Environmental impact of a food product can be evaluated, from its production to its disposal, through what is known as life cycle analysis. This includes climate effects, utilisation of natural resources and impact on biodiversity, but the method is associated with major uncertainties when assessing potential damage from environmental pollutants.

“There are about 3000-6000 chemicals that we have good information on today. At the same time, global daily operations have a turnover of around 80 000-100 000 substances. Many are not toxic in the volumes they are used, but it is problematic that we do not know more about them," says Oskar Nyberg, PhD in ecotoxicology at the Department of Ecology, Environment and Plant Sciences.

The introduction of new substances to the market is faster than our ability to examine their individual effects. Existing studies and results are based on a few organisms tested in closed environments, making it difficult to determine their broader effects in nature.

“70 to 80 percent of the available ecotoxicological data are based on a handful of standard organisms. 'When evaluating environmental pollutants in a life cycle assessment, the models used describe how individual species are affected, in order to extrapolate the effect is in an ecosystem,” says Oskar Nyberg.

New methods for better damage assessment

In cases where chemical releases have been evaluated, complete information on the substances is often missing, which underestimates the environmental damage. In his thesis, Oskar Nyberg investigated how to make better and more reliable forecasts in damage assessments, without having to conduct new animal tests. A new method is to use machine learning to estimate the uncertainties in previous studies and to predict the impact of unknown chemicals by comparing them with the properties of known chemicals.

“Such analyses can identify statistical relationships between the physical structure of chemicals and various toxicological properties. It becomes a screening that can indicate whether a substance is potentially toxic and which chemicals should be prioritised for further investigation," he says.

In January, Oskar Nyberg defended his doctoral thesis "Protecting food with poison: Exploring ecotoxicity of agrochemicals and pharmaceuticals". Photo: Isabell Stenson.

Despite an extensive library of toxicological studies, data quality is often uneven and insufficient. Oskar Nyberg therefore does not consider these assessments enough to draw definite conclusions or set limiting values in policy decisions.

“In order to make informed decisions about which product A or B is best, it is important to know the level of uncertainty in the background data,” he says.

“Life cycle assessments need to take into account that software does not yet react if there is no toxicological information for certain substances. Here, I would like to see a flag function that warns when information on certain chemicals is missing.”

Antibiotic resistance needs to be included

Interest in these questions grew when Oskar Nyberg studied Egyptian fish farms. There he found antibiotic-resistant genes in ponds fertilized with manure from poultry farms that use large amounts of antibiotics. Although resistance is a known threat to animal and human health, there was a need to develop methods how to evaluate such environmental risks in life cycle assessments.

“It is known that antibiotics reaching the environment can promote resistance in bacteria, and that resistant bacteria can spread genetic material to related bacteria. Evaluating resistance development in the environment according to the presented methodology allows for a novel and niched application in food production, as well as in cases of sewage treatment and pharmaceutical production,” he says.

According to him, this knowledge gap can be closed by shifting the focus of toxicological models from the point where antibiotic concentrations become toxic to animals, to the point where concentrations cause bacteria to develop resistance, which usually occurs at much lower concentrations.

In his research, he has developed a database and an app that allows us to consider the uncertainties in the estimated environmental impacts based on quality and availability of the underlying data. But Oskar notes that this is only a small piece of the puzzle in understanding the effects of chemical use in food production.

“We only have fragmented information about how harmful chemicals are. We still have a long way to go before we can clearly and reliably account for whether toxic effects to the environment are larger or smaller for a particular food product, even though we release huge amounts of potentially toxic chemicals into the environment in order to protect our food,” he concludes.

Read the thesis: 

Protecting food with poison: Exploring ecotoxicity of agrochemicals and pharmaceuticals

Text: Isabell Stenson