Peatland restoration pathways to mitigate greenhouse gas emissions and retain peat carbon

Peatlands play a crucial role in the global carbon (C) cycle, making their restoration a key strategy for mitigating greenhouse gas (GHG) emissions and retaining C. This study analyses the most common restoration pathways employed in boreal and temperate peatlands, potentially applicable in tropical peat swamp forests. Our analysis focuses on the GHG emissions and C retention potential of the restoration measures. To assess the C stock change in restored (rewetted) peatlands and afforested peatlands with continuous drainage, we adopt a conceptual approach that considers short-term C capture (GHG exchange between the atmosphere and the peatland ecosystem) and long-term C sequestration in peat. The primary criterion of our conceptual model is the capacity of restoration measures to capture C and reduce GHG emissions. Our findings indicate that carbon dioxide (CO2) is the most influential part of long-term climate impact of restored peatlands, whereas moderate methane (CH4) emissions and low N2O fluxes are relatively unimportant. However, lateral losses of dissolved and particulate C in water can account up to a half of the total C stock change. Among the restored peatland types, Sphagnum paludiculture showed the highest CO2 capture, followed by shallow lakes and reed/grass paludiculture. Shallow lakeshore vegetation in restored peatlands can reduce CO2 emissions and sequester C but still emit CH4, particularly during the first 20 years after restoration. Our conceptual modelling approach reveals that over a 300-year period, under stable climate conditions, drained bog forests can lose up to 50% of initial C content. In managed (regularly harvested) and continuously drained peatland forests, C accumulation in biomass and litter input does not compensate C losses from peat. In contrast, rewetted unmanaged peatland forests are turning into a persistent C sink. The modelling results emphasized the importance of long-term C balance analysis which considers soil C accumulation, moving beyond the short-term C cycling between vegetation and the atmosphere. Supplementary Information The online version contains supplementary material available at 10.1007/s10533-023-01103-1.


Importance of peatlands in a changing climate
Peatlands cover only about 3% of the Earth's terrestrial surface (Gorham 1991) but play a crucial role in the global carbon (C) cycle.They act as significant C stores and are sources or sinks for greenhouse gases (GHG) like carbon dioxide (CO 2 ), methane (CH 4 ) and nitrous oxide (N 2 O; Frolking et al. 2011).These gases contribute to climate change and are considered crucial anthropogenic GHGs.Due to their high C density, peatlands are globally recognized as vital C reservoirs (Gallego-Sala et al. 2018) accounting for about 21% of the global soil organic C stock, estimated ~ 3000 Pg (Leifeld and Menichetti 2017).Peatlands also serve as substantial stores of organic N, with Northern peatlands alone accumulating 8-15 Pg N. When including tropical peatlands, the global estimations reached up to 26 Pg N (Swenson et al. 2019).
Undisturbed peatlands are currently a C sink (~ 0.1 Pg C y -1 ), a moderate source of CH 4 (~ 0.03 Pg CH 4 y -1 ), and a very weak source of N 2 O (~ 0.00002 Pg N 2 O-N y -1 ) (Frolking et al. 2011).However, anthropogenic activities, primarily agriculture and forestry drainage (up to 20% of global peatlands), result in net CO 2 emissions (~ 0.1 Pg C y -1 ), reduced CH 4 emissions (10% smaller than in natural conditions), and increased N 2 O emissions (~ 20 times higher than in natural peatlands).Consequently, subsidence and soil degradation contribute nearly 6% of global anthropogenic GHG emissions (Wichtmann et al. 2016).Most likely, the global peatland's GHG balance has turned to a C source, a slightly smaller CH 4 source, and a larger (but still small) N 2 O source (Frolking et al. 2011).In Europe, 46% of the remaining peatlands have degraded to the point where peat is no longer actively forming (Swenson et al. 2019).
As typical wetlands, peatlands are severely threatened by drainage, climate change, fires and groundwater extraction (Fluet-Chouinard et al. 2023).However, their restoration is beneficial, enabling C capture and sequestration, and minimizing N 2 O emissions (Leifeld and Menichetti 2017).
The northern hemisphere has experienced the highest warming during winter and early spring (Ljungqvist et al. 2016), leading to more frequent freeze-thaw cycles.Annual precipitation has also increased, particularly during the cold half-year in northern regions (Ljungqvist et al. 2016).Conversely, there are many indications of growing frequency of droughts during the warm season (Chiang et al. 2021).An increase in flash floods is also predicted (Zheng et al. 2022).These changing climatic conditions, droughts, rapid fluctuations in groundwater level (Mander et al. 2021), flash floods (Schindler et al. 2020), and soil moisture conditions (Pärn et al. 2018;Evans et al. 2021;Huang et al. 2021) can create hot spots and hot moments of GHG emission in peatland ecosystems.

