This report provides an assessment of the geomorphological feasibility of sand dune reactivation at Kenfig Burrows. The report is based on a review of relevant environmental data, published and unpublished literature, air photograph and LiDAR interpretation, a site visit and discussion with local CCW personnel in mid March 2011, and laboratory analysis of a limited number of sand samples collected during the site visit.

Several factors need to be considered before any decision is made regarding possible dune reactivation works, including the potential impacts of existing habitats and various stakeholder interests (e.g. grazing rights), the possible impacts on neighbouring areas outside the NNR, and the requirements for consents / planning permission. While recognizing the importance of these issues, the present study is concerned only with the geomorphological feasibility of dune reactivation.

At the present time, the Kenfig dunefield is a sediment starved system which experiences a moderate wind energy regime. High vegetation density is favoured by relatively light grazing pressure, a high regional water table, dispersed visitor pressure, and historically low rates of shoreline erosion. The greatest potential for creating areas of mobile dune exists close to the shore in the northern half of the site where a relatively large ‘reservoir’ of sand is stored within the frontal dunes. Existing blowouts in this area could be artificially enlarged with the aim of creating long sand transport corridors which link up with existing stable, or largely stable, parabolic dunes further inland. The potential for sand mobilization on a large scale would be enhanced if the vegetation and dune re-profiling works are carried out in combination with a programme of local beach or dune sediment nourishment.

Kenfig Burrows is located between Port Talbot and Porthcawl on the southeastern side of Swansea Bay (Figure 1). In terms of topographic setting the dunefield lies to the southwest of a major upland area and northwest of a smaller area of high ground located to the west of Bridgend (Figure 2). The northern boundary of the dune system is defined by the River Kenfig (Afon Cynffig), although dunes, now largely built upon, continue on the north side of the river as Margam Burrows, Aberavon Burrows and Baglan Burrows. The southern side of the Kenfig dune system is formed by high ground inland of Sker Point, part of which is also covered with a veneer of wind-blown sand. To the east and southeast of the dunefield is an escarpment over which blown sand has begun to climb (Figure 3a). At the northern end of the escarpment dunes and sand sheets have reached up to 3.2 km inland from the shore. Along the northern part of the dunefield eastward movement of blown sand has been impeded by the Kenfig River.

Most of the dune complex is underlain by Keuper Marls of Triassic age, but in the area around Kenfig Pool the Marls are overlain by glacial till (Price & Brooks, 1980). Kenfig Pool is a large (28 ha) permanent lake with a maximum depth in the centre of about 4 m. It probably has a long history, and documentary records provide evidence that it was in existence at least by 1365, when poachers were prosecuted for taking fish when the fishing rights were held by the monks of Margam Abbey (Steers, 1946). According to Evans (1960), a 16th century map shows a water course called the Pool Gutter which ran from Pool westwards to the sea. However, according to Orr (1911, 1912), in the early 19th century the Pool was essentially a marshy area which drained northwards towards the Kenfig River. This stream may have been blocked by migrating dunes later in the 19th century, and during the 1940’s periodic surface overflow again flowed westwards directly to the sea (George, 1943, cited in Steers, 1946). Comparison of aerial photographs taken in 1946, 1971 and 1992 showed that the size of the Pool progressively decreased slightly over that period, with accompanying development of marginal Phragmites reedbeds and Salix carr woodland (Jones, 1995; Jones et al., 1996).

In the area to the west of Kenfig Pool blown sand is restricted to a relatively narrow zone near the shore. The surface deposit between these dunes and the Pool are mapped as “alluvium” on the British Geological Survey 1:50,000 sheet. The absence of dunes in this area may partly reflect permanently wet ground conditions which impeded aeolian sand transport.

Archaeological and historical evidence suggests that dunes were of relatively limited extent until Norman times, when a significant settlement existed close to the position of Kenfig castle, adjacent to an old course of the River Kenfig (Higgins, 1933; Evans, 1960). Livestock grazing was practised in the surrounding area under a commons type regime which has survived to the present day. Rabbits were introduced by the Normans and were abundant until the outbreak of myxamatosis in 1954.

