s41467-026-68668-x.pdf

Type: Document | Status: ready
← All sources

Since the dataset we used is largely focused on temperate forests, our analyses may be further improved after filling the gaps for boreal and southern Europe. Similarly, further information on stand age, forest size and connectivity could improve our ability to understand the links between forest carbon stocks and biodiversity temporal and spatial patterns. Concluding remarks Our work represents a step towards understanding the biodiversity- carbon stock relationships and dynamics in European forests, for which we did not find a univocal relation. Among the carbon pools, lying and standing deadwood were proven as much more relevant than living biomass in explaining the variation in European forest species richness. Aboveground living biomass carbon stock is highly relevant in global forest assessments in relation to the wide amount of available data; however, it cannot be a proper proxy for forest biodiversity. Regional and national forest inventories, as well as forest restoration initiatives (e.g., REDD+), should account for the deadwood carbon pool since it provides prominent support to several taxonomic groups. The achievement of biodiversity conservation targets may be jeo- pardized by climate change mitigation policies not accounting for the conservation of deadwood in forest ecosystems, as its extraction for fuel or wood products could compromise its support to species diversity. Methods Data We used a recently built database encompassing 12 European countries and resulting from the harmonization of 32 datasets containing both field-sampled multi-taxon biodiversity information (i.e., at least three groups of organisms among animals and either plants or fungi), and tree-level stand structure data. Sampling units (i.e., concretely delimited forest areas of known geographical coordinates) were clustered in forest sites (i.e., an environmentally homogeneous geographical area) 29, where forest is de fined as an area with a tree cover equal or greater than 40%. We analyzed a total of 140,146 observations of 3520 species in 7971 plots (see Supplementary Fig. 1). Most sampling units comprise two or more functional and taxonomic groups resulting in a total of 2615 spatially distinct multi-taxon plots across 99 sites (Fig. 5). We used data on six taxonomic and functional groups: saproxylic beetles, birds, epi- phytic and epixylic bryophytes and lichens, wood-inhabiting fungi and vascular plants. The sampling protocols used for each group used comparable approaches, with different efforts defined by the original researchers based on the site environmental and biological context 66:v a s c u l a r plants and wood-inhabiting fungi were sampled in areal plots or blocks of plots rising to an overall sampling area in the order of hundreds of square meters; epiphytic and epixylic lichens and bryophytes were sampled on all or part of the standing living trees and deadwood items in each plot; birds were sampled by point counts mostly during time frames from 5 to 20 minutes 67; saproxylic beetles were sampled through window-flight interception traps (1– 6 in each plot), in some cases also with emerging traps and Winkler extractors 68.S p e c i e s names and higher taxonomic information were checked automatically using the R-packages“taxize” 69 and, for vascular plants,“WorldFlora”70 and either corroborated by experts or checked against the GBIF database (https://www.gbif.org/). Data preparation A complete assessment of plot-level species richness is almost impracticable due to the multiple constraints of sampling activities as opposed to the vastness and complexity of biological diversity 71.A general rule is that, as the number of encountered individuals increa- ses, species richness exhibits a nonlinear increase 72.U n t i lac o o r d i - nated forest biodiversity monitoring network using standardized protocols is put in place 48, the use of harmonized data from studies using different sampling efforts requires the standardization of rich- ness data. Rarefaction operates by down-sampling data to achieve an equal sampling effort, i.e., the one of the smallest samples. A com- plementary approach is to extrapolate species richness to a larger sample size through an estimated asymptotic species richness 72,73.W e estimated the expected species richness for each site for each taxo- nomic group using the “estimateD” function within the R package 35/g113N 40/g113N 45/g113N 50/g113N 55/g113N 60/g113N 10/g113W 5/g113W 0/g113 5/g113E 10/g113E1 5 /g113E2 0 /g113E2 5 /g113E 30/g113E Number of taxa sampled 1 2 3 4 5 6 Number of sampling units 6 25 50 100 200 Fig. 5 | Distribution of the sampling sites in Europe. Gray areas are covered by forests with a tree cover greater than 40% according to the European Forest Institute Forest Map of Europe88. The number of taxonomic groups sampled at each site is represented by different shades of red, while the number of sampling units is indicated by the dot size. Source data are provided as a Source data file. Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 6 “iNEXT,” set for sample-based incidence data74. Finally, a comparable measure of species richness (scaled species richness) was computed as the ratio between observed species richness for each taxonomic group (i.e., the quantity of sampled species) in each speci fics a m p l i n gu n i t (alpha diversity) and the expected species richness within each specific site (gamma diversity). Notwithstanding the importance of incorpor- ating various measures of species diversity, we were not able to analyse species abundance due to the partial harmonization of this data within the database. Stand structure data were thoroughly checked and corrected for possible errors (e.g., related to measurement units). Missing data for standing tree heights, diameters and lengths of lying deadwood were estimated based on other available measurements for the specifict r e e or fragment (e.g., unit