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, 17 January 2019

Penfro Formation -- a lexicon correction needed

A type locality for the Penfro Till Formation?  Cemented till exposed at Black Mixen, Lydstep.

A type locality for the Penfro Till Formation?  Cemented and stained till near the coast at Ceibwr, overlain by cemented glacifluvial gravels and fresh grey-blue till.

I have sent this message off to BGS with a view to getting the lexicon record corrected, with respect to the Penfro Formation.    I don't object to the name -- some name has to be used for the ancient till deposits, and "Penfro" is as good a name as any.  But the record is misleading since neither West Angle nor Llandre should be used as a type location.

So what are the candidates for type localities?  I am suggesting Ceibwr and Black Mixen (Lydstep) since these are the two sites where an old cemented till can be seen in close proximity to a fresher till.  Other sites are Witches Cauldron and Whitesands -- but I need to do some more checking at both......

We'll see how we get on.  It would be good to visit the sites in the company of a BGS geologist, if we can organize that.

Penfro Till Formation.

I cannot work out where the type locality is supposed to be.  Is it Pencoed?
The two "partial type sections" referred to are highly misleading, and should be removed from the Lexicon.  West Angle does NOT reveal an ancient till (Penfro Till) beneath the raised beach.  Both Dixon and Bowen have misinterpreted the stratigraphy.  There is a till at the site, but it is of late Devensian age, and rests ABOVE the silt and clay series assumed to be interglacial.  Llandre is a useless site too -- there are apparently glacifluvial gravels there, but there are no stratigraphic markers that can be used to tie in with any regional stratigraphy.

The only two convincing sites where an ancient till can be seen in close proximity to a more recent till are Black Mixen (Lydstep) and Ceibwr.  Please use these as type localities instead.  What is the process for correcting the lexicon?

Wednesday, 16 January 2019

The curse of lithostratigraphy

I have spent a great deal of time on this blog celebrating stratigraphic correlations and stressing their importance in understanding what happened during the Quaternary in West Wales and in SW England in particular.   These are just a few of my posts:

Without stratigraphic matching and comparisons, we would not be able to work out common histories of events and environmental conditions -- but there are limitations and even dangers that we should be aware of.  There are different sorts of stratigraphies -- and we have already, on this blog, bemoaned the fact that our understanding of the Quaternary in Wales has probably been set back by decades by this weird thing called "aminostratigraphy" -- involving the correletion of layers of sediment in different locations according to the "amino acid ages" given to contained in shell fragments.  When you do that sort of thing with an immature dating technique, you are asking for trouble --and in Wales we have had trouble in bucket-loads........

When we talk of "lithostratigraphy" we are on safer ground -- after all, this is the basis of most solid rock correlation and dating, although fossil content, geomagnetic signatures and degree of metamorphism are among the other factors used by geologists in sorting out (for example) the Permian from the Pre-Cambrian.  Here is the Wikipedia definition:

The lithology of a rock unit is a description of its physical characteristics visible at outcrop, in hand or core samples or with low magnification microscopy, such as colour, texture, grain size, or composition.  It may be either a detailed description of these characteristics or be a summary of the gross physical character of a rock.  It is the basis of subdividing rock sequences into individual lithostratigraphic units for the purposes of mapping and correlation between areas. In certain applications, such as site investigations, lithology is described using a standard terminology such as in the European geotechnical standard Eurocode 7.

In Pembrokeshire the lithology of a rock if often a simple indicator of where it lies in the sequence. Even the least experienced geologist can differentiate between the red sandy rocks of the Old Red Sandstone, the fossiliferous grey limestones of the Carboniferous Limestone, and the black shiny anthracites of the Coal Measures.  Most can see the differences between igneous and sedimentary rocks, although metamorphics can be more difficult.  But when it comes to shales and  mudstones, most of us would have difficulty in differentiating, in hand specimens, between Ordovician samples from Ceibwr and Millstone Grit samples from near Haverfordwest.

