Accounting for forest condition in Europe based on an international statistical standard

Forest ecosystems, covering 35% of Europe's land area, are vital for biodiversity and climate change mitigation. Despite forest expansion and biomass accumulation in Europe, pressures like eutrophication, drought, and tree cover loss continue to degrade forest condition. This degradation negatively impacts biodiversity, rural economies, and the delivery of ecosystem services. Sustainable natural resource use is crucial for long-term economic competitiveness and security, making forest protection and restoration essential objectives of the European Green Deal. The SEEA EA framework, adopted by the UN Statistical Commission in 2021, provides a spatially-based statistical framework for tracking biophysical and economic information about ecosystems, linking it to economic and human activity. This study operationalizes a method to account for forest condition using spatially-explicit, frequently updated datasets, providing a large-scale monitoring system to establish conservation actions and restoration priorities. It also presents the first large-scale test of SEEA EA guidelines on ecosystem condition accounting, using most of the European continent as the accounting area.

The paper references existing literature on forest health and ecosystem assessments in Europe, noting a lack of a common, internationally adopted standard for assessment. It highlights the importance of the SEEA EA framework for providing a consistent, comparable methodology acceptable across different sectors. The review also covers existing literature on forest productivity, climate change impacts on forests, the importance of biodiversity indicators (specifically threatened bird species), and the effects of forest disturbance regimes. Studies on forest disturbance, canopy mortality, and biomass use are cited to provide context for the observed changes in forest condition. The role of soil organic carbon and its relationship with forest health is also discussed, along with the challenges of using various vegetation indices to represent productivity. Existing literature on indicator selection, weighting systems in index construction, and the challenges of defining reference conditions in the context of secondary vegetation and changing climate are reviewed.

The assessment of forest ecosystem condition strictly adheres to the SEEA EA framework. This involves defining an accounting area and forest typology, selecting condition variables, establishing reference levels, aggregating variables into an index, and assessing uncertainty and parameter sensitivity. Data sources include regularly updated datasets from initiatives like Copernicus. **Definition and Delineation of Forest Ecosystems:** The study used the EU's biodiversity strategy ecosystem typology and the CORINE Land Cover data (CLC) for 2000 and 2018, delineating forest ecosystems based on CLC classes (broad-leaved, coniferous, mixed forests, and transitional woodland/shrub). These were intersected with 11 biogeographic regions (excluding the Anatolian region due to data limitations), creating a typology of 44 forest subtypes. **Assessment of Forest Condition:** A three-step approach was employed: (1) Selection of seven forest condition variables (vegetation water content, soil organic carbon, threatened forest bird species richness, tree cover density, forest productivity, forest connectivity, and landscape naturalness), guided by the SEEA ecosystem condition typology and criteria such as directionality and data availability. (2) Definition of reference conditions using maximum variable values in primary or least-disturbed protected forests (IUCN categories Ia, Ib, and II). The minimum value, representing a degraded state, was derived from the ambient distribution within each forest type. (3) Aggregation of variables into an index using indicator-specific weights (determined through indicator ranking). Uncertainty analysis examined potential bias due to the use of primary/protected forests as reference conditions and differences in environmental conditions between reference and non-reference sites. A parameter sensitivity analysis assessed the impact of changes in reference levels and weights on the index values. **Data Sources and Processing:** Detailed descriptions and sources for each variable are provided. Data processing involved methods such as three-year averaging for NDWI to reduce interannual variability, Gaussian kriging for spatially interpolating LUCAS soil organic carbon data, generalized linear modeling for predicting threatened bird species richness based on various predictor variables, and the use of Forest Area Density (FAD) and landscape naturalness derived from CORINE Land Cover data.

The analysis covered 1,964,211 km² of forest in Europe. Forest condition exhibited a patchy distribution, with high conditions in eastern Alps, Carpathians, Scandinavia, and along the Black Sea, and lower conditions in the Atlantic plain, British Isles, and Iberian Peninsula. Most of the forest area (63%) experienced an increase in condition between 2000 and 2018, although the average increase was limited (4.3%). However, 37% of the area showed declining condition, more pronounced in northern Scandinavia, the Carpathians, the Balkans, the northern Apennines, and the Iberian Peninsula. The average forest condition increased from 0.566 in 2000 to 0.585 in 2018 (a 1.9% growth). This upward trend was consistent across most forest types (except for Macaronesian forests, which declined significantly). Forest condition varied significantly among forest types (0.31–0.78), with lower conditions observed in transitional woodland and shrub compared to broad-leaved, coniferous, and mixed forests. Regional differences were mainly driven by species richness of threatened forest birds. Forests in the Black Sea, Alpine, Continental, and Boreal regions generally scored above the European average, while Atlantic, Mediterranean, and Macaronesian forests had below-average conditions. The sensitivity analysis showed that the forest condition index was relatively stable to parameter variations (below 2.5% change for a 10% parameter perturbation), with soil organic carbon weight having the highest impact. The uncertainty analysis highlighted the potential bias due to the use of primary/protected forests as reference conditions. A significant portion of Europe's forests are secondary, and the reference sites used differed in environmental conditions (e.g., higher altitudes, colder temperatures). The uncertainty assessment resulted in a map assigning four uncertainty levels (low, low-medium, medium-high, high) to different forest types.

The findings demonstrate that European forests are in moderate condition compared to undisturbed forests, with a slow rate of improvement offset by declines in certain areas. Forest productivity and connectivity are relatively high, but the distance from the reference state highlights ongoing pressures. Improving soil carbon levels and conserving/restoring threatened bird species are essential for enhancing forest condition. This requires improved management practices and restoration efforts, along with extended recovery periods to approach natural conditions. The study's limitations in using only readily available data, such as omitting variables like deadwood and age structure, are acknowledged. The choice of indicator weights and the definition of reference levels introduce uncertainty. The selection of primary forests as references, when many areas lack such forests, creates potential bias. The study's results are relevant to policy initiatives like the European Green Deal and the proposed nature restoration law, providing a consistent framework for monitoring and guiding restoration efforts.

This study provides a comprehensive assessment of European forest condition using the SEEA EA framework, offering a robust, spatially explicit, and internationally comparable baseline. The findings highlight the need for continued forest restoration and improved management practices to enhance forest condition and ecosystem services. Future research could focus on incorporating additional data (e.g., deadwood, age structure), refining weighting systems, and developing more dynamic reference conditions to account for climate change and other environmental shifts. The integration of forest management practices into the accounting framework also needs further exploration.

The study's reliance on readily available data limits the inclusion of certain variables (deadwood, age structure, detailed species composition). The weighting system used for the condition index, while tested, could be subject to further refinement. The uncertainty analysis identified limitations due to reliance on primary or protected forests as reference sites, which may not fully capture the variation in forest conditions across Europe. The assumption of a boreal reference condition for the Arctic region introduces potential bias due to ongoing changes in vegetation.