive networks, a body of literature which we had not previously considered. The addition of a novel competitor in a competitive network can lead to an 'enemy of my enemy is my friend' scenario, which could potentially explain why the control plots were suddenly perceived as better habitat after the invasion of C. baileyi. We made our phrasing clearer that we do not have data on perception of patch quality but that our results simply suggest possible changes in perception. Finally, we added a new analysis to the supplemental material showing that density-dependent habitat selection in not a good explanation for the shift.
