Main

Natural climate solutions (NCS), a suite of land-management options including protecting, restoring and sustainably managing ecosystems1, provide readily implementable approaches to sequester carbon in terrestrial pools and/or reduce GHG emissions2. NCS are a complement to, not a substitute for, emissions reductions related to decreased use of fossil fuels and decarbonization of energy and transportation sectors. They are expected to slow the rate of global warming in the next 10–40 years, in parallel with the development of mitigation technologies and energy transformation3.

Following the first submission in 2015, China submitted its updated Nationally Determined Contributions (NDCs) in 2021. This included a ‘net-zero’ commitment, with a peak in CO2 emissions before 2030 and net carbon neutrality being achieved by 20604. Over the past decades (especially since 1999), China has strengthened its efforts in ecological restoration and protection5. While not all the projects were originally designed for carbon sequestration and climate change mitigation, collectively they made a substantial contribution to China’s terrestrial ecosystem carbon sink6,7,8. Some adaptive action strategies such as China’s Achievements, New Goals and New Measures for Nationally Determined Contributions9 and Master Plan of Major Projects of National Important Ecosystem Protection and Restoration (2021–2035)5 have been released, proposing a series of strategies for ecosystem management in the next decades to support implementation of the NDCs. So far, the contributions to emissions reductions from recent projects equivalent to NCS activities have not been fully quantified. More important, the mitigation potential of future NCS activities and their contribution to China’s climate goals is largely unknown2.

In this Article, we quantified 16 distinct pathways (see definitions and descriptions in Supplementary Tables 1 and 2) to provide an integrated evaluation of the climate mitigation capacity of previous projects during 2000–2020 and the future NCS potentials during 2020–2030 and 2020–2060 by conserving, restoring and improving management of forests, croplands, grasslands and wetlands in China. The evaluation of the mitigation capacity of previous projects for 2000–2020 is a foundation to understand the historical contribution and a baseline for setting feasible boundaries for future additional NCS pathway projections (Supplementary Tables 1 and 2). We first provide estimates of the historical mitigation capacity of projects during 2000–2020, as well as the future ‘maximum additional mitigation potential’ (MAMP) of NCS by additional stewardship options under biophysical constraints in the next 10 (by 2030) and 40 (by 2060) years, measured in CO2 equivalents per year (CO2e yr−1). We then constrain NCS levels by mitigation cost thresholds. We note that the essence of NCS is to seek new opportunities and approaches with additionality over the baseline (2000–2020 in this study).

Overall NCS potential

During 2000–2020, the total mitigation capacity of the nine pathways related to previous projects was 0.6 (0.5–0.7) PgCO2e yr−1 (95% confidence interval (CI)) (Fig. 1a), counterbalancing 8% of the mean annual fossil CO2 emissions in the same period (7.5 PgCO2 yr−1), in which the increased carbon sequestration from ecosystems (0.5 PgCO2 yr−1, excluding CH4 and N2O) is about 54% (28–80%) of the overall land carbon sink in China (0.18–0.51 PgC yr−1) (ref. 10). During 2020–2030, the maximum potential of NCS (MAMP) of the 16 pathways is projected at 0.6 (0.3–1.0) PgCO2e yr−1 (Fig. 1b), which is 6% of the mean annual CO2 emissions during the same period (assuming the industrial CO2 emissions at year 2030 to be around 10–12 PgCO2 yr−1) (ref. 11). The MAMP is projected at 1.0 (0.6–1.4) PgCO2e yr−1 during 2020–2060 (Fig. 1c).

Fig. 1: Historical mitigation capacity and future NCS potential in China.
figure 1

a, Historical capacity during 2000–2020. b,c, Future potential during 2020–2030 (b) and 2020–2060 (c). Light to dark colours of the bars indicate mitigation levels at the mitigation costs of US$10, US$50 and US$100 (MgCO2e)−1, respectively. Black lines represent the 95% CI. Ecosystem service benefits (air, biodiversity, soil and water) linked with each NCS pathway are indicated by coloured short bars. Pathway names: Reforestation (RF), Avoided Forest Conversion (AVFC), Improved Plantation (IMP), Natural Forest Management (NFM), Fire Management (FM), Biochar (BIOC), Cover Crops (CVCR), Cropland Nutrient Management (CRNM), Improved Rice Cultivation (IMRC), Avoided Grassland Conversion (AVGC), Grazing Optimization (GROP), Grassland Restoration (GRR), Avoided Coastal Wetland Impacts (AVCI), Avoided Peatland Impacts (AVPI), Coastal Wetland Restoration (CWR), Peatland Restoration (PTR). See Supplementary Tables 1 and 2 for pathway definitions and extent of application. Note only nine pathways have mitigation capacities during 2000–2020, and the other seven pathways (in grey) do not have the historical capacities as no activities have been taken before 2020.

