OR/15/017 Field survey - May 2014
|Tappin, D R, Long, D, Carter, G D O. 2015. Shetland Islands Field Trip May 2014 - Summary of Results. British Geological Survey Internal Report, OR/15/017.|
In May 2014 we returned to the Shetlands to further investigate the sediments previously identified on Mid Yell and also to carry out reconnaissance surveys on the islands of Fetlar, Unst and Mainland. We planned to acquire further cores and samples of the possible tsunami sediments and, of particular importance, we were to investigate and determine the stratigraphic relationships between the three event beds previously identified in 2013 at Mid Yell, with a primary objective to determine whether these beds could be found in one vertical succession at one location. We also planned to survey between Whale Firth and Mid Yell Voe to determine whether this area, a narrow isthmus, was inundated by the Storegga tsunami.
On Mainland we revisited locations where Storegga tsunami deposits had been previously identified; mainly around Sullom Voe and southward to Dury and Garth voes (Fig. 10). We also carried out reconnaissance in the north and central areas to determine whether there were any other exposures of tsunami sand in coastal peats not previously found. We visited locations where lake cores had sampled sands dated at 8,200 cal yr BP, coeval with the Storegga event and where younger sediments dated at ~5000 cal yr BP had been found (Bondevik et al., 2005).
Around Sullom Voe (Fig. 10) the tsunami sediments are preserved in coastal peat sections (Bondevik et al., 2005). This was the area where the Storegga tsunami sediment was first identified in 1988 (Fig.11). On the east side of the Voe at Scatsta (Fig. 12), a 6–10 cm thick, massive, medium-coarse muddy sand is exposed in a small channel cut into the peat (Figs. 13–16). The channel lies just above sea level and below the information board (Figures 13 and 14). The enclosing peat is structureless and immediately below the sand there is birch branch in the peat (Figure 14). Similar sediments were once visible at Garths Voe to the east of Scatsta (Birnie, 1992) but now have been covered/destroyed by new road construction.
The section at Garths Voe is described in detail in Burnie (1992) and dating of Salix wood in peat underlying the sand (Fig.11) show this to be 8980–8523 cal yr BP (Smith, 1993). Above the sand Betulawood is dated at 5987–5745 cal yr BP. The sample material dated seems to represent different botanical communities (clearly seen on Fig. 11), with the older organic layer beneath dominated by Cyperaceae (Sedges) and Salix (Willow) and above the sand Betula (Birch) becoming dominant and Salix almost absent. The birch was initially associated with ferns, but then heaths appeared at the site, and at some time after 5130 cal yr BP shrubs or trees disappeared and heath- and sedge-dominated blanket peat communities predominated.
Subsequent dating of peat above and below the sand at Scatsta and Garth’s Voe by Smith (1993b) further confused the issue as these returned dates of 6715–6448 cal yr BP and 6635–6548 cal yr BP for the lower contact, and 6266–5946 cal yr BP and 4406–4087 cal yr BP for the upper.
Further dating was carried out by Bondevik et al. (2005) from samples at the Houb on the west side of Sullom Voe. The sand layer is interpreted to be deposited by the same tsunami as described from Garths and Maggie Kettle’s Loch (Bondevik et al., 2003). Just above the sand at the Houb there is a horizon in the peat containing remains of birch trees; from here a root in growth position was 14C dated to 7720–7960 cal yr BP (7025±60 14Cyr BP). Seeds from the peat immediately below the sand layer were dated to 7820–8030 cal yr BP (7120±60 14C yr BP). Based on these dates Bondevik et al., (2005) concluded that the sand layer was deposited by the Storegga tsunami.
We visited the northern point on Mainland at Isbister (Figs. 17 and 18) where the beach at Sand Voe, a north facing inlet, is backed by a thick till, but with no evidence of sand. This area is mainly grassland that suggests a minimal peat cover; till and bedrock form low coastal cliffs. Behind the beach is the Loch of Flugarth (Figure 17) which has the potential to preserve tsunami sediments, but lake coring would be necessary to prove this.
Examination of various inlets on the eastern side of Mainland (Fig. 19) revealed no evidence of tsunami deposits and an absence of coastal peat, with till and bedrock exposed along the shoreline. Sites visited included Colla (Fig. 20), Firth, Vidlin (Fig. 21), Swining, Dury and Garth Lochs.
Both at Vidlin Voe and Colla Firth, the valley sides are sculptured, there is little peat and the pastureland sweeps down to the sea. We didn’t visit Dury Voe where Bondevik et al. (2005) found a sand layer in numerous peat outcrops along the shore and some 400m inland at the Ayre of Dury. The sand layer is 1–5 cm thick, with a sharp lower boundary. It is fine to medium-grained and rests on ca 1m of peat. The sand fines inland. It was dated at between 1540–1820 cal yr BP (1745±60 14C yr BP) and 1290–1420 cal yr BP (1460±50 14C yr BP).
Loch of Benston (Fig. 22) and Garth Loch (Fig.23) were cored and sampled by Bondevik et al (2005) who identified two sand units (Fig. 24). At Garth Loch the lower unit was dated at 7970–8210 cal yr BP (7320±70 14C yr BP) and attributed to the Storegga landslide. The upper sand they dated to 5280–5490 cal yr BP (4645±65 14C yr BP) and attributed to a younger event, not defined. At Loch of Benston the Saksunarvatn Ash was found below the lower sand (Fig. 25). The Ash has previously been described in lake sequences on Shetland (Bennett et al., 1992). It is dated elsewhere to ca 9000 14C yr BP (Birks et al., 1996) and 10,18±760 ice core years (Grönvold et al., 1995).
On the return journey home we visited Eshaness Headland, NW Mainland, site of the famous cliff top storm deposited boulders (Hall et al., 2006) (Figs. 25, 26 and 27). Here boulders up to >1m3 have been lifted up from the cliff face and thrown onto on the cliff top ~22m above sea-level (Hall et al., 2008).
