A New Menu Of Monitoring Techniques

Researchers at Wageningen University & Research want to get the most complete picture possible of biodiversity in new agricultural systems. That is why they are bringing together all kinds of technologies, from automatic DNA analysis to heat maps made with drones. The goal is a single dashboard that connects all these measurements.

Biodiversity is high on the agenda. Under the heading of nature-inclusive agriculture, scientists and farmers strive to grow food in systems that also provide space for a diversity of plants, animals, insects and ecosystems. This should lead to better nature conservation, more pollination of plants by bees, pest control in the form of natural enemies and more resilient food systems in times of climate change.

The big question is: do these new forms of agriculture also have the effect on nature that we expect? To answer that question, animal ecologist Arjen de Groot and his team at Wageningen Environmental Research are developing a new set of monitoring methods. They are investigating how DNA analysis, automatic image and sound recognition, drone photography and other remote sensing techniques can complement each other to measure biodiversity quickly, reliably and on a large scale.

A range of new technology
The research object of the scientists is the Ketelbroek food forest, near Groesbeek. A food forest is a planted forest, in which, for example, fruit trees, nut trees, sweet chestnuts and edible plants are grown. Such a food forest is a 'nature-positive' form of agriculture, with plenty of room for biodiversity. 'Ketelbroek consists of at least seven vegetation layers, from the canopy of the trees to the fungi in the soil. In between, all kinds of insects, many different plants, and birds and mammals find a place,' De Groot explains. 'It is a complex system, and that is one of the reasons why we are testing the new methods here.' Moreover, the Dutch government wants to give this form of agriculture a boost.

Several teams are busy scanning the forest. Equipment is being installed that filters fragments of DNA from all the forest inhabitants from the air. Drones take photos of a few hectares with a resolution that is so high that you can distinguish the leaves of individual bushes. 'We use that to create a 3D model of the forest.' Audio loggers — small boxes that hang from trees — record the sounds of the forest: the singing of birds, the squeaking of bats. 'Based on those recordings, audio recognition software can then produce a species list.'

A handful of experts
'Most current forms of biodiversity monitoring rely on people,' De Groot explains. 'For example, to count bird species or bats, experts have to spend several weeks in the field. They record which species they observe by sight and hearing, according to official protocols.' Samples are also sent to the handful of experts who can interpret them. Insects, for example, are captured in all layers of vegetation and delivered to specialists who are sometimes already retired. Sometimes dissection is even involved. Each group, from birds to plants, has specific methods, and all methods take a lot of time and labor.

De Groot: 'A lot of research is being done into new agricultural methods and monitoring the effects of these is essential. If we want to meet this demand for data, we are running up against the limits of the current methods. We already see many projects that choose not to measure at all, or only very selectively.' If the food forest is to become a fully-fledged agricultural alternative, it is essential to understand how it works, or as De Groot puts it: 'which buttons we can turn.'

Analyses by supercomputers
The research into the new monitoring methods has been underway for about half a year now. The researchers are still busy collecting and analyzing data, but De Groot can already elaborate on one of the most important techniques: DNA metabarcoding. Scientists all over the world contribute to databases with, for example, insect DNA. The researchers want to use these to assess samples on a large scale.

It works like this. Like primates and humans, insects or plants have DNA in common, but there are also differences. The pieces of DNA on which species differ from each other are called barcodes by researchers. 'By determining the exact code of such a piece of barcode DNA, we can recognize which species are in a sample — even if it is a mixture of all kinds of DNA fragments.' DNA used to be read out by a specialized company, but that is no longer necessary. De Groot and his colleagues have a device in the lab the size of a USB stick. 'We use it to quickly read out the DNA codes ourselves.'

A greater challenge is comparing all those codes with the reference collections. 'One sample can yield two million codes, and we sometimes have hundreds of samples. Just imagine how much data that is! We use the supercomputers of Wageningen University & Research to analyze that. We also maintain good contact with our colleagues from the Wageningen Data Competence Center . Their work is very relevant to our research.'

Making broth from insects
'In our assessment, we prefer methods that are not disruptive,' says De Groot. ' Filtering DNA from the air , for example, prevents us from having to catch animals or comb the forest. The same goes for image recognition. We are working on an insect trap that automatically releases an animal when the camera system has recognized it.' Unfortunately, not all data can be collected in this way. If the researchers are forced to catch animals, they prefer to leave them intact as much as possible.

'In the past, insects were ground into a single homogeneous liquid, like soup in a blender, to prepare them for DNA analysis', De Groot explains. 'That meant we couldn't study the insects again later and connect those results to the species data from the DNA research.' The scientists are now looking at a new method, in which a liquid slightly affects the insect bodies to extract the DNA, but otherwise leaves them intact. 'As if you were making broth from them.' A taxonomist could then look at those bodies, for example. Science for discovering new species still depends on the eye of the specialist, says De Groot. 'That's the only way you can really get the most out of a sample.'

All data from these different methods and techniques must eventually come together on a dashboard. Graphs and other data presentations must then offer the user an overview of the biodiversity in a food forest. This is not just about the numbers or distribution of species. The aim is also to connect different datasets. 'For example, we want information about where in the forest it is warmest or most humid and combine that data with the data about the species that occur there.'

Such combined datasets are not self-evident. 'The colleagues who work with drones and the DNA researchers have often worked together very little, let alone that their data is coordinated. In this project, we bring their perspectives together. Specialised colleagues ask all teams: what does a dataset look like for you? They want to develop one system in which all those different data come together and are compared with each other, that is one of our most important challenges.'

In two years, De Groot wants to be able to offer a new menu of monitoring techniques. Which methods a scientist or grower ultimately uses depends on their goal. It is clear that these techniques are necessary. Certainly in the research into new forms of agriculture that should ensure more biodiversity, more data is needed.