Earlier estimation of developments in the CO2 storage capacity of forests: categorising the results of the German National Forest Inventory
© plainpicture / Alexander Schönberg
The change in the sink capacity of forests in Germany is mainly due to extreme conditions such as drought and storms, followed by tree damage caused by beetle infestation. As the huge loss of the forest sink in recent years only became known when the current fourth National Forest Inventory (BWI-4) was published, an essential information basis for policy decisions on climate protection for forests was missing. One of the most important factors in achieving Germany's climate policy targets is the role of forests as a natural carbon sink. Additional relevant policy decisions on climate protection will be made in the years ahead, which is why well-founded, reliable and timely assessments of forest development over the next five years are important.
"The climate protection function of forests is failing" and sink targets of -25 million tonnes of CO2-equivalents (Mt CO2-eq) in 2030 for the land use sector (LULUCF) "are no longer achievable". These or similar statements are being made following the publication of the BWI-4. In our view, these statements are not sufficiently nuanced with a view to forests:
- The loss of forest sinks in Germany is primarily a result of collapsed spruce stocks, which were mainly planted in unsuitable locations in the 1950s.
- The forest sink was created in the period from 2012 to 2017. It vastly decreased up to 2022. Forest modelling can help us understand how the sink will develop.
The role of spruce stocks
The National Forest Inventory is carried out in Germany every 10 years. The results of the BWI-4 for the period 2012 to 2022 were recently presented to the Federal Ministry of Food and Agriculture (BWI-4 brochure). Almost 80,000 survey points are now being used to record tree species, determine tree diameters and heights, measure deadwood and record other biodiversity-relevant structures such as woodpecker cavities. The time series of the national forest inventories – along with the interim inventories with 5-year increments – makes it possible to analyse the development of forest growing stock, habitat quality and carbon removals for climate protection.
The following periods are distinguished below: the period 2002-2012 relates to BWI-3; the period 2012-2017 relates to the interim inventory CI-2017 and the first half of BWI-4 (results on the changes relate to 2013 to 2017); the period 2017-2022 corresponds to the second half of BWI-4 (results on changes relate to 2018 to 2022). Data on BWI-3, CI-2017 and BWI-4 are freely available. This data was used to calculate results for the periods 2012-2017 and 2017-2022 since detailed data on the survey points for CI-2017 and BWI-4 are not available.
In particular, the years 2018 to 2020 saw very severe natural disasters. Prolonged drought and storm events, followed by bark beetle calamities, led to a decline in timber stocks. Three aspects can be observed here:
- The mortality rate for spruce rose to 9 to 12 percent in the 2017-2022 period. For other tree species, the increase in the mortality rate was significantly lower, particularly for oak and beech (see Annex 1).
- According to BWI-4, tree growth decreased by 14.9 percent in the 2012-2022 period. According to the interim inventory, only a slight decline in tree growth of 1.8 percent was observed in the 2012-2017 period. In the 2017-2022 period, a decline in tree growth of over 25 percent can therefore be expected, resulting from the poor conditions for growth in this period (see Annex 2).
- According to calculations carried out by Thünen Institute, timber production increased sharply overall between 2018 and 2021, reaching levels of over 80 million cubic metres per year. A significant share of this was fallen timber (Annex 3), which is why the share of regular harvesting activities was lower than in the years before the damage caused to forests. As a result, less wood was harvested on average in stocks without severe damage. The trees were able to continue to grow and increase their stock. This partially buffers the loss due to the increased mortality rates.
Figure 1 shows the annual change in growing stock in the forest for deciduous and coniferous trees. Values above zero mean that the growing stock has increased in the net balance and the forests comprise a carbon sink. Negative values indicate a decrease in growing stock or a source of CO2 (see also Annex 4).
Figure 1: Change in growing stock by tree species group, Source: Authors’ own graph with BWI-4 data, DT-HL = deciduous trees with a high longevity such as lime species; DT-LL = deciduous trees with a low longevity such as poplar species.