On Yell the previous visit in 2013 identified a number of potential locations where tsunami sands are preserved or where they are suspected (Fig. 28). The schist and gneiss bedrock of the islands has resulted in excellent acidic conditions that have allowed a thick peat cover to form over the entire island. The regional geological guide (Flinn, 1994) suggests that previously 97% of Yell was peat covered and that currently 63% of the island has more than half a metre thickness of uncut peat. Flinn notes “layers of sand and /or gravel, generally up to 10cm thick, are commonly visible in the peat along the banks of streams high up in the hills. Such occurrences extend down to the shores without change of character, and all appear to be the result of streams overflowing their banks as a result of heavy falls of rain.” Such an interpretation is in conflict with the presence of marine diatoms within the sand layers (Sue Dawson, pers comm), which on the coast may well be the result of tsunami inundation.
Mid Yell Voe
At Mid Yell (Fig. 30) we returned to the sites examined in June 2013. 14C age dates of peat and twig samples at this location (Mid Yell 7a, Figs. 4 and 5) range from 8120±40 to 5800±40 cal yr BP. On the north side of the Voe, exposed intermittently along a ~50 m peat bank up to 2–3m high we found only one continuous 5–10 cm thick sand bed within the peat (Figs. 32 and 33) which we traced along section; the sand is laminated, poorly sorted, micaceous and organic rich with clay or peat clasts. Both upper and lower contacts are diffuse, but show no significant erosional relationships with the peat. The bed varies in elevation from 0.5 m above the beach (in the east) to approximately one metre below (centre), indicating that the sand is draped over a pre-existing, uneven, topography. Birch wood is buried within the peat both above and below the sand layer.
We dug two pits along the section (Figs 33, 34) to review the results from the fieldwork in 2013. East of the headland (See Figs 32, 33 and 34) at Pit east the undulating sand bed in the peat described above is at beach level. The bed is ~5–7 cm thick, muddy, fine-coarse grained, with diffuse upper and lower boundaries.
Within the section to the east of Pit east we also found a sand injection (Neptunian) dyke or downward penetrating sand (Fig. 35).
In the centre of the section, beneath the cliff (Headland of 2013 in Fig. 34) where there are stones on top forming a small headland (identified as the same location sampled in 2013 – Mid Yell ), clearing the peat revealed contorted sands at beach level (Fig. 36). A gouge core taken at the back of the pit penetrated the contorted sand, a sand layer at 0.39–0.44m and, at the base, a lower sand layer ~0.25 m thick, silty, but predominantly medium- to coarse-grained, micaceous with lithics up to 3cm diameter (Fig. 37). The basal sediment becomes clayey and may be till. In the peat above the exposed (upper), contorted, sand layer there are birch branches and logs up to 12cm in diameter (Fig. 36). A gouge core above this location on the top of the peat cliff sampled thin sand within the peat, just above the base.
At the location of the western pit (Fig. 34) there was no sand in the peat at beach level, but at the base of the pit we found sand at ~ 1m depth that we interpret as a continuation of the sand bed traced to the east. The sand is up to 10 cm thick, the base is sharp and undulating with faint internal lamination; it is very micaceous with organic clay or peat clasts. Above the sand there is a peat bed (~50 cm) with sparse twigs and then a peat layer with contains numerous twigs and logs (10–15cm diameter). A gouge core taken at the base of the pit sampled a sand layer 60–90cm extending below the base of the pit. Lack of core penetration suggests this sand immediately overlies till or bedrock. Above beach level the peat has a different texture (?Sphagnum moss blanket peat) and extends to top of the cliff which has a hummocky topography and is, covered with modern grass.
To the east of the main Mid Yell section and the eastern pit, there was a site with laminated sand within the peat (Figs. 40 and 41). The sand was up to 10 cm thick with sharp contacts at both the base and top. The relationship of the sand with the sand bed in the main section to the west is uncertain because of cliff collapse between, but both thickness and internal structure suggest that the two beds are not connected.
Landward of east pit we gouge-cored a traverse inland to trace the extent and elevation of the sand within the peat (Figs. 40 and 42). The sand was traced in two cores to 10 m inland (Fig. 42). At site 1, the sand was found 1.14 m below ground level, with base peat at 1.64m. At Site 2 the sand was sampled at 0.74m and the base peat at 1.14m.
A reconnaissance along the south side of the Voe found opposite the main exposure on the north bank further sand beds within the low peat banks (Fig. 43).
South west Mid Yell
At the west of Mid Yell Voe (see Fig. 30) in the stream bed at the head of the Voe and in a peat cut on the south bank there is a thin (cm) sand laminae (Figs. 44 and 45) very similar to the sand a Basta Voe (see below).
Mid Yell – comments
Our re-evaluation of the sections at Mid Yell found evidence for only one laterally continuous sand bed within the peat section on the north side of the Voe that is most likely deposited from a tsunami. This sand is exposed within the peat cliff and can be traced 50m along section. The change in elevation of the sand we attribute to depositional draping over an uneven topography; a mode of formation that is strong evidence for a tsunami origin. The two other sands found at this location, one at the base of the peat section sampled in a core, and the second at the back of the beach, are probably not from tsunamis.
We consider the sands at the base of the peat section are more likely to be either weathered till, fluvial sands deposited before peat accumulation, lacustrine deposits (Fraser Milne pers. Comm.) or a combination of all three processes. Our arguments against a tsunami origin are that the basal sands are thicker than other tsunami sands found elsewhere on Shetland (except for the 1500BP event), there are a number of individual beds, and the sands do not lie within the peat, a significant characteristic of other tsunami sands elsewhere, such as those at Whale Firth and Sullom Voe. With regard to the upper sand, again this is not laterally extensive and also that it is contorted and with a similar structure to some other, laminated sands, considered not to be from tsunami, found at other locations such as at the Loch of Vatsetter. Comparison of our results with those from the 2013 survey is difficult because the two section logs illustrated in Fig. 46 and 47 (Mid Yell 7a and 7b) are quite different with two intra peat sands in 7a and only one in 7b.
