Scottish Journal of Geology Investigation of coastal environmental change at Ruddons Point, Fife, southeast Scotland

: Ruddons Point, on the Firth of Forth coastline, Scotland, is a laterally extensive terrace of glacial and marine sediment deposits raised above current sea level, situated near to Kincraig Point, a key site that records a series of stepped erosional platforms carved into the local bedrock, interpreted as post Last Glacial Maximum paleoshorelines. The deposits at Ruddons Point continue inland, with exposures of the raised sands and gravels cut by the local river, the Cocklemill Burn. The site provides an opportunity to examine the depositional history through the Late Devensian and Holocene. Geophysical survey aided in interpreting characteristics of subsurface sediments such as the transition between the younger saltmarsh sediments and older underlying sands and clays below, which slope in a northerly direction. A chronology obtained through OSL dating spans from ~29 ka for sands and clays at an elevation of -0.66 mOD to surface windblown sands at < 300 yrs, at an elevation of 8.45 mOD. A basal peat, dated by radiocarbon dating to the early Holocene at ~9.2 k , Cytheropteron spp. , Palmoconcha guttata , Paracytheridea cuneiformis , Robertsonites , Sarsicytheridea bradii , Semicytherura undata , Finmarchinella angulate , Finmarchinella finmarchica , Hemicytherura clathrate and marine/inner


Investigation of coastal environmental change at Ruddons Point, Fife, southeast Scotland Study Area
The field study location centres on Ruddons Point, situated in south Fife, on the northern coastline of the Firth of Forth (Fig. 2). The Ruddons Point promontory separates the beaches of Shell Bay to the south and Largo Bay to the north. The terrace lies approximately 1 km to the northwest of Kincraig Point terraces which are themselves a key type example of eroded shorelines where four clear platforms are cut into the local, igneous bedrock (Fig. 3). Previous researchers Smith et al., 1969;Jardine, 1982) have commented on the potential age and correlation of the Kincraig Point erosional terraces as being related to the Main Perth Shoreline, a shoreline thought to have been formed during the Perth Re-advance circa 17.5-14.5 ka (McCabe et al. 2007), and Late Glacial Shorelines, however no absolute dating has been undertaken to confirm this. The Ruddons Point sand, shell and gravel deposits sit above a unit of glacial till and underlying clays which themselves have been deposited unconformably on Carboniferous-Permian age basaltic intrusive sequences (Chapman, 1976). These intrusives sit within the wider bedrock geology of the study site which is composed of Carboniferous limestone and coal (Forsyth and Chisholm, 1977). These Ruddons Point deposits were noted in as early as 1835, when W. J. Hamilton remarked on their position, suggesting a "subsequent elevating movement has taken place" for them to be raised above sea level. The raised deposits have an uneven surface due to erosion, wind-blown sand A C C E P T E D M A N U S C R I P T accumulation and vegetation cover but reach approximately 6.7 mOD (metres above Ordnance Datum) at the coast and up to ~9.5 mOD further inland. Mean high water spring (MHWS) tides across the field site reaches ~2.7 mOD, based on tidal records from the closest permanent tide gauge located at the Port of Leith (The National Oceanography Centre, 2021). The deposits of Ruddons Point outcropping to the south of Largo Bay, and the underlying glacial till and clay have not previously been dated, however research has been undertaken further inland along the Cocklemill Burn by Tooley and Smith (2005), and through an undergraduate dissertation by Overshott (2004), the results from which were presented in a field guide by Dawson and Dawson (2007). Tooley and Smith (2005) used radiocarbon dating on a sequence of the sediments, whereas Overshott (2004) focused on OSL. Both studies identified horizons which could be associated with the Main Postglacial Shoreline and Blairdrummond Shoreline, although the radiocarbon dates obtained by Tooley and Smith (2005) were generally older than the calculated OSL ages at similar elevations. Figure 3 shows the sampling locations visited in the current study. This study aims to examine coastal change at five locations across the field site to expand upon the previous research conducted along the Cocklemill Burn and to further current understanding of the ice extent history and sea level changes in this region of Scotland since the LGM. The raised deposits exposed on the coast at Ruddons Point (Locality 1) have not previously been dated. The raised terrace bank extends inland where deposits are exposed in section by a small river, the Cocklemill Burn, running through the field site. Deposits in this area, ~0.5 km east of Locality 1, have previously been dated through Tooley and Smith"s (2005) and Overshott"s (2004) studies, however with differing results, and therefore these raised deposits were also investigated in this study, differentiated from the more coastal Ruddons Point section with the name Cocklemill Burn section and cores (Localities 4 and 5). The previous field studies by Overshott (2004) and Tooley and Smith (2005) were undertaken in the vicinity of Locality 5. Further coring was undertaken in two more sites within the study area, through clays close to the mouth of the Cocklemill Burn where it meets Largo Bay beach (Coastal Core, Locality 2) and in the saltmarsh adjacent to the raised deposits (Saltmarsh Core, Locality 3).

