The Scottish Government

« Previous | Contents | Next »

Listen

Module 5 Estimates of Carbon Loss from Scenarios of Accelerated Erosion of Peats

5.1 Introduction

This module reviews the processes that have been implicated in the erosion of peats and organo-mineral soils and how these may change into the future. Estimates of the extent of erosion are also presented and discussed and areas requiring further research have been identified.

The literature, and indeed evidence, pertaining to the erosion of organic-rich soils focuses heavily on blanket peat or mire soils rather than organo-mineral soils such as peaty podzols, peaty gleys or peaty rankers. All the available evidence suggests that the latter group of soils are much more stable than blanket peats and there are few instances of erosion on them. More details are given in Table 5.1. Against this background, the remainder of this section will focus on peat erosion and it is likely that the processes that contribute to peat erosion are also applicable in some degree to organo-mineral soils.

This report covers a brief review of the processes and causal factors of erosion of peats, assessments of the extent and likelihood of erosion of peats, approaches to define, identify and quantify peat erosion and losses, an appraisal of potential changes to the causal factors and an assessment of likely changes to erosion risk and ends with conclusions and a summary.

5.2 Background

Soil erosion is a natural process which occurs in all soils to a greater or lesser extent. Soil erosion becomes of concern when the rate exceeds "natural" or "background" rates which can be considered as broadly equal to the rate of formation of new soil material by weathering processes. Soil erosion at rates exceeding background values is termed "accelerated erosion" and, in humid temperate climates such as that of Scotland, much of the erosion is usually the result of human activities that lead to removal of the protective vegetation cover. However, mass movements such as landslips also occur in the over-steeped, glaciated hillslopes of Scotland (Ballantyne, 1991). The material eroded from soils is a major contributor to the sediment load carried by streams and rivers and can disrupt transport links, therefore erosion has implications both on- and off-site.

The major processes of soil erosion are:

i) Water erosion including gullying, rilling and sheet erosion
ii) Mass movements such as landslides
iii) Wind erosion.

Erosion of cultivated mineral soils has been the subject of considerable research efforts and the review of evidence within Towers et al.. (2006) provides an up-to-date summary of our current understanding of the processes involved. These events tend to be localized and, although their short term impact can be dramatic, in most circumstances they are relatively easy to rectify. In contrast, upland erosion can cover large areas and is more difficult to repair.

Erosion of peat ranges from a few scattered rills, hags or gullies within an otherwise pristine bog, to bogs where there has been complete removal of the surface vegetation with large areas of bare exposed peat. Two examples are given in Figures 1 and 2. In Scotland, the pattern represented by Figure 5.1 is much more common whereas that represented by Figure 5.2 is restricted to specific areas of serious erosion. In some places, the original peat cover has been almost completely removed and only remnant "hags" remain (Figure 5.3)

Figure 5.1 Typical pattern of gully erosion in blanket peat

Figure 5.2 Severely eroded peat with extensive bare ground (black areas)

Figure 5.3 Remnant of peat, indicative of a much more extensive cover in the past.

5.3 Causes of peat erosion

5.3.1 Erosion processes

There are various ways to describe different erosion processes, features and patterns in peat. Perhaps the most influential was that of Bower (1960) who identified five main types of erosion system on the basis of morphology and pattern. These were considered to result from the operation of two main processes: water erosion and mass movements. Water erosion produces dissection systems which develop onto and into the peat mass whereas sheet erosion occurs on the peat surface as the vegetation breaks up. On steeper slopes mass movements may result and under very wet conditions bog bursts and peat slides can occur.

The dissection process is much more prevalent and extensive compared to bog bursts and slides, produces a very distinctive landscape, but operates over very long time periods. Any changes to that landscape happen very slowly compared to peat slides and bog bursts and for that reason, the latter are studied much more intensively as the evidence of the erosion event and its associated impact is much more immediate. Brief reviews of both processes are given below.

5.3.2 Dissection processes

Drainage dissection was considered by Bower to be the more important erosive process both spatially and in terms of volume of peat removed (Bower, 1960). Indeed understanding drainage dissection is arguably the key to understanding peat loss in the uplands (Holden et al., 2006). Two types of dissection system associated with water erosion were identified by Bower, which produce different types of gully patterns. The key differences in the pattern of gullies produced by the two dissection processes are a function of peat depth and slope angle, both of which themselves are inter-related. The frequency and complexity of the gully pattern is highest on high level, very gently sloping upland plateau. Bower's classification has been widely employed in subsequent studies on peat degradation albeit with some suggested modifications to it, e.g. Tomlinson (1981).

Although peat erosion produces a quite dramatic landscape of hags, gullies and shallower channels, poor connectivity between them often means that peat that is removed does not always find its way to the drainage system but is redistributed within the peat mass (Evans & Warburton 2005). The connectivity between these different components can change with time and can alter the amount of sediment delivery. Extensive revegetation of the gully floors can have similar effects.

The trigger for erosion of organic and organic-mineral soils has been the subject of much debate and research for a number of decades. Burning, either accidentally or deliberately, grazing pressure by sheep and/or deer, air pollution, peat cutting, drainage and recreational pressure have all been cited in the literature as either causing or exacerbating erosion of upland soils. Exceptional weather events are also contributory factors and there are a number of studies that suggest that climatic perturbations during the previous millennium (Burt et al., 1997. Holden et al., 2006) caused desiccation at the peat surface and provided the initial trigger for peat erosion. Subsequent anthropogenic factors such as burning, grazing and atmospheric pollution have then impacted on the initial disturbed surface and exacerbated the erosion effect.

