Status of SOC in EU - NextGen C task, intermediate aboveground biomass products

European forest above-ground biomass, before and after disturbance fusion — NextGenC Annual above-ground biomass (AGB) density maps of European forests, 1990–2024, at 300 m in EPSG:3035 — the same biomass field before and after acute-disturbance fusion, each with a per-pixel uncertainty layer. Produced for the NextGen Forest Carbon (NextGenC) reporting effort. This is a demonstration / illustrative product — the intermediate biomass maps that feed the project's eventual carbon-input and soil-carbon maps (not yet built). See Interpretation & limitations below before using the maps quantitatively. Contents 4 multiband GeoTIFFs: AGB_before_fusion.tif and AGB_after_fusion.tif (AGB density, t ha⁻¹) and their _sd.tif uncertainty (1 SD). 35 bands each = years 1990–2024 (band n = year 1989 + n; year also stored as the layer name, e.g. y1990). thumbnails/ — PNG previews (1990–2024 time-average), including the supporting disturbance layers that are used to build the "after" map but are not deposited. README.md (full file & band documentation), LICENSE, and MD5SUMS.txt. Before vs after. The two AGB maps are identical over the CCI era (2005–2024); they differ only in 1990–2004, where the "after" map has acute-disturbance dynamics carved into the per-pixel pattern (the "before" map uses a frozen average template there). From 2005 on, disturbance information is not in the biomass layer but in the separate products, which are not part of this deposit. Raster specification.- CRS: ETRS89-LAEA Europe (EPSG:3035), equal-area.- Resolution 300 m; grid origin (900000, 5500000) m, all bands and both products pixel-aligned on one grid; grid 21667 × 14000 (before, full FAO-EU footprint; after covers the 25-country disturbance-atlas footprint on the same grid).- Int16 with scale_factor 0.1 (true value = stored × 0.1; precision 0.1 t ha⁻¹); NoData = −9999 (non-forest / outside coverage); units tonnes of biomass per hectare. Cloud-Optimized GeoTIFF, DEFLATE-compressed, with overviews; readable as an ordinary GeoTIFF. Method. ESA CCI radar biomass is first decompressed for SAR saturation (Bayesian saturation model calibrated on ICP basal-area plots). It then supplies the spatial pattern and per-pixel year-to-year change, while each country-year is rescaled so its total matches the FAO FRA 2025 national above-ground biomass (a hard constraint on country totals only) — giving the "before fusion" map. Acute disturbances are then detected with the European Forest Disturbance Atlas (occurrence + agent); their biomass loss is taken from CCI where it exceeds a per-stratum noise floor, and carved into the pixel pattern in the pre-2005 period (where no annual CCI signal exists), giving the "after fusion" map. Uncertainty is propagated from the Step-1 sources. Interpretation & limitations. This is an intermediate, demonstration-grade product: several steps are deliberately simple, and their approximations are meant to be absorbed by the downstream Bayesian calibrations (NextGenC Steps 2–3), which are not yet run. Consequently the maps are not the final carbon product; the _sd layers propagate Step-1 sources only and will change once the Bayesian steps run; the disturbance detection thresholds, the mortality-fraction transfer and the biomass scaling are first-order placeholders; the carved disturbance is above-ground biomass only (killed roots and the fate of disturbed biomass are handled later); FAO smoothing makes national totals piecewise-linear between the FRA epochs while per-pixel variability is preserved from CCI; the CCI saturation ceiling is structurally uncertain at the highest biomasses; and the "after fusion" map covers 25 EU countries (Cyprus and Malta absent from the disturbance atlas) whereas "before fusion" covers all FAO-reported EU countries, so their extents differ slightly. Where this fits in NextGenC, and what comes next. This biomass release is Step 1 of a three-step spatial pipeline whose goal is annual carbon-input forcing maps and a soil organic carbon (SOC) status map for EU forest soils. The maps here are the spatial and temporal backbone that the next two Bayesian steps turn into carbon fluxes and soil carbon — and it is there that the provisional simplifications of this demonstration become calibrated, uncertainty-quantified estimates. Step 2 (carbon-input maps) partitions the standing biomass into component pools (foliage, branches, stems, coarse and fine roots) and converts them to annual litter and mortality input via component-specific turnover rates, calibrated against ICP Forests litterfall; unobserved components such as roots are carried as explicit wide priors, and the acute-disturbance loss delivered here is routed by agent × component × severity into soil / removed / combusted fractions released through agent-specific lag kernels. Step 3 (SOC status) calibrates a soil decomposition model against ICP soil-carbon data to convert the carbon-input series into a soil organic carbon map. Key assets of the Bayesian approach: staged calibration with clean cuts (each step calibrated on its own data, no parameter identified from two datasets at once); end-to-end posterior uncertainty propagation, so the provisional per-pixel SD here is later replaced by a full budget that widens honestly where data are sparse; conservation by construction, with simplex-constrained fate fractions and lag kernels enforcing mass balance in fate and time inside the model; priors that absorb the deliberately simple deterministic choices made in Step 1 rather than treating them as fixed truth; honest treatment of unobserved components as wide priors; and stratification by forest type, species, soil and climate. National totals remain anchored to FAO throughout, and the carbon-input product is framed as a gross-input diagnostic (input = stock × turnover) rather than a closed NPP−losses budget. Credits & attribution. Project: NextGenC (NextGen Forest Carbon), Natural Resources Institute Finland (Luke). Companion to a separate Finland SOC pilot dataset within the same reporting effort (different data and methods). Source data (not redistributed; please credit if reused): ESA CCI Biomass (ESA Climate Change Initiative); FAO Global Forest Resources Assessment 2025; European Forest Disturbance Atlas — Viana-Soto & Senf (2025), ESSD 17:2373, Zenodo 13333034, CC-BY-4.0; Copernicus HRL Forest Type (European Environment Agency). License: Creative Commons Attribution 4.0 International (CC-BY-4.0). Contact: [e-mail]