Sub-lethal pesticides might still kill ecosystems

Understanding how plants and pollinators interact isn’t easy when humans interfere. Spraying crops with lethal doses of pesticide is an excellent way to break any link between pollinators and plants. But what happens if the dose is sub-lethal? Research in the lab has shown sub-lethal pesticide doses can damage mangle relationships. A new study by Robert Gegear and colleagues has scaled up these sub-lethal stresses to the population level.

Rather than go into the field and poison insects, Gegear’s team created an agent-based model called SimBee. By changing parameters modelling a bee’s memory and information processing, they could examine different scenarios. They looked at how pesticide stresses affected bee abundance, plant diversity and the stability of plant-pollinator relationships. Effects of pesticides might not be visible immediately. Running the experiment as a simulation meant that the team could fast-forward into the future. They saw how pesticides affected the ecosystem over 20 seasons.

SimBee is an Agent-Based Model. Each bee runs a series of rules mimicking the interaction between bees and plants at a microscale. By running these calculations over and over, scientists can look for emerging patterns at the macroscale. In SimBee, each bee uses memory and information processing to decide how to forage for nectar. Once they gather nectar, they deposit it at the colony. The amount of nectar in the colony at the end of the season sets the number of bees in the simulation for the next season.

SimBee doesn’t just simulate bees, though. Plants are assigned pollen points. The model tracks visits, and each plant can produce a seed from a bee visit. Whether or not a plant produces a seed after a pollinator visit depends on what proportion of compatible pollen the pollinator is carrying. This is up to a maximum of six seeds in the model. At the end of the season, 40% of the seeds, randomly selected, germinate to become next year’s plants, unless the model has a constant number of plants.

Modelling both pollinators and plants allows the model to examine the complexity in the relationship between the two.

Looking at the effects of pesticides on bee abundance, the scientists set the number of plants to constant and let the bee population vary. For the second test, on plant diversity, bees were a constant. In this model, it was the seed set for the plants that changed. For system stability, both bees and plants were set to vary based on each other’s success.

The results of the model are worrying.

Even with just 25% of the pollinators suffering from impaired memory, bee populations declined. For the plant diversity experiment, things were slightly different. For the first six seasons, there was no apparent effect. However, after seven seasons, plants risked becoming locally extinct, even with a constant supply of bees.

Put these two results together, and something odd happens. Pollinator numbers increase. While the bees do well, the plants do not. “…as the proportion of impaired bees increased, so did heterospecific pollen transfer, resulting in reduced seed set and the eventual loss of plant species,” write Gegear and colleagues. “Under the 50% impaired condition, heterospecific pollen transfer increases alone were sufficient to drive the loss of at least one plant species in the 20 virtual seasons; most simulation runs showed a loss of two species.”

If you remove species from the system, you remove pollen interference, and the remaining plants have more chance of success. The more plants there are, even if they’re from fewer species, the more food there is for the virtual pollinators. It’s almost like that which does not kill you makes you stronger, from the pollinator’s perspective. But this is a simulation, and there’s no such thing as a free lunch.

The team say that increased numbers of impaired bees cause more damage to plant life. “Under 75% and 100% bee impairment conditions, all simulation runs ended with either a single plant species or complete system collapse,” they state. 

Comparing their results to reality, the team note: “Our model represents a closed ecosystem with a simplified virtual season; thus, exact predictions of expected rate of decline for a particular species in the real world were not possible. However, the model predicted that cognitively impaired foragers at these frequencies would cause a 50% reduction in bee abundance in 4–6 years, which is consistent with declines reported for the endangered rusty patched bumblebee (Bombus affinis) and other bumblebee species at risk in North America.”

Given the accuracy of their pollinator predictions, the plant results are a worry. “Previous modeling approaches to plant species extinctions in pollination networks were based on the assumption that pollinator loss is necessary to initiate plant extinction events. However, we found that changes to pollinator behavior alone could cause the loss of plant species even when pollinator populations remained stable… or even increased over time…”

Coextinction is unnecessary in many simulations for plants to disappear, but SimBee’s pollinators are generalists. If a plant critical to the survival of a specialist pollinator were to disappear, then coextinction becomes more likely. While less lethal pesticides might seem like a good idea, their effects on pollinators may mask a collapse elsewhere in the ecosystem.