length/height, unit diameter and unit ID). This process also took into account additional measurements of living and dead elements available for the sampling unit, site, and forest cate- gory. We used the predictive mean matching method of the mice function (R-package“mice” 75) for these estimations. Overall, we mod- eled 25% of the standing trees’heights. Since some protocols included only one diameter measure for lying deadwood, 70 and 78% of the second (smaller) diameter of respectively, stumps and logs were imputed. Stump heights were derived for 11% of records, and log lengths for 25%. We used a set of equations to calculate tree/deadwood fragment’s volume and thus biomass and carbon stocks in tonnes per hectare for each sampling unit for the three aboveground carbon pools (standing living trees, standing deadwood, and lying dead- wood) (Fig. 6). For the living trees volume estimation, we applied two different equations depending on the tree species. We used a generalized allo- metric volume equation developed for multiple tree species at the continental scale 76 (originally presented as Eq. 2 in the reference publication and hereafter referred to as Eq.1). Species-specific β0 and β1, which are scaling coef ficients, were extracted from Tab. 4 of the same publication for species where this information was available. The coefficients were selected based on the associated climatic zone of each sampling unit, covering 75% of the standing trees. For the remaining species (i.e., 25% of standing trees) the generic cone volume formula was used (Eq. 2). V i refers to the volume of the stem of an individual tree (m3), dbh refers to the diameter at breast height (cm in Eq. 1, as reported in the reference publication, and m in Eq.2), hi refers to the height of the individual tree (m). V i = β0/C1 dbhβ1 ð1Þ V i = ðdbh 2 Þ 2 /C1 Π /C1 hi 3 ð2Þ The volume V i (m3) of deadwood fragments was calculated through the cone formula (Eq.2) or truncated cone formula (Eq.3)f o r Alive Deadwood Standing deadwood LogsSnags Stumps Truncated cone formula Living trees Lying deadwood Heights, Diameters, Lengths (m) Volume per unit (m³/ha) Biomass per unit (t/ha) Carbon stock per unit (t C/ha) Eq. 3 Cone formula Eq. 2 Volume * WBD Eq. 5 (Di Cosmo et al., 2013) Biomass * 0.485 Eq. 7 (Martin et al., 2021) Climate – Species related Eq.1, Eq. 2 (Muukkonen et al., 2007) Volume * WBD * BEF Eq. 4 (Penman et al., 2003, Kattge et al., 2020 Biomass * 0.48 Eq. 6 (Penman et al., 2003) Carbon stock per plot (t C/ha) Standing deadwood Lying deadwoodLiving trees Fig. 6 | Workflow of the carbon stocks assessment for each sampling unit. Equation 1 is obtained from literature76; Eqs. 2 and 3 refer to the formula of the volume of the cone and truncated cone respectively; Eqs.4 and 577 use wood basic density (WBD) derived from existing literature77,79 or, when not available in the former paper, from TRY database78 and different biomass expansion factors (BEF) based on the climatic conditions of the sampling unit (boreal or temperate) and the dominant species group (coniferous or broadleaf); Eq.780 uses an improved carbon fraction of 0.485 for deadwood. Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 7 standing and lying deadwood, respectively. Dbh is the diameter at breast height (m), hi refers to the height of the individual dead tree (m), r is the radius at the small end of the fragment (m), obtained from the minimum diameter reported in the database for the associated deadwood unit, andR i st h er a d i u sa tt h el a r g ee n do ft h ef r a g m e n t( m ) , obtained from the maximum diameter reported in the database for the associated deadwood unit. V i = ðr2 i + R2 i + Ri /C1 riÞ/C1 Π /C1 hi 3 ð3Þ We then calculated the volume per hectare (m³/ha) of each unit by multiplying its volume (m3)b y1 0 , 0 0 0( m2, the area of one hectare) and dividing it by the sampling unit size (m2). The biomass of each living tree Bi (t/ha) was estimated by multi- plying its volumeV i (m3/ha), by the corresponding wood basic density (WBD) and biomass expansion factor (BEF), following Eq. 4.T h i s approach is derived from Eq.3.2.3 of the IPCC good practice guidance for land use, land-use change, and forestry 77. WBD values were obtained from IPCC standards 77,s p e c ifically Tab. 3 A.1.9– 1, for avail- able species. When species-specific values were unavailable, WBD was obtained from the TRY database 78, using genus or species-level esti- mates depending on data availability. BEF accounts for the additional biomass components not included in stem volume measurements, such as branches and foliage, allowing for a more accurate estimation of the total tree biomass. Different BEF values were applied to broad- leaf and coniferous species, depending on the climatic zone of the sampling unit (boreal or temperate), as specified in Tab. 3 A.1.10 of the same guidelines 77. Bi = V i /C1 WB D /C1 BEF ð4Þ The biomass of each deadwood unit Bi (t/ha) was estimated by multiplying its volume V i (m³/ha) by the corresponding wood basic density, WBD, following Eq.5. This approach is derived from the same reference for living trees77. WBD values were obtained from Tab. 4 of the reference publication79 which provides density values (t/m³) based on decay stage, deadwood type (stump or log), and group (conifer or broadleaf). Bi = V i /C1 WB D ð5Þ The carbon stock Ci (t C/ha) of each living tree was estimated by multiplying its biomass Bi (t/ha) by the carbon fraction of dry matter (0.48) as in Eq. 6, following IPCC standards77. Ci = Bi /C1 0:48 ð6Þ The carbon stock associated with each deadwood unitCi (t C/ha) was obtained by multiplying the unit biomass Bi (t/ha) by an average carbon fraction of dry matter factor as in Eq. 7, following reference literature80. Ci = Bi /C1 0:485 ð7Þ Finally, the plot-level stock for each carbon pool (standing living trees, standing deadwood and lying deadwood) was obtained by summing the carbon of each tree/fragment (t C/ha) in each sampling unit (Eq. 8). In