Methodology Field Methods
The field site preserves a complex depositional history with key features including: (1) ample examples of raised deposits composed of sand, shell and gravel horizons; (2) exposed glacial till and clay at the shore and exposed at the mouth of the Cocklemill river and (3) preserved buried stratigraphy, including a basal peat, beneath the saltmarsh. To investigate each of these aspects of the field site, a multidisciplinary study was undertaken utilizing initial geological descriptions and geophysical investigation, followed by sampling for OSL profiling and dating, radiocarbon dating and complimentary microfossil analyses, with the overarching aim of building a depositional interpretation and chronometric framework for the field site.
Geographic coordinates for each site location, as well as the elevation of each raised marine deposit section and the ground surface of each core location were keyed into Ordnance Survey (OS) elevations using a Trimble dGNSS, positioning system with a R8 Base and R6 Rover affixed to a 2 m antenna. Stratigraphic changes and sample locations were recorded relative to both its depth within the section or core, but also relative to Ordnance Datum Newlyn (mOD).
Detailed sedimentological descriptions were carried out on cleaned vertical sections through the deposits. Using a Wink vibracore drill system, four additional subsurface cores were collected to extend the sedimentological description into the subsurface and to allow for additional sampling. Magnetic susceptibility readings were taken at ~5 cm intervals through 7.5 m of the Cocklemill Burn section using a handheld SM-30 magnetic susceptibility meter (ZH Instruments).
An electromagnetic (EM) ground conductivity survey was undertaken across the foreshore and saltmarsh adjacent to the raised deposits to improve understanding of the subsurface stratigraphy. The equipment used included the EM conductivity meter (CMD Explorer, GF Instruments) integrated with the Trimble dGNSS positioning system. The CMD Explorer was operated with coils in the vertical orientation and recording was set to continuous GPS measurement with ~10 m line spacing between lines. The apparatus measures ground conductivity over three different depth ranges through the subsurface (C1 being the shallowest at approximate depth down to ~2 m, C2 to ~4 m and C3 the deepest to ~6 m). To complement the EM ground conductivity survey, a direct current (DC) resistivity tomography line was acquired in a transect across the centre of the saltmarsh, using an ABEM SAS4000 Terrameter. The DC Resistivity method penetrates deeper into the subsurface than the EM conductivity survey with a maximum survey range of approximately 12 m. Metal probes were positioned with 1 m spacing along a transect running broadly N-S through the site. The equipment automatically ran a series of iterations of potential resistivity between the probes to create a 2-D electrical tomographic pseudo-section of the subsurface.

OSL sampling in the field
Sections 1 and 4, Ruddons Point and Cocklemill Burn, respectively were profiled and sampled for OSL. Sections were first cut and cleaned by hand shovel, then covered with opaque black cover, with final cleaning taking place under this cover. For initial characterization, bulk sediment samples were taken at a high spatial sampling interval (5 to 15 cm between each sample) through each profile. These samples were subjected to rapid luminescence measurements using a SUERC portable OSL reader (Munyikwa et al., 2021). At Cocklemill Burn, additional samples were collected in small copper tubes for more targeted laboratory analysis. At Ruddons point, samples for dating were collected by inserting steel tubes sealed with black tape, with corresponding dosimetry material for radionuclide concentration measurements. Infield gamma spectrometry was obtained at the Ruddons Point and Cocklemill Burn sections using a handheld Gamma Surveyor Vario (GF Instruments). Dating samples and associated dosimetry material were collected from the drilled cores under light-safe laboratory conditions.

