Scottish Government

Chapter 6 Soil Erosion

This chapter outlines the major processes of soil erosion relevant to Scotland and their significance for soil functioning and summarises the policy instruments relevant to the control of soil erosion. The major part of the chapter collates the evidence for the occurrence, spatial location and rates of soil erosion in Scotland and evaluates the major controlling factors of soil erosion, the severity of the threat and the gaps in the available evidence. Finally, the likely trends in soil erosion in the future decades are summarized.

6.1 Summary

• Soil erosion is a natural process which occurs in all soils to a greater or lesser extent. The major processes considered are water erosion, mass movements and wind erosion, although tillage displacement is increasingly recognised as a significant contributor to soil erosion rates.
• Datasets such as the NSIS and LCS88 have been used to determine the location and extent of soil erosion but these require care in their interpretation. However, the episodic occurrence of soil erosion causes significant difficulties for the systematic collection of data on erosion.
• A number of models have been produced that predict in relative terms the inherent risk of erosion by water occurring on both mineral and organic soils and of sediment loss.
• Minimising soil erosion is a GAEC requirement that farmers must meet to receive their single farm payment. The incidence of soil erosion on cultivated land may thus decrease, but this needs to be monitored.
• There is concern that predicted changes in the intensity and duration of heavy rainfall events may increase the risk of extreme erosion events. Such events pose a significant threat to river water quality.
• In upland Scotland, erosion of peat covers the largest area and this may increase under climate change. The factors controlling peat erosion and implications of peat erosion for soil carbon storage need further research.
• Predicted changes in the intensity and duration of heavy rainfall events may increase the risk of major events such as landslides. GIS analysis could be performed to identify specific features, such as settlements or transport links, most at risk from landslides.

6.2 Introduction and Description of Threat

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. Based on estimates of soil renewal rates, Kirkby (1980) proposed a soil loss tolerance value of 0.1 mm year ^-1 for the UK.

Soil erosion at rates exceeding background values is termed "accelerated erosion". In humid temperate climates such as that of Scotland, much accelerated erosion is 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-steepened, 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 to be discussed in this section are:

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

Water erosion is initiated when the rainfall rates or rapidly melting snow exceed the infiltration capacity of the soil and water runs off the surface carrying detached soil particles in suspension or as bedload. Transport is generally more effective when water flow occurs in defined channels and sheet erosion, rills and gullies effectively represent a hierarchy of severity of erosion at a site. Soils are most susceptible to erosion when they are bare, but erosion can occur in cropped fields where there is partial cover or along areas deliberately left bare such as tramlines in cereal fields. Much soil erosion in Scotland is relatively limited in its spatial occurrence and often much of the eroded topsoil is trapped at downslope field boundaries or deposited on gentle within-field slopes. Mass movements generally occur under conditions of saturation. The sediment:water ratio of mass movement sediments is much greater than that during water erosion and individual occurrences can be much larger in scale. Wind erosion is less common in humid temperate climates but can affect soils in high mountain environments and soils left bare following cultivation, particularly those of fine sandy to silty texture.

Impacts of soil erosion on soil functions and related threats

Food and other biomass production

One major impact of soil erosion is that it generally involves loss of the productive and fertile topsoil leading inevitably to a potentially significant threat to the biomass production of the soil. During the 1930s, drought in the Midwestern United States led to the "dust bowl", during which severe wind erosion of cultivated topsoils occurred across parts of Oklahoma, Texas, Kansas, Colorado and New Mexico. The dust bowl was the trigger for the foundation of the US soil conservation service, which led to the development of cultivation methods aimed at reducing soil erosion. Soil losses in Scotland tend to occur on a fairly localized scale and eroded soil is often trapped at field boundaries such as walls and hedges. In some instances farmers simply move eroded soil back upslope. Given the relatively low frequency of erosion events and the short transport distances of eroded soils any threat to the biomass production function by soil erosion in Scotland must be viewed as small. It remains the case, however, that cultivation practices and exposure of bare soil during the winter months can provide the conditions for locally severe erosion particularly in areas with sandy textured soils.

Environmental Interactions

Other on-site effects of soil erosion include loss of the carbon storage function of soils as soil losses generally come from the more organic topsoil layers. The potential implications of soil erosion for carbon storage have been discussed recently (Quinton et al., 2006). Grieve (2000) showed that carbon losses of almost 50% can occur in upland soils with a peaty surface horizon when the protective vegetation cover is lost. Also in upland areas, the incidence of erosion on peaty soils may be a contributory factor to declines in the carbon concentrations in upland soils which have been reported recently and also to the increases in fluxes of DOC from upland peaty catchments. Further research to quantify the links between soil erosion and the carbon storage function of soils will undoubtedly be needed.

