How are geographic ranges structured and how do they erode

Whether an ecologist focuses on habitat use, source-sink, or patterns of abundance across a species' geographic range, the observed patterns are the result of factors operating on various spatial and temporal scales. Here I argue that we can gain a better understanding of the distribution of individuals by integrating theories and observation from different scales into a coherent and general theory of how ranges are structured. Anderwartha and Birch's multipartite theory accomplishes many of the things that we would want from such a general theory. In the first chapter I revisit the multipartite theory and suggest how it can be modified to include recent developments. I then focus on the main impediment to applying the multipartite theory: the difficulty of defining the local population in practice. Finally, I introduce the novel concept of range erosion. Range erosion emphasizes the role that both range structure and the distribution of threats play in determining observed patterns of range loss or collapse.;The next two chapters focus on forest elephants. In Chapter 2, I examine four different models of how forest elephants move through their environment and the effects of temporal resolution on our understanding of movement. In Chapter 3, I test models based on the cumulative impact of multiple roads versus models based on distance to the nearest road, river, or populated place and find that models based on the cumulative impact outperformed the other models. This chapter also shows that the impact of roads on abundances is two-tiered, affecting both the probability that an area will be occupied and the expected abundance given that an area is occupied.;Chapter 4 focuses on modern range erosion in 43 species of wide-ranging mammals. I test whether variation in the vulnerability of species in different portions of its geographic range is an important predictor of range loss. Our results suggest that vulnerability is an important part of any explanation of range loss and that biomes are a better proxy for variation in vulnerability than position in the range. I close with a fifth chapter that highlights the connections between the preceding chapters.