Full size image

The realization and importance of diverse NCS pathways comprehensively reflects a nation or region’s stewardship of certain ecosystems, land-cover history, status of the ecosystems, land ownership, human intervention, technological progress and governance12. In the context of China, the land area that can be used for restoration (for example, reforestation, grassland and wetland restoration) (Extended Data Figs. 1 and 2), the technology and land practices that can be applied (for example, water and nutrient management in croplands and grazing management in rangelands) and the land-transfer risk control (for example, avoiding loss of natural land cover to cropland or building) are the direct determinants of ecosystem restoration, management and protection, respectively1,13.

NCS potential by pathway

Historical and future fluxes of individual pathways have contributed to the overall estimates presented in the preceding (Fig. 1a–c and Supplementary Tables 1–3).

Forest ecosystems

Forest pathways are most important for the historical mitigation capacity and future potential in China (Fig. 1a–c and Supplementary Table 3). During 2000–2020, the carbon sink in managed forests (including reforestation, natural forest management and fire management) was estimated to be 446 TgCO2 yr−1, contributing 73% of the total flux of the nine NCS pathways (Fig. 1a). Of this total, more than 55% was contributed by reforestation (247 TgCO2 yr−1 on an area of 44.7 Mha) and 42% by natural forest management (restricted timber logging) (189 TgCO2 yr−1 on an area of 41.3 Mha). Fire management made a small contribution of 3% (10 TgCO2 yr−1), in which 9 TgCO2 yr−1 was contributed by fire prevention and 1 TgCO2 yr−1 by prescribed fire.

In the future, the total MAMP from the five forest pathways (including two new pathways) would be 436 TgCO2 yr−1 by 2030 and 627 TgCO2 yr−1 by 2060, representing 64–68% of the total NCS of the 16 pathways (Fig. 1b,c and scenario setting in Supplementary Table 2). Considering the constraints of water stress, the farmland red line (the minimum acceptable arable land area in China), urbanization and the 26% forest coverage goal by 2035 (similar to the historical value in 1700)14,15, the maximum additional area for future reforestation after 2020 will be limited to be 31.8 Mha (Extended Data Figs. 2 and 3), similar to the estimate in ref. 16, and the MAMP would be 77 TgCO2 yr−1 and 235 TgCO2 yr−1 by 2030 and 2060, respectively (assuming that 21.7 Mha be reforested by 2030 and 31.8 Mha by 2060). The lower NCS potential of reforestation during 2020–2030 compared with the historical capacity in 2000–2020 is due to the lower carbon flux rate in the 10 yr forest and the smaller reforested area. The magnitude of CO2 sequestration during 2020–2060 is comparable to the historical value, with the forests getting older (Fig. 1a–c and Supplementary Tables 2 and 3). The MAMP of natural forest management would be almost double (336 TgCO2 yr−1) within the next 40 years because a larger area of natural forests (79.1 Mha) will be restricted from logging (clearcutting of natural forests for fuelwood and timber is prohibited by law in China since 2020), although the carbon sequestration rate of these unlogged forests would slightly decrease as they would shift from younger (mean age 43 years in 2010) to older (mean age 63–93 years in 2030–2060) age classes with a lower uptake rate (Extended Data Fig. 4)17. Natural forest management would account for 77% of the total NCS of forest pathways by 2030 and 54% by 2060, when reforestation will have a higher NCS potential (37%) compared with the historical period. Only 5–9% of the total CO2 flux resides in improved plantations, avoided forest conversion and fire management. Improved plantations by reducing harvest and improving wood residue treatments could contribute a net sink of 17 and 49 TgCO2 yr−1 by 2030 and 2060, respectively (Fig. 1b,c). The risk of forest conversion to other land uses would be relatively small (0.01 Mha yr−1). Reducing conversion from forest to cultivated or developed land will contribute 6 and 7 TgCO2 yr−1 of emissions reduction by 2030 and 2060, respectively. The mean annual burnt area due to wildfire was greatly reduced (from 0.89 Mha yr−1 during 1949–1999 to 0.12 Mha yr−1 during 2000–2017) (Extended Data Fig. 5) since the implementation of a fire-prevention policy in 198718, and future wildfires may also be maintained at a low level (<1 TgCO2 yr−1).

Forest volume is estimated to increase by 8.26 billion m3 (equivalent to 7.34 PgCO2 of carbon sequestration) by 2030 over the 2000 level due to reforestation and natural forest management according to our projections, indicating that achieving the goal specified in China’s NDCs (6.0 billion m3 over the 2005 level)4 is achievable.

