Nature is changing faster than our ability to notice it. Species disappear quietly, ecosystems shift unevenly, and much of the living world remains hidden in soil, water, soundscapes and night-time movement. We live in a world saturated with data, yet, biodiversity, the complex web of millions of interacting species, is still surprisingly hard to see, measure or fully understand.
Elaine van Ommen Kloeke, Program Manager of ARISE at Naturalis Biodiversity Center, is working to change that. ARISE has built a comprehensive DNA reference database of all multicellular Dutch species, and continues to expand it. This DNA- and AI-powered technology facilitates research and biodiversity monitoring at scale.
In this conversation, she explains why biodiversity has historically been so difficult to measure, how technology can help scale ecological knowledge, and why data alone will never solve the biodiversity crisis.
Could you briefly introduce yourself and your work?
“I’ve been working at the knowledge institute Naturalis for the past five years as a program manager of a large scale research infrastructure programme. Most people know us for our museum, which hosts one of the largest natural history collections worldwide, but behind the curtains a lot of research is taking place, with over 300 dedicated researchers.
We also build large-scale research infrastructure. In fields like physics or astronomy, you have facilities like CERN or the James Web telescope, where institutions work together to answer bigger questions than one group could do on their own. This has never happened for biodiversity until a few years ago. ARISE is one of those examples. It’s the largest biodiversity infrastructure programme in the Netherlands and quite unique in Europe. The aim is to recognize any multicellular species in any location using new technology. That includes everything from tiny algae to the mammals you see outside if you are lucky enough to spot them.”
What motivated you to do this work?
“That’s going back into my history. I have always been drawn to nature and I did a PhD in Ecology. But I never viewed myself as a typical researcher, I want to do everything around science: how you propel it, and how you build a bridge between scientific output and the real world. How can we actually make an impact? Throughout my career that has been my driving force.
So when this job opportunity came, with my experience in building digital platforms combined with my ecology background, it felt like the perfect way to combine the world of building products, infrastructure, hardcore data management systems and how that can be done for biologists who are usually focused on their own specific research questions.”
Many people associate technology mainly with negative impacts on nature. This is quite a broad question, in your experience, how can technology help research and thereby protect biodiversity?
“This is always a bit of a conundrum. On the one hand, we know what's going wrong with the world. Habitat loss, the way we use the land, climate change, pesticide use, and all kinds of other things have a negative impact on biodiversity and the livelihood of the natural world. So, in that sense, we don’t need more data or technology to be aware of these issues.
The challenge is that as long as we cannot make it tangible, people will find it hard to address such issues.
You can accurately measure Co2 or nitrogen levels, making it possible to predict weather patterns and climate change. But with biodiversity we are talking about thousands or millions of species in a given country. How do you know those species are there? How do you know they have been there forever? How many are there? Until now, that hasn't been measurable at scale.
In the Netherlands, we monitor between 5-8% of all species through all kinds of great initiatives. And that’s in a small, densely populated country with many active foundations and volunteers who monitor birds or beetles every year. Even with that network, we only cover a small fraction of species and a very limited area. That means if you really want to scale up and look at the tiniest bugs or flies or other hidden biodiversity in the soil, air, or water, then you need something beyond human observation. That's where technology comes in.
On the basis of environmental DNA (eDNA) we can recognise species, since each species has a unique DNA. Even in this room, there could be traces of fungi, insects, and maybe cat hair on my clothes. So you can detect if a species is present by remnant DNA.
To be able to do that you first need reference data. You have to find them, catch them, extract DNA, create a barcode, maybe record a sound or take an image. This is exactly what the ARISE platform now provides, a comprehensive reference database so that we can use technology like eDNA and sensors to quickly recognize species.”
Would that be possible without technology like the one used in ARISE?
“No, not really. The eDNA part is uncovering something we cannot see or touch. I’m working with the Westerdijk Institute, the largest fungal institute in the world, and even there they are only able to study a small percentage of existing fungi, probably less than 10%.
What we can do is to detect small ‘blips’ of DNA. Over time, those signals can reveal patterns and even predict changes under specific conditions. That’s crucial for understanding how biodiversity is doing. Of course you can be sad about large mammals disappearing due to habitat loss. But if you want early signals of whether an ecosystem is improving or degrading, the smaller species are often the best predictors. So those are exactly the ones you want to monitor.