Peatland restoration
Rewetting is a crucial step for conservation or sequestration of C in peatlands previously affected by drainage (Günther et al. 2020).Likewise, rewetting can affect nitrogen cycle and reduce cumulative N 2 O-N emissions by up to 70% in European peatlands (Liu et al. 2022).Usually, peatland restoration and rewetting are considered as synonyms, but in this paper we consider rewetting as increasing water table level compared with previously drained status.In case of restoration we expect that both water table and vegetation are manipulated in the way that enables to achieve ecosystem status similar to pre-drainage.Restoration pathways depend on factors such as initial vegetation of the drained area (Heger et al. 2022;Schaller et al. 2022), nutrient status of the residual peat layer (Kreyling et al. 2021) and the expected time frame for achieving planned ecological and socioeconomic benefits of the restoration.
Rewetting may transform former peat extraction sites into mires, paludicultural land, wet forests or shallow waterbodies.Ecologically, the preferred pathway is to restore them as mires (Wilson et al. 2022), which can be achieved by restoring water level, by a combination of rewetting and the application of peat moss layer transfer technique (Gonzalez-Sargas & Rochefort 2019), or to establish shallow waterbodies in hydrologically complex sites (Christen et al. 2016).
Restoration of drained forests is usually achieved by blocking ditches and restoring water levels (Grand-Clement et al. 2015).Regulating the water regime in restored sites is effective for mitigation of GHG emissions (Järveoja et al. 2016b).
Peatland restoration is vulnerable to hydroclimatic conditions, particularly in the temperate zone.Changing precipitation pattern, increasing temperature and decreasing snowpack are expected to contribute to more frequent extreme events like droughts and torrential rains, resulting in increased vulnerability and interannual variability (Alm et al. 1999;Drollinger et al. 2019).In addition, it is also important to consider the potential substantial losses of dissolved and particulate C from drained and restored peatlands when estimating C budgets (Billett et al. 2010;Rosset et al. 2022).
Due to global warming, northern peatlands are projected to experience increased GHG emissions, particularly during non-growing period (Rafat et al. 2021), while Garcin et al. (2022) highlight a lack of knowledge of hydroclimatic vulnerability of peat C in tropical peatlands.Wet-dry seasonality of GHG fluxes is expected from tropical peatland soils (Inubushi et al. 2003).However, the overall impact of climate change on GHG fluxes needs to be better understood.Current understanding suggests that changes in soil temperature, photosynthesis, and soil moisture drive alterations in net C fluxes (Rafat et al. 2021).

Restoration versus afforestation of peatlands
Peatland restoration involves rewetting, whereas afforestation of drained peatlands and maintaining their drained condition cannot be considered equivalent to restoration.The key issue lies in the difference in the time scale and the discrepancy in distinguishing between short-term C capture in ecosystem (often observed in studies on GHG exchange between the atmosphere and ecosystem) and long-term C sequestration in soil.When peatlands are drained, C loss from the peat can offset the benefit of long-term CO 2 sequestration achieved by afforestation (Jurasinski et al. 2020(Jurasinski et al. , 2023)).
Currently, there is insufficient evidence on the long-term benefits of active afforestation of drained peatlands to mitigate climate change (Jurasinski et al. 2023).Afforestation on drained peatland forests and some former peat extraction areas may provide short-term benefits for climate change mitigation (Mäkiranta et al. 2007;Samariks et al. 2023).However, this approach does not account for the value of long-term C storage in peat.Similarly, intensive forestry on drained peatlands will not restore the peatland ecosystem's flora, fauna, and functions (Haapalehto et al. 2017;Loisel & Gallego-Sala 2022).
In addition, afforested drained peatlands are more susceptible to wildfires (Kohlenberg et al. 2018;Zheng et al. 2023).These risks are exacerbated by more frequent and more intense droughts in the boreal zone (Walker et al. 2019), resulting in losses of burnt wood and substantial C losses from burnt and burning peat layers (Liu et al. 2023).The impacts of severe fires have been devastating drained areas of formerly tropical peatland forests in Southeast Asia (Page et al. 2009).
Restoration-versus-afforestation of peatlands is being debated during the legislation procedure of the European Union Nature Restoration Law (NRL) (Jurasinski et al. 2023).
The general aim of this paper is to assess shortterm C capture (GHG exchange between the atmosphere and the peatland ecosystem) and analyse the fluxes in the context of long-term C sequestration in peat of restored (rewetted) peatlands and afforested drained peatlands.As a specific objective we implement a conceptual model that compares changes in the long-term C budget and climate warming mitigation potential in the restored and afforested peatlands.