When it comes to the Quaternary, things are also very difficult, since different events or climatic / environmental conditions will give rise to deposits with similar lithologies or appearances but with widely differing ages (an old stratified slope deposit might well look very similar to a fresh one).  So it might be difficult to sort out what is what unless there are very clear stratigraphic sequences on display which can be used for lateral correlations.  The Devensian Irish Sea till, for example, always occurs in the same stratigraphic situation (where we can see it) but it differs markedly in its physical characteristics, from one place to another, depending on what the "parent material" might have been, and depending on the glaciological conditions that obtained at the time of deposition. In cases like this my own instinct is to minimise the number of lithostratigraphic labels I use, since it would not help anybody if I was to refer to the Abermawr till, the Druidston till, the Newport till and the Gwbert till (and many others) by different names.  When I was writing my doctorate thesis in 1965, I tried always to refer to the Irish Sea till, but I did qualify that by referring to two sub-categories -- the coastal facies (made mostly of marine muds containing sea shells)  and the land facies (made largely of locally derived rock fragments in a sandy and silty matrix). 

So to the bigger problems, and to the reason for my use of the word "curse."

Unfortunately there are some geomorphologists who have made a specialism out of making lithostratigraphic correlations and publishing an endless stream of "lithostratigraphic regional and national correlations" in which the correlations keep on changing.  Confusion reigns, in a certain section of the specialist literature.  Prof DQ Bowen has probably published more than anybody else in this field, and some of the terms he has applied to some of the deposits have, as far as I know, never been used by anybody else.   However, in some cases, because he is the "specialist", others have deferred to him and his work,  without first checking carefully on reliability.  So we have the use of these terms:  formations, members and beds.  That's all very fine, and conforms to geological good practice, but it gets a bit mystifying as far as Joe Public is concerned.  In SW Wales we find things labelled as belonging to the St Asaph Formation and the Elenid Formation (correlated with Oxygen Isotope stage 2) and then lower in the sequence the Penfro Formation (correlated with OIS 16).  

I have referred to the Penfro Formation before, and have expressed strong concerns about its reliability as a label:

If we look at the definitive entry for this "ancient till"  in the BGS lexicon, under the Albion Glacigenic Group as the "parent unit", we find that the type localities are West Angle Bay and Llandre Quarry in Pembrokeshire.  That is an extraordinary error on the part of the geologists, since there is NO ancient till exposed at either site, and the sands and gravels at West Angle are demonstrably Devensian -- and therefore have nothing to do with the Albion Glacigenic Group. 

The critical central part of the West Angle exposure.  Here we see the erosional contact between the older deposits (partly interglacial) to the south and the younger (partly glacial) deposits to the north.  Dixon, Bowen and others might not have seen this contact clearly, and so they assumed (erroneously) that some glacial deposits are older than the grey silt and clay series.

 There are admittedly some glaciofluvial gravels at Llandre that might be old, but neither DQB nor anybody else has ever published a full description of them, and there is nothing at the site to tie the gravels into a regional sequence.

Type localities need to be stable, accessible and clearly tied into a regional stratigraphy.  Llandre is useless as a type locality for anything;  if West Angle is used as a type locality for anything, it should be for the Devensian or the Ipswichian, and most definitely not for the Penfro Formation or the Anglian glaciation.

Then we come to another problem of lithostratigraphy.  Once somebody has studied a region in detail and published what might be deemed to be "definitive" studies of the sedimentary sequence, with "formation" or "member" labels attached,  and type localities identified, there is a tendency for those labels to be over-used or applied in a non-critical fashion.  I became very much aware of this during my short visit to the Isles of Scilly a couple of years ago, as I looked at innumerable coastal sections on all of the main islands.  In a QN article I described and interpreted the deposits as best I could, given the limitations of time and resources:

Right from the start, I had problems in fitting the sequence and the characteristics of the deposits into the stratigraphic sequence devised by James Scourse in 1991 and subscribed to (with minor modifications) by others since then.  There were two problems, one relating to the choice of type localities, and the other to the choice of stratigraphic labels.  I had problems with the "Watermill Sand and Gravel" on the basis that Watermill Cove is not a very good type locality and that the raised beach sediments are neither sandy nor gravelly.  I had problems with the "Scilly Till" type locality at Bread and Cheese Cove since the till there is associated with glacigenic structures which are not all that widespread and since other till exposures elsewhere are perhaps more typical of the northern fringes of the islands.  The "Porthloo Breccia" I can live with, in that it is always in the same stratigraphic position and since Porthloo is a good type locality.  I had problems with the "Tregarthen Gravel" and the "Hell Bay Gravel"  labels since they did not seem to have consistent characteristics and since some exposures were not at all gravelly.  The labels seemed to me to be surplus to requirements.  I was not too sure about the "Bread and Cheese Breccia" either,  since in places it had exactly the same characteristics as the Porthloo Breccia.  And neither was I sure about the "Old Man Sandloess" since it too had rather variable characteristics from one exposure to another. 

Clay-rich till exposed at Chad Girt, in a rocky gully on White Island, Isles of Scilly.  Should this be called "Scilly Till", or something else.....?

In my QN paper I used these Scourse terms sparingly, just where I thought they might help in understanding the stratigraphic sequence -- but I have been heavily criticised for occasionally placing the deposits I observed into the "wrong" category in the "accepted lithostratigraphic sequence" !!  Hmmm.  As far as I am concerned, I have not accepted anybody's sequence, and will not accept any sequence unless it seems to me to be either accurate or useful.  In ice wastage environments depositional circumstances -- at exactly the same moment in time -- from place to place can vary enormously, and that needs to be recognized.   It is a singularly fruitless exercise to become worked up about the application -- or non-application -- of predetermined labels when the only questions worth asking  are these:  What happened?  And when?

Hence my scepticism about lithostratigraphic labelling.  There are bad labels, and bad type localities.   If we become obsessed by labels and type localities, we can easily be dragged into a scenario in which we "force" the sediments we find into predetermined boxes -- and in the process lose scientific objectivity.  If labels are worth using, they will survive.  If some of them are causing more trouble than they are worth, they should be dumped.

Tuesday, 15 January 2019

UK Quaternary Domains

This is an interesting report, which outlines how the geologists of the BGS have gone about defining the details of British Quaternary stratigraphy -- using the same principles as they do for solid rock units. They are, after all, geologists and not geomorphologists..........  

ENTWISLE, D C, WILDMAN, G. 2010. Creation of the Till Thematic Layer. British Geological Survey Internal Report, IR/10/041. 14pp.

The pap above is a very interesting one  -- designed to appeal to a geographer like me!  It's a sort of landscape regions map, but with Quaternary surface characteristics and sediments added.

One thing of interest is the apparent acceptance of the BGS geologists that the Anglian ice pressed against the coasts of Somerset, Devon and Cornwall all the way from Weston super Mare to the Isles of Scilly.  So they must accept the presence of Quaternary sediments and landscape features all the way along the cliffline.  That is, of course, in tune with what many geomorphologists have said over the years, although there are many others (frequently cited on the pages of this blog) who vehemently disagree and who claim that the Anglian ice edge was located somewhere far to the west in the outer reaches of the Bristol Channel or in the Celtic Sea. (This latter view looks more and more untenable, especially as evidence accumulates for a Devensian Irish Sea Glacier that pushed far to the south of the Isles of Scilly.)  So the Anglian ice edge as shown on this map appears, in some places at least, to coincide with the position of the Devensian ice edge.

It follows that the Devensian ice limit, as shown for Wales on this map, has to be very unreliable indeed.

I'll shortly do another post on this, and on the Penfro Till Formation (supposedly of Anglian age) which the BGS lists in its lexicon, apparently on the basis of very dodgy evidence supplied by Prof Dai Bowen.