A number of factors have been implicated as the causes of peat erosion but most workers recognise the complexity and inter-relationships between them. By far the most work on peat erosion is from the Southern Pennines and there must be care in how the findings from these studies are extrapolated to other areas; the Southern Pennines receives a lower annual rainfall than much of the remainder of blanket mire in the UK so climate influences differ but, equally important, they are in close proximity to a population of more than 5 million people and the recreational and air pollution pressures that that entails (Tallis, 1997).

5.3.3 Climate

There is still considerable debate about the balance between whether peat erosion is a man-induced process or whether it is also partly a natural process representing the end point of a cycle of accumulation and build up of organic matter. There is some evidence from the Southern Pennines suggesting that climatic perturbations at key points of the last millennium provided the initial 'trigger' for peat erosion and that the much of the present landform stems from that period. Between AD 1150 to AD1300 (the early medieval warm period), there was a drier than average period causing drying out of pools, shrinkage and cracking of the peat and crusting of bare peat surfaces. From AD1300 to AD 1740 during which time the Little Ice Age occurred, there were generally colder winters and wetter summers with resultant volumes of run-off increasing erosion risk (Burt et al., 1997). Wishart & Warburton (2001) also suggest that some of the gully systems on the Cheviot Hills may be at least 500 years old.

Presently much of the eroded peat resource is at high altitude and in general terms, the higher the peat is above sea level the greater is the likelihood of it being eroded and the proportion of it that is eroded (Grieve et al., 1994). This suggests that the current climate is also a factor in peat erosion. The positive correlation between altitude and severity of gully erosion, suggesting that climatic factors, including the incidence of frost, high winds and more frequent and heavier rainfall events are important, was recognised as long ago as 1949 by Osvald and also by Bibby (1982). Although some areas of eroded peat are found as low as sea level, it is usually in very exposed situations such as Lewis, Shetland and the western seaboard of Scotland. Analysis of a number of sites from throughout the UK and Ireland suggest that climatic influences may be the most important influence in causing erosion (Rhodes & Stevenson, 1997).

Summer drought has the potential to bring about large and possibly irreversible changes to peats through desiccation of the soil surface, leading to cracking and increased vulnerability to erosion (various studies cited in Holden et al., 2006). Such drying also makes the peat more vulnerable to wildfires.

5.3.4 Burning

Burning is a commonly used management tool to control semi-natural vegetation on heathlands and to a lesser extent on blanket mires. If these fires take place at the appropriate time (October to April), are well managed and not allowed to become too hot and destroy the root mat, they should not expose the peat to erosive forces (Rhodea and Stevenson 1997). However in some areas, damage is caused by fires as a result of heavy recreational use, often in summer and these fires are unmanaged (Anderson 1997). Rhodes and Stevenson (1997) found that vegetation burning was implicated on only one site out of seven sites studied although the authors do speculate that the burning may have helped perpetuate and enhance erosion. Imeson (1971) and Yallop (2006) contend that heather burning has been a causal factor in soil erosion, and Yallop (pers. comm.) suggests that there is a lot of additional evidence, admittedly most of which is apocryphal, to support this view. The balance of opinion is that fire in general is not a key trigger in initiating peat erosion although accidental (or deliberate) sporting- and recreation-related fires do have the potential to cause serious damage in single events. More work is required to determine the role that fire plays in soil erosion, but if practiced properly.

5.3.5 Grazing

Grazing by domestic (largely sheep) and wild animals (largely red deer) can alter the ground vegetation on peat (which in itself is viewed as degradation of the resource) and when carried to its ultimate extreme, create bare patches of peat which are exposed to climatic and other influences to promote further erosion. This process is very similar to the poaching of wet soils by grazing animals within fields but a key difference is that the latter can be rectified reasonably well and quickly whereas restoration on peat is much more difficult, expensive and takes much longer.

Heavy grazing by sheep causes a decline in the cover of heather and other ericoids and replacement by tussock forming graminoid species such as Eriophorum Angustifolium, Molinia Caerulea and Nardus Stricta. Nardus in particular has short rhizomes and hence has poor soil binding qualities. Thus peat or other soils with high organic rich surface horizons supporting a Nardus-dominated vegetation become more prone to erosion (McKee & Skeffington, 1997). Birnie & Hulme (1990) noted that much of the peatland vegetation in Shetland showed evidence of grazing modification and peat erosional features are widespread. Biologically unsustainable grazing levels have been identified as the cause of the continuing degradation of both the vegetation and peat resource in Shetland. The balance of evidence is that while heavy grazing may exacerbate and accelerate peat erosion the original trigger for its initiation probably predates the intensive use of the hills for grazing, around two centuries ago (Rhodes & Stevenson 1997).

5.3.6 Atmospheric pollution

Skefflington et al. (1997) suggest that acid deposition since the start of the Industrial revolution about two centuries ago may be one of the causal factors for peat erosion in the Southern Pennines. Sphagnum has almost completely disappeared from these bogs thereby preventing further accumulation of organic matter at the soil surface; although the reason for this loss is not well understood, it is thought to be linked to acidification. Another aspect of acid deposition on blanket bogs and whether it has caused a reduction of fertility of these systems is the leaching out of base cations. In a comparison of 8 bogs across the UK from NW Scotland to SW England, there is some evidence, but no more than a trend, that the two Southern Pennine peats have lower exchangeable Mg and K than the others. Exchangeable Ca is lower at one of the sites. However, none of the levels of base cations correlated significantly with sulphur deposition. It may be that atmospheric deposition does not cause erosion per se, but the huge loss of Sphagnum spp. due to acid deposition, provided the impetus for other factors to take effect. Cresser et al. (1997) suggests that further work is required on the possible effects that acid deposition has had on the physical properties of peat.

