COS 51-7
Determination of landscape-level impact and habitat factors of invasive beaver distribution in Tierra del Fuego

Tuesday, August 11, 2015: 3:40 PM
338, Baltimore Convention Center
Jonathan J. Henn, Austral Center for Scientific Research (CADIC-CONICET), Argentina
Christopher B. Anderson, Austral Center for Scientific Research (CADIC-CONICET), Argentina
Gaston Kreps, Austral Center for Scientific Research (CADIC-CONICET), Argentina
Guillermo Martinez Pastur, Austral Center for Scientific Research (CADIC-CONICET), Argentina
Background/Question/Methods

Invasive beavers in Tierra del Fuego are a research and management priority. However, important baseline information is lacking, including understanding the extent and distribution of their impacts at the landscape scale. Using Google Earth© images, we classified land cover in 175 1.5km2 plots evenly distributed throughout the Argentine side of Tierra del Fuego. In each plot we mapped any signs of beaver impact in the form of flooding caused by damming and forest cut for foraging. Estimates of impact upon each land cover class and ecoregion of the island were calculated based on these maps. Finally, spatial distribution of beavers was analyzed using join count analysis and Moran’s I, while a logistic regression model was fit using environmental variables including vegetation cover, forest type, stream characteristics, and topography to predict beaver occurrence across the island.

Results/Conclusions

In total, beavers have impacted approximately 31,476 ha of land. The majority of these impacts were found in forested areas, where 3.5% of the island-wide forest has been impacted. Most of these impacts are due to forest cutting, but flooding also had an important effect, especially in peat bogs. Spatial analysis indicated that the distribution, but not extent of beaver impacts are spatially clustered and that most impact occurred in the south-central mountainous area of Tierra del Fuego. The best model for predicting beaver presence includes variables related to water (peat bog cover and stream presence), forage availability (forest type cover), and topography (slope and elevation) and was able to correctly distinguish plots with beaver from plots without beaver 85% of the time. These initial results support previous assertions that this invasion is the largest alteration to sub-Antarctic forests in the Holocene and can be used to develop future maps based on habitat- and landscape-scale conditions to assist with the orientation of control, eradication, and restoration efforts currently being planned.