Carbon carrying capacity in primary forests shows potential for mitigation achieving the European Green Deal 2030 target

International climate agreements aim to limit warming to well below 2 °C, but operationalizing land-sector mitigation requires rethinking carbon accounting, particularly the choice of reference level to evaluate stock changes. Current approaches often use projected management-based reference levels that do not capture the ecological potential of forests nor differentiate ecosystem condition, leading to undercounting historical carbon stock debt and masking emissions from harvesting. The authors propose using an ecologically based reference level: the carbon carrying capacity (CCC) represented by primary forests—defined as naturally regenerated forests dominated by ecological processes with minimal direct human disturbance. CCC is the long-term, landscape-level carbon stock sustained by species life histories, environmental conditions, and natural disturbance regimes, excluding direct anthropogenic impacts. Europe’s long land-use history complicates identification of primary forests, but these forests retain the highest ecosystem integrity and carbon stocks for given forest types. The study addresses two questions: (1) What is the CCC of European forest types as represented by extant primary forests, and how do these stocks compare with global default values and models? (2) What is the mitigation potential from protecting primary forests (avoiding losses) and restoring secondary forests (gaining stocks toward CCC)?

The paper situates its approach within ongoing debates about land-sector accounting under the Paris Agreement and EU LULUCF rules, noting that current methods inadequately distinguish ecosystem condition and net out harvest emissions across entire forest areas. Prior global models (e.g., GlobBiomass, GeoCarbon) underestimate high biomass forests due to sensor saturation and calibration limitations, particularly above ~125 MgC ha⁻¹ and in broadleaf or mixed types. IPCC default biomass values (updated in 2019 from 2006 guidelines) improved differentiation by Global Ecological Zones (GEZs) and forest types, but are based on limited datasets in Europe and few primary forest studies, leading to large uncertainties (~90%) and data gaps. Europe’s land-use legacies—including conversion to production forestry—have reduced biomass stocks relative to potential. Empirical and modeling studies indicate mature and old-growth forests can maintain sinks and high stocks; large-diameter trees disproportionately drive biomass; and wood product pools store far less carbon and for shorter durations than standing forests. These strands motivate using CCC from primary forests as a reference for mitigation accounting and restoration targets.

Study scope and data: The authors compiled tree inventory data for primary forest sites across Europe’s boreal, temperate, and subtropical (Mediterranean) GEZs: 7,982 sites (1,818 research sites from existing studies, 6,015 national forest inventory sites filtered for primary/natural forest ≥100 years old and some protection status, and 149 literature sites), totaling 288,262 trees across 27 countries. Site representativeness was evaluated against environmental space (elevation, mean annual temperature, water availability index) within current forest cover per GEZ. Forest types were classified as conifer, broadleaf, or mixed (CORINE Land Cover 2018). Measurements and carbon components: At each site, DBH (and sometimes height), species/forest type, living and dead trees, and coarse woody debris (CWD) were measured. Biomass of individual trees (above- and belowground) and dead standing trees/CWD was estimated using regional/species-specific allometric equations, wood density, carbon concentration, and decay-class adjustments. Litter and soil organic carbon (SOC) were not consistently measured in the field; SOC was extracted from the FAO GSOC global map. Dead and belowground biomass were included to estimate total ecosystem biomass carbon. Tree size distribution and harvest scenarios: Using individual-tree data, the authors derived distributions of tree density and biomass by DBH class and calculated cumulative biomass contributions. They quantified the proportion of total biomass in large trees and simulated harvest maturity scenarios by imposing DBH thresholds (broadleaf 50–80 cm; conifer 40–60 cm; mixed 50–70 cm) and limiting maximum tree sizes to those thresholds while keeping tree density constant, estimating foregone biomass as percentage of primary forest biomass. Spatial comparison and reference levels: Global spatial datasets (forest extent; GlobBiomass and GeoCarbon for aboveground biomass; GSOC for soils) were used to extract modelled biomass at site locations and across mapped primary and secondary forest areas per GEZ. The CCC was represented by mean biomass carbon density from primary forest sites within each GEZ. Current carbon stock (CCS) of secondary forest was represented by global modelled biomass. Ratios of site-based CCC to modelled values were applied spatially: (1) to update total carbon stock in mapped primary forest (CCC), and (2) to estimate potential gains in secondary forests by scaling CCS up to CCC. Aggregation was reported for biomes (boreal, temperate, subtropical). Statistical comparisons (e.g., t-tests) evaluated differences between site and modelled distributions. The analysis covers existing forest area only (excludes cleared potential natural forest).