Nitrogen deposition onto organic soils is harmful to them in terms of altering species composition and hence the habitat value of the resource but is unlikely to increase the risk of physical erosion. Indeed the decline in sulphur deposition and increase in N deposition has been linked to revegetation of parts of the North Pennines and hence providing a stabilizing influence (Evans and Warburton 2005) and may be responsible for increased DOC loads in streams (see Module 4).

5.3.7 Drainage

During the 1950s and 1960s, there were extensive efforts to drain blanket peat for agricultural purposes, although their effectiveness was never properly assessed (Holden et al., 2006). Artificial drainage may have altered the natural hydrology of peatland systems and thereby potentially making them more susceptible to erosion forces, but there is no clear evidence that this is the case. Certainly in Scotland, areas of eroded peat have not been subjected to artificial drainage and areas that have, based on visual observation in the field and from air photographs, do not show any evidence of erosion.

5.3.8 Other factors

Previous studies (Tallis et al., 1997) have identified factors that cause peat 'degradation' but in a different sense from peat erosion. These include the domestic and commercial extraction of peat as a fuel, the fertilization of peat for agricultural improvement, forestry development, reseeding, and the increase in the use of ATVs (all terrain vehicles) by land managers. Peat extraction and trafficking by vehicles have the potential to increase the amount of bare soil exposed to other agents but there is little hard evidence that this has happened to any great extent. Forestry operations in the past, particularly crop establishment, were very intrusive, but there is no literature to suggest that this caused any increase in the physical movement and erosion of peat. Peat is also very susceptible to damage by trafficking either by walkers or by the use of All Terrain Vehicles ( ATVs) and this is a growing threat to the peat resource, albeit in specific local contexts.

5.2.9 Peat slide and bog bursts

Bog bursts and slides are generally relatively small in extent and Warburton et al. (2004) has produced a comprehensive review of the processes involved. Sites of peat mass movement share a number of common characteristics which predispose them to failure:

• a peat layer overlying an impervious or very low permeability base
• a convex slope or a slope with a break of slope at its head
• proximity to local drainage either from seepage, groundwater flow, flushes, pipes or streams
• connectivity between surface drainage and the peat/impervious interface

There have been a number of bog bursts and slides reported in Scotland, e.g. Bower (1960) and Acreman (1991), but most have been reported from the Pennines (Nolan and Birnie 2006, Holden et al. 2006). An example is illustrated in Figure 5.4. There is evidence of a marked increase in the frequency of landslides recorded in British peatlands and the associated frequency of triggering events such as intense rainstorms. There are obvious implications for this trend continuing should predicted climate change scenarios become reality.

Figure 5.4 A bog slide in Skye, September 2005. The human figure on the left hand side, near the top of the slide, indicates the scale. Photo courtesy of A J Nolan

Nolan & Birnie (2006) in their report on a peat slide on Skye in 2005 discussed whether this type of event was possible to predict. As Warburton et al. (2004) concluded that 'the prospect of predicting the location and timing of such events is still a long way off', they suggest a more practical approach might be to develop a field assessment method which will provide an indication of the vulnerability of any slope to failure. It is the actual risk of failure - the product of inherent vulnerability and the exposure to trigger events - that makes it very difficult to predict. Vulnerability is a relatively static factor and is therefore easier to assess, exposure is not. They suggest a set of easily observable factors that should be considered to assess vulnerability of any slope to failure. These include:

• the presence and depth of a peat surface horizon
• the steepness and form or slope
• degree of humification of the peat
• general hydrological characteristics of the site
• if the peat does fail, are there natural anchoring points to allow it to stop?

Whilst not suggesting that this is a fully developed rigorous decision support tool, further development and refinement of the underpinning logic provides some promise.

5.4 Extent of peat erosion

5.4.1 Scotland

The extent of peat erosion has been quantified to a much greater extent in Scotland than in Wales, but there are difficulties in defining, identifying and delineating peat erosion consistently. Without a systematic image analysis approach, it is virtually impossible to get a completely consistent definition or delineation of 'eroded peat'. For this reason, there is no defined density of hags or gullies for an area to qualify as eroded peat within either the 1: 250 000 scale soil map of Scotland ( MISR 1984) nor the Land Cover of Scotland 1988 ( LCS88) ( MLURI 1988) datasets. Areas delineated as eroded peat in both these datasets include large areas that are not actively eroding, i.e. the areas that lie between the gullies that in most instances comprise most of the area. In essence what was delineated was a recognisable area of land that exhibited erosion features but with no hard and fast definition.

A number of data sources have been used ( MISR 1984, MLURI 19995 and Grieve at al 1994, 1995) to assess the extent of peat erosion and have been summarized in Towers et al (2006). The LCS88 dataset shows that just less than 6% of Scotland had eroding blanket bog which is approximately 34% of all areas of blanket bog identified. This compares with around 7.5% of Scotland or 31% of all peat categories as calculated from the NSIS. Given the differences in the datasets (point vs polygon), there is good correspondence between these figures. Given that a map unit of eroded blanket bog will have substantial areas of both bare and vegetated (that is, uneroded) peat, in many cases, the vegetated area will dominate, the actual area of eroded and bare peat will be considerably less than 6%. This also holds true for calculations based on the NSIS. The largest areas of eroded peat are along the Caithness/Sutherland border, on Lewis and Shetland and in the Monadhliaths between Loch Ness and Strathspey.