Laboratory characterisation
Simple laboratory screening of samples to assess luminescence sensitivities and approximate stored doses, then the subsequent full OSL dating were carried out in the CERSA Luminescence Laboratory at the School of Earth and Environmental Sciences, University of St Andrews. Samples were opened in subdued red-light conditions and wet-sieved to isolate the 90-250 μm fraction. For initial laboratory screening, each sample underwent a hydrochloric acid (HCl) treatment for 10 minutes to remove carbonates, followed by a 40% hydrofluoric acid (HF) etch for 40 minutes to remove the outer surface of the grains, before a final HCl treatment for 10 minutes to remove fluoride precipitates. The grains were then dispensed to 10 mm diameter as ~2mm monolayer (small aliquot), greased, steel discs for loading into Risø TL-OSL DA-20 reader for analysis. For dating samples, grains underwent a heavy liquid (lithium polytungstate, LST) density separation to preferentially concentrate quartz grains, in addition to treatment with HCl, HF and HCl. The samples were re-sieved at 150 μm, and quartz dispensed in the 150-250 μm grain fraction.
A single aliquot regeneration (SAR) dose protocol was used to estimate equivalent doses (D e ) (Murray and Wintle, 2000), with the protocol modified for CERSA423-427 to include a hot bleach at the end of the natural and regenerative dose cycles. Aliquots returning unacceptably high recuperation % or recycling ratios were rejected from further analysis. The central age model (CAM) (Galbraith et al. 1999) was used to calculate final D e determinations for all samples.
Environmental dose rates for all dating samples were calculated from radionuclide concentrations (U, Th, K, Rb) measured by inductively coupled plasma mass spectrometry (ICP-MS), carried out within the University of St Andrews StAIG lab (ICP-MS) and Actlabs, Canada. Additionally, for the Ruddons Point section, high-resolution gamma spectrometry (HRGS) measurements were made on bulk sediment samples at the University of Stirling Environmental Radioactivity laboratories.

Radiocarbon Analyses
A total of 10 radiocarbon dates were collected for analysis by accelerator mass spectrometry (AMS) at key horizons across the raised marine deposits and selected cores. All radiocarbon ages are presented in Table 2. in both conventional yrs. BP and calibrated yrs. BP. Calibrations have been carried out in the Calib Rev 8.1.0, using either the IntCal20 (Stuiver et al., 2021) or Marine20 (Heaton et al. 2020) calibration curves, and stated with one and two σ error ranges. All shells and shell fragments were calibrated using Marine20 calibration with a local ΔR of -141 ± 57, obtained from http://calib.org/marine/ (based on Harkness, 1983) to compensate for the marine reservoir effect (MRE). Both humin and humic acid fractions of the peat sample were dated and calibrated using IntCal20. Five shell samples were submitted to BetaAnalytic, Miami, three shell samples were submitted to QUB CHRONO and one peat sample was submitted to SUERC for dating of both the humin and humic acid fractions.

Microfossil Analysis
Microfossil analysis was undertaken on 19 samples from a 2.2 m interval through the raised sands of Ruddons Point (Locality 1). Samples were broken into very small pieces by hand and put in a ceramic bowl to be thoroughly dried in an oven. They were then soaked in hot water with a spoonful of sodium carbonate added to help remove the clay fraction and facilitate release of the organic component. After soaking overnight, each sample was washed through a 75 micron sieve with hot water, the residue being decanted back into the bowl and dried. Because of the organic content of most of the clays/silts, this process had often to be repeated twice to achieve a good, clean breakdown. After breakdown and drying, the samples were stored in plastic bags and later picked of their microfaunal content under a binocular microscope. A selection of microfossils were picked out into faunal slides and recorded; the organic remains are recorded on a presence (x)/absence basis but the abundance of each ostracod and foraminiferal species is recorded semi-quantitatively (present/several specimens, common, etc.).

Electrical and Electromagnetic Survey
The extent of the EM conductivity survey across the field area is presented in Fig. 4. The surface represents results from the deepest penetration to depths of ~ 6 m into the subsurface, however the conductivity patterns show similar patterns at all three depths. Areas of higher conductivity (~300-400 mS/m) are identified in the central saltmarsh, and lower conductivity (between ~0-250 mS/m) along the beach front and close to the edge of the raised terraces. The conductivity signature also appears to reduce as you move inland to the east. The conductive high in the central saltmarsh could be indicative of elevated levels of saline water within the saturated saltmarsh. Therefore, the survey results may be showing a gradual decrease in salt water as you move inland from the coast. The DC Resistivity survey, running broadly S-N through the central saltmarsh (Fig. 5.) shows that resistivity generally increases with depth and identifies a potential slope of the underlying sediments in a northerly direction, with the northerly slope most prominent in the resistivity bands ranging from ~2.3 to ~25.9 ohm-m, represented in green, yellow, gold and orange). The upper dark and light blue bands (~0.2 to 1.0 ohm-m) have no discernible slope, which suggests these represent saltmarsh sediments infilling over the underlying tilted sediments. The transition from saltmarsh silts into underlying sands and clays can also be seen in the core at Locality 3 (Fig. 11). The upper green resistivity bands (~2.3 ohm-m) outcrop at surface level at the southern end of the transect, which suggests that this could be representative of the diamict and clays which also outcrop at the coast and therefore may continue beneath the saltmarsh and dip in a northerly direction. The lowest red band (resistivity of >25.9 ohm-m) does not exhibit a tilt towards the north and therefore may represent the underlying bedrock geology. The DC resistivity transect also reveals the presence of a depression in the resistivity layers which is being interpreted as an infilled paleo river channel running through the field site.