Soil erosion has significant off-site effects on surface waters through:

• silting up and reduced capacity of water-supply reservoirs
• loss of fish spawning areas through the deposition of fine sediment on river-bed gravels
• contamination of river waters by nutrients (mainly phosphorus) or pesticides adsorbed on eroded sediment particles

These effects are now considered within river basin management under the Water Framework Directive ( WFD), and this has led to many policy initiatives which currently protect soils such as the Forests and Water Guidelines. This is a trend which is likely to continue in future as the implementation of the WFD continues.

Biodiversity

As with food and biomass production, loss of fertile topsoil through severe erosion can have a significant impact on soil biota and on ecosystem functioning.

Provision of a platform

Erosion can damage the built infrastructure by undermining foundations and depositing sediment. The landslides which were triggered by extreme rainfall during the summer of 2004 caused significant damage to parts of the trunk road network.

Provision of raw materials

Large scale erosion of peat represents the major threat to the soil's function in providing raw materials.

Protection of cultural heritage

Soil erosion poses a threat to archaeological features such as buried cropmarks, as loss of topsoil will ultimately lead to plough damage to such features preserved in cultivated soils.

6.3 Policy

Farmers receiving direct payments are expected to maintain their land in Good Agricultural and Environmental Condition ( GAEC) as described in The GAEC Framework for Scotland (2005). Measures to reduce the likelihood of soil erosion include the following requirements: "All cropped land over the following winter must, where soil conditions after harvest allow, have either crop cover, grass cover, stubble cover, ploughed surface or a roughly cultivated surface. Fine seedbeds must only be created very close to sowing." There are also voluntary codes of practice such as the PEPFAA code and the Farm Soils Plan that give advice on how to minimize the risk of soil erosion.

The sustainable use of soil is embedded within the Forestry Standard, the document that sets out the Forestry Commission's stance on sustainable forestry and guidance is given with the Forests and Soil Conservation Guidelines (Forestry Commission 1998, update in press). In addition, the Forests and Water Guidelines include several measures to prevent the movement of eroded soil from a site to watercourses.

6.4 Evidence

6.4.1 Current Status - State of Soil

Evidence of threats to the current status of the soil resource has been gathered from a number of sources:

• National Soils Inventory for Scotland ( NSIS): Although not explicitly designed as a survey of soil erosion, evidence of soil erosion was recorded at Inventory points in the field. The major advantage of this approach was the national coverage and objectivity of the grid sampling scheme, but sites were generally visited from April until October when erosion in lowland areas is least evident.
• Land Cover of Scotland (LCS88): This survey, which was based on aerial photograph interpretation, recorded the extent of eroded blanket bog and montane vegetation, but provides no evidence of other forms of erosion in the uplands or of erosion in lowland areas ( MLURI,1993).
• Commissioned surveys of erosion incidence: SNH commissioned a survey of erosion in upland Scotland based on further interpretation of LCS88 aerial photographs. Results relate to a stratified random sample of 20% of the upland area. A similar field-based survey of erosion in upland areas in England and Wales commissioned by DEFRA provides further data on the threat of erosion to upland soils. The Forestry Commission has also undertaken a survey to monitor the degree of soil erosion explicitly associated with harvesting operations.
• The Scottish Road Network Landslides Study (Scottish Executive 2005): This provides the most recent analysis of the occurrence and risk of landslides in Scotland. Although the study was focussed on assessing the potential for landslides occurring adjacent to the road network, many of the principles and procedures outlined in the report can be extrapolated to the wider landscape.
• Case studies reported in scientific papers: These range in scope from reports of specific incidences of erosion, generally following extreme precipitation events to surveys of erosion within defined areas. From the perspective of determining the extent of the erosion threat nationally, the major disadvantage of such case studies is sampling bias, as these focus on more extreme events. However they do allow quantification of the potential severity of the problem.
• Long-term river records: Mean net erosion rates within river catchments can be determined from sediment loads carried by rivers. The harmonised monitoring data maintained by the Scottish Environment Protection Agency ( SEPA) includes records of river flow and suspended sediment concentration measured at approximately monthly intervals since the early 1970s. These form the basis of an assessment of changes in net erosion rates over the last three decades, but do not quantify soil redistribution within catchments by erosion and deposition.
• Modelling of erosion susceptibility: The potential risk of soil erosion in Scotland has been modelled using two different procedures. Lilly et al. (2002) used a rule-based approach to identify the inherent geomorphological risk of soil erosion while Anthony et al. (2006) used a process-based model to predict sediment and nutrient movements to waterbodies.