For deciduous trees, the stock build-up decreased slightly from the 2002-2012 period to the 2012-2017 period, despite an increase in deciduous tree stocks due to forest conversion. In the 2017-2022 period, the stocks of deciduous trees also increased and comprised a carbon sink. However, the growing stock build-up roughly halved. This was due to a combination of a decline in tree growth and a slightly increased mortality rate for deciduous trees.
The change in conifer stocks is strongly influenced by the negative development of spruce. Their stock already decreased in the 2002-2012 period, mainly due to windthrow caused by storms in the 2002-2007 period. In the 2012-2017 period, there were significantly fewer natural disturbances and the spruce stands showed a strong build-up. In the 2017-2022 period, very high stock levels were lost due to increased mortality and decline in spruce growth. Spruce stocks thus became a high source of CO2. The other conifers, like pine, were less affected by natural disturbances. However, their stock build-up was very low in the 2017-2022 period.
It should be noted that primarily the loss of spruce stocks led to the forests in Germany becoming a source of CO2 in the 2017-2022 period. Deciduous tree stands were still a carbon sink during this period – despite the extreme conditions.
Forest models can provide reliable forecasts
Only now, with the BWI-4, are reliable data for forests available for the 2017-2022 period. Under the United Nations Framework Convention on Climate Change (UNFCCC), Germany is obliged to report its national greenhouse gas emissions. The German government's national greenhouse gas inventory (NIR 2024) already contains statements on the emissions and carbon removals of forests in the 2017-2022 period.
Figure 2 shows the greenhouse gas balance of living biomass, i.e. living trees, up to 2022. Up to 2017, the GHG balance of living trees is based on data from forest inventories and timber production statistics. The results of the interim inventory (CI-2017), which covers the 2012-2017 period, were updated for 2018 to 2022. The increase in timber production was taken into account. Results of the BWI-4 were not yet available for reporting. Using this simple modelling approach, a sink capacity of -27.4 to -38.8 million tonnes of CO2-eq was reported for living trees in forests in Germany for 2018-2022. This extrapolation lacks assumptions on the increased mortality of trees and expected decreases in growth due to unfavourable growing conditions like drought.
In the FABio-Forest model, which was developed at Oeko-Institut, we adjusted the assumptions in the greenhouse gas inventory whitin a short study (dotted green line in Figure 2) and included annual reported mortality rates from the forest condition survey (dotted green line in Figure 2) and estimated values for the decrease in growth (solid green line in Figure 2). With these assumptions, we were able to show that the living trees in Germany’s forests in the 2017-2022 period became a source of 25.6 million tonnes of CO2-eq per year on average. In the extrapolation from 2022 onwards, we replicated the development in the “with measures scenario” of the German government's 2024 Projection Report (shown in Figure 1).
Figure 2: GHG balance of living trees in the forest, reported and modelled results, Sources: Hennenberg et al. (2024), NIR 2024 and "with measures scenario" (MMS-2024); ESRR = timber removal taken into account based on the felling recalculation by Thünen Institute; WZE-Mort = consideration of mortality from the Forest Condition Survey; corrZW = 50% decrease in growth was assumed for conifers and 25 % for deciduous trees for 2018 to 2022; gS = low natural disturbance comparable to the 2013-2017 period.
How well can the results of BWI-4 be estimated using FABio-Forest?
The BWI-4 brochure presents first results on the carbon balance of the forest area. The biomass of living trees – as shown in Figure 3 – is both above-ground and below-ground. The change in sink capacity amounted to -260 million tonnes of CO2-eq in the 2012-2017 period and +132 million tonnes of CO2-eq in the 2017-2022 period (conversion factor: Mt C * 44/12).
Figure 3: Carbon stock in Germany's forests and wood products, Source: BWI-4 brochure, p. 46 (translated from German to English). Living trees are the sum of above-ground and below-ground biomass.
Converted into annual emission levels, the forest sink amounted to -52.1 million tonnes of CO2-eq per year on average in the 2012-2017 period (Figure 4 and Table 1). The recalculations based on CI-2017 and BWI-4 therefore show an improvement in sink performance of around -6 million tonnes of CO2-eq per year in the 2012-2017 period. By way of comparison, the level previously reported in the NIR 2024 based solely on the CI-2017 was -45.8 million tonnes of CO-eq.
