The Swedish Forest Agency uses a model for calculating annual gross and net felling. The model is based on consumption statistics, import and export statistics, stocks statistics and data on felled but not removed whole trees. The consumption statistics consist of consumption of sawlogs, pulpwood roundwood for energy purposes (fuelwood) and other roundwood consumption.
The survey is part of Sweden's official statistics and has product number JO0312.
The annual statistics include preliminary and final felling statistics expressed in cubic metres standing volume (m3sk, the trunk volume above the stump including top and bark but excluding branches) and net felling in the unit cubic metres solid volume excluding bark (m3f ub). The gross felling is presented by county and the net felling by assortment of stemwood. The statistics also include a forecast for the current year's gross felling.
The statistics also consist of felling data from the SLU (Swedish University of Agricultural Science) National Forest Inventory. These statistics refer to information on felled and thinned area and the distribution of felling by tree species. SLU National Forest Inventory's data is based on a sample-based inventory and the data are reported as five-year averages.
The trunk volume above the stump including top and bark but excluding branches.
The trunk volume except bark, top and branches.
Total felled trunk volume above stump, including both removed trunks and felled trunks not removed from the forest. Usually expressed in m³sk.
Total felled trunk volume above the stump, excluding not removed felled whole trees. Expressed in m³sk.
Total felled trunk volume above the stump, excluding bark, tops and not removed felled whole trees and not removed stem parts. Expressed in m3fub.
Sawlogs are rough logs from trees that will be sawn to sawnwood products. The category includes roundwood for beams, sleepers, and veneer sheets of softwood. As of 2018, even roundwood for industrial production of timber piles, posts, veneer sheets and matches of hardwood are reported under the category of sawlogs.
Pulpwood is roundwood used to produce pulp and in the next step paper and board products. As of 2018, roundwood to produce fiberboard is reported in the pulpwood category.
Fuelwood refers to stemwood used for energy purpose. Fuel wood includes traditional firewood used in for example detached houses and chips made of stemwood used in heat and power plants.
Other wood consists other assortments than sawlogs, pulpwood and fuelwood. Before 2018, the category consisted of roundwood for production of props, posts, charcoal, veneer sheets and matches of hardwood, wood wool, mining timber and agricultural timber excluding sawlogs. As of 2018, we do not include roundwood for the industrial production of timber piles, posts, veneer sheets and matches made of hardwood or fiberboard in the category. See sawlogs and pulpwood.
Felling where the stand density is lowered below 0.3. The purpose is usually to achieve new forest (regeneration felling).
Thinning of forests with removals of trees.
Stand management thinning of young forests without actual removal of trees.
Harvesting of seed trees and other trees not included in final felling, thinning and pre commercial thinning.
Individual private owners, estates and not limited companies (sole trader).
The State, state-owned companies, other public owners, private-sector limited companies and other private owners such as religious associations including the Swedish Church, privately owned foundations and funds, profit and non-profit associations, profit driven community groups.
Forest land which, according to established criteria, can produce at least, an average, one cubic metre standing volume per hectare and year.
This includes productive forest land, bogs, mountains, mountains, mountain coniferous forests, fields, natural pasture and other land (such as roads, power lines and storage sites).
Developed areas including urban areas such as parks, industrial sites and various facilities with other use than forestry.
How we make the statistics
The Swedish Forest Agency's gross felling model
The Swedish Forest Agency uses a top-down model for calculating annual gross and net felling. The model is based on consumption statistics, import and export statistics, inventory statistics and data on felled but abandoned trees (left in the forest).
The consumption statistics consist of the forest industry's consumption of roundwood (sawlogs and pulpwood), consumption of roundwood for energy purposes (firewood) and other roundwood consumption. The consumption data are adjusted with data on imports and exports of roundwood and changes in stocks to calculate net felling.
Finally, the gross felling is calculated by adding the net felling with data on felled trees that have been felled but left in the forest during felling.
The Swedish Forest Agency also makes a forecast for gross felling for the current year during the second half of the year. The gross felling model is also used for this, but instead of actual wood consumption, it is based on expected consumption. This expected consumption is based on forecasts for the production of sawn timber and pulp in the forest industry and forecasts in the energy sector. Consumption figures are used to estimate how much wood is expected to be used for the forest industry's forecasted production.
