How to capture biodiversity loss

I am currently positioned at the edge of Tau Waterhole in South Africa’s Madikwe Game Reserve. On the other side of the muddy pond, a lone bull elephant is flicking his tail. He stands gazing at the arid landscape beyond.

Sadly, I haven’t been posted to South Africa to research this article but am sitting in my London kitchen watching a live stream of the pachyderm on Explore.org. The website broadcasts day and night from various wildlife hotspots worldwide, allowing anyone to drop in and watch nature.

The stream at the Tau Waterhole is just one example of how technology is helping us get a better picture of life on Earth. And the need for this monitoring is greater today than ever. Scientists and campaigners have been alerting the world to the destruction of species for decades, yet their warnings are becoming ever starker. Concerns about a ‘sixth mass extinction’, caused by human activity, have now entered mainstream consciousness.

Biodiversity refers to the variety of life in a habitat – from the tiniest single-cell organism to fungi, vegetation, insects, humans and, indeed, the elephant at the waterhole. A high level of biodiversity is generally seen as desirable.

Biodiversity is under threat. A 2019 report from the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services alerted the world to the dramatic decline in abundance of species numbers. Average populations of native species have dropped by at least 20 per cent since 1900 and certain kinds of species, such as amphibians, have declined by a shocking 40 per cent.

The scientific consensus is clear that these population declines are almost exclusively caused by human activity and the arrival of invasive species (which are often linked to humans too). Not only is this decline a tragedy in itself, but it is also likely to affect humans negatively. Life on Earth sustains humanity, and we get many so-called ‘ecosystem services’ from nature. From the insect pollination, which practically all agriculture relies on, to cleaning the water we drink or the oxygen we breathe, we rely on biodiversity for our survival. Indeed, sustainably supporting biodiversity is one of the top ‘Technology Critical Targets’ – a list of seven challenges that engineers have set their sights on to give life on Earth the best shot at thriving.  

If we are to preserve life on Earth and reverse this depressing trend, we first need to understand how much biodiversity is out there. By analogy, our science around climate change relies on the collection of data about the presence of greenhouse gases in the atmosphere. Without this information, we could not build models to predict the impacts of temperature rises.

The trouble is, we are a long way from having such a clear view of global biodiversity levels, so we struggle to say accurately how much there is, what rate it is declining at, or how different interventions could reverse the trend. Researchers, campaigners, governments and international organisations are working at different levels to try and tackle this concern.  

“One morning I went out to find Pavarotti. I was driving in the bush but drove into an aardvark burrow and my jeep got stuck.” Kasim Rafiq is a researcher from Liverpool John Moores University who studies large predators. As part of his PhD research, he was in Botswana’s Okavango Delta, where he was searching for a large one-eared leopard the scientists had dubbed Pavarotti.

Rafiq had been tracking the leopard when he had his aardvark misadventure. While he was figuring out what to do next, another jeep full of tourists on safari drove by and offered to help him out of the ditch. As he was talking to the tourists, one of them mentioned they had seen a large one-eared leopard earlier that morning. Rafiq asked to see their photos. Sure enough, it was Pavarotti – but two kilometres away from where he’d been looking.

This chance encounter led to an idea. What if tourists on safari uploaded pictures to a central database with associated GPS metadata? By using a form of citizen science, researchers would be able to collect data from hundreds of sources and would not need to spend so long travelling around solo looking for animals.

During a pilot study, Rafiq worked with several tourist groups visiting Botswana’s magnificent nature reserves. He then spent time working through the pictures to build up a detailed view of how many creatures were out there and where they were roaming.

Rafiq says his pilot study and the ‘citizen science’ delivers data on wildlife location and abundance comparable to more traditional research methods. It also requires a lot less legwork, saves money and could produce data over longer time frames, since researchers can’t stay in the field forever.

This is one example of the many technologies being used to collect information about biodiversity. Until relatively recently, the only way scientists could get an idea of the number of plants, animals, insects, or anything else in the field was to spend months observing them and counting what they saw. This meant data was almost always limited. Yet by using new technology it is getting cheaper and faster to collect a lot more information.

Talia Speaker is a programme officer at WildLabs, a partnership of conservation groups including WWF launched in 2015. She explains that WildLabs is a platform where researchers and technology experts can meet one another, share best practice, and get advice or support.

Speaker says there is a plethora of technologies used in the field today, but the three most common are camera traps, acoustic monitors and bio loggers.

Camera traps have been around for decades but have become significantly cheaper and better over time. Usually attached to a tree, a motion sensor on the camera trap notices when an animal walks by and takes a photo. Newer models have thermal lenses to capture animals at night and are a “good way to get data on rare and elusive species”, says Speaker.

Another common tool is bio acoustics. These recorders can be placed on trees, in the ground or even underwater to record noise coming from all directions. Researchers then listen in and figure out what is living nearby based on what they hear. Modern bioacoustics tools can be programmed to only capture recordings of species that the researcher is interested in. For instance, they could ‘ignore’ birdsong if the researcher was only interested in monkey calls.

Then there are bio loggers, a kind of tracking device. The trackers are attached to creatures, and transmit data showing where the animal has been so we can learn more about its behaviour. Again, bio loggers have been around for years, but Speaker says they are getting smaller, cheaper and provide more information than just location. Like Apple Watches, they can even record things like the animal’s heart rate. And the tiniest chips currently available can fit on the backs of individual bees!