If we venture to the other side of ARISE, which is more about technology like camera traps, radars, and sound recording. Such sensors usually capture species we can see as humans, but even there, technology reveals much more than human observation. For instance, we tested an entire camera network in the dunes and the ecologist involved said: now that it's running 24 hours a day, rain, no rain, coffee, no coffee, we see so many more species than if we were walking around. Let me give you some cool examples.
For instance, they hadn’t seen a barn owl in the area for over 20 years. They were unaware it was even there, but it was caught on camera several times. That tells you something about the health of the ecosystem.
Another important one was the reintroduction of bunnies. Bunnies are highly important for dune areas because they interact with the grasses that stabilise dunes. They have been dying from all kinds of viruses. So as they reintroduced the rabbits they put up enclosures to protect them from predators. But the bunnies kept disappearing. The cameras revealed two things: a predator dug under the fence within a night, and the rabbits were escaping themselves by climbing over the fence. The field managers were amazed by what the camera revealed.
On the one hand we can see and record the animals, but at the same time, the technology allows you to see so much more than you would be able to as a human.”
Did you personally have a moment in which you were really convinced that biodiversity and technology can reinforce each other?
“I always embraced technology because it enables scaling. What drove me further is that we have experts in taxonomy that dedicate their lives to exploring a given species, naming it, and uncovering new ones. But the sad truth is that most of these taxonomic experts are in their seventies. That knowledge is disappearing.
At the same time, this community also showed the most resistance to technology because taxonomy requires expertise, and investments should be made in this type of research and knowledge. And indeed, you can never replace taxonomy, but we also need to acknowledge that such expertise cannot scale to the needs we have on a national or global scale. Now, slowly there's a growing recognition that we need to have technology to support that expertise, to get information quicker, and maybe even reignite the interest in the field.
We need the experts to help us and to tell the technology what it needs to know. For example, a computer can never recognise a cat unless it's been trained properly and can distinguish between a robot cat, actual cat, or a dog that looks like a cat. With insects, it can come down to the direction of a single hair on a rear leg for instance. One way, it’s one species. Another way, it’s something entirely different. That’s knowledge only an expert can provide. So it’s a collaboration. Technology will always be a tool, an assistant, helping experts do their job.”
You already touched upon ARISE and what it does, referring to it specifically as infrastructure. Could you explain what kind of problem it solves? What is it doing?
“That’s why I use the word infrastructure. Even though people usually think of roads and bridges when using that word. Yes, it's a project in the sense that we have a specific timeline and budget. But what we are building is the infrastructure that will remain and develop over time.
I call it infrastructure because it's an ecosystem of services that can be used by researchers, or even commercial parties in the future. It focuses on species recognition. Can we quickly recognize a species in the field, in a sample, in a given experimental plot?
We are already providing this for dozens of researchers. We had a project for example, where they were building green walls, covering part of a building with vegetation to improve biodiversity. Well, how can you tell biodiversity has actually improved? You could have a human go and count the insects or you use sensors and eDNA that give you more detailed information.
We help make biodiversity measurable for anyone who wants to improve it. It’s not a single product. It’s a combination of machines, tools, data management systems, services, front-end interfaces, all working together.”
You’re using an ecosystem to study an ecosystem.
“Yes, basically.”
When you look ahead, say five years, what do you think ARISE could change?
“I refer to ARISE on the one hand as a project and on the other as a thing. Because it will continue. We have already connected to other projects that build on top of it and connect with other infrastructures. So that’s super cool to see.
So hopefully it changes and continues to adapt and develop. That's how products and services work, if you don't adapt, you'll become obsolete.”
Biodiversity itself can change tremendously in five years. Is there any particular kind of change you are anticipating or preparing the project for?
“Absolutely. We take into account not only native species, but also invasive and pest species. That’s important, for example, for farmers who need to know if certain pest species are emerging. Invasive species often indicate larger flows of species, sometimes due to climate change, sometimes due to economic trade. Detecting those early is crucial. With these systems, we can detect species even before they are widely observed.”
With these tools you generate a lot of data. How do you make it understandable without oversimplifying it for the sake of making it understandable?
“I like this question. As a knowledge institute you have to set boundaries. What we distribute is the data itself. The data we generate we call occurrence data: was a specific species observed at a specific time and place?
This data goes to the global biodiversity system, GBIF. It’s open data, and can be picked up by anyone. It's hard for us to interpret something without knowing the full context, so we leave that kind of interpretation to the experts in the field. That’s where we need humans, ecologists and field managers specifically.
We want to make it measurable and accessible, so experts can have reliable data at their fingertips.”
