Predicting the risk of bear-human encounters through remote sensing
Viewing coexistence with wildlife as a challenge for society as a whole
In Japan, incidents of damage caused by encounters between bears and humans surged across the country in 2025, with the appearance of bears in urban areas becoming a serious issue in northern Tohoku, especially in Akita and Iwate. Why is the risk of encounters currently on the rise? And where is danger likely to grow in the future? This article introduces research by Professor Wataru Takeuchi of the Institute of Industrial Science, The University of Tokyo, which integrates remote sensing and machine learning to show the probability of bear appearances and how the risks may change moving forward.
What can we learn about bears from satellite images?
The term “remote sensing” refers to mounting sensors on artificial satellites or aircraft to measure the characteristics of objects from a distance, without making physical contact. Weather forecasting using the Himawari meteorological satellite is one such example. A key feature of this technology is its ability to capture a wide range of information at once, including weather conditions and land use, over vast areas.
Takeuchi, together with graduate student Didier Delgorge, conducted research using remote sensing techniques based on satellite imagery to predict the distribution of encounters between Asiatic black bears and humans. How can satellite images reveal the distribution of bears?
“Even though we are using satellite images in our research, that doesn’t mean we can count bears from space. Some studies directly observe animals using satellite imagery—such as research on penguins in Antarctica—but Asiatic black bears generally live solitary lives in forests covered by dense vegetation, so it is hard to directly perceive individual bears from satellites. What we are using satellite images to observe are human land-use patterns, vegetation, and other factors that shape bear habitats.”
Risks caused by the disappearance of buffer zones
One factor behind the growing severity of conflict between bears and humans is the decline in population management capacity caused by the aging and decreasing number of hunters. Another major factor is the decline in the number of people who live in and make use of rural areas. Rural landscapes actively managed by human hands used to function as buffer zones, separating the deep mountain habitats of bears from human living areas and helping to limit encounters between the two.
However, depopulation has led to an increase in abandoned farmland and poorly maintained forests. Tracts of land overgrown with vegetation provide bears with excellent hiding places, turning them into new areas of activity and habitat. This process is bringing human living areas and bear activity zones ever closer together.
“By analyzing satellite imagery, we are able to identify land-use patterns, discover the presence of abandoned farmland, distinguish between planted forests and natural forests, and even determine the distribution of tree species that serve as food sources for bears. We conducted our analyses by combining this macro-level information with micro-level data such as bear sighting reports, numbers of damaging incidents, local population figures, and fluctuations in acorn harvests.”
Visualizing the blurring boundaries in Akita Prefecture
The figure below shows an analysis of the relationship between the state of vegetation and the risk of encounters between Asiatic black bears and human in 2025. The left shows the analysis results for Toyama Prefecture, while the right shows those for Akita Prefecture. Red indicates a high risk of encounters, and green indicates forested areas.
In Toyama Prefecture, the risk around Toyama Station (★) is low, and the high-risk areas are scattered along the boundaries between urban districts and farmland, as well as within forested areas. By contrast, in Akita Prefecture, a large hotspot (high-risk area) has formed within Akita City, and even the area surrounding Akita Station (★) falls within the high-risk prediction zone. These findings highlight how, in Akita, the boundary between human living areas and bear habitats has become extremely blurred.
“One possible reason for this difference lies in topography. In Toyama Prefecture, the boundary between the plains that contain farmland and urban areas and the surrounding mountainous regions is topographically well-defined. In contrast, Akita Prefecture is characterized by forests and urban areas that adjoin each other in a mosaic-like pattern. The findings from Akita Prefecture may suggest that bears are no longer relying solely on forests, and are also beginning to make use of urban environments, attracted by the discarded food and fruit trees found there—in other words, they may be adapting to life in urban areas. However, we need to further verify this moving forward.”