Locality 1: Ruddons Point raised marine deposits
Description. The Ruddons Point terrace has a large cut bank exposure along the modern shore face to the south of Largo Bay (Fig. 6). At the base of the exposed raised deposits is a glacial till, which extends below the modern beach level. The soft, fine-grained, grey clay matrix contains angular clasts of various lithologies. The till unit appears to pinch out towards the NE as you move laterally along the exposure. Deposited on top of the till is >3 m of sand and gravel and cobble units (~3.57-6.77 mOD). The lowest ~0.60 m of sand is dark brown in colour, fine grained and organic-rich, with evidence of burrowing and bioturbation. There is also evidence of thin (~1-2 cm thick) gravel horizons within the sand. The darker sand transitions at ~4.17 mOD to a lighter beige sand unit (~0.70 m in thickness) with less evidence of disturbance. The upper 0.12-0.15 m of this unit is laminated. A sharp contact separates these sands from the overlying unit of cobble-sized clasts, which is ~0.45 m thick, although the thickness is variable along strike in the exposure. This gravel and cobble horizon is supported by a coarse sand matrix that contains heavily weathered molluscs. Individual clasts are sub-rounded, the largest being ~18-20 cm in the longest axis and are comprised predominantly of igneous and metamorphic lithologies such as basalt and quartzite. Overlying this unit is a light beige fine sand, ~0.35 m in thickness, and at ~5.67 m OD, the sand transitions to a darker beige, ~0.25 m in thickness. Another sharp contact separated this sand from an overlying gravel unit. The average clast size is smaller than the lower unit, which contained both gravel and cobble-sized clasts. The unit has an abundance of marine shells, including intact limpet, periwinkle, cockle and razor clam shells. In some areas of the exposure, there appears to be an imbrication in the gravel and cobbles in an easterly direction. Directly above this gravel, at ~6.20 m is a shell supported unit, ~0.10 m in thickness, containing numerous whole, unbroken shells, predominantly gastropods The upper most stratum of the section is a cap of modern sandy soil until the surface at 6.77 mOD, with prevalent root systems throughout, which is variable in thickness.
OSL chronology. Initial in-field screening of the sands suggests a slight increase in net OSL and IRSL intensities with depth through the section, with the exception of the lowest part that consists of darker, bioturbated sands. The darker sands show an initial inversion to lower signal intensities between ~4.20-3.80 mOD. Below this, intensities again increase with depth, through the lowest ~0.25 m of the section, between ~3.80-3.55 mOD. OSL and IRSL depletion return the lowest values in the sands directly below the overlying gravel units as well as in the darker sands at the base of the section. The ratio of IRSL: OSL is also variable through the section, ranging from ~0.10 in the darker sands of the lowest 0.60 m of the section to ~0.19 in the sands between the gravel units ( Fig. 7). Subsequent laboratory analyses suggest a less complex depositional sequence, with a slight increase in stored dose with depth; this suggests that the reduction in OSL and IRSL intensities through 4.20-3.80 mOD above, is partly influenced by grain colour. The paired stored dose estimates are often divergent, suggesting for the most prominent of these, that the sediment was poorly bleached at deposition.
All OSL ages are presented in Table 1. Five OSL dating samples were collected through the section, in each of the sand units. The sample highest in the sequence yielded an OSL date of 3.57 ± 0.88 ka (at 5.77 mOD), with the underlying sample dated to 2.75 ± 0.78 ka (at 5.44 mOD). Below the lower gravel the three OSL samples also show an age inversion with depth, dated to 3.87 ± 0.77 ka (4.78 mOD), 2.75 ± 0.59 ka (4.45 mOD) and 2.06 ± 0.49 ka (4.17 mOD), with the lowest sample taken just above the transition from the beige sands to the darker sands below. The error ranges of the upper four measurements overlap and have a weighted combined age of 3.14 ± 0.37 ka. The lowest sample in the section is potentially underestimating the depositional age.
Radiocarbon chronology. Through the Ruddons Point raised marine deposits, five marine shell samples were radiocarbon dated; two from the upper gravel, one from the lower gravel, and a further two shells from the base of the section, above the glacial till (conventional and calibrated dates presented in Table  2). Dated shells from the upper gravel returned comparable ages 2738-3140 cal BP and 2755-3171 cal BP (all dates reported with 2σ age range). These dates would suggest deposition of the gravel shortly deposition of the lower sands (dated above to 3.14 ± 0.37 ka). From the lower gravel, a bivalve dated to 7447-7800 cal BP. This is at odds with the younger OSL ages surrounding this gravel horizon. This might suggest the shell was redeposited during the high-energy deposition of the gravel. The shell was largely intact, which suggests that it was not transported far. Shell fragments collected from the base of the section above the till are dated to 7884-8266 cal BP and 7931-8300 cal BP.
Microfossil analyses. Results from the microfossil analyses are displayed in Table 3. The lowermost microfossil samples collected from the Ruddons Point profile section (at 3.60-3.00 m depth, ~3.50-4.10 mOD) indicate the sediment is an inhospitable environment for the microfauna, as within these samples only coal/charcoal, megaspores and plant debris were identified. A slight exception, at 3.30m depth, there were very rare quantities of marine molluscs and ostracods. At 2.90 m foraminifera were identified, including the Ammonia batavus, Cibicides lobatulus, Elphidium macellum, Planorbulina mediterranensis and Elphidium margaritaceum. All samples collected above 2.80 m depth within the section were considered to represent a more hospitable environment for microfauna, with common and diverse microfauna, derived from marginal marine and inner shelf environments, found in all of the samples from