Is there a particular area where you expect faster or more accurate biodiversity data to change real-world decisions? What about related to policy?
“Yes and no. Policy can be stringent. It could be that policy decided a number of years ago to monitor certain species, with a specific methodology. And it can be very hard to change policy.
But you can only change policy by starting to do things differently. You need to measure and show that, for example, looking at a broader system adds value.
What we are seeing now, after 5-6 years working with eDNA, is that slowly it’s being picked up by policymakers. Where you see more willingness to invest is actually industry, industry is quite keen to improve biodiversity and make a difference. Many companies and local governments that manage land are very willing to improve biodiversity and are actively looking for tools to measure it.
That's great, but we have to be realistic, we can't do everything. Interpretation is key, but a lot of things still need to happen. It might sound strange for me to say, but biodiversity in itself should also not be the aim. Biodiversity is not just a number, it's an interpretation of what should be present in a specific ecosystem.
Sometimes fewer species does not mean less important. People can become enthusiastic about the numbers, but biodiversity is about context, uniqueness, ecological function, and the interaction between species.”
Could this research influence policy more broadly speaking? Has that already happened or do you expect it to happen?
“I think so. You already see carbon credit systems emerging. For biodiversity, similar reporting structures are being discussed, for example around land degradation. But it always comes back to measurement, if you can’t measure biodiversity, you can’t report on it in a meaningful way.
Satellite data is widely used and it’s a good predictor of vegetation type, especially for large industries like palm oil or agricultural companies. But you can't detect the tiny species. That’s the value of combining technologies. Satellite images give you the large-scale overview. eDNA and local sensors give you insight into what’s happening at ground level.
I believe that most people want to do a good job, without biodiversity our ecosystem and planetary health is declining. You can’t expect soil to produce and produce and produce and expect it not to degrade. You need to reinvest in that soil. And the only way to do that is to show people what they are doing so they can make a decision based on that.”
I’m curious how this measure of carbon and biodiversity can be implemented in future politics?
“I don’t want to oversimplify, but sometimes if we don’t do that there is no action. People get lost. Imperfect measurement or analysis can still lead the way to something positive.
You need different interests and perspectives to come together, science, policy, industry, landowners. It’s an ecosystem of opinions and needs, and data is one of the building blocks that helps move everything in a direction.
A colleague of mine once said: we biologists won’t solve the biodiversity crises, we can measure it all we want, but the biodiversity crisis will have to be solved by the whole society. So we need to make sure that they have the tools, the information and enthusiasm to get going.”
What are the main challenges, ethical considerations or risks in using AI for biodiversity? How do you design safeguards?
“We already work with AI with 10 different models for image and sound recognition. Legally, privacy issues mainly concern humans. So, we have safeguards to identify and filter out humans. In that sense, there aren't major ethical issues with the AI itself. We have experts to train and improve the model. So even if they’re not perfect, there are hundreds of volunteers helping us improve the system.
Ethically speaking, we consider the sensitivity of species.
You can imagine, if we spot a wolf, some people might be enthusiastic but there might be some that want to take matters in their own hands as they don’t want a wolf nearby. The same goes for very rare species like orchids or particular birds. If a particular bird is spotted the whole bird community may come to take a picture, so we protect those occurrences and filter them differently. We have the data but with sensitive species we are careful when reporting to the outside world. That is more the ethical side of things.”
Talking about the outside world, what would you advise people who want to make an impact on biodiversity but do not work specifically in the field?
“First and foremost, get involved. When it comes to biodiversity, all initiatives can help. You can become more aware of your surroundings and how you can help a particular species. For instance, planting flowers for bees and pollinators, or not using pesticides in your garden. That’s already a real impact on a local scale. Nothing is too small. Everything can help.
Second look at what’s available. On the research side, we produce a lot of tools and data, but we are terrible at informing the outside world of its existence. If you look at GBIF, it’s a huge repository, but it needs to be picked up by enthusiasts that can turn it into a useful tool.
Biodiversity will never be of such economic value that we can earn money from it. And we shouldn’t. It’s the core of our existence; without it we can’t live. We need to embrace that and give it a different value, not just economics. We should see it as the value of what makes us alive. And that’s the most important message; don’t treat it as something you can make money from but use it as something that defines who we are and what our existence means.”
Acknowledgements
We would like to sincerely thank Elaine van Ommen Kloeke for sharing her time, insights and vision. Her work at Naturalis and ARISE demonstrates how technology can help make the invisible visible, and help society better understand and protect biodiversity.