The new age dating at Mid Yell is somewhat similar to the initial dating at Garth’s Voe (Birnie, 1992) with an older date (8120±40 cal yr BP) at the base that suggests a Storegga age, and a younger, mid-Holocene age (5800±40 cal yr BP) below an upper sand layer (Fig. 47). Age dating of peat is well known to be compromised by contamination, but the 2013 identification of two upper sands within the peat at Mid Yell 7a is problematic although the upper sand could be the contorted sand bed we describe above. If so then the intra peat sand at 0.60m depth would equate with Storegga. Perhaps comparison to the Sullom Voe does provide an explanation; here later dates of ~5000 BP, in the absence of positive identification of multiple and overlying sand beds as sampled in Loch cores, led to an interpretation that the younger dates were erroneous and due to contamination (Bondevik et al., 2005).
The laminated sands (Fig. 44) to the east of the main section at Mid Yell seem different (laminated) to the massive sands at the main section to the west around the headland, suggesting a different origin, but these sands appear similar to those at the base of west pit. There is further discussion of this below in the broader context of evidence from other locations, and a storm origin proposed only on the evidence from Mid Yell.
Whale Firth south (1)
The section at the southern end of Whale Firth was as described in 2013 (Figs. 48 to 51) and now dated at 4760±30 14C cal yr BP. West of the jetty there is a persistent sand bed up to 12 cm thick within the peat that can be traced 50 m along the coast. The sand bed has an undulating, but sharp, base and a fairly planar top; in places the sand is faintly laminated. The sediment is up to granule size, poorly sorted and organic rich and there are clasts of peat and clay. There was no evidence for the major rip up of the size clasts of the size seen at Sullom Voe, however, the enclosing peat is again very woody.
At WP 703, we sampled the sand bed with the Russian corer (Fig. 52).
Further to the north at WPs 669–671 there were further disrupted and laminated sand beds (Figs 48, 52–53) similar to those seen at the east end of Mid Yell Voe section. The sand was just above beach level and formed discontinuous lenses with a combined thickness if 10-12 cm. At one location there was a cobble of mica schist enclosed. The peat inland of these sites was very hummocky, possibly due to extensive peat flows.
Central Whale Firth (2)
We also visited a small bay (Grimister) half down the loch (WPs 622–624) Figs. 28, 29 and 55. There was no peat at shore level, and the pasture land descended to the sea terminating at low rock cliffs; the beach was bouldery, but the valley feeding into the Firth was regarded as a possible coring site (later sampled in August 2014).
Comments – Whale Firth
The significant thickness of the sand in the peat at the head on the Firth suggests a major tsunami event — ?Storegga, but there is an absence of rip up clasts as seen at Sullom Voe. There is only one 14C date that suggests this to be the younger ‘Garth’ tsunami but the thickness and the absence of an older sand (as in the lakes) suggest that the age may be due to contamination as at Sullom Voe. The discontinuous sands north of the main sand are at about the same level and may be laid down from a tsunami but also from other mechanisms such as storms or peat bursts (see discussion below).
Mid Yell core traverse
To investigate the possibility that the Storegga tsunami (or any other event), inundated the col between Mid Yell Voe and Whale Firth we cored between the two locations (Fig. 56).
No definite indications of tsunami sediments were encountered. The results from the sites are presented in Milne et al. (2014).
Basta Voe is an enclosed inlet on the east side of Yell (Fig. 57). Previous research has identified three sand layers in the low coastal peat cliffs (Fig. 58); but only the upper one has any great areal extent (Bondevik et al., 2005; Dawson et al., 2006). In a low coastal cliff, on its west side towards its northern termination we confirmed the presence of a thin (1–2cm) layer of sand within the upper section of peat (Figs. 59 and 60) as previously described. The sand is very fine-grained, very well sorted. The enclosing blanket peat shows several variations in composition, possibly reflecting environmental change. There is a dried peat layer about 25cm above the sand layer. The sand is mainly massive, but at one coastal site had a flame structure. The sand at the coast was traced landward in an extensive peat cutting extending orthogonally from the coastal peat cliff to the road (a distance of ~25m). The cutting reveals the sand layer (with possible climbing ripples — Fig 56) rising up slope at a fairly uniform depth below present surface of the peat and penetrating inland 70–80 m. The sand layer creates a line of weakness at which the peat blocks break as they dry out. Cut peat on the other side of the road also shows the same sand layer.
Walking up the west side of Dalsetter stream at the head of the Voe beyond the old bridge and up the Burn of Gossawater (Fig. 57), we traced a sand layer in the peat similar to that at the coast ~ 600m inland (limit is WP 615 on Fig. 57), thinning laterally up the valley side. The highest elevation is at +9 m OD (Dawson et al., 2006). The sand is coarsest in the centre of the valley.
We looked for sands on the west bank of the Voe farther south of the sands at the head (Figs 63 and 57, WPs 643 and 644) but found none. At WP 643 (Fig. 61) bedrock was overlain by grey till and one to two metres of peat. At WP 644, only peat was exposed (Fig. 62). The peat is massive and there is no evidence of tree material at the Basta Voe sites, perhaps indicating the peat to be much younger here than at other locations at Mid Yell.
Comments - Basta Voe
The upper sand at Basta Voe has been dated previously at ~1500 cal yr BP and its origin proposed as from a locally generated tsunami (Bondevik et al., 2005; Dawson et al., 2006). The restricted areal extent of the lower sands led Dawson et al. (2006) to propose a storm origin for these. They traced the upper sand around the head of the Voe and suggested a local source. These interpretations are not disputed.
