Differential matrix use by Neotropical birds based on species traits and landscape condition.

There is considerable uncertainty about the role of human-modified habitats in supporting species in fragmented landscapes. This is because few studies sample outside of native habitats in the "matrix". Those that do, often fail to sample landscapes in a way that accounts for the confounding effects of native habitat pattern and species detection biases that can obscure species responses. We employed multi-species hierarchical occupancy models to determine the use of human-modified habitats by Neotropical birds in landscapes that were similar in forest amount and configuration but surrounded by a matrix of agriculture (predominately pasture), bauxite mining (surface mining for aluminum), or suburban development in central Jamaica. We found that the vast majority of bird species used the matrix: with the highest mean occurrences for open-associated, followed by generalist, and lastly forest-associated species. Migrant species had higher mean occurrences in all matrix types relative to resident species. Contrary to our expectation, mean occurrence for the entire species community, and for forest-associated and migrant species, were highest in bauxite, intermediate in suburban, and lowest in agriculture. Open-associated species had higher occurrences in both bauxite and agricultural matrices, whereas generalist species had higher occurrences in suburban matrices. Additional behavioral observations indicated that Neotropical birds used matrix areas, particularly scattered trees, to acquire food, and secondarily, as movement conduits. Matrix use patterns reflected the differential availability of potential resources and structural connectivity across the three landscape types, but only for those species adapted to open/edge environments and with generalized habitat requirements. Patterns of matrix use by forest specialists reflected the differential levels of degradation of the native forest; thus, we propose that higher matrix use for forest-dependent species may be induced by diminished within-forest resources. These results underscore that effective management of human-modified matrices requires in-depth understanding of the trade-offs between the benefits available in the matrix and the impacts on the disturbance of native habitats. This article is protected by copyright. All rights reserved.