total, plot-level stock data were assessed for 2445 plots for standing living trees, 1458 plots for standing deadwood, and 1654 plots for lying deadwood. C tot = Xn i ðCiÞ ð8Þ The three carbon pools did not show clear patterns of covariation. With the exception of unmanaged forests, mostly displaying high amounts of carbon across the three pools, a great variation of the three carbon stocks was evident for different forest categories and man- agement regimes (Figs. 7 and 8). Statistical analyses The carbon stock for the standing living trees, standing deadwood and lying deadwood carbon pools in each plot were logarithm base 10 transformed to achieve a normal distribution. We ran six BRTs to assess the relationship between the scaled species richness of each Silvicultural System UNM SEL SHW CWS RCL SCL Fig. 7 | Three-dimensional scatter plot of the relationship between above- ground carbon stocks across different silvicultural systems.Each point repre- sents a sampling unit, colored according to the silvicultural system: unmanaged (UNM), selection cutting (SEL), shelterwood (SHW), coppice with standards (CWS), retention clearcutting (RCL), simple clearcutting (SCL). Carbon stocks are logarithm base 10 transformed for better visualization and comparison with the partial dependence plots of the result section. An interactive version of the plot is available at https://zenodo.org/records/15096050. Source data are provided as a Source data file. Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 8 taxonomic group as response variable and carbon stocks, along with other selected predictors at the plot level (Table 2). The selection of predictors was based on their ecological rele- vance for the diversity of the selected taxonomic groups, as well as their availability and consistency across the entire dataset. The forest category81 was included to account for: broad-scale climatic condi- tions, biogeographical patterns, and effects related to canopy features. Forest categories are defined by dominant tree species, which play a significant role in shaping community patterns by in fluencing struc- tural attributes, light availability and soil characteristics 82,l e a d i n gt oa different spatial and temporal resource distribution for several taxo- nomic groups. The sampling units in our database encompassed a wide range of forest categories, including boreal forests (1), hemi- boreal forests and nemoral coniferous and mixed broadleaved- coniferous forests (2), alpine coniferous forests (3), acidophilous oak and oak-birch forests (4), mesophytic deciduous forests (5), beech forests (6), mountainous beech forests (7), thermophilous deciduous forests (8), broadleaved evergreen forests (9), mire and swamp forest (11), non-riverine alder, birch, or aspen forests (13), and plantations and self-sown exotic forests (14) 81. To better capture broad-scale climatic conditions, elevation (in meters) was included as a predictor variable in the model. Elevation is a key driver of temperature, precipitation, and soil properties, which shape species composition and biodiversity Table 2 | Variables used in the boosted regression trees. Name, type, role in the model and source (BRTs) Variable Type Selection References Scaled species richness Continuous Response 74 C stock standing alive Continuous Predictor From raw data C stock standing dead Continuous Predictor From raw data C stock lying dead Continuous Predictor From raw data Elevation Continuous Predictor From raw data Forest category Categorical Predictor 81 Silvicultural system Categorical Predictor 84 Protocol Categorical Predictor 66 1 2 3 4 5 6 7 11 14 Very Low Low Average High Very High Very Low Low Average High Very High Very Low Low Average High Very High Forest category C stock stand alive C stock stand dead C stock lying dead Silvicultural system UNM SEL SHW CWS RCL SCL Fig. 8 | Alluvial plot of the distribution of sampling units across forest cate- gories of aboveground carbon stocks.Each color represents a silvicultural sys- tem: CWS coppice with standards, RCL retention clearcutting, SEL selection cutting, SHW shelterwood, SCL simple clearcutting, and UNM unmanaged. Forest categories are numbered as follows: 1 = boreal forest, 2 = hemiboreal forest and nemoral coniferous and mixed broadleaved-coniferous forest, 3 = alpine con- iferous forest, 4 = acidophilous oak and oak-birch forest, 5 = mesophytic deciduous forest; 6 = beech forest; 7 = mountainous beech forest, 11 = mire and swamp forest, and plantations and self-sown exotic forest (14). Carbon stocks were categorized into five levels (very low to very high) using quantile-based classification, where each category represents 20% of the data distribution. Sampling units with missing values for one or more carbon stocks were excluded from the graph to improve clarity. Source data are provided as a Source data file. Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 9 patterns83. The decision to include elevation rather than broad-scale climatic variables was based on exploratory analyses showing that, within our dataset, mean annual temperature (MAT) and mean annual precipitation (MAP) were definitely homogeneous within most forest categories (interquartile range for MAT and MAP was 2.11 °C and 292 mm, respectively). Only mountain forests, i.e., alpine coniferous forests (category 3) and mountainous beech forests (category 7), showed a substantially higher variability, up to 3.83 °C and 664 mm. Based on these patterns, we deemed elevation as an ecologically meaningful predictor able to capture residual climatic variation in the database that is not accounted for by forest categories. The silvi- cultural system for each sampling unit was included as a categorical predictor. This variable refers to the method by which trees in a forest are harvested and replaced through regeneration or planting, resulting in distinct forest stands for both structural and compositional char- acteristics. Different silvicultural practices have