Correction: The area of tall forests is not declining in Finland and Sweden

The World Resources Institute published an Insights article in June 2024 highlighting that the area of tall Nordic forests has declined by 20% over the period of 2001-2021. In this response we show that the conclusion is completely wrong. Drawing conclusions about forest development based solely on remote sensing data can lead to highly erroneous conclusions and should be avoided whenever possible.

The conclusion of the Insights article is based on a study partly funded by the World Resources Institute. The Insight article states that the largest decline in the area of tall forests was seen in Denmark, Finland, Norway and Sweden. This contradicts Figure 6B of the original article, which clearly shows that virtually all of the claimed loss of tall forests occurred in Finland and Sweden.

In the study the 77^th Airborne Laser Scanning (ALS) canopy height percentile (P77) was used as a surrogate for tree heights in the Nordic countries. Actual trees measured in the field were not used at all. Tall tree forests were defined as areas with a P77 value ≥15 meters. We translated this P77 ≥ 15m definition into a forest height attribute using actual field-measured data from Central Finland. Our field plot size was 30×30 meters, which corresponds to the plot size used in the study, i.e. approximately Landsat pixel size. We calculated the P77 values from the ALS data and compared these values with the basal area weighted mean heights (BWH) of the field-measured trees. BWH is a commonly used height attribute of forests in the Nordic countries.

Our analysis (Figure 1) shows that the P77 values correspond well to the values of BWH minus 3 meters. In other words, the definition of P77 ≥ 15m can be translated into the forest height attribute of BWH ≥ 18m. From now on we will use the definition of BWH ≥ 18m for tall forest.

We estimated the area of tall forests using National Forest Inventory data. National Forest Inventories provide time series of field plot data from which reliable estimates of forest development can be obtained. The estimations were harmonized for Finland and Sweden: A field plot was considered as tall forest if its BWH ≥ 18 m. 

In Finland, our estimation covers the period 2004-2021. The estimates are not annual, but averages of measurements made during 2-to-5-year periods; the estimates were assigned to the middle year of the measurements. 

In Sweden, our estimation covers the period 2001-2021 and the estimates are 5-year moving averages. Estimates for both countries include all productive forest land outside formally protected areas.
 

Figure 2 shows the area of tall forests based on time series of National Forest Inventory data. In Sweden, the area of tall forests has increased from 5.1 to 6.4 million hectares from 2001 to 2021. In Finland, the corresponding area has increased from 3.4 to 4.7 million hectares from 2004 to 2021. 

In other words, the relative increase was about 25% in Sweden and 35% in Finland. This strongly contradicts the claim that the area of tall forests in the Nordic countries has declined 20% over the last two decades.

Estimates from the National Forest Inventory data clearly show that the area of tall forests in both Finland and Sweden has increased substantially over the period of 2001-2021. The reason for an opposite trend in the above-mentioned study remains unsolved. However, it is obvious that analyses of forest development based solely on remote sensing data can lead to highly erroneous conclusions and should be avoided whenever possible.

Petteri Packalen (petteri.packalen@luke.fi), Minna Räty, Annika Kangas, Kari T. Korhonen
The Finnish National Forest Inventory, Natural Resources Institute Finland

Jonas Fridman (jonas.fridman@slu.se), Cornelia Roberge, Mats Nilsson, Göran Ståhl
The Swedish National Forest Inventory, Swedish University of Agricultural Sciences