Protect young secondary forests for optimum carbon removal

nature climate change Article https://doi.org/10.1038/s41558-025-02355-5 Protect young secondary forests for optimum carbon removal Received: 2 September 2024 Accepted: 7 May 2025 Published online: 24 June 2025 Check for updates Nathaniel Robinson 1,2 , C. Ronnie Drever 3, David A. Gibbs Kristine Lister4,5, Adriane Esquivel-Muelbert 6,7, Viola Heinrich Philippe Ciais 10, Celso H. L. Silva-Junior 11,12, Zhihua Liu 13, Thomas A. M. Pugh 6,7,14, Sassan Saatchi13,15, Yidi Xu 10 & Susan C. Cook-Patton 16,17 , 4 , 8,9 Avoiding severe global warming requires large-scale removals of atmospheric carbon dioxide. Forest regeneration offers cost-effective carbon removals, but annual rates vary substantially by location and forest age. Here we generate grid-level (~1-km2) growth curves for aboveground live carbon in naturally regrowing forests by combining 109,708 field estimates with 66 environmental covariates. Across the globe and the first 100 years of growth, maximum carbon removal rates varied 200-fold, with the greatest rates estimated in ~20- to 40-year-old forests. Despite a focus on new forests for natural climate solutions, protecting existing young secondary forests can provide up to 8-fold more carbon removal per hectare than new regrowth. These maps could help to target the optimal ages and locations where a key carbon removal strategy could be applied, and improve estimates of how secondary forests contribute to global carbon cycling. With climate change intensifying globally1 and a narrowing window in which to act2, meeting the 1.5 °C warming target requires steep emissions cuts alongside large-scale atmospheric carbon dioxide (CO2) removals3. Natural climate solutions can provide cost-effective, scalable carbon removals4,5, with forest cover restoration as a particularly prominent strategy6–8. However, carbon removal rates can vary substantially by location and as forests age, meaning newly regenerating forests may not provide substantial carbon removal for years9. Thus, understanding this variation is critical for leveraging forest restoration for effective climate mitigation, and can guide policymakers and project developers as they integrate carbon removal strategies with other key objectives—such as biodiversity conservation, livelihood support and socio-economic priorities. Despite a common emphasis on tree planting10, resources are insufficient to plant trees at the required scale3,11. Instead, greater focus is needed on natural forest regeneration10,12, regrowth on cleared lands after disturbance8, which can be highly effective at capturing carbon and restoring biodiversity and other ecosystem services8,13–15. Existing estimates of potential carbon removal by natural forest regrowth fail to capture sufficient variation across space and stand age. The Intergovernmental Panel on Climate Change (IPCC) Tier 1 default removal rates distinguish only two secondary forest age classes: young (≤20 years) and old (21–100 years)16,17, at the level of continent and ecozone. Cook-Patton et al.8 improved the spatial resolution with a global 1-km2-resolution map for young (≤30 years) forests, but did not address how removals change as forests mature. Another recent The Nature Conservancy, Worthington, MA, USA. 2CIFOR-ICRAF, Nairobi, Kenya. 3Nature United, Toronto, Ontario, Canada. 4World Resources Institute, Washington, DC, USA. 5Nicholas School of the Environment Master of Environmental Management Program, Duke University, Durham, NC, USA. 6School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, UK. 7Birmingham Institute of Forest Research, University of Birmingham, Birmingham, UK. 8Remote Sensing and Geoinformatics Division, Helmholtz Centre for GeoSciences (GFZ), Potsdam, Germany. 9School of Geographical Sciences, University of Bristol, Bristol, UK. 10Laboratoire des Sciences du Climat et de l’Environnement, Université Paris-Saclay, Gif sur Yvette, France. 11Instituto de Pesquisa Ambiental da Amazônia - IPAM, Brasília, Brazil. 12Programa de Pós-Graduação em Biodiversidade e Conservação, Universidade Federal do Maranhão (UFMA), São Luís, Brazil. 13CTrees, Pasadena, CA, USA. 14Department of Physical Geography and Ecosystem Science, Lund University, Lund, Sweden. 15NASA Jet Propulsion Laboratory, Pasadena, CA, USA. 16The Nature Conservancy, Arlington, VA, USA. 17Smithsonian Environmental Research Center, Edgewater, MD, USA. e-mail: ; 1 Nature Climate Change | Volume 15 | July 2025 | 793–800 793 Article Spatial and temporal variation of carbon removal rates For each grid cell and forest age, we estimated annual carbon removal rates based on yearly changes in aboveground biomass. Rates typically start low, increase and then decline in older forests (Fig. 1a). To assess temporal variation, we calculated the difference (range) between each cell’s maximum and minimum annual removal rate over 100 years of regrowth. Globally, this range spans 0.006–4.61 MgC ha−1 yr−1 (mean ± s.d.; 0.51 ± 0.42 MgC ha −1 yr−1). The Mediterranean forests, woodlands and scrub biome shows the smallest range (0.19 ± 0.23 MgC ha−1 yr−1), indicating relatively stable rates through time, whereas tropical and subtropical moist broadleaf forests exhibit the greatest range (0.98 ± 0.45 MgC ha−1 yr−1). We also examined the maximum carbon removal rate per grid (Figs. 1b and 2a) and found that it varies more than 200-fold across Nature Climate Change | Volume 15 | July 2025 | 793–800 Removal rate (MgC ha–1 yr−1) a Annual removal rate versus age by biome 1.6 Biome Tropical and subtropical moist broadleaf forests Tropical and subtropical dry broadleaf forests Tropical and subtropical coniferous forests Mediterranean forests and woodlands Temperate broadleaf mixed forests Temperate conifer forests Tropical and subtropical savannas Temperate savannas Boreal forests/taiga 1.2 0.8 0.4 0 0 b 25 50 Age (years) 75 100 Maximum removal rate versus age at maximum rate by ecoregion 4 Biome Tropical and subtropical moist broadleaf forests Tropical and subtropical dry broadleaf forests Tropical and subtropical coniferous forests Removal rate (MgC ha–1 yr−1) effort18 showed removal rates declining through time, but only at the level of global forest types (for example, broadleaf deciduous forests). Recent remote sensing-based approaches integrate time since disturbance with biomass maps to estimate biomass accumulation as forests age19,20, but are constrained by <40 years of satellite data, which limits estimates in older stands, and rely on coarse space-fortime substitutions19. Consequently, current research limits our ability to spatially and temporally assess the medium- and long-term climate mitigation potential of allowing naturally regenerating forests to mature. To better understand atmospheric CO2 removal by natural forest regeneration, we mapped live aboveground carbon (AGC) density through time in stands aged 1–100 years using eight times more (...truncated)