varying effects on forest structure, leading to different resource and habitat availability for several taxonomic groups 15,20,35,67.T h ec l a s s ification used in this study was based on a harmonization effort on the same database by Trentanovi et al. 2023 84. We included the following terms referring to the “form of treatment” (Tab. 2), explicitly CWS = coppice with stan- dards, RCL = retention clearcutting, SEL= selection cutting, SHW = shelterwood, SCL = simple clearcutting, and UNM = unmanaged. The sampling protocol was included as a categorical predictor to account for variations in data collection methodologies across different countries and forest types, as differences in sampling methods could influence diversity estimates, potentially introducing inconsistencies across the database. For a detailed overview of the different protocols, see Tab. 2 in Burrascano et al. 29. The distribution of sampling units across forest categories, silvicultural systems and carbon stocks highlights the wide range of forest structures and management approaches represented in the database, with a certain in fluence of management strategies on carbon storage dynamics (Fig.8). BRTs effectively model complex non-linear relationships and enable reliable identi fication of relevant variables and interactions, with a strong predictive performance and robustness against over- fitting, missing data, and collinearity 85. The model setting involves finding the optimal combination of four main parameters (“number of trees”, “learning rate”, “tree complexity” and “bag fraction”)t om i n i - mize predictive error: the number of trees refers to the total count of individual decision trees, influencing both predictive performance and model stability; the learning rate controls the contribution of each tree to the final model, providing a balance between accuracy and the risk of overfitting; tree complexity determines the maximum depth of each tree, affecting the model’s ability to capture complex patterns in the data; and the bag fraction specifies the proportion of the training set randomly selected for each iteration 85.W eu s e dt h e“gbm.step” func- tion from the “dismo” package86 to test various BRT settings by com- bining different learning rates (0.001, 0.0025, and 0.005), tree complexity values (2, 3, and 5), and bag fraction values (0.5, 0.75), and assess the relative importance of each explanatory variable, calculated as the total reduction in model deviance attributable to that variable, averaged across all trees and rescaled to a total of 100% 85.W eu s e dt h e Gaussian error distribution. We selected parameter combinations that resulted in models with over 1000 trees and evaluated their performance using three metrics, following the approach of Napoleone and colleagues 87:( i )e x p l a i n e d deviance (D²), calculated asD²= 1 −(residual deviance/total deviance). This metric ranges from zero to one, where a value of one indicates a perfectfit, and is considered a generalization of the traditionalR²u s e d in regression analysis 87; (ii) the cross-validated mean correlation coefficient, which measures the average correlation between the training and test datasets; and (iii) self-statistics, which assesses the correlation between predicted and observed values. These correlation metrics are commonly used to evaluate the efficiency of BRTs 22,87. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Data availability The data used for this study are available upon request using the data explorer within the Cost Action website (https://www.bottoms-up.eu/ en/results/data-explorer.html). Source data are provided with this paper. Code availability T h ec o d eu s e di nt h i ss t u d yi sa v a i l a b l ea tC o d eO c e a n :https://doi.org/ 10.24433/CO.9463772.v2. References

  1. Köhl, M., Linser, S. & Prins, K. State of Europe’s Forests 2020.https:// doi.org/10.13140/RG.2.2.12881.76643(2020).
  2. Sanchez Lopez, J. et al. Brief on the role of the forest-based bioec- onomy in mitigating climate change through carbon storage and material substitutionhttps://publications.jrc.ec.europa.eu/ repository/handle/JRC124374(2021).
  3. Profft, I., Mund, M., Weber, G.-E., W e l l e r ,E .&S c h u l z e ,E . - D .F o r e s t management and carbon sequestration in wood products.Eur. J. For. Res. 128,3 9 9–413 (2009).
  4. European Commission. Eurostat Wood products-production and trade. Available at: https://ec.europa.eu/eurostat/statistics- explained/index.php?title=Wood_products_-production_and trade (2024).
  5. FAO. Global Forest Resources Assessment 2020: Main Report https://doi.org/10.4060/ca9825en(FAO, 2020).
  6. European Commission, COMMUNICATION FROM THE COMMIS- SION TO THE EUROPEAN PARLIAMENT, THE COUNCIL, THE EUR- OPEAN ECONOMIC AND SOCIAL COMMITTEE AND THE COMMITTEE OF THE REGIONS. New EU Forest Strategy for 2030 COM/2021/572final https://eur-lex.europa.eu/legal-content/EN/ TXT/?uri=celex:52021DC0572(2021).
  7. European Environment Agency. State of Nature in the EU: Results from Reporting under the Nature Directives 2013 2018. EEA Report No 10/2020https://doi.org/10.2800/088178(Publications Office of the European Union, LU, 2020).
  8. Betts, M. G. et al. Producing wood at least cost to biodiversity: integrating triad and sharing–sparing approaches to inform forest landscape management.Biol. Rev. 96,1 3 0 1–1317 (2021).
  9. Paquette, A. & Messier, C. The role of plantations in managing the world’s forests in the Anthropocene. Front. Ecol. Environ.8, 27–34 (2010).
  10. Franklin, J. F. & Lindenmayer, D. B. Importance of matrix habitats in maintaining biological diversity.P r o c .N a t l .A c a d .S c i .U S A106, 349–350 (2009).
  11. Nagel, T. A. et al. Can triad forestry reconcile Europe ’s biodiversity and forestry strategies? A critical evaluation of forest zoning.Ambio 54,6 3 2–641 (2025).
  12. Nordén, B. & Appelqvist, T. Conceptual problems of ecological continuity and its bioindicators.Biodivers. Conserv.10, 779–791 (2001).
