Some of the ideas discussed in this blog are published in my new book called "The Stonehenge Bluestones" -- available by post and through good bookshops everywhere. Bad bookshops might not have it....
To order, click

Thursday 16 June 2011

Devensian erratic flushing from South Wales

This map, which I have used before on this blog, shows the maximum extent of glacier ice in South Wales during the Devensian or last glaciation, around 20,000 years ago.  If -- as seems possible -- a scatter of Pembrokeshire erratics was left across the South Wales coastlands at the end of the Anglian Glaciation, or later glaciations, these erratics might well have survived in central and South Pembrokeshire and on the western tip of Gower, because these areas were ice-free.  However, a series of Welsh Outlet Glaciers flowed down the main valleys from the uplands of mid-Wales during the Devensian, in the valleys of the Towy, Taf, Tawe, Neath, Rhondda, and xxx, to name but a few.  These glaciers expanded out to the present coastline and indeed beyond it, taking on the character in some cases of piedmont lobes.  Within the limits of these lobes, the glacial deposits that remain are of Welsh origin, with till that is recognizably different from the Irish Sea till, and with a suite of erratics transported from the north.  In such circumstances, the chances of any Irish Sea till surviving were very small -- and most if not all of the erratics left during the previous Irish Sea Glaciation must have been "flushed" or moved southwards, into the area now submerged in the Bristol Channel.  If there are bluestones that were laid down in a trail from Preseli to Somerset, they will be found, if anywhere, on the sea bed.

More on erratic trains and glacier bed entrainment

This is an interesting photo, taken near the edge of the Greenland Ice Sheet.  We can see three different types of erratic train.  On the right there are two moraine trails which can be traced back to mountain fronts which stand clear of the ice.  These trails combine into one near the edge of the photo.  Lower down, on the right side of the photo, there is a discernible moraine which can be traced back to a "rise" in the ice surface -- attributable to an isolated upland which is perhaps an accumulation centre although it is really a part of the ice sheet.  There must be some erosion going on here, with debris entrained and carried away from the location where erosion is active.

The third and longest erratic train or moraine ridge spreads out at the bottom right of the photo where the ice is decelerating and spreading laterally.  But notice how thin and well defined the trail is on the surface of the ice sheet, stretching far into the distance.  The trail appears to pass between two upland areas, but it has some of the characteristics of a medial moraine.  But we can see no obvious source for all this debris, which appears to be feeding into the glacier more or less continuously.  But the source must be BENEATH the ice, and not above it -- and this illustrates the type of entrainment which I envisage for the bluestone erratics that have found their way to Stonehenge.

Wednesday 15 June 2011

On the Trail of Bluestone Erratics (3): Transport

Following my posts on entrainment, it's time to turn my thoughts to the processes which operate in glaciers which enable them to transport debris of all sizes (including massive rocks sometimes weighing hundreds or even thousands of tonnes) from one place to another. 

Once boulders, slabs or monoliths of rock are trapped on, within or under a glacier they will be moved at varying rates, depending upon the mechanics of glacier flow.  These mechanics vary in intensity and relative importance from one glacier to another, about which more in a moment.  But first, we can say that the simplest mode of transport occurs when rockfalls or avalanches from mountainsides above a flowing glacier introduce either continuous or intermittent (pulsed) supplies of debris onto the ice surface.  The Foothills Erratic Train, which I have described several times in this blog, started life as a big avalanche of rock onto the surface of a dynamic valley glacier in the Rockies;  but there must have been many more avalanches from more or less the same place, because the erratic train is strung out over a great distance and formed over six centuries.  We do not know whether the avalanche debris spread out as an apron over much of the glacier surface, or whether it was restricted to the valley side.  In the latter case, we can envisage a classic lateral moraine situation such as we see on hundreds of valley glaciers in all of the high mountain areas of the mid and high latitudes.  In most cases, however, the supply of debris is seasonal if not continuous from all parts of the valley side, with frost shattering and gravity playing key roles.

 Before and after.  The Sherman Glacier landslide of 1964 covered much of the lower part of the glacier with a thick apron of rockfall debris.  This was a catastrophic event resulting from a powerful earthquake.

Down-valley, where valley glaciers converge, a lateral moraine will often be joined to another at the tip of a spur to become a medial moraine.  Complex glaciers, which have several tributary glaciers flowing into them, will sometimes take on a striped appearance with a number of medial moraines moving along in parallel with one another, as on the Barnard Glacier in Alaska;  but if one or more of the glacier tributaries exhibits surging behaviour, the moraines will be displaced or pushed out of position in a manner which is instantly recognizable by glaciologists.  These strange surface patterns are far more frequent than was once supposed -- and surges can move erratic trains many miles to the right or left of their predicted courses. 