Here is another useful BGS Report:

An overview of the lithostrati- graphical framework for the Quaternary and Neogene deposits of Great Britain (Onshore).
British Geological Survey Research Report RR/04/04 38pp.

This map shows the glacigenic groups in Southern Britain, including the "Abion Glacigenic Group" outside the putative Devensian limit and the "Caledonia Glacigenic Group" within the Devensian limit.  The proto-river pattern is much discussed.  Note that the ice of the Anglian Glaciation is assumed to have affected both shores of the Severn Estuary, as far out as Flatholm, Steepholm and Brean Down.

You can see the relationship between the maps above and the "traditional" map showing ice edges and glacier flowlines -- after Boulton, Wright, Zeuner and many others.  The "Older Drift" / "Newer Drift" distinction is of course perpetrated by the use of maps such as this.  in reality, as ever, things are much more complicated than the maps suggest......

Wednesday, 9 January 2019

Ceibwr -- Quaternary stratigraphy and related events (2)

Further to my earlier post about the exposures now to be seen at Ceibwr, here are some more photos, from the flanks of the rocky spur that separates Channel 1 from Channel 2.  Again, no complete sequence is to be seen, but at the northern tip of the spur we can again see TWO tills of different ages, one of them solidly cemented and the other fresh, separated by stained and bedded gravels:

Sediments resting on an abraded rock surface, parts of which are still affected by waves during extreme storms coinciding with high tides. The rock surface, cut across thin-bedded shales, is at bottom left.  Resting on it is a stony deposit with cobbles and larger erratics which is interpreted as an ancient cemented till.  It is stained by iron oxide and manganese oxide cement, and has a slightly purple tint in its colouring. Above it we see fine-grained glaciofluvial gravels with a foxy red or brown colouring.  the gravels contain occasional larger cobbles.  The grey-blue uncemented till seen beneath the turf to the right of the photo is clearly younger than the stained and cemented deposits.

Close-up of the youngest (uncemented) till deposit, showing relatively stone-free Irish Sea till at the base, with a more stony layer above (at the position of the trowel), and a sandy and silty colluvium / sandloess close to the ground surface.

Close-up of the cemented stratified gravels in the same section, with traces of what might be frost "churning" under permafrost conditions and possible ice-wedge casts.  Alternatively these may be primary sedimentary structures, related to downslope flows of saturated materials in a dead-ice environment.

In a separate exposure just a few metres away, we see cemented till at the base, capped by a manganese oxide cemented hardpan, with foxy-red glaciofluvial gravels above.  Then, above a sloping erosional contact we see fresh blue-grey till and colluvium -- apparently redeposited and mixed up by downslope movement.

The deposits described above rest on this wave-washed platform.  It is difficult here to distinguish modern erosional features from old ones -- but abraded surfaces do seem to pass beneath the deposits.

Exposure on the eastern flank of the rocky spur, showing pseudo-stratified slope breccia made of local mudstones, beneath fresh and uncemented grey-blue till.  Slope deposits are found in the exposures only in locations where there are adequate upslope outcrops of solid rock and a steep gradient.  This is an indication that periglacial or permafrost conditions preceded the arrival of glacier ice at this location -- exactly as seen at many other North Pembrokeshire sites.

River cutting upstream from the footbridge, in the main Ceibwr valley.  Cemented foxy-red gravels are seen here beneath blue-grey till with abundant clasts derived from local slope breccias.

Steep-sided channel 2.  The rocky spur is on the right.  Note the rock-cut slope near the cars.

Apparent meltwater erosional features beneath an overhang on the flank of the rock spur.  The location is near the white vehicle shown in the above photo.  The smoothed rock surfaces have been damaged by subsequent weathering processes, suggesting great age.