  13. Nordén, B. et al. Effects of ecological continuity on species richness and composition in forests and woodlands: a review.Ecoscience 21, 34–45 (2014).
  14. Dormann, C. F. et al. Plant species richness increases with light availability, but not variability, intemperate forests understorey. BMC Ecol. 20,4 3( 2 0 2 0 ) .
  15. Burton, J. I., Ares, A., Olson, D. H. & Puettmann, K. J. Management trade-off between aboveground carbon storage and understory plant species richness in temperate forests.Ecol. Appl. 23, 1297–1310 (2013). Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 10
  16. Hilmers, T. et al. Biodiversity along temperate forest succession. J. Appl. Ecol. 55,2 7 5 6–2766 (2018).
  17. Lassauce, A., Paillet, Y., Jactel, H. & Bouget, C. Deadwood as a surrogate for forest biodiversity: meta-analysis of correlations between deadwood volume and species richness of saproxylic organisms.Ecol. Indic. 11,1 0 2 7–1039 (2011).
  18. Parajuli, R. & Markwith, S. H. Quantity is foremost but quality matters: a global meta-analysis of correlations of dead wood volume and biodiversity in forest ecosystems. Biol. Conserv. 283, 110100 (2023).
  19. Parisi, F. et al. Linking deadwood traits with saproxylic invertebrates and fungi in European forests-a review.iForest Biogeosci. For.11, 423 (2018).
  20. Asbeck, T. et al. Biodiversit y response to forest management intensity, carbon stocks and net primary production in temperate montane forests.Sci. Rep. 11, 1625 (2021).
  21. Anthony, M. A. et al. Fungal communi ty composition predicts forest carbon storage at a continental scale.Nat. Commun. 15, 2385 (2024).
  22. Sabatini, F. M. et al. Trade-offs between carbon stocks and biodi- versity in European temperate forests.Glob. Change Biol. 25, 536–548 (2019).
  23. Bernal, B., Murray, L. & Pearson, T. Global carbon dioxide removal rates from forest landscape restoration activities.Carbon Balance Manag. 13,2 2( 2 0 1 8 ) .
  24. da Silva, L. P. et al. Natural woodlands hold more diverse, abundant, and unique biota than novel anthropogenic forests: a multi-group assessment.Eur. J. For. Res. 138, 461–472 (2019).
  25. Felton, A., Lindbladh, M., Brunet, J. & Fritz, Ö. Replacing coniferous monocultures with mixed-speciesproduction stands: an assess- ment of the potential benefits for forest biodiversity in northern Europe. For. Ecol. Manag. 260,9 3 9–947 (2010).
  26. Goded, S. et al. Effects of eucalyptus plantations on avian and herb species richness and composition in north-west Spain.Glob. Ecol. Conserv. 19, e00690 (2019).
  27. González-Montelongo, C. & Pérez-Vargas, I. Looking for a home: exploring the potential of epiphytic lichens to colonize tree plan- tations in a macaronesian laurel forest.For. Ecol. Manag. 453, 117541 (2019).
  28. Brancalion, P. H. S., Hua, F., Joyce, F. H., Antonelli, A. & Holl, K. D. Moving biodiversity from an afterthought to a key outcome of forest restoration.Nat. Rev. Biodivers.1 –14 https://doi.org/10.1038/ s44358-025-00032-1(2025).
  29. Burrascano, S. et al. Where are we now with European forest multi- taxon biodiversity and where can we head to? Biol. Conserv. 284, 110176 (2023).
  30. Pan, Y., Birdsey, R. A., Phillips, O. L. & Jackson, R. B. The structure, distribution, and biomass of the world’sf o r e s t s .Annu. Rev. Ecol. Evol. Syst. 44,5 9 3–622 (2013).
  31. Kebrle, D., Hofmeister, J., Kodet, V. & Hošek, J. Key structural factors and their thresholds for promoting bird diversity in spruce- dominated production forests of central Europe.For. Ecol. Manag. 550, 121522 (2023).
  32. Stokland, J. N. The saproxylic food web. in Biodiversity in Dead Wood (eds Jonsson, B. G., Stokland, J. N. & Siitonen, J.) 29–57 (Cambridge University Press, 2012). 3 3 . J a c o b s ,J .M . ,S p e n c e ,J .R .&L a n g o r ,D .W .I nfluence of boreal forest succession and dead wood qualities on saproxylic beetles.Agric. For. Entomol. 9,3 –16 (2007).
  33. Ranius, T., Eliasson, P. & Johansson, P. Large-scale occurrence patterns of red-listed lichens and fungi on old oaks are influenced both by current and historical habitat density.Biodivers. Conserv. 17,2 3 7 1–2381 (2008).
  34. Müller, J. et al. Effects of fores t management on bryophyte species richness in central European forests.For. Ecol. Manag. 432, 850–859 (2019).
  35. Prescott, C. E. The in fluence of the forest canopy on nutrient cycling. Tree Physiol. 22,1 1 9 3–1200 (2002).
  36. Luyssaert, S. et al. Old-growth forests as global carbon sinks.Nature 455,2 1 3–215 (2008).
  37. Paillet, Y. et al. Biodiversity differences between managed and unmanaged forests: meta-analysis of species richness in Europe. Conserv. Biol. 24,1 0 1–112 (2010).
  38. Keeton, W. S., Whitman, A. A., McGee, G. C. & Goodale, C. L. Late- successional biomass development in northern hardwood-conifer forests of the northeastern United States.For. Sci. 57, 489–505 (2011).