Figure 4 and Table 1 below show the GHG balance of living trees for the 2017-2022 period. On average, this was a source of 26.4 million tonnes of CO2-eq per year according to the BWI-4 brochure. In the NIR 2024, an average sink of -32.2 million tonnes of CO2-eq per year was reported for this period. This means that previous reporting overestimated the sink capacity of living trees in Germany by 58.6 million tonnes of CO2-eq per year.
Figure 4: Comparison of the GHG balance of living trees based on the results of the BWI-4, NIR 2024 and the FABio-Forest modelling, Sources: Authors’ own graph based on data from the BWI-4 brochure, NIR 2024, Hennenberg et al. (2024). With / w/o WZE mortality = with / w/o taking into account the mortality rate from the annual Forest Condition Survey; with / w/o corrZW = with/without correction of increases in 2018-2022 by a decrease of 50% for conifers and 25% for deciduous trees.
Table 1: Change in stock by tree species group, Sources: Authors’ own presentation based on data from the BWI-4 brochure, NIR 2024, Hennenberg et al. (2024).
Figure 4 and Table 1 also show the results of the FABio-Forest modelling. The comparison with the results of the BWI-4 and the values reported in the NIR 2024 shows that the results from FABio-Forest for the 2012-2017 period are close to the results of the NIR 2024. This was to be expected since only the results of CI-2017 were used to determine the parameters of the FABio-Forest model.
For the 2017-2022 period, FABio-Forest models an average source of 25.6 million tonnes of CO2-eq per year. This value is very close to the BWI-4 values.
The results of the NIR 2024 can also be well-simulated in FABio-Forest for the 2017-2022 period if the assumptions made there are used to determine the parameters.
Projection results with FABio-Forest provide high directional reliability
The comparison clearly shows that the GHG balance of BWI-4 could be projected significantly better with FABio-Forest than with the modelling approach used in the NIR 2024. This was possible because, in FABio-Forest, mortality data from the Forest Condition Survey was taken into account and assumptions were made about the decrease in increment.
However, first comparisons at the level of the tree species groups show that, in FABio-Forest, the change in forestry stocks among, for example, coniferous trees is modelled too high for pine and too low for spruce. One of the reasons for this is that the growth functions of the model are not yet able to take sufficient account of climate effects. In the ongoing DIFENs project funded by the Forest Climate Fund, a climate-sensitive control of increment is currently being implemented in FABio-Forest. The mortality of tree species could also improve the projection by means of climate-sensitive functions. Results from satellite data analyses like those in the FNEWs project can be incorporated as additional information. It should be noted that the current version of the FABio-Forest model can be used to project results with a high degree of directional certainty for the development of the carbon storage capacity of living trees in forests in Germany.
The future of forest (ensemble) modelling
The next interim inventory for some of the BWI survey points will take place in 2027. Results for the 2022-2027 period can be expected in 2029, i.e. one year before the important target year, 2030, for the LULUCF sector in the German Federal Climate Change Act and the EU's LULUCF Regulation. For the years ahead, reliable predictions are needed for forest development in the 2022-2027 period – before 2029. Model-based analyses with forest models such as FABio-Forest can be very helpful in this regard.
FABio-Forest is one of several forest models that can be used to model forest development in Germany. Thünen Institute uses the Matrix Model for forest modelling as part of the German Government’s Projection Report. As Figure 2 shows, the results in FABio-Forest and the Matrix Model are very comparable if the applied assumptions are harmonised. The strengths of these two empirical models are that they are very good at controlling the level of timber production and the intensity of forest management. There is another group of forest models which is process-based; these include the 4C, FORMIND and FORMIT-M models, which model tree growth as physiological processes. These models have the advantage of modelling the effects of climate change on forest development. It would be desirable to use several forest models to carry out ensemble modelling for different climate and timber production scenarios. The modellers would jointly coordinate the data used and the assumptions made in order to ensure that the model runs are as comparable as possible. This would facilitate the identification of the effects of the assumptions made and more reliable corridors of expected forest development in Germany.
