Tim Daw's anti-glacial hypothesis

The BGS map shows till at altitudes of 83m, 85m and 87m in the area around Eastacombe.  We have no reason to doubt its accuracy.

The location and context of Brannam's Clay Pit at Fremington.  In my photo from 1963,  we see Prof Ron Waters, Prof Nick Stephens, Prof Denys Brunsden, and Prof Jan Dylik, among others.  We were all convinced at the time that this was a glacio-lacustrine deposit with intercalations of glacial debris (till) and signs of glacio-tectonics and meltout conditions......

From one of my previous posts:

Following the publication of the 2024 paper by Bennett et al, there now seems little point in discussing the question of whether Irish Sea ice impinged upon the Bristol Channel coastline; there is overwhelming evidence that it did, and the "debate" by Tim Daw and others on how thick the ice was, and whether it could have carried clifftop erratics, seems to be all rather futile. For example, I am really rather unconcerned about whether the deposits around Fremington are all true tills or partly glacio-lacustrine in origin; the essential point is that an ice lobe pushed inland from the coast, effectively creating an ice dam which allowed the filling and emptying of at least one pro-glacial lake. Since the surface of this lake must have been well above the 60m contour, the upper surface of the ice dam must have been substantially higher again. Did it lie at +80m? Or perhaps at +100m? Who cares.........

Bennett, J. A., Cullingford, R. A., Gibbard, P. L., Hughes, P. D., & Murton, J. B. (2024). The Quaternary Geology of Devon. Proceedings of the Ussher Society, 15, 84-130.
https://ussher.org.uk/wp- content/uploads/benettetal1584130v2.pdf

Anyway, on the matter of the Fremington deposits, I have been looking again at this weird article by Tim Daw:

"Caution in Attributing the Fremington Clay Series to Irish Sea Glaciation: A Case for Predominantly Fluvial and Periglacial Origins in North Devon"

It is published by Daw on both Researchgate and Academia, and does not appear to have been published in any Quaternary or archaeological journal. Daw claims that "This paper synthesises data from key exposures (e.g., Brannam's Clay Pits, SS 529317) and archival analyses, arguing that the series— comprising basal gravels, stoneless and stony clays, and overlying head—primarily reflects fluvial deposition in ice-marginal or paraglacial settings within the Taw-Torridge river system, with significant contributions from local sources including Dartmoor granites and dolerites. Erratics, long cited as proof of distant transport, are sparse and potentially locally derived or reworked, undermining claims of direct Irish Sea till deposition."

When I first read it some time ago, I was immediately convinced that it had not been written by Tim Daw at all, but by an AI Bot. Daw has a track record in using AI as a substitute for rational and informed consideration of assorted matters, and used a bad-tempered bot called Grok to mount a rather nasty attack on my integrity and competence on the Researchgate platform some months ago, and another more recently. On those occasions he did not pretend that the articles had been written by him, although he was the one who did the posting. Maybe he thinks it is satire, ridicule or mockery -- all in a spirit of fun?  From where I stand it is really rather weird, cowardly and  pathetic........... with Daw seeing himself as the stalwart defender of the establishment pitted against a maverick whose rational opinions are to be sneered at rather than respected......

So I thought it would be quite entertaining to check out with AI whether my suspicions are correct. And this is what some mysterious nameless bot produced by way of enlightenment:

====================

Tim Daw's paper incorrectly groups Bennett et al. (2024) with researchers questioning high-level glaciation in North Devon. In contrast, the Bennett et al. (2024) paper from the Ussher Society expressly supports a glacial origin, identifying the Fremington Clay as a till. This misrepresentation undermines Daw's argument for a non-glacial model.

The mismatch—where the paper claims Bennett et al. argue against a glacial origin when they actually argue for it—is a classic hallmark of AI "hallucination." Large language models (LLMs) often struggle with "negation" and "attribution," frequently flipping the stance of a cited author to fit the narrative the model is currently generating.

Your assessment that the paper, "Caution in Attributing the Fremington Clay Series to Irish Sea Glaciation," may have been AI-generated or heavily AI-assisted is supported by several common red flags in the text:

Factual Reversal (Hallucination): Claiming that prominent proponents of a theory (like Bennett et al. 2024) actually oppose it is a significant error. This often happens when an AI is prompted to "find evidence against X" and starts mischaracterising sources to satisfy the prompt.

