4.5 Holocene Terrestrial Change

4.5.1 Geomorphological change

Figure 19 shows the localities where something is known of physical changes in the Holocene to the landscape of Argyll and Bute: the number is distressingly small. We can divide this collection into analyses of colluviation (slope and soil erosion and subsequent re–deposition) and alluviation (fluvial erosion and deposition). The two can be closely linked.

Figure 19: Map of records of Holocene geomorphological change © Richard Tipping: Carradale, Kintyre (Tipping et al. 1994); Dalness Chasm, Glen Etive (Brazier et al. 1988); Dunadd (Miller and Ramsay 2001; Housley et al. 2004; Housley et al. 2010); Islay (Cressey unpublished); Loch Etive (Howe et al. 2002); Oban region (Rhodes et al. 1992; Macklin et al. 2000); Torbhlaren (Tipping et al. 2011).

4.5.1.1 Colluvial activity

Tipping et al. (2011) obtained perhaps the most securely dated record of slope and soil erosion in Scotland at Torbhlaren in the lower Add valley. Peat forming on the valley–side above the floodplain received varying amounts of soil eroded from sandy glacio–marine terraces above. An alluvial fan on the slope began to form from c. 7500 cal BC and in one colluvial record, colluviation also commenced within the Mesolithic period, at c. 5110 cal BC. Despite being only a couple of hundred metres apart, the three dated sediment cores suggest that the phases of most intense soil erosion were diachronous. Sediment delivered at high energies to one borehole between c. 4800 and c. 4250 cal BC has no counterpart in an adjacent borehole, though comparable high–energy events might be recorded in both at c. 3800–c. 3200 cal BC and c. 3550 and c. 3400 cal BC. Further high–energy mineral sediment movement occurred between c. 200 cal BC and c. cal AD 50 and after c. cal AD 700.

In the Oban region, Rhodes, Rumsby and Macklin (1992) and Macklin et al. (2000) analysed soil erosion from geochemical signals. Rhodes et al. (1992) report analyses for a range of elements at Gallanech Beg but with no discussion of the data while Macklin et al. (2000) report only potassium (K) in three additional peat sequences at Lon Mor, Bhuilgh bhith and Cnoc Philip. They argue that soil erosion was greatest in the transition to farming between c. 3500 and c. 2000 BC, but also in the same paragraph argue for soil stability between c. 3600 and c. 2000 cal BC. Finally, Cressey (unpublished) used geochemical signatures to trace the development of a medieval landscape at Lochs Finlaggan and Ballygrant on Islay.

In upper Glen Etive, the Dalness Chasm debris cone, a large fan–shaped sediment pile formed from eroded material high above, formed between earliest Holocene deglaciation and c. 2200 cal BC. Around cal AD 1600 the clearance of hazel scrub by people, including by fire, is thought to have re–activated the lower slope of the fan by fluvial erosion (Brazier, Ballantyne and Whittington 1987) though the date of this reworking is coincident at least with the 'little ice age' (Ballantyne 1991). Innes (1983) argued that over–grazing of upland slopes since the late 18th century greatly increased talus and debris cone formation.

4.5.1.2 Fluvial activity

In Loch Etive itself, Howe et al. (2002) described how terrestrial sediment sources have dominated sedimentation in this sea loch throughout the Holocene, with accelerating sediment accumulation rates in the last c. 200 years originating, probably, from changing land use from, for instance, woodland management for the iron–smelting industry.

In and near Kilmartin Glen, Tipping et al. (2011) described Neolithic and Bronze Age river development and channel change in the lower River Add at Torbhlaren. There was increased flood frequency/magnitude at around 2400 cal BC and major changes in river behaviour at c. 1240 cal BC but unchanging channel dimensions through the Holocene suggest that discharge did not significantly vary.

To the west, Housley and co–authors 14C dated one of many peat–filled abandoned channels of the lower Add near Dunadd to the BC–AD boundary. No significance can be attached to this age estimate, though, without also dating other channels. The peat filling this channel was terrestrial (Miller and Ramsay 2001; Housley et al. 2004; Housley et al. 2010). It would be interesting to know the altitude of this channel in light of the discussion on early historic Relative Sea Level change (above) but this is not given.

Carter and Tipping (1991–2) and Tipping, Carter and Haggart (1994) mapped the terrace surfaces along the lower Carra Water on Kintyre. Most are related to glacial or relative sea level stages but the extensive Rhonadale Wood Terrace is fluvial. It began to accumulate between 4500 and 3500 cal BC when sufficient sediment from earlier–formed terrace fills or soils on slopes filled the river enough to promote frequent flooding and the creation of a floodplain. Its surface was abandoned by the river as it incised at some time in the last few hundred years.

