Mapping forest-based natural climate solutions

The study addresses how to identify, map, and prioritize feasible locations and actions for forest-based natural climate solutions (NCS) to mitigate climate change. While decarbonizing energy systems remains paramount, NCS—through protection, restoration, and improved land management—offer deployable and scalable mitigation opportunities. Estimating mitigation potential is complicated by heterogeneous landownership, varying feasibility of management strategies, and previously limited availability of consistent, high-resolution datasets on forest carbon stocks and fluxes. The authors propose a spatially explicit, decision-relevant framework that incorporates land management restrictions to identify where NCS actions are appropriate and likely to deliver atmospheric benefits. They focus on the cool, wet, and high-carbon-density coastal temperate rainforests of western North America, which have high sequestration rates, relatively low risk of wildfire-driven reversals compared to many other ecozones, but high risk of disturbance from timber harvest. The paper aims to demonstrate a generalizable mapping approach using recent global datasets to quantify NCS opportunities in this 51 Mha ecoregion and to estimate the potential contribution of improved forest management and conservation to 2030 land-based climate commitments by the United States and Canada. The study defines sequestration as uptake through forest growth and carbon losses as stock decreases from disturbances, noting that losses to harvested wood products and operational emissions are not included in their emissions accounting, which relies on globally consistent models.

Prior work highlights the significant mitigation potential of forests and the importance of actions such as improved forest management, conservation, and protection of at-risk carbon stocks. However, many assessments have relied on ambitious, generalized scenarios and lacked high-resolution, spatially explicit consideration of landownership and management restrictions, leading to potential overestimation of opportunities. Emerging global datasets now provide consistent, higher-resolution estimates of aboveground biomass and carbon fluxes with uncertainty, enabling improved regional planning. Previous mapping and cluster analyses have linked potential emissions reductions to land management strategies across administrative regions, but until recently, a paucity of globally consistent flux and stock data hindered robust ecoregional quantification. The study builds on this literature by integrating land tenure and designations with carbon models to better align mitigation planning with on-the-ground feasibility and policy levers, acknowledging evolving science, regional variability, and risks (e.g., wildfire) to carbon durability.

Study area: The coastal temperate rainforests of western North America (approximately 51 Mha) were delineated using subnational datasets: Alaska Level 2 Coastal Rainforests; British Columbia Cool Hypermaritime and Highlands Ecodivision; and, in Washington and Oregon, ecoregions combined with fire regime groups to select areas with long return intervals or low–mixed severity regimes. The WA/OR extent was further refined using forest productivity and Biophysical Settings/Potential Vegetation Types to retain areas with high carbon potential, overlaid with 12-digit hydrologic units.

Spatial hierarchy and NCS action categories: A jurisdictionally nested database was built by integrating authoritative land ownership and designation datasets (country, state/province, landowner type, and specific management designations). Lands were classified into NCS Action Categories based on feasibility of additional carbon mitigation through management: NCS1 (available for action; e.g., private industrial forests, public timber tenures, National Forest areas open to harvesting); NCS2 (not currently available; administratively protected, e.g., riparian buffers, wildlife reserves); NCS3 (not available; permanently protected, e.g., national/provincial parks, wilderness). Where explicit spatial layers were lacking (e.g., some riparian buffers), regulations were mapped using GIS-derived buffers on hydrography.

Ownership and designation data: For Alaska, sources included BLM National Surface Management Agency polygons, Tongass suitability and available stands, Haines State Forest vegetation polygons, State Forest boundaries, University of Alaska Land Grant Trust, and Alaska Mental Health Trust parcels. NCS2 included riparian buffers via NHD and existing forest carbon projects, and PAD-US GAP 2; NCS3 included PAD-US GAP 1. For British Columbia, Generalized Forest Cover Ownership was supplemented with Aboriginal Lands of Canada Legislative Boundaries, Tree Farm Licenses, Timber Supply Areas; NCS2 included riparian buffers using CanVec and lakes/coasts per BC Fish Protection Act and various reserves; NCS3 used CPCAD IUCN I–III. For Washington, datasets included NGDA, PAD-US, Northwest Forest Plan (NWFP) Land Use Allocations, county tax parcels; NCS2 encompassed NWFP reserves and PAD-US GAP 2, with riparian/wetland buffers per Washington Forest Practices Rules; NCS3 included congressional reserves and PAD-US GAP 1. For Oregon, BLM and ODF ownership/management, USFS Region 6, parcel datasets, TIGER/Line AIANNH, and NWFP LUAs informed classification; NCS2 included BLM wilderness, FWS study areas, NWFP reserves; NCS3 included PAD-US GAP 1 and riparian buffers per Oregon Forest Practices Act and Private Forest Accord.

