The species has declined by 80-85 per cent in the past three decades – the new technology could help the elephants and the wider climate crisis, say researchers.

By

Published: Friday, 16 August 2024 at 10:22 AM


Imagine an elephant living in a rainforest of the Congo Basin in West Africa. It consumes vegetation in the forest understorey, creating space for larger, woody trees to grow, and this increases the quantity of carbon that is stored within that habitat.

In today’s world, that carbon has a value – indeed, scientists have calculated that the carbon sequestration and other ecosystem services, such as seed dispersal, of a single elephant is worth $1.75 million and that it removes carbon dioxide equivalent to the emissions from more than 2,000 cars.

But while having a positive value in global terms, elephants have a negative impact closer to home. They come out of their forest home and raid crops grown by local communities, resulting in conflict that leads to income losses, injuries and even fatalities.

African forest elephants gather at sunrise at Dzanga Bai, Dzanga-Sangha Special Reserve, Central African Republic. Credit: Andy Isaacson, WWF US

The global conservation group WWF wants to come up with a way of making the presence of forest elephants – now recognised as a separate to African savannah elephants, and a species that has declined by 80-85 per cent in the past three decades – a benefit to these people. To do that, it says, it needs to count the number of elephants in an area with much greater accuracy.

It is working with the tech giant IBM to train artificial intelligence (AI) to identify individual elephants from camera-trap images. According to Dr Thomas Breuer, WWF Germany’s senior project manager for Central and West Africa, they will do this by taking photographs of already-known elephants and feed those into a computer programme.

“The AI then creates an algorithm [to identify elephants] – we don’t know what it is using to do this, whether it’s the eyes or what,” says Breuer. It is believed that elephants have features, such as their tusks and wrinkles on their trunks or around their eyes, that make patterns like fingerprints, and it is these patterns the AI will exploit. IBM says this could take about a year to develop. 

African forest elephant
An African forest elephant in Loango National Park, Gabon – the elephants here are known to cross rivers and lagunes (large water bodies) and forage on beaches. Credit: Christiaan van de Hoeven, WWF Netherlands

Once individuals can be identified, conservationists will be able to make much more accurate estimates of population numbers within particular areas. “If community A has 100 elephants and community B has 10, the community A should be rewarded more than B,” says Breuer.

In other parts of Africa, some conservation projects have devised a scoring method whereby villages accumulate points for having rare or dangerous animals (such as large carnivores) on their land. This method could also be used to reward them for the presence of elephants, Breuer believes. 

African forest elephant mother and calf in Dzanga Bai
A forest elephant mother and calf in Dzanga Bai, a forest clearing in Dzanga Sangha Protected Area, southwestern Central African Republic. Credit: Carlos Drews, WWF

But many conservationists believe the long-term solution is for companies looking to offset the impact they have on their planet to invest in the carbon sequestrated by the elephants through their day-to-day activity and behaviour.

Breuer accepts there are many challenges to this concept, but – one day – it could be the solution to living alongside species such as forest elephants and maybe many other large mammals. 

African forest elephants
Forest elephants in Odzala-Kokoua National Park, Republic of Congo. Credit: Jaap van de Waarde, WWF Netherlands

Main image: Forest elephants at sunrise at Dzanga Bai, Dzanga-Sangha Special Reserve, Central African Republic. Credit: Andy Isaacson, WWF US

More amazing wildlife stories