Assessing strategies to reconcile agriculture and bird conservation in the temperate grasslands of South America.

Globally, agriculture is the greatest source of threat to biodiversity, through both ongoing conversion of natural habitat and intensification of existing farmland. Land sparing and land sharing have been suggested as alternative approaches to reconcile this threat with the need for land to produce food. To examine which approach holds most promise for grassland species, we examined how bird population densities changed with farm yield (production per unit area) in the Campos of Brazil and Uruguay. We obtained information on biodiversity and crop yields from 24 sites that differed in agricultural yield. Density-yield functions were fitted for 121 bird species to describe the response of population densities to increasing farm yield, measured in terms of both food energy and profit. We categorized individual species according to how their population changed across the yield gradient as being positively or negatively affected by farming and according to whether the species' total population size was greater under land-sparing, land-sharing, or an intermediate strategy. Irrespective of the yield, most species were negatively affected by farming. Increasing yields reduced densities of approximately 80% of bird species. We estimated land sparing would result in larger populations than other sorts of strategies for 67% to 70% of negatively affected species, given current production levels, including three threatened species. This suggests that increasing yields in some areas while reducing grazing to low levels elsewhere may be the best option for bird conservation in these grasslands. Implementing such an approach would require conservation and production policies to be explicitly linked to support yield increases in farmed areas and concurrently guarantee that larger areas of lightly grazed natural grasslands are set aside for conservation.

Birds of a eucaliptos woodlot in interior São Paulo

Some 255 birds were recorded between 1982-2001 in and near a 2314-ha "Horto" of old eucalyptus plantations with native understory and a lake, near Rio Claro, in central São Paulo, Brazil. This is close to the 263 recorded in and around a ten-times smaller nearby 230-ha woodlot of semideciduous forest. Different species were 44, for a total of 307 in both areas. One hundred and fifty nonvagrant forest and border species were recorded in 1982-86, a number close to the 152 in the small native woodlot. With dry years and logging of plots in 1985-93, 21 of the 150 species were lost, 42 species decreased in numbers, 49 were stable, 19 increased (15 being border species), and 5 entered (one of dry forest and 4 of borders), so 129 species remained in 1996-2001 compared to 133 in the native woodlot. Open-area birds were 33, versus 50 in better-checked grassy swales in sugar cane near the natural woodlot, for a total of 53. Several species, like some border ones, did not enter the open but isolated and mowed interior lake area, or took years to do so. Water and marsh birds were 46 versus 40 in smaller creeks and ponds near the natural woodlot (total, 55) but many were migrants or infrequent visitors using distant areas, and perhaps should be counted as 0.1-0.9 "local species" rather than "1" species. Use of this more accurate method would reduce waterbird totals by 14 "species" in the Horto and by 11 around the native woodlot. I also recommend longer censusing at the edges in large woodlots or many edge species will be recorded only in small fragments of habitat. Several species increased and others decreased with occasional cat-tail and water-lily cleanups at the lake. A forested corridor between the Horto and natural woodlot is recommended, with old eucalyptus left to provide flowers for hummingbirds.

Cerca de 255 espécies de aves foram registradas entre 1982-2001 dentro e próximo do Horto Florestal, em 2314 ha de plantações antigas de eucaliptos e sub-bosque nativo com um lago, próximo a Rio Claro, no centro do Estado de São Paulo, Brasil. Esse registro é próximo das 263 espécies registradas dentro e próximo de uma mata natural dez vezes menor, com 230 ha. As espécies diferentes foram 44, para um total de 307 em ambas as áreas. Cento e cinqüenta espécies não-vagantes de mata e borda foram registradas em 1982-86, assim como 152 na mata nativa pequena. Com os anos secos e a retirada de madeira em 1985-93, 21 das 150 espécies desapareceram, 42 espécies diminuíram em números, 40 permaneceram estáveis, 19 aumentaram (15 sendo espécies de borda) e 5 surgiram (uma de mata seca e 4 de borda), portanto, 129 espécies permaneceram em 1996-2001 se comparado com 133 na mata nativa. As aves de áreas abertas foram 33, versus 50 em áreas de canaviais junto à área florestada, para um total de 53. Várias espécies, assim como algumas de borda, não entraram em áreas abertas do lago, isoladas e capinadas, ou levaram vários anos para entrar. As aves aquáticas e de brejo foram 46 versus 40 em pequenos riachos e lagos próximos à área natural (total 55), mas muitas eram migrantes ou visitantes infreqüentes usando áreas distantes, e talvez devessem ser contadas como 0,1-0,9 "espécie local", em vez de "1" espécie. O uso deste método mais refinado reduziria os totais para aves aquáticas de 14 "espécies" no Horto e de 11 nos arredores da área nativa. Também, recomendo efetuar o censo por longos períodos nas margens de grandes áreas, ou muitas espécies de borda serão registradas somente em pequenos fragmentos de habitat. Várias espécies aumentaram e outras diminuíram com as ocasionais limpezas do lago e retiradas de taboais e lírios-de-água. Um corredor de mata entre o Horto e a área natural é recomendado, com os velhos eucaliptos se encarregando de proporcionar flores para os beija-flores.