Kirkabister 1 and Burra Ness (Kirkabister 2)
Kirkabister is located southeast of the sands at the head of Basta Voe, on the end of the head land (Fig. 63 – WPS 616–618). In the cliff section (WP 616. Fig. 63) southeast of the old farm buildings at the back of the beach, there is a sequence of till overlying interbedded, peat, sand, gravel and clay (Fig 64).
The upper section in the coastal cliff is mainly soft peat with no sand (Fig. 64) approximately two to three metres thick, that overlies a more consolidated section at the base of the cliff where muddy peat is interbedded with laterally variable polymict gravels and mainly thin beds of grey sandy clay shown in detail in Figs. 65 and 66.
The section in Figure 65 (Located in Fig. 64) shoes a lower varicoloured grey laminated sequence of sandy clay (~3 cm) not seen in Fig 66, overlain by ~10 cm of brown muddy peat, within which (near the top) is a 2–4 cm thick pale grey-brown, muddy sand with a ?climbing ripple feature similar to the sand at Basta Voe (see Fig. 60); this unit is traced quite clearly into Fig 66. Overlying the muddy peat (see Fig 66) is an up to 10 cm bed of pale grey brown sandy clay. From Fig 64 it appears that this bed varies in thickness. The grey sandy clays are similar to the Storegga sand/clay at Maryton, near Montrose on mainland Scotland (Figure 67).
Kirkabister 2 – Burra Ness
On the north side of Burra Ness (the headland east of Kirkabister – Fig 63; WPs 617–618) there are storm beach deposits (large stones) and back-beach (?aeolian) sands underlain by peat within which is a thin (2–3cm) sand layer – slightly muddy (Figs. 68 and 69). The peat surface on the beach may have been exposed in recent storms. These deposits were first discovered in 2013.
Comments on Kirkabister and Burra Ness
The section at Kirkabister is complex and requires more research to validate any tsunami origin of the sediments at the base of the coastal cliff. The lower section beneath the peat is quite unique; there could be two tsunami sediment deposits here, with the grey, sandy clays apparently very similar in appearance to the Storrega deposit on the east coast of Scotland at Maryton. There seemed to be considerable lateral variation in sediment thickness of the upper sandy clay. The absence of any birch wood compromises this interpretation however.
At Burra Ness (or Kirkabister 2) there was no strong evidence for tsunami deposits, except for the sand beneath beach level, however the similarity in appearance of the back-beach sands and the sands within the peat may suggest a common mode of formation.
Loch of Vatsetter
The Loch of Vatsetter (Fig. 70) is on the east side of central Yell, cut off from the sea by a tombolo. On the southern side of the Loch at the base of a low coastal peat cliff there are limited exposures (20 m long) and up to 30 cm thick, of thinly (cm) laminated, fine-grained sands interbedded with peat (Figs 71 and 72) (WP 648 in Fig. 70).
The laminated sands units appeared to be preserved as channel fills (Fig. 71). In places the laminae are deformed (Figs. 72 and 73).
The laminated sediments have a contorted base suggesting infilling of an irregular surface. The upper surface is cut and overlain by darker sand free peat (Fig. 72). The peat with the laminations is interpreted as a fallen or transported block.
The location is at the coastal margin of a series of channels (Fig. 70) which may explain the features observed. The peat at this location is formed of a number of layers (Figs. 74 and 75): at the top there is one metre of fibrous peat overlying 70 cm of peat with Betula branches on clayey peat with large Betula branches at beach level; the block with the laminae seems to have slipped over this sequence.
We augered landward of the edge of the peat cliff but found no evidence of sand.
We investigated Salt Wick on the outer west coast of central Yell (Figs. 70 and 76) for evidence of sand in the peat, but found none. On the storm beach at this location there are only a few patches of peat cliff extending to the shoreline; there is no evidence of old peat or sands, suggesting rapid erosion of the cliff.
Comments – Vatsetter Beach
The sands at Vatsetter Beach are preserved in displaced peat blocks, with the displacement possibly associated with the streams at this location. Their origin is uncertain, but their very localised extent and the absence of any tsunami sediments at the locality discounts a tsunami origin. They are thus of either of marine origin or deposited by stream action; the absence of any extent landward suggests the former; further examination of their mineralogy is required to confirm this.
Other locations on Yell
We carried out a reconnaissance of the coastline of Yell Island (mainly in the east, north and south) to investigate for other sites where tsunami sands might be preserved (Fig. 77).
Gutcher (HU549993) (Fig. 78). No sign of potential preservation at Gutcher even in coastal inlet (Loch of Gutcher) at the ferry terminal.
The coast along Bluemull Sound (Fig. 77) didn’t show promise. It was noted that this area had had improvements to the agriculture quality of the land such that blanket peat was rare and that worked fields extended to the coast.
Wick of Breakon (HP528051). Here (Fig. 79) there are extensive dunes that appear to sit directly onto bedrock with no evidence of peat. The sands here seem richer in carbonate than on other beaches.
Mouth of the Gloup Voe at Kirks of Gloup ((HP507050) (Fig. 80) there is an extensive storm beach of cobbles and larger clasts.
Wick of Whallerie. There is a thin till on top of the bedrock either side of the beach but no peats at the coast to preserve any potential tsunami deposit. There possibly was peat behind the storm beach but the ground rises up quickly and almost certainly the peat has been cut to produce pasture land.
Southern end of Gloup Voe (HP506030). Here there is some peat but much wind-blown sand (Fig 81). It was noted that the Gloup inlet has several ayres at the northern end and at low tide the water is very shallow (Fig. 82). The position of tracks on either side suggests that it is easily fordable at low tide. The inner part of the Voe appears much deeper. If a tsunami wave entered this inlet it is likely to have become very broken up at the entrance and any wave reaching the southern end to be quite confused.
North Ay Wick (approx. HU537867) (Fig. 83). Looked at coastal sections from car, appeared to be bedrock and glacial and not promising for suitable peat sections.