6.4.2 National Soils Inventory for Scotland ( NSIS)

The National Soils Inventory for Scotland ( NSIS) is an objective sample of Scottish soils. Soil and site conditions of 3094 locations throughout Scotland (although the NSIS points in the Orkney Islands were not sampled) were available and the presence or absence of erosion features was recorded at 2845 of these points (Table 6.1). The vast majority of sites (86%) were not eroding at the time of the sampling. Gullying and rill erosion were the most frequently recorded erosional features. Rill erosion was recorded at 100 sites with either peat soils or organo-mineral soils suggesting that gully erosion was either extending upslope or at an incipient stage. Only 9 sites with erosion were on cultivated land (<0.5%). This may partly reflect the timing of data collection which would have been mainly during the period April to October, as any winter erosion would have been remediated prior to the site visit. Case studies also indicate that erosion of cultivated soils is localised in its occurrence; national monitoring schemes are therefore unlikely to record it adequately.

Table 6.1: Number of inventory points with erosion features recorded grouped by land cover type.

Wind erosion was primarily found on mountain tops (soils were alpine or oroarctic podzols) or coastal links soils (regosols). The occurrence of erosion features at survey points was also compared with the soil type found at each point. Peat covers around 22% of the land area of Scotland and, according to the NSIS, almost a third of the peat sites visited were eroding.

Overall, the NSIS provides an objective sample of 5 main types of erosional features and covers all of Scotland (apart from the Orkney Islands). The regular grid pattern allows areas of erosional features to be determined. However, although the presence or absence of erosion was recorded at each site, the actual area and depth of gullying was not quantified and the severity of erosion cannot therefore be determined. The usefulness of the survey is also constrained by the prolonged period over which the data were collected (around 10 years), the timing of the sampling (spring to autumn) and the lack of strict definitions for each erosional feature.

6.4.3 Land Cover of Scotland 1988 (LCS88)

The Land Cover of Scotland 1988 (LCS88) dataset comprises 126 main categories of land cover identified from air photographs of approximately 1:25,000 scale. The photographs were taken primarily in 1988 though a few were taken in 1989. The photographs were interpreted by a team of skilled interpreters who had extensive field experience in matching tonal patterns on air photographs with vegetation communities in the field. Erosional features were identified in only two categories of land cover: eroded blanket bog and eroded montane vegetation.

The LCS88 dataset shows that just less than 6% of Scotland had eroding blanket bog which is approximately 34% of the total area of blanket bog identified. This compares with around 7.5% of Scotland or 31% of all peat categories as calculated from the NSIS. The area of erosion in the montane zone is calculated as around 3% of Scotland from the LCS88 but only 0.5% from the NSIS. This may be explained by the fact that all montane land cover classes were considered to have erosional features, which may not always be the case. Given that a map unit of eroded blanket bog will have substantial areas of both bare and vegetated (that is, uneroded) peat, the actual area of eroded and bare peat will be less than 7.5%. This also holds true for calculations based on the NSIS.

6.4.4 Commissioned surveys of erosion incidence

Grieve et al. (1994; 1995) quantified the area of erosion from aerial photographs in a 20% sample of the Scottish uplands. In total, approximately 12% of the sampled area had erosional features recorded. Peat erosion accounted for the greatest extent, 6% of the sample area, and the extent was very similar to that derived from analysis of the LCS88 and NSIS data. It must be emphasized however that these data do not indicate that 6% of the upland area of Scotland has been lost to erosion, but simply that erosion has affected peat soils in 6% of the area. The actual area of eroded peat will be significantly smaller than this.

Figure 6.1 Totals of a) deer and b) sheep numbers in Scotland in recent decades (From Hunt, 2003; Scottish Executive Abstract of Scottish Agricultural Statistics 1982-2003).

6.1a) Scottish deer population totals.

6.1b) Scottish sheep population totals.

The greatest occurrence of peat erosion (20% of samples) was found in the Monadhliath Mountains. However the occurrence of the most severe class erosion was greatest in Eastern parts of the country, particularly the eastern Grampians, where land use and grazing pressures were considered to be greatest. This spatial association of erosion and land use pressure, together with the trend of increasing deer numbers in the last four decades (Fig. 6.1a) indicates a significant area of concern over erosion in the uplands. Grazing by sheep is also often implicated in soil erosion in the uplands and the number of sheep grazed in Scotland rose by almost 2,000,000 in the 1980s before leveling off, and then more recently declining following the outbreak of Foot and Mouth Disease in 2001 (Fig 6.1b).

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 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.

Although the areas affected vary, the very clear message emerging both from national data sets and from systematic studies of soil erosion at the national or regional scale is that the incidence of soil erosion in upland areas is greatest on peat soils. Given the significance of peat soils for terrestrial carbon storage and biodiversity, erosion of peat soils in upland areas of Scotland remains a subject of major concern.

