Biodiversity conservation in tropical agriculture: New and old world perspectives

The four chapters within this dissertation focus on critical unanswered questions regarding the conditions under which the tropical countryside can sustain native biodiversity, both in isolation and in conjunction with protected areas. The first chapter addresses one of the highest-profile species in conservation: the tiger. The tiger conservation potential of a protected area is inevitably affected by land use choices within the larger unprotected landscape. This chapter presents a tiger population model for the species, centered on the Indian subcontinent, to determine whether a strategy focused exclusively on protected areas would lead to a better outcome for the species than a strategy that gave some consideration to unprotected landscapes. The research shows that the vast majority of protected areas in the subcontinent are incapable of supporting tigers without significant attention paid to the landscapes around them. The second chapter concentrates on the long-term potential of tropical countryside to sustain native diversity. To address this concern, we examined patterns of avian persistence within a human-dominated landscape in southwest India with a very long history of continuous agricultural production (> 2000 years). This chapter shows that the tropical countryside can support biodiversity over the long term, as almost all of the members of the pre-cultivation bird species pool still are found in the region. The third chapter focuses on the response of bird communities to fine-grained variation in land cover, within the tropical countryside, and on simple remote-sensing measures that correspond with that response. The chapter indicates that multiple components of bird communities responded to this fine-grained variation and that simple remote-sensing tools have the potential to predict this response. The last chapter expands on the previous analysis to examine the effect of local land cover variation in the tropical countryside on a wide range of taxa: birds, mammals, bees, herpetofauna, moths, and plants. The same remote sensing measures utilized in the previous work also strongly correlated with community variation across all taxa. This suggests that simple remote sensing metrics can used to predict the response of many disparate components of native biodiversity to local land cover change.