 This classic photo shows how the lateral moraines of the Barnard Glacier are transformed in to medial moraines on every spur end.  This debris almost all comes from frost shattering on the rocky slopes above the glacier -- and most of it remains on the glacier surface.

Moraine loops formed during surges of the Susitna Glacier tributaries

As long as this lateral / medial morainic material continues to move along on the glacier surface beneath the firn line, where the glacier surface is melting, it will remain at the surface and will not be greatly modified by the processes of abrasion.  In other words, the edges of the boulders, slabs and pillars will remain sharp and angular, and will not be rounded off.  When we see medial and lateral moraines emerging at a calving glacier snout, we are often amazed to see that they are superficial, sitting on the ice surface and with hardly any debris beneath them, deeper within the ice.  Medial moraines can sometimes penetrate deeper, because when two glaciers come together there will be some disruption and redistribution of surface materials as two ice streams are blended into one. 

If a rockfall introduces debris onto a glacier surface above the firn line, it will quickly be buried by accumulating snow, and will effectively travel downwards as well as forwards, into the body of the glacier.  In this case the debris might be scattered or dispersed within the ice, and might also be subjected to the processes of abrasion as smaller particles move past larger ones and leave them polished, scratched or even gouged.  They can also be broken in circumstances where pressure variations occur within the ice.  In very rare circumstances (where there is a great deal of bottom melting beneath a glacier) large blocks may even pass down from the surface of a glacier all the way to the bed........ and the chances of such blocks remaining "fresh"and sharp edged in such circumstances would be very low indeed.

 Moraines on the surface of glaciers to right and left of Malmberget, East Greenland.  Look at the glacier on the right -- the medial moraine appears at the surface below the firn line.  In the higher part of the glacier it is submerged beneath snow and firn -- but its origin can clearly be traced back to the peak on the far horizon, with snowfields on either side of it.

We can say with some confidence that the Darwin Boulders of Tierra del Fuego  (which are found in two erratic trains of limited length, and do not extend back to the place of origin) were probably transported supraglacially (on the surface of the ice) following short-lived pulses of rock supply associated with avalanches of valley-side collapses.  We can also say that the Foothills Erratic Train was a result of intermittent but ongoing avalanches or collapses on a high mountainside, with massive piles of rock debris transported initially along the edge of a valley glacier and then along the contact zone between the Cordilleran Ice Sheet (in the west) and the Laurentide Ice Sheet (in the east).  From the "debris input point" to the eventual ice edge hundreds of kilometres to the south, the ice surface must always have been melting rather than accumulating, since the erratics that were dumped in a long string (as shown in a previous post) show remarkably few signs of glacial erosion.

The Foothills Erratic train, stretching for about 600 km and originating near Jasper in British Columbia.

So far so good.  What this tells us is that erratics that are entrained into or onto a glacier do NOT generally become scattered as "erratic fans" since ice tends to flow in a linear fashion with very little lateral spreading, except in exceptional circumstances where there may be a glacier snout lobe or piedmont.  Erratics tend to be transported, in ideal conditions, in a straight line, or on a curve.  But as we can see when we look at the map of the Foothills Erratic Train, there are some thickenings and thinnings of the line which might represent either increases or reductions in the rate of debris supply from the "source mountainside" -- and occasionally there are sharp changes in direction or kinks which probably indicate that for some reason either the Cordilleran Ice Sheet or its counterpart to the east experienced a pulse (or maybe a surge from a distant upland glacier) which displaced the contact zone between the two more or less parallel ice streams.

Erratic fans are in fact best interpreted as series of diverging erratic trains of differing ages. 

Where fans of erratics are found in the landscape the explanation may be that entrainment from a particular rock source continued over a long period of time during which the direction of ice movement changed -- sometimes by more than 45 degrees.  Sometimes the erratic train created during the early phase of a glaciation is buried by later sediments so that the erratics are not displaced or moved sideways as later erratic trains are let down either to right or left.  But more commonly, I suspect, the erratics left in a landscape may be of many different ages, having been entrained originally in one glacial episode and then again (maybe more than once) in later glaciations.  If we had been able to track their movements, I suspect that we would have seen something akin to a zig-zag pattern, ie something that appears to be truly erratic or eccentric.  However, nature is actually rather well organized, so the pattern is not chaotic or random; it is subject to analysis like everything else, and capable of explanation. Each point in the zig-zag would represent a change in the direction of ice flow or a temporary pause (maybe lasting for a whole interglacial!) in the movement of the boulder;  each segment of the line would represent a transport phase when ice movement was more or less stable or consistent.