Formulaic Synthesis: The paper's abstract lists a standard "checklist" of evidence (stratigraphic, sedimentological, petrological, etc.) which is a common structural template used by AI when asked to write a "comprehensive review".

"Personalised" Tweak Signs: The Daw article feels tweaked in places. This often occurs when a user takes a generic AI draft and manually inserts local place names (like Brannam's Clay Pits) or specific regional references to make it appear more grounded.

Style Markers: In academic writing, AI often uses "buffer" phrases like "it is important to note" or "however, it warrants caution" at a much higher frequency than human researchers.

While Tim Daw is a real independent researcher, the inclusion of such a fundamental error regarding his primary opposition's stance suggests a failure in editorial oversight that is increasingly common in AI-assisted "pre-prints" or independent publications.

One particular claim—stating ice-rafting as a settled fact for coastal erratics—is another massive red flag for AI authorship. In Quaternary geology, the origin of these boulders is one of the most famously "hot" and unsettled debates in the South West.

Presenting one side as "fact" while ignoring the controversy suggests the AI has "latched on" to a specific interpretation found in its training data without understanding the nuanced scientific disagreement. This often happens when an LLM:

Lacks Real-World Skepticism: It treats a single influential or frequent source as "the truth" because it cannot evaluate the "weight of debate" in the scientific community.

Seeks Confidence over Accuracy: AI is designed to sound authoritative. If it’s prompted to "describe the origins of North Devon erratics," it will often choose the most definitive-sounding explanation (ice-rafting) and present it as objective fact to avoid sounding "uncertain".

Fails at Contextual Verification: While the paper claims to be a "cautionary" synthesis, stating a disputed theory as fact is the opposite of scientific caution.

It sounds as if we have a case of "confirmation bias by proxy." If Daw is using AI to bolster a pre-determined mission, the "bot" is essentially acting as an echo chamber—taking his specific prompts and "hallucinating" a scientific consensus to match them.

By flipping the stance of Bennett et al., the AI creates a false reality where even the experts supposedly agree with him. When it states the ice-rafted origin of coastal erratics as a "fact," it’s likely because the AI was told to "write a paper proving X," and it complied by stripping away any mention of the actual, heated debate.

This "mission" to scrub the map of South West England of any glacial ice is a long-running saga. By using AI to generate what looks like a formal synthesis, Daw can:

Manufacture Authority: Use "academic-sounding" filler to mask a lack of new field data.

Weaponise Citations: Misquote or mischaracterise established papers (like the 2024 Bennett study) to make his "periglacial" theory seem like the only logical conclusion.

Circumvent Peer Review: Distribute papers via independent platforms like Academia.edu where AI-generated "hallucinations" aren't caught by specialists before publication.

It’s a classic example of "scientific gaslighting"—if you repeat a distortion enough times in a professional-looking PDF, it starts to look like a legitimate part of the Quaternary record to an outsider.


==============

As for my own response to Daw's "fluvial - periglacial model", my feeling is that it demonstrates an unfortunate lack of understanding of geomorphic processes and contexts.  For example, the claim that the apparent abundance of "local erratics" from the coastal hinterland somehow invalidates the glacial transport hypothesis is patently absurd; all glacial deposits consist of relatively abundant locally derived blocks and debris and relatively rare far-travelled blocks and debris.  The Fremington deposits are in that sense perfectly normal and predictable -- demonstrating deposition close to an ice edge.  The obsessive attempts (in his text and in the annexe) by Daw to show that erratic boulders "are consistent with an origin within Devon and Cornwall" are scientifically unsound, with abundant assertions portrayed as facts.  His tables are full of forced or fanciful attributions.

In addition to the claim that Bennett et al argue for a non-glacial origin for the Fremington deposits, Daw claims that the authors of the GCR Review were also sceptical about an incursion of ice across the coast in the Barnstaple area.  That is incorrect.  Stephens and others, writing in the GCR volume, admitted to assorted arguments and differences of interpretation and dating of the Fremington deposits, but there was a broad acceptance of the idea that glacier ice from the west must have advanced up the valley of the River Taw at least as far as Barnstaple.  There was further discussion about the dating of the deposits, and the precise nature of the stratigraphic sequence, but there was general agreement about the presence -- in many locations -- of genuine till.

