Understanding how diversity responds to habitat heterogeneity in a landscape is a central issue for implementing effective conservation plans. In this study, we analyzed how the composition and abundance of neotropical bird assemblages vary among habitats in the Ñeembucú Wetlands Complex, the largest wetland system in Paraguay. Bird surveys were conducted during one year within dominant habitats in the landscape: riparian forests, natural grasslands, savannas, and anthropized sites. The Ñeembucú Wetlands Complex showed a high bird richness (209 species) that contrasted with a pattern of dominance by a small set of 16 species that comprised half of the abundance. This set of oligarchic species consists of generalist species that exploited a wide variety of habitats and were well adapted to human disturbance, contributing to an important overlap in the composition of assemblages.  However, despite this overall similarity, there is still significant differentiation in bird assemblages, especially between habitats with the most contrasting vegetation physiognomy. Riparian forests and savannas showed higher diversity values than the more structurally homogeneous grasslands. Also, all natural habitats showed, in general, higher diversity values than anthropized sites. Oligarchic species had higher importance in anthropized environments, which also had the most uneven distribution of abundance between species, indicating the dominance by a smaller number of species and the loss of ecological diversity as farmland and urban development increase.

Study area

The study areas are located in the Humid Chaco ecoregion, in the Ñeembucú Department (southern Paraguay), and belong to the Ñeembucú Wetlands Complex. The regional climate is tropical humid with a marked seasonality: the mean annual precipitation averages 1400-1500 mm but with a dry season in winter. The average temperatures are around 16-17°C in winter (May-August) and 25-26°C in the warmer season (December-March). The altitude of the study area ranges between 50 and 124 m above sea level (a.s.l.).

Four dominant habitat types were considered in the landscape: riparian forests, natural grasslands, savannas, and anthropized habitats. Riparian forests are directly or indirectly associated with water, as they are subject to periodic flooding during the summer; they can be found along the edges of rivers, streams, and lakes or forming forest islands surrounded by wetlands and floodplains. Primary forest patches have a main canopy that is up to 25 m high, dominated by Enterolobium contortisiliquun, Syagrus romanzoffiana, Aspidosperma quebracho-blanco, Libidibia paraguariensis, Microlobius foetidus, and Guazuma ulmifolia.  Savannas are also subject to periodic flooding and are found on soils with high clay and salt content. The dominant species is the palm Copernicia alba, with a sparse or absent shrub layer, and an herbaceous cover dominated by Hyptis lappacea, Pluchea sagittalis, and Chromolaena christiaeana. In the paleobasins of the Paraná River, where soils are sandy, the landscape is dominated by grasslands of Andropogon spp., Coleataenia prionitis, Elionurus muticus, Sorghastrum pellitum, and Schizachyrium spp., and sometimes with the presence of scattered woody species such as Cecropia pachystachya and E. contortisiliquum. The selected anthropized count points included modified sites with a clear human influence that were mainly located on ranches and the periphery of small townships and villages. They generally comprised individual houses with yards, barns, orchards, small pastures, and unpaved roads.

Field sampling and data collection

In the terrestrial habitats identified in the study area, bird surveys were carried out during 158 point counts between October 2020 and July 2021. Forty-three points were selected in riparian forests, 30 in natural grasslands, 50 in Copernicia savannas, and 35 in anthropized sites.  The number of points per habitat was based on their importance within the landscape, but not directly proportional due to accessibility constraints and security issues. The sampling points were distributed along three transects in the districts of Alberdi, Guazú-Cuá, San Juan Bautista, and Tacuarás Villa Franca and were separated by a minimum distance of 500-600 m to avoid counting the same individuals twice. Although it was initially planned to sample each point count three times, for security reasons 41% of the points (11 of 43 points in riparian forest, 14 of 30 in natural grasslands, 23 of 50 in savannas and 17 of 35 in anthropized sites) could only be sampled twice; the rest were sampled three times. At each point count, all birds seen or heard within a radius of 50 m from the center of each site within 5 min of arriving at the sample point were counted. Bird sampling began at dawn, from 6:00 to 10:00 a.m., and during the afternoon, from 4:00 to 7:00 p.m. Of the total of 408 inventories, 143 were carried out in the dry season (May-September), and 265 in the wet season (October-April). Adverse weather conditions (heavy rains) were avoided, and all bird sampling was carried out by the same person (Fátima Ortíz).