FABio-Forest was used to project future forest development for three scenarios. The results of this are shown in Figure 5. The assumptions made range from low natural disturbances (comparable to the assumptions made in the German government’s Projection Report) to high natural disturbances. For the latter, it is assumed that extreme conditions like those in 2018 to 2021 occur on average every five years.
Figure 5: Sensitivities of GHG balance of living trees in Germany’s forests to low, medium and high natural disturbances, Sources: Hennenberg et al. (2024), NIR 2024. gS = low natural disturbances, mS = medium natural disturbances, hS = high natural disturbances.
Thus, the statement “the climate protection function of forests is failing” is an oversimplification. It is possible that forests in Germany will be a carbon sink in 2030. Only with an earlier and more reliable projection of the forest carbon sink capacity will it be possible to assess whether the forest sink – along with further emission reductions and sink enhancements from ongoing and planned measures in the LULUCF sector such as peatland protection, the creation of agroforestry areas and use of more wood for construction – will be sufficient to achieve the LULUCF sink target of -25 million tonnes of CO2-eq in 2030.
Dr Hannes Böttcher is a forest researcher and works on climate protection policy in the land use sector. He coordinates the Biogenic Resources and Land Use subdivision in the Energy & Climate Division in our Berlin office. With Dr Klaus Hennenberg, biologist and expert on forest modelling and sustainability assessments in Darmstadt, Böttcher developed the forest development model FABio-Forest. Dr Mirjam Pfeiffer is an expert on ecosystem dynamics and forest modelling and also works in our Darmstadt office. Judith Reise is a biologist who researches and advises on biodiversity and climate protection in terrestrial and marine ecosystems in our Berlin office.
This study benefited from funding for the projects DIFENs (Waldklimafonds FKZ: 2220WK32A4) and LANDMARC (European Union Horizon 2020 research and innovation programme, Grant Agreement No. 869367).
Further information
Model description of FABio-Forest in the "Reference scenario of wood utilisation and forest development in the UBA BioSINK project" by Oeko-Institut (in German)
According to the Forest Condition Survey (WZE), tree mortality has increased since 2018. For spruce in particular, the mortality rate due to abiotic causes (drought, storm) and biotic causes (calamities) rose very sharply, reaching values of 9% to 12% between 2018 and 2022. Pine and other conifers and deciduous trees likewise experienced increased mortality rates; these were lower than 3%. Moderate increases in mortality rates were recorded for beech and oak with values below 1%. For oak, there was no difference compared to historical values, though biotic causes have increased.
Figure 6: Mortality rate due to biotic and abiotic factors, Source: Authors’ own graphs based on data from the Forest Condition Survey (WZE).
In the BWI-4 period (2012-2022), the increment of trees decreased by an average of 14.9% across all tree species compared to the BWI-3 period (2002-2012). According to the Carbon Inventory (CI-2017, period 2012-2017), the decrease amounted to only 1.8 %. If it is assumed that this decrease results from the ageing of forest stocks, a decrease of approx. 3.6% would have been expected in the BWI-4. The decrease in increment of approx. 11% is likely to be related to the poor growing conditions in the period 2017 to 2022. The exact proportion compared to the CI-2017 is 25 % to 30%. However, this is a rough categorisation. A precise evaluation requires detailed data from the BWI-4 and the CI-2017 (which have not yet been published).
Table 2: Analyses of the increment of the tree species groups according to BWI-3, CI-2017 and BWI-4, Source: Authors’ own compilation based on data from the forest inventories. DT-HL = deciduous trees with high longevity; DT-LL = deciduous trees with low longevity. In the BWI-4, ash, maple, birch and alder are shown separately.
Figure 7: Development of timber production including the share of damaged timber in millions of cubic metres harvested from 1990 to 2021, Source: Reise et al (2024). Note: Drought as a cause of damage has only been recorded since 2020.
Table 3: Changes in stock according to tree species groups, Source: Authors’ own compilation based on forest inventory data. DT-HL = deciduous trees with high longevity; DT-LL = deciduous trees with low longevity.