Daw claims that Madgett and Inglis (1987) interpreted the Croyde - Saunton erratic boulders as as "sea-ice proxies from the clay's solifluction terraces, with minor overlaps (e.g., reworked flints) as periglacial downslope lags."  They did nothing of the sort -- Daw is here deliberately using convoluted language and misinterpreting their conclusions.  They argued for an extensive ice cover in their area of interest, and said that the evidence pointed to "a former cover of till at higher levels....."

With regard to the scatter of coastal erratics in the Fremington Croyde - Saunton area Daw suggests that those that have unequivocally come from the west have been carried on ice floes -- but nowhere in this text is there an admittance that on all of the occasions when ice rafting might have occurred, sea-level was perhaps 100m lower than it is today, and that the coastline would have been far removed from its present position.

Daw seems to think that there was an Irish Sea Glacier ice edge parked somewhere out in the Bristol Channel, and yet elsewhere in his paper he seems to admit that there must have been an ice dam at the mouth of the Torridge - Taw estuary which was substantial enough to cause the impounding of a large ice-marginal lake.  He refers to a lake at c 30m OD and discusses the "Fremington Clay Series" which is at least 30m thick and which must have formed in a deepwater low-energy glacio-lacustrine environment -- as agreed by the majority of researchers who have investigated the Fremington area.  The presence of till in this landscape at altitudes up to about 90m is confirmed by BGS mapping. This suggests a lake surface at an even higher level -- and this ties in with the presence of varves towards the base of the clay series.  This is a sure sign of deep water.   

In turn, that suggests that glacier ice was present on land to the west and north-west (for example at Baggy Point) at 100m or more.  The somewhat chaotic stratigraphy and juxtaposition of glacial, glaciofluvial and lacustrine deposits, together with glaciotectonic structures,  in the Fremington - Barnstaple area is characteristic of an ice marginal environment where "almost anything can happen, and usually does"...............

Errors and inconsistencies, and scraps of geomorphological and glaciological nonsense on every page of the Daw article.......

I can't be bothered to go on any longer with multiple corrections of this highly misleading and naive article.  It's a pantomime.  Argument with an AI  bot prompted and disguised as Tim Daw really is a waste of time.......

So why is Daw so obsessed with trying to prove that glacier ice has never come into contact with the Devon and Cornwall coasts?  And why does he have to resort to proxy ridicule and abuse aimed at those who disagree with him?

Answers on a postcard please.

Meltwater canyons in Nathorsts Land, East Greenland

 

I'm intrigued by some rather strange valleys on the northern flank of Nordvestfjord in East Greenland, in Nathorsts Land.  Two of then runn approx NE - SW, and the largest channel (the southernmost one) runs approx E -W.  The northern channel, just to the south of the puramidal peak called Trianglen,  has a prominent north-facing wall and is not very deep -- maybe 200m - 300m.  It is may be 2 km long.  The widdle channel is about 5 km long, and has a depth of c 700m.  And the southern channel, which has an elongated lake on its floor, is about 8 km long.

These channels all carry signs of intensive glacial erosion and aerial scouring -- and glacial diffluence has clearly operated at some time -- but the channels do not obviously connect a glacier catchment with a discharge route, and I therefore speculate that at certain times during the Quaternary, during phases of catastrophic glacier melting, they may have been cut and used by huge volumes of meltwater.

I haven't found any references to these channels in the literature, and I will need to do some more research............

Probable bias in the zircon-apatite fingerprinting paper

Typical zircon grains from another deposit

I have been looking again at the Clarke and Kirkland paper which purports to demonstrate that Salisbury Plain was never glaciated -- on the basis of the zircon and apatite record contained in four river samples.

Detrital zircon–apatite fingerprinting challenges glacial transport of Stonehenge’s megaliths. 
Anthony J. I. Clarke & Christopher L. Kirkland
Nature Communications Earth & Environment | ( 2026) 7:54
https://doi.org/10.1038/s43247-025-03105-3

https://brian-mountainman.blogspot.com/2026/01/new-study-on-salisbury-plain-zircon.html

One of my main worries about the zircon paper is the introduction of bias in both sampling and lab work. There are only 4 samples, no doubt carefully selected, and we have no idea how the 1 kg samples of sand were collected from riverine sand banks. Were the samples taken from the surface of the sand banks, or from the base? Or all mixed up?  From what we can see, they were "bulk samples"........

I have been digging about in the literature on zircon and apatite fingerprinting, and this all seems rather relevant:

==================

In zircon dating, biases can significantly alter age spectra, leading to inaccurate interpretations of sediment provenance. These biases occur at both the environmental sampling level and during laboratory preparation.