During the 13th century a large quantity of sand apparently arrived on the Kenfig shore and the town of Kenfig was progressively engulfed by transgressive dunes between the late 13th and 15th centuries (Higgins, 1933). The castle (built in 1152) was still occupied in 1403 but engulfed by sand shortly afterwards. The old town of Kenfig was abandoned and physically relocated around 1470. In 1528 the old church (built in 1154), glebe and adjoining land were buried by sand (Steers, 1946). Between 1514 and 1573 sand encroached on the river and main coast road, and by 1700 had reached the crest of the escarpment behind the coastal plain (Higgins, 1933; Steers, 1946). The major episodes of sand blowing and dune movement were evidently associated with periods of the Little Ice Age in which there were high frequencies of severe storms and storm surges.

No long-term weather data are available for the Kenfig area. The nearest open coast station with more than ten years of wind records is Pembrey in Carmarthenshire. Saye (2003) and Pye & Saye (2005) reported a sand drift potential (DP) at Pembrey, calculated, using the method of Fryberger & Dean (1979) and Met Office data for the period 1992-2000, of 1935 Vector Units (classified as moderate by global standards), the a resultant drift potential (RDP) of 1435 Vector Units. The resultant drift direction at Pembrey is 54° (towards the NE) and the wind regimes is classed as wide unimodal, with an RDP/DP ratio of 0.74 (Figure 4). Although no wind data are available for the Kenfig area, it is evident from an examination of the general topography and dune morphology that the local wind regime differs somewhat from that at Pembrey. Kenfig is partially sheltered from southwesterly by the high ground of the Southwest Peninsula, and consequently winds from a westerly direction assume greater relative importance. The local pattern of airflow is also significantly influenced by the topography inland from the coast, with funnelling of winds around the areas of higher ground and along the valleys (Figure 2).

In the southern part of the area the higher ground escarpment reaches the coast at Sker Point and extends away from the shore in a northeasterly direction. The proximity of the high ground causes a reduction in average wind velocities in the southern part of the dunefield compared with the northern half, where there is also local acceleration of airflow towards Pyle and the Kenfig River valley. The effect of these local topographic influences is that parabolic dunes at the northern end of the dunefield have long axes oriented almost SW-NE, those in the centre have long-axis orientations of WSW-ENE and W-E, while those in the extreme south have long-axis orientations of WNW ESE and NW-S (Figure 5).

The dunefield consists of several generations of transgressive parabolic and elongate parabolic dunes, the arms of which separate a series of dune slacks, some of which contain standing water during the winter months. The migration of successive dunes has led to partial reworking of older dunes, leaving residual mounds and ridges on the margins of the deflation corridors.

In the southern part of the dunefield the dune crests typically reach 18 to 19 m OD, but in the higher energy northern area they typically reach 21 to 24 m. In the area formerly known locally as ‘The Desert’ on account of the presence of a large area of bare sand until stabilization works were undertaken in the late 1960’s, the highest dune crest reaches to almost 30 m OD.

A number of sections across the dunefield, taken from the LiDAR digital elevation model of the area, are shown in Figures 6 to 10. A north-south section along the highest point of the frontal dunes is shown in Figure 11. In general the frontal dunes in the north are much higher in the north due to greater sand availability and stronger winds. One major blowout in the frontal dunes is present at chainage 1200 to 1300 m.

Swansea Bay and the wider Bristol Channel experience a macro-tidal regime. At Porthcawl the tidal range is 8.9 m of mean spring tides and 4.2 m on mean neap tides, creating a relatively wide foreshore at Kenfig Sands (sometimes referred to Sker Sands). The elevation of mean high water spring tides is approximately 4.6 m OD and that of mean high water neap tides approximately 2.2 m OD.

Waves recorded at Scarweather Sands, approximately 6 km west of Porthcawl, approach the coast mainly from the west and southwest (Figure 12). The resultant approach direction is from the west southwest and the resultant energy transfer direction is to the east-north-east (79°). Waves approaching Kenfig Sands and Margam Sands undergo some refraction as they pass over the Kenfig Patches and North Kenfig Patches, resulting in local wave focusing and variations in alongshore breaker height. A high proportion of wave crests approach almost normal to the shore but there is low net longshore drift towards the north. During storms, offshore significant wave heights can occasionally exceed 5 m but the probability of their coincidence with still water levels exceeding 4 m is only about 1 in 20 years (Sutherland & Wolf, 2002).

