Peatlands have been under severe threat from various anthropogenic factors such as pollution, drainage and intensive utilisation for food and fibre production and many are now strong net sources of carbon to the atmosphere (Evans et al., 2021b). Peatlands are important sources of fluvial carbon including particulate organic carbon (POC), dissolved organic carbon (DOC) and dissolved gases (Billett et al., 2015; Rosset et al., 2022). Previous studies have suggested that the relative roles of these fluvial forms are typically ∼15 %–40 % of CO2 equivalent net ecosystem exchange (Dinsmore et al., 2010; Roulet et al., 2007; Billett et al., 2010). However, POC flux is particularly high from peatlands where vegetation cover is partial (Evans et al., 2006) and in these systems POC can contribute >80 % of the fluvial flux (Pawson et al., 2008) while a lack of vegetation will also be associated with a reduced terrestrial C uptake across the peatland and potentially to enhanced direct losses to the atmosphere. Given such large potential contributions to C losses, it is critical that more studies acknowledge the POC pathway in the carbon budget. Previous studies have suggested that both DOC and POC are metabolised to CO2 in the fluvial system to some degree, with current best estimates between 50 %–90 % conversion for POC and 80 %–100 % for DOC (Evans et al., 2013). However, most studies focus on terrestrial gas fluxes or aquatic DOC fluxes. Hence, the various pathways for POC storage, transport or transformation to CO2 are not well studied (Palmer et al., 2016). A large body of research has examined the geomorphological controls on peat erosion, especially from the perspective of sediment load (Reviewed in Li et al., 2018a). However, as governments and landowners look to quantify and reduce greenhouse gas emissions from peatlands, focus should turn to quantifying those that result from peatland erosion.

The onset of peatland erosion can be traced back over a thousand years (Evans and Warburton, 2011). It is hypothesised that there is a “threshold process” whereby the peat changes from a stable, intact state to an unstable, erosional state. Some propose that erosion is a natural termination after thousands of years of peat accumulation resulting in instability of the peat mass (Conway, 1954; Pearsall, 1956; Colhoun et al., 1965). Others argue that much of the erosion has resulted from anthropogenic pressures, including burning (Yallop et al., 2009), overgrazing (Wilson et al., 1993), artificial drainage installation (Worrall and Evans, 2009; Holden et al., 2007), and atmospheric pollution (Yeloff et al., 2006). Climate change can potentially enhance rates of erosion through more extreme weather events (e.g. Cotterill et al., 2021). For example, drought can cause desiccation of the peat, and impact by heavy rain can cause “wind splash” impacts and rapid overland flow, contributing to destabilisation and transport of the peat (Warburton, 2003). In contrast, reduced frost days and fewer freeze-thaw cycles could reduce erosion caused by needle ice (Li et al., 2018b). Interactions between climate change and erosion are complex and often ecosystem/biome or region specific. Nonetheless, erosion of peat is projected to change in the coming century with sediment yields projected in different regions decreasing or increasing by -1.27 to +21.63 tha-1yr-1 by 2800 (Li et al., 2017). It is therefore important to understand the links and feedbacks between climate, land use, peat erosion and CO2 emissions.

In the past century, most of the peat erosion and post-erosion POC research has been conducted in the UK. However, peat erosion is a pressing or emerging problem for peatland systems around the world, with potential for massive carbon losses and climate feedbacks (Fig. 1). Potential erosion hotspots are occurring in different environmental and management contexts around the world from drained forestry sites which may have relatively low areas of exposed peat (Marttila and Klove, 2010) to industrial extraction sites with almost complete bare peat cover (Campbell et al., 2002). At the extreme end of erosion, collapse of inland permafrost systems in the arctic and boreal regions (Swindles et al., 2015) can cause localised rapid erosion and movement of soil carbon via thaw slumps (Lamoureux et al., 2014; Pizano et al., 2014), with potential for high emissions as the mobilised carbon becomes available to decomposer organisms in freshwater environments (Li et al., 2024). In contrast, arctic permafrost coastal erosion and coastal-adjacent thaw slumps, which are occurring at an alarming rate in response to rapid warming around the Arctic Ocean, are depositing carbon directly into the ocean (Lantuit and Pollard, 2008; Lantuit et al., 2012). Equally, In the tropics of Asia, coastal erosion of peatlands is causing large direct fluxes of peat to the ocean (Kagawa et al., 2024), but there are also examples of inland peat erosion in Asia which will generate POC that is primary processed in terrestrial systems (Wang et al., 2019).

