Establish a SnapshotSafari collaboration and reap the benefits of an already-established and well-tested data collection and processing protocol.
Camera trap networks are revolutionizing the fields of ecology and conservation. Over the last two decades, hundreds of studies have adopted camera trapping as a non-invasive sampling tool that produces robust and accurate data at relatively low labor and cost investment. Cameras provide distinct advantages over direct observation, obtaining permanent records that are unaffected by human presence from locations that are often difficult to monitor. From these efforts, we can obtain species inventories, model species distribution, predict population dynamics, and gain insights into animal behavior. Ultimately, these data enable us to gain an understanding of the ecological underpinnings of a system, test hypotheses of community ecology, and inform management decisions.
Biodiversity Monitoring by the Serengeti Camera Trap Project
Funding from the National Science Foundation and National Geographic produced the world’s largest and longest continuous camera trap survey, SnapshotSerengeti, which covers a 1,100-km2 area in the center of Serengeti National Park. Over 200 camera traps have operated continuously from June 2010 until the present day, collecting millions of images of 48 different species ranging in size from mongooses to elephants. Cameras are attached to trees or on metal poles, and are activated by sensors that detect movement and variations in heat. Batteries and SD cards are changed every two months, and images are transferred to the University of Minnesota (UMN) at six-month intervals, where they are processed and uploaded into the Minnesota Super-Computer Institute. Volunteers from around the world view the images at www.snapshotserengeti.org, and classify the species present in each photo, count the number of individuals, record presence of offspring, and describe basic behaviors.
The SnapshotSerengeti camera trap grid recorded animal abundance and distribution over a sufficient proportion of the Serengeti to provide an unprecedented picture of population status and trends for mammalian, avian, and reptilian species. Most importantly, ongoing data collection can be used to measure impacts of environmental perturbations such as extreme weather events or illegal incursions inside the park. SnapshotSerengeti expands the questions that camera traps can address in ecological research and demonstrates that citizen science can effectively process large quantities of ecological data. Three doctoral students were trained as a result of this project, and Snapshot data has been used in the thesis work of an additional three post-doctoral, doctoral, and masters' students. Since the Snapshot dataset was published only last year (2015), at least eight research groups have used these data in studies that have resulted in two NSF proposals, six published papers, and numerous conference presentations, and we anticipate increasing interested in this data product.
Several doctoral students were trained at the University of Minnesota as a result of this project, and Snapshot data has been used in the thesis work of post-doctoral, doctoral, and masters students at other institutions as well. Multiple research groups have used these data in studies that have resulted in two NSF proposals, six published papers, and numerous conference presentations, and we anticipate increasing interested in this data product.
Proposed Monitoring for SnapshotSafari
Our goal is to establish a network of dozens of camera-trapping grids in national parks and reserves throughout southern and eastern Africa. We are well on the way, with 14 sites from six countries actively collecting data using our protocols, and many more coming online throughout 2018 and 2019.
Management Implications: Within reserves, long-term camera trapping efforts would provide unparalleled information on the status and trends of wildlife populations which may supplement or replace existing monitoring programs, in addition to being beneficial for security purposes. Consistent background monitoring provides an extra tool to adaptive management that enables reserves to measure perturbations across a system that result from management interventions. Besides providing information relevant to the management of the reserve itself, studies across reserves would enable the development of mathematical models for structured decision making that compare how the implementation of different management decisions changes state variables across different ecological systems.
Opportunities for Publicity: A SnapshotSafari camera trap network would continue the SnapshotSerengeti institution of citizen science involvement and public education. Over the last six years, SnapshotSerengeti has involved over 140,000 citizen scientists from around the world, who identify animals in millions of camera trap images. We engage the public through interactive discussion boards (“Snapshot Talk”), blogging, and other social media outlets (Facebook, Instagram, Twitter). Moreover, the website and blog posts have enabled thousands of K-16 students to classify photographs, investigate databases, and formulate and test their own research questions. Working with the National Geographic Society and the Howard Hughes Medical Institute, we are developing teaching tools that use Snapshot to introduce a wide range of students, from elementary to college undergraduates, to African ecosystems, the process of developing and answering a scientific question, and using statistics and the scientific method. As an example, we used SnapshotSerengeti to develop a teaching module for fifth graders called “Savanna Web”, which targets common core and next generation learning goals in the area of earth science, ecosystems and food webs. The module ends with a skype call from the classroom to one of our team members in Serengeti; students are able to ask questions directly to the researcher about their learning. We anticipate using new material from the wider diversity of southern African camera networks to build upon these initiatives.
Benefit to Individual Reserves & Supporting Organizations: As research and management outputs are published, participating reserves and funding organizations will develop an international reputation for contributing to scientific knowledge. In addition, a long-term large-scale camera-trap survey would be sure to generate thousands of captivating photographs showing animals engaging in behaviors that are seldom seen from a vehicle. The candid images generated by SnapshotSerengeti have appeared in books, press, and television documentaries. The appeal of the individual photographs encourages online volunteer participation, and particularly exciting images are often discussed and admired in our associated blogs and chat groups. Imagery from individual reserves would further increase public awareness of the reserve itself and be used to attract tourists or volunteers, and continue to engage these visitors after their departure.
Advantages of Snapshot Monitoring: One of the primary advantages of establishing a biodiversity monitoring program in line with the Snapshot design – beyond the exponentially greater scientific understanding of ecosystem processes generated by consolidating data collection efforts across a wide variety of reserves – is that the methodology and product have been validated by 6+ years of camera trapping in Serengeti. We have demonstrated that the deployment of a systematic grid of camera traps spaced according to our procedure produces population estimates and trends comparable to those generated from other surveying methods, such as ground transects or aerial surveys (see: Supporting Literature). Furthermore, we have developed a system for processing image data, cultivated a relationship with >100,000 citizen scientists, developed a platform for utilizing the processing power of the online volunteers, and validated that the consolidated effort of these untrained volunteers produces identifications that at ~97% accurate overall. Our database structure, cleaning codes, consolidation algorithms, and past experience will be used to streamline the extraction of data product from raw camera trap imagery.
Research Process: Camera traps are deployed according to the research design discussed in [Background]. UMNLC will work with funding organizations, NGOs, and reserves to raise money to deploy and maintain long-term camera-trap grids. Current UMN student Sarah Huebner will be available to assist staff in deploying camera-trap grids if requested. South African students or volunteers, under the supervision of reserve management, will maintain the camera trap grid and transmit images to the UMN Lion Center.
Images from camera-traps are cleaned, compressed, and uploaded to the citizen-science organization Zooniverse, where they are placed online for species classifications by volunteers. Once all the images have been viewed by 10-30 different volunteers, the data is condensed into a summary reports whereby information on each species can be easily extracted. These data will allow researchers and reserve managers to generate estimates of species abundances, track changes in animal numbers through time, and improve security at each reserve. In addition, these data will be made available for scientific and educational purposes
- Snapshot Camera Trap Collaboration Overview
- Snapshot Camera Trap Deployment Guidelines
- Snapshot Camera Trap Metadata Guidelines
- Snapshot Camera Trap Database Guidelines
- Snapshot Camera Trap Directory Structure Guidelines