Finland is number one in forest artificial intelligence and Joensuu is its lodestar

Finland is the only country in the world whose forests have been mapped down to every single tree, stump and stem. When there is enough data, artificial intelligence can help prevent forest fires and wildfires, reduce carbon dioxide emissions and improve biodiversity.

Through the ages, forests have been a source of support and security for Finland: they not only serve as an economic guarantee and the driving force of Finnish society, they also offer a spiritual landscape. However, there is no thick spruce stand or remote wilderness in Finland that artificial intelligence has not found. Indeed, Finland is the only country in the world whose forests have been mapped down to nearly every single tree. Soon, even twice over.

Laser scanning of forests is a way to map what lies within. In practice, laser scanning involves remote sensing with drones and airships, among others. It enables artificial intelligence to analyse and describe in greater detail the types and locations of forests in Finland, what has been done in them over time and what should be done – or not done – in the future.

Even before laser scanning, Finnish forests have been relatively well mapped. Surveys conducted on foot by human means can only achieve a fraction of the accuracy with which modern technology maps forests. For example, Joensuu has played a key role in the development of forest surveys, serving as a pioneer of research, education and development in the forest sector.

The best forestry expertise in the world

Joensuu’s status as a pioneer in the world of forest technology and knowledge is no accident, but rather the product of decades of development work. The expertise found here is among the best in the world and we intend to remain at the forefront going forward, as well.

“As early as the 1980s, Joensuu made a strategic decision to invest in forest resource inventorying and planning methods. It led to an accumulation of top global expertise in the field in the Joensuu region,” explains Lauri Sikanen of the Natural Resources Institute Finland (Luke).

Since then, a dense forestry sector ecosystem has grown in the city, where the University of Eastern Finland, Luke, the European Forest Institute, Karelia University of Applied Sciences and dozens of technology companies cooperate. In 2023, this ecosystem officially became Forest Joensuu, and in November 2024, it received the Internationalisation Award of the President of the Republic, which is granted to Finnish companies that have gained international success, and to the communities behind them.

Indeed, the largest forest machine factory in the world, John Deere, is located in none other than Joensuu. Data collected from forests and the artificial intelligence taught using it will enable, for example, the development of autonomous forest machinery in the future.

“We have a unique situation where research, education and industry in the forest sector are all under the same umbrella. There are no such hubs anywhere else,” says Seppo Tossavainen, Head of Economic and International Affairs for the City of Joensuu.

Forest, mathematics and love

It was love that originally brought Tuomo Kauranne to Joensuu decades ago. When he could not find any work suitable for a mathematician, he had to come up with something himself. This led to the establishment of Arbonaut, a 30-year-old technology company in Joensuu that develops precision digital solutions for forest management.

Arbonaut’s own history extends far beyond the forest. The company had once operated, among others, an online shop with the S-Group and developed positioning functions for mobile phones. However, the time was not right for them and Google ultimately took the lead in mobile map and positioning services. As a result, Kauranne and Arbonaut had to find another direction. In Joensuu, the forest was the natural answer.

Machine vision and artificial intelligence algorithms had already been used by Arbonaut in the 1990s, but the current operating model for forests is more recent, only dating back to this millennium. Now the company operates in more than 40 countries and has been a pioneer in the utilisation of artificial intelligence and machine vision in forest mapping.

“Our task is to refine basic data as accurately and as usefully as possible. In practice, we transform the data collected with aircraft, satellites, drones and airships into sensible and usable information,” explains Kauranne.

There are approximately 20 billion trees growing in Finnish forests. Almost all of them can be given their own “ID card” with the help of artificial intelligence, which specifies, for example, the species, height, volume and location of the tree. Data is used not only in forest management but also, for example, in climate change management as well as the assessment and safeguarding of biodiversity.

“Nature is very sneaky – even when you are looking for similarities, there are always differences. Artificial intelligence can be taught to learn based on the variations in the data, thus allowing it to develop itself in the analysis of the data,” says Sikanen.