At the present time beach levels are low due to a shortage of sand supply. In the southern and central parts of the area the beach consists of a steeply sloping upper section, composed of gravel and cobbles, and a low-angle lower section composed of fine to medium sand (see photographs in Appendix 1 and beach profiles in Figures 13 to 15). Much of the sand area exposed at low tide remains wet, limiting the rate of aeolian transport towards the dune system. A number of rock ‘scars’ and outcrops of weakly consolidated muddy sand, locally with a poorly developed peat developed at the surface, are often exposed on parts of the foreshore along the southern part of Kenfig sands. Towards the northern end of the system a gravel storm beach is also present but is fronted by a higher sandy beach which provides a source of windblown sand. There is a clear association along the length of Kenfig Sands between beach width, beach elevation, dune height and sand volume of the frontal dunes (Saye et al., 2005; Pye et al., 2007).

Steers (1946 p 177) reported that most of the dune system between the southern end of Kenfig Sands and Kenfig Pool was partially vegetated with Ammophila arenaria, although Agropyron junceum was rare and Elymus arenaria entirely absent. Salix repens covered the alluvial plain immediately seaward of the Pool. Steers also reported that a typical drift line vegetation assemblage was present along the beach - dune interface, with Cakile maritima, Arenaria peploides and Salsola kali. Today, strandline and embryo dune communities are confined to small areas near the mouth of the River Kenfig. There is very little input of new windblown sand from the beach to the remainder of the frontal dune system. Owing to lack of windblown sand accumulation, areas of yellow dune with marram (Ammophila arenaria) are scarce, occurring mainly in the northern part of the dune system (Hurford, 2006).There are a small number blowouts close to the shore where there is greater exposure to wind action, but lobes of avalanching sand on the leeward side are very poorly developed and generally restricted to localised areas with high trampling pressure.

Analysis of historical maps has shown that only limited changes in the position of the dune toe, mean high water mark and mean low water mark have taken placed since 1876, although the overall trend has been one of slow long-term erosion (Pye & Saye, 2005; Figure 16). Sand eroded from the southern part of the dune frontage and has mainly moved alongshore in a northerly direction and has accumulated in frontal dunes along the northern-most third of the frontage and near the mouth of Kenfig River. The areas where new sand accumulation has occurred are largely vegetated by marram communities while the older, more stable dunes in the south are dominated by closed Festuca grassland, with extensive dry Salix, Calluna and Salix-dominated slacks further inland (Hurford, 2006).

Between 1933 and 1973 approximately 4.5 x 105 tonnes of sand was mined from the beach and the foredune areas (Carr & Blackley, 1977). Sand has also been dredged from nearshore and offshore areas over a long time period. These activities may have had a significant negative impact on the sediment budget of the beach, though direct evidence is lacking.

In 2000 approximately 10,000 cubic metres of sand dredged from the approaches to Neath Harbour was placed on the lower foreshore off Sker (Kenfig) Beach as part of an experimental nourishment operation (BP Chemicals Group Ltd, 2000). It is unclear what proportion of this material has found its way onto the upper beach and into the dunes, but no new foredunes or transgressive dunes have been created as a result.

In the mid 1990’s much of the dune frontage was cliffed, and in places fronted by a relatively narrow (2 to 5m) cobble storm beach. In the past 15 years the cobble beach has grown in size, providing greater protection against wave action during storm conditions. The source(s) of the cobbles and shingle are uncertain; they include rocky sub-tidal outcrops and submerged boulder deposits around Sker Point and the Kenfig Patches, and the rock armour structure of Port Talbot Tidal Harbour which was built between 1966 and 1970.

The Kenfig dunes have been grazed since medieval times, chiefly by sheep and rabbits, but with some cattle. Grazing pressures were historically heavy, but after World War II the practice of grazing declined and by the late 1990’s was limited to about 100 cross bred ewes which were free to roam across the entire dune system. In 1999-2000 the number of sheep was increased to 200-500 by 2000 but grazing intensity remained ineffective for purposes of ecological management (CCW, 1999, 2008).

Rabbits were introduced to Britain by the Normans and ‘farmed’ in warrens throughout the medieval period, resulting in locally heavy grazing pressure on the vegetation which at Kenfig, and elsewhere, encouraged active sand blowing. The rabbit population was drastically reduced as a result of the major myxamatosis outbreak in 1954 and has never fully recovered. In the late 1990’s the total number of rabbits in the area was estimated to be less than 500 (CCW, 1999). From the 1970’s onwards vegetation progressively spread and stabilised the dunefield. By the mid 1990’s bare sand represented only 2.5% % of the total area and there was virtually no active sand movement more than 100 m inland from the shoreline (Davies, 1995; Jones, 1995, 1996). The situation today remains much the same. Most of the dune area is now stabilised by a thick sward of Festuca rubra and species diverse communities are restricted mainly to a number of intra-dune areas which have been artificially mown since the early 1990’s. Invasion by bracken (Pteridium aquilinum) is a problem in some areas but spread of sea buckthorn (Hippophae rhamnoides) has been successfully controlled since the late 1980’s.