Cropland ecosystems

Cropland NCS showed promising opportunities for climate mitigation via improved cropland management. During 2000–2020, the total mitigation from the three cropland pathways was 84 TgCO2e yr−1 (Fig. 1a), 14% of the overall historical flux of the nine NCS pathways. Cropland nutrient management decreased N2O emissions by 79 TgCO2e yr−1 from reduced nitrogenous fertilizer use and improved fertilizer production technology (Supplementary Table 3 and Extended Data Fig. 6). Cover crops (on 4.5 Mha of winter fallow) contributed 5 TgCO2 yr−1 through soil carbon sequestration. The mitigation capacity of biochar was minimal (0.5 TgCO2 yr−1) as the amount of crop residues used for biochar production is small (0.11% of the total of 708 Mt yr–1).

In the future, the MAMP of the cropland pathways (including a new pathway of improved rice cultivation) could reach 104 TgCO2e yr−1 by 2030 and 223 TgCO2e yr−1 by 2060 (scenario setting in Supplementary Table 2) (Fig. 1b–c). These values are 1.2 and 2.6 times as large as the historical value, respectively, contributing 16% and 23% to the overall NCS potential of the 16 pathways. Specifically, the potential additional N2O emissions reduction through improved N management could be 46 TgCO2e yr−1 by 2030 and 153 TgCO2e yr−1 by 2060, assuming that our scenario of reducing N fertilizers by 13–49% would not affect crop yield19. The contribution from cover crops could increase to be 25 TgCO2 yr−1 if the maximum area of 15.1 Mha of winter fallow is applied with cover crops. Biochar would remain as a less important NCS pathway (0.1 TgCO2 yr−1). By comparison, rice cultivation is extensive in China (30.5 Mha), and improved rice cultivation as a new NCS pathway could contribute 33–44 TgCO2e yr−1 in CH4 emissions reductions in the next 10-40 years through improved soil management and drainage processing20.

Grassland ecosystems

Grazing optimization and grassland restoration was estimated to be a net sink of 80 TgCO2 yr−1 during the period 2000–2020, contributing 13% to the overall mitigation capacity of the nine pathways (Fig. 1a), of which 59 TgCO2 yr−1 was from grazing optimization (fencing or sowing on an area of 99.6 Mha (ref. 21)) and 21 TgCO2 yr−1 from grassland restoration (on 5.8 Mha cropland) (Fig. 1a and Supplementary Tables 2 and 3). In the future, the total additional MAMP from grassland ecosystems is estimated at 68–82 TgCO2 yr−1 during 2020–2060 (8–11% of the overall NCS potential) (Fig. 1b,c and scenario setting in Supplementary Table 2). Grassland grazing optimization could be applied on a smaller area in the future (15.6 Mha by 2030 and 62.4 Mha by 2060) and consequently would contribute a smaller NCS potential of 8 TgCO2 yr−1 and 29 TgCO2 yr−1, respectively. For grassland restoration in cropland over steep slopes, the additional area available in the future would be only 0.54 Mha (Extended Data Figs. 1 and 2). Thus, the MAMP of grassland restoration would remain constant at 3 TgCO2 yr−1 during 2020–2060. Cropland expansion is the major cause of grassland conversion in China at a rate of 0.53 Mha yr−1. Avoided grassland conversion to cropland will provide a new mitigation opportunity, reducing an additional CO2 emission of 50–57 TgCO2 yr−1 (Fig. 1b,c).

Wetland ecosystems

The mitigation capacity from wetland projects during 2000–2020 was negligible (<1 TgCO2e yr−1) compared with the other pathways (Fig. 1a), but shows great potential as NCS in the next decades (35 and 52 TgCO2e yr−1 by 2030 and 2060, respectively (Fig. 1b,c), contributing 4–5% to the overall NCS potential of the 16 pathways), especially for peatland (Supplementary Table 3). Avoided peatland loss could potentially reduce CO2 emissions 24 TgCO2 yr−1 over the next 40 years (on 0.09–0.38 Mha of area). Peatland restoration by soil re-wetting could contribute an additional 4 and 14 TgCO2e yr−1 mitigation potential (the net effect of decreasing CO2 emissions and increasing carbon accumulation but increasing CH4 emissions) by 2030 and 2060 (0.05 and 0.19 Mha), respectively (Fig. 1b,c). Coastal wetland restoration could sequester 2 and 4 TgCO2 yr−1 (0.11 and 0.38 Mha). Avoided coastal wetland impacts (including salt marshes and sea grass beds) could reduce CO2 emissions 6 and 10 TgCO2 yr−1 by 2030 and 2060 (0.10 and 0.41 Mha), respectively. Although the loss rate of coastal ecosystems and wetlands is small (<0.01 Mha yr−1), they need extra attention because the carbon flux per unit area released from wetland loss is substantial (Fig. 1b,c and Supplementary Tables 2 and 3).

