Approximately 20% of the Brazilian Amazon has now been deforested, and the Amazon is currently experiencing the highest rates of deforestation in a decade, leading to large-scale land-use changes. Roads have consistently been implicated as drivers of ongoing Amazon deforestation and may act as corridors to facilitate species invasions. Long-term data, however, are necessary to determine how ecological succession alters avian communities following deforestation and whether established roads lead to a constant influx of new species.  

We used data across nearly 40 years from a large-scale deforestation experiment in the central Amazon to examine the avian colonization process in a spatial and temporal framework, considering the role that roads may play in facilitating colonization.

Since 1979, 139 species that are not part of the original forest avifauna have been recorded, including more secondary forest species than expected based on the regional species pool. Among the 35 species considered to have colonized and become established, a disproportionate number were secondary forest birds (63%), almost all of which first appeared during the 1980s. These new residents comprise about 13% of the current community of permanent residents.

Widespread generalists associated with secondary forest colonized quickly following deforestation, with few new species added after the first decade, despite a stable road connection. Few species associated with riverine forest or specialized habitats colonized, despite road connection to their preferred source habitat. Colonizing species remained restricted to anthropogenic habitats and did not infiltrate old-growth forests nor displace forest birds.

Deforestation and expansion of road networks into terra firme rainforest will continue to create degraded anthropogenic habitat. Even so, the initial pulse of colonization by non-primary forest bird species was not the beginning of a protracted series of invasions in this study, and the process appears to be reversible by forest succession. 

We generated the avian regional species pool (n=725 species) for the Biological Dynamics of Forest Fragments Project (BDFFP; 2°20′ S, 60°W), ~80 km north of Manaus, Amazonas, Brazil. To do so, we used a few simple criteria: 1) the species must have been previously recorded in the Amazon (total ~1300 species), 2) for terra firme species, we only included birds that are known from the Guiana area of endemism, and 3) we imposed distance cutoffs of ~500 km from the BDFFP for resident species and ~1000 km for migratory species. We then curated the resulting list by hand to ensure that the final list matched current knowledge. These criteria necessarily mean that all species that have already been detected from the BDFFP are included. 

To this list, we added three additional columns. The "BDFFP list" column denotes all species that have been recorded at the BDFFP between 1979 and 2017 (n=407 species; see Rutt et al., 2017). Another column includes a list of the birds that are part of the "core forest avifauna" at the project (n=268 species), or those species that are regularly found in primary terra firme forest. Terra firme species are only designated as part of the core avifauna if they reached a relative abundance of rare, uncommon, or common during one of the last two avifaunal inventories (Cohn-Haft et al., 1997; Rutt et al., 2017). Finally, the last column categorizes each species by habitat according to the Parker et al. (1996) databases. When appropriate, we used the first (primary) habitat type that was listed therein; however, we made adjustments if the primary code suggested the species occurred in habitat not found in the central Amazon (e.g., montane forest, temperate grassland). In those cases, we accepted secondary or tertiary habitat codes. We then collapsed these 22 categories (21 distinct habitats plus ‘Edge’) for the regional species pool into a more manageable seven that adequately captured habitat diversity in the immediate vicinity of the BDFFP: aquatic, primary forest, riverine, secondary forest, white sand, palm, and grassland/pasture (see Appendix 1 in the paper).

Taxonomy follows the South American Classification Committee (J. V. Remsen, Jr. et al., 2018). See the Methods section of the paper for further details.

References

Cohn-Haft, M., Whittaker, A., & Stouffer, P. C. (1997). A new look at the "species-poor" central Amazon: the avifauna north of Manaus, Brazil. Ornithological Monographs, 48, 205-235.

Parker, T. A., Stotz, D. F., & Fitzpatrick, J. W. (1996). Ecological and distributional databases. In D. F. Stotz, J. W. Fitzpatrick, T. A. Parker, & D. Moskovits (Eds.), Neotropical Birds: Ecology and Conservation (pp. 113-407). Chicago, IL: University of Chicago Press.

Remsen, J. V., Jr., Cadena, C. D., Jaramillo, A., Nores, M., Pacheco, J. F., Pérez-Emán, J., . . . Zimmer, K. J. (2018). A classification of the bird species of South America. American Ornithologists' Union. http://www.museum.lsu.edu/~Remsen/SACCBaseline.htm

Rutt, C. L., Jirinec, V., Johnson, E. I., Cohn-Haft, M., Vargas, C. F., & Stouffer, P. C. (2017). Twenty years later: an update to the birds of the Biological Dynamics of Forest Fragments Project, Amazonas, Brazil. Revista Brasileira de Ornitologia, 25(4), 259-278.