Peat erosion is clearly progressing in a variety of contexts and at different rates, but in every case it will be exacerbated by climate change and associated extreme weather events (Zhao et al., 2024). This is why IPCC reporting of emissions needs to move towards a more nuanced understanding of POC turnover than the broad downstream POC-CO2 conversion rate of 70 % (based on UK examples, IPCC, 2014). Depending on the context, biome and global location, estimated emissions resulting from peat erosion could vary significantly from currently reported rates.

The 2013 IPCC wetlands supplement (IPCC, 2014) present a general calculation for a POC emissions factor (EFPOC) for all peatlands and drained organic soils. This generic model, although primarily based on evidence from the UK, was designed for any peatland soil that had suffered significant disturbance that lead to bare peat, including drainage, burning, peat extraction and conversion to arable land as follows (IPCC, 2014): EFPOC=POCFLUX BAREPEAT×PEATBARE×FracPOC-CO2 Where: POCFLUX BAREPEAT is the POC flux per area of bare peat surface, PEATBARE is the area of bare peat and FracPOC-CO2 is the conversion of POC to CO2 following export from the peatland.

Mapping of bare peat extent at high resolution is progressing (Macfarlane et al., 2024) but the underpinning data for estimating emissions associated with bare peat are highly uncertain (Evans et al., 2013) as the flux depends on specific fluvial mixing events in time and space (Palmer et al., 2016). We argue that this calculation has two major sources of uncertainty which are critical to resolve to confidently quantify emissions that arise from peat erosion. Firstly, the flux of POC from bare peat at the source is only one part of peat volume loss – quantification of the relative contribution of direct CO2 loss, subsidence and erosion to surface retreat rates will give rise to better quantification POC loss via erosion. Eroded peat will potentially be processed and mineralised in multiple environments, from headwater streams, floodplains to rivers and the ocean (Evans et al., 2013; Zhou et al., 2021). Therefore, the second source of uncertainty is the fraction of eroded peat/POC that is converted to CO2. This can be addressed by considering the environments and organisms that interact with it while in transit over various timescales.

Erosion systems in peatlands are complex, often starting with narrow “V” shaped gullies which over time become “U” or trapezoidal-shaped gullies as lateral erosion predominates once erosion reaches the base of the peat (Evans and Warburton, 2008), with eroded peat moving through this system on its way to the headwater streams (Fig. 3). Based on sediment flux data compiled by Li et al. (2018a), we calculated 0.7 tCO2eha-1yr-1 loss of POC at the outlet of eroding blanket bogs (by applying a 53 % conversion factor for carbon content and a median sediment yield of 35 tkm-2yr-1). This rate is over six times lower than the POC flux that would be lost from peatlands based on our SRR rates estimates (Table 1). Due to scale-dependency of processes (see Li et al., 2018a, for a detailed discussion on this), direct comparison such as this should be cautioned against. Nevertheless, this difference raises important questions about how peat carbon is processed within the peatland system with potential important implications for carbon budgets in eroding peatland systems. Peat can transit through wider “U-shaped” gullies in a matter of hours; however, a proportion is likely to resediment within the system for years where it will be further oxidised by the microbial community. How long peat sediment is retained in the system and the conditions it is exposed to will determine its decomposition rates, how much moves on out of the system and how much is retained for the long term (Fig. 3).