Consumption of round wood in industry
Data on the consumption of roundwood in the forest industry comes from an annual survey conducted by Biometria (formerly SDC). Biometria compiles data on the consumption of round wood and other wood raw materials in the wood fibre industry and the wood mechanical industry. This refers to sawmills, pulp industries and wood panel industries. As of 2018, match manufacturers and pole industries are also included in the survey.
Data on the raw material consumption of small sawmills (which consume less than 2,000 m3fub of round timber per year) are based on the latest sawmill inventory and updated assessment by the trade association Småsågarnas Riksförbund.
Fuel wood
Since 2013, data on the extraction of wood for energy purposes has been obtained from the Swedish Energy Agency's survey "Production of unprocessed wood fuels". Data for previous years are based on data from Forest Impact Calculations 1999. From 2013 onwards, the information on fuel wood includes, in addition to firewood mainly used in detached houses, also wood consisting of stemwood chips.
Other wood
Other wood refers to wood other than timber, pulpwood and fuel wood. The information is based on an older assessment. As of 2018, roundwood for industrial production of piles, posts, veneers and matches made of hardwood or fibreboard is no longer included in the category.
Import and export of round wood
The Swedish Forest Agency uses data from Statistics Sweden's foreign trade statistics.
Stock change
The data comes from the Swedish Forest Agency's survey, Stocks of softwood sawlogs, pulpwood and pulp chips. The change in stock of softwood sawlogs and pulpwood is calculated as closing stock at the end of the year minus opening stock at the beginning of the year. Information on stocks of hardwood sawlogs is missing, which is why no adjustment is made for stock changes for hardwood sawlogs.
Felled whole trees left in the forest
Data on felled whole trees left behind are part of the gross felling and the statistics are retrieved from the Swedish National Forest Inventory. The annual volume of felled trees left in the forest is calculated based on the ratio between a five-year average of trees left behind according to the Swedish National Forest Inventory and a five-year average value of the Swedish Forest Agency's net felling.
Felling according to the Swedish National Forest Inventory
The Swedish National Forest Inventory's felling estimates are based on an inventory of a sample of Sweden's land area. As for felling data, these are estimated from stumps on sample plots that have been felled. The stump inventory involves the registration of all stumps from the last felling season with a diameter of at least 5 cm. The felling season is defined by the dates of bud cracking during the previous and current tax year. In addition to measuring stumps from felled trees, the type of felling and all other data registered on the Swedish National Forest Inventory's sample plots are stated.
Gross felling by owner class and county
To distribute the gross felling by owner class and county/region, data from the Swedish Forest Agency's action surveys are used. The data are reported as three-year averages
Reliability of statistics
In the gross felling model, the Swedish Forest Agency strives to use the best available input sources. Several different data sources are used, and the reliability of the statistics is largely linked to the uncertainties in these data sources. If incorrect values are used, it has a direct impact on the result in the gross felling model. Calculated gross felling volume can be both overestimated and underestimated. At present, there is only an average error estimate for felled whole trees and for parts of the firewood statistics. In addition to selection errors, there may also be systematic errors. It is therefore not possible to estimate the size of the total error in the calculations. Instead, here follows a description of various uncertainties that can be noted in the sources used.
Import and export of round wood
The gross felling calculation is sensitive to whether the statistics on roundwood trade are correct or not. Increased imports can reduce the need for domestic wood in industry and thus felling. Over the years, official foreign trade statistics from Statistics Sweden have mainly been used in the gross felling calculation. Statistics Sweden's data on imports and exports of roundwood are reported in the unit cubic metres (m3f). Based on a 2021 interview survey to those who provided information on the import and export of round wood to Statistics Sweden's unit, the unit in the calculation is consistently assumed to be cubic meters of solid measure under bark (m3fub).
Statistics Sweden's statistics for foreign trade within the EU are based on a sample survey in which companies are asked to report monthly how much they have imported and exported. Trading in companies that are not included in the survey is modelled by Statistics Sweden. Foreign trade with countries outside the EU is registered in mandatory customs declarations, which constitute input data to Statistics Sweden's statistics. As far as the Swedish Forest Agency is aware, Statistics Sweden does not conduct any extensive quality checks of reported values.
In Biometria's report (see below), import data is collected regarding the amount of round wood consumed in the wood fibre and wood mechanical industry that is imported. The data at Biometria is in m3fub.