During the Second World War, Kenfig, in common with many other Welsh dune areas, was used for military training. This also helped to maintain a high proportion of bare sand in the area. Aerial photographs taken during the period 1946 - 48 show that much of the site was comparatively mobile, with extensive areas of bare sand and active dunes (Figure 17). Significant areas of bare sand remained until the late 1960’s, when efforts were made to stabilize the largest areas using brushwood fencing and marram planting. Stabilization works continued to be undertaken until the late 1980’s.

Approximately 480 ha of the Kenfig system was designated as a Site of Special Scientific Interest by the Nature Conservancy in 1954. Most of the land is owned by the Kenfig Corporation Trust, which was established following the abolition of the old borough of Kenfig in 1886. Following a High Court case in 1971, the Trust obtained a ruling that they, and not the Margam Estate, were rightful owners of the Kenfig common lands which included the sand hills (Griffiths, 2002). Shortly afterwards, in 1977, a Local Nature Reserve was established under the management of Mid Glamorgan County Council, and later transferred to Bridgend County Borough Council. The Kenfig Dunes and Pool area was designated as a National Nature Reserve in 1989. The SSSI, NNR, and surrounding areas at Kenfig and neighbouring dunefield at Merthyr Mawr were declared as a candidate SAC in March 1995. Amongst the species of designated importance are Annex II fen orchid and petalwort which are now restricted in their occurrence to a few wet-slack marginal areas (Jones & Etherington, 1992; Jones et al., 1995. The extent of the SSSI, NNR and SAC is shown in Figures 3b,c & d).

During the late 1960s a three mile long haul road was constructed along the frontal dunes in order to transport stone from the Cornelly quarry for use in construction of the Port Talbot Outer Harbour. The track remains today, but is narrower than formerly, and appears to have little effect on the mobility of sand or its transfer from the most seaward dunes to the more landward areas.

The site experiences heavy recreational pressure but visitors currently have had relatively little impact on the overall level of dune activity. Trampling by humans and horses has created pathways and small areas of bare sand which are generally seen as beneficial from a biodiversity perspective. The unauthorized use of motorbikes and quad bikes has created small areas of bare sand in a number of places. Two small blowouts near the south-eastern margin of the dunefield, close to the M4, owe their continuing existence to the activities of children. Another blowout on the Sker Farm property, south of the NNR, has been active since the 1930’s, possibly assisted by trampling of cattle which still graze the area (Houston & Dargie, 2010).

In the absence of vegetation disturbance, natural deflation and dune migration, since the 1980’s many slacks have been invaded by Salix repens, sea rush (Juncus maritima) and scrub, resulting in loss of rare orchids (such as Liparis loeselii) and petalwort (Petalphyllus ralfsi). Since 2001 a number of attempts have been made to increase the area of bare sand subject to seasonal waterlogging, involving excavation of sand using earth-moving equipment. The spoil has been placed as mounds in the surrounding area but no attempts have so far been made to initiate mobile dunes. The most recent work was undertaken in late December 2010. Small blowouts in the frontal dunes near the haul road have also been artificially enlarged in an attempt to create active dunes, but with limited success.

In 2006 CCW erected a 3.7 km long fence across the central part of the NNR, enclosing 1.9 km2 of dunes on the northern side of the fence (35% of the total area) where cattle grazing was introduced. Fifty cattle and calves were initially allowed to graze the northern area between May and October. In 2009 about 20 Highland cattle were also introduced and allowed to remain throughout the winter. The land on the south side of the fence remains grazed only by c. 300 sheep from the end of October until beginning of March. To date, the additional grazing pressure appears to have little impact on the area of bare, mobile sand.

