Date of Award


Degree Name

Biological Sciences


College of Science

Type of Degree


Document Type


First Advisor

Anne Axel

Second Advisor

Jayme Waldron

Third Advisor

Suzanne Strait


The ability to assess the quality or ecological value of a landscape is a useful tool in effective conservation and natural resource management. Ecosystem biodiversity and functionality are reduced when human activities cause habitat alteration and/or fragmentation. Compromised habitats are unlikely to support naturally occurring communities or healthy ecosystem functions. Conservation scientists have historically concentrated their efforts on large, intact habitats with little human disturbance; the conservation value of small and/or disturbed areas has been comparatively overlooked. One limitation may be a lack of tools to identify disturbed areas with conservation value. Techniques that allow researchers to assess the habitat quality of disturbed-but-functioning areas with conservation potential are especially useful in a threatened ecosystem like the tropical dry forests (TDFs) of Madagascar.

Madagascar has an extremely high number of endemic species -- species unique to the large island nation -- as well as widespread impacts from human activities. This has resulted in large swaths of disturbed habitat across the vast majority of the island’s landscapes. A variety if forest types are found in Madagascar including: evergreen (coastal forest, low altitude, mid altitude, and lower montane rainforest, montane scrubland, and woodland), deciduous (coastal, western seasonally deciduous, and southern dry deciduous forest), mangrove, and marshland (DuPuy and Moat 1996). The southern tropical dry forest makes up about 1400 km2 or 15% of the forest cover in Madagascar. This forest is used for agriculture and forest product extraction but is mostly impacted by livestock grazing. Given that the vast majority of remaining forests in this region are unprotected, it is likely that the forests in the region are grazed and have been for many years.

We evaluated the effects of livestock on the tropical dry forest through vegetation, satellite, lemur occupancy, and acoustic sampling at 24 sites within the Beza Mahafaly Special Reserve. Sites were divided between two forest types (gallery and dry deciduous) and three levels of grazing intensity (ungrazed, moderately grazed, and heavily grazed). Canopy and understory measurements varied significantly by forest class and grazing intensity. Land cover classification using satellite imagery resulted in highly accurate assignment of forest class and grazing presence, but not grazing intensity. Lemur occupancy varied by forest class, yet grazing presence resulted in lower occupancy in only one of two lemur species sampled. Biological sound activity was significantly different by season, yet did not differ by forest class or grazing intensity. The effects of grazing on vegetation and wildlife appear to be greater in the dry deciduous forest than in the gallery forest. Remote monitoring methods used in this study are promising for future monitoring schemes and land management decisions.


Animal feeding - Madagascar

Tropical dry forests - Madagascar