The research that brought Impact Forestry to life

Key takeaways

• Gaps in data and the complexity of forest ecosystems make it difficult to predict how mature forests store carbon over time
• By combining scientific research with forestry expertise, we developed a solution that provides clear insights into the carbon potential of mature forests and empowers landowners to make informed decisions about management
• This article focuses on Estonia, explaining how we developed forest models and refined an approach that can be applied across different markets

Sustainable forestry is an essential tool in the climate fight. While planting new forests captures additional carbon, improving the management of existing forests increases their ability to store carbon over time—making them an equally vital part of the solution.

In 2022, we introduced landowners to the opportunity of earning carbon credits through afforestation. Now, we have developed a solution for existing forests (Impact Forestry). Our vision is clear: to protect Europe’s mature forests by empowering landowners with financial incentives.

Developing this new solution required us to us to ask critical questions, such as:

• What are the outcomes for the forest if it is allowed to continue growing without harvesting?
• How does the biomass of different tree species vary across forest types when left unharvested beyond the minimum legal harvesting age?
• Will the trees keep accumulating biomass, or will their biomass eventually decline?

However, getting answers to these questions was more challenging than anticipated.

This article takes you behind the scenes of the investigation phase of our product development and the hurdles we overcame to bring our idea for Impact Forestry to life.

Challenges in forest data

Researching what happens when forests grow past their legal harvesting age posed immediate challenges. We looked at all tree species, and asked questions like:

• What happens to a 150-year-old pine forest in the cladonia habitat type (a common habitat type for pine in Estonia) if left to grow?
• Will it continue to grow and sequester additional carbon from the atmosphere for decades to come?
• Or will it gradually decay and release its stored carbon back into the atmosphere?

However, the available data on tree growth primarily covers the early decades of development; since commercially managed forests are typically harvested long before reaching 100 years of age, there has been little incentive to look beyond this timespan.

Consequently, there is limited research on the natural lifespan and growth patterns of older trees, leaving a gap in long-term forest modelling beyond 100 years.

This complexity called for a collaborative, multi-disciplinary approach.

Collaborating with scientists to find answers

To address this gap in the data, we collaborated with forest and ecosystem scientists to create empirical growth curves for Estonia's main tree species. Our goal was to understand how tree biomass changes with age across various habitat types and environmental conditions.

We are replicating this approach in every new country we enter, working with local data sources and scientists to ensure each country has growth models that most accurately reflect the local conditions.

Given the complexity of forest ecosystems, we knew there wouldn’t be a single, definitive criterion of truth. Forests are intricate systems where living (biotic) and non-living (abiotic) components interact dynamically and often in non-linear ways. Factors such as climate, soil quality, water availability, and human activity further influence these ecosystems, making them highly sensitive and variable.

To ensure our results were as accurate as possible, we decided to consult a range of widely used data sources:

• Alternative Assessment of Estonian Forests: Provided growth predictions up to 101 years
• Estonian Statistical Forest Inventory: Provided estimates to 150 years, though considered conservative for some species like pine and spruce
• Estonian Forest Registry: A comprehensive dataset covering all forest properties in Estonia, both private and state-owned

To predict forest growth, we needed models that accounted for the complexity of real-world conditions. So, in collaboration with local data and forestry scientists, we analysed 1.7 million forest stands from the Forest Registry to develop nonlinear biomass prediction models, considering factors like site quality, tree density, and more.

Why nonlinear models?

Unlike traditional models that treat all trees of the same species as identical, nonlinear models reflect the complex, real-world relationships between tree growth and environmental factors.

For example, a spruce tree growing near a riverbank, receiving an abundance of sunlight will grow differently from a spruce tree further from the river bank in shady conditions. Nonlinear models account for these site-specific variables, making them more effective for predicting biomass. In addition, they provide:

1. Biological accuracy: Tree growth is nonlinear—young trees grow rapidly, but as they mature, their growth slows down and biomass may even decline. Nonlinear models are more suited to capture these dynamics compared to linear ones
2. Flexibility: They account for diverse factors like species differences, site conditions and competition, which linear models cannot handle effectively
3. More realistic: They avoid oversimplification, capturing complex scenarios like biomass decline in mature trees due to decay or competition

Applying deterministic formulas for growth curves

To refine our predictions further, we applied the deterministic forest growth calculation formulas from the “Forest Management Guidelines” to the Forest Registry data.

The results varied considerably. Together with scientists, we developed a balanced approach that realistically represents the most likely biomass growth curves for each forest stand. These are our final growth curves for the most common tree species, categorised by site index (1a to Va) in Estonia:

Figure 1: Our final growth curves for the most common tree species in Estonia

What was the most challenging part?

Every tree’s growth is influenced by its specific conditions, and considering these diverse factors is not easy. This is mainly because some factors are impossible to measure, others are costly to track, or the sometimes the quality of data simply isn’t consistent enough.

Nature is complex. This is why analysing a variety of models and methods, rather than relying on one way only is a more comprehensive way to go. For example, if we predict a tree will grow 2 cm in diameter, but it actually grows 1.8 cm, that’s a small error.

At the stand level, we look at the average growth of all the trees. While some trees may grow slightly more than predicted and others slightly less, these variations generally balance each other out. So, even if our prediction for an individual tree is incorrect, we can still make informed decisions for the whole stand.

Final solution and what's next

We have already applied this approach in Latvia and Sweden, and will follow the same process in all new markets:

• Gather and analyse regional forest data
• Validate existing forest growth models
• Collaborate with local scientists and forestry experts to refine biomass predictions
• Develop precise, site-specific growth curves
• Ensure landowners receive the most accurate and actionable insights

This is just the beginning of how science, technology, and forestry expertise can come together to deliver meaningful results to the climate fight.

At Arbonics, our mission is to scale nature-based solutions through technology. Together with scientists, we developed software that predicts forest growth for the next 40+ years, assessing whether a given forest type is suitable for a carbon project or not and providing landowners with actionable insights. To our knowledge, no one else has software with this level of precision and capability.

To learn more about our proprietary technology, head to this article.