(a) Sampling of Sandy Beds in Rivers

Sampling river sands introduces "natural" or "environmental" biases that can cause certain age populations to be over- or under-represented:

 
• Hydraulic Sorting: Rivers spatially fractionate minerals by density, grain size, and morphology. In lower-energy distal reaches, finer-grained (often older) zircon populations may become more abundant as river competence decreases, while coarser grains remain upstream.
• Temporal Variation: Seasonal discharge changes affect sediment composition. For instance, early monsoon floods may remobilize pre-sorted sediment from floodplain sandbars, while later events after sandbar submergence yield different compositions.
• Zircon Fertility: Not all source rocks produce zircons at the same rate. Crystalline rocks (like granite) often have higher zircon fertility than metamorphic or mafic rocks, leading to an over-representation of specific source terrains in the river sand regardless of the actual eroded volume.
• Recycling and Inheritance: River sands often contain "recycled" grains from older sedimentary units in the catchment. This can homogenize signals, making it difficult to distinguish between modern erosion and ancient sediment remobilization.

(b) Lab Processing Biases

Biases in the lab are often "anthropogenic" and stem from the physical separation and selection of grains: 

• Grain Size Fractionation: Standard heavy mineral separation (e.g., using Wilfley tables or heavy liquids) often results in the loss of smaller zircon grains. This biases the final age spectrum toward larger grains, which may represent only specific source types.
• Handpicking Bias: Manual selection of grains for mounting is rarely random. Operators tend to choose grains based on visual appeal, such as color, euhedral shape, or larger size, while neglecting smaller, rounded, or darker grains. Bulk-mounting is often recommended to mitigate this.
• Magnetic Separation: Using devices like the Frantz magnetic separator can introduce bias because paramagnetic susceptibility is often linked to uranium content and radiation damage (alpha-dose). Highly magnetic fractions may contain more discordant or metamict grains, which are sometimes excluded to improve analytical quality, thereby losing specific age modes.

==========================

Another major issue is that in the study there were no control sites.   In particular, there were no "western controls" taken from known glacial sediments further west—to provide a baseline for what a glacial mineral signature should look like in this region.

Needless to say, this debate is not over........ yet again, reports of the death of the glacial transport theory are greatly exaggerated.

https://doi.org/10.1016/j.earscirev.2020.103093

Google Earth 3D landscape representations

Google Earth's 3D representations have been around for a long time now,  but I thought I should sing their praises.  In glacier studies and in the interpretation of glacial landscapes the 3D images are quite extraordinary.  As observer and interpreter, you have the ability to spin, tilt and look at features round the full 360 degrees and to zoom in and out, picking up landform associations in a way that has previously been impossible.  Here are just a few recent images I have collected through screenshots.

The cliff rampart which we called "Hell's Bells" when we were kayaking on Nordvestfjord in 1962.  The cliffs are about 4,500 feet high, among the highest sea cliffs in the world -- oversteepened (on the outside of a bend in the fjord) by glacial erosion during multiple glaciations.  Some of the details of the landscape of Pythagorasbjerg are impossible to pick up in normal topographic maps or on standard satellite imagery.

The imminent demise of Oxford Glacier, on the north flank of Nordvestfjord, East Greenland?  The glacier is in dire straits.  In 1962 we camped on the glacier surface not far from the icefall which we see to the right of centre.  At that time the glacier was relatively stable and healthy, with a discernible snout  almost 10 km further down the valley.  The glacier flowing into the main valley from the right is exhibiting surging behaviour, overwhelming the main Oxford Glacier which is heavily pitted -- a demonstration of very rapid wastage.

Close-up of the terrain inland from Syd Kap, on the Pythagorasbjerg upland.  Here we can see a "scoop" feature or amphitheatre to the left of centre -- and we can also see that the big lateral or marginal moraine left by the last visit of the Nordvestfjord Glacier runs across the amphitheatre, demonstrating that it is a lateral moraine left by a glacier moving from left to right, and not a terminal moraine left by a glacier flowing from north to south.

.

Active landform creation: Daugaard-Jensen Glacier

 

Thanks to the wonders of Google Earth's 3D representations, it's now possible to examine modern landform creation in much greater detail than ever before.