Locality 2: Coastal core
Core description. The coastal core was driven through till close to the mouth of the Cocklemill Burn ( Fig.  8.). The core reached a depth of 1.27 m from 2.23-0.96 mOD. There are no lithic fragments and only sparse amounts of shell fragments within the core. The clay within the bottom ~0.20 m of the core contains noticeably less sand than the sediment above and is noticeably more plastic. Overlying this clay, is a fine, well sorted light beige sand horizon, approximately 6-7 cm thick (1.80-1.74 mOD). The contacts between this sand layer and the clays above and below it are sharp. The overlying clay is dark grey-brown sandy clay with a lack of any lithic fragments. The topmost 0.11 m of the core is a dark brown, finegrained, organic rich sediment ( Fig. 9.).
OSL chronology. Initial field-based OSL screening revealed that the upper most sediment (top ~0.11 m) had distinctly lower OSL and IRSL signal intensities than the sediments below, suggesting a discontinuity in age between the surface sediments, which could have been significantly reworked and bleached by both coastal and fluvial processes. This trend was reproduced in the subsequent laboratory characterization, with the paired aliquots from the three uppermost samples returning a small stored dose compared to the samples from the underlying sediments. A full dating sample was collected from close to the base of the core at 1.05 mOD (CERSA447) and OSL dated to 28.09 ± 1.37 ka, with an over dispersion (σd) of 21 ± 3 % (Table 1). Over dispersion is a measure of variability in the calculated equivalent dose (D e ). An example of the OSL signal and regenerative curve for an aliquot within the sample is shown in figure 10. No samples were collected for radiocarbon dating within this core.