Forest mask and data processing: A 30 m North America Land Change Monitoring System (NALCMS) 2015 land cover mask encompassing forest, shrubland, and grassland classes was used (to capture recent clearcuts). Data were resampled and aligned as needed.

Carbon stocks: Aboveground biomass (AGB) and associated standard errors were sourced from the ESA Biomass_cci global AGB map for 2018 (100 m, resampled to 90 m), and national/regional models for Canada (Matasci et al., 30 m), U.S. Pacific Northwest regional (Hudak et al., 30 m), and contiguous U.S. (Williams et al., 30 m). Units were converted from Mg ha⁻¹ to Mg per pixel, then to MgC (×0.47), then to MgCO2e (×3.67). Totals were summarized by ownership polygons; SE rasters were aggregated by summing, squaring, and square-rooting, with 95% CIs derived from SE × 1.96.

Carbon fluxes: Global spatially explicit carbon flux maps (Harris et al.) provided average annual sequestration and emissions for 2001–2021. Flux rasters (average MgCO2e pixel⁻¹ over 21 years) were converted to average annual values and summarized by ownership and jurisdiction within NCS1 areas. Flux uncertainties were not available at this analysis level.

2030 opportunity scenario: Historical timber harvest variability (coefficient of variation ~10%) was used as a conservative target to estimate reduced emissions from improved forest management and conservation. Applying a 10% reduction to the region’s average annual forest carbon losses (91 MtCO2e yr⁻¹) yields an estimated 9.1 MtCO2e yr⁻¹ reduction, compared to U.S. and Canada’s combined 2030 land-based commitment of 175 MtCO2e yr⁻¹.

Assumptions and exclusions: Losses to harvested wood products and operational emissions (harvest, milling, transport) were excluded from emissions accounting; analysis assumes historical flux patterns persist through 2030 and does not model leakage. Land governance constraints and administrative buffers were incorporated where mappable; some protections not captured in global/national protected area databases were mapped from regulations.

• Extent of NCS opportunity: 40% of the coastal temperate rainforest ecoregion (13 Mha) is potentially available for NCS action (NCS1), containing 4,900 ± 640 MtCO2e of aboveground forest carbon, representing 45% of regional and 0.5% of global aboveground forest carbon stocks.
• Carbon balance (2001–2021): Average annual carbon sequestration of 130 MtCO2 yr⁻¹ versus average annual carbon losses of 91 MtCO2e yr⁻¹, yielding a net sink of 36 MtCO2 yr⁻¹ in NCS1 forests. Average carbon density of NCS1 forests is 370 MgCO2e ha⁻¹, higher than 90% of the world’s forests.
• 2030 mitigation potential: A 10% reduction in average annual forest carbon losses via improved forest management and conservation could reduce emissions by about 9.1 MtCO2e yr⁻¹, equivalent to 5.2% of the combined 2030 land-based climate commitments of the U.S. and Canada (175 MtCO2e yr⁻¹).
• Ownership patterns: Of carbon stocks available for NCS action, 61% are on public lands, 32% on private lands, and 7% on Indigenous ownerships. Public lands generally have higher carbon density than private lands.
• Jurisdictional insights: • Oregon and Washington show the largest recent emissions from forestry on NCS1 lands, with average annual carbon losses of 45 MtCO2e yr⁻¹ (OR) and 27 MtCO2e yr⁻¹ (WA) from 2001–2021. • Private lands dominate losses: in Oregon, private NCS1 lands contributed 82% of average annual losses; in Washington, private lands contributed 78%, with industrial private forests showing net losses of 3.8 MtCO2e yr⁻¹ and private non-industrial forests showing a net gain of 5.7 MtCO2e yr⁻¹. • Remaining private carbon stocks available for action are substantial in WA and OR (640 ± 8 MtCO2e and 630 ± 31 MtCO2e, respectively), together ~26% of NCS1 stocks; public forests in these states contribute an additional ~18% of NCS1 stocks. • British Columbia manages the largest share: 5.1 Mha (40% of NCS1 stocks), totaling ~2000 ± 610 MtCO2e, with average annual carbon losses of 11 MtCO2e yr⁻¹. Privately managed BC forests (280 ± 130 MtCO2e) lose ~7.0 MtCO2e yr⁻¹ on average. Indigenous lands in BC hold ~190 ± 100 MtCO2e and are expanding as treaties progress. • Alaska: Indigenous landowners manage most NCS1 area and stocks (~110 ± 15 MtCO2e) with the highest cumulative annual losses (0.6 MtCO2e yr⁻¹). Federal Tongass National Forest has the highest carbon densities in the state (~360 MgCO2e ha⁻¹). State lands exhibit the highest per-area loss rates.
• Fire context: Coastal temperate rainforests have lower relative forest cover loss to fire (13% of forest cover loss from 2001–2021) and high carbon densities compared to other ecozones, supporting durability of mitigation relative to many boreal and drought-prone forests.
• Model comparisons: The global AGB model predicts higher stocks than national/regional models at jurisdictional scales—by ~23% (BC), 15% (WA), and 7% (OR)—and higher than the Pacific Northwest regional model by ~13% (WA) and 2% (OR), with stronger convergence at southern latitudes.

