Bees and agriculture – monitoring studies test the real life

Bees and agriculture – monitoring studies test the real life

Monitoring studies are what I stand for when it comes to improve agricultural landscape for bees. Last week, I already explained what they are: a test for reality and a possibility to answer “what-if…”-questions. Before the registration of pesticides, there are extensive tests for all products. However, there are often problems that come to life only after the registration. This shows the pitfalls of the current system: on the one hand, after all the studies and modelling before registration, we feel safe. A lot of intelligent people worked hard to evaluate the risks. They regulated how to apply pesticides, for what crops and when.

Unfortunately however, even the most intelligent person can’t consider all situations. Impossible. Neonicotinoids were considered safe for bees because they shouldn’t be exposed – the coated seeds disappeared in the soil, after all. The history proved this conviction wrong. Working with model species is another problem. Again, it is impossible to test all species. Though it would be a great improvement not using only honey bees, also including single solitary bee or bumblebee species wouldn’t give total security. And this is the point monitoring studies come in.

Different conditions produce different effects

The studies before registration are important, let me state this first of all. They help to exclude the worst risks and they already try to represent reality to a certain extent. But, the devil is often in the detail. What seems safe in one setting may be false in others. Let me give some examples.

I participated in some studies that looked for the risks of neonicotinoids for honey bees. For instance, we examined if with improved techniques during sowing, reduced the risks that dust from the coated seeds could deposit on flowers. For this, we cut down part of an oilseed rape field and sowed coated maize in this area. We also put some honey bee colonies directly in the field (not at its edge, as beekeepers would do) before sowing. By this, we wanted a “worst case scenario”, with colonies directly exposed to any dust coming up during the process. In front of some of these colonies, we put a grid covered with fine gauze, as well as at the edge between maize and oilseed rape to sample the dust.

Colonies directly between a maize and a oilseed rape field. The grid in front of the colonies is covered with gauze to catch the dust during the sowing process.

These colonies stayed in the field for a few weeks, we also overwintered them and observed how they did until the next spring. And, though you might not want to hear this, they did well. We didn’t look into details like orientation. We just observed the development of the colonies during almost a year and took dozens of samples of flowers, bees, larvae, wax, honey and pollen to examine the residues. There were also other studies however, that didn’t find effects on honey bees. A very good and thorough one was published a few years ago by Maj Rundlöf and her colleagues.

Monitoring studies to assess wild bees

In this study, they tested what effect seed coating in a flowering crop has on honey bees, bumblebees and solitary bees. While the honey bees did fine, bumblebees and solitary bees did not. They were less abundant and solitary bee nesting and bumblebee colony growth was affected. This shows that the “model organism” honey bee may work well under some conditions, but doesn’t always give all necessary information.

Wild pollinators usually aren’t part of the testing in the registration process. This is one of the points the Leopoldina criticizes in its latest discussion paper. But sometimes, it happens. In one of these studies, we monitored the bees in apple orchards before and after application of a sprayed insecticide (not a neonic) over two years. We assessed ten orchards, describing the landscape they were in, neighbouring crops, the whole pesticide regime the growers used etc. At the end, there were some orchards with many solitary bees and others with less. At some orchards there were plenty of nests, at others it was really hard to find them.

All of these farmers used insecticides and also other pesticides. What was different, was the structure of the landscapes. More wildflowers in and around the orchards meant more bees. Patchy vegetation or earth field paths were perfect nesting grounds for all the mining bees that pollinated the apple flowers. The more of these structures were present, the more bee species and individuals we found in the orchards.

Field paths with patchy vegetation, a hedge and wild flowers – a good bee habitat with stable populations.

Combining standards and complexity

Writing this post, I’m more than aware of all the possible criticism you could have on the studies. Unfortunately, there is nothing like a perfect study. We never cover every question with a single study. What monitoring studies do though, is giving an idea of how different situations affect the consequences of a certain practice. Regulators don’t like them because they’re not “standardized”, they’re always different. But this is exactly their strength: they approach the “what if”-questions. It’s all the small, not standardized details that may change the effect a certain pesticide has. It’s the explanation why some scientist did find risks for bees and others didn’t.

Standard studies and monitoring studies don’t exclude each other. They are complementary. No study will ever represent the “complete truth” or exclude every possible risk. We have to acknowledge this. A colleague once said that statistics has to follow reality, not vice versa. Same is true in this context. Of course it would be great to have the perfect model to predict the safety of pesticides only by computer models. Without the risk of a rainy season, organisation issues or a farmer not reading the label. But this would produce only a false sense of security and control. Instead, the unforeseen factors and risks that arise during a monitoring study can help to improve the models and increase the standards of these totally non-standard studies.


Leave a Reply

Your email address will not be published. Required fields are marked *

Sign up with your email address to receive news and updates!

  • linkedin