The water table within the Kenfig dunefield is relatively high and significant areas of the dune slacks are flooded during the winter months. The slacks have traditionally supported a rich wet slack flora. In 1985 a network of 130 plastic-line dip-wells was installed on an approximate 200 m grid across the dunefield and water levels were subsequently monitored at fortnightly intervals over a number of years (Jones & Etherington 1989; Jones, 1995). The data showed that the water table within the dunefield is broadly dome-shaped, attaining the highest elevation of c. 11 m OD just to the northwest of Kenfig Pool and sloping seaward both to the northwest and southwest (Figure 18)). The steepest gradient was located approximately two thirds along the seaward boundary of the dunefield, suggesting preferential groundwater flow in that direction. The water table is believed to be mainly rain-fed but there may be some contribution from regional aquifers. Mean monthly rainfall in coastal areas of South wales is approximately 120 mm between October and January, falling to 50 to 60 mm between April and July (Figure 19). Mean monthly temperatures (and evaporation) reach a minimum of about 8°C in January and February and a maximum of 19 - 21°C in July and August (Figure 19).

Water levels in a sub-set of 20 dip-wells are still monitored, mainly on a monthly basis (Low, 2011; Figure 20). Two sections across the central part of the dunefield showing the maximum and minimum water levels recorded since 2001 are illustrated in Figures 21 & 22. The general surface of the water table slopes seawards, with a maximum vertical range of about 1.5 m (Figures 23 & 24). During wet winters some slacks contain standing water for several months, while in other areas the surface sand is kept moist by capillary rise. Only in the driest summers does the water level fall sufficiently to allow the surface sand to dry out. As shown experimentally by Hotta et al. (1985), moisture contents as low as 1% can significantly raise the threshold for aeolian sand transport.

Saye (2003) collected a suite of 15 sand samples from the frontal dunes at Kenfig (locations shown in Figure 25) and analysed them using laser diffraction to determine the particle size distribution. The results are summarised in Table 1. The most seaward dunes in the south of the area were found to be relatively coarse (median size of 537 to 711 um) and moderately well sorted, but samples from the most seaward dunes further north were found to consist mainly of moderately well sorted medium sands (median size 292 to 473 um). Samples taken further back in the frontal dunes were also coarse at the two southernmost sampling transects and medium to fine further north (Figure 26). The fluid threshold wind velocity required to mobilize sand in the southern area would be of the order of 40 cm/s whereas that required in the north of the area would be much lower, of the order of 20 to 25 cm/s (Pye & Tsoar, 2009, p106).

As part of the present study, twelve sediment samples were collected from the inland dunes and two samples from the beach (locations shown on Figure 25). The samples were dry-sieved at ‘quarter phi’ intervals using a mechanical shaker and the raw data processed using the GRADISTAT software package (Blott & Pye, 2001). The results are also summarised in Table 1. The great majority of the dune samples can be classified as unimodal, very well sorted fine sands. Small quantities of coarse sand and very fine gravel-sized fragments of shell material are present in some samples. One sample of weakly consolidated muddy sand taken from the southern - central part of Kenfig beach also contained 6.3% mud. In a dry, bare condition, the threshold fluid wind velocity required to initiate aeolian transport of the fine inland dune sands would be of the order of 20 cm/s.

Analysis of the samples taken from the frontal dunes by Saye (2003) showed that the sands have an average calcium oxide content of 1.93 to 2.93% (mean = 2.75%; Table 2), which is approximately equivalent to 3.8% calcium carbonate. Most of this represents broken shell although some detrital limestone grains are also present.

The Kenfig Burrows dune sediments have a high silica content (88 to 95%, Table 2) and the principal mineral present is quartz, with subsidiary feldspar, calcite and heavy minerals (Saye, 2003). They are compositionally similar to other dune systems in Swansea Bay, reflecting a common marine sediment pool source within the Bay which is derived primarily from Pleistocene glacial deposits (Culver, 1977; Carr & Blackley, 1980). Swansea Bay is essentially a closed system which receives very little sediment from adjacent areas alongshore or offshore.

The potential problem posed by the loss of mobile dune systems across Wales as a whole has been noted for some time (Rhind & Jones, 2009). If current projections of sea level rise prove to be correct, coastal erosion can be expected to lead to significant loss of dune habitat, and especially ‘embryo dune’ and frontal ‘yellow dune’, at sediment starved locations such as Kenfig (Saye & Pye, 2005). Unless the frontal dunes are able to roll back as shore erosion proceeds, the sand stored in the frontal dunes areas will be lost to the sea as erosion proceeds. Consequently, there is interest in identifying potential methods whereby inland migration of frontal dunes can be encouraged, and where landward limit of present dunefields can be allowed to move inland, thereby maintaining or increasing the sand dune area (Rhind & Jones, 2009).