6.4.5 The Scottish Road Network Landslides Study

This study (Scottish Executive, 2005) recognised five distinct types of landslide varying in the nature of displacement and the liquid content of the sediment. However, most landslides that occur in Scotland are of the flow type, a spatially continuous movement of a material as a viscous fluid. Flow type landslides include the particularly dramatic ones that occurred in the summer of 2004 and disrupted parts of the Scottish trunk road network. It should be noted that this type of event is neither new nor uncommon; events have been recorded since 1744 and more recently events have occurred throughout the Northern, Western and Central Highlands although not on the scale of those of 2004.

The key contributory factors to debris flow occurrence were identified from a comprehensive literature review and a workshop of experts from different fields. The factors influencing debris flow occurrence were categorised into three types:

1. Hazard factors affecting debris flow occurrence. These include topographic factors (slope angle, height, aspect), geological, geotechnical and hydrogeological factors (geological formation, landslide history, likelihood of earthquakes, shear strength, void ratio, relative density and permeability and surface drainage), meteorological factors (rainfall, snow melt) and vegetation and land use.

2. Hazard factors affecting debris flow runout. These include slope angle, height and magnitude (affecting the volume of material delivered to deposition zone), channel characteristics and vegetation and land use.

3. Factors affecting exposure to debris flow hazards. The key factor in relation to the exposure that results from a debris flow is whether or not the flow reaches a vulnerable element in this case a trunk road or associated infrastructure. Clearly, if there is no possibility that the flow will reach a trunk road (or associated infrastructure) then both the hazard and the hazard ranking become, for the purposes of this study, zero. Factors such as road usage and emergency response times were assessed here but they are not particularly relevant to the present study.

A GIS screening analysis of likely debris flow occurrence and subsequent exposure and hazard was carried out. As a first pass, there are three critical factors that could be obtained rapidly and remotely from a GIS to assess whether these conditions are in place:

• A source area where the slope angle is greater than 26° and less than 50°.
• A run-out zone where the slope angle is greater than 8°.
• A trunk road is present within either of the above zones.

It should be noted that peat can flow at much lower angles than these and it would be appropriate also to perform an alternative first pass in which a search is carried out for all trunk roads passing through areas of peat.

The identification of high risk areas was then interfaced with an assessment of exposure to that risk, primarily road usage. In this way some key stretches of the trunk road network were identified as being of high perceived hazard. Management and mitigation options were also identified. Overall the report highlights the importance of water as the main contributory factor triggering debris flow events. Climate change models for Scotland in the 2080s indicate that summer precipitation will decrease but winter precipitation increase. However, summer storms are believed to be at least partially responsible for triggering the events of August 2004, and climate data may not give a full picture of the relationship between precipitation and landslides.

6.4.6 Case studies reported in scientific papers

Table 6.2 (Davidson and Grieve, 2004) summarises data from event-based studies of soil erosion from a range of sites in Scotland. Most instances follow high-magnitude, relatively rare, rainfall events and affected soils which were bare, for example, due to the cultivation of winter cereals. Many studies also linked incidence to direction of cultivation ( e.g. Davidson and Harrison, 1995). Although these data indicate that erosion of lowland soils does occur, there has been some debate over whether erosion rates should be considered significant. Frost and Speirs (1996) considered that on soils derived from soft sediments such as glacial tills, the impact of soil erosion on in-field soil quality was questionable. However, other on-site effects such as threats to underlying archaeology are also relevant (Davidson et al., 1999), and off-site effects, particularly the movement of eroded soil to water courses are of major significance.

Table 6.2: Event-based studies of erosion of lowland agricultural soils in Scotland.

Studies in England and Wales underline the importance of winter cereals for soil erosion and the greater occurrence of erosion on sandy textured soils (Brazier, 2004; Quinton and Catt, 2004; Evans and Brazier, 2005). Evans (2005) reported that it is unlikely that the severity of erosion in England and Wales has increased within the last 20 years but the evidence also underlines the need for better schemes for monitoring erosion.

Evidence of absolute rates of soil erosion over longer time periods is available from studies using the caesium isotope ¹³⁷Cs as a tracer. ¹³⁷Cs was added to soils from atmospheric testing of atomic weapons and events such as the Chernobyl incident in 1986. The isotope is strongly adsorbed to soil clay particles. Redistribution of soil particles following ¹³⁷Cs deposition provides an estimate of absolute rates of soil redistribution during the last 40 years. Bowes (2003) mapped soil losses from soil ¹³⁷Cs inventories in 25 m square cells along transects across cultivated fields on sandy soils in east central Scotland. Net annual soil redistribution in all the fields sampled ranged from losses of 3 kg m ^-2 to gains of 6 kg m ^-2. However it must be stressed that these losses apply to small areas. Field boundaries effectively trap much of the eroded soil within the field and net losses from the field were much lower. This may partly explain why ¹³⁷Cs estimates of erosion rates are often greater than data obtained from surveys of larger spatial units (Brazier, 2004).