Spatial distribution of the NCS

The climate mitigation from the 16 NCS pathways shows large spatial differences across the country. Inner Mongolia, Heilongjiang, Sichuan and Yunnan are the top four provinces with the highest levels of historical achieved capacity and future potential (Fig. 2). Except for some provinces in Northwest and East China, natural forest management and/or reforestation contributes the most to the overall NCS. For Xinjiang and Qinghai, grazing optimization had the biggest contribution to the historical flux, and grassland and peatland management will be important NCS opportunities in the future. In some provinces in Central and East China (including Henan, Hubei, Hunan, Shandong, Anhui, Jiangxi and Jiangsu), cropland nutrient management and improved rice cultivation are also important options for future NCS. Improved plantations are non-negligible in Guangxi province.

Fig. 2: Historical mitigation capacity and future NCS potential by each NCS pathway across the provinces of China.
figure 2

The numbers above each of the three bars for each province represent the total flux (TgCO2e yr−1) in the periods of 2000–2020, 2020–2030 and 2020–2060, respectively. AH, Anhui; BJ, Beijing; CQ, Chongqing; FJ, Fujian: GD, Guangdong; GS, Gansu; GX, Guangxi; GZ, Guizhou; HB1, Hebei; HB2, Hubei; HLJ, Heilongjiang; HN1, Hainan; HN2, Henan; HN3, Hunan; IM, Inner Mongolia; JL, Jilin; JS, Jiangsu; JX, Jiangxi; LN, Liaoning; NX, Ningxia; QH, Qinghai; SC, Sichuan; SD, Shandong; SH, Shanghai; SX1, Shaanxi; SX2, Shanxi; TJ, Tianjin; XJ, Xinjiang; XZ, Xizang; YN, Yunnan; ZJ, Zhejiang. Base map from the National Earth System Science Data Center, National Science and Technology Infrastructure of China (http://www.geodata.cn).

Full size image

At the national level, the dominant type of the NCS was restoration and improved management before 2020, but it would shift towards improved management in future decades, with protection becoming more important (Fig. 3a–c, the type of NCS defined in Methods). At the provincial level, ten NCS types were further identified as the predominant options, including six single types and four mixed or combined types (Fig. 3e,f), indicating the need for diverse NCS activities.

Fig. 3: NCS pathway types at the provincial level for the three periods.
figure 3

a–c, Proportional distribution of NCS types during 2000–2020 (a), 2020–2030 (b) and 2020–2060 (c). Four types of pathways are classified: ‘Restoration’ (>50% restoration), ‘Protection’ (>50% protection), ‘Management’ (>50% management) and ‘Mixed’ (no pathway type contributes the majority). The red star represents the national barycentre of NCS types. d–f, The provinces are further classified into ten groups on the basis of the ecosystem type of pathway contributing the most mitigation potential during 2000–2020 (d), 2020–2030 (e) and 2020–2060 (f). Mixed types are not specified ecosystem types as they contain many combinations. Base maps in d–f from the National Earth System Science Data Center, National Science and Technology Infrastructure of China (http://www.geodata.cn).

Full size image

Cost effectiveness

By using marginal abatement cost (MAC) curves, the climate mitigation at three cost thresholds of each NCS pathway was estimated. During 2000–2020, only 125 TgCO2e yr−1 (20% of the total) was achieved at a mitigation cost below US$10 (MgCO2e)−1 (Fig. 1a–c). In the future, 165 TgCO2e yr−1 (26%) and 298 TgCO2e yr−1 (30%) out of the total NCS will be achievable by 2030 and 2060, respectively, at US$10 (Mg CO2e)−1. At a mitigation cost of US$50 (MgCO2e)−1, 354 TgCO2e yr−1 (58%), 390 TgCO2e yr−1 (62%) and 641 TgCO2e yr−1 (65%) of the total NCS are available during 2000–2020, 2020–2030 and 2020–2060, respectively. For mitigation costs up to US$100 (Mg CO2e)−1, 406 TgCO2e yr−1 (66%), 573 TgCO2e yr−1 (90%) and 898 TgCO2e yr−1 (91%) are achievable. Natural forest management is the most cost-effective option at all of the three mitigation costs and will achieve the highest MAMP (336 TgCO2 yr−1) at US$100 (Mg CO2e)−1 during 2020–2060 (Fig. 1a–c). The portion of achievable NCS at the three cost thresholds is similar to that of the United States but higher than that of Canada. The US total NCS is 1.2 PgCO2e yr−1 during 2015–2035 (19% of industrial CO2 emissions of the 1990–2018 level), 25%, 76% and 91% of which would be achieved at US$10, US$50 and US$100 (MgCO2e)−1, respectively22. For Canada, the total NCS is 78 TgCO2e yr−1 during 2021–2030 (100% of all heavy industry emissions of the 2018 level); 12%, 34% and 51% of which would be available at CAN$10, CAN$50 and CAN$100 (MgCO2e)−1, respectively23.