A number of options for dune reactivation at selected Welsh dune sites were identified by Houston & Dargie (2010) and subsequently discussed with stakeholders and conservation managers. The areas identified as potential reactivation sites within the Kenfig NNR are shown in Figure 27. One proposal was to remobilise a 950 m length of frontal dunes by stripping of the vegetation. Stakeholder feedback was that it probably would not be appropriate to try to develop a project at this location and on this scale, although one or two smaller breaches in the frontal dune ridges might provide routes for sand movement and might stand a better chance of success in the north of the area near the river mouth where more sand is available and the dunes have always been more mobile. Proposals to remobilise a number of individual parabolic dunes inland from the shore towards the north of the site were also considered. Stakeholder feedback was that this might be worth trying but a cautious, staged approach would be most appropriate.

The turf stripping and wet slack creation projects which have been undertaken at so far at Kenfig have mainly been biologically driven, particularly in terms of creating replacement wet dune slack habitat suitable for colonization by fen orchids and petalwort. These attempts have had varying success, with bare sand areas in some of the earlier ‘scrapes’ being almost completely re-vegetated within a very few years. The present study has therefore focused on the geomorphological factors which might influence the selection of the most suitable sites for future intervention, with a view to creating self-sustainable slack environments in association with new mobile dunes.

There are two essential requirements for active sand dunes: (a) a significant supply of sand and (b) sufficient wind energy to transport the sand and concentrate it into dunes (Pye, 1983, Pye & Tsoar, 2009). Other environmental factors, such as precipitation, temperature and humidity can have a significant influence on rates of windblown sand transport, but cannot prevent it under conditions of high wind energy. Vegetation is not necessary to build dunes, although its presence can have a major impact on rates of windblown sand transport and the form of the dunes which develop.

The historical evidence suggests that the episodes of major sand blowing and dune migration which characterised the period 1250 to 1800, leading to burial and abandonment of the original town of Kenfig, were associated with a high rate of sand supply and strong winds (high frequency and magnitude of storm events). This period was one of relatively low temperatures in the northeast Atlantic and northwest European area, and has often been referred to as ‘The Litte Ice Age’. Climate modelling results suggest that, during this period, marked thermal and pressure gradients across the northeast Atlantic resulted in a high intensity of cyclonic activity and strong winds both at altitude (the jet stream) and at the surface. The effects of a high frequency of relatively severe storms are likely to have included high rates of coastal erosion in exposed areas, a lowering of the average wave base, and landward reworking of sediments from offshore areas towards the coast. Although storm waves usually lead to beach erosion and sediment movement towards the nearshore zone, this material is often stored in nearshore bars and is available to be moved back onshore, together with new sediment brought onshore from deeper water, when fair-weather wave conditions return.

There is now strong evidence from many dunefields in western Europe that stormy periods during the Little Ice Age were associated with major phases of transgressive dune activity, while warmer periods with fewer severe storms were generally associated with episodes of widespread dune stabilization (e.g. Clemmensen et al., 2000). Even within the present century, there is evidence of high levels of active sand blowing at many British coastal dune locations during the relatively cool period between the 1950 and 1970’s, with widespread dune stabilization during the warmer period which began in the 1980’s. Several factors have undoubtedly contributed to the stabilization trend during the past 30 years, including reduction in grazing pressures reduced disturbance from vehicular traffic, increased inputs of nutrients (especially nitrogen) which stimulate vegetation growth, and more effective dune management following the designation of many dune areas as sites of nature conservation importance. However, the role of natural climatic variation cannot be ignored.

At the present time, Kenfig is a sediment starved system. Input of new sediment from rivers and erosion of coastal rock / sediment exposures is relatively low, and most of the readily available sediment on the floor of Swansea Bay has already been reworked landwards to accumulate in the dunes and estuaries of the area. Under present climatic conditions, there are relatively few south-westerly and westerly gales when waves can move sediment from deeper waters in the Bristol Channel towards the shore, and when winds are strong enough to create large blowouts in the well-vegetated dunes. In time this situation may change back towards conditions which existed in the mid- 20th century, although an early return to more extreme ‘Little Ice Age’ conditions seems unlikely in view of future global climate change projections.