For the year 2023, Biometria reports an import of round wood of 6.80 million m3fub, while Statistics Sweden reports a preliminary import of 7.26 million m3fub, i.e. a difference of about 0.4 million m3fub. However, the data are not fully comparable as changes in the stock of imported wood are not considered in Biometria's data. For the years 2017–2023, the differences between Biometria's and Statistics Sweden's total imports of round timber have varied from -0.5 to +0.5 million m3fub.
Stock of round timber
The Swedish Forest Agency conducts a stock survey with the aim of capturing stocks of softwood sawlogs, pulpwood and wood chips on four occasions a year. The survey covers all purchasing companies, pulp industries and major sawmills. Up to and including 2018, a sample of sawmill companies with a turnover of between SEK 5 and 50 million was examined. As of 2018, a comprehensive survey of these companies was also conducted. As of 2020, only companies with a turnover greater than SEK 20 million are included. In case of non-response, answers are imputed.
Biometria's survey for sawmills, pulp and board industry
Data from Biometria is based on a survey of sawmills (which consume more than 2,000 m3 of round timber per year) and of the pulp and board industry. As of 2018, the manufacture of poles and matches is also included in the survey.
As of 2020, the data for sawmills and other wood mechanical industries are based on administrative survey data from Biometria's timber accounting system (VIOL), while data for the pulp industry have been collected via questionnaires, as before.
From 2020 onwards, wood consumption data in sawmills and other wood mechanical industries refer to the measured volume. This includes volumes that were measured during the year but were in stock at sawmills at the end of the year. When calculating the gross felling, this is considered through a correction based on the Swedish Forest Agency's stock statistics for sawlogs.
Measured volumes of sawlogs may include logs, which can be delivered that the sawmill chooses to resell. Biometria adjusted for 317,000 m3fub of resold volumes in 2023. The volumes may have been larger, which in that case means an overestimation of wood consumption and gross felling.
Data on the raw material consumption of small sawmills (which have consumed less than 2,000 m3fub of sawlogs per year) are derived from the most recent sawmill inventory conducted in the early 2000s and an updated assessment by the trade association Småsågarnas riksförbund. In total, roundwood consumption in small sawmills is estimated to amount to 650,000 m3fub per year.
Fuel wood
The Swedish Energy Authority's study on unprocessed wood fuels consists of several different studies. A total survey of the Swedish Tree Fuel Association's member companies and a selection of other companies. In addition, data from the survey of the energy use of firewood in detached houses are included.
The survey has a relatively high response rate, and the responses are reviewed. Coverage errors and measurement errors may occur.
The Swedish Forest Agency recalculates the data presented to m3f ub with the conversion rates 2.04 MWh/m3f (stemwood chips), 2.05 MWh/m3f (tree part chips) and 2.25 MWh/m3f (firewood) and 0.88 m3fub/m3f. It is possible that some of what is reported in the Swedish Energy Agency's survey does not have so much bark, it disappears in the handling of the wood. The Swedish Forest Agency's recalculation may thus mean that we underestimate the volume of firewood. The conversion rates between energy measures (MWh) and volume measures (m3f) are also a source of uncertainty.
In the case of tree chips, the calculations assume that 50 percent of the total weight of a tree that is 15 cm in diameter is stemwood. This is a rough estimate as the proportion of stemwood is higher in larger trees and lower in smaller ones. Wood chips make up a low proportion (about 3 percent) of the total firewood felling, which is why the uncertainty has a marginal impact on the total felling of firewood.
The Swedish Forest Agency's investigations on measures in forestry
The survey on large-scale forestry is almost comprehensive and the answers are assumed to be of good quality. The distribution of the measure volumes per county is also considered to be reliable. The data for small-scale forestry are more uncertain. Up to and including 2017, the data were collected through interviews, and from 2019 onwards via a web-based survey. The collected data is of acceptable quality, but the survey consists of a random sample, which is why sampling and non-response errors together with measurement errors affect the reliability of the statistics.
At the county level, the estimation of most variables becomes too uncertain for a single year. Therefore, county-by-county data are reported as three-year moving averages. The relative mean error varies between counties, small counties generally have larger errors than large ones.