Besides climate mitigation, the NCS measures bring co-benefits, including improving biodiversity, air purification, soil enrichment and water retention24,25 (Fig. 1). The value of these co-benefits is not accounted for in our estimation of cost effectiveness of the NCS, but they will save economic costs and improve human well-being26.

Calibration upon the global assessment for China

A relative comparison between our estimates and those in ref. 1 can be made in terms of implementation extent and flux rate (per unit area) within a given period (Extended Data Fig. 7 and Box 1). In ref. 1, the estimated MAMP of reforestation for China is as high as 1,257 TgCO2e yr−1 (2010–2030), five times our estimate (247 TgCO2 yr−1, 2000–2020), mainly because they used a much higher value for carbon sequestration rates (4.14 MgC ha−1 yr−1). They assumed that 25–30% of total NCS by reforestation can be achieved at US$100, that is, 319 TgCO2 yr−1 (refs. 27,28). We estimated that 55% (136 TgCO2 yr−1) of the total NCS (247 TgCO2 yr−1) can be realized at US$100 during 2000–2020 on the basis of our MAC curves using actual data from regions that implemented reforestation in China (Extended Data Fig. 7). The CO2 sink of natural forest management in 2000–2020 is 189 TgCO2 yr−1 in this study, versus only 35 TgCO2 yr−1 in ref. 1. This large difference is because we separated natural forests into different age classes with different sequestration rates (40–90 years) to obtain a weighted mean carbon sequestration rate (0.71–1.25 MgC ha−1 yr−1), whereas ref. 1 set sequestration rates to a mean value (0.14–0.39 MgC ha−1 yr−1) uniform across all forests. Much higher NCS estimates of biochar are reported by the Natural Conservancy27 (65 TgCO2e yr−1 at the price of US$100) and ref. 29 (500 TgCO2e yr−1), who assumed that 14% or 100% of China’s crop residues can be used to produce biochar in the future, but this fraction is actually less than 1%, and it will be a gradual process to transform crop residues to produce a large amount of biochar, requiring a change of regime of crop residue use. Based on this, biochar NCS is very small (<1 TgCO2e yr−1) in both historical and future estimates in our study (Extended Data Fig. 7). The differences between our study and the global studies indicate that baseline setting as well as the use of regional data can greatly impact NCS assessment. Thus, well-calibrated national assessments are necessary to improve upon rough global estimates. For further comparative analysis, see the corresponding sections in each of the individual pathways in the Supplementary Information.

The implications

To achieve the future maximum potential of NCS and enhance the co-benefits of multiple ecosystem services, it is critical that integrated land planning and management optimization is applied at different jurisdictional levels and spatial scales, especially since the NCS pathways will be implemented on such a large land area nationwide (Extended Data Figs. 1 and 2). Specifically, overall land-use planning should focus on systematic layout (total area control) of land restoration30 and replacement31 at the national level on the basis of accurate land surveys, emphasizing coordination between restoration/protection and management at the provincial or regional level and highlighting livelihood and realization of social justice at the local level by implementing ecological compensation26. The case of the Loess Plateau of China—the pioneer area of the Grain for Green Project—has proved that food security can be maintained through improved management practices such as improving nutrient use efficiency, water-saving irrigation and construction of terraces and check dams32 while at the same time capturing carbon in vegetation and soil and reducing soil sediment in the Yellow River33,34. We also suggest that the government should set a percentage target of NCS based on the resource status of each region/province, continuing to strictly implement ecosystem protection policies as the damage to ecosystems will outweigh the benefits of managing and restoring without protection35. Alongside protection, improving management of ecosystems and promoting ecological restoration should be implemented for maintaining the long-term benefits of NCS. Our spatial maps clearly show different opportunity options in each province throughout the country. At the national or provincial level, since natural resource abundance and economic capital are limited, it is important for policymakers to transform their thinking from ‘expanding the extent of ecosystems’ to ‘improving the benefits per unit land area’, maintaining the overall strength of investment, fair and effective fund distribution (for example, via improving ecosystem services payment system)36 and management and technological creativity (for example, improving the fertilizer production process or developing digital fencing).