Conceptual framework
To provide a comprehensive understanding of the framework and to organize the rich source material, we have adopted a three-stage system.These stages refer to peatlands with different water regimes: natural (pristine), drained, and restored (rewetted) ecosystems.Within each stage of water regime, we further divided them into blocks based on primary Vol:.( 1234567890) land use types (Fig. 1).Estimated fluxes of all three GHGs-CO 2 , CH 4 and N 2 O-are presented using a general three tier scale (high, medium and low fluxes).Flux values are represented by arrows, with upward-pointing arrows indicating emission and downward-pointing arrows indicating capture or uptake of the corresponding gas.In addition, we have indicated the averaged lateral losses of dissolved organic C in water (Fig. 1).
Although our main focus lies on temperate and boreal peatlands, we also briefly discuss tropical peatland forests due to their significance as global hot spots of GHGs, and their large-scale escalating disturbance and destruction.In characterizing GHG fluxes in natural and drained peatlands, we primarily relied on literature sources supplemented by unpublished original data from our research group.However, for  For restored (rewetted) peatlands, due to the scarcity and less systematic nature of the data, we present a selection of available GHG flux data in Table 1.The categories of restored (rewetted) peatlands are as in Fig. 1: wet forests, paludiculture, shallow lakes and open mires.Upon no accurate distinction between the mentioned categories in the source data we added a general category -rewetted peatlands -to both Table 1 and Fig. 2. The latter one presents average flux values (Fig. 2A-C) and GHG balances in CO 2 equivalents (Fig. 2D and Fig. 2E).In the more general Fig. 1, paludiculture is shown as a single category.To compile this information, we referred to both literature sources and results from local research projects in Estonia.For our literature search, we utilized well-known indexing systems such as Web of Science (Clarivate Analytics), Scopus, and Google Scholar.We used following keywords and their combinations: "rewetted peatland(s)", "restored wetland(s)", "restored peatland(s)", "rewetted forest(s)", "paludiculture", "Sphagnum farming", "greenhouse gases, "carbon dioxide ", methane, "nitrous oxide", "tropical peatlands".In addition, we looked for references from cited rewetting-related papers.
To explain the dilemma of peatland restoration and afforestation in a long-term perspective and mitigation of climate warming, we created a conceptual model that helps to characterize the changes in the C budget under different management practices (Fig. 3).
The resulting climatic effects of these management options are strongly time-dependent and rely on difference in forming above or below ground carbon pool and carbon turnover rate.
Our conceptual model of long-term C stock dynamics over 300-year period is based on the approach proposed by Minkkinen & Laine (1998).We developed it further to include C stock dynamics in nutrient-poor mires (open bog and bog forest), drained bog forest, and potential land use scenarios after milled peat extraction (restored bog, naturally regenerated unmanaged forest and managed forest).In modelling, we assumed equal C stock 686 t ha −1 in any peatland as initial status, corresponding to C stock with 2 m peat depth in pristine bog (Supplementary Table 1).As approximation, the stem, branches and coarse root biomass were assumed to have stable ratio (branches 12% of stem biomass and coarse roots 19% of above ground biomass) in any age group, wood density (kg m 3 ) and C content (%) values were derived from literature (Kask and Pikk 2009;Külla 1997).Initial value of stem biomass was set 3 m 3 ha − 1 for pristine open bog, 100 m 3 ha −1 yr −1 for pristine bog forest and 0 m 3 ha −1 for any other land use classes (initially harvested peatland forest or treeless peat extraction site) with annual stem biomass increment by 0 m 3 ha −1 yr −1 for open bog, 0.1 m 3 ha −1 yr −1 for stable continuous cover bog forest, 4 m 3 ha −1 yr −1 for drained peatland forest and naturally regenerated forest on peat extraction site (initial value 0 m 3 ha −1 yr −1 until year 20, thereafter gradually increasing until 4 m 3 ha −1 yr −1 by year 50 and starting to decline to 0.1 m 3 ha −1 yr −1 by year 160 as equal to pristine bog forest) while mean annual increment of 6 m 3 ha −1 yr −1 was assigned to managed forest on drained former peat extraction site because of higher fertility of deeper peat layers (usually fen peat) after removal of Sphagnum peat.In pristine bog forest, restored bog and unmanaged peatland forest on former peat extraction site no harvest and timber removal is considered.Harvesting cycle is assumed to be 100 years in both restored managed peatland forest and drained managed peatland forests, while only stem biomass is removed.Fine root biomass was considered as part of C turnover and part of net ecosystem exchange (NEE).NEE values for modelled ecosystems are based on original data from Estonian study sites (Table 1), Estonian national GHG inventory, and literature (Salm et al. 2012;Wilson et al. 2016a).A conceptual scheme presents modelled longterm C stock dynamics in pristine mire ecosystems, restored and afforested drained peatlands (Fig. 3).