Integrating land management restrictions with high-resolution carbon stock and flux models yields a generalizable, decision-oriented framework to identify where forest-based NCS are both suitable and feasible. This approach reduces overestimation of opportunities that can arise when protections and management constraints are ignored; for example, 30% of forests in the study area are administratively protected from harvest yet often not represented in global and national protected area datasets. The availability of semi-annual flux data and stock estimates with uncertainty at fine spatial resolution supports annual, ecoregional-scale NCS assessments. The results validate use of global AGB maps in regions lacking detailed local models (e.g., Alaska), given their correspondence with national and regional estimates. Comparisons with USFS FIA inventories indicate similar ownership-driven patterns of flux, despite differences in magnitudes due to methodological and data coverage differences. The analysis suggests that a modest, historically grounded reduction in harvest-related losses (10%) can deliver meaningful emissions reductions (9.1 MtCO2e yr⁻¹) by 2030, contributing notably to national land-based targets. Implementation will be local and stand-scale, requiring tailored silvicultural strategies, community engagement, and alignment with economic and policy contexts. Rural and Indigenous communities are pivotal co-producers of NCS projects that meet both carbon and community-defined objectives; equity and consent are central. While coastal temperate rainforests exhibit relatively low wildfire-driven reversal risks, other disturbances (insects, wind) and socioeconomic drivers must be considered. Coarse-scale driver analyses corroborate that forestry is the dominant cause of loss within NCS1 areas (~78%), reinforcing the focus on improved management and conservation.

The study presents a replicable spatial framework that overlays land ownership and management designations with high-resolution forest carbon stock and flux models to map feasible forest-based NCS opportunities. Applied to the coastal temperate rainforests of western North America, the analysis identifies 13 Mha available for action holding 4,900 ± 640 MtCO2e and shows that a conservative 10% reduction in harvest-related carbon losses could reduce emissions by ~9.1 MtCO2e yr⁻¹, equating to 5.2% of the combined 2030 land-based climate commitments of the U.S. and Canada. The approach supports strategic planning, policy design, and investment decisions, especially when co-produced with rural and Indigenous communities to ensure equitable, durable outcomes. Future work should evaluate specific stand-scale management actions, incorporate additional NCS pathways (e.g., restoration and enhanced growth of regenerating forests), explore synergies and trade-offs among NCS strategies, refine flux estimates where regional data allow, and improve representation of evolving Indigenous governance and land rights.

• Emissions accounting excludes transfers to harvested wood products and operational emissions (harvesting, milling, transport), potentially underrepresenting total sectoral emissions.
• Flux uncertainty estimates were not available at this analysis level; stock and flux magnitudes vary across methods (remote sensing vs. plot-based), geographies, and datasets.
• Older AGB datasets may misestimate stocks in recently harvested areas; results should be treated as iterative and validated with inventories (e.g., FIA) where possible.
• The scenario assumes historical flux patterns persist through 2030 and does not model market or policy-induced leakage; large-scale implementation feasibility is not assessed.
• Mapping of land use restrictions relies on available datasets and regulatory proxies (e.g., buffered hydrography); governance strength, enforcement, and complex tenure arrangements may limit applicability elsewhere.
• Indigenous lands are represented using governmentally recognized ownerships; this underrepresents traditional territories and evolving governance and consent processes.
• Disturbance risks (e.g., insects, wind) remain and could affect carbon durability despite lower relative wildfire risks in the region.
• Climate benefits of IFM and conservation vary by site productivity and practice; per-hectare effects are not resolved in this ecoregional assessment.