A partial resampling of the NSIS is due to start in 2007, and this may provide an indication of whether there has been any change in the incidence of erosion on different soils since the initial survey. However the intensity of erosion was not recorded and the lack of strict definitions for each erosional feature suggest that any new information may be of limited value.

Grieve et. al. (1994, 1995) quantified the area of erosion from aerial photographs in a 20% sample of the Scottish uplands. Peat erosion accounted for the greatest extent, 6% of the sample area, and the extent is very similar to that derived from analysis of the LCS88 and NSIS data. It must be emphasized again however that these data do not indicate that 6% of the blanket peat area of Scotland has been lost to erosion, but simply that erosion has affected peat soils in 6% of the area.

5.4.2 Wales

The total area of peat in Wales is between 63 500 (Rudeforth et al. 1984) and 78 000 (Taylor and Tucker 1983) hectares. The two surveys use slightly different criteria to define peat but both estimates exclude mineral soils with peaty surface horizons (organo-mineral soils). Both equate approximately to around 3.4% of the total area of the country, the figure given by CEH (2003). The distribution of peat in Wales is shown in Figure 1.2.

There is no indication of the extent of eroded peat within Wales (Rudeforth et al. 1984, CEH 2003) although both publications allude to the existence of eroded peatlands. Holden et al. (2006) contend 'that many organic soils in England and Wales are severely degraded and that most organic soils in England and Wales are degraded in some way, even if not severely'. McHugh et al. (2000) state that most erosion on upland soils was associated with peat soils, with increasingly lower incidences of erosion on wet organo-mineral and dry mineral soils similar in broad terms to that observed in Scotland (Table 5.1). It is recommended that a systematic assessment be made of the Welsh peat resource in order to get a robust estimate of the area of eroded peat.

The recent DEFRA-commissioned study of the extent of soil erosion in the upland areas of England and Wales (McHugh et al.., 2002) was a ground-based survey measuring similar parameters to those measured in the SNH-commissioned air photo study of upland Scotland (Grieve et al.., 1994, 1995). McHugh et al. (2002) found that the extent of degraded soil represented around 2.5% of the area surveyed, a smaller percentage than that reported by Grieve et al.. (1995). McHugh et al. (2002) quantified only the area of degraded soil, and thus the results are not directly comparable with the area of soil affected by erosion computed in the Scottish study. However, both studies provide clear evidence that erosion is most extensive on organic (peat) soils. Organo-mineral soils contributed about one third of the sites that displayed erosion.

5.5 Approaches to define, identify and quantify peat erosion and losses

5.5.1 Modelling of erosion susceptibility

The risk of soil erosion occurring in Scotland has been modelled following two different procedures.

• Lilly et al. (2002) used a rule based approach to identify the inherent geomorphological risk of soil erosion by overland flow
• Anthony et al. (2006) adopted a more process based approach integrating water balance models with understanding of soil erosion processes at a field scale and was designed to predict sediment and phosphate movement to waterbodies as part of a diffuse pollution screening tool.

Both approaches relied to varying degrees on national datasets of soil texture, Hydrology of Soil Types ( HOST) class and digital elevation models ( DEM) and produced output at a fairly crude scale (1km grid cells). These approaches are described in detail in Towers et al (2006).

Lilly et al. (2002) developed an erosion classification system that assumes that all soils are bare and that erosion can be modeled using the inherent characteristics of the soil to absorb water from rainfall or snowmelt. Separate decision rules were developed for mineral and organic soils and applied to each 1km grid cell in the national scale 1:250 000 soil map of Scotland which was overlain with a GIS coverage of slope categories derived from a DEM. A summary of the results are shown in Figure 5.5.

Figure 5.5 Proportion of inherent geomorphological erosion risk.

Figure 5.5 shows that the majority of mineral soils fall into the moderate risk category (just under 30% of Scotland) while organic soils are predominantly in the moderate or high risk categories (˜24 and 26% respectively). Much of the mineral soils at moderate risk of erosion by overland flow are to be found in the agricultural lowlands (Figure 5.6). However, it has to be stressed that this classification is the inherent risk and not the actual risk of erosion.

This work has been extended for upland soils where the frequency of disturbance to the vegetation cover was added. Thus soils with a semi-natural vegetation cover that was unlikely to be removed would be deemed to have a low risk of erosion even if the soils were highly susceptible. Land uses such as forestry or those involving muirburn would have a greater likelihood of bare or disturbed soil and be at risk. There is a very marked reduction in the risk of erosion occurring in upland Scotland when vegetation cover is taken into account. Areas with the highest susceptibility to erosion fell from 26 (from Figure 5.5) to 8% while the area with the lowest susceptibility rose from 2.5 (from Figure 5.5) to 42%.

The diffuse pollution screening tool (Anthony et al 2006) was developed for SEPA and EHS Northern Ireland and covered a wide range of potential pollutants amongst which was suspended sediment. This model differed from that of Lilly et al. (2002) in that it attempted to predict sediment yields. Most of the area under organic or organic-mineral soils have predicted sediment losses of less than 50 kg/ha/annum and on closer examination of the database, 84% of this area has predicted sediment losses of less than 10 kg/ha/annum and 65% has predicted losses of less than 1 kg/ha/annum. It may be that the particulate organic carbon loss is being underestimated through a lack of understanding of the processes involved compared to the removal and transport of mineral sediment.

