Was there a Devensian calving bay in the Celtic Sea?

The widely cited reconstruction of the Devensian Irish Sea Glacier edge in the Celtic Sea.  It seems to break all the rules of glaciology -- so how reliable is it?  It seems to have been "invented" so as to explain the presence of Devensian till on the north coast of the Scilly Islands.

A representation of the Celtic Deep in St George's Channel, between SE Ireland and Pembrokeshire.  If Devensian ice scoured out this channel, or at least flowed through it and along it, it must have maintained a substantial surface gradient on its long profile.  That argues for quite thick ice in St George's Channel, if the ice really did extend far to the SW of the Scilly Isles.

The elongated sand ridges in the Celtic Sea.  Note also the position of the -130m contour, thought to represent the maximum sea level lowering during the Devensian (Last Glacial) Maximum of ice extent.  After Scourse et al.

Celtic Sea -- generalised -100m and -130m contours.  The map also shows the Celtic Deep and the zone of elongated sand ridges.

As I have noted on this site before, I have a major problem with this interpretation of  great lobe pressing far out to the south-west, since nothing like it exists in the glaciated regions of today.  Intuitively, one feels that in a deep embayment like this, with anchoring peninsulas in S Ireland and Cornwall, there must have been a concave ice front or calving embayment, if thr ice edge was close to the grounding line.  If the grounding line was far to the SW, and if the whole area occupied by the Irish Sea glacier was dry land at the time, then a lobe is a little more plausible.  Just a little......

So where was the grounding line?  I believe it was somewhere near the -90m contour, since the thick sediments (often 20m or more) give us an artificially shallow sea today, and since there must have been some isostatic depression at the peak of the glacial maximum in this region.  Even 20m of isostatic depression would have been sufficient to push the grounding line well to the NW of the present -130m contour. 

The proposed Celtic Sea calving embayment at the peak of the Devensian Glaciation, c 23,000 years ago.  The ice edge would have been approx at the grounding line at a time of isostatic depression in the Celtic Sea.

This is glaciologically much more likely than the strange glacier lobe proposed by Clark and others, and it does allow for Devensian ice to reach the north coast of the Scilly Isles.  However, this will only have been possible if there was much more ice in the eastern part of the Celtic Sea and in the approaches to the Bristol Channel than shown by Clark and others.  Lundy Island and Caldey Island become quite critical in our understanding of where this eastern limit might have been.  We also overcome the "Celtic Deep problem" and explain why Pembrokeshire was apparently not completely inundated by Irish Sea Ice.  All of the ground evidence shows that the Irish Sea Glacier ice edge crossed the St Davids Peninsula, pressed up against the flanks of Mynydd Preseli, and affected the coasts of western Pembrokeshire without pressing far inland.  This is all completely consistent with a calving bay in the Celtic Sea.

That strange ice lobe needs to be abandoned, since it accords neither with glaciological principles nor with the field evidence.  Instead, we need to move the SW ice margin northwards by maybe 200 km and eastwards by maybe 100 km.  I rest my case.  Come forth, all you geomorphologists, and argue with me.......