Annual forest biomass dynamics, harvesting, and litter inputs from 1985 to 2023 reconstructed by using NFI and satellite data

This data was reconstructed using an integrated framework for annual forest biomass dynamics, tree stand growth and harvesting losses, and litter inputs by combining National Forest Inventory (NFI) data from 2805 sites from Finland of years 1985, 1990, 1995, and 2023 with satellite‐derived basal area observations from the period of 2005 to 2023. The workflow produced spatially and temporally explicit biomass and litter input time series (1985–2023) by tree species and tree components (such as foliage, branches, stem and stumps, coarse- and fine- roots), distinguishes biomass increase as a growth and biomass decrease as harvesting losses, and converts turnover/death rates based litter inputs from living trees and residues from harvesting thus total plant litterinto chemically and physically resolved AWEN fractions of three size group (non-woody, fine-woody and coarse woody litter) suitable for soil carbon modelling and greenhouse‐gas reporting. The reconstructed data is deliberately inventory‐conservative The NFI data defined biomass magnitude and plot specific forest stand structural composition (distribution of the biomass to species and their components), while satellite derived data on the plot basal area were used exclusively to enhance temporal resolution, detect growth and disturbance timing, and guide interpolation between inventory years. No direct biomass prediction from satellite variables was applied. The dataset spans 1985–2023 and represents a fully mass‑balanced annual biomass and litter accounting dataset, suitable for: soil carbon modelling (e.g. Yasso‑type models), ecosystem carbon balance analysis, national greenhouse‑gas inventory applications. The analysis is decribed in detail in the included manuscript "BIOMASS_LITTER_NFI_SATELLITE manuscript ZENODO.pdf" and can be replicated by running the included R code "litter_sites_v57_01.07.26.R". The input files requires biomass data from NFI and BA from satellite observations are located in data folder "input files and codes.zip" . These files were prepared in a workflow by [e-mail] and [e-mail] "biomass_fineroot.csv" # NEW biomass with fine roots "kasvillisuustk_kooste_13022026.csv" # OLD metadata & site attributes "kuviotiedot_pysyvät_MSNFI.xlsx" # Satellite basal area The biomass files can be replicated by running the sourcefile.R code from "biomass_workflow" folder in input files and codes. More detail description of the biomass estimation is in the readme.pdf file in the same folder. In addition the main R script sources a code provided by [e-mail]("process_fixed_data_ulit.R") # providing a base_year, reporting_year, awen_trees, and litter_rates0 - information contained in the "fixed_data" folder of "input files and codes" The output dataset provides the following biomass and litter data files: The biomass and litter data file "tree_biomass_litter_per_site_species_component_year.csv" and the files containing site level and component level litter data distributed to AWEN fractions for soil C modelling with Yasso model ("tree_litter_per_site_year_awen.csv" and "tree_litter_per_site_year_by_component.csv"). The "tree_biomass_litter_per_site_species_component_year.csv" is a site–year–species–component level dataset containing reconstructed forest biomass stocks, biomass changes, harvest losses, and litter inputs derived from an integrated NFI–satellite analysis. The "tree_litter_per_site_year_awen.csv" and "tree_litter_per_site_year_by_component.csv" datasets represent chemically resolved litter inputs to soil, separated by tree species, biomass component, and litter size class. File name Purpose Resolution 1. tree_biomass_litter_per_site_species_component_year.csv Core biomass and litter output site × year × species × component 2. tree_litter_per_site_year_awen.csv Aggregated AWEN‑resolved litter carbon inputs site × year × species × size × AWEN 3. tree_litter_per_site_year_by_component.csv Component‑level AWEN litter inputs site × year × species × component × AWEN Key interpretation notes All biomass and litter flows are mass‑balanced annually. Satellite data inform temporal dynamics, not biomass magnitude. File 2 is derived by aggregating File 3 across components. File 1 is the source dataset for all litter outputs. Units and conventions Biomasses are expressed in tons of dry matter of biomass and litter is expressed in tons of carbon. All variables are expressed per hectare. Biomass: t DM ha⁻¹ Carbon: t C ha⁻¹ yr⁻¹ Coordinates: projected CRS consistent with NFI input README description of variables in all the data files: Shared identifiers and spatial variables (all files) Variable Description site Harmonized site identifier linking NFI and satellite data region Modelling region (1 = South, 2 = North) year Calendar year (1985–2023) koealatunnus_MUSTIKKA MUSTIKKA plot identifier koealatunnus_VANHA Legacy NFI plot identifier koealatunnus_BIOSOIL BIOSOIL plot identifier x, y Primary projected coordinates x_kasvika, y_kasvika Alternative coordinate reference Biological classification variables Variable Description species Species group (1 = Pine, 2 = Spruce, 3 = Broadleaved) species_f Factor version of species component Tree component (1 stem, 3 branches, 4 foliage, 6 stump, 7 coarse roots, 10 fine roots) comp_f Factor version of component size Litter size class (nwl non‑woody, fwl fine woody, cwl coarse woody) compound AWEN fraction (a, w, e, n) Biomass variables (File 1) Variable Description (units: t DM ha⁻¹) bm_comp Observed or reconstructed component biomass bm_species Species‑level biomass bm_total Observed total biomass (NFI years only) bm_smooth_pred Smoothed annual total biomass predicted from PPA species_bm_smooth Smoothed species biomass comp_bm_smooth Smoothed component biomass (primary stock variable) Biomass ratios and structure preservation (File 1) Variable Description species_ratio Observed species fraction of total biomass spcomp_ratio Observed component fraction within species mean_species_ratio Site‑mean species biomass ratio (from NFI) mean_component_ratio Species‑mean component ratio (from NFI) Basal area variables (File 1) Variable Description ppa Observed or harmonized basal area (m² ha⁻¹) ppa_smooth Smoothed annual basal area time series Biomass change and harvest variables (File 1) Variable Description comp_prev Component biomass in previous year comp_change Annual biomass change gain_comp Annual biomass growth loss_comp Final biomass loss used for litter accounting loss_strict Year‑to‑year loss in satellite era total_loss, per_year_loss Redistributed NFI‑interval losses Stem biomass losses represent removals; non‑stem losses remain as harvest residues. Litter variables Variable Description litter_rate Component‑specific turnover rate (yr⁻¹) litter_turnover Litter from living biomass turnover (t DM ha⁻¹ yr⁻¹) litter_harvest Litter from harvest residues (non‑stem only) mlitter Total litter dry mass (File 3) Cha Litter carbon input (t C ha⁻¹ yr⁻¹; 50% C assumed) proportion AWEN fraction applied README description of variables in the INDIVIDUAL data files: File 1. tree_biomass_litter_per_site_species_component_year.csv site Unique site or plot identifier corresponding to harmonized NFI/satellite locations. region Regional class used in modelling. 1 = South 2 = North Used to apply region‑specific PPA corrections, biomass models, and litter turnover rates. year Calendar year (integer), continuous from 1985 to 2023. species Tree species group code. 1 = Pine‑dominated species 2 = Spruce‑dominated species 3 = Broadleaved species species_f Factor version of species with labels Pine, Spruce, Birch. component Tree biomass component code. Code Component 1 Stem (incl. bark) 3 Branches 4 Foliage 6 Stumps 7 Coarse roots 10 Fine roots comp_f Factor version of component with descriptive labels. Biomass stocks (NFI‑based and reconstructed) bm_comp Observed or reconstructed component biomass(Mg dry matter ha⁻¹). Derived from NFI data and redistributed from annual total biomass when interpolated. bm_species Observed or reconstructed species‑level biomass(sum of components within species). bm_total Observed total site biomass in NFI years (sum of all components and species). Only populated at inventory years; differs from smoothed predictions. bm_smooth_pred Smoothed annual total biomass prediction(Mg ha⁻¹), estimated from reconstructed basal area (PPA) using the log–log biomass model. This variable defines the annual biomass trajectory between NFI years. species_bm_smooth Smoothed annual species biomass, obtained as: species_bm_smooth=bm_smooth_pred×mean_species_ratio comp_bm_smooth Smoothed annual component biomass, obtained as: comp_bm_smooth=species_bm_smooth×mean_component_ratio This is the primary biomass stock used for growth, loss, and litter calculations. Biomass ratios (structure preservation) species_ratio Observed species biomass fraction at a given site and year: bm_species\bm_total spcomp_ratio Observed component fraction within species: bm_comp\bm_species mean_species_ratio Site‑specific mean species biomass ratio, averaged over all NFI observations. Assumed constant for annual redistribution. mean_component_ratio Species‑specific mean component ratio, averaged over inventory years. Ensures mass balance and stable structural allocation. Basal area (structural driver) ppa Observed or bias‑corrected basal area (m² ha⁻¹) from NFI or satellite data. ppa_smooth Smoothed annual basal area time series. Linear interpolation for NFI period Smoothing (e.g. rolling median / GAM) applied only in satellite period Used as the structural input to the biomass–PPA model. Biomass change