The Glaciation of Rhosyfelin

A glacial erratic with a distinct orange-coloured patina -- exposed among the litter on the floor of the 2012 dig at Rhosyfelin

I'm often asked what the evidence of glaciation is in the Rhosyfelin area.  Well, there is plenty of it about.  The upper part of the Brynberian Valley (in which the rocky crag is located) has been glaciated on at least two occasions, and probably more.  The glacial episodes about which we know most were the Anglian (about 450,000 years ago) and the Devensian (reaching its peak around 20,000 years ago.)  The former was almost certainly the time of the GBG (Greatest British Glaciation) and in the latter the Irish Sea Glacier was very extensive and may even have extended beyond the Devensian limit in some areas.  The Wolstonian / Saalian Glaciation was probably extensive, but traces of it are very difficult to differentiate from those of the Devensian, as Chris Clark, Phil Gibbard and many others have pointed out.  So let's put that to one side for the moment, while allowing the possibility that it might have done some work, at the very least, in modifying the landscape of West Wales.

Suggested limit for the Irish Sea Ice in the Devensian period.  The glacier came in from the north and north-west.  The dotted area shows the suggested extent of the Preseli ice cap, which would have been contemporaneous.

During both the Anglian and Devensian glaciations, the ice of the Irish Sea Glacier came into this area from the N and NW.  First, what do we know about the Anglian ice extent?  It has been accepted for more than 100 years that the ice inundated the whole of Pembrokeshire and that it flowed up the Bristol Channel  -- and for well over 50 years it has been known that it impinged upon the coasts of Devon, Cornwall and Somerset.  This is indisputable -- and if any archaeologist or geomorphologist tells you otherwise, do not believe them. So the ice was very thick indeed -- maybe up to 600m thick -- as it flowed across this area.   Was entrainment of blocks of stone or monoliths possible at this time from the crags at Rhosyfelin?  Yes -- without a doubt.  I have covered this possibility -- and indeed likelihood -- many times before on the blog. Large glaciers and ice sheets can and do entrain large blocks on the up-slope flanks of upland barriers where glacier bed conditions are just right.  So entrainment was more likely to have happened on the northern flank of Preseli than either on the summit or on the southern flank, where the glacier bed at the peak of glaciation was most likely to have been frozen to its bed. This is a matter of ice physics or glaciology -- not a lot to do with geology, although obviously tors and rocky crags will have made perfect entrainment locations.  One of the problems with Rhosyfelin is that we do not know how much bigger the pre-glacial crag might have been.  It might well have been 10m or more higher than it is today.  Thousands of tonnes of the famous foliated rhyolite might have been removed.  At the moment, we have no way of telling -- but we are working on it.........

The deep gorges of the Brynberian and Nevern Valleys.  Rhosyfelin is the rocky spur on the western flank of the valley near the base of the photo.  These valleys have many features in common with the subglacial meltwater channel complex of Cwm Gwaun and Nant y Bugail.  There are other rocky crags in the Nevern Valley at Felin y Gigfran.

There is also a strong possibility that the river gorge around Rhosyfelin and Pont Saeson was cut initially during the Anglian Glaciation by torrents of glacial meltwater flowing subglacially or on the edge of a wasting ice mass, as the Irish Sea Glacier was breaking up.  In addition to the main gorge, there is a strange subsidiary channel which is both looped and humped -- and this is typical of the subglacial meltwater channels in North Pembrokeshire.  Sometimes these channels are simply swept clean of debris by the sheer force of meltwater flowing through them, at very high hydrostatic pressures.  I would not fancy the chances of any Anglian deposits (either glacial or fluvioglacial) surviving in the area, given that almost half a million years has elapsed -- but I would not rule this out entirely.

This brings us to the extraordinary flat face of the Rhosyfelin cliff -- revealed in spectacular fashion as a result of the recent dig.  This does not look to me like a quarried face, and I am currently convinced that it is a long glaciated face, fashioned above all else by moving ice.  I have not seen any striae on the face -- but I have not spent enough time there, and one day I'll go back and take a closer look. Of course, if there are striations on the face, they are just as likely to be Devensian as Anglian -- given that the ice has returned for a second (or third, or fourth) time to the same area.  One intriguing feature is that the direction of ice movement -- as far as we know -- was broadly from the NW towards the SE, which was perpendicular to the alignment of the Rhosyfelin rocky ridge.  So the natural direction for subglacial meltwater flow would have been up the Brynberian Valley rather than down it, if the dictates of hydrostatic pressure were followed.  It is more likely, given these facts, that meltwater flow was marginal, along or near the ice front of a melting glacier.