I think it's important to get these images onto the public record, since this is going to be a very important Pleistocene site.  Overall, the stratigraphy is confirmed as follows:

Modern soil -- c 20 cms. Uncemented
Sandloess and colluvium -- c 50 cms. Uncemented. Holocene?
Clay-rich Irish Sea Till -- up to 2 m thick. Blue-grey colour.  Uncemented. Late Devensian?
Brecciated slope deposits -- up to 50 cms thick. Uncemented.  Early / Middle Devensian?
Clay-rich colluvium -- c 20 cms thick. Uncemented but stained / gleyed. Ipswichian interglacial?
Glaciofluvial gravels -- c 1.5m thick. Stained and cemented. Foxy red colour.  Anglian?
Stony till -- up to 1m thick. Stained and cemented by iron oxide and manganese oxide precipitates. Foxy red / purple / black colouring.  Anglian?
Brecciated slope deposits -- up to 1 m thick.  A pre-glacial Anglian phase?

Note that there are no traces here of the raised beach (Ipswichian) which we see at Poppit, Abermawr, Porthclais and many other locations. 

Further work is needed on the abraded rock surfaces which are exposed by occasional storm wave assaults at the exit of Channel 2.  Are these surfaces ice-moulded?  Are they simply old channel floors cut largely by meltwater erosive forces?  Or are some of them -- at various altitudes -- affected by present-day and ancient coastal processes?

Flattish rock surfaces, eroded across dark Ordovician mudstones, at the head of the Ceibwr tidal creek, near the limekiln.  Are we looking here at the bed of the meltwater channel, or are present-day coastal processes responsible?  However these surfaces have been cut, it seems that the sediment fill in the channel, between HWM and LWM, is not more than a few metres thick.

Monday, 7 January 2019

"Till boulder" on James Ross Island

Now for something completely different -- a boulder made of "diamictite" (let's call it till) in a boulder train on James Ross Island, Antarctica.  It did not fall onto a glacier surface, and so it must have been excavated by over-riding ice in a deeply frozen state -- and then transported rather delicately by the glacier........

I have never seen this phenomenon before!  Further info here:

Glacier front geomorphology

Click to enlarge.  This is one of the most fabulous images I have ever seen of a receding glacier front.  (I don't like referring to a "retreating glacier" since glaciers do not turn round and go marching back up the valleys whence they came.......   They keep on moving forward, but when the rate of melting on the surface and at the snout exceeds the forward movement, then they have a problem.......)

There is no point in trying to describe everything visible in this image -- but there is a whole course in glacial geomorphology captured here.

I don't even known where this photo was taken -- but I suspect it might be somewhere in Patagonia.

Friday, 4 January 2019

Ceibwr -- Quaternary stratigraphy and related events (1)

Following recent visits to Ceibwr, I think I'm getting the story sorted out at long last.  It's not easy, since there is no single location where the full sequence of sediments can be seen.  There are more than a dozen exposures in and around the bay, so extrapolations are necessary.......

These are the key locations:

SN107457  --  roadside exposures of cemented glaciofluvial gravels

SN107457  --  gully with potholes (channel 3) near coast

SN108457  --  largest exposure of cemented gravels etc -- near tip of spur

SN110456  --  plug of Irish Sea Till and other deposits in mouth of valley (Channel 1)

SN109456  --  intake for Channel 2 (hump)

SN108456  --  meltwater erosion traces on rock wall of Channel 2

SN109456  --  hummock of fresh glaciofluvial gravels filling entrance to Channel 2

SN110455  --  riverside exposure -- cemented gravels beneath Irish Sea Till

Herewith a photo gallery with annotations in roughly the order of the list above. We concentrate here on the features in and around the Ceibwr Gully -- I'll deal with the other exposures in a later post.

Roadside exposure.  Approx 1m of cemented iron-stained glaciofluvial gravels with horizontal bedding, resting on vertical strata of Ordovician shales.  A classic unconformity........

Roadside -- cemented gravels.  Here they appear to be plastered against the side of the shale bedrock ourcrop.  The location is about 10m above the floor of Channel 2, which we see at the right edge of the photo.   It's most reasonable to propose that the gravels were emplaced by water flowing along an ice edge, at a time when the channel was filled by wasting ice.