Greenhouse gas fluxes in restored peatlands
Figure 1 illustrates a conceptual view of estimated GHG fluxes in natural, drained and restored (rewetted) peatlands.In restored peatlands, CO 2 fluxes play the main role in climate impacts, while CH 4 emissions are moderate and N 2 O fluxes are low making them less significant.Lateral losses of dissolved and particulate C in water can account up to a half of total C budget.Among the different restored peatlands, paludiculture has the highest estimated carbon dioxide h −1 ; chambers, GC; range between average annual values.
h −1 ; chambers, GC, average annual ± se.According to our analysis, all restored peatlands were C sinks.However, there was significant variation of the data, mainly due to limited data availability and differences in the age since rewetting, as well as variation in plant cover development (  open peatlands and paludicultural sites was − 928, − 534, and − 528 kg CO 2 -C ha −1 yr −1 , respectively (Table 1).These findings are consistent with the CO 2 flux values reported by Günther et al. (2020) for rewetted peatlands.Due to their diversity, paludicultural ecosystems showed the most significant variation in CO 2 -C fluxes ranging from high C capturing to moderate emissions, with the highest values observed in Phalaris arundinacea and Poa spp.plantation on rich fen peat (Kandel et al. 2020).Shallow lakes established on flooded peat extraction sites generally emitted low to moderate levels of CO 2 except extensively vegetated eutrophic sites (Minke et al. 2016).In all rewetted peatlands, the average annual CH 4 fluxes were at moderate level, with the highest values observed in shallow lakes (451 kg CH 4 -C ha −1 yr −1 ) and followed by rewetted forests, open bogs/fens and paludiculture (218, 163, and 122 kg CH 4 -C ha −1 yr −1 , respectively; Table 1).Average annual N 2 O fluxes ranged from low to moderate levels (0.01 to 4.45 kg N 2 O-N ha −1 yr −1 ), except in Phalaris paludiculture plantations on poor and acid peat where liming and fertilization resulted in fluxes of up to 19 kg N 2 O-N ha −1 yr −1 under high groundwater levels and 28.1 kg N 2 O-N ha −1 yr −1 under low water levels (Maddison et al. 2016; Table 1).
In management of drained peatlands, global warming potential (GWP) of GHG-s should be considered to avoid making decisions based solely on short-period benefits that may overlook the long-term climate cooling effect.Drained peatlands are known to be persistent CO 2 emitters over the long term, while rewetted peatlands as resilient re-established mire ecosystems effectively contribute to mitigating climate change, even considering radiative forcing of increased CH 4 emissions (Günther et al. 2020) and decreased N 2 O emissions.Figure 2D and E demonstrate the climate effect of rewetted peatland ecosystems in GWP 100 and GWP 500 timeframes.Mires are the only terrestrial ecosystems capable to continuously sequester atmospheric carbon in the long term (Gorham et al. 2012;Cobb et al. 2020), i.e., the carbon sequestration itself is more important than its compound (methane vs carbon dioxide).
For restored (rewetted) peatlands see literature sources in Table 1.

