Empty Forests and Extinction Filters

Conservationists face a challenge: more than a quarter of the world's mammal species are threatened with extinction. In an ideal world, we would know which species are vulnerable to what threats and how we can most effectively reduce species losses. The challenge is that many species are rare and unstudied, and we lack a clear understanding of what determines their vulnerability. This project seeks to remedy this by using wildlife photos and cutting-edge analyses and theory. Our main goal is to understand and predict vulnerabilities of tropical forest mammals to extinction. To accomplish this goal, the international project team is working on describing occupancy, activity and extinction patterns of mammal species across the tropics at different temporal and spatial scales. Additionally, the team is exploring how these patterns are affected by different extrinsic (e.g., ecological and anthropogenic factors) and intrinsic (e.g., species characteristics) factors. The project relies on camera-trap data collected by the Tropical Ecology Assessment and Monitoring (TEAM) Network, the largest standardized tropical forest monitoring system. This database comprises more than 200 mammal species, and over two million photos. The following are the project’s main accomplishments during the past year: 1. PhD student Andrea F. Vallejo Vargas has successfully defended her PhD thesis. Her thesis provides a comprehensive analysis of the diel activity patterns of tropical forest mammals and the factors explaining variation therein. All articles in the PhD thesis are either accepted or under review in high impact scientific journals. 2. Our article describing a novel statistical approach for quantifying and categorizing both diel and lunar behavior based on camera trap data has been accepted in Proceedings of the Royal Society B. 3. We expanded our network of collaborators to include researchers studying sensory ecology. We are now collaborating on an analysis that seeks to relate genes coding for opsins (light sensitive proteins) in mammals to their diel and lunar activity patterns to better understand the evolution of color vision. This new collaboration was made possible by our work on diel and lunar activity patterns in tropical forest mammals. 4. We have begun a new collaboration with U. Wageningen master’s student Hasna Afifah) to study mammal distribution and activity in relation to fruiting events in tropical forests Sumatra. As part of this collaboration Hasna Afifah has visited our lab to work together and learn our analytical approaches. 5. Project members (postdoctoral fellows) have continued developing their skills with Bayesian methods and are now proficient in hierarchical models that can address a range of questions. 6. We started expanding our hierarchical analytical framework to allow the analysis of acoustic data. Until now, we have relied primarily on camera trapping data, but the methods are readily applied to other non-invasive ecological field data. 7. We continued expanding models for Bayesian multispecies occupancy analysis, including a new dynamic occupancy model which resulted in the publication of a scientific article and several popular science articles. 8. Results from the project continue to be included in the curriculum of several courses at NMBU (conservation biology, global change ecology, ecological research methods, statistics, tropical ecology). 9. We continued and strengthened the Interdisciplinary Conservation Network that brings together experts from across NMBU and beyond. This is now a popular meeting place where researchers from all disciplines (philosophy, economics, ecology, social sciences, law, etc.) meet and exchange ideas about national and international conservation topics. 10. We have taken on a leading role in the coordination of the Tropical Forest Arena, an open arena for information exchange and debates in Norway on issues related to tropical forests. 11. We continued updating the project website.

Conservationists face a challenge: more than a quarter of the world's mammal species are 'threatened with extinction' (see http://www.iucnredlist.org/). In an ideal world, we would know which species are vulnerable to what threats and how we can most effectively reduce species losses. The challenge is that many species are rare and unstudied, and we lack a clear understanding of what determines their vulnerability. We address this challenge by making use of a unique opportunity. This opportunity exists in the form of over 3.4 million photographs recorded by automatic cameras ('camera traps'). These images were recorded in forests across the tropics and are the result of an immense systematic multi-year effort. The resulting data--comprising 244 mammal species, most of which are elusive and poorly known--along with cutting-edge analyses and theory, will clarify what makes species vulnerable. Using novel analytical approaches, we will explore the pattern of local species extinctions and re-colonizations, as well as where and when animals are active. This will permit a systematic characterization of animal presence and activity patterns across forests and continents that has never been previously achieved. How the species respond to local conditions and threats will be examined and the best means to predict these outcomes determined. One key proposal--the extinction filter hypothesis--is that history will help. Simply put, the idea is that species are more resilient to threats and conditions if they have survived similar threats and conditions in the past. We will test this, along with other ideas, to develop better guidelines for predicting which species are vulnerable to which threats under which circumstances. The project draws on not only a vast network of data but also of expertise. We foresee ground-breaking improvement in understanding and predicting species vulnerabilities, advancing both scientific knowledge and conservation practice.