Otters Wick (HU525856, WP 651) (Location Fig. 77). Looked at coastal sections from car, appeared to be storm beach with bedrock and glacial deposits and not promising for suitable peat sections.
Wick of Gossbrough (approx. HU530833) (Location Fig. 77). Looked at coastal sections, where there is an extensive area of cliffs of laminated wind-blown sands showing cross cutting dunes. These extend along several hundred metres of the beach. Below the sands are bedrock and glacial deposits.
Burra Voe, near Hamnavoe (approx. HU494802 WPs 652) (Location Fig. 77). Checked peats at the top of the beach at the ayre. Here there are large blocks/rafts of peat up to 1.5m high. They show a sequence of peats including layers of Colluna (heather) within smoother sphagnum moss peat.
Checked site near fish farm at mouth of Aris Dale (river. approx. HU486808. WPs 653 and 654) (Location Fig. 77). Peat 1–2m high sits on top of sands and gravels (Fig 78). There are a few wisps of sand within the peat in the basal few centimetres of the sequence. At one location there is a curious horizon of large schistose boulders about 60cm above the base of the peat (Figs 841 and 85). Armouring of peat on Yell by beach boulders is report by Flinn (1994), who considered these to reflect storm events that allow preservation of peat during marine transgressions.
A day was spent on Fetlar (Fig. 86) conducting a reconnaissance survey for tsunami sediments. Fetlar is mainly composed of obducted oceanic crust, predominantly ophiolites, so soils are thin, peat is not well developed on the island. The land is mainly pasture.
On the southwest coast we visited the Wick of Triesta and Papil Water (WPs 686–688), here the coastal sand dunes directly overlie bedrock (Figs. 87 and 88).
On the southeast coast at Funzie Bay (Fig. 86) there is a storm beach composed of cobbles and boulders (Fig. 89).
We carried out a reconnaissance visit to Unst to survey potential coastal sites (see locations on Figs 91 and 92) where tsunami sands might be preserved. Previous visits to the island (Smith, 1993b) had not found sands in the coastal peats, but deposits in cores had been found and dated at Norwick, Burragarth and Snarravoe (Fig. 91). Like Fetlar, the south eastern part of Unst is oceanic crust and composed of Ophiolite and gabbro, so peat is not present here, but seems to be well developed elsewhere, where the tsunami sands were cored.
At Norwick the peat is 8 m thick and two sand units were sampled at about 4 m depth beneath ground and dated 6840±40 years 14C (Smith et al., 2004). At Site 2, the lower part of the Burn of Norwick valley landward of coastal dunes is widely occupied by peat (Smith, 1993; Smith et al., 1995). In the centre of the valley at its seaward end, the peat rests upon a thin (ca 10–20 cm) layer of grey, micaceous, sand with angular clasts and organic fragments, beneath which is a laminated, grey-green, silty clay. The top of the silty clay is relatively uniform, lying at -3.38m (-4.70 m) to -4.17m (-5.49 m), with the top ca 25 cm not laminated, and contains a scatter of angular clasts (the largest 4 cm in its greatest dimension). Above these basal deposits, the peat is woody at its base and above this contains silty horizons. The sand continues up-valley forming a tapering wedge in the peat. Throughout, the sand contains peat intraclasts as well as many organic fragments. In the sand, the pollen includes Chenopodiaceae and Plantago maritima,, which may indicate a marine origin, given that the sites probably lay a considerable distance from the shore at time of deposition. There are exceptionally high percentages of corroded pollen and spores in the sand and at the top of the underlying grey-green, silty clay. A 14C (AMS) date of 6840±40 (7748–7589) was obtained for the base of peat overlying the sand layer at borehole 140.
At Burragarth the northwestward-facing bay (Lunda Wick) is backed by dunes, landward of which a peat moss occupies the lower end of Vinstrick valley. At the base of the sequence lies a sand and gravel surface, overlain to seaward by a grey, silty clay which forms a gently sloping surface and thins out landward, reaching a maximum observed altitude of –3.65m (-4.88 m). Above the silty clay is a widespread layer of grey micaceous sand, up to 0.75 m thick, containing fragments of vegetation (roots, stems and twigs). Above the sand, peat with horizons of organic silt and sand extends to the surface. Towards the head of the valley, the sand layer attains an altitude of at least 4.11 m (2.88m) (Burragarth E., Table 4). The sand shows high values for corroded pollen and spores and at the top contains Chenopodiaceae (see Fig. 10, below). There are no diatoms in the fine sand, but the silty clay contains a few broken and eroded pennate forms, mainly Pinnularia viridis. Pinnularia occurs widely in Storegga tsunami deposits in eastern Scotland, incorporated into the tsunami sand as the tsunami moved across underlying freshwater sediments. The presence of Pinnularia fragments in the silty clay at this site could indicate re-working of the upper layers of that deposit as freshwater (possibly peaty) sediments either down valley or towards the valley sides were reworked. A radiocarbon date of 7215±60 (8165–7878) was obtained for the base of the peat at 4 m depth, overlying the fine sand in borehole 108 (Smith, 1993).
Coring at Snarravoe sampled sands at about one metre depth dated at 7785–7974 cal yr BP, and correlated with the Storegga event (Bondevik et al., 2005).
Nor Wick Beach (WPs 691–692) (Fig. 93). Very boulder, storm beach, no peat, with pastureland over rock cliffs that descend to the sea.
Burra Firth (WP 693) (Fig. 94). Mainly rock cliffs at the coast with a sandy beach at the head of the loch.
Quoys (WP 694) (Fig. 95). Above Loch of Cliff, mainly pasture land, no peat.
Houland (WP 696) (Fig. 96). Pasture land, no peat.
Ham of Muness (WP 697) (Fig. 97). Pastureland, gravel/cobble beach, thin peat.
Sand Wick (WP 698) (Fig. 98). Pasture overlying till on bedrock, no peat.
