The symbols on the map show designated moraines, lakes and other features which are mostly assigned to the Late Devensian. The yellow area is the area assumed to have been ice-free during the LGM. However, that is still a matter of debate, and I am now convinced that the designation of the ice-free enclave or corridor in Pembrokeshire is based on unreliable evidence. There are other questions about Gower and the Vale of Glamorgan.
Note that there was no synchroneity between the maximum LGM extent of ice on different parts of the contact zone. The BRITICE-CHRONO team thinks that the Irish Sea ice might have begun its retreat earlier on the western flank of the Welsh uplands than on the eastern flank.
Anyway, this will do for now -- let's assume that the line may be reasonably reliable to +/- 10 km or so..........
So what went on along the contact line between the Irish Sea Ice stream and the Welsh ice forced southwards in Cardigan Bay and across West Wales? There are analogies in the literature, relating to the ice streams as reconstructed for the Laurentide Ice Sheet. See this:
Ice streams in the Laurentide Ice Sheet: Identification, characteristics and comparison to modern ice sheets
Volume 143, April 2015, Pages 117-146
Martin Margold, Chris R. Stokes and Chris D.Clark
Ice stream activity scaled to ice sheet volume during Laurentide Ice Sheet deglaciation
C. R. Stokes, M. Margold, C. D. Clark & L. Tarasov
Nature 530,322–326(18 February 2016)
Utting, D., Atkinson, N., Pawley, S. et al. (1 more author) (2016) Reconstructing the confluence zone between Laurentide and Cordilleran ice sheets along the Rocky Mountain Foothills, southwest Alberta. Journal of Quaternary Science, 31 (7). pp. 769-787. ISSN 0267-8179
Palaeo-ice stream landsystem
By Jacob Bendle