variables comp_prev Component biomass in the previous year: B_t−1 comp_change Annual component biomass change: ΔB=B_t−B_t−1 gain_strict Positive component biomass change: max⁡(ΔB,0) Represents growth. loss_strict Negative component biomass change (absolute value): max⁡(−ΔB,0) Represents mortality + harvesting in satellite era. NFI‑period loss redistribution variables These variables are used to redistributing biomass losses between sparse NFI years (1985–1995). total_bm Total site biomass used to compute NFI interval losses. prevy Start year of NFI interval. nexty End year of NFI interval. total_loss Total biomass loss over the NFI interval: max⁡(Bprevy−Bnexty,0) n_years Length of the NFI interval in years. per_year_loss Annualized biomass loss during NFI interval: total_lossn_years comp_share Proportion of component biomass within total site biomass in a given year. loss_comp Final component‑level biomass loss used in litter accounting. Uses redistributed NFI losses before 1995 Uses loss_strict after 1995 gain_comp Final component‑level biomass growth (used mainly for diagnostics). Litter parameters litter_rate Component‑specific turnover rate (yr⁻¹). Derived from region‑ and species‑specific parameter tables. reduction_rate Rate applied to adjust litter input under harvesting (currently identical to litter_rate, kept for extensibility). Litter fluxes litter_turnover Litter input from living biomass turnover: comp_bm_smooth×litter_rate Occurs every year regardless of management. litter_harvest Litter input from harvest residues: 0 for stems (removed from site) Equal to loss_comp for all other components Represents roots, stumps, branches, amd foliage retained on site. Interpretation summary In practical terms: comp_bm_smooth = best estimate of standing biomass gain_comp and loss_comp = annual biomass dynamics litter_turnover = biological litter input litter_harvest = management‑induced litter input litter_turnover + litter_harvest = total litter to soil File 2. tree_litter_per_site_year_awen.csv NOTE: VALUES ARE ALREADY AGGREGATED ACROSS TREE COMPONENTS. This file provides site‑level annual litter carbon inputs, already aggregated across biomass components but resolved by: tree species, litter size class, AWEN chemical fraction. Each row represents a unique combination of: site × year × species × size × AWEN compound Variables Identification and spatial variables siteHarmonized site identifier linking NFI, satellite, and soil datasets. regionRegional classification used in modelling. 1 = South 2 = North yearCalendar year (1985–2023). koealatunnus_MUSTIKKAMUSTIKKA system plot identifier (if available). koealatunnus_VANHALegacy NFI plot identifier. koealatunnus_BIOSOILBIOSOIL inventory identifier. x, yPrimary spatial coordinates (projected). x_kasvika, y_kasvikaAlternative coordinate reference from KASVIKA dataset. Biological classification speciesTree species group code: 1 = Pine‑dominated 2 = Spruce‑dominated 3 = Broadleaved sizeLitter size class used in soil carbon modelling: nwl = non‑woody litter (foliage + fine roots) fwl = fine woody litter (branches + coarse roots, stem+bark of the living trees) cwl = coarse woody litter (stumps) compoundAWEN chemical fraction: a = acid‑soluble w = water‑soluble e = ethanol‑soluble n = non‑soluble (lignin‑rich) Carbon input variable ChaAnnual litter carbon input (t C ha⁻¹ yr⁻¹) for the given species, size class, and AWEN fraction. Calculated as: Cha=0.5×L_s,c×p_s,c,k where L_s,c is total litter dry mass and p_s,c,k is the AWEN fraction. File 3: tree_litter_per_site_year_by_component.csv Purpose This file provides fully disaggregated component‑level litter inputs. Each row represents a unique combination of: site × year × species × biomass component × AWEN compound Variables Identification and spatial variables (Same definitions as File 1) site, region, year koealatunnus_MUSTIKKA koealatunnus_VANHA koealatunnus_BIOSOIL x, y x_kasvika, y_kasvika Biological classification speciesTree species group code (1 = Pine, 2 = Spruce, 3 = Broadleaved). componentTree biomass component code: Code Component 1 Stem (incl. bark) 3 Branches 4 Foliage 6 Stumps 7 Coarse roots 10 Fine roots sizeLitter size class derived from component for the living trees turnover litter: foliage, fine roots → nwl branches, coarse roots, stem+bark → fwl stumps → cwl compoundAWEN chemical fraction (a, w, e, n). Litter variables mlitterTotal litter dry mass input (t DM ha⁻¹ yr⁻¹) from the given component, including: turnover litter from living biomass, and harvest residues (non‑stem components). proportionAWEN fraction proportion applied to this species–component combination. ChaCarbon input (t C ha⁻¹ yr⁻¹) for the specific component and AWEN fraction: Cha=0.5×mlitter×proportion Interpretation notes Stem litter components (component = 1) only contribute through turnover, not harvest residues.