Estimates of erosion rates using ¹³⁷Cs budgets have also led to increased recognition of tillage displacement as an important contributor to net erosion within cultivated fields. While tillage erosion is not as important as water erosion in most landscapes (Van Oost et al., 2005), it can lead to significant localized soil thinning within fields, for example on convex slopes. Such soil thinning represents a particular threat to archaeological cropmarks in particular areas (see chapter 9). Davidson et al. (1998) estimated that a mean erosion rate of about 0.5 mm year ^-1 could lead to damage to a cropmark site on a convex slope in Perthshire within a few decades.

Case studies of relevance to the erosion of soils in the uplands have considered the effects of forestry management and grazing by herbivores on erosion. Most erosion from forest areas occurs is linked to disturbance during the planting and harvesting phases (Stott and Mount, 2004), often originating in ditches and stream banks. However, this is a temporary phenomenon and a recent study in Wales has shown that recovery occurs within 4 years as revegetation of exposed banks occurs (Stott, 2005). Several studies have shown that the Forests and Water Guidelines are generally effective in limiting soil damage and minimizing the effects of forest operations on sediment inputs to streams (Nisbet et al., 2002) although Carling et al. (2001) highlighted several areas where further research is needed, most notably in understanding the long-term sustainability of soil structure through several forest crop rotations.

The importance of wind in peat erosion has also been highlighted in studies in the Northern Pennines (Warburton, 2003). Wind erosion of peat normally occurs when the peat surface dries and cracks. Surface layers can then become detached and be blown in strong winds. However, Hulme and Blyth (1985) observed erosion of these surface layers from erosion channels by water during a severe thunderstorm. The thickness of these layers ranged from 1 to 20 mm and it was reported that almost all were removed during the one hour storm. Birnie (1993) assessed erosion rates from an exposed hill peat in Shetland using pins driven into the peat surface. Although the results were variable and complicated by the loss of pins through trampling by stock, he estimated erosion losses of between 1 and 4 cm year ^-1 due to the effects of trampling and rubbing by sheep as well as geomorphic and weathering processes.

The effect of grazing pressure on erosion rates on the Trotternish Ridge in north-east Skye was monitored between October 1998 and October 2003 (Waterhouse et al., 2004). The land is steep with inherently high erosion rates and is used as common grazing by crofters. Using erosion pins, average gully extension was measured at 105 mm a ^-1. Gully areas where rabbits were active showed the greatest erosion as a result of both burrowing and overgrazing. Trampling by sheep prevented revegetation of disturbed sites though the natural geomorphic processes also contributed to soil creep and mass movement.

6.4.7 Long-term river records

Data from harmonized monitoring provides some insights from which long-term trends in net soil erosion losses from catchments may be inferred, although the low frequency of sampling (monthly) and gaps in data sets limit the usefulness of these data for assessing sediment fluxes. Figure 6.2 shows trends in suspended sediment concentrations since 1975 for rivers draining two catchments in E Lothian and Fife. If there were any drastic change in erosion in these intensively arable areas we might expect to see this reflected in greater suspended sediment concentrations and there is no clear evidence of such a trend in these graphs. Suspended sediment concentrations in rivers are strongly related to discharge (flow). For the River Eden, the regression of concentration on discharge is highly significant (r ² = 0.391, p < 0.001), but date was not significant when added as an additional variable, suggesting that the relationship between sediment concentration and discharge has not changed over the 25 year period.

Figure 6.2: Trends in suspended sediment concentrations ( SS) for two Scottish rivers.

a)

b)

6.4.8 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, whilst
• 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, HOST class and digital elevation models ( DEM) and produced output at a fairly crude scale (1 km ² grid cells).

The inherent geomorphological risk approach (Lilly et al., 2002) assumes that all soils are bare and that erosion can be modelled using the inherent characteristics of the soil to absorb water from rainfall or snowmelt. The likelihood that a soil would erode was determined according to the erosive power generated by soils becoming saturated and initiating overland flow. The steepness of the slope would determine the erosivity of this flow. As much of the information available on the mechanisms of soil erosion is for mineral soils, a pragmatic approach to assessing the likelihood of erosion on upland organic and organo-mineral soils had to be adopted. Anecdotal evidence suggested that peats were more susceptible to erosion than peaty soils (peaty gleys, peaty podzols and peaty rankers) even where slope and site conditions were similar. Thus peats (organic soils) were simply assumed to be at high risk (Figure 6.3).

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 were 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 greater risk. When vegetation cover was taken into account the areas with greatest susceptibility to erosion fell from 26 to 8% while the area of land with the lowest susceptibility rose from 2.5 to 42%.

Figure 6.3: Distribution of the modelled inherent erosion susceptibility to overland flow.