Predicted risk maps showing the distribution of human–bear conflict risk in 2025
Credit: Wataru Takeuchi Laboratory
Simulating future risks based on population dynamics
One of the major achievements of this research lies in its ability to not only perceive past and present conditions, but also predict future encounter distributions. Takeuchi’s laboratory combined research data with future population projections over the next roughly 30 years—including age structure, gender, and migration rates—and estimated future risks using the MaxEnt model, a statistical method for predicting species distributions. Findings indicated that, in Toyama, the risks are likely to increase gradually across a broad area, whereas in Akita, they may rise sharply within more limited regions.
“Understanding how risks change spatially in step with demographic shifts is extremely important for achieving sustainable management in the future. Historical trends show that as populations decline, the number of people managing forests also decreases, resulting in an increase in abandoned farmland. Municipal populations are strongly affected by changes in social structure. Consider that even in municipalities where the population is projected to fall by half over the next 20 years, the number of bears will not decline rapidly. Only by taking such correlations into account can effective long-term measures be devised.”
Toward evidence-based social decision-making
Population decline, demographic concentration in urban areas, and the deterioration of rural functions—these trends will continue to accelerate. Without effective countermeasures, the risk of encounters with bears will only keep rising. However, maintaining buffer zones and managing wildlife entails a substantial social burden. Rather than placing this burden on specific regions alone, it needs to be addressed by society as a whole.
“As the term evidence-based policy making (EBPM) suggests, we are entering an era in which science is applied in social decision-making processes. We have to decide how to prioritize and allocate our limited funds, time, and human resources. I believe that evidence generated through research can provide a valuable basis for making such decisions.”
We need to balance environmental conservation with economic viability, and decide how much risk we are willing to tolerate with regard to encounters with wildlife. We should consider these issues––not only to accommodate human society––but also in a way that takes a comprehensive view of the entire ecosystem.
“Rather than relying on ad hoc measures or leaving the issue to municipalities in areas where bears frequently appear, it is essential to pursue comprehensive initiatives with a long-term perspective. The first step toward becoming a mature society is to accurately understand the current situation and reflect on what each of us can do, and we hope that our research can help provide data necessary to achieve that.”
Possibilities provided by citizen science and the accumulation of data
Takeuchi’s laboratory engages in a wide range of activities as a core hub for remote sensing in Asia, from making practical use of satellite data through to conducting field surveys. The common theme running through this research, which spans a diverse range of fields—including the environment, disasters, agriculture, and urban studies—is the relationship between human activity and nature.
Human-elephant conflict risk assessment under coupled climatic and anthropogenic changes in Thailand(2022) Credit: Wataru Takeuchi Laboratory
Takeuchi teaches the importance of monitoring continuously and accumulating data. Even images from the past can begin to reveal new insights when analyzed using the latest technologies. It is the long-term accumulation of data that holds the key to highly accurate future predictions. In addition to a broad range of data, Takeuchi also emphasizes the value of information that can only be obtained in the field.
“With the spread of social media, it has become easier to collect reports of wildlife sightings. Posts from citizens on-site about wildlife—as well as about disasters and infrastructure anomalies—provide important clues for research and countermeasures. These sorts of citizen-driven efforts are known as ‘citizen science.’ We hope to convey its value and inspire more people to take part by proactively communicating how the collected data is being put to use.”
There is a concept known as “One Health,” which recognizes that the health of humans, animals, and the environment is deeply interconnected. Efforts to maintain healthy ecosystems ultimately comes full circle, contributing to human well-being and happiness. The more people there are that take an interest in and reflect on the type of society we should build—in coexistence with wildlife—the clearer our path toward a future that we can all embrace will be.
“Human-elephant conflict risk assessment under coupled climatic and anthropogenic changes in Thailand”, is published in Science of the Total Environment Volume 834 (2022) at DOI: 10.1016/j.scitotenv.2022.155174

Featured Researcher
Wataru Takeuchi (Professor, Institute of Industrial Science, The University of Tokyo)
Special Field of Study: Remote Sensing for Environment and Disaster
Comments
No comments yet.
Join by voting
How did you feel about the "Possible Future" depicted in this article? Vote on your expectations!
Please visit the laboratory website if you would like to learn more about this article.
Share