Lunda Wick Bay (WP 699-700) (Figs. 99, 100). Pasture on till overlying bedrock; Sand and rocky (storm) beach.
Comments – Unst
No peat was observed along the coast, and no evidence of tsunami sands, that contrasts with the peat sections and tsunami sands recovered in the coastal valleys at Nor Wick and Burragarth, and in the lake cores at Snarravoe.
All of the main coastal exposures on Yell visited in 2013 where possible tsunami sediments had been identified previously were visited and all sediments present, examined. Tsunami sediments were confirmed at Basta Voe, Mid Yell Voe and Whale Firth. At Kirkabister more work is required to validate the sediments origin and at Burra Ness the sediments are probably not from a tsunami. The one site at Sullom Voe, Scatsta where tsunami sediments are proven provided a context to the Yell sediments, allowing the identification of both similarities and differences in the sediments at the different locations (Sullom Voe and Yell). The observations made during the fieldwork reveal a number of significant aspects of the different sands observed and their relationships to the peat which may be further developed to establish improved discriminating criteria for identifying their origin and (possible) different ages.
Relationships between the tsunami sands and the peat
The field work confirmed preservation of tsunami sands in coastal peat at the Yell locations noted above. A significant aspect of the preservation of the sands (or lack of them) is that they were deposited, and are now preserved, in thick coastal peats. Preservation of the tsunami sediments is thus dependent on peat accumulation which is primarily controlled by the underlying rock type. Significant peat accumulations are found on till or siliceous rocks such as Old Red Sandstones and metamorphic schists and gneisses. It is these rocks which form the bedrock of Yell, much of northern Mainland and northeast (Nor Wick) and southwest (Barragarth) Unst. This explains in large part why the potential for finding tsunami sediments along coastal exposures at these locations is so high. The large areas of serpentine in Unst and Fetlar produce a very poor soil able to support few plants but not significant accumulation of hill peat; again explaining why tsunami sediments are not found there.
The fieldwork identified significant features of the peat/tsunami sand relationships. The first is that, at Whale Firth, Mid Yell and around Sullom Voe, where the tsunami sands are definitely identified, they lie towards the base of the peat sequence, resting on a significant underlying peat thickness (of at least 0.5m) usually with a greater peat thickness above. The second feature is that, at all these locations (Whale Firth, Mid Yell, and around Sullom Voe) the peat within which sand is preserved tends to be very ‘woody’, with a significant proportion of large birch roots and branches.
At Whale Firth, the single sand bed is preserved towards the base of the peat; it is up to 12 cm thick and very similar in lithology (except for an absence of large rip-up clasts), thickness and extent to the sands at Sullom Voe. The sand was traced inland to an elevation of 8-9 m above sea level (Milne et al. (2014); all evidence that supports a tsunami origin.
At Basta Voe the sediments were as described previously, with three sands present, but only the uppermost sand traced significantly far inland and up the stream slope, which supports a tsunami origin. The peat/tsunami sand relationship at Basta Voe is different to the locations at Whale Firth, Mid Yell and around Sullom Voe. The sand is preserved in the upper part of the peat section, the peat is a ‘blanket’ deposit, and not ‘woody’.
Age dating of the tsunami sands
There are no established ‘unique’ sedimentological criteria for identifying tsunami sediments, for example, that allow their discrimination from other high energy processes such as storms, although work is progressing on this (Costa et al., 2015; Costa et al., 2012). For the most part, recent tsunami deposits are identified from their immediately recorded source mechanisms, such earthquakes in the instances of the Indian Ocean, 2004 and Japan, 2011. Palaeotsunami deposits are also recognised from their relationship to an identified source mechanism, with the relationship usually established from a positive correlation between the age of the source mechanism and the age of the deposit; an excellent example is of course the correlation in ages of the coeval tsunami sand deposits in Scotland with the Storegga landslide off Norway.
Until recently most tsunami sands in Scotland have been identified or been assumed to result from the Storegga landslide off Norway; because no other submarine landslides, either off Norway, or off Scotland, have associated coeval sediments that indicate a tsunami was generated. This does not mean, however, that these landslides did not generate tsunamis, it is just the preservation potential of these sediments is very low. It is also now apparent that on Shetland age dating of the peat in which the tsunami sediments are preserved may be subject to contamination that compromises what should be straightforward correlations with putative sources.
Thus, when the most likely origin of the 1500 cal yr BP sediments at Basta Voe was identified as from a tsunami, it was somewhat contradictory that, conversely based on the foregoing, no immediate source was identified. Subsequent dates of 5–6000BP from Yell in 2013 were further confusing because, although there were submarine landslides of approximately this age off Norway (e.g. Traenadjupet), identifying these sources was compromised by the limited extent of the deposits (none being found anywhere else) that indicates the source to be local. Thus from Shetland, there was the possibility of two local tsunamis, but no source mechanism.
Based on the evidence from Basta Voe (with additional confirmation of an event of this age from farther south at Dury Voe on Mainland), both sedimentological and 14C dating, although there is no obvious source mechanism, it seems most probable the sediments here are from a locally generated tsunami; yet to be found.
The dating evidence from Mid-Yell and Whale Firth for a mid-Holocene tsunami requires further and careful consideration. There is strong evidence for a mid-Holocene (~5000 year) event from lake cores farther south on Mainland where two events are recorded within two stratigraphical sequences at two very close locations; sedimentological and age dating evidence is strong; as with Basta Voe however, there is no obvious source mechanism.