A similar exercise was undertaken by Vinten et al. (2005) for Scottish Parishes where winter cereals were grown. Soils under winter cereals are thought to be at greater risk of erosion than those under spring cereals due to the prevalence of bare soil during the winter when the soils are at or near field capacity. They combined the proportion of soils with a moderate or high erosion risk in a parish as predicted by Lilly et al. (2002) with that of the proportion of winter cereals grown in a parish and calculated the area of land at risk of erosion (Table 6.3). This has not however been extended to include other crops.

Table 6.3: Soil erosion classes in Agricultural Land Use Categories with tilled land.

The diffuse pollution screening tool developed by Anthony et al. (2006) 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. The screening tool incorporated a landscape connectivity index based on soil runoff characteristics, slope angle and shape and soil texture. The landscape connectivity model was then combined with a water balance model and a model of soil detachment and transport within the PSYCHIC modeling framework ( DEFRA Project PE0202). The map output is at a resolution of 1 km ² grid cells and the amount of sediment includes point source contributions from sewage treatment works discharges and septic tanks (Figure 6.4 and Table 6.4)

Table 6.4: Modelled total annual sediment losses (tonnes per year) to surface waters, by source.

Figure 6.4: Spatial distribution of the total modelled annual sediment loss from point and diffuse sources to ground and surface waters in Scotland.

Overall erosion models provide a useful indication of the geographical variation in erosion susceptibility, highlighting areas potentially at risk and areas where the environmental function provided by soils where rainfall infiltrates may be inadequate to protect surface waters. They also provide a framework for evaluating the effects of changing rainfall or cropping on potential erosion rates. However, the predictions from soil erosion models are often insufficiently verified against field data (Brazier, 2004). There is thus a clear need for targeted measurements of actual erosion rates for calibration of model predictions to increase confidence in the outcomes from soil erosion modelling.

6.4.9 Current status of threat and gaps in data/evidence

There is a substantial body of evidence on the processes of soil erosion and the factors controlling these processes in lowland Scotland. Severe erosion of cultivated soils is highly episodic and localized in its occurrence. Studies have highlighted the importance of land use practices which leave soil surfaces bare during the winter months and the particular susceptibility of sandy-textured soils to erosion in the lowlands. Long-term river records do not suggest an increase in such erosion in recent decades although monthly sampling is too infrequent to prove this conclusively. Compliance with GAEC conditions and the requirements of the WFD should also minimize soil erosion, although collection of data on the incidence of erosion will necessary to monitor the effectiveness of such policy instruments. In a similar manner, case studies relevant to the uplands have also underlined the success of policy instruments such as the Forests and Water Guidelines in reducing the transfer of eroded soils to water courses.

In the uplands, surveys of erosion incidence both in Scotland and in other parts of Great Britain have shown that peat soils are the type with the greatest area of erosion, linked in part to their inherent susceptibility and in part to pressures such as large numbers of grazing animals. This must be seen as a significant threat to upland soils, given the importance of peat soils particularly for carbon storage. Surveys have provided useful data but there remains a significant gap in our knowledge of the current status of erosion on peat soils, since the surveys reviewed here are based on aerial photography from the late 1980s. There are also uncertainties over the mechanisms of erosion in peat soils as well as the links between land use and peat erosion. The evidence for such a link is based on a very limited analysis of the spatial association of severe erosion and greater stocking densities. This would benefit from a more rigorous analysis in order that stocking densities consistent with maintaining the integrity of peat soils can be established. A greater understanding of the mechanisms leading to the initiation of erosion in peat soils would also be of benefit in this respect.

There is also a need for better determination of erosion rates so that predictions from erosion models can be properly calibrated and verified. This could be achieved by a limited programme of field monitoring at research sites together with determination of mean long-term rates for specific soil types and land uses using techniques such as ¹³⁷Cs inventories.

6.5 Future trends

Policies such as the GAEC requirements, the Forests and Water Guidelines and the Water Framework Directive are explicitly directed towards minimising soil erosion and should reinforce recent trends in reducing erosion of agriculture and forest soils. However, future trends in soil erosion linked to extreme weather conditions are more difficult to predict. Climate models suggest an increase in rainfall for Scotland, and major flood events have attracted considerable media attention in the last few years. A recent analysis predicts that for the UK as a whole, event magnitudes at a given return period will increase by 10% for short-duration (1-2 day) events and by up to 30% for longer frequency (5-10 day) events (Ekstrom et al., 2004). Analysis of extreme rainfall data for the period 1961 to 2000 indicate that event magnitudes have significantly increased particularly for Scotland. For example the 50 year rainfall event in Scotland has become an 8 year event for Eastern Scotland and an 11 year event for Southern Scotland during the analysis period (Fowler and Kilsby, 2003).

