Sleek ducks with long thin serrated bills designed to catch fish, the Brazilian merganser is one of the most threatened wildfowl species in the world.
WWT has been supporting The Terra Brasilis Institute, a local conservation organisation which is passionate about conserving the cerrado habitat in and around the Serra da Canastra National Park, and the mergansers that live there. Cerrado is an extensive savannah criss-crossed by gallery forests, which are evergreen corridors that have formed along the rivers and wetlands.
We helped write the action plan for conserving Brazilian mergansers. Drawing on our long experience of waterbird monitoring, we helped Terra Brasilis establish a capture and marking programme using leg rings and radio transmitters. It was the first time Brazilian mergansers had ever been caught for science and the ongoing programme is helping us to understand how the birds use the densely forested rivers.
We know that the birds are extremely sensitive to changes to the fast flowing rivers where they live and fish. Clearing cerrado to make way for maize and soya farming often leads to rapid soil erosion on steep slopes. This in turn causes the rivers to silt up, creating big problems for mergansers. Poorly managed livestock grazing and mining activities can have the same effect. The proliferation of hydro-electric dams in Brazil is also a worry: impounded lakes replace fast-flowing rivers, but are no substitute for a merganser. WWT, with its Brazilian partners, is planning to work with local people to demonstrate the benefits that soil protection can bring for agriculture, rivers and wildlife. WWT scientists are teaming up with the University of São Paulo to use satellite images to understand why Brazilian Mergansers have disappeared from so many areas, and why they survive in the places they do. We are also supporting the Federal University of Minas Gerais to examine the genetics of the species, so that we can understand whether mergansers are able to move between their different isolated remnant populations.