Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Susan Parks


African forest elephant, Loxodonta cyclotis, passive acoustic monitoring, poaching, remote sensing, vocalization


The African forest elephant (Loxodonta cyclotis) is a critically endangered and cryptic species that inhabits the rainforests of Central Africa. Forest elephant populations are severely threatened by poaching for the ivory trade, and an improved understanding of forest elephant behavior and habitat use, and of the anthropogenic pressures that threaten their existence, is essential for conservation of the species. However, their remote tropical rainforest habitat poses logistical constraints on research and makes forest elephants very difficult to observe and study visually. Limited data collection methods have also inhibited our ability to understand the determinants of poaching activity that is driving forest elephants toward extinction. This dissertation addresses forest elephant behavior, ecology, and conservation questions that span multiple scales by capitalizing on the advantages of passive acoustic monitoring (PAM) to detect elephant vocalizations and gunshots. At the finest scale, Chapter 1 examines forest elephant vocal repertoire use at a forest clearing in the Central African Republic and discusses implications for PAM. The different vocalization types of the repertoire varied in the generality or specificity by which they were used by certain age-sex classes of elephants. An understanding of these patterns is important for PAM of forest elephants, as they determine the population (or subset) that is detected and sampled. At the intermediate scale, Chapter 2 examines forest elephant landscape-scale response to individual poaching events detected in a PAM study system. Elephants within 10 km of gunfire events responded to poacher presence (before gunshots were fired) and to gunshots themselves, exhibiting behavioral changes in either vocal activity, site usage, or both. These results suggest that, in addition to the outright killing of targeted individuals, poaching activity affects the general population of elephants across the landscape. At the broadest scale, Chapters 3 and 4 used detections of elephant vocalizations and gunshots to analyze the distributions of forest elephants and poaching events across a 50-sensor PAM grid spanning 1250 sq. km of rainforest in Republic of Congo, for a period of over 3 years. To elucidate the determinants of these distributions, elephant and gunshot detection data were combined with habitat and landscape variables quantified using satellite remote sensing. In Chapter 3, variation in poaching risk depended primarily on factors related to poacher accessibility, such as distance to major rivers and logging roads. These results can guide the allocation of anti-poaching patrol effort to cover high-risk areas at times of increased vulnerability. Chapter 4 examined the habitat resources and anthropogenic pressures (e.g., poaching and logging) that influence forest elephants’ use of the landscape. Elephant occurrence probabilities decreased over the 3 years of the study and were seasonally dependent, increasing in the wet season. Ongoing logging activity deterred forest elephants from using nearby sites, but previously logged areas provided important habitat resources. By leveraging remote sensing methods to expand the scale and resolution of data collection, this dissertation aimed to advance our understanding of forest elephant behavior and ecology, and confronted questions that will improve conservation efforts to protect the species from extinction.


Open Access