Research Group -Marion Adeney
Marion Adeney is co-advised by Drs. Stuart Pimm
and Norm Christensen.
Currently on a NASA Earth Systems Science Fellowship, I am interested in tropical ecosystems, land use change, fire, remote sensing of wetlands and conservation. My research focuses on fire, effectiveness of Amazonian protected areas, and seasonally flooded white sand savannas (campinas) of the Amazon.
Amazonian flooded campos
Amazonian flooded campos, often called white sand campinas (also campos naturais, campinarana, bana, varillál) are a seasonally flooded woodland/savanna-like vegetation found largely on sandy soils in the Amazon. Part of a spectrum of Amazonian savannas, they host a unique community of plants and animals, equipped to deal with harsh conditions of flooding, nutrient scarcity and fire. Many occur in small patches, like islands, surrounded by forest. An exception is the enormous area in the north-western Brazilian Amazon, which drains into the Negro and Branco watersheds. Largely unstudied because of their extreme remoteness and inaccessibility, white sand campinas continue to yield surprises, including at least six new bird species discovered in the last decade. My research on campinas focuses on large-scale patterns of flooding and fire, which are important drivers of these systems, as well as vegetation classification and habitat modeling for birds.
Vegetation and bird habitat modeling
The dearth of scientific work in these unique ecosystems was recently highlighted by the surprising discovery of a new jay (Cyanocorax) in white sand campinas within 150 km of Manuas (M. Cohn-Haft, pers. comm.). The jay occurs in a complex of white sand ‘islands’, located between the Madeira and Purus rivers, south of Manaus, Brazil. In collaboration with Dr. Cohn-Haft of Brazil’s National Institute for Amazonian Research (INPA), and the GEOMA project of INPA and Brazil’s National Institute for Spatial Research (INPE), I am creating a vegetation / habitat map and a model for bird habitat in these white sand campinas. Because of plans for large-scale infrastructure, including the repaving of highway BR-319, which bisects the interfluve, this area is of immediate conservation concern. Anthropogenic fire, mining for sand and agriculture are all immediate threats when white sand campinas become accessible. Through collaboration with the GEOMA project, I hope that my work can contribute to conservation planning in this sensitive area.
Flooding and fire
The proposed origins of white sand campinas vary from successional vegetation recovering from past disturbance by indigenous peoples, to in situ weathering of soil, to sand deposited in ancient dunes. Origins likely vary by location, but patterns of fire and flooding are also probable drivers of these unique systems. White sand campinas are saturated for several months of the year, and in some areas become flooded, with water reaching depths of a meter or more. The length and amount of saturation affects vegetation growth and fire may also play an important role. I am using the MODIS fire product and daily MODIS surface reflectance, to examine patterns of fire and flooding in the white sand campinas of the Rio Negro, from the year 2000 to present.
Patterns of fire in the Brazilian Amazon
In the Brazilian Amazon, fire closely accompanies roads and deforestation. Using satellite fire data (hot pixels), I looked at how well protected areas prevent fire, especially when they are bisected by roads. Over ten years of data, both inside and outside of protected areas, the vast majority of hot pixels occurred within ten km of roads, in areas of high human impact, and in ENSO years. However, all types of protected areas (including fully protected parks, areas that allow limited habitation and use, and indigenous lands) have much lower rates of fire than unprotected areas. This is true not only at borders, but also along roads that bisect protected areas. In fact, the difference in fire rates inside versus outside protected areas was greatest near roads, meaning protected areas have the most impact in especially vulnerable places. There are regional differences, but on the scale of the Brazilian Amazon, there is no clear difference between types of protected areas. This paper is in preparation.
Fire and bird communities in Sumatra
After a fire burns in a tropical forest, what happens to the birds? I looked at changes in bird community composition along a gradient of burn severity in a humid lowland rainforest in Sumatra, Indonesia. I used mist nets, and five years of point count data, starting directly after a fire burned in 1997. Bird community composition changed in direct relation to the severity of the fire’s impact. In badly burned areas, diversity was higher, but the bird community contained many more generalist species. Forest specialists, especially understory birds, declined dramatically with increasing burn severity.
I did this work for my master’s degree in Conservation Biology at Columbia University (NYC), advised by Dr. Joshua Ginsberg of the Wildlife Conservation Society.
Adeney, J. M., J. R. Ginsberg, G. J. Russell, and M. F. Kinnaird. 2006. Effects of an ENSO-related fire on birds of a lowland tropical forest in Sumatra. Animal Conservation 9:292-301.(pdf available)
E-mail: marion.adeney@duke.edu