Photo: Jon Andersson
Our goal
Sweden is home to an important part of the EU's remaining more natural and near natural forests and forests that have not previously been clear-cut, so-called continuity forests. These forests are unique environments for biodiversity, with many species associated with older and old forest, dead wood and continuity of tree cover and of different structures and substrates. They are also very important for the climate through their large carbon stocks in the soil and biomass, and their resilience to disturbance.
There is a long list of environmental targets that relate to forests, at UN, EU and national levels. In order to achieve these targets, the forests that currently hosts high conservation values must be possible to locate and demarcate in a functional way, and it is urgent! We have developed and collected data for this map service, to present an overall picture of where the most biologically valuable forests are, or may be. But we also want to provide a good overview where there are valuable concentrations, or clusters of forests worth preserving with functionality, and where restoration is needed. The mapping also contributes to make inventory efforts more effective, by delineating the areas with the greatest potential.
Our vision
Our vision is that the Forest Monitor will grow into a mapping of additional parts of Europe's forests, outside of Sweden, a mapping where most of the natural heritage in the forest is included.
Our vision also includes the development of a fully automated interactive system, with planning tools, as well as relevant monitoring tools.
Photo: Tor Tuorda
Photo: Jon Andersson
Our method
In short, we use various remote sensing, AI, as well as forest history research, historical maps and already existing modern open source mapping, to create our own data. This data is used to map areas of potential older forest and continuity forest. To validate, we use, among other things, field inventory (large reference areas of entire landscape sections of thousands of hectares), as well as open data from various inventories. By using a combination of data modeling (species and habitat), and existing open data on conservation and natural values, from authorities, researchers, municipalities and non-profit nature conservation organisations, we then classify these areas into:
1. high conservation values
2. probable conservation values
3. potential older forest and/or continuity forest (depending on regional forestry history)
The degree of accuracy of natural values in the classification thus depends on whether the areas' natural and conservation values are confirmed, i.e. usually mapped in the field (high conservation values), otherwise able to be validated with supplementary assessments (probable conservation values), or are only the result of a remote sensing (potentially older forest or continuity forest). The last category likely contains significant areas of forest with natural and conservation values, but these values are currently unknown to us at the Forest Monitor. When it comes to areas with high conservation values, our data presents an underestimation. Regarding the area of older forest and continuity forest, our data currently presents an overestimation of the area, especially in southern Sweden and in the County of Gävleborg.
Southern Sweden 1.0
The map layer of potential older forest (including continuity forests) in the southern half of Sweden, has been built by and classified by skogsmonitor.se, and has relative accuracy to show older forest that was not felled during the most recent intensive period of use of clear-cutting as a completely dominant method, from 1950 and onward. The mapping also excludes forest areas smaller than 0.5 hectares, which means that certain forest fragments with continuity forest, older or old forests are not included in the mapping. This is because they are small or consist of fragmented smaller pieces. The mapping of certain forms of deciduous forest with conservation values ​​also has limitations in the 1.0 version. The maping of Kalmar and Örebro County has been refined and has a lower degree of error. Note! This is our first version, and in some parts of the landscape there is a relatively large over-mapping of older forest, which is due, for example, to method limitations, as well as clear-cuts that were missed by the Swedish Forest Agency's data showing logged areas. Some other not accurate data may also occur due to other potential sources of error, which are due to limitations in this version of the remote sensing analysis. We are currently working with an improved method and will present new, more precise data during 2023.
Photo: Viktor Säfve
Photo: Jon Andersson
Northland and the counties of Dalarna and Värmland 1.0
The map layers for potential older forests and continuity forests (forests that were not felled during the great clear-cut era from the 1950s onwards) in the northern half of Sweden are based on the Swedish Environmental Protection Agency's remote sensing project that was created between the years 2016 - 2023. These map layers have been modified, and then in their entirety classified according to the three conservation/natural value classes, by us at skogsmonitor.se. Until we have done our own remote analyzes over the rest of Svealand and the Norrland region, the Swedish Environmental Protection Agency's mapping will continue to be the foundation to the layer showing potential old forest and continuity forest in these parts of the country. Note that parts of the Swedish Environmental Protection Agency's mapping of potential continuity forest have a higher accuracy (the region near the mountains, Norrbotten, Västerbotten, Jämtland, Dalarna and Västernorrland), while Gävleborg and Värmland so far have a lower accuracy.
More on precision and data quality
Note that the mapping consists both of data based on field inventory (very high precision) and data based only on remote sensing and AI, automated aerial image interpretations and change detection analysis based on a series of satellite images.
Our validations show a high degree of accuracy for areas that have only been mapped with remote sensing in the counties where the data has been updated, and a relatively high, but varying degree of accuracy for the non-updated counties.

​
The data layer of potential older forests and continuity forests that the remote analysis points out, contains various degrees of error. It is important to point out that both clear-cuts and young managed forests are sometimes incorrectly mapped as potential older forest or continuity forest, especially in those parts of the country that are not updated using AI (Skogsmonitor's data) or visual inspection (EPA:s data). 

Bit, despite varying degrees of over- and mismapping, a large percentage of the mapping is correct even in these counties.
Therefore, we assess that the 1.0 map layer of potential older forest and continuity forests, even in the counties that have not been updated, is useful for locating older forests and continuity forests.


Continuity forest or not? 

The different layers of the map vary in accuracy as to whether an older forest is a continuous forest or not.

Forest history differs between different parts of the country. While forest that was not logged during the great clear-cutting era (from 1950 onwards) is usually continuity forest in northwestern Sweden, the mapping in Norrland's coastal region, and in Svealand and Götaland contains both older forest with disrupted continuity historically, as well as old forest and continuity forest.
Forests are continuously affected by logging, forest management and other exploitation, which is why there may be a lag that affects the accuracy of our data.

We are constantly striving to refine and update!

The map shows the counties that are updated with new data and those that are waiting to be updated. The counties that are updated are all counties in Norrland (dark green) except Gävleborg (light green). In Svealand and Götaland, only Kalmar County, Örebro County (red) and Dalarna County (dark green) have been updated. Orange and red colors show Skogsmonitor's own data in the southern part of the country. Green colors show the area where the mapping is based on processed and classified data from the Swedish Environmental Protection Agency.
Photo: Jon Andersson
Continuity forest
Continuity forest is defined by the Swedish Forest Agency (2011) as: “A forest that has natural values whose occurrence is explained by the fact that for a long time there have been suitable forest habitats and substrates in this particular forest or in its vicinity“.
The definition includes forests that have never been clear-cut and includes the majority of all forests (particularly in the northwest of Sweden) generated before clear-cutting was introduced as common practice, and on a large scale.
Many species are associated with continuity forests, and these forests have a unique biological diversity. The Swedish Species Information Center at the Swedish University of Agricultural Sciences states that: "Lack of continuity forests, i.e. forests that have never been clear-cut, is one of the main reasons why forest-dwelling species are listed on the Swedish Red List." and that: "In order to reverse the trend of declining populations, unprotected forest environments, with habitats for red-listed species, need to be preserved in the long term throughout the country.".