Such predicted increases in the frequency of extreme rainfall events suggest that soil erosion incidence linked to extreme weather conditions may increase in the next few decades. It is therefore expected that erosion events similar to those tabulated previously (Table 7.2), landslides such as occurred in the autumn of 2004 and bog bursts like those seen on Shetland in 2003 and other significant erosion of peat may increase in frequency in the future. However, climate models generally predict average conditions and the errors on the predictions are substantial. Models are incapable of predicting localised summer storms. It is therefore difficult to suggest policy changes by which such the effects of such events may be avoided. Agricultural and forestry policies aimed at reducing the erodibility of soils and runoff from soils will undoubtedly limit the impact of extreme events. Predicting and mapping the areas most susceptible to such events will also enable better targeting of protective measures for infrastructure such as roads.

6.6 Conclusions

• Soil erosion at rates exceeding background rates is termed accelerated erosion. In Scotland accelerated erosion has generally been the result of human activities which lead to removal of the protective vegetation cover. Tillage also increases soil erosion rates through downslope displacement of soil.
• Accelerated erosion of soils has important implications for a number of soil functions. Loss of the more organic topsoil represents a threat to both the biomass production and carbon storage functions of the soil. Erosion also represents a threat to the filtering and buffering function of the soil, leading to important off-site effects such as contamination of rivers with sediment and associated nutrients and pesticides.
• Current policies aimed at limiting soil erosion include the expectations within the GAEC Framework and the Forests and Water Guidelines. The Water Framework Directive also places emphasis on reducing loss of soils to surface waters. Voluntary codes such as PEPFAA (Prevention of Environmental Pollution From Agricultural Activity) and the Farm Soils Plan will also help reduce erosion.
• In the lowlands, studies of individual erosion events highlights the susceptibility of bare sandy soils during the winter months, and the significance of practices such as cultivating directly up and down slopes. National inventories do not adequately quantify such events, which pose a significant threat to the quality of surface waters and to archaeological cropmark features. We therefore recommend monitoring of erosion events to ensure the success of agricultural policies designed to reduce erosion.
• Surveys of erosion in the uplands have shown that peat soils are particularly susceptible by erosion, with estimates that some 6% of upland Scotland is affected. Erosion of peat soils in upland Scotland poses a significant threat to the carbon storage function of these soils, and peat erosion remains a subject of major concern. Further research into the link between climate and land use change and peat erosion and the significance of peat erosion for carbon loss is urgently needed.
• The evidence also highlights the importance of extreme rainfall events in the occurrence of events such as the severe landslides which occurred in the late summer of 2004.
• Models of soil erosion provide an indication of geographical variations in erosion and areas potentially at high risk. There is a need for better quantification of actual erosion rates to calibrate the models and validate their outcomes.
• Climate change models predict that Scotland will become warmer and wetter particularly during the winter months. If the incidence of extreme rainfall events increases, there are likely to be more frequent occurrences of erosion events, including erosion of bare cultivated soils, landslides and large scale peat erosion. Hazard mapping offers a possible means of targeting protective measures.

6.7 References

Anthony, S., Betson, M., Lord, E., Tagg, A., Panzeri, M., Abbott, C., Struve, J., Lilly, A., Dunn, S., DeGroote, J., Towers, W. and Lewis, D. (2006) Provision of a screening tool to identify and characterise diffuse pollution pressures: Phase II. Final Report WFD19 (230/8050). Sniffer.

Ballantyne, C.K. (1991) Holocene geomorphic activity in the Scottish Highlands. Scottish Geographical Magazine. 107: 84-98.

Birnie, R.V. (1993) Erosion rates on bare peat surfaces in Shetland. Scottish Geographical Magazine. 109: 12-17.

Bowes, J. (2003) Soil erosion on buried archaeological sites in arable areas: a modelling approach. Unpublished PhD thesis, University of Stirling.

Brazier, R. (2004) Quantifying soil erosion by water in the UK: a review of monitoring and modelling approaches. Progress in Physical Geography. 28: 340-365.

Carling, P.A., Irvine, B.J., Hill, A. and Wood, M. (2001) Reducing sediment inputs to Scottish streams: a review of the efficacy of soil conservation practices in upland forestry. Sci. Tot. Environ. 265: 209-227.

Davidson, D.A. and Grieve, I.C. (2004) Trends in soil erosion. Scottish Natural Heritage Commissioned Report F00AC106.

Davidson, D.A. and Harrison, D.J. (1995) The nature, causes and implications of water erosion on arable land. Soil Use and Management. 11: 63-68.

Davidson, D.A., Grieve, I.C., Tyler, A.N., Barclay, G.J. and Maxwell, G.S. (1998) Archaeological sites: assessment of erosion risk. Journal of Archaeological Science. 25: 857-860.

Duck, R.W. and McManus, J. (1987) Soil erosion near Barry, Angus. Scott. Geog. Mag. 103: 44-46.

