Past major changes in sea level have had significant implications for global and shelf sea tidal dynamics. Changing shelf sea tidal dynamics impact on tidal elevation amplitudes, the location of tidal mixing fronts, dissipation, shelf sea biogeochemistry and sediment transport. Some of these major changes are reflected in the geological records of shelf seas and therefore proxy data may be used to constrain tidal model outputs. This study explores a new geological proxy for northwest European shelf sea (NWESS) tidal dynamics by quantifying the relationship between modelled tidally-modulated bed shear stress (BSS) and observed seabed sediment grain size. A grain size tidal current proxy (GSTCP) has been developed by comparing tidal model output of BSS with observational data on present-day grain size data from the Irish Sea. This new proxy is shown to reproduce large-scale sediment distribution in the present-day Irish Sea, and the relationship is applied to the NWESS to predict sediment distribution across the shelf, over a range of (palaeo) time slices. A new three-dimensional palaeotidal model has been developed, which incorporates dynamic palaeotopography from the latest glacial-isostatic adjustment model for the region (Bradley, 2011). The tidal evolution of the NWESS is simulated using 1 ka time slices, from 21 ka BP (taken to be the approximate time of the Last Glacial Maximum) to present-day. Model outputs of BSS from the new palaeotidal model were compared with outputs from existing palaeotidal models, which were developed using different glacialisostatic adjustment models (Peltier, 1994; Lambeck, 1995). The new glacial-isostatic adjustment model produced significantly different relative sea level signals across the shelf, and hence there were differences in the timing of the major changes in modelled BSS between simulations. Sediment grain size evolution profiles were generated for five BGS UK shelf sediment cores, using laser particle diffractometry and radiocarbon dating techniques. Predictive grain size profiles were generated by applying the GSTCP to the modelled evolution of BSS at the sediment core locations, and were compared with the observed grain size profiles. In general, the GSTCP reproduced the observed trends in grain size variations in three of the four sediment cores, although tended to over-predict the grain size in all core locations. Despite the limitations of the GSTCP for reproducing observed sediment classifications at specific sites, the GSTCP was applied to the regional model output of BSS to generate predictive maps of seabed sediment types on the shelf, in 1 ka time slices from 21 ka BP to the present-day. Such maps of sediment distribution are useful for a number of applications, including for physical (e.g. morphodynamic) modelling and biological studies (e.g. habitat mapping). The new palaeotidal model was also used to estimate changes in the position of the shelf sea tidal mixing fronts with sea-level rise since 21 ka BP. Prior to this work, the only proxy data used to constrain palaeotidal model simulations was from one sediment core from the Celtic Deep. The timing of stratification at the core location had been used to validate a palaeotidal model; however, considerable variation in the timing of onset of stratification at the core locations considered here was predicted by different palaeotidal model simulations. Further, since the position of the tidal mixing front is very sensitive to the value used for the critical contour, it is suggested that using the timing of stratification at an isolated site is too sensitive a parameter for validating palaeotidal models. The GSTCP does not fully resolve the changes in observed grain size at the core locations, thus the proxy is considered unsuitable for constraining palaeotidal model output in this context. The main limitations are the lack of consideration of sediment availability and supply, the limited spatial extent of the geological data (i.e. few sediment cores) and the absence of wave-induced BSS. Future work should consider combined wave and current induced sediment transport, and the feedbacks between evolving morphodynamics and hydrodynamics, over a range of timescales. The new proxy can be used to approximate the large-scale sediment distribution over the NWESS for the present-day, and has been applied to palaeotidal model output to predict the evolution of large-scale patterns in sediment dynamics across the shelf over the last 21 ka. The predictive map of the large-scale distribution of seabed sediment grain size is useful for a number of applications, including for considering the marine aggregate resource, for marine habitat mapping, and for including spatially-varying bed roughnesses in hydrodynamic- and morphodynamic models.