Repeated drone-based remote sensing approaches are highlighting the amount of redeposition of peat sediment within peatlands. These approaches not only measure the geomorphic loss of peat on gully walls but also the vertical accumulation of peat in wider gullies downstream (Li et al., 2019; Glendell et al., 2017). One of the next important questions to answer is what is the mean residence time of peat particles within the system and how much decomposition occurs while the sediment is in this new environment? We know that areas that contain wide gullies are strong sources of CO2 (Artz et al., 2022), as are areas of bare peat in former peat extraction sites (Rankin et al., 2018). However, bare gullies were measured to be only marginal sources of CO2 with low metabolic activity in general (Dixon et al., 2014; Gatis et al., 2019). Using eddy covariance at high temporal resolution, conversely, Artz et al. (2022) observed peaks of ecosystem respiration at their eroded sites, suggesting that there are abiotic triggers that cause a pulse of CO2 flux from the peat. Therefore, it is important to collect more data across a range of bare peat systems and conditions to understand the likely loss rates via CO2 in these gully complexes.

Vegetation cover on the gully floor is a key control on the rate of resedimentation, as higher cover of vascular plants and bryophytes will hold up the flow of water and suspended sediment (Crowe et al., 2008; Harris and Baird, 2019; Milner et al., 2021), resulting in a reduced POC yield at the outflow (Evans et al., 2006). Erosion systems often reach a revegetated-base state with low POC flux rates at the outlet (Evans et al., 2006), however, it is not clear whether the trapped peat sediment is lost as CO2 higher up in the peatland system. The turnover of resedimented material may depend on the vegetation cover of the gully floor, and its potential to trap freshly produced POC. Vegetated gullies were found to be “hotspots” for ecosystem respiration in an eroding blanket bog (McNamara et al., 2008). Another naturally revegetated site had almost neutral CO2 exchange as a result of high primary productivity being balanced by ecosystem respiration rates that were almost four times higher than in a bare gully system (Dixon et al., 2014). For both examples much of the ecosystem respiration flux will be from plant respiration, linked to vigorous growth associated with revegetated gully bases, however a significant amount could be from the resedimented peat. We understand that where peat is deposited on downstream floodplains (Alderson et al., 2019), it is turned over rapidly in aerobic conditions (Alderson et al., 2024) and up to 80 % of POC deposited on a floodplain downstream of a peatland catchment will be mineralised to CO2 within 30 years (Evans et al., 2013). Similar emission rates could be occurring in revegetated gully floors upstream of these floodplains. Microbial activity is likely to be lower in the gullies than downstream floodplains due to higher altitude (and therefore lower temperature), higher moisture (often saturating and anoxic) and potentially different plant communities and associated microbial communities. However, CO2 flux from redeposited POC within peatlands could still be substantial and this knowledge gap could be resolved by measuring the isotopic signature of the CO2 produced to partition into autotrophic and heterotrophic sources.

Depending on the extent of bare peat within a peatland, and the local slope and wind conditions, erosion can be the dominant pathway for carbon loss (Evans et al., 2006). Peat that is lost through erosion has potential to be degraded to CO2 at various stages on its transit as POC. Due to the complex biophysical processes and interactions that cascade from peat erosion there is very high uncertainty around the emissions that occur as a result. There is a risk of both under- and overestimating emissions from peat erosion, depending on the characteristics of the site. These uncertainties feed through to uncertainties in national peatland emissions reporting (IPCC, 2014) and estimates of emissions reductions that can be achieved through bare peat restoration and revegetation. There are mismatches in data on sediment/carbon fluxes at certain points and across scales on the journey of peat after it is eroded. These mismatches can be addressed by applying scalable measurements at key junctures in the peat sediment's transition out of the peatland. These measurements should be biologically and biogeochemically focussed on the processes that mineralise peat sediment to CO2, and ultimately cause the emissions that we need to quantify. Climate change is driving increasingly severe erosive forces across all peatlands, from increased storminess to decreasing permafrost stability. Therefore, emissions arising from peat erosion are likely to have an increasingly important role in the carbon balance of all peatlands.