  39. Naudts, K. et al. Europe ’s forest management did not mitigate cli- mate warming. Science 351,5 9 7–600 (2016).
  40. Persiani, A. M. et al. Stand structure and deadwood amount in flu- ences saproxylic fungal biodiversity in mediterranean mountain unmanaged forests.iForest 9, 115–124 (2016).
  41. Atrena, A. et al. Quality of substrate and forest structure determine macrofungal richness along a gradient of management intensity in beech forests. Ecol. Manag. 478, 118512 (2020).
  42. Doer fle r ,I . ,G o s s n e r ,M .M . ,M ü l l e r ,J . ,S e i b o l d ,S .&W e i s s e r ,W .W . Deadwood enrichment combining integrative and segregative conservation elements enhances biodiversity of multiple taxa in managed forests.Biol. Conserv. 228,7 0–78 (2018).
  43. Bujoczek, L., Bujoczek, M. & Zi ęba, S. How much, why and where? Deadwood in forest ecosystems: the case of Poland.Ecol. Indic.121, 107027 (2021).
  44. Burrascano, S. et al. It’s a long way to the top: plant species diversity in the transition from managed to old-growth forests.J. Veg. Sci.29, 98–109 (2018).
  45. Hofmeister, J., Ho še k ,J . ,H o l á ,E .&N o v o z á m s k á ,E .D e c l i n ei n bryophyte diversity in predominant types of central European managed forests.Biodivers. Conserv.24,1 3 9 1–1402 (2015).
  46. Hodgetts, N. et al. An annotated checklist of bryophytes of Europe, Macaronesia and Cyprus.J. Bryol. 42,1 –116 (2020).
  47. Burrascano, S. et al. Towards an effective in-situ biodiversity assessment in European forests.Basic Appl. Ecol. 84, 121–132 (2025).
  48. Bouget, C., Larrieu, L., Nusillard, B. & Parmain, G. In search of the best local habitat drivers for saproxylic beetle diversity in temperate deciduous forests. Biodivers. Conserv. 22, 2111–2130 (2013).
  49. Kiraly, I., Nascimbene, J., Tinya, F. & Odor, P. Factors in fluencing epiphytic bryophyte and lichen species richness at different spatial scales in managed temperate forests.Biodivers. Conserv.22, 209–223 (2013).
  50. Boch, S., Prati, D., Hessenmoeller, D., Schulze, E.-D. & Fischer, M. Richness of lichen species, especially of threatened ones, is pro- moted by management methods furthering stand continuity.PLoS ONE 8, e55461 (2013). 5 2 . D e t t k i ,H . ,K l i n t b e r g ,P .&E s s e e n ,P .A .A r ee p i p h y t i cl i c h e n si ny o u n g forests limited by local dispersal?Ecoscience7,3 1 7–325 (2000).
  51. Nascimbene, J., Marini, L. & Odor, P. Drivers of lichen species richness at multiple spatial scales in temperate forests.Plant Ecol. Divers. 5,3 5 5 –363 (2012).
  52. Medina, N. G. et al. Species ric hness of epiphytic bryophytes: dri- vers across scales on the edge of the mediterranean.Ecography37, 80–93 (2014).
  53. European Commission. Regulation (EU) 2024/1991 of the European Parliament and of the Council of 24 June 2024 on nature restoration and amending Regulation (EU) 2022/869 (Text with EEA relevance) Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 11 PE/74/2023/REV/1http://data.europa.eu/eli/reg/2024/1991/oj (Publications Office of the European Union, 2024).
  54. Larsen, J. et al. Closer-to-Nature Forest Management. From Science to Policy 12 https://doi.org/10.36333/fs12(2022).
  55. Rybár, J. & Bo šeľa, M. Trade-offs or complementarity between bio- mass production and biodiversity in European forests: a review. Cent. Eur. For. J. 69,2 0 1–213 (2023).
  56. European Commission. Proposal for a REGULATION OF THE EUR- OPEAN PARLIAMENT AND OF THE COUNCIL amending Regulations (EU) 2018/841 as regards the scope, simplifying the compliance rules, setting out the targets of the Member States for 2030 and committing to the collective achievement of climate neutrality by 2035 in the land use, forestry and agriculture sector, and (EU) 2018/ 1999 as regards improvement in monitoring, reporting, tracking of progress and review COM/2021/554final. https://eur-lex.europa. eu/legal-content/EN/TXT/?uri=celex:52021PC0554(Publications Office of the European Union, 2021).
  57. European Commission. Proposal for a REGULATION OF THE EUR- OPEAN PARLIAMENT AND OF THE COUNCIL on a monitoring fra- mework for resilient European forests COM/2023/728final https:// eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52023PC0728 (Publications Office of the European Union, 2023).
  58. Giuntoli, J. et al. The quest for sustainable forest bioenergy: win-win solutions for climate and biodiversity.Renew. Sustain. Energy Rev. 159, 112180 (2022).
  59. Camia, A. et al. The use of woody biomass for energy production in the EU. https://data.europa.eu/doi/10.2760/831621, JRC122719 (Publications Office of the European Union, 2020). 6 2 . V e r k e r k ,P .J .e ta l .Forest-Based Climate Change Mitigation and Adaptation in Europe. https://efi.int/publications-bank/forest- based-climate-change-mitigation-and-adaptation-europe(2022) https://doi.org/10.36333/fs14.
