The goal of the Novel Biodiversity Metric: Small Mammal Track Analysis using AI, Morphometrics and Traditional and Local Ecological Knowledge (TaLEK) project is to develop a new technique for a cost-effective and community-generated metric to monitor mammal biodiversity in Africa.  The benefits for project partners include increased engagement and capacity for local students and indigenous groups, as well as new and transformative actions for African 30×30 goals planning. Assessing impacts on biodiversity is key and as yet there are currently no reliable and cost-effective metrics to do this. Activities will begin by developing a reference database of small mammal species across different environments. Team members will live-trap small mammals and identify species before letting them run over a track plate to collect their footprints before release. An AI pipeline will be developed to identify these tracks by species to produce a rapid biodiversity health index. With a classification model in hand, team members will validate the technique in field environments of high biodiversity interest (seasonal rivers, desert, and denuded areas) where comparisons can be made with available data about biodiversity for comparison (bioblitz studies, existing large mammal biodiversity etc.)

Project partners in Namibia will be the main implementers, but the project aims to build regional linkages through collaboration with counterparts in neighboring Botswana. Open data access will encourage sharing and communication across borders. Government partners in both countries will be encouraged to collaborate on biodiversity monitoring using this common technique and shared data.

WildTrack is collaborating with the Tswalu Foundation and Oppenheimer Generations Research and Conservation to administer this critical project.

Goals for the Novel Biodiversity Metric: Small Mammal Track Analysis using AI, Morphometrics and Traditional and Local Ecological Knowledge (TaLEK) Project include:

To inventory the small mammal biodiversity at sites in Namibia and Botswana that are both disturbed and undisturbed.
The goal is be to build a species list of small mammals using TEK and publications by academic experts in the field. An indicator of success will be that the species list conforms to the existing lists, but may include new species, or new range data.

To build a library of footprints from these species.
The goal is to test several of the current track-plate techniques to identify those that are most effective at capturing small mammal prints. This depends both on encouraging animals to walk over the plates, but also capturing clear images of those prints. Success will be driven by both.

To train for local use an AI and TEK-based algorithm to identify the species.
The goal will be to import the plates into our AI pipeline (currently working with footprints of larger animals on natural substrate), train object recognition to identify the footprints, box and label each print and then develop the classification algorithms. Success will be when the algorithm classification matches the one given by the experts in TEK and the academic publications.

To use this metric to measure biodiversity health at sites identified as high priority as proof of concept.
Success will be measured by comparing the results of species ID during live-trapping trials where animals are identified and released, with those where the track plates are analyzed blind by AI pipeline collaborators.

Drs. Zoe Jewell and Sky Alibhai co-founded WildTrack (wildtrack.org) in 2011 to address a widespread need for less invasive and more cost-effective tools to monitor endangered species. WildTrack’s multi-award-winning footprint identification technology (FIT), based on traditional tracking skills, has now been developed for species ranging from black rhino to Polar bear and mountain lion. This approach works in close collaboration with conservation efforts to engage local communities and particularly those with expert tracking skills. WildTrack is now developing an AI pipeline to accelerate the delivery of key data to conservation decision-makers. Jewell and Alibhai are Adjunct Faculty at Duke University’s Nicholas School of the Environment and Associate Academics at the Centre for Compassionate Conservation of the University of Technology in Sydney, Australia.