The evidence on the tsunami source from Yell at this stage, is inconclusive. The sand at Whale Firth is remarkably similar to the sands at Sullom Voe (except for the absence of large rip-up clasts); it lies within a woody peat, so a Storegga source seems to be most likely, except the one date here is mid-Holocene (4760±30 cal yr BP). There is no sequence, similar to those cored at Loch Benston and Garth Loch, which would provide a firmer context (two sand layers at one site with different ages). Thus the identification in 2013 of three sand layers within one stratigraphic sequence at Mid Yell promised to provide a firmer basis for a mid-Holocene event. It is thus unfortunate that the results from our 2014 visit do not confirm the presence of more than one tsunami sand at this location. The thick sand at the base of the peat, because of its’ thickness and stratigraphic position (at the base of the peat) where it overlies till or bedrock (not within it as seen elsewhere at Whale Forth and Sullom Voe), is probably not a tsunami deposit. We suggest it is weathered bedrock, mineralogenic sand laid down before the peat formed (as described by Flinn, 1994), a lacustrine sediment (Fraser Milne, Pers. Comm.) or a combination of all three processes. Only one tsunami sand therefore is confirmed at Mid Yell, its origin interpreted as such because it is preserved fully within the peat and because of its lateral extent along the section; its variable elevation, high in the east, but below ground level in the west supporting a tsunami origin.
Our interpretation of a single, laterally extensive mid-peat sand is hard to reconcile with the logs of the Mid Yell 7a and 7b logs reproduced in Dawson et al. (2013). The sequence we identify most resembles Mid Yell 7b with an upper sand preserved in woody peat and overlain by fibrous peat. The lithology of the sand we describe within the woody peat is different to the sand at Whale Firth (and hence Sullom Voe) in that it is thinner, muddier, the upper and lower peat contacts are not as sharp, and in places it is contorted. Not being able to relate our results to 14C dates of Mid Yell 7a compromises our interpretation of the age dating from this section. If we are correct that the basal sand in Mid Yell 7a is not from a tsunami, then the lower of the overlying two sands should be Storegga and the upper a mid-Holocene event based on the peat age of 8120±40 cal yr BP for the lowest peat and the 5800±40BP age from the peat immediately below the upper sand. From a similar location, Milne et al. (2014) identify in a gouge core two 5–6 cm intra-peat sands above the basal sand in Mid-Yell Log 7 (contorted sands location).
The other dates from Mid Yell of 6060±40 cal yr BP at Mid Yell 1 (by the road) and 780±30 cal yr BP from Mid Yell 3 (at the west end) do not really provide any further confirmatory evidence on the likelihood of additional tsunami events, except that at these locations there is only one sand within the peat, that perhaps confirms a single laterally extensive event.
Tsunami sands at other locations?
At Kirkabister, on the south side of the headland, a complex series of peat/clay/sand/gravel interbeds at the foot of a low coastal peat cliff are of uncertain origin. One clay bed is similar to the Storegga sediment at Maryton on the east coast of Scotland. The absence of a significant thickness of peat beneath the sediments together with no evidence of woody peat above or below the sand compromises the interpretation of their origin from a tsunami of 5000 or 8200 years BP. The clastics could be from a younger tsunami — ?1500BP, but their lithology is completely different.
No further convincing evidence of tsunami sands was found at any other location visited on Unst, Fetla, Yell or Mainland.
The ‘laminated’ sands
Other sand layers and lamina we discovered at Whale Firth, Mid Yell Voe, Vatsetter and Basta Voe are probably not tsunami deposits. The two lower sand laminae at Basta Voe are probably storm deposits because of their limited extent (Dawson et al., 2006). At Whale Firth, Mid Yell and Vatsetter the laminated (cm) sands in the peat are of uncertain origin, although several observations suggest they are not from a tsunami; at both Whale Forth and Yell the sands are at a similar elevation to the tsunami sands identified at these locations. Differences between the laminated and tsunami sands include; i) their laminations within the peat and ii) their limited lateral extent both along the coast and inland. At Vatsetter, tsunami sands are absent, but here the laminated sands are preserved in displaced blocks (seaward of hummocky peat morphology) at the mouth of a series of small streams. The laminated sands at Whale Firth were also associated with streams and an, inland, hummocky peat topography. There are no streams at Mid Yell, but the sands here are also preserved in displaced blocks. At all three locations the laminated sands were not preserved in woody peat, although the sands in the displaced blocks at Vatsetter were overlying woody peat with large birch logs that were well exposed in the coastal cliff just to the west. The ‘overthrust’ relationship of the blocks to the woody peat at Vatsetter demonstrates lateral movement towards the coast.
Alternative origins of the laminated sands (discounting tsunami) include deposition from streams or some form of peat block displacement, such as from bog bursts (see Appendix 2), both of which would account for their limited lateral coastal extents and local associations with streams and translated blocks; but a lack of landward extent of sand in the gouge coring (which might be expected) does not support either of these interpretations (see Appendix 2 for further discussion). A marine or wind-blown origin is also possible, but if this was the explanation, why are the sands not more common along the coast? Further research is required to determine the origin of these sands, but in the context of this report, the most important conclusion is that they were not deposited from a tsunami.
There is undoubtedly strong evidence for tsunami sands within the coastal peat exposures on Yell resulting from the May 2014 fieldwork, that confirms some, but not all, of the conclusions of field visit results from 2013. The evidence is supportive for tsunami sands at both Whale Forth and Mid Yell, but the age of these deposits and hence their source mechanism, requires further investigation. At Mid Yell there is good evidence for one laterally extensive deposit that drapes a previously existing undulating peat surface. The basal sand, on the basis of the peat stratigraphy is considered either to be weathered bedrock or monomineralic sands (lacustrine?? Pers. comm. Fraser Milne) laid down before the peat development.
The preliminary 14C dating suggests that in Shetland there might be three tsunami events, but only 8150 BP Storegga has a positively identified source, confirmed by coeval deposits on mainland Scotland. The absence of coeval deposits on mainland Scotland for the younger events compromises somewhat the tsunami interpretation for the two younger events as it is more usual that there is a tsunami source mechanism, but no corroborating sedimentary evidence, rather than vice versa with sedimentary evidence but no source.