Did the Irish Sea Glacier cover this area during the Devensian?  I am increasingly convinced that it did (see the map above), and that the ice edge was somewhere to the south of Rhosyfelin, on the northern face of Preseli.  This assumption is based on a lot of evidence, including the presence of apparently fresh striations on Carn Alw, the presence of apparent marginal channels along the hillsides near Carn Goedog, and the fresh moraine near Tafarn Bwlch and Waun Mawn.  There must have been a great lobe of ice pressing in from the N and NW and filling the depression in which Brynberian and Rhosyfelin sit.  As indicated in earlier posts, some geomorphologists have speculated that there was a marginal lake in this depression either prior to the maximum ice advance, or shortly afterwards.  We await detailed stratigraphy and analysis of the sediments in this depression -- all will be revealed in due course.  There is a layer at the base of the sediment sequence exposed during the 2012 excavation which looks as if it might be lacustrine -- I did not have time to examine it properly.  It's only a few centimetres thick --  was it laid down at the bottom of this large proglacial "Lake Brynberian", or maybe on the bed of a much smaller lake confined within the valley, held up my a plug of melting ice somewhere to the north of Rhosyfelin?

The exposures on the flank of the 2012 dig.  There are many layers in the sediments, although the boundaries between them are diffuse.  The finest-grained deposits are at the base of the section, and may indicate lacustrine or slackwater deposition.  Note the accumulated litter in the foreground, where quite large blocks have accumulated beneath the cliff face.

What about the Late Devensian history of the Rhosyfelin area?  Well, we know that there is a great litter of broken debris beneath the cliff face, as we would expect beneath any steep rock face in Pembrokeshire.  If analogies around the Pembrokeshire coast are anything to go by, this litter must have accumulated after about 20,000 BP, with a late peak of frost shattering and periglacial activity around Zone III or the Younger Dryas, around 13,000 - 11,000 BP.  After the melting of the Irish Sea Glacier in this area, we might well have had 10,000 years or more of oscillating cold climate during which considerable damage might have been done to the Rhosyfelin crag -- so it is very difficult to imagine what it might have looked like during the glacier melt episode.

Several points emerge from an examination of the 2012 dig:

1.  The litter of broken rock seems to me to be entirely natural -- including the placement and positioning of the "big Rhosyfelin orthostat."

2.  I think that the dig might now have got down to the top of a till layer.  There are signs of erratic blocks in the litter -- including at least five big blocks of quartz (which might of course be rather local) and a stone with a yellowing surface patina which does not look like a lump of Rhosyfelin rhyolite (see photo at the top of this post).

Four quartz erratics exposed on the floor of the dig.  We do not currently know how far these might have travelled.  The are sub-rounded, and have been subjected to some wear during transport.

3.  I have not seen any sign thus far of any fluvio-glacial material, although I would not rule this out when the dig goes a bit deeper.

4.  A very intriguing feature of the exposed surface of the litter is the degree of rounding or smoothing that appears on the edges of many of the stones.  Naturally, this foliated rhyolite breaks down into fragments and slabs with rather sharp edges -- so something must have been responsible for the wear on the edges.  The only two candidates, realistically, are overriding ice and meltwater flow.  The former is, I think, a possibility, given that the till layer is now being exposed; the stones which we now see at the dig surface may have come from some distance away, or they might just have been overridden in situ. Some accurate provenancing is needed to sort this out!   Another explanation is that there has been a long process of rounding of stone edges by snowmelt, maybe during the many centuries during which frost shattering and scree accumulation were going on.  (I have to admit that I'm a bit confused here -- more work is needed.)

The rounded upper surface of one of the local rhyolite slabs exposed on the floor of the 2012 dig.  Many stones of all sizes are rounded in similar fashion.  What caused this rounding?  Overriding ice, or glacial meltwater?

5.  The thickness of accumulated slope deposits here is quite surprising, given the relatively short length and limited height of the slope to the NW of the dig site (that is the orientation of the steepest part of the slope.)  As mentioned above, the deposit with the finest and most consistent grain size is near the base of the 2m or so of exposed deposits, suggesting a temporary lacustrine or alluvial situation with relatively low-velocity water flow.  Above that, the sequence of deposits is complex, with a great deal of soil and slopewash material mixed with angular fragments, most of which are presumably rather local in origin.  There are several layers which appear rich in organic material -- and for which radiocarbon dates will presumably be obtained.  It will be interesting to see what comes out of these investigations.

In summary, we have traces of at least two glaciations here, with most of the major erosion and landscape development attributable to the earlier one, and most of the deposition of superficial deposits attributable to the later one.  There is undoubtedly a great deal more to be discovered, and I hope that MPP has some good geomorphologists on board to help him in the task of unravelling what actually went on.

If MPP wants to make use of the above material, he is welcome to do so, as long as he acknowledges the source.