Challenges and next steps

Climate change and ecosystem spatial shifts may bring great uncertainties to the carbon sequestration rates estimated on the basis of the existing spatial locations of the samples. The feedback between climate change and ecosystem change (such as forest albedo, particularly for the temperate zone and the boreal forest on the northeast) may also reduce the mitigation potential of the NCS37. These aspects, related to long-term ecosystem change and climate feedback, should be further explored for better estimating future NCS potential. In addition, due to limited data availability, biomass and soil carbon or other GHGs were not fully considered. Other pathway options, including actions to avoid carbon loss due to natural disturbances (such as pests, diseases and climate extremes), or NCS potential from managing the urban green infrastructure, livestock and aquaculture sectors, need be quantified. In addition, the inclusion of co-benefits brought by NCS will greatly change the assessment of cost effectiveness. Finally, the implementation of an NCS plan will also be affected by national economic development level, population change, policies and investment, pandemic events (such as COVID-19) and other socioeconomic factors.

In this work, we showed that China’s NCS could counterbalance 6% (11–12% if including legacy effect) of industrial carbon emissions between now and 2030. This modest proportion highlights that energy transformation and the rapid application of low-carbon technologies are urgent priorities for climate mitigation. Nevertheless, NCS pathways can still make an important contribution through additional carbon sequestration or avoided GHG emissions, buying time for the development and implementation of new technologies. Multilevel governance strategies that consider the spatial heterogeneity of the NCS pathways are needed to increase the synergies among multiple NCS pathways for realizing climate mitigation and co-benefits of ecosystem services.

Methods

This study presents a comprehensive quantification of carbon sequestration as well as CO2/CH4/N2O emissions reductions from terrestrial ecosystems based on multiple sources of data from literature, inventories, public databases and documents. The pathways considered ecosystem restoration and protection from being converted into cropland or built-up areas, reforestation, management with improved nitrogen use in cropland, restricted deforestation, grassland recovery, reducing risk from forest wildfire and others. Here we describe the cross-cutting methods that apply across all 16 NCS pathways. The definitions, detailed methods and data sources for evaluating individual pathways can be found in the Supplementary Information.

Cross-cutting methods

Baseline setting

We set 2000 as the base year because the large-scale national ecological projects, such as the Grain for Green Project, were started since then. We first evaluate the historical mitigation capacity during 2000–2020, which is the first 20 years of implementing the projects. From this procedure we can determine how much mitigation capacity has been realized through the previous projects in the past two decades and to what extent additional actions can be made after 2020. Relative to the baseline 2000–2020, we then evaluate the maximum potentials of the NCS mitigation in the future 10 (2020–2030) and 40 (2020–2060) years, corresponding to the timetable of China’s NDCs: carbon peak before 2030 and carbon neutrality by 2060.

The settings of baseline in this study are different from the existing assessments (2000s–2010s as a baseline and 2010–2025/2030/2050 as scenarios)1,22,23,27,28. Baseline sets the temporal and spatial reference for NCS pathway scenarios, which may have a great impact on the NCS estimates. Notably, NCS actions during 2000–2020 will have a great impact in the future periods, which we refer to as the ‘legacy effect’. The legacy effect itself, mainly reforestation, is independent of being assessed, but it is conceptually attributed to natural flux and excluded from future NCS potential estimates.

Maximum potential

The MAMP refers to the additional CO2 sequestration or avoided GHG emissions measured in CO2 equivalents (CO2e) at given flux rates in a period on the maximum extent to which the stewardship options are applied (numbers are expressed as TgCO2e yr−1 for individual pathways and PgCO2e yr−1 for national total) (Extended Data Fig. 1 and Supplementary Table 2). ‘Additional’ means mitigation outcomes due to human actions taken beyond business-as-usual land-use activities (since 2020) and excluding existing land fluxes not attributed to direct human activities1. The MAMP of CH4 and N2O are accounted by three cropland and wetland pathways (cropland nutrient management, improved rice cultivation and peatland restoration). We adopt 100 yr global warming potential to calculate the warming equivalent for CH4 (25) and N2O (298), respectively38,39 because these values are used in national GHG inventories, although some researchers have argued that using the fixed 100 yr global warming potential to calculate the warming equivalents may be problematic because they cannot differentiate the contrasting impacts of the long- and short-lived climate pollutants39. Because the flux rate of the GHG by ecosystems may vary with the time of recovery or growth, the MAMP may also change for different periods even given the same extent.