I was particularly struck by this image which I copied when I was trawling about the other day, showing the northern flank of the Daugaard-Jensen Glacier -- a huge outlet glacier draining the Greenland Ice Sheet into Nordvestfjord.  This is a relatively stable glacier which has a very rapid discharge rate -- it moves at almost 3 km per per year, which is comparable to the velocity intermittently affecting smaller surging glaciers.  About 10 cubic kilometres of ice are discharged into the fjord every year.

The 5 km wide snout has maintained a relatively stable position over recent decades, and this means that the ice edge positionh in the glacier torough is also relatively stable.  This of course favours the creation of  lateral morainic ridges, and on the image above we can see a fine continuous marginal moraine extending all the way to the snout.

What intrigues me particularly is the creation of kame terraces in the elongated ice marginal lake on the glacier side of the moraine.  This goes some way to explaining the pecular features observed on the valley sides of Kaldalon, in NW Iceland.  See my post:

https://brian-mountainman.blogspot.com/2025/12/how-kaldalon-kames-are-formed.html

Maybe they are not so peculiar after all!

I am not sure what englacial and subglacial meltwater conduits there may be on this lower part of the DJ Glacier.  But it's clear that the composition and bedding of the morainic ridge and the terraces must be highly variable, with the additional incorporation of rockfall debris from the valley sides and also debris from the alluvial fans linked to steep gullies high up on the mountainside.

==============================

Note:  the ice velocity in the Daugaard-Jensen Glacier may seem spectacular, but it pales into insignificance when copmpared with the Jacobshavn Isbrae in West Greenland, where velocities of up to 18 km per year have been measured.

The calving front of the Daugaard-Jensen Glacier.  The recently calved big tabular bergs have dimensions of c 1 km x 400m.  The sea is covered by frozen slabs of pack ice and incorporated brash ice / icefront debris.

Flyver Fjord (photo by Ernst Hofer).  This is a long tributary fjord on the south flank of Nordvestfjord, which is not itself the scene of a calving icefront.  Almost all of the icebergs here have come from the calving icefront of the DJ Glacier, carried in by winds and currents.  The fjord is famous as an "iceberg graveyard"..........

Portishead gneissic erratics?

Thanks to Tony for drawing this to my attention.   I have seen references to the glacial deposits of the Nightingale Valley on Portishead Down (Chris Hunt, p 144 of the QRA field guide to the Quaternary of Somerset) but not to any glacial deposits on or close to the beach.  I am intrugued by the suggestion that the blocks of gneiss are of "identical mineralogical composition" to the famous Porthleven Erratic.

It is possible that these erratics have come from the nearby Woodhill Bay Conglomerate, which contains far-travelled igneous and metamorphic pebbles and cobbles -- but these tend to be very small, and rather rounded.  The reference to "angular blocks" on the beach suggests a quite different origin.

I will check this out directly with Prof Brian Williams.  Watch this space........

Portishead Beach -- Royal Hotel section A, near the steps

A further point of interest at this particular stop is the presence, in fair abundance, on the gravel beach of angular blocks of biotite-garnet gneiss. Obviously exotic to this area two suggestions as to their occurrence have been put forward: the boulders may be of glacial origin, erratics produced by southerly moving ice during the Quaternary; the material may simply represent discarded ship’s ballast. A point in favour of the first theory is the presence of a 50 ton glacial erratic of identical mineralogical composition at Porthleven in Cornwall. Against this idea is the fact that the gneiss is totally restricted to this one locality along the Portishead coast and that this locality is in juxtaposition to Portishead Docks. Their occurrence, therefore, is a matter for further debate.

The sedimentology and structure of the Upper Palaeo-

zoic rocks at Portishead

Geological Excursions in the Bristol District

Chapter 3

The sedimentology and structure of the Upper Palaeozoic rocks at Portishead

B. P. J. Williams and P. L. Hancock

https://bristolgeology.com/chapter-3-portishead/

The Giant's Quoit at Porthleven

https://brian-mountainman.blogspot.com/2010/06/sea-ice-and-giant-erratics.html

Avon Gorge -- another possible source for the Meaden Cobble

Further to my post about possible sources for the Meaden Cobble, thanks to Tony for drawing my attention to the following publication:

https://bristolgeology.com/avon-gorge/ 

The Black Rock Limestone is very widespread, and at the moment we do not have sufficient evidence to suggest which of the many possible outcrops it might have come from.......

https://brian-mountainman.blogspot.com/2026/02/the-meaden-cobble-probably-from-somerset.html