Human impact on fish habitats detected with AI

A fundamental challenge exists in ecology when trying to understand why a species lives in some locations but not others. Natural factors such as the climate and the availability of food are decisive in determining where each species can best live. Yet, research often overlooks how humans alter the available areas within this optimal habitat. Freshwater ecosystems are some of the worst impacted by human activities. As such, it is crucial to understand where humans are negatively impacting freshwater organisms within its optimal habitat.

In a new study published in Nature Communications, researchers from the Department of Aquatic Ecology and Evolution at the Institute of Ecology and Evolution at the University of Bern, in collaboration with Eawag, show that almost 90% of potential freshwater fish habitats in Switzerland are negatively impacted by human activities. These include unnatural features in watercourses such as artificial banks and barriers that hinder the migration of species in rivers and lakes. "The research highlights the importance of knowing how much a species' habitat is affected by human activities in order to identify the factors that pose the greatest threat to populations," says lead author Conor Waldock from the Institute of Ecology and Evolution at the University of Bern.

Uncovering hidden life using artificial intelligence

To identify the areas of fish species' natural habitat where humans have the greatest impact on fish populations, the researchers compiled data on the populations of nine fish species throughout the Aare-Rhine catchment. They then linked this data with numerous environmental factors, representing both natural and human-caused influences. To find correlations between the environmental factors and the occurrence of fish populations, the researchers used a research approach based on machine learning. However, this traditional approach could only show where species are most likely to occur, but not what factors enable or prevent their occurrence there. "Therefore, in a next step, we applied 'Explainable Artificial Intelligence' which showed for 15,000 river catchments in Switzerland which environmental factors determined whether the respective location is good or bad for the individual species," explains Waldock.

Potential habitats for freshwater fish are large but under threat

The research results show that 40% of all river basins would be suitable habitat for the endangered Spirlin (Alburnoides bipunctatus) studied in the study, due to the natural conditions for survival, growth and reproduction of the species. However, most of this suitable habitat is affected by human activities that have a negative impact on the species, as the study shows. "These results also provide new insights into the possible causes of the decline of the Spirlin," says Waldock. Similar results were found for the other fish species studied with an average of around 90% of potential habitats being negatively impacted by human activities. In around half of the river locations identified as potential fish habitats, multiple human threats act together to negatively impact fish populations.

"It is worrying that in such large areas where species could potentially live, the habitat quality for these species is worse than expected due to past and present human interventions in the rivers," says Waldock. These areas within a fish species' optimal habitat that have been negatively impacted by humans are referred to as the species' 'shadow distribution', a new term introduced in this study. "Without revealing the shadow distribution of species, we may not have realised the full potential of biodiversity in Swiss river ecosystems. With this scientific knowledge, we are in a better position to act on behalf of degraded biodiversity," Waldock continues. "There are many research groups investigating how the environment impacts the distribution and threats of species. However, this new study is probably the first to separate human impacts from natural factors to provide more nuanced information for species conservation" says Ole Seehausen, co-author and Professor of Ecology and Evolution at the University of Bern.

Scientific foundation to promote biodiversity

Building on the current study, the research group, together with multiple partners in the Institute of Political Science at the University of Bern and the Swiss Competence Center for Fisheries, is now developing an approach to identify the most important locations and measures for the conservation of biodiversity in river ecosystems. This approach will jointly consider the areas where species currently exist and the newly discovered shadow distributions where they could potentially live, as well as the locations where these species will be most affected by the impacts of climate change. "The limited resources available for conservation should be strategically invested to maximise the benefits for biodiversity. We hope that the scientific knowledge we have gained here will provide directly relevant information to environmental decision-makers," concludes Waldock.

This study is part of the project LANAT-3 "Stopping the loss of freshwater biodiversity - despite climate change". This project is funded by the Wyss Academy for Nature through the implementation program with the canton of Bern (Office for Agriculture and Nature) and by the Federal Office for the Environment (FOEN).