Estimates of particulate organic carbon ( POC) release from UK uplands vary from 0.1 tonne/sq km/annum in intact Scottish peatlands (Hope et al. 1997) to about 100 tonne/sq km/annum in the southern Pennines (Evans et al. 2006). POC flux is largely controlled by the geomorphic processes that determine the transport and deposition of eroding organic material. There is some debate about the fate of POC (or eroded carbon). If it is rapidly altered to gaseous and dissolved forms in the fluvial system the carbon may be made available in a climatically active form (Pawson et al. 2006). Holden et al. (2006) argue that POC loss in severely eroding peatlands is a major, if not the single largest component of the carbon budget. Most sampling programmes have ignored POC removal from organic soils and Holden et al. recommend that further research is required on POC fluxes in degraded peatland systems.

Overall, erosion models provide a useful indication of the geographical variation in erosion susceptibility and highlighting areas potentially at risk. They also provide a framework for evaluating the effects of changing rainfall patterns or intensities. However, the predictions from soil erosion models are often insufficiently verified against field data (Brazier, 2004) and model predictions cannot therefore be seen as a substitute for measurements of actual erosion rates or of the occurrence of erosion events.

5.5.2 Surveys vs. sediment budget analysis

There have been a number of surveys, using subtly different methodologies, with the objective of quantifying the extent and severity of peat erosion. They are sampled based and have variously used National Soils Inventory ( NSI) sites in England and Wales (McHugh, 2002) and 5 km X 5 km squares in Scotland (Grieve et al., 1995). It should be emphasised that both surveys were seeking to identify and quantify all types of upland erosion including debris flows, landslides and sheet erosion as well as different forms of peat erosion.

There has been some relatively recent debate about the relative usefulness of such surveys specifically by Warburton et al. (2003) on the McHugh survey. They argue that the methods used lacked adequate definition, that there was a lack of geomorphological context to the results in that there was a lack of appreciation of the context of the 50m plots within the larger sediment system and no linking of erosion loss to timescale. They argue that an approach based on sediment budget analysis provides reliable estimates of erosion and deposition, an indication of how upland sediment systems operate, and the relative significance of slope and channel processes. They are also critical of the emphasis on the assessment of the amount of bare ground and a lack of awareness of the significance of stream channels as the main factors controlling the sediment flux from upland areas. They advocate a number of appropriately placed sediment budget assessments in the uplands rather than a rigid countrywide survey.

5.5.3 Use of air photographs/remote sensing

There is a strong temporal dimension to peat erosion processes and there is some strong evidence that the current pattern of gullies and hags may have been initiated several centuries ago. Two studies have used air photographs from two different time periods to compare and contrast and whether any significant changes in erosion extent could be tracked over that time period. McHugh (2000) found that between 1946 and 1989 both peat and mineral erosion appeared stable with approximately 60% of the erosion features in both categories showing no change in erosion. Conversely, between 55 and 60% of erosion caused by landusers and grazing animals deteriorated further in the same period. When subdivided into individual five-year periods, it was clear that observations of recovered erosion consistently equalled or exceeded those of deteriorated erosion. Air photographs also indicated that peat and mineral erosion were not important sources of continued erosion or of newly initiated degradation between 1946 and 1989. However the report acknowledges that one of the main difficulties associated with assessing losses from eroded peat may be the continued removal of material but without an associated increase in the areal extent of the feature.

Wishart & Warburton (2003) compared oblique photographs from the 1920s to those taken at the present day and a 1951 vertical air photograph with one taken in 1983. The former demonstrate that little change had occurred in the form and position of the peat margin during this period with only small areas around the projecting areas of the margin being lost. Comparison of a second sequence of oblique photographs of the same scene with an 80 years gap indicates considerable re-vegetation of previously bare peat with no increased erosion features. The comparison of vertical air photography between 1951 and 1983 showed no major changes in the pattern and dimensions of dissections systems. The only significant change is some limited re-vegetation and vegetation destruction and erosion along public rights of way. Apart from local adjustments, general changes in gully development are not detectable by this method over these timescales and strongly suggest that gullies must be much older and develop either over long time periods or during periods of greater landform change. Field work carried out in 1999 showed that the gully pattern had not changed from 1983, implying little or no change for 48 years. The authors concluded that blanket mire geomorphology in their study area, the Cheviot Hills, had remained largely unchanged for over 100 years with peat loss occurring steadily but at a rate imperceptible in terms of landform change. This concurs strongly with the observation and conclusion made by McHugh: the spatial extent of erosion is not changing, but it is highly likely that erosion has continued where it was originally initiated. Based on comparisons with the Southern Pennines, the authors speculate that the gully systems on the Cheviot may be at least 500 years old.

LIDAR ( Light Detection and Ranging) technology is a remote sensing technique that uses laser light in much the same way that sonar uses sound, or radar uses radio waves . The journey time of the laser beam, from leaving the instrument to it's return after reflection, is measured and - knowing the speed of light - a distance can be computed between the aircraft and the ground . The technology has been used to evaluate peat morphology and hydrology on National Trust Properties (Haycock 2003) and it provides a potentially powerful tool for monitoring our peat resource more widely. LIDAR potentially provides a much more sensitive method of measuring and monitoring peat depths within gully systems and hence rates of removal and also subtle changes in gully dimension. Neither of these have proved possible by conventional remote sensing techniques.