Duck, R.W. and McManus, J. (1988) A flood deposit in an emptied Scottish reservoir. Scott. Geog. Mag. 104: 167-170.

Evans, R. (2005) Monitoring water erosion in lowland England and Wales - a personal view of its history and outcomes. Catena. 64: 142-161.

Evans, R. and Brazier. R. (2005) Evaluation of modelled spatially distributed predictions of soil erosion by water versus field-based assessments. Environmental Science and Policy. 8: 493-501.

Frost, C.A. and Speirs, R.B. (19840 Water erosion of soils in south-east Scotland - a case study. Research and Development in Agriculture. 1: 145-152.

Frost, C.A. and Speirs, R.B. (1996) Soil erosion from a single rainstorm over an area in East Lothian, Scotland. Soil Use and Management. 12: 8-12.

Grieve, I.C. (2000) Effects of human disturbance and cryoturbation on soil iron and organic matter distribution and on carbon storage at high elevations in the Cairngorm mountains, Scotland. Geoderma. 95: 1-14.

Grieve, I.C., Davidson, D.A. and Gordon, J.E. (1995) Nature, extent and severity of soil erosion in upland Scotland. Land Degradation and Rehabilitation. 6: 41-55.

Grieve, I.C., Hipkin J.A. and Davidson D.A. (1994) Soil erosion sensitivity in upland Scotland. Scottish Natural Heritage Research, Survey and Monitoring Report, No.24.

Hulme P.D. and Blyth A.W. (1985) Observations on the erosion of Blanket peat in Yell, Shetland. Geografiska Annaler. 67A: 119-122.

Hunt, J.F. (2003) Impacts of Wild Deer in Scotland - How Fares the Public Interest? Report prepared for WWF Scotland and RSPB Scotland.

Kirkbride, M.P. and Reeves, A.D. (1993) Soil erosion caused by low-intensity rainfall in Angus, Scotland. Applied Geography. 13: 299-311.

Kirkby, M.J. (1980) The problem. In M.J. Kirkby and R.P.C. Morgan (ed.). Soil Erosion. John Wiley, Chichester.

Lilly, A., Birnie, R.V., Hudson, G. and Horne, P.L. (2002) The inherent geomorphological risk of soil erosion by overland flow in Scotland. Scottish Natural Heritage Research, Survey and Monitoring Report No. 183.

McHugh, M., Harrod, T. and Morgan, R. (2002) The extent of soil erosion in upland England and Wales. Earth Surface Processes and Landforms. 27: 99-107.

MLURI. (1993) The Land Cover of Scotland 1988. Macaulay Land Use Research Institute. Aberdeen.

Nisbet, T.R., Welch, D. and Doughty, R. (2002) The role of forest management in controlling diffuse pollution from the afforestation and clearfelling of two public water supply catchments in Argyll, West Scotland. Forest Ecology and Management. 158: 141-154.

Quinton, J.N. and Catt, J.A. (2004) The effects of minimal tillage and contour cultivation on surface runoff, soil loss and crop yield in the long-term Woburn Erosion Reference Experiment on sandy soil at Woburn, England. Soil Use and Management. 20: 343-349.

Quinton, J.N., Catt, J.A., Wood, G.A. and Steer, J. (2006) Soil carbon losses by water erosion: experimentation and modelling at field and national scales in the UK. Agriculture Ecosystems and Environment. 112: 87-102.

Scottish Executive (2005) The Scottish Road Network Landslides Study. M G Winter ( TRL Limited), F Macgregor and L Shackman (Scottish Executive) (eds)

Stott, T. (2005) Natural recovery from accelerated forest ditch and stream bank erosion five years after harvesting of plantation forest on Plynlimon, mid-Wales. Earth Surface Processes and Landforms. 30: 349-357.

Stott, T. and Mount, N. (2004) Plantation forestry impacts on sediment yields and downstream channel dynamics in the UK: a review. Progress in Physical Geography. 28: 197-240.

Van Oost, K, Van Muygen, W., Govers, G., Deckers, J. and Quine, T.A. (2005) From water to tillage erosion dominated landform evolution. Geomorphology. 72: 193-203.

Vinten, A.J.A., Towers, W., King, J.A., McCracken, D.I., Crawford, C., Cole, L.J., Duncan, A., Sym, G., Aitken, M., Avdic, K., Lilly, A., Langan, S. and Jones, M. (2004) Appraisal of rural BMP's for controlling diffuse pollution and enhancing biodiversity. SNIFFER

Warburton, J. (2003) Wind-splash erosion of bare peat on UK upland moorlands. Catena. 52: 191-207.

Waterhouse, E.C., Henderson, D.J. and Birnie, R.V. (2004) Effects of changes in grazing pressure of sheep on erosion and vegetation cover on Trotternish, Isle of Skye. Report prepared for SEERAD. Macaulay Institute. Aberdeen.