Given that both new sediment supply and wind energy in eastern Swansea Bay are relatively low, the prospects for large-scale dune reactivation are relatively poor. The greatest chances of success in any dune reactivation schemes lie in areas where the largest quantities of sand are available and the wind velocities are highest; i.e. in the dunes closest to the shore in the north-western part of the Kenfig dunefield.

Based on an analysis of LiDAR data, aerial photographs, groundwater level data and preliminary field survey, four possible areas for experimental reactivation schemes have been identified; the locations are shown on Figure 27 and enlargements of each area in Figures 28-31.

Area A represents an area of maximum natural sand accumulation under present coastal process conditions. There are also a number of small, partially vegetated blowouts in the area which could be enlarged (Figure 28). Vegetation stripping and minor bull-dozing in this area could potential create one or more large blowouts with an active sand ‘nose’, although there would be loss of marram-dominated ‘yellow dune’ habitat. Significant dune migration within Area A could also impact on the Kenfig River, although unless the quantities windblown sand were very large blocking of the river mouth would be very unlikely and sand entering the river would be carried back onto the southern end of Margam Sands.

Two of the proposed areas (B and C, Figures 29 & 30) include areas of high frontal dune ridge with ample areas of relatively low-lying ground behind over which migrating dunes could pass without adverse impact on infrastructure outside the Reserve. There is already a significant blowout in the seaward dune ridge within Area C which could be enlarged artificially to create a continuous wind ‘corridor’ linking up with the large parabolic dune located about 700 m inland. Although there are significant reserves of sand stored within the frontal dunes in this area, the rate of sand movement, and the potential for transgressive dune growth and migration, could be enhanced by sand nourishment of the upper beach in this area, possibly using dredgings from Neath Harbour or Port Talbot Harbour.

The fourth area (Area D) is an area where active sand blowing took place on a significant scale until the late 1980’s, and where patches of bare sand still remain near the windward crest of the main parabolic dune. This area has a relatively long (c. 1 km) unobstructed wind ‘corridor’ to the west and there are significant quantities of dry sand (i.e. above the winter time water table) stored in mounds on the margins of the deflation corridor and in the frontal dunes at its western end. Reactivation of this large dune would require breaching of the frontal dune ridge and artificial building up of he arms to focus the wind flow towards the dune ‘nose’. Again, the rate of sand movement and dune migration could be enhanced by artificially adding sediment to the upper beach and seaward end of the deflation corridor.

Consideration has also been given in this study to the possibility of including part or the whole of the area known as ‘The Desert’ as a potential dune reactivation area. This was identified as a candidate site by Houston and Dargie (2010) and was active until largescale dune stabilization works were undertaken in the late 1960’s. It has the advantage that it has already been disturbed by human activities, is not seen as being of priority conservation importance within the Reserve, and has a large, unobstructed fetch to the west / west-southwest. However, it is considered probably too far (> 2 km) from the shore to allow the establishment of a continuous sand transport pathway from the beach, and there are already a number of slack recreation areas within the deflation corridor which could be lost. The main body of the dune at the eastern end of the deflation corridor also lies relatively close to the boundary of the Reserve and the reestablishment of sandblow on a significant scale could lead to invasion of neighbouring land and the Kenfig River channel. For these reasons, this area has not been included on the list of possible sites for initial trials.

The success of any reactivation trial is likely to depend partly on the scale of the work undertaken and also on whether significant quantities of additional sediment could be provided to the system in the form of beach and/ or dune nourishment. Much will also depend on weather conditions during and following the initial works. Reactivation would be favoured by periods of dry, windy weather which include periods when the water table falls to the point where surface sand is not kept wet by capillary rise. Artificial lowering of the water table would assist the reactivation process, but this would probably be ruled out on grounds that it would have an adverse impact on the wet slack vegetation communities. If present / recent conditions weather patterns persist into the near future, it is likely that some on-going management would be required to prevent re-stabilization of the deflation corridors by algae and higher vegetation. This could take the form of surface ploughing / tilling on an annual or seasonal basis.

It is anticipated than any initial works would seek to create an area of bare sand equivalent to approximately 5% of the area of the SSSI, with the intention that this would be increased to about 10% through the operation of natural processes over a 10 year time period. CCW stated objective is to achieve a minimum of 10% of mobile sand in order to assist achievement of favourable conservation status.