Modelling approach
To estimate the carbon balance of afforested drained and restored rewetted peatlands we developed a conceptual model.Based on the availability of data from our research projects and similar studies in other countries we chose hemiboreal ombrotrophic peatlands as a modelling example.Figure 3 presents estimated long-term dynamics of C stock of afforested drained and restored rewetted hemiboreal ombrotrophic peatlands and their reference pristine bog ecosystem.In a 300-year perspective, considering stable climate conditions, drained forest peat (former bogs) will lose 1.8 and 1.0 t C ha −1 yr −1 , respectively, due to peat mineralization.In afforested peatlands under continuous drainage, C losses from peat reach 0.9 t C ha −1 yr −1 , whereas ecosystem C losses (the peat + vegetation budget and 100-year timberharvesting regime) are 0.4-0.6 t C ha −1 y −1 .In comparison, fifty years after rewetting the naturally regenerated unmanaged peatland forest show decreased C loss, which is due to lower peat mineralization and C accumulation in biomass.In 160 years they achieve C dynamics similar to the pristine bog forest-with a moderate C sequestration rate of 0.1 t C ha −1 y −1 .Restoring a peat extraction site to bog ecosystem would become C neutral in nearly 20 years, and onwards continuous mean annual C sequestration of 0.22 t C ha −1 y −1 is assumed (Supplementary Table 1).
Peatland forest drainage can give a significant increase in short-term C capture in biomass (Lohila et al. 2011), however, long-term dynamics in C stock and peat mineralisation remain largely unknown.
Restoration of crop plantations established on tropical peatlands and other dramatically altered peatlands differs from restoration of temperate and boreal bogs because the source community is predominantly swamp forest where peat-forming material is predominantly wood.Above ground and below ground litter formation under the reforestation and management with moderate drainage can more easily compensate peat mineralization than in temperate and boreal areas (Couwenberg et al. 2010).

Conclusions
Understanding the dynamics of GHG fluxes caused by land-use change is essential for successful peatland restoration.Our analysis identified several contradictory research results and gaps in a deep understanding of these processes.Notably, there is a lack of GHG flux data for most of tropical drained and restored peatlands, with the exception of oil palm plantations.Another important issue is retention of C in restored peatlands, where we can differentiate between shortterm C capture (GHG exchange between the peatland and atmosphere) and long-term C capture (accumulation in soil).Based on our analysis of literature sources and own research results from Estonia, rewetting of former peat extraction areas for further management or conservation is the only viable approach for long-term C sequestration.In contrast, afforestation combined with continuous peatland drainage may have short-term economic benefits but leads to C losses in the long term.
Uncertainties in long-term estimations of C storage and GHG flux dynamics are remarkably high and do not allow for exact predictions.However, even educated guesses can be valuable for decision making on further management of peatlands.To make accurate estimations, it is crucial to investigate the full combined impact of hydroclimate change, microbial processes, and vegetation on GHG emissions from restored peatlands.

Fig. 1
Fig. 1 Conceptual figure showing main generalized types of peatland ecosystems from natural and drained to potentially restored peatlands.Proposed set-up for greenhouse gas emission from the main compartments of ecosystems is shown.Arrows indicate literature-based mean values of greenhouse gases CO 2 , CH 4 and N 2 O fluxes (upward and downward arrows indicate emissions and captures of the corresponding gas,

Fig. 2
Fig. 2 Average annual values of greenhouse gas fluxes in main groups of restored (rewetted) peatlands.A-C Annual average values of CO 2 -C (NEE), CH 4 -C and N 2 O fluxes.D: greenhouse gas balance in CO 2 equivalent GWP 100 values (25

Table 1
Greenhouse gas fluxes in main types of boreal and temperate rewetted peatlands Type vegetation period average values.
GC, average annual values.

Restored to bog Unmanaged forest Unmanaged forest soil Managed forest Managed forest soil Bog Bog forest Managed forest Managed forest soil Milled peat extraction field Pristine Drained bog
; Mander