  60. Puletti, N., Giannetti, F., Chirici, G. & Canullo, R. Deadwood dis- tribution in European forests.J. Maps 13,7 3 3–736 (2017).
  61. Rondeux, J. & Sanchez, C. Review of indicators and field methods for monitoring biodiversity within national forest inventories. Core variable: deadwood. Environ. Monit. Assess. 164,6 1 7–630 (2010).
  62. Ranius, T. et al. The effects of logging residue extraction for energy on ecosystem services and biodiversity: a synthesis.J. Environ. Manag. 209,4 0 9–425 (2018).
  63. Burrascano, S. et al. Handbook of fie l ds a m p l i n gf o rm u l t i - t a x o n biodiversity studies in European forests.Ecol. Indic. 132, 108266 (2021).
  64. Bouvet, A. et al. Effects of forest structure, management and landscape on bird and bat communities.Environ. Conserv.43, 148–160 (2016).
  65. Janssen, P. et al. Are biodiversity patterns of saproxylic beetles shaped by habitat limitation or dispersal limitation? A case study in unfragmented montane forests.Biodivers. Conserv.25, 1167–1185 (2016).
  66. Chamberlain, S. A. & Szöcs, E. Taxize: taxonomic search and retrieval in R. F1000Res 2,1 9 1( 2 0 1 3 ) .
  67. Kindt, R. WorldFlora: an R package for exact and fuzzy matching of plant names against the World Flora Online taxonomic backbone data. Appl. Plant Sci. 8, e11388 (2020).
  68. Pärtel, M., Szava-Kovats, R. & Zobel, M. Dark diversity: shedding light on absent species. Trends Ecol. Evol. 26, 124–128 (2011). 7 2 . C o l w e l l ,R .K .e ta l .M o d e l sa n de s timators linking individual-based and sample-based rarefaction, extrapolation and comparison of assemblages.J. Plant Ecol. 5,3 –21 (2012).
  69. Chao, A. & Jost, L. Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size.Ecology 93,2 5 3 3–2547 (2012). 7 4 . H s i e h ,T .C . ,M a ,K .H .&C h a o ,A .i N E X T :a nRp a c k a g ef o rr a r e f a c t i o n and extrapolation of species diversity (hill numbers).Methods Ecol. Evol. 7,1 4 5 1–1456 (2016).
  70. Buuren, van, S., Groothuis-Oudshoorn & mice:, K. Multivariate imputation by chained equations in R.J. Stat. Softw.45,1 –67 (2011).
  71. Muukkonen, P. Generalized all ometric volume and biomass equa- tions for some tree species in Europe. Eur. J. For. Res 126, 157–166 (2007).
  72. Penman, J. et al. Intergovernmental panel on climate change: Good practice guidance for land use, land-use change and forestry(Insti- tute for Global Environmental Strategies for the IPCC, 2003).
  73. Kattge, J. et al. TRY plant trait database –enhanced coverage and open access. Glob. Change Biol. 26, 119–188 (2020).
  74. Di Cosmo, L., Gasparini, P., Paletto, A. & Nocetti, M. Deadwood basic density values for national-levelcarbon stock estimates in Italy.For. Ecol. Manag. 295,5 1–58 (2013).
  75. Martin, A. R., Domke, G. M., Doraisami, M. & Thomas, S. C. Carbon fractions in the world ’s dead wood. Nat. Commun. 12, 889 (2021).
  76. EEA. European forest types — European Environment Agency https://www.eea.europa.eu/publications/technical_report_2006_ 9 (2006).
  77. Barbier, S., Gosselin, F. & Balandier, P. In fluence of tree species on understory vegetation diversity and mechanisms involved— ac r i - tical review for temperate and boreal forests.For. Ecol. Manag.254, 1–15 (2008). 8 3 . D a n i ,R .S . ,D i v a k a r ,P .K .&B a n i y a ,C .B .D i v e r s i t ya n dc o m p o s i t i o no f plants species along elevational gradient: research trends.Bio- divers. Conserv.32, 2961–2980 (2023).
  78. Trentanovi, G. et al. Words apart: standardizing forestry terms and definitions across European biodiversity studies.For. Ecosyst. 10, 100128 (2023).
  79. Elith, J. & Leathwick, J. Boosted Regression Trees for Ecological Modeling. R documentation. Software (2017). 8 6 . H i j m a n s ,R .J . ,P h i l l i p s ,S . ,L e a t h w i c k ,J . ,E l i t h ,J .&H i j m a n s ,M .R .J . Package ‘dismo’version 1.3-8 https://CRAN.R-project.org/ package=dismo(2022).
  80. Napoleone, F., Giarrizzo, E. & Bu rrascano, S. Habitat conservation state and plant diversity respond to different drivers in semi-natural grasslands.J. Veg. Sci. 32,e 1 3 0 5 5( 2 0 2 1 ) .