Locality 3: Saltmarsh Core
Core description. The saltmarsh core was collected through the intertidal saltmarsh which infills the lowlands at approximately high tide level. The core extended to a depth of 3.57 m, between 2.71 mOD and -0.86 mOD and reveals both clay and sand rich horizons (Fig. 11.). The lower most 0.33 m of the core consists of a sandy brown clay with rounded to sub-rounded pebbles up to 1 cm in length and evidence of small coal fragments. Overlying this, is a lithic sand with coarse to very coarse lithic fragments and medium grained sand which fines upwards through the core. This lithic sand has a grey- blue colour which transitions to a brown green colouration at ~1.10 mOD. At ~0.70-0.45 mOD there is a brown clay horizon within the sand with mm-scale, sub-rounded lithics, and cm-scale gravel. The sand overlying this sand is dark brown/beige and contains shell fragments and small lithic fragments. Above this is a coarse, black, highly organic layer (~6 cm thick) and the upper most ~1.00 m consists of a dark brown, fibrous and waterlogged saltmarsh clay with a high organic content. The clay contains a small amount of angular shell fragments (<2 %) and an absence of sand and pebbles.
OSL Chronology. Initial rapid screening of the core using portable OSL displayed an overall increase in OSL and IRSL net signal intensities with depth, suggesting an age progression within the core (Fig. 11).
To ascertain the age of the lowest sediment in the sequence, which is also the lowest elevation sediment cored in the field study, a dating sample was colle cted from the lower most clay (CERSA446 at -0.66 mOD). A second sample was collected from the lithic sand stratigraphically above the clay (CERSA445 at 0.08 mOD) to explore the age range between the clay and overlying sand. Results show that the lithic rich sand of CERSA445 date to 18.51 ± 1.28 ka, and the underlying clay to 29.12 ± 3.18 ka. The σd within the lithic sand is 28 ± 5 % but is much higher in the underlying clay at 46 ± 8 %. In this situation there is no evidence of bioturbation however there is potential for either incomplete bleaching at the point of deposition and there are numerous lithic fragments and pebbles within the clay which could be causing variability in the beta dose rates.