The ‘maximum’ is constrained by varied factors across the NCS pathways. We constrain forest and grassland restoration by the rate of implementation, farmland red line and tree surviving rate (Extended Data Fig. 2). Surviving rate here is the ratio of the area with increased vegetation cover due to reforestation to the total reforestation area. The farmland red line refers to ‘the minimum area of cultivated land’ given by the Ministry of Land and Resources of China. It defines the lowest limit, and the current red line is ~120 Mha. It is a rigid constraint below which the total amount of cultivated land cannot be reduced. From this total amount, there is provincial farmland red line. This red line sets a constraint on the implementation of the NCS pathways associated with land-use change. We set the future scenario of farmland area that can be used for grassland or forest restoration on the basis of the provincial farmland red line. Basic farmland is closely related to national food security. By 2050, China’s population is predicted to decrease slightly, but with economic development, the per capita demand for food may increase40. We assume that the food production in the future can meet the food demand via increasing agricultural investment and technological advancement. The N fertilizer reduction scenario is set to be below the level 60%, under which crop yield is not significantly affected19, because N fertilizer is surplus in many Chinese croplands. For timber production, we assume that the demand for timber can be met if the production level is maintained at the level of 2010–2020 (83.31 million m3 yr−1). As deforestation of natural forests is 100% forbidden since 2020, the future timber will come mainly from tree plantations. For grazing optimization, we assume that livestock production is not affected by grassland fencing due to refined livestock management such as improving feed nutrient and fine-seed breeding41.

The areas of historical NCS implementation during 2000–2020 were estimated using statistical data, published literature and public documents, with a supplement from remote-sensing data. The flux rates were obtained either by directly using the values from multiple literature sources or from estimates using the empirical formulae. For the estimates of future NCS potential, the flux rate and extent of the pathway were determined on the basis of the baseline (2000–2020). The extent is assumed to be achieved by using the same rate but limited by the multiple constraints stated in the preceding unless the implementation scopes have been reported in national planning documents. We estimate the legacy effect by multiplying the implementation area in the past by the flux rates in the future two periods.

Saturation

The future mitigation potential that we estimate for 2030 and 2060 will not persist indefinitely because the finite potential for natural ecosystems to store additional carbon will saturate. For each NCS pathway, we estimate the expected duration of the potential for sequestration at the maximum rate (Supplementary Table 3). Forests can continue to sequester carbon for 70–100 years or more. Restored grasslands and fenced grasslands can continue to sequester carbon for >50 years. Forest-fire management and cover crops can continue to sequester carbon for 40–50 years or more. Sea grasses and peatlands can continue to sequester carbon for millennia. Avoided pathways do not saturate as long as the business-as-usual cases indicate that there are potential areas for avoided losses of ecosystems. In this case, sea grass and salt marsh would disappear entirely after 64 years, but it would be 100–300 years or more for forest, grassland and peatland.

Estimation of uncertainties

The extent (area or biomass amount) and flux (sequestration or reduced emission per area or biomass amount in unit time) are considered to estimate uncertainty of the historical mitigation capacity or future potential for each NCS pathway. We use the IPCC approaches to combine uncertainty42. Where mean and standard deviation can be estimated from collected literature, 95% CIs are presented on the basis of multiple published estimates. Where a sample of estimates is not available but only a range of a factor, we report uncertainty as a range and use Monte Carlo simulations (with normal distribution and 100,000 iterations) to combine the uncertainties of extent and flux (IPCC Approach 2). The overall uncertainties of the 16 NCS pathways were combined using IPCC Approach 142. If the extent estimate is based on a policy determination, rather than an empirical estimate of biophysical potential, we do not consider it a source of uncertainty.

MACs

The economic/cost constraints refer to the amount of NCS that can be achieved at a given social cost. The MAC curve is fitted according to the total publicly funded investment and total mitigation capacity or potential during a period. The MAC curves are drawn to estimate the historical mitigation or MAMP at the cost thresholds of US$10, US$50 and US$100 (MgCO2e)−1, respectively. The trading price in China’s current carbon market is ~US$10 USD (as the minimum cost43), and the cost-effective price point44,45 to achieve the Paris Agreement goal of limiting global warming to below 2 °C above pre-industrial levels is US$100 (as the maximum cost). A carbon price of US$50 is regarded as a medium value1,46. For the pathways of reforestation, avoided grassland conversion, grazing optimization and grassland restoration, we collected the statistical data of investments in China from 2000 to 2020 and estimated the affordable MAMP below the three mitigation costs. Due to data limitations, the points used for fitting the MAC curve are values for cost (invested funds) and benefit (mitigation capacity) in each of the provinces. We rank the ratio of benefit to cost in a descending order to obtain the maximum marginal benefit for MAC by assuming that NCS measures are first implemented in the region with the highest cost/benefit rate. We refer to the investment standard before 2020 as the benchmark and estimate the cost of each pathway for the future periods with discount rates of 3% and 5%, respectively. The social discount rate 4–6% is usually used as a benchmark discount value in carbon price studies in China compared with lower scenarios (for example, 3.6%)46,47. In a global study for estimating country-level social cost of carbon, 3% and 5% are used for scenario analysis48. Note that the mean value from the two discount rates was used in presenting the results. For the other pathways where investment data cannot be obtained, we refer to relevant references to estimate MAC. All the cost estimates are expressed in 2015 dollars, transformed on the basis of the Renminbi and US dollar exchange rate of the same year. The year 2015 represents a relatively stable condition of economic increase over the past decade (2011–2020) in China (the increase rate of gross domestic product (GDP) in 2015 is similar to the 10 yr mean). In the cases when the MAC curves exceed the estimated maximum potentials in the period, we identify the historical capacity or the MAMP as limited by the biophysical estimates.