5.5.4 Field measurements of rates of peat loss

There have been rather few systematic studies into the causes of blanket mire degradation in Scotland (Coupar et al., 1997) resulting in a serious deficit of data relating to rates of peat erosion. Birnie (1993) provides the only measured loss rates in Scotland and they range from between 1-4 cm/annum. Measurements of erosion rates were made on bare peat surfaces at two hill top sites in Shetland between 1982 and 1987 and represent the worse case scenario for maximum erosion risk in the Shetland context. Hulme & Blyth (1985) suggest that up to 20 mm of peat can be removed during one storm, so the results produced by Birnie concur, in broad terms, with this estimate. It must be emphasised that this rate of peat loss only refers to situations where the surface of the peat is already bare and where there is no surface vegetation to bind the upper surface. In parts of Shetland, bare peat surfaces can extend to 15% of the blanket peat area (Birnie & Hulme, 1990), so the problem can be locally serious. Nevertheless, the vast majority of the peat resource is not bare and is not actively eroding.

In Wales, rates of peat erosion have only been calculated on Pumlumon (Yeo, 1997). Francis (1990) recorded average rates of 16 mm/yr whilst Robinson & Newton (1986) recorded rates of 30 mm/yr. Francis estimated that only between 19 and 44% of the observed surface recession could be attributed to sediment loss, the majority being caused by peat wastage due to oxidation, shrinkage, consolidation and compaction of the peat. Desiccation during summer was identified as a more important erosive agent than frost heave during the winter.

Other studies have taken place in the Pennines and they all show similar scales of loss to the Shetland and Wales studies. Tallis (1998) states that peat can be lost at rates up to 5cm/year either from the whole surface or along lines of preferred water flow. Here again, losses of this scale can occur only once bare peat is exposed so it is important to maintain vegetation cover to prevent new erosion. Anderson et al. (1997) have differentiated between rates of surface-lowering (mm/year) on gently sloping peat flats (28.7, range 16.0 -53.1, n=10), gully sides (11.9, range 7.8 - 20.4, n=5) and gully floors (5.5, range 2.5 - 8.1, n= 3). Warburton (2003) studying the combined effect of water and wind ('wind-splash erosion') on a sparsely vegetated peat surface calculated losses of 0.46 and 0.48 tonnes/hectare in two different years, corresponding to an annual average annual surface lowering of approximately 0.5 mm, equivalent to a surface lowering of approximately 3mm in the field when the density of the surface peat is taken into account. These losses are much lower than those recorded in other studies, but most of those took place in gully situations where the erosive energy of flowing water is much greater. Warburton's study took place on a 3 ha relatively flat, sparsely vegetated peat with relatively few hags.

Holden et al. (2006) have tabulated the results from sixteen studies from England and Wales, including those cited above, that have measured peat erosion rates using erosion pins. The majority (9) report rates between 10 and 30 mm/annum and three report rates below 10. The highest rate of 73.8 mm/annum is based on a one year study. Whilst we cannot be definitive about peat losses from bare peat surfaces due to small scale site heterogeneity and flow patterns and local anthropocentric pressures, losses of between 10 and 30 mm/annum is a reasonable figure based on evidence from throughout the UK. Because of the complexity of the gully systems and the fact that some removed peat is simply re-distributed, it is very difficult to estimate carbon losses with any confidence.

There is considerable qualitative evidence indicating that in recent decades the rate of peat loss has declined for many upland UK sites. Evans & Warburton (2005) have reported a three fold reduction in sediment loss compared to 40 years ago in a small catchment in the Pennines. Similar trends were by Wishart & Warburton (2003) and Clement (2005).

5.5.5 Long-term river records

Data from harmonized monitoring provides useful insights from which long-term trends in net soil erosion losses from catchments may be inferred. Mean net erosion rates within river catchments can be determined from sediment loads carried by rivers. The harmonised monitoring data maintained by the Scottish Environmental Protection Agency ( SEPA) includes records of river flow and suspended sediment concentration measured at approximately monthly intervals since the early 1970's. These form the basis of an assessment of changes in net erosion rates over the last 3 decades, but do not quantify soil redistribution within catchments by erosion and deposition.

Figure 5.6 below demonstrate the trends of suspended sediment concentrations in a selection of five catchments where a large proportion of the catchment comprises soils with an organic surface horizon. There is no real evidence of a trend of rising sediment loads in any of the catchments. Each catchment has a few abnormally high readings, but there is no real indication that they are increasing in number or intensity. If any trend is apparent it could be interpreted as one of decreasing sediment loads on the river Nairn, but this could be down to more careful management of the predominantly sandy soils on the lower cultivated land rather than any influence that the more organic soils might be exerting.

The River Findhorn is a predominantly upland catchment in eastern Scotland which was the one of the areas identified by Grieve et al. (1995) and in the LCS88 dataset with the most significant evidence of severe peat erosion but even here there is no evidence of a rising trend. These data are imperfect, in that they do not capture every extreme event, they do not differentiate between the mineral and organic carbon particulates within suspended sediments and most of the rivers in the Harmonised monitoring data are relatively large (in a Scottish context) and some of the sediment removed near the head waters is likely to be redeposited before the measuring gauge near the estuary. More gauging stations are required in subcatchments dominated by organic soils including measurements of particulate soil organic carbon.