A mid-Holocene tsunami is indicated by the lake core data from Mainland Shetland and preliminary age dating available from the sediments on Mid Yell supports the presence here of this event., However the deposits on Mid Yell are very similar to those in coast sections around Sullom Voe which are interpreted as sourced from Storegga; these deposits (as Sullom Voe) were also initially dated as mid-Holocene in age, with dates similar to those recovered from Mid Yell and Whale Firth, so does contamination also explain the younger dates at these locations? Further work is required.
The sands and dating at Basta and Dury Voes support another (younger) tsunami event; unfortunately only the first location was visited. The young age dating and the blanket peat in which the sand is preserved that is not ‘woody’ compared to coastal sections at Whale Firth, Mid Yell (and at Sullom Voe) suggest that the deposit seems most likely from a tsunami much younger then Storegga, and with a local source, as yet unidentified.
An important result of the 2014 fieldwork has been the identification of a possible relationship between the tsunami sands in the coastal peat exposures and the peat stratigraphy. Further examination of this relationship may provide a context to the age dating that will help in differentiating between the (mid and early Holocene) 14C ages. The age period over which the tsunami sands were deposited experienced major climatic changes that are reflected in the vegetation (see Burnie, 1988), thus the peat palynology may provide an answer. The tsunami deposits at Whale Firth and Mid Yell were deposited in woody peat with a significant proportion of Birch. The youngest deposit at Basta Voe was deposited in blanket peat with no wood; the work by Birnie at Dury Loch suggests that this change is reflected in the palynology. The striking similarity between the Mid Yell/ Whale Firth deposits and those at Sullom Voe, both in sedimentology and stratigraphy suggests that they are the result of the same event. In isolation; 14C dating may not solve this problem, but further research on peat stratigraphy just might.
With regard to the deposits at Kirkabister, their position at the base of the coastal peat sequence is at variance with the intra-peat tsunami deposits found elsewhere on Yell and on Mainland and suggests that may not be from the same tsunami source mechanism(s) as the sediments elsewhere on Shetland; further research is required.
The origin of the laminated sands found at Whale Firth, Mid Yell and Vatsetter is still uncertain, but they are probably not deposited from a tsunami, which is our primary interest here. Other mechanisms for emplacement include peat slides, flash floods, aeolian.
Postscript; Immediately after this report was finalised, age dating of peat sections at Whale Firth and Mid Yell Voe confirmed that the sands preserved in the woody peat here are of Storegga age, ~8200 cal yr BP.
- BONDEVIK, S, MANGERUD, J, DAWSON, S, DAWSON, A R, and LOHNE, Ø. 2005. Evidence for three North Sea tsunamis at the Shetland Islands between 8000 and 1500 years ago. Quaternary Science Reviews, Vol. 24, 1757–1775.
- BIRNIE, J (editor). 1992. Garths Voe. Geological Conservation Review No. 6.
- SMITH, D E. 1993a. Norwick, Unst; Burragarth, Unst; Sullom Voe, Mainland.
- SMITH, D E. 1993b. Norwick, Unst; Burragarth, Unst; Sullom Voe, Mainland. 52–56 in The Quaternary of Shetland. BIRNIE, J, GORDON, J, BENNETT, K, and HALL, A M (editors). (Quaternary Research Association Field Guide.)
- BONDEVIK, S, MANGERUD, J, DAWSON, S, DAWSON, A, and LOHNE, Ø. 2003. Record-breaking height for 8000-year-old tsunami in the North Atlantic. EOS, Transactions American Geophysical Union, Vol. 84, 289–293.
- BIRKS, H H, GULLIKSEN, S, HAFLIDASON, H, MANGERUD, J, and POSSNERT, G. 1996. New Radiocarbon Dates for the Vedde Ash and the Saksunarvatn Ash from Western Norway. Quaternary Research, Vol. 45, 119–127.
- GRÖNVOLD, K, ÓSKARSSON, N, JOHNSEN, S J, CLAUSEN, H B, HAMMER, C U, BOND, G, and BARD, E. 1995. Ash layers from Iceland in the Greenland GRIP ice core correlated with oceanic and land sediments. Earth and Planetary Science Letters, Vol. 135, 149–155.
- HALL, A M, HANSOM, J D, WILLIAMS, D M, and JARVIS, J. 2006. Distribution, geomorphology and lithofacies of cliff-top storm deposits: Examples from the high-energy coasts of Scotland and Ireland. Marine Geology, Vol. 232, 131–155.
- FLINN, D. 1994. Geology of Yell and some neighbouring islands in Shetland. Memoir of the British Geological Survey Sheet 130 (Scotland) HMSO for the British Geological Survey, 130, 119.
- DAWSON, A G, DAWSON, S, and BONDEVIK, S. 2006. A Late Holocene Tsunami at Basta Voe, Yell, Shetland Isles. Scottish Geographical Journal, Vol. 122, 100–108.
- SMITH, D E. 1993b. Norwick, Unst; Burragarth, Unst; Sullom Voe, Mainland. 52–56 in The Quaternary of Shetland. BIRNIE, J, GORDON, J, BENNETT, K, and HALL, A M (editors). (Quaternary Research Association Field Guide.)
- COSTA, P J, ANDRADE, C, CASCALHO, J, DAWSON, A G, FREITAS, M C, PARIS, R, and DAWSON, S. 2015. Onshore tsunami sediment transport mechanisms inferred from heavy mineral assemblages. The Holocene.
- COSTA, P J M, ANDRADE, C, DAWSON, A G, MAHANEY, W C, FREITAS, M C, PARIS, R, and TABORDA, R. 2012. Microtextural characteristics of quartz grains transported and deposited by tsunamis and storms. Sedimentary Geology, Vol. 275–276, 55–69.
- DAWSON, A, and COSTA, P J M. 2013. Field Report 2. Will climate change in the Arctic increase the landslide tsunami risk for the UK? Shetland Fiedlwork June 2013