Matrix quality determines the strength of habitat loss filtering on bird communities at the landscape scale

Habitat loss represent a major threat to biodiversity, however, the modulation of their effects by the non-habitat matrix surrounding habitat patches is still undervalued. The landscape matrix might change community assembly in different ways. For example, low-quality matrices can accentuate environmental filtering by reducing resource availability and/or deteriorating abiotic conditions but they may also over limit dispersal of organisms and make communities more prone to ecological drift. To understand how matrix quality modulates the effects of habitat loss, we quantified the relative importance of environmental filter and ecological drift in bird occurrences across more local (400 m buffer around sampling sites) and broader (2 km focal landscapes) gradients of habitat loss embedded in low- and high-quality matrices. We used a trait-based approach to understand habitat loss filtering effects on forest specialist and habitat generalist bird occurrences. We found that low-quality matrices, composed mainly of low-productive pasturelands, increased the severity of habitat loss filtering effects for forest specialist birds, but only at the landscape scale. Bird occurrence was in general higher in high-quality matrices, i.e., more heterogeneous and with low-contrasting edges, indicating the role of the matrix quality on attenuating species extinction risks at the landscape scale probably due to mass effect. Moreover, forest specialists presented a strong negative response to habitat loss filtering across different functional traits, while generalists presented a high variability in traits response to habitat loss. We raised evidence in supporting that landscape habitat loss filtering may be relaxed or reinforced depending on the quality of the matrix, evidencing that matrix quality has a strong impact in modulating community assembly processes in fragmented landscapes. In practical terms, it means that improving matrix quality may help in maintaining the high diversity of birds even without any increase in native forest cover.