Tree by Tree: How We Simulate Forests

When a company supports a Pina climate protection project, there’s a question behind it that isn’t easy to estimate: How will this forest develop over the next few decades? The answer comes from a simulation that goes far beyond a rough projection. This article explains in simple terms how it works—without requiring any prior knowledge of forestry or computer science.

A digital twin of the forest—tree by tree

Many growth projections in forestry are based on average values per hectare. In other words, they treat a forest as if the same “average tree” were growing everywhere. In a real forest, things are different: Every tree has neighbors, and they help determine how well it grows. A tree that is closely surrounded by others receives less light and space for its crown and therefore grows more slowly than a free-standing tree of the same species and size.

Pina’s simulation captures precisely this effect. It records every single tree, including its location, height, trunk diameter, crown, and species. And it calculates how close its neighbors are and how large they are. Technically, this is called a tree-by-tree and distance-dependent growth model. It is based on TreeGrOSS, a recognized model developed by the Northwest German Forest Research Institute (NW-FVA). The result is a virtual forest in which each tree grows based on its actual neighborhood.

A forest is not considered as a whole, but is divided into what are known as stands. A stand is a section of forest that is fairly uniform in itself—for example, with similar tree species, similar ages, and similar soil. For this reason, it is planned and managed as a single unit. A forest area usually consists of many such stands, and the simulation calculates each one individually.

What does the simulation take into account? The data set

A simulation is only as good as its data. Pina uses three sources:

• The forest inventory. For each stand, the following data are recorded: tree species, age, area, site productivity (site class), timber volume, and average height and diameter. Using this data, Pina generates a virtual forest with individual trees that match the measured values. Each simulation starts from this unaltered initial state.
• Regional reference data. The growth region (an area with similar growing conditions) and yield tables determine how a tree species typically grows in that region.
• Climate data. For each region, there is a climate time series covering each year through 2100: temperature, precipitation, and water balance. This data determines the climate risk—that is, how likely it is that trees will die due to heat or drought. In this way, climate factors are incorporated into the forecast, including for the coming decades.

Step by Step: How the Simulated Forest Grows

The simulation runs in increments of usually five years. In each increment, the following occurs in each stand:

• Mortality: Some trees die, for example, due to competition, old age, or climate-related risks.
• Competition: For each tree, the extent to which its crown is crowded by neighboring trees is calculated. A free-standing tree faces little competition, while a tree wedged between others faces a great deal.
• Growth: Every tree increases in height and diameter. The rate of growth depends on the species, location, and, specifically, this competition. Trees under pressure grow more slowly.
• Canopy Development: The tree canopies are updated. They determine the level of competition in the next step.
• Regeneration: Where light filters through gaps in the canopy, young trees regrow on their own. Shade-tolerant species such as beech or fir can do this even under a dense canopy, while light-loving species such as oak or pine need a clearing.

Forest Management in the Simulation: Thinning, Harvesting, Planting

In addition to natural growth, the simulation also models the actions taken by foresters. They typically intervene every five years:

• Harvest: Mature trees that have reached their target diameter are harvested. The target diameter is the diameter at which a tree species is considered ready for harvest.
• Thinning: In addition to harvesting, individual trees are removed so that the remaining trees have more space and light. The simulation can do this in two ways. Either the weakest trees are removed so that the strongest ones can grow better, or particularly promising “future trees” are specifically selected and their strongest competitors are removed. The second method is called selective thinning and is common in modern German forestry practice.
• Planting: If gaps appear in the forest after harvesting and not enough young trees regrow on their own, the simulation plants new trees of the desired species. This continues until the area is once again densely forested.

There are clear limits to these interventions. As a rule, no more timber is harvested than a stand can sustainably yield (see next section). Habitat trees are never felled. These are mostly old trees with hollows, cracks, or deadwood that serve as habitats for many animals, such as woodpeckers, bats, and rare beetles. And when individual trees are harvested, the canopy remains dense enough that the forest stays stable and can regenerate.

The Target Scenario: the Forest Development Type (WET)

In the simulation, each stand is assigned a target scenario: a desired mix of tree species, such as a mixed beech forest with fir, sycamore maple, and a smaller proportion of spruce. This concept originates from German forest planning and is called a “forest development type,” or WET for short. It is defined for each stand and describes what kind of forest should develop at that location over the long term—for example, a climate-resilient mixed forest instead of a spruce monoculture. 

Planting and regeneration are guided by this goal. Species that are still missing from the target composition are specifically planted. In natural regeneration, the process is initially left to chance. Once a species has reached its target proportion, no more young trees of that species will grow. In this way, the simulated forest moves step by step toward its target state, rather than being left to its own devices.

How much wood can be harvested? The Gerhardt Method

To ensure that a forest remains a long-term CO₂ sink, no more timber should be harvested than can regrow. The simulation calculates this amount for each stand individually using a classic forestry method known as the Gerhardt method.

To do this, the simulation allows the stand to grow a little into the future on a trial basis, without any intervention. This shows how much it actually grows. In addition, the simulation compares whether the stand contains more or less wood than would be typical for its location and age. If it contains less, less wood may be harvested so that it can rebuild itself. If it contains more, a little more may be harvested.

However, this adjustment does not happen all at once, but is spread out over 40 years. For example: If a stand is short 40 m³ of timber per hectare, the harvest is not reduced by 40 m³ all at once, but by about 1 m³ each year. This ensures that forest management remains consistent and predictable.

The result is an upper limit per stand and time period that the harvest generally must not exceed. Because a forest’s age, timber volume, and growth change over the decades, this upper limit is recalculated every five years.

Why You Can Trust the Results

What does this have to do with CO₂ credits? The simulation shows how the forest—and thus its CO₂ storage capacity—develops. The credits are based on this. That’s why it’s not just important that a forest is likely to grow, but that every single figure behind it is verifiable and reproducible. Two features of the simulation ensure this:

• Reproducibility: Each simulation starts with a random value (“seed”). Randomness is built into the simulation because, for example, the death of individual trees and natural regeneration depend in part on chance. Using the same input data and the same seed produces exactly the same forest, tree by tree and step by step. A result can therefore be recalculated and verified at any time.
• Scientific Basis: The growth model follows recognized forestry methods established in Germany, including TreeGrOSS from the NW-FVA and traditional yield tables.

The purpose of this attention to detail: Anyone who relies on a Pina project is relying on a transparent calculation, not an estimate.