  81. Kempeneers, P., Sedano, F., Seebach, L., Strobl, P. & San-Miguel- Ayanz, J. Data fusion of different spatial resolution remote sensing images applied to forest-type mapping.IEEE Trans. Geosci. Remote Sens. 49, 4977–4986 (2011). Acknowledgements This article was funded by the EU Framework Programme Horizon 2020 t h r o u g ht h eC O S TA s s o c i a t i o n(www.cost.eu): COST action CA18207: BOTTOMS-UP–biodiversity of temperate forest taxa orienting manage- ment sustainability by unifying perspectives. The authors are thankful to all those experts and stakeholders contributing to the data collection. Part of the data for the Czech Republic were collected thanks to the funding of the Czech Science Foundation (grant GACR 22–31322S). The author Sebastian Kepfer Rojas was supported by Novo Nordisk Foun- dation grant NNF20OC0059948. The experimental site OPTMix in Centre-Val de Loire, France, was installed and equipped by INRAE EFNO thanks to the support of the region, the French National Forest Office, and belongs to the network ANAEE-F. The author Elena Haeler was supported by Program ‘Pilotprojekt zur Förderung der ökologischen Infrastruktur in Pärken’of the Federal Office for the Environment (FOEN) and the Wildnispark Zurich Foundation. Author contributions L.B. and S. Bu contributed to the data preparation, analysis and writing of the introduction, methods, results and the discussion of the manuscript. E. Hae, Y.P., and N.K. contributed greatly to the structuring of the paper. Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 12 F.N. and E.A. contributed to the methodological section. S. Bo and T.L. contributed to the discussion. F.C. worked on the harmonization of the database we used for this research. C.A., F.A., C.B., G.B., A.C., P.D.S., I.D., D.D., M.F., P.G., M.G., J.H.-C., E. Hol, J. Hof, J. Hos, I.G.M., P.J., K.J., S.K.R., D.K., T.K., J. Ma, A.M., M.M., J. Mu, B.N., P.O., Z.P., P.S., T.S., K.S., M.S., A.T., F.T., G.T., M.U., K.V., M.W., and W.W.W. contributed to the manu- script through fundamental feedback to the proposed draft. All authors have been involved in the collection of the data and building of the dataset. Competing interests The authors declare no competing interests. Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-68668-x. Correspondence and requests for materials should be addressed to Lorenzo Balducci. Peer review information Nature Communications thanks Mikko Pelto- niemi and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Reprints and permissions information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jur- isdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence toshare adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creativecommons.org/licenses/by-nc-nd/4.0/. © The Author(s) 2026 1Sapienza University of Rome, Rome, Italy. 2Austrian Research Centre for Forests (BFW), Vienna, Austria. 3Lessem, INRAE, University Grenoble Alpes, Saint Martin d’Heres, France. 4Department of Land, Environment, Agriculture and Forestry, Università degli Studi di Padova, Legnaro, PD, Italy. 5University of Göttingen, Silviculture and Forest Ecology of the Temperate Zones, Göttingen, Germany. 6UR EFNO, INRAE, Nogent sur Vernisson, France. 7WSL, Swiss Federal Institute for Forest, Snow and Landscape Research, Birmensdorf, Switzerland. 8Department of Forest Science, Agriculture Academy, Vytautas Magnus University, Kaunas, Lithuania. 9Council for Agricultural Research and Economics (CREA), Research Centre for Forestry and Wood, Arezzo (AR), Italy. 10Forest & Nature Lab, Department of Environment, Ghent University, Gontrode (Melle), Belgium. 11Institute of Biology and Environmental Science, Vegetation Science & Nature Conservation, University of Oldenburg, Oldenburg, Germany. 12Department of Botany and Zoology, Faculty of Science, Masaryk University, Brno, Czechia. 13Institute of Plant Sciences, University of Bern, Bern, Switzerland. 14DIFAR, University of Genova, Genova, Italy. 15Center for Macroecology, Evolution and Climate, University of Copenhagen, Globe Institute, Copenhagen, Denmark. 16Department of Botany, Faculty of Science, University of South Bohemia, České Budějovice, Czechia. 17Department of Forest Ecology, Faculty of Forestry and Wood Sciences, Czech University of Life Sciences Prague, Praha-Suchdol, Czechia. 18Ecological Services, Hořovice, Czechia. 19Department of Plant Biology and Ecology, University of the Basque Country UPV/EHU, Bilbao, Spain. 20Evolutionary Ecology and Conservation Genomics, University of Ulm, Ulm, Germany. 21Department of Geosciences and Natural Resource Management, University of Copenhagen, Copenhagen, Denmark. 22Department of Ecology and Environmental Sciences, Faculty of Sciences, Palacký University Olomouc, Olomouc-Holice, Czechia. 23School of Agricultural Forest and Food Sciences HAFL, Bern University of Applied Sciences, Zollikofen, Switzerland. 24Institute of Botany of the Czech Academy of Sciences, Průhonice, Czechia. 25Czech University of Life Sciences in Prague, Prague, Czechia. 26Bavarian Forest National Park, University of Würzburg, Rauhenebrach, Germany. 27Norwegian Institute for Nature Research NINA, Oslo, Norway. 28HUN-REN Centre for Ecological Research, Institute of Ecology and Botany, Budapest, Hungary. 29Institute of Research on Terrestrial Ecosystems (IRET) of the National Research Council (CNR), Florence, Italy. 30Faculty of Ecology and Environmental Sciences, Technical University in Zvolen, Zvolen, Slovakia. 31INBO-Research Institute for Nature and Forests, Geraardsbergen, Belgium. 32Albert-Ludwigs-University Freiburg, Freiburg, Germany. 33Department of Life Science Systems, School of Life Sciences, Technical University of Munich, Terrestrial Ecology Research Group, Freising, Germany. e-mail: [email protected] Article https://doi.org/10.1038/s41467-026-68668-x Nature Communications| (2026) 17:1976 13
Page 2 of 2