Locality 4: Cocklemill Burn raised marine deposits and cores
Sedimentological description. The Cocklemill Burn raised deposits are composed of sands, gravels, organic rich layers containing coal fragments, and shell hash horizons, topped by windblown sands and sandy soil (Fig. 12. (a)). The base of the section from ~1.78-2.38 mOD displays cycles of poorly sorted, medium to coarse sands and units of well-rounded gravels (~2-4 cm in size) with lithologies of sandstones, basalt and quartz. Between the gravel rich units is a 2-3 cm thick bed of coarse-grained organic rich material and a coarse grained sand. Overlying the gravels, from 2.38-2.63 mOD are medium to coarse-grained sands with organic matter in discontinuous layers.
From ~2.63-2.98 mOD area a series of sand beds, which display flaser bedding with visible lenses of darker, coarser material. They do not form a continuous layer and are laterally variable. There are a scattering of larger cm scale pebbles and shell fragments visible within the sands. From 2.98-3.63 mOD are interbedded sands with numerous cm scale gravel beds topped by a gravel bed, ~0.15 m thick, with rounded clasts (ranging from 1-3 cm in scale). Above the gravels, from ~3.63-4.83 mOD the sands are fine to medium grained, with further occurrence of a thin (2 cm thick) black organic layer and thin (2-5 cm thick) gravel beds. Occasional orange iron staining within the sands is associated with infilled burrows and dark patches of organic material. There are also irregularly occurring pebbles and shell fragments within the sands. At ~4.83-5.03 mOD is a gravel and shell hash bed ~0.20 m thick. The sands above this shell/gravel bed (5.03-7.73 mOD) contain shell fragments, with cycles of shell rich beds. The sand units show further evidence of burrowing and horizons of heavy bioturbation infilled by dark organics. The upper most horizons of the marine deposits, from 7.73-8.03 mOD, consist of shell hash and an organic coal-rich layer which varies in thickness laterally across the section. This manifests as two thin distinct coal bands but when followed laterally it combines with a shell hash layer to form a thicker band of shell fragments, larger whole shells and fine and cm chunks of coal fragments. From ~8.03-9.34 mOD are fine to medium sands with evidence of shell fragments and many modern roots within the sands.
The core taken at the base of the section at Locality 4 spans a depth of 2.45 m, from 2.64-0.19 mOD. The bottom ~1.60 m contains sand with discontinuous organic rich layers. Above this is a shell horizon, followed by ~0.80 m of sand with thin gravel layer. The second Cocklemill core taken at Locality 5, approximately ~80 m to the west, spans a depth of 1.22 m, from 2.56-1.34 mOD. At the base of the core is a dark brown, compacted, friable peat, above which is a soft, grey coloured, carse clay. The carse clay and peat were not noted in the Cocklemill core at Locality 4, despite recording sediments at a similar elevation. Above the carse clay is a medium grained sand containing angular shell fragments and discontinuous organic rich bands, but no lithic fragments, similar to the sediments in the core at Locality 4. A thin (1-2 cm thick) shell layer, 0.80 m from the top of the core separates sands which are grey-beige in colour from more orange-beige sands above it.
OSL chronology. A step-change in IRSL and OSL net signal intensities occurs across a break at ~1.35 m depth (8.04 mOD), with the intensities increasing by two orders of magnitude (Fig. 13.). This trend was reproduced in the subsequent laboratory screened OSL dataset with the stored dose "jumping" from sub-Gy to ~ 5 Gy. Moving further down the section, the stored doses from ~1.50 to ~3.14 m depth are in the region of ~4-6 Gy, after which the stored dose appears to broadly increase with depth until a depth of ~4.60 m (~4.74 mOD) after which point the stored dose drops off from ~10 Gy to ~6 Gy for the rest of the section. This implies that the environmental dose rates are different above and below the depth of 4.60 m, or that there is a difference in luminescence sensitivities, which could reflect variations in characteristics such as mineralogy or grain size.
The upper most OSL dating sample (LL2) was taken at 0.89 m depth from the top of the section (8.45 mOD) and provided a date for the upper sand of 0.28 ± 0.07 ka. This young age corroborates the results from the field-based measurements of OSL and calibrated OSL screening, which identified lower signal intensities and stored dose in the upper sands of the section. The next three samples; LL7 at 2.01 m depth (7.33 mOD), LL12 at 2.23 m depth (7.11 mOD) and LL14 at 3.00 m depth (6.34 mOD) in the section were dated to 4.34 ± 1.24 ka, 5.58 ± 0.99 ka, and 4.46 ± 0.66 ka, respectively, with a combined age of 4.73 ± 0.50 ka. OSL sample LL19 at 4.30 m depth (5.04 mOD) was dated to 7.66 ± 2.38 ka, although it has a large associated error. This sample was taken above the break in stored dose identified in the calibrated OSL results and also coincides with a change in magnetic susceptibility readings. The lower section yielded two younger ages of 6.41 ± 1.45 ka (LL20, 4.74 mOD) and 5.63 ± 1.77 ka (LL22, 3.69 mOD). Initial OSL screening of both of the cores at Locality 4 and Locality 5 did not show an obvious increase in signal intensity with depth (Fig. 14, Fig. 15). Likewise, there was no noticeable increase in stored dose through the calibrated OSL in the core at Locality 4. At Locality 4, a full dating sample (CERSA473) was taken near the base of the core, at an elevation of 0.49 mOD, which yielded an age of 6.02 ± 0.53 ka. These three lowest OSL ages are within error of each other, and the combined age is 6.03 ± 0.48 ka.
Magnetic susceptibility. Magnetic Susceptibility was used to complement OSL screening in identifying mineralogical changes within this section, a technique also used by Muñoz-Salinas et al. (2011). Magnetic Susceptibility readings taken at approximately every 5 cm throughout the Cocklemill Burn section range from 0.242 to 4.810 x10 -3 SI Units (Fig. 13.) Radiocarbon chronology. Two marine shells were radiocarbon dated, at 1.85 mOD from within the core, and in the upper most shell horizon of the raised deposits at 7.78 mOD. The lower shell fragment was dated to 8343-8766 cal BP, whilst the upper sample dated to 5585-5961 cal BP. These radiocarbon samples sit within the luminescence profile for Cocklemill Burn, however are discordant with the OSL chronology, with both dates being older than the OSL ages stratigraphically below them. In the core at Locality 5 (Fig. 15), a shell fragment from the shell rich layer (1.77 m OD) was radiocarbon dated to 8679-9181 cal BP. This shell fragment is similar in age to the shell fragment dated in the core at locality 4 at a similar elevation. Both humin and humic acid fractions of the basal peat sample collected were radiocarbon dated. The resulting dates fall within a part of the radiocarbon calibration curve with multiple potential results with corresponding probabilities in parenthesis ( Table 2). The humic acid fraction of the peat was dated with the highest probability to 9085-9292 cal BP (93%) whilst the humin fraction of the peat was dated to 9023-9147 cal BP (64%).