Let's move on to consider erratics carried on the bed of a glacier.  Here it is quite possible for movement to be more erratic still, with abundant stops and starts and changes of direction.  If a large boulder remains on the glacier bed while it is transported over a great distance, that implies that the glacier is warm-based and that it is melting on its bed.  This melting provides water for lubrication, and facilitates the faster movement of the glacier,  the erosion of the bed, and the abrasion and even the splitting or crushing of the boulder itself -- so it might reduce in size, and it would almost certainly have most of its rough edges knocked off.  If the rate of melting is very fast, then a lodgement till might be plastered onto the glacier bed, and large boulders might be effectively buried and incorporated into this thickening till layer.  if that were to happen, the onward transport of the erratic would stop.  But if a large boulder does keep moving, it might be carried on a sinuous course by the plastic deformation of basal ice around and over obstacles on the bed; it might "stick" for a while and then be released and moved on; and it might eventually come to a complete and final halt on the bed if it is wedged up against a large and immovable rock outcrop or hill, so that the ice then simply moves over it as it moves over a drumlin or any other streamlined form (of which there are many).

Finally, to erratics which are carried within the body of a glacier.  Now things become much more complicated, if they are not complicated enough already.  If big erratics are carried up into a glacier by ice subject to compressive flow (in other words, if the ice is decelerating or slowing down because it has encountered an obstacle), there may be internal deformation of the ice if it is close to the pressure melting point.  However, if the ice is cold the stresses will result in shearing or faulting, with shear planes carrying relatively fast-moving ice up and over masses of ice that are slowed by friction at the bed.  If an obstacle to the glacier's progress is small, a single shear-plane (which may be many hectares or square kilometres in extent) may suffice to enable the glacier to maintain its forward motion, but if an obstacle is substantial (ie a subglacial range of hills or mountains) then there may be multiple or "stacked" shear planes, with each slice of the glacier moving faster than the slice immediately below it.   A cross section through such a glacier would look like a multiple-layer cake, with sponge slices (the ice layers) separated by thin cream layers (the shear planes).  Other analogies can be found in the field of geology.

What happens to the erratic boulders that are trapped in the ice in this sort of scenario?  It's difficult to be quite sure, since these environments are incredibly difficult to observe.  We can obtain some clues from the snouts of glaciers like the Thomson Glacier and White Glacier on Axel Heiberg Island, which I visited many years ago.  Here there are massive thrusts at the glacier snout.  A huge moraine at the snout acts as a barrier to the forward movement of the ice, and so the advancing glacier has pushed new ice up and over the barrier via a series of immensely complex thrusts or shear planes.  Some of these are short-lived and operate only for a while, before being intersected and sliced up by later episodes of shearing.  Debris which is in contact with these shear planes has to be substantially modified by abrasion, gouging and crushing processes, because the environment is a very dynamic one.  So boulders will show much evidence of smoothing and scratching by the tools which have been fellow travellers.  But what happens to boulders which escape the attentions of these shear planes and which travel within the "slices" of glacier ice?  In theory, they could remain almost unmodified, transported for maybe hundreds of kilometres within the glacier without ever coming into contact with other debris.

 Shear planes appearing at the surface on the snout of White Glacier.

If you think this is all too complicated, rest assured that what I have written above is a simple summary of what goes on during glacial transport -- to get a more realistic picture you need to read a few of the thousands of detailed articles about transport processes in the glaciological and geomorphological literature........

So let's apply some of the above to the situation in South Wales and Southern England, in the domain of the Irish Sea Glacier and the glaciers coming southwards from the Brecon Beacons and the South Wales Coalfield.  On this blog I have published many maps showing the idealised directions of ice movement as far as we can work them out from striae on rock surfaces and from the distribution of erratics on the ground surface and in sediments.  One such map (based on the work of Kellaway) is shown below, together with a map of known or assumed erratic trails in South Wales.  Both maps are composite, blending evidence from a number of different glacial episodes, collected by a number of different researchers.  The erratic trails which appear smooth were in all probability much more complex, with Pembrokeshire erratics (for example) moved SE, SSE or ESE at various stages.  Those erratics that were carried across Carmarthen Bay towards Gower and the Vale of Glamorgan would also have been moved S, SE or even SW at various times by large outlet glaciers moving down towards the Bristol Channel Lowlands from the Welsh Ice cap.  These erratics would have followed zig-zag courses.