Figure 5.6 Trends in suspended sediment concentrations ( SS) from selected catchments.

Other methods include the reconstruction of erosion rates from sediment storage sites such as lakes (Yeloff et al., 2005) or using tracers such as Cs-137 (Walling et al., 2005).

Holden et al. (2006) argue that what is missing in all these approaches is an attempt to provide a national picture of peat loss. They recommend a spatial approach to this issue to gain a better understanding of the erosional dynamics of peat systems in different areas each defined by their topography, hydrology and vegetational characteristics.

5.6 Future trends

Given the uncertainty over the initial causes and subsequent factors involved in peat erosion, it is difficult to be too definitive about potential future trends. However, some commentary is offered on the potential future role of the key triggers into the future. There is a much greater environmental awareness of the role of blanket peat in terrestrial carbon budgets and climate change mitigation strategies and there has been significant changes made to policies and to management guidelines on activities that might adversely affect our peatland resource. These are outlined below but there may also be influences that are much more difficult to control.

5.6.1 Burning

If best guidance is adhered to, the Muirburn Code (Scottish Executive 2004), the Heather and Grass Burning Code ( MAFF 1994 on behalf of the Welsh Office Agricultural Department), and the 2005 consultation on it, should prevent any future burning of blanket mires or peats. The exception to this is peat where heather constitutes more than 75% of the vegetation cover but these are extremely rare. Heather burning is normally carried out on peaty podzols that support heather-dominant vegetation, and one of the objectives of the guidance is to prevent or minimise damage to the underlying soils

The main future threat to blanket peat from fire are likely to be those related to accidental or deliberate fires started by recreational users. Recreational pressure is growing constantly and although the pressure may be greatest in the Pennines due to the proximity of large urban centres, the value of peatlands for bird watching and a 'wilderness experience' is growing. This may lead to increased visitor numbers even on more remote bogs and an increased risk of wildfire occurrences. Fire risk may also be increased if climate change produces long warm dry spells causing desiccation of the peat surface.

5.6.2 Grazing

There is increasing recognition that many of Britain's uplands have suffered from over-stocking by domestic livestock and more recently by wild herbivores such as red deer. In recent years, the number of sheep in Scotland has declined quite dramatically from a peak of almost 10 million in 1991 to 7.6 million in 2006, representing a drop of almost 25%. Similar trends have been found in Wales. In 1993, there were 11, 256,000 sheep, rising to 11, 768, 500 in 1999 but falling sharply to 9,475,000 in 2005, representing a fall of around 20%. Cattle are confined predominantly to enclosed land and mineral soils so equivalent numbers have not been presented. It is very difficult to predict whether these trends will continue but as farming continues to move from an industry focused on production to one that is expected to deliver a range of environmental benefits, certainly any reversal of this trend is highly unlikely.

Estimates of the number of red deer in Scotland are difficult to obtain and often contentious but an estimate of 'wild deer' from the Deer Commission for Scotland's website is 600 000. In their Vision Document ( DCS 2000), the estimate of 350 000 red deer has been made. This is lower than a figure of 450 000 made by the RSPB and WWF in a report published in 2003, although it is made clear that this figure is an estimate. The same document indicates that this represents a trebling of red deer numbers over a 40 year time period at the same time as their range has declined in extent due to afforestation. The Deer Commission for Scotland is committed to achieving sustainable levels of deer in the uplands and the best prediction that can be made is that deer numbers will decline, perhaps modestly and slowly, from their current levels. More information can be found at the DCS and WWF websites below.

http://www.rspb.org.uk/scotland/policy/deer/index.asp (accessed 8th November 2006)

http://www.dcs.gov.uk/i_faq.htm (accessed 8th Nov 2006)

5.6.3 Acid deposition

Acid deposition has been implicated as one of the causal factors of peatland erosion, in particular in the Southern Pennines and might be one explanation for the severity of the problem in that area. The UK's emissions of SO ₂ peaked in 1970 and have declined by almost 70% since 1990 (Fowler 2001). In the context of International transboundary agreements and protocols for acid emissions, this causal factor has decreased considerably and is likely to continue to do so. Although N deposition has yet to show similar reductions, it may as car engine technology improves and the use of N fertilisers decreases (Holden et al. 2006).

5.6.4 Climatic factors

Climatic perturbations have been implicated in the past as a causal trigger of peat erosion and a factor in current erosion processes and there appears to be a growing consensus that many of the gully and hag systems seen today are not recent phenomena (say within the last 100 years), but may have been initiated towards the start of the last millennium, long before the bogs would have subjected to the anthropocentric influences since the Industrial Revolution. Prolonged periods (several decades to a century or two) of climatic conditions that differ from the 'norm' over the last millennium, have been identified as key trigger periods for the initiation of peat erosion (Tallis, 1997; Burt et al., 1997). There is also a strong correlation between areas of peat erosion, altitude and current adverse climate (Osvald, 1949; Rhodes & Stevenson, 1997).