Glacial sedimentation
The oldest dated sediments in the field area are the clays within the coastal (locality 2) and saltmarsh (locality 3) cores which yield ages in the range of ~28-29 ka (CERSA446 and CERSA447 have a combined age of 28.25 ± 1.26 ka). These dates suggest that the clays were deposited during the last glacial period, with ice reaching this part of Fife prior to these dates. The glacial till exposed along the shore face was most likely laid down by ice, and push structures on the till surface also hint at the role of significant glaciation during the deposition of the till. At the ice maximum over Scotland, thought to occur by 27 ka, the ice extended far to the east of the field site, joining with the Fennoscandian Ice Sheet (FIS) and therefore the ice at Ruddons Point may have persisted for some time after the age range of the dated sediments.
Previous 19th and early 20th century accounts describe clays directly overlying till in the region (referred to as boulder-clay) and suggests that this clay was deposited in an Arctic climate. Brown (1867) describes an Arctic shell clay in the nearby area of Elie and likens it to that found at Errol, in the Tay Valley. Based on identification of shell species within the clay, the clay was interpreted to have been deposited in an Arctic cold climate when sea level was around 150-200 ft. (~45-60 m) higher than today. Brown (1867) goes on to postulate that the clay deposited directly above the boulder-clay at Elie, could be largely contemporaneous with the development of the boulder-clay itself. This observation would support the older dates of ~28 and 29 ka which were obtained from clays at the base of the coastal (Locality 2) and saltmarsh (Locality 3) cores. However, the clay at the base of each of these cores did not contain any obvious fossils or fossil fragments on observation, such as is described by both Brown (1867) and Geikie (1902).

A C C E P T E D M A
N U S C R I P T

Timing of deglaciation in southeast Scotland
The overlying lithic sands/sandy clay from the saltmarsh core dating to ~18 ka at 0.08 mOD suggest these sediments were deposited much later than the pre-glacial clay below and point to deposition in the early post glacial landscape. There is still uncertainty as to the timing of initial deglaciation in the east of Scotland. Researchers who have focused on mapping the ice extent across the North Sea have suggested that the BIIS and the FIS were joined until 18.5 ka with ice extending into the North Sea (Sejrup et al. 2016). Sissons et al. (1966) describe low-lying marine sediments at a similar elevation to those at Ruddons Point, which they map as underlying the carse clay, peat and high and main buried beaches in the Grangemouth area of the Inner Forth as Late Glacial in age, however these sediments have not been dated. At a higher elevation along the Tay valley, radiocarbon dates from +20 mOD raised clay deposits, known as the Errol Clay Formation dated the ice retreat from the Wee Bankie Moraine and subsequent marine sedimentation to ~14.5 ka BP (Peacock, 2003). Raised marine silts at Lunan Bay recorded at ~+22 mOD have been dated earlier at 17720 ± 50 BP, suggesting that ice had already retreated westwards of this point at this time (McCabe et al., 2007). The BIIS is thought to have been highly dynamic in the North Sea Basin (Evans et al. 2021). The North Sea Lobe (NSL) of the BIIS was fed by the Firth of Forth ice stream, which also had a strong influence on the retreat of the NSL from 19.9-16.5 ka cal BP (Roberts et al. 2019). Dating of offshore glaciomarine sediments suggests that the deglaciation of the Forth had occurred by 16.9-16.0 ka cal BP (Roberts et al. 2019), with final retreat by 15.8 ka (Evans et al. 2021).
In this study, the environment setting at ~18.5 ka has been interpreted as a time of sea level transgression, when the sea advanced inland due to the receding of the ice sheet westwards. During this time, modelled RSL curves from the Forth Valley suggests RSL was approximately +40 m above present sea level and the Tay Valley RSL curve suggests RSL would have been in excess of +25 m above present sea level (Shennan et al. 2018, Fig. 1). Forsyth and Chisholm (1977) describe the Arctic clays at Elie as some of the oldest of the late glacial marine deposits and suggest they may have been disturbed by ice-push before the deposition of overlying sands. At St Monans, approximately 7.5 km east of the Largo Bay field site, Geikie (1902;Chapter 22, p310) describes sands and clays on the 100 ft. terrace as being "contorted", which suggests they have been overridden by ice. Geikie (1902) draws comparison between these clays of the 100 ft. terrace at St. Monans to the Arctic clays at Elie. This comparison could provide further evidence that these old ages from ~18.5-29 ka, obtained from the Locality 2 and Locality 3 cores, show that sediment pre-dating the last glacial advance are preserved here. In this study, no evidence of glaciotectonism was specifically noted within the cores themselves, however potential ice-push structures were identified on the exposed clay surface close to the site of the coastal core (Fig. 8).
This study has not identified and dated sediments from the Late-glacial period, the period between 14.7-11.7 ka (Walker and Lowe, 2019), however previous researchers have suggested the higher (un-dated) raised rock platforms of the nearby Kincraig Point could have been formed during this time e.g. Cullingford and Smith, 1966), based on correlations to other shorelines in the region and the widely described "100 ft sea" marine incursion, which has long been documented across Fife (Geikie, 1902;Rice, 1961). However, these platforms