- Dataset and coverage: 7,982 sites, 288,262 trees, 27 countries across Europe’s boreal, temperate, and subtropical GEZs. - Biomass stocks by country/type: Aboveground plus belowground and dead biomass varied by an order of magnitude, from 21 MgC ha⁻¹ (alpine birch, Sweden) to 346 MgC ha⁻¹ (mixed spruce–fir–beech, Bosnia-Herzegovina). Dead biomass peaked at 113 MgC ha⁻¹ in Czech conifer forest after severe windthrow. Mountain systems tended to have higher biomass; boreal zones were lower than temperate/subtropical. - Primary forest site data vs global models: Frequency distributions showed consistently higher site-based aboveground biomass carbon densities than GlobBiomass and GeoCarbon for all forest types, with significant differences (P < 0.0001 across categories). Cumulative biomass from models relative to site data was lower by factors: all forest types 0.57 (GeoCarbon) and 0.59 (GlobBiomass); conifer 0.64 and 0.65; broadleaf 0.53 and 0.61; mixed 0.28 and 0.46. Field data at site locations averaged ~3× higher than modelled biomass. - Within GEZs: Site biomass for primary forests was similar to models in boreal forests, about 2× higher in temperate, and 2.5–3× higher in subtropical forests. Compared with IPCC default values, site biomass was notably higher, including more than double in temperate oceanic forests, and similar/slightly lower in temperate continental forests. - Large trees dominate biomass: Across primary forest sites, 50% of total live biomass carbon was contained in trees with DBH ≥ 60 cm (median), representing only ~15% of tree numbers. Inclusion of dead and belowground biomass added on average 16% and 22%, respectively, to aboveground living biomass. - Foregone carbon due to harvesting: Simulated harvest maturity thresholds indicated foregone biomass of 12–21% at higher DBH thresholds and 46–52% at lower thresholds, illustrating mitigation potential from allowing trees to grow larger/older. - European forest area context: Europe has 174 Mha of forest; 1.07% is mapped as primary forest (78% of this in the boreal). The total aboveground biomass carbon stock across all forests is ~9,861 TgC; primary forests contribute ~1.02% of this stock (~100.8 MtC). - Updated CCC for primary forests: Total carbon stock in mapped primary forest updated using site data is 1.6× higher than global model estimates (temperate and subtropical roughly double or more; boreal similar). - Potential gains in secondary forests: Estimated CCC in secondary forest areas is 2.29× their current CCS, implying an additional 12,659 MtC (46,415 MtCO₂) of potential stock gain through restoration/growth. - Annual removal equivalence: Assuming linear growth to CCC over 150 years, potential removals equate to ~309 MtCO₂e per year (or ~464 MtCO₂ per year over 100 years), additional to the current EU forest sink (289 MtCO₂ in 2021), and comparable to the EU Green Deal 2030 LULUCF target of 310 MtCO₂ net removals. - GEZ-specific CCC vs CCS (aboveground biomass MgC ha⁻¹): Boreal mountain CCC 29.7 vs CCS 33.1; Boreal coniferous CCC 44.3 vs CCS 42.2; Temperate continental CCC 145.5 vs CCS 61.6; Temperate mountain CCC 158.0 vs CCS 76.7; Temperate oceanic CCC 147.5 vs CCS 45.8; Subtropical mountain CCC 72.9 vs CCS 42.3; Subtropical dry CCC 178.4 vs CCS 186.8.

The study demonstrates that primary forests in Europe store substantially more carbon than indicated by global biomass models and many default values, thereby underestimating their mitigation value. Underestimation arises from sensor saturation and calibration limitations in remote sensing at high biomass, lack of differentiation among forest functional types, and biases in plot sampling (small area sampled, spatial bias, limited representation of some forest types/regions). Allometric equations often underrepresent large trees, further biasing estimates downward; including dead and belowground components increases total stocks substantially. Large, old trees disproportionately contribute to ecosystem carbon storage, yet are scarce in production forests managed on short rotations or selective harvest. Protecting and restoring primary forests thus avoids significant, long-lived emissions and maintains a resilient carbon reservoir with lower risk of loss. Mature and old forests can continue to function as carbon sinks due to structural complexity and ongoing growth, contributing substantially to the global forest sink. Restoration of secondary/degraded forests towards CCC offers large additional removals, potentially matching EU LULUCF targets within existing forest area. However, wood provisioning needs and other land-use demands must be balanced. Strategies such as strict protection of remaining primary/old-growth forests, extending rotation lengths, retaining large trees, restoring coppice systems, reducing waste, enhancing long-lived wood products, and improving substitution effects can enhance carbon storage while supporting other ecosystem services. The CCC-based reference level provides an ecologically grounded accounting framework to quantify past carbon stock losses, foregone mitigation from managing below CCC, and future gains from restoration, improving transparency and policy relevance.

Retaining and increasing forest carbon stocks—especially by protecting primary and old-growth forests—and avoiding emissions should be prioritized in climate and biodiversity policy under the European Green Deal. Europe’s remaining primary forests are limited and only about half are strictly protected; enforcing strict protection and managing buffer zones are urgent. Restoration of existing forests by allowing continued growth, targeted active measures, and reconnecting fragments can deliver large additional removals and co-benefits for biodiversity and ecosystem services. Using CCC as an ecological reference level clarifies mitigation accounting by highlighting avoided losses, foregone carbon under current management, and potential gains. With appropriate governance and management, restoration of existing forests could on its own achieve removals comparable to the 2030 EU Green Deal LULUCF target, while broader emissions reductions and sustainable wood use strategies proceed in parallel.

- Site representativeness: Primary forest sites are spatially biased and may often occur on less productive or extreme sites (steep slopes, shallow soils, high elevations), potentially yielding conservative CCC estimates; conversely, some sites may include productive remnants. Some forest types/regions (e.g., subtropical dry/mountain, coastal boreal and temperate oceanic) are underrepresented. - Mapping uncertainties: The mapped area of primary forest may be overestimated in some countries (e.g., Sweden, Finland, Iberian Peninsula). Misclassification influences both total CCC estimates for primary forests and the area available for secondary forest restoration, potentially amplifying mitigation estimates. - Component data gaps: Soil carbon was derived from global maps due to insufficient field data; improved SOC measurements would reduce uncertainty. Dead biomass and belowground pools are variably measured and may still be underestimated. - Model and method limits: Remote sensing models underestimate high biomass; plot-based allometry underestimates large-tree biomass; spatial upscaling via ratios introduces uncertainty. - Harvest simulation assumptions: Foregone carbon estimates assume constant tree densities and size caps at harvest thresholds, and landscape-scale age distributions in managed forests could lower mean stocks, implying conservative estimates. - Temporal dynamics: The conversion of potential stock gains to annual removal rates assumes linear (or simplified) accumulation over 100–150 years; actual growth trajectories are likely nonlinear and climate-sensitive. - Scope: Estimates address existing forest areas only and exclude potential reforestation/afforestation of cleared lands.