SLU National Forest Inventory's estimate of gross felling
The statistics in tables 03–06, 08 and 10 in the Swedish Forest Agency's statistical database come from the SLU (The Swedish University of Agricultural Sciences) National Forest Inventory. The Swedish National Forest Inventory's estimates of five-year averages for the gross felled volume for the whole of Sweden have an average error of about 4 percent. At certain time intervals, estimates are made of the statistical uncertainty. In addition to random sampling errors, there are systematic errors, among other things. caused by undiscovered stumps from felled trees. Studies have shown that the stump inventory underestimates the felled volume by about 7 percent, an error that is compensated for in the calculations. It has not been possible to prove any corresponding systematic errors in the estimates of felled area. Analysis of the control assessment carried out at the Swedish National Forest Inventory has found that the seasonal assessment and classification of felling measures are very accurate.
Since 2003, the model-calculated felling on four occasions (2006, 2007, 2014 and 2015) has been just within or slightly outside the Swedish National Forest Inventory's estimates when the margin of error has been considered. A 95% confidence interval does not mean that the true value is always within the range, just most of the time. For the most recent period (2018–2022), the model-calculated felling is below the estimates of the Swedish National Forest Inventory by 2.4% (2.3 million m3sk). The difference is within the margin of error for the Swedish National Forest Inventory's estimate. However, the periods are not fully comparable as the Swedish Forest Agency's model calculation refers to the calendar year and the Swedish National Forest Inventory's felling season.
In addition to differences caused by sampling errors and periodicity, there are differences caused by differences in methods and data sources. As previously described, the Swedish Forest Agency's calculation model is based on several different sources, each of which has known and unknown uncertainties. The advantage of the model calculation in contrast to the Swedish National Forest Inventory is that it has higher timeliness in that it provides an annual estimate, and that already the year after the reference year, and that it provides a forecast for the current year. The calculation model is also based on production, foreign trade, energy and inventory statistics, which provides good comparability and interoperability with several other statistical areas.
An advantage of the NFI's felling estimates is that the NFI has quality control of the entire statistical production chain, from field inventory to calculations, and that the survey has been carried out using the same method for a long time. It thus provides reliable statistics for the long-term development of felling and is comparable with other official statistics from the Swedish National Forest Inventory.
The estimates of the Swedish National Forest Inventory are important from a quality point of view for the annual model-calculated gross felling statistics, and over a longer period, the two estimates should follow each other.
Development work for better quality
In 2019 and 2020, the Swedish Forest Agency and the Swedish National Forest Inventory conducted a joint quality study of the two gross felling models. The study resulted in several conclusions that formed the basis for a development work to improve the quality of input data to the Swedish Forest Agency's gross felling model.
In 2021, as part of the development work, Statistics Sweden investigated whether import and export data on round wood refer to cubic metres under or on bark. Another development work that was completed in 2022 concerned the conversion rate between fixed measure under bark (m3fub) and forest cubic meters (m3sk). The revision of the conversion ratio meant that the previous conversion ratio of 1.2 between the units m3fub and m3sk was revised to 1.188. The statistics have been revised from year 2000 onwards to take account of the new conversion rate.
Average felled area for final fellings
The average area for final felling greater than or equal to 0.5 hectares is based on the Swedish Forest Agency's register for felling notifications and registered information on felling conducted that can be linked to these notifications. Felling conducted is registered within the Swedish Forest Agency's supervisory activities with support for remote sensing-based methods where satellite images taken at different times are compared to identify and delimit felling. The operations are comprehensive, but there are uncertainties, and these can partly be assessed by a comparison between the felled area according to this register and the SLU National Forest Inventory's estimates of the final felled area.
The estimates of the Swedish National Forest Inventory exceed the Swedish Forest Agency's register by 21 percent for the latest five-year average. This can be explained by the fact that the Swedish Forest Agency's register only covers fellings greater than or equal to 0.5 hectares. The estimates of the Swedish National Forest Inventory also include the area that is conducted in smaller fellings. But there are also other potential explanations for differences between these two statistical sources. These may be due to measurement technical differences in how the felled area is delimited, that fellings over 0.5 hectares have been missed in the Swedish Forest Agency's documentation for various reasons, and random variations in the Swedish National Forest Inventory's estimates. It has not been possible to clarify how much of the total difference consists of different partial explanations.
More documentation on the quality of the statistics can be found in the quality declaration for the statistics below (only in Swedish).