Additional mitigation required to meet Paris Agreement NDCs

On 28 October 2021, China officially submitted ‘China’s Achievements, New Goals and New Measures for Nationally Determined Contributions’ (‘New Measures 2021’ hereafter) and ‘China’s Mid-Century Long-Term Low Greenhouse Gas Emission Development Strategy’ to the Secretariat of the United Nations Framework Convention on Climate Change as an enhanced strategy to China’s updated NDCs (first submission in 2015). The goal of China’s updated NDCs is to strive to peak CO2 emissions before 2030 and achieve carbon neutralization by 2060. It specified the goals to include the following: before 2030, China’s carbon dioxide emissions per unit of GDP are expected be more than 65% lower than that in 2005, and the forest stock volume is expected to be increased by around 6.0 (previously 4.5) billion m3 over the 2005 level. In the ‘New Measures 2021’9 and ‘Master Plan of Major Projects of National Important Ecosystem Protection and Restoration (2021–2035)’5, many NCS-related opportunities are proposed to consolidate the carbon sequestration of ecosystems and increase the future NCS potential, including protecting the existing ecosystems, implementing engineering to precisely improve forest quality, continuously increasing forest area and stock volume, strengthening grassland protection and recovery and wetland protection and improving the quality of cultivated land and the agricultural carbon sinks.

Industrial CO2 emissions

The historical CO2 emissions data from 2000 to 201749,50 are used as the benchmark of industrial CO2 emissions during 2000–2020. For future projections, we use the peak value of the A1B2C2 scenario (in the range of 10,000 to 12,000 Mt) in 2030 from ref. 11. We assume that CO2 emission increases linearly from 2017 to 2030.

Characterizing co-benefits

NCS activities proposed in the future measures or plans may enhance co-benefits. Four generalized types of ecosystem services are identified: improving biodiversity, water-related, soil-related and air-related ecosystem services (Fig. 1). Biodiversity benefits refer to the increase in different levels of diversity (alpha, beta and/or gamma diversity)51. Water, soil and air benefits refer to flood regulation and water purification, improved fertility and erosion prevention, and improvements in air quality, respectively, as defined in the Millennium Ecosystem Assessment52. The evidence that each pathway produces co-benefits from one or more peer-reviewed publications was collected through reviewing the literature (see the details for co-benefits of each pathway in Supplementary Information).

Mapping province-level mitigation

The data for extent of implementing forest pathways are obtained from the statistical yearbook and reported at the province level. To be consistent with the forest pathways, the other pathways were also aggregated to the provincial-level estimate from the spatial data. If the flux data were available in different climate regions, the provinces are first assigned to climate regions. When a province spans multiple climate zones, the weight value is set according to the proportion of area, and finally an estimated value of rate was calculated (for fire management, some grassland and wetland pathways). For the forest pathways, we first collected the flux-rate data from reviewing literature and then averaged these flux rates to region/province. The flux rates for reforestation and natural forest management were calculated separately by province and age group. Similarly, specified flux rates are applied for different times after ecosystem restoration or conversion for other pathways.

Classification of NCS types

Three types of NCS pathways were classified: protection (of intact natural ecosystems), improved management (on managed lands) and restoration (of native cover)35. In our study, four (AVFC, AVGC, AVCI, AVPI), eight (IMP, NFM, FM, BIOC, CVCR, CRNM, IMRC, GROP) and four (RF, GRR, CWR, PTR) NCS pathways were identified as protection, management and restoration types, respectively (Supplementary Table 1). These pathways can be further divided into groups of ‘single’ type or ‘mixed’ type according to their contribution to individual pathways. Specifically, in a certain area, when the mitigation capacity of a certain pathway accounts for more than 50% of the total, it is regarded as a single or dominant NCS type; if no single pathway accounts for more than 50%, it is a mixed type, named by the top pathways whose NCS sum exceeds 50% of the total mitigation capacity.