One interesting fact to have emerged from both field research and glacier modelling exercises in recent years is that the big ice masses of the British Isles showed a sort of "pulsing" behaviour during the Devensian, with advances and retreats coinciding with "binges" (growing or accumulation episodes) and with "purges" (rapid evacuations or advances) respectively.  These binge and purge episodes were generally out of phase in various sectors of the British-Irish ice sheet. The amazing "animated" models created by glaciologists in Aberystwyth and other UK universities illustrate this perfectly.  The scenario during the Devensian is that of expanding glaciers flowing down towards the Bristol Channel from the South Wales uplands in an early glacial phase; the arrival and expansion of the Irish Sea Glacier (or a Celtic Sea Piedmont Glacier) during the peak glacial phase, around 20,000 years ago; and finally one or more phases of valley glacier advance during a late glacial phase.  We have to assume that this scenario, or something like it, was followed during each of the big glacial episodes to have affected western Britain..........

Back to the erratics.   The coastal lowlands of South Wales must have been affected during the Ice Age by a number of episodes involving the Irish Sea Glacier and a number of episodes involving Welsh ice, pouring out of the South Wales valleys and perhaps forming piedmont glaciers, as shown by Prof David Bowen on many of his maps.  In such circumstances, what chance might there have been for glacial erratics from Pembrokeshire or Ireland to have survived in situ either on the Gower Peninsula or in the Vale of Glamorgan?  Not a lot.  It would, in my view, have been comparatively rare for Pembrokeshire erratics to have been dumped inland of the present coastline; most will have been dumped in the area now inundated by the sea, in Carmarthen Bay and the Bristol Channel.  Those that were dumped on the area that is now dry land would, in all probability, have been "flushed out" towards the south by the South Wales Valley glaciers in the later phases of each glaciation.

In my view, it is extraordinary that there are ANY Pembrokeshire erratics left on the Gower or the Vale of Glamorgan; but against the odds there are indeed a few -- in particular igneous erratics from the St David's Peninsula found in western Gower.  Other erratics from Pembrokeshire have also been found at Pencoed and near Cardiff; some of these are in the "Storrie Collection" held at the National Museum of Wales.

Has this got us anywhere?  I hope it has.  Maybe the foregoing will help to explain why I roll my eyes and sigh when people say to me "If the bluestones at Stonehenge really are glacial erratics, why isn't there a continuous trail of them all the way from Preseli to Salisbury Plain?"  If only the theory of glacial transport was that simple.......... but it isn't, as I keep on trying to explain to the archaeologists.

Sunday 12 June 2011

The Sarsens -- quarried or just collected?

The new report by Field and Pearson seems to come down quite heavily against the idea that the sarsens at Stonehenge were collected from the Marlborough Downs and transported very laboriously all the way to the site of the monument.  They see that hypothesis as unnecessary and redundant.  They suggest that the stones were collected from the vicinity of Stonehenge -- and they make the point that this suggestion has been made many times before, by many Stonehenge investigators over the past 200 years or more.  Further, they suggest that the builders were indecisive, or changed their minds about the stone settings they were trying to build, over and again.  They also seem to accept that the sarsen monument was never finished, or completed in the manner envisaged by Anthony Johnson and the great majority of modern archaeologists who have worked on Stonehenge.  Another attractive idea is that the sarsen monument was intended to be horseshoe-shaped, rather than circular.

Most of that is what I have been saying as loudly as possible -- without any amplification from the archaeology establishment or its journals -- for the last few years.  We get there in the end.....

The really interesting area of debate on the sarsens now centres on the question of whether they were quarried, or excavated out of the ground, or whether they were simply collected up from a surface litter.  Field and Pearson seem to favour the idea that the stones were soft when they were dug up, on the basis that the shaping of the stones (which is extensive) could have been done easily if the stones were fresh, but not very easily if they had been lying around on the surface for many thousands of years.  They cite the fact that there are hollows in the Stonehenge area that might have been "sarsen quarries" -- now very much denuded. 