A second exposure of horizontally bedded gravels, a little way uphill.  These are also solidly cemented, but the exposed surface is covered with mosses and lichens.

The small gully (Channel 3) with potholes, on the western flank of the exit of Channel 2 where it reaches the coast.  The potholes contain many large submerged erratics, including some that are far-travelled, of igneous origin.

Rock surface on the flank of the gully, showing classic pothole erosional features and surface staining by iron and manganese oxide.

This might just be the oldest Pleistocene deposit ever recorded in Wales.  It's a cemented slope breccia, high on the western flank of Channel 2, with the gully (Channel 3) directly below.  The gully may or may not be younger than the deposit -- more investigations are needed.  There is no stratification here -- this is a typical rockfall or slope collapse deposit, which probably has nothing to do with frost processes or periglacial conditions.  The cemented till and cemented gravels lie ABOVE this deposit.

In the same exposure, above the gully, we see TWO tills of vastly different ages.  The lower till is cemented, and is most reasonably interpreted as of Anglian age. Above that we see stratified glaciofluvial gravels, also cemented -- with several different facies.  Above that, uncemented grey-blue clay-rich till with many of the characteristics of Irish Sea Till -- probably of Devensian age. This is a hugely significant photograph -- I need to visit the site again in the company of other Pleistocene specialists.......

Remnants of cemented slope breccia with erratics (ancient till?), at the water line within one of the potholes. The gully might  have been cut -- or at least enlarged -- in the Anglian glaciation, before the deposition of the concreted till.

Close-up of the concreted till (Anglian?) on the slope above the gully.  It incorporates broken slabs of local bedrock, a wide range of erratics (including igneous rock types), and cobbles and pebbles of many different sizes.  Some of these rounded stones have been glaciofluvially transported, and some may even have been derived from raised beaches that are now destroyed.   The deposit is matrix-supported.  This might be a flow-till.

Above the glaciofluvial gravels, near the end of the trowel handle, there is a layer of clay-rich colluvium up to 20 cms thick, stained orange and red by pedological processes, and above that a "churned" layer of slope breccia.

The churned slope breccia deposit, made of shale fragments derived from a rock outcrop immediately upslope.  This deposit is uncemented.

The head of the gully is plugged with c 2m of cemented red-stained till and glaciofluvial gravels (seen here in the lower half of the photo) and grey-blue uncemented Irish Sea till with abundant erratic boulders above.  I'm pretty certain that these deposits represent two distinct glacial episodes.

Above the double till layers we see c 50 cm of sandy and silty colluvium / sandy loam which is elsewhere called Sandloess.  There are occasional erratics within it, probably redeposited from outcrops upslope.   This must have accumulated during the Holocene / post-glacial period.  The modern soil horizon is seen at the top of the section.

On piecing things together, the full Pleistocene stratigraphy here seems to be as follows:

Modern soil -- c 20 cms.  Uncemented
Sandloess and colluvium -- c 50 cms.  Uncemented.  Holocene?
Clay-rich Irish Sea Till -- up to 2 m thick.  Uncemented.  Devensian?
Brecciated slope deposits -- up to 50 cms thick.  Uncemented
Clay-rich colluvium -- c 20 cms thick.  Uncemented but stained / gleyed.  Ipswichian interglacial?
Glaciofluvial gravels -- c 1.5m thick.  Stained and cemented.  Anglian?
Stony till -- up to 1m thick.  Stained and cemented. Anglian?
Brecciated slope deposits -- up to 1 m thick

I need to consult with other geomorphologists on all of this, but I'm pretty sure that this set of deposits ties in rather neatly with the old till / new till situation described from Lydstep:

The cementation process is crucial to the interpretation of these deposits -- are the cemented deposits all very old, and are all of the uncemented deposits relatively young?

Next, I must go back to the Witches Cauldron and check out the extremely complicated cemented deposits there.  More clues will undoubtedly emerge.......

I'll do another post shortly on the other exposures at Ceibwr, which broadly confirm the sequence described above.