It’s indisputable that technology has made our lives easier and often times better. And more and more frequently technology is making it into our work in conservation. Just look at camera traps (trail cameras) and how much of a game-changer they were for remote surveying of large areas to assess wildlife populations. One of the latest emerging technologies to change the world in recent years is AI, and one organization has decided to use it to help conservation. The team at Tech4Conservation with software developers at WildME, with assistance from our researchers and our extensive predator database, have created the African Carnivore Wildbook, or ACW. The ACW is an open source platform that uses the power of AI to individually identify animals from timestamped and georeferenced photographs, a task that can take researchers months. It stores all of this information in a searchable database that researchers have access to for free.
Why is identifying individuals even important I hear you ask? Well, if you want to be able to answer some key ecological questions relating to occupancy, survival, movement and dispersal, reproductive success and modelling populations where all individuals are considered explicitly, then you need to successfully identify and follow the lives of individuals. We recently published a paper describing a case study using an example of African wild dogs: An AI-based platform to investigate African large carnivore dispersal and demography across broad landscapes
This study investigates how the ACW is facilitating the monitoring of the African wild dog, one of the more challenging carnivores to monitor effectively due to the large distances that they can travel, especially when they disperse from where they were born to new areas to try to establish new packs. By collecting photographs from multiple sources across multiple countries, from researchers and tourists alike, the ACW enables individuals to be tracked across time and space easily. The ACW encourages collaboration between research institutions across countries by ensuring that when a photograph is being matched, it is compared to data from the whole database regardless of who owns it. If a match is made, collaborations are forged between different research institutions that are following the same individuals. In this way, long distance movements are captured. An example of just this occurred when in 2018 tourists saw and photographed some wild dogs in northeastern Botswana and sent the photos to us here at BPC. We loaded their photos to ACW and stored the individuals marked 'unknown origin'. Across the border in Zimbabwe, a different wild dog research organization, Painted Dog Conservation (PDC), uploaded their wild dog photos to ACW. One individual known to PDC that had been missing since 2013 was matched with the photos captured by tourists in 2018 in Botswana. These two independent sightings were 5 years and over 200km apart in different countries and owned and archived by different conservation organisations. Without ACW, this match likely would never have been made.
Figure: Examples of two long-distance, transboundary dispersal events documented by means of photographic evidence archived by independent research institutions and tourists.
While using ACW still comes with a few challenges, such as the difficulty in matching left- and right-hand side photos, and the need for multiple sightings to get a full pack composition, the benefits far outweigh the shortfalls. One main take-home message of this paper is to encourage more organisations to use ACW. The more photos that are uploaded, the broader the area that is covered, the more matches and potential long-distance movements that can be found! Carnivore species live long and have the potential to move far, and the only way to track this is through monitoring individuals. Only now are we reaching the capacity to do this effectively.
Comments