The content on this page was translated automatically.

09/30/2022 | Podcast

Weather stations for biodiversity

Using mass methods, Prof. Dr. Birgit Gemeinholzer and her team are trying to record what species richness is present at a particular location. To do this, they are setting up stations that resemble weather stations but can do much more - including "face recognition" of moths.

Birgit GemeinholzerImage: Sonja Rode.
Prof. Dr. Birgit Gemeinholzer.

If you can't hear the audio automatically, deactivate the mute, adjust the volume or activate autoplay in your browser.


It's early in the morning, relatively early in the morning, and I've arranged to have coffee with Professor Gemeinholzer, who is our new professor of botany, having arrived with a whole host of incredibly interesting projects in her luggage, including one on recording species diversity. There are some very exciting things that you have already briefly outlined for me, Ms. Gemeinholzer, and about which I would now like to ask you. Good morning, dear listeners, good morning, Ms. Gemeinholzer.

Good morning, Mr. Mense. I'm glad you're here.

Ms. Gemeinholzer, I once took part in an insect count a few years ago. We walked across a meadow in a stooped position and looked to see what we could find in the way of beetles and ants and so on. Of course, that was relatively time-consuming. We then reported this to the nature conservation association and they summarized it. The idea was to get an idea or an overview of how many insects are now present in a particular area. You have, I think, a better idea.

Yes, we are working on: not only what insects are in an area, but also what plants do the insects visit at what time of year? So, what is the food like? Like the availability for insects and whether insects have actually become rare because the food plants have become rare. So we try to analyze not only the insects. We do that in cooperation with the Entomological Society Krefeld and with the Museum König in Bonn. But we also try to analyze which pollen is on the insects, which vegetation is really important for the insects in this season. Because the plants can be determined throughout the year. This is relatively well done in Germany. You know a lot about plant biodiversity, but you don't know much about ecosystems, their functions. That's what we're working on.

And you have developed a method to automate that. Can you explain to us what that looks like?

Exactly. Until now, this, to really routinely analyze this interaction in high throughput, was not possible. Because we have lacked the methods. The insects are very difficult. You need experts for each of the different groups. They are usually scattered somewhere in Germany and usually they don't have the time. We from botany want to determine the pollen. Even they don't have many characteristics to really assign to the individual species. But what we have discovered in the meantime: We can identify, based on the genetic code, based on the DNA, which species either the insects are, so taking bulk samples of insects, analyzing mixed samples of insects, and we can identify the plant traces on the insects. So we know at what time of year we have what insect-plant interactions. We have agreed internationally on certain regions on the DNA, because the DNA is incredibly long. We want to take only a very small piece of the DNA and on the basis of this small piece, which we can read, this is called sequencing ... with this sequencing we get a piece of the DNA and we test it against a reference database.

It's like if you search for Mr. Schmitt on Google, for example, and you type that in and Google automatically finds all the Schmitts that you want. And that's how we get it in the reference database, that we take a piece of the DNA and we blast those nucleotides against the reference database and then we try to find identical sequences. And then it tells us which potential candidates are just similar to my sequence that I found on my unknown insect, or on the pollen that was on the insect. So we can really identify food chains.

Okay, that's interesting. If I understand it correctly, this way of collecting the samples is something novel that you once compared to a weather station. Maybe you can explain that to us and our listeners how that works....

Yes, right now it's so difficult because the botanists go to a certain place at a certain time of year. We never have overlapping data. But what we need to be able to maybe monitor biodiversity shifts in the age of climate change, so to be able to detect those shifts as well, are biodiversity weather stations. That means that at the same time, at the same place, the same sensors automatically collect insects, plants and other animal groups, so that we can really say whether the entire biotic communities are still intact or whether they are shifting and how they are shifting. So we're building biodiversity weather stations, similar to climate stations, and we hope that maybe in the future we can even then model biodiversity changes.

That is, these weather stations, they automatically capture insects, for example.

Exactly. At the moment, there are only prototypes. They are now collecting at regular intervals and we are looking at which groups of organisms we collect at which intervals and how, and which individuals we have found with them.

And the stations I read, they also have ears and eyes and even noses.

So the botany and the Museum König and the entomologists association in Krefeld, we are responsible for the insects and for the plants in addition. But then there are groups that do work with photo traps for larger mammals and birds.

Bats, for example?

Exactly. Then there are sensors for image recognition experiences, moth scanners. What moths come up against a screen at night or can you detect those photographically?  Is there three-dimensional recognition? That's where we're trying to take advantage of algorithms from facial recognition from Google. And with that, we try to generate recognition features.

So facial recognition for moths, so to speak. That's not bad. What conclusions have you been able to draw from the data so far? Do you already have any results?

So far, we are working on prototype development and analyzing to what extent the different traps and sensors influence each other. For example, if the machine changes, what influence does that have on the sensors? Or what kind of influence do environmental noises have? If the meadow is mowed next to the sensor system, can we filter that out? At the moment, we are still very busy with all these side effects. That's why we only have these prototypes, which are located at two sites in Germany. And the third, which we are currently trying to set up.

Will it come to northern Hesse, are we lucky to have one here?

Unfortunately not at the moment, because I'm very new here in Kassel, so I didn't write here. At the moment the locations are in Bonn, near Berlin and in Hamburg.

Okay. You have another project, as I have seen, which does not deal with the present or the conclusions for the future, but with the past and draws conclusions from it. You have already mentioned the keyword pollen, that is one keyword, the other keyword bumblebee. What is behind this?

Exactly, in our automatic traps we try to analyze which plant traces are carried into the traps by the insects? That's a mixed sample, and we can only tell which plants are used at which times of the year. But now we are actually also interested in the individual. So: which bumblebee visits which plant at which times of the year? We're trying to do that in a historical and recent context. Recent is what we have here now. And historical is what has been in the past. We collect bumblebees in natural history museums.

You actually go to the archives or the exhibition rooms of the museums and look at what kind of bees - and bumblebees in that case - still have pollen on their legs?

Exactly: have it on their bodies or carry it on their legs. And we collect them, we can sample them. Then we look: What did they collect in the past and what do they collect today? We have now completed a project where we collected again at exactly the same location at the same time. We have found quite different preferences now. A new project is just trying to analyze whether simply climate change has shifted the flowering dates of the plants, so we are now collecting a period of time to analyze whether the food supply has really changed or how that relates. So we're trying to analyze whether this food that we're just providing to insects in nature reserves is just such a necessity food because they can't find anything better, or whether that's really what they would prefer and that's what they used to prefer.

The diet has changed, you can say that already?

I'm guessing yes.

Yes, thank you very much. We are curious about that. Thank you, Ms. Gemeinholzer, for the interview and for the delicious coffee.

Thank you for being here.


Interview: Sebastian Mense