Current thinking suggests that we may be entering into a period of unprecedented warming. Current models suggest a possible warming of up to 3.5 °C in summer and 2.5 °C in winter by the 2080s (Hulme et al., 2002 quoted in Scottish Executive, 2006) in a high emissions scenario. In addition, summers are predicted to become drier across Scotland, particularly in the south and east, with more severe extreme rainfall events. Similar scenarios are presented for Wales (Hulme et al., 2002). If this combination of events were to become reality, it is feasible to speculate an increase in the processes observed by Hulme & Blyth (1985) and described by Warburton (2003). Essentially these processes involve the physical removal of exposed, bare dry crusted peat surfaces, often within peat hags but also on more open exposed bare ground by wind and/or water. The combination of a warmer climate causing increased desiccation of the peat surface and an increase in severe rainfall events is likely to be the principal cause in any increase in peat erosion.

This view is supported by Holden et al. (2006) who state that the most significant risk to the physical integrity of peatlands in England and Wales is posed by climate change and they also make the same link to previous climatic perturbations. Peat bogs are located where they are largely governed by climatic influences and the major influences that cause physical erosion, i.e. water and wind, are also climatically controlled. Holden et al. also suggest that climate change may lead to a decrease in frost frequency and an extended growing season, which should promote revegetation and reduce sediment supply. There are therefore both potentially positive and negative effects of climate change on peat erosion trends and there is huge uncertainty which, if either, might predominate. Bragg & Tallis (2001) also conclude that particular care should be given to the management of our blanket peat landscapes at the present time of climatic uncertainty.

Intense rainfall events have been cited as the key trigger for bog bursts and slides (Nolan & Birnie, 2006). Most climate change scenarios suggest an increase in the frequency and duration of such events but the actual timing of these is impossible to predict except within short (days and hours) timespans and even then there is no certainty that any burst or slide will actually occur. Based on available evidence, there is potential for an increase in bog bursts and slides in the context of future climate scenarios.

5.6.5 Other factors

Blanket bog is protected under Natura 2000 and many of the major bogs are protected by a number of designations including SSSI, NNR, SAC and SPAs. Such protection should prevent potentially damaging management practices such as overgrazing, injudicious burning, extensive peat extraction (except for domestic use in certain parts of Scotland) and ploughing that may trigger or exacerbate erosion processes. Past practices that may have been making peat more vulnerable to erosion processes such as drainage is no longer being carried out and indeed is being actively reversed in many areas including the extensive peatlands of Caithness and Sutherland. Liming was practiced in the past to increase biomass production for grazing sheep, but is strongly discouraged now for the effect that it has on the blanket bog vegetation. It might have a role in encouraging vegetation growth in peatland restoration projects but as higher soil pH encourages faster organic matter turnover, it should not be used on blanket bogs.

5.7 Conclusions

It is not possible to arrive at a definitive rate of peat erosion due to hydrological and topographical differences between bogs. However, of the studies conducted, losses in the range of between 10 and 30 mm/ha/annum from bare peat surfaces can be taken as a reasonable estimate.

At present most of the factors implicated in peat erosion processes are moving in a direction favourable to reducing the rate of peat erosion. However there is huge uncertainty over the potential impact that a changing climate might have on peat erosion. Current scenarios suggest a warmer climate generally with more extreme conditions; a hot dry spell causing desiccation at the peat surface followed by an intense rainfall event provides one of the key triggers for enhanced peat erosion. Peat erosion has been linked to significant climatic perturbations in the past and there is a consensus that we may be on the cusp of a similar climatic shift at the present time. Past evidence suggests that climate change may have a significant impact on the physical integrity of organic and organo-mineral soils and in particular, on deep blanket peat.

This review has also identified gaps in our understanding and recommends additional research in the following areas:

• An estimate of the extent and location of peat erosion in Wales
• A better understanding of the physical processes that causes peat erosion
• The role of particulate organic carbon in carbon budgets in eroded peatland systems
• A spatial approach to gain a better understanding of the erosional dynamics of peat systems in different areas each defined by their topography, hydrology and vegetational characteristics.
• The potential role of LIDAR technology in monitoring peat erosion rates in a number of different bog systems

5.8 Summary

• Most work has taken place on erosion of blanket mires or peat compared to soils with shallower organic rich surface horizons such as peaty podzols or peaty gleys (organo-mineral soils)
• Peat erosion results from the complex interaction of climatic and anthropogenic influences acting over a long period of time
• Considerable debate surrounds the processes that underpin the development of peat erosion, but there is a growing consensus that it is not a recent phenomenon and that climatic shifts in the last millennium may have initiated the erosion. Other factors now operate to perpetuate the erosion processes
• There have been relatively few studies on the rates of peat erosion from bare surfaces with most having taken place in the Pennines; the few results from Scotland and Wales are reasonably consistent with these. Rates vary according to local circumstances but the majority of studies suggest losses of between 10 and 30 mm/annum. It is not possible from current knowledge or data how this translates into carbon losses due to peat erosion
• There is no hard evidence that the concentration of suspended solids in Scottish rivers have increased between the mid 1970s and the present day.
• Currently most of the factors implicated in peat erosion are such that erosion should halt or decline. The exceptions to this are national and local climatic change that may lead to an increased risk.
• More strategically placed monitoring stations in sub-catchments dominated by organic and organo-mineral soils, would inform us of trends in particulate organic carbon displacement and movement. Separation of the organic and mineral particulate within suspended sediments would indicate whether there was any change in their proportions.
• More studies are required in order to develop an understanding of the mechanics of erosion and the production of a model to predict erosion risk of carbon rich soils.
• Studies on monitoring the extent of peat erosion over time have found the rate of change is too sensitive for the use of aerial photography.
• Lidar technology, coupled with ground truthing surveys, offers potential for future monitoring of peat erosion rates.

« Previous | Contents | Next »