Rising atmospheric greenhouse gas (GHG) concentrations, primarily CO₂, CH₄, and N₂O, are driving global warming towards irreversible consequences. Land-use change and ecosystem degradation have converted natural ecosystems from net GHG sinks to sources. Ecological restoration, assisting the recovery of degraded ecosystems, is proposed as a crucial climate stabilization strategy. Ambitious international targets exist for restoring significant land areas, making it imperative to understand the impact of restoration on GHG emissions to inform effective policies and improve IPCC guidance on GHG inventories. Forests, grasslands, and wetlands, covering substantial portions of the Earth's land surface, play significant roles in regulating the global carbon cycle. Previous studies on the effects of ecological restoration on GHG emissions have yielded inconsistent results, influenced by ecosystem type, restoration methods, and restoration age. This study aims to comprehensively quantify the impacts of ecological restoration on CH₄, N₂O, and net ecosystem CO₂ exchange (NEE) across forest, grassland, and wetland ecosystems at a global scale, explore the relationship between GHG emissions and restoration age, and identify key influencing factors.

Existing research indicates that forests act as net carbon sinks, but deforestation and disturbances counteract this effect. Afforestation and reforestation can alter biomass accumulation and soil properties, affecting GHG fluxes; studies show varied outcomes concerning CH₄ and N₂O emissions. Grasslands are generally efficient CH₄ and CO₂ sinks but N₂O sources; degradation impacts GHG emissions. Wetland ecosystems are efficient CO₂ sequesters, with drainage causing GHG emission shifts. However, the impacts of wetland restoration on GHG balances remain controversial, with studies reporting both net GHG sources and sinks. While numerous studies exist at local or regional scales, a comprehensive global analysis of restoration's effects on the three major GHGs across diverse ecosystems is lacking, hindering the refinement of IPCC guidelines and GHG inventories.

This meta-analysis compiled a global dataset from 253 peer-reviewed articles, focusing on GHG emissions from ecological restoration projects. The search encompassed multiple databases using keywords related to restoration, GHGs, and ecosystem types (forests, grasslands, wetlands). Inclusion criteria for studies involved field experiments with paired control sites or chronosequences, at least three replicates per treatment, yearly or growing-season measurements, and reporting of at least one GHG. The dataset included GHG fluxes (CH₄, N₂O, GPP, ER, NEE), environmental factors (longitude, latitude, MAT, MAP), restoration age, and soil properties (WT, ST, WFPS, Eh, BD, pH, SOC, TN, NH₄⁺, NO₃⁻). Data were extracted from publications, and standard deviations were estimated where missing. Meta-analysis employed Hedges' d to calculate weighted response ratios (RRd) between restored and control ecosystems, using MetaWin 3 software with a categorical random effects model and bootstrapping for 95% confidence intervals. Statistical analyses included ANOVA to compare GHG fluxes and soil variables, and mixed meta-regression to examine relationships between GHG fluxes and environmental factors. GWP was calculated to assess the overall greenhouse effect.

The meta-analysis revealed significant effects of ecological restoration on GHG emissions. Forest and grassland restoration significantly decreased CH₄ emissions (RRd of -2.3 and -1.6, respectively), increasing CH₄ uptake by 90.0% and 30.8%, primarily due to changes in soil properties like reduced bulk density and increased water-filled pore space. Conversely, wetland restoration significantly increased CH₄ emissions (RRd of 2.9), a 544.4% increase, mainly attributed to elevated water table depth. Forest and grassland restoration showed no significant effect on N₂O emissions, but wetland restoration significantly reduced them (RRd of -2.9), a decrease of 68.6%. Wetland restoration significantly reduced net ecosystem CO₂ exchange (NEE), shifting ecosystems to net CO₂ sinks within approximately 4 years. Forest NEE decreased with restoration age; afforestation and reforestation sites transitioned to net sinks in 3-5 years, whereas clear-cutting and post-fire sites took 6-13 years. Overall, forest, grassland, and wetland restoration decreased global warming potentials by 327.7%, 157.7%, and 62.0%, respectively, compared to controls. Analysis of soil properties showed that forest restoration increased SOC, NH₄-N, and DOC but reduced soil temperature, WFPS, and pH. Grassland restoration increased WFPS, soil moisture, and biomass, reducing bulk density and NO₃-N. Wetland restoration increased water table depth and SOC but decreased bulk density and redox potential. Significant correlations were found between GHG response ratios and specific soil properties across ecosystems. For instance, CH₄ response ratios were positively correlated with WFPS in forests and negatively correlated with bulk density in forests and grasslands; wetland CH₄ emissions were positively correlated with water table depth. N₂O emissions were positively related to NH₄⁺-N and NO₃⁻-N concentrations. Analyses of restoration age showed significant relationships with GHG fluxes, illustrating the time-dependent nature of restoration's effects on GHG emissions. GPP and ER exhibited positive exponential relationships with restoration age, with GPP/ER surpassing 1 (indicating a CO₂ sink) after various timeframes depending on restoration type.

The findings highlight the substantial potential of ecological restoration for GHG mitigation. The contrasting responses of CH₄ to restoration across ecosystem types underscore the importance of considering the specific context when designing and implementing restoration projects. Increases in CH₄ uptake in restored forests and grasslands are linked to changes in soil properties that promote CH₄ oxidation. The significant increase in CH₄ emissions from restored wetlands is due to elevated water table depths, creating anaerobic conditions favorable for methanogenesis. The reductions in N₂O emissions from restored wetlands are likely due to decreased soil NH₄⁺ concentrations, reducing substrates for microbial processes producing N₂O. The observed temporal patterns of NEE demonstrate the time-dependent nature of restoration's effect on the carbon cycle. Early-stage restoration may involve CO₂ release due to decomposition, but the establishment of vegetation and increased photosynthesis eventually lead to net CO₂ uptake. The significant reduction in GWP across all restoration types affirms that ecological restoration can be a valuable tool for climate change mitigation. These results contribute to a more nuanced understanding of GHG dynamics in restored ecosystems, refining previous assumptions and providing more accurate data for GHG inventories and policy decisions.

This study provides a comprehensive global meta-analysis demonstrating the significant potential of ecological restoration in mitigating GHG emissions. Findings indicate that afforestation, reforestation, wetland rewetting, and grassland restoration through grazing management or cropland conversion are effective strategies. The strong influence of restoration age highlights the need for long-term monitoring and consideration of temporal dynamics in GHG assessments. Future research should focus on improving the understanding of microbial mechanisms driving GHG fluxes in restored ecosystems and refining the modeling of these dynamics for improved prediction and policy guidance.

While this meta-analysis used a large dataset, inherent limitations exist. Data heterogeneity across studies, including differences in methodology, site characteristics, and reporting practices, could introduce bias. The reliance on published data limits the representation of certain restoration types or regions. Furthermore, this study primarily focuses on three major GHGs; additional analyses including other GHGs could provide a more complete picture.