On their own, these technologies are most useful for researchers who are, for instance, following a particular pack of apes or investigating the mating calls of certain birds. However, Speaker says there are efforts to make more use of this data through aggregation.

She points to AI for Earth, which is backed by Microsoft. It uses image recognition that can classify millions of photos from countless camera traps around the planet. Able to scan through far more pictures than a human ever could, the hope is this sort of AI will tell us more about animal populations than any individual camera trap could.  

The drawback of sensors is that they only tell us about plants and animals that happen to be near a camera trap, acoustic recorder, or tourist jeep at a particular time. To get the bigger picture, we need to use multiple sources of information.

Since the 1970s, Nasa has been collecting detailed images of land use across the planet. Today anyone can visit the Earth Observatory website to view maps showing how vegetation cover has changed in Brazil, say, or the extent of coral bleaching off the coast of Australia. Satellite images are therefore one important tool in letting us see what impact humans are having on the planet. But of course, satellite imagery can only give us a very general view of change.

A more comprehensive approach might be found in the analysis of DNA. Darren Evans, a professor of ecology at Newcastle University, describes how in recent decades there has been a global effort to classify the DNA of all life on Earth to create a ‘barcode’ of life. Thanks to this effort, scientists can now collect a sample of DNA, analyse it in the lab, and then find out what species it belongs to through reference to the International Barcode for Life.

To illustrate how helpful this is, Evans explains that researchers can now collect a water sample from a pond and use the Barcode for Life to get a list of every species whose DNA is present in the sample. Living things are continually shedding cells, so these traces can tell us pretty accurately which creatures are currently in the pond or have passed through.

In his own research, Evans is using DNA to learn how ecosystems work more precisely. Just by collecting pollen from a single flower, he not only gets the flower’s DNA, but also the DNA of any bee or insect that has landed on it. This information can then give us a clearer idea of which species rely on which other species.

The development of DNA barcode systems is extremely powerful and can tell us which species inhabit an area much faster than any method previously developed. But it has limits. Evans explains that it doesn’t tell us how many individual organisms are present and may also reveal creatures that no longer live there (a water sample could contain traces of long extinct organisms).

When researchers collect biodiversity information, many upload it to global databases like GBIF (Global Biodiversity Information Facility), that hold vast amounts of data from every corner of the globe. This is good, but making sense of it all is incredibly difficult.

One organisation trying to help here is Montreal-based GEO BON, which works to make sense of all this data so it can be used in meaningful ways. “The challenge is less the price of technology, and more human resources and how to analyse the data available,” explains Adriana Radulovici, executive secretary at GEO BON.

For example, as part of their UN commitments, all national governments must produce annual reports on biodiversity within their borders. This is reasonably straightforward for wealthier nations, but others lack access to relevant data, skills, or knowhow. GEO BON can support them with finding data they need, using appropriate models and showing how to analyse this data.

When it comes to tackling biodiversity loss, it would be very useful to have models showing change over time, just as we have climate-change models based on carbon emissions data. Such a model would reveal how many plants and animals exist at present, and how various human interventions might either exacerbate or reverse nature’s decline.

At present, no such model exists, but there are efforts to achieve this goal. One of the most advanced is the Madingley Model, based in Cambridge. Mike Harfoot, who helped build the model, explains that it uses mathematical representations of processes that all organisms go though. The model is ‘populated’ with all kinds of plants and animals found in all ecosystems, rising from the smallest multi-cell organisms up to the biggest trees and herbivores. All these living creatures are constrained by rules governing life – energy usage, feeding, reproduction, death and movement. It also contains climatic data and various Earth system processes such as ocean currents.

Using the model, scientists can explore hypotheses. What, for instance, would happen if all the large carnivores were removed from an ecosystem? What would happen to herbivores if grasslands were given over to agriculture? How would an ecosystem survive if a significant proportion of insects disappeared?

Harfoot says the model is still fairly crude but is already being used by scientists around the world. For example, the model recently looked at biodiversity levels in forests as rising amounts of vegetation are removed (to replicate deforestation). It showed that biodiversity in tropical forests remains resilient for a significant amount of time, but past a certain threshold, it collapses very fast. In other locations, the decline of biodiversity with removal of vegetation follows a steadier downward gradient, while in some places even a small amount of vegetation removal means the whole ecosystem collapses almost instantly.  

Looking forward, technology is expected to become even more accurate. The European Commission, for instance, is currently in the process of building Destination Earth, a high-​resolution digital twin that aims to replicate our planet in digital form.

This immense project is in the early stages, and in its first iteration will mainly focus on climate change and extreme weather. It will be open access, so any scientist could use it for their own research. For example, someone studying plankton could combine their data with Destination Earth, to see how warming ocean waters will affect plankton distribution.

As researchers and international organisations collect and make sense of more data, it should become easier to get a true view of the current state of Earth’s biodiversity. This could then be used to decide which interventions will most effectively reverse its decline.

Back at the Tau Waterhole, the sun has set and the elephant that initially caught my eye has moved on. The camera shows a scene devoid of life, but for a few insects.

Scientists’ efforts to capture more information about life on Earth will hopefully mean magnificent creatures remain a part of the landscape in all their rich variety.