I don't have any strong ideas on this, since I don't know the landscape or the sarsen geology details, but I do wonder whether large sarsens littering the surface might have had hardened upper faces and relatively soft downward faces -- allowing the shaping and smoothing of one face but not the other.  It has been noted many times that the outer faces of the sarsens (ie the surfaces facing outwards in the stone settings) are rough and untreated, whereas the inner faces (towards the centre of the monument) are often smoothed or shaped in some way.  This is often explained as a cosmetic or aesthetic feature, created for effect by the builders.  But maybe this architectural / design feature was an instinctive response by the builders to the one rough hard face / one smooth soft face phenomenon?

Worth thinking about.......

Friday 10 June 2011

The times they are a'changin'..........??

One of the excellent EH plans featured in the new publication

I have come across this excellent and wide-ranging report by David Field and Trevor Pearson.  Very interesting indeed -- and in advance of a careful reading the following things stand out from the pages:

1. An admission that the sarsen stones might well have come from the immediate locality of Stonehenge, and that the idea of sarsen-collecting expeditions to the Marlborough Downs is dubious and probably unnecessary.

2.  An acceptance that the bluestones MIGHT be glacial erratics (although the authors still defer to Profs D and W and don't want to stray too far from the party line on this.....)

3.  They have at least read my book, although it would have been nice if they had spelt my name correctly......

4.  An acceptance of the idea that the Stonehenge stone monument was probably unfinished, and that the builders went through many changes of plans and probably ran out of stones.

I'll come back to all of this in much greater detail, but does this mean that AT LAST the people working for English Heritage are prepared to challenge the long-held belief system established by Atkinson and others and perpetrated by Darvill and Wainwright?  Maybe, maybe.......

ISSN 1749-8775

Sunday 5 June 2011

Roches Moutonnees

 Above -- a very old sketch of a roche moutonnee hill in Sweden.  Below -- a very large roche moutonnee in Yosemite National Park, USA

 Those pictures on Google Images are a real hotch-potch of all sorts of things, only some of which I would classify as roches moutonnees or streamlined forms. Above are two illustrations that DO fit the bill.  The up-glacier side is on the right in both photos -- with a much gentler slope that has been smoothed and abraded by ice under pressure as it rises.  In contrast, the down-glacier (left) sides have been steepened and broken up by plucking and quarrying on a grand scale.  The processes are still being debated, but here the basal ice will have been under tension -- there may well have been a cavity, and freezing and thawing on the glacier bed might have been instrumental in the processes of dragging large blocks of stone away from the hillside.

The assymetric long profile is typical -- features which do not look like this are better referred to as whaleback forms.  They are much more irregular and "hummocky", and in whaleback territory it is sometimes difficult to discern the direction of ice movement.

I have seen roches moutonnees ranging in size from whole mountains to very small features no more than 5m long and 2m high -- but the processes responsible for them are essentially the same in all cases.

Streamlined hills on the Somerset Levels?

 I have just received this message from Alex Gee, for which many thanks!

I am writing to ask your opinion (as an expert in glacial geomorphology)  on the following observations.  For some years my friends and I have been digging a cave on the south flank of the Mendip Hills (above Wells).  After digging we often stop to admire the magnificent view across the Somerset levels. 

The Somerset Levels are marked in blue on this map

 Over time I have noticed that a lot of the small rounded hills on the levels, have a convex  asymmetric profile, their longitudinal axis being roughly east to west.  The westerly hill slope is gently graded (grass pasture used for cattle grazing), the easterly slope having a steep gradient (being unfarmed and wooded).  In plan form some are crudely tear drop shaped. Their length to height ratio is approximately  4 or 5:1,  the summits of the hills are generally at the eastern end, and 4/5 to  3/4 of the hills length to the east.   

 The more I look at them, the more the thought has grown that their assymetric shape (and orientation), could be due to easterly movement of glacial ice from the Bristol Channel?   Could the hills be a form of roche moutonnee?  There is no obvious structural explanation of their form. The hills are mainly composed of horizontally bedded Rhaetic and lower Jurassic strata, lying unconformably on Triassic  Mercia Mudstone (playa lake deposits) and Triassic Dolomitic Conglomerate. The slopes cut across the contact indiscriminately. This rules out the role of stratal dip.

Yesterday, I pointed out the morphology of the hills to my friend; who is a geologist and Mining engineer( ex-Cambourne school of mines).   His immediate comment was " If this was north of Watford then, I'd say it was due to glaciation".

This is an interesting observation, which I have never seen before -- and even Geoff Kellaway, who assembled much evidence for glaciation in Somerset, did not (as far as I know) mention possible streamlined hill shapes.  Some more work is obviously needed on this........