Reducing world forest fires with artificial intelligence

One possible task of forest artificial intelligence is to prevent and extinguish wildfires and, in particular, forest fires. Massively destructive wildfires have become more common around the world due to climate change, causing damage not only to nature, but also people and the climate.

Artificial intelligence can identify in advance risk factors that accelerate forest fires, such as dry grass, tree branches and other highly flammable material.

“Vertical fire breaks are vital to forest fire prevention. Tree trunks don’t burn, and if a gap is cleared between the branches and the ground, the fire cannot climb along the trunk. This makes it easier to control and extinguish fires. Canopy fires are more problematic, as they are easily pushed by the wind and they move so fast that there is no escape,” explains Kauranne.

Arbonaut has also carried out laser scanning of forests in Portugal, France and Spain, but nowhere else is this as systematic as in Finland. However, there is potential, and the Joensuu locals hope that the benefits of artificial intelligence will be understood globally. “Amazingly little is currently being invested in wildfire prevention in the countries of southern Europe or, for example, Canada and the United States,” Tossavainen points out.

“At worst, forest fires already cause a huge amount of greenhouse gas emissions globally, so combating them would also be vital for mitigating climate change.

From a climate perspective, forest artificial intelligence also plays a major role in emissions trading and carbon sink calculation. For example, Arbonaut and Luke have been calculating Finnish carbon sinks, which were revealed to be a source of carbon emissions thanks to more accurate calculations.

Artificial intelligence also helps to preserve biodiversity, as data can be used to more effectively identify valuable sites and there is even tree-specific data. “Biodiversity and efficient harvesting are also not necessarily in conflict, as areas that are often important to biodiversity are not good for harvesting,” says Lauri Sikanen from Luke.

“With the help of data, valuable sites can be identified in advance and barred from harvesting. They are often difficult for machines, as the trees are perhaps smaller, the stands are more diverse, there are differences in elevation, moisture and other factors.

Seppo Tossavainen and Tuomo Kauranne present Hiil Oy’s charred wood, which is a sustainable building material.
Forest AI development is just beginning

Although artificial intelligence is already doing a lot in forests, its development is only just beginning. There is also a lot more going on in the forest sector, and there is, for example, a project under way in Joensuu to combine the forest industry and the hydrogen economy. A P2X Solutions hydrogen plant, which focuses on renewable energy, is planned for the city, and the decision to build it will possibly be made this year.

If the carbon dioxide that has so far been dumped into the air by energy companies burning wood can be captured and combined with hydrogen, this will produce biomethanol, or biofuel. This would simultaneously cap the smokestacks and produce a new, valuable product refined from the forest.

Another significant project is UNITE Research Council of Finland Flagship of Science, which is funded by the Research Council of Finland. Approximately 150 researchers are involved in the project. Its goal is to make Finland a leading centre of expertise in forest bioeconomy and forest data, spatial data technology and gamification. In addition, an enormous number of other projects are under way in the Joensuu region, such as the creation of digital twins for forest roads.

The guiding idea behind this development is that, in decision-making concerning the forest, we would also be able to take more account of the rest of creation – not forgetting forest owners and their income,” says Seppo Tossavainen.

“I believe that artificial intelligence can also be used to find alternative ways to bring income to forest owners, and forests don’t necessarily need to be felled to earn money from them. Digital data can be used to target measures and income generation better than before.

In the future, Joensuu aims to be a digital hub, focusing not only on forests, but also more extensively on environmental modelling, which also applies forest solutions to, among other things, digital border security.

“We could have a data pool that brings together detailed data on forests, waterbodies, settlements and climate. It would be a growth platform for innovations and business activities,” says Tossavainen.

Key actors in the forest sector, such as Metsähallitus, the Finnish Forest Centre or large forestry companies, are already using artificial intelligence in, for example, forest resource inventorying and decision support systems. With digitalisation, the range of uses for artificial intelligence is constantly expanding. Over the next few years, we will see what kind of opportunities the world’s first forest artificial intelligence really brings – to Joensuu, Finland and the world as a whole.

Joensuu is one of the InnoCities. The key areas of development in the city’s innovation activities are forest bioeconomy and photonics.