These isolated reports are not nearly enough to write an informed history of landscape change. There is, perhaps, the suggestion that Neolithic colluviation and alluviation were significant, and this would be different to other regions of the British Isles but the evidence is limited. The region needs such a history because changes in human settlement and land uses are implied in the interpretations of some of these physical changes although abrupt climate change provides an additional strong hypothesis. It is very likely that geomorphic changes in the highlands and on steep slopes such as debris cone formation were climatic in origin (Ballantyne 2004) because pastoralism is likely to have made few impacts, but colluviation in lowland valleys as at Torbhlaren need not have been. Yet here Mesolithic and Neolithic colluviation occurred in the absence of evidence for major human activities in the pollen record (Tipping et al. 2011). The prevailing paradigm in Holocene alluviation in the British Isles is also for a climatic causation (Macklin, Johnstone and Lewin 2005) but many case studies do not explore the timing or scale of human activities, biasing somewhat the conclusions. Where both hypotheses are equally appraised, human activity emerges as important. Argyll and Bute should provide an important testing ground for this debate.

4.5.2 Pedological change (Figure 20)

There were three archaeological sites known to Donald Davidson and Stephen Carter in 1997 that allowed some assessment of soil type and land use quality in the past: An Sithean on Islay (Barber and Brown 1984), Cul a' Bhaile on Jura (Stevenson 1984; Whittington 1988) and at Achnacree (Soulsby 1976; Carter and Dalland 2005). At all three, soils had become acid podsols by the Bronze Age. Impoverished soils were probably the norm in Argyll and Bute in later prehistory. Nevertheless, An Sithean (Canmore ID 37374) and Achnacree contain two of the best–explored later prehistoric field systems in Scotland, though at An Sithean the system is less coherent than first appears (Barber and Brown 1984; Halliday 2015, 284). Beneath the Caisteal nan Gillean II shell midden on Oronsay, the soil was also acid and leached (Paul 1987). Barber (1981) argued that soils at the monastery of Iona (Canmore ID 21649) were anthropic plaggen soils, a deliberate introduction as agricultural improvement: it would be good to test this idea further.

Figure 20: Map of records of Holocene pedological change © Richard Tipping: Achnacree (Soulsby 1976; Carter and Dalland 2005); An Sithean (Barber and Brown 1984); Cul a' Bhaile (Stevenson 1984; Whittington 1988); Iona (Canmore ID 21649) (Barber 1981); Moine Mhor (Housley, Clarke and Campbell 2007); Oronsay (Canmore ID 37804) (Paul Paul 1987).

The Moine Mhor is a raised moss, formed on an impermeable substrate of marine clay following mid–Holocene sea level fall, probably over a very long time (above). Such mosses spread laterally: the Moine Mhor at its fullest might have covered 1800 ha compared to 1200 ha today. The archaeological potential beneath the moss remains undiscovered despite ground penetrating radar survey (Housley, Clarke and Campbell 2007).

The blanket peat cover at Achnacree has been 14C dated, though poorly (Carter and Dalland 2005), to between c. 1550 cal BC and c. cal AD 250. The moss has formed on well drained glaciofluvial sands and gravel, so that a palaeo–hydrological signal for peat inception must be suspected, somewhat obfuscated by the great range of the available calibrated 14C assays. The Moss of Achnacree remains one of the most significant later prehistoric archaeological landscapes in the British Isles: its paleo–environmental potential has only inadequately been explored despite the attentions over several decades of many leading archaeologists (see Carter and Dalland 2005). The glaciofluvial sand and gravels are studded with kettle holes, some large but many small, basins containing complete Holocene sediment records which are key. Carter and Dalland (2005) rightly indicate that we would now have less confidence in the stratigraphic integrity of pollen within soils. The small–diameter kettle holes must record, free of disturbance, pollen–analytical data on later prehistoric agricultural activities. Attention has focused on the Bronze Age field system/s at Achnacree, but the kettle holes must also contain a record of the earliest agriculture on substrates known elsewhere to have attracted early Neolithic agriculture (Fairweather and Ralston 1973; Murray et al. 2009), very close to the chambered cairn at Achnacreebeag (Ritchie 1970) which plays such a significant role in Alison Sheridan's (2007, 2010) chronology of earliest Neolithic introductions.

Blanket peat is not common in Argyll and Bute, but the region has peaty gleys in abundance because of a clay–rich till (boulder clay), extensive poorly drained plateaux and a wet climate. Nothing is known of their development but they represent a major impediment to agricultural potential. It is important to understand the timing of this critical pedological change in the region and its relation to climate change and human activities.