A single dashboard for all biodiversity data

Biodiversity has become a policy priority, with scientists and farmers embracing ‘nature-inclusive agriculture’ as a paradigm for efforts to move towards food production systems that also provide space for a diversity of plants, animals, insects and ecosystems. The idea is that this should lead to better nature conservation, more plant pollination by bees, the use of natural enemies to control pests, and more resilient food systems in an era of climate change.

The big question is: are these new forms of agriculture actually delivering the desired effects for our natural environments? To answer that question, Dr Arjen de Groot and his team Wageningen Environmental Research are developing a new set of monitoring methods. They’re investigating how DNA analysis, automatic image and sound recognition, drone photography and other remote sensing technologies can complement each other to measure biodiversity quickly, reliably and on a large scale.

A range of new technology

The scientists are using the Ketelbroek food forest, near Groesbeek, as their research site. A food forest is a landscaped forest planted with things like fruit trees, nut trees, sweet chestnuts and edible plants. It’s a ‘nature-positive’ form of agriculture, with plenty of room for biodiversity. “Ketelbroek consists of at least seven layers of vegetation, from the tree canopy down to the fungi in the soil. Between those layers, you’ll find all sorts of insects, many different plants, and birds and mammals,” explains wildlife ecologist De Groot. “It’s a complex system, and that’s one of the reasons we’re using this site to test the new methods.” The Dutch government also wants to encourage this type of agriculture.

A number of different teams are surveying the forest. They’re installing equipment that can filter the air to capture fragments of DNA from all of the forest’s inhabitants. Drones are being used to take pictures of a few hectares of forest, at such high resolutions that it’s possible to make out the leaves of individual bushes. “This will enable us to create a 3D model of the forest.” Audio loggers – small boxes that hang from trees – record the sounds of the forest: birdsong, and the squeaking of bats. “Audio recognition software can then produce a list of species based on those recordings.”

A handful of experts

“Most existing approaches to monitoring biodiversity rely on humans,” explains De Groot. “Counting bird species or bats requires experts to spend several weeks out in the field. They identify species by sight and sound, in line with official protocols.” Samples are also sent to the handful of experts who can analyse them. Insects, for example, are caught in every layer of vegetation and are delivered to specialists, some of whom have already retired. Sometimes, this even involves dissection. There are specific methods that come into play for each group, from birds to plants, and all those methods take up a lot of time and effort.

“There’s a lot of research going on into new methods of agriculture, and it’s essential to monitor the effects of those methods,” says De Groot. “If we want to satisfy this demand for data, existing approaches to monitoring can only take us so far. We’re already aware of projects that, for this reason, are opting not to bother with measuring at all, or are doing so only very selectively.” If the food forest is to become a viable alternative option in agriculture, it’s essential that we understand how it works. As De Groot puts it, we need to know “which levers we can pull”.

Analyses by supercomputers

The research into the new monitoring methods has been underway for about six months now. The researchers are still busy collecting and analysing data, but De Groot can already shed some light on one of the key technologies: DNA metabarcoding. Scientists all over the world contribute to databases for things like insect DNA, for example. The researchers plan to use these for large-scale sample assessments.

- Unfortunately, your cookie settings do not allow videos to be displayed. - check your settings

How can you measure biodiversity in an innovative, solid manner in a complex system like a food forest? This is the question animal ecologist Arjen de Groot and his team are trying to answer. In the Ketelbroek food forest, he is investigating how DNA analysis, automated image and sound recognition, drone photography and remote sensing technologies can complement one another to accurately measure biodiversity at a large scale.

Like primates and humans, insects or plants share some DNA, but there are also differences between them. These bits of DNA that differ between species are referred to by researchers as a barcode. “By determining the exact code of a piece of barcode DNA, we can recognise which species are included in a sample – even if it’s a mixture of all kinds of DNA fragments.” In the past, DNA reading was often performed by a specialised company, but that’s no longer necessary. De Groot and his colleagues have a device at the lab the size of a USB stick. “This device enables us to quickly read the DNA codes ourselves.”

But comparing all those codes to the reference collections is a bigger challenge. “One sample can yield up to two million codes, and we sometimes have hundreds of samples. You can imagine how much data that is! We’re using WUR’s supercomputers to analyse it all, which involves having a close working relationship with colleagues at the Wageningen Data Competence Center. Their work is very relevant to our research.”

Making insect broth

“When making our assessments, we prioritise methods that aren’t too disruptive,” says De Groot. “Filtering DNA from the air, for example, means we don’t have to catch animals or scour the forest. The same applies to image recognition. We’re working on an insect trap that automatically releases a creature once the camera system has recognised it.” Unfortunately, not all data can be collected this way. However, when the researchers do have to capture samples, they like to leave them intact if at all possible.

“In the past, preparing insects for DNA analysis involved grinding them into a homogeneous liquid, like soup in a blender,” explains De Groot. “This meant that we couldn’t subsequently study the insects again and connect those results with the species data from the DNA research.” The scientists are now looking into a new method, using a mildly abrasive liquid which extracts DNA from insect bodies but otherwise leaves them intact. “It’s a bit like making a broth.” Those insect bodies can then still be examined by a taxonomist, for example. But the science of discovering new species still relies on scrutiny by specialists, according to De Groot. “It’s the only way you can really extract all the value out of a sample.”

The plan is for all the data from those different methods and technologies to eventually come together on a dashboard. Graphics and other forms of data presentation will provide users with an overview of the biodiversity within a food forest. This overview is not just about ascertaining population numbers or the diversity of species, but also aims to connect different datasets. “For example, we want information on which parts of the forest are warmest or most humid, and to then combine that information with the data on the species found there.”

Combined datasets such as these are not all that common. “Our colleagues who work with drones will not typically have collaborated much with the DNA researchers, and they’re even less likely to have aligned their data with each other. This project brings their perspectives together. Specialist colleagues are asking all the teams: what does your dataset look like? They want to develop a single system to bring together and compare all these different types of data, and doing that is one of our key challenges.”

In two years, De Groot wants to be able to offer a new menu of monitoring technologies. The actual methods that an individual scientist or grower eventually ends up using will depend on what they’re trying to achieve. However, it’s clear that there’s a need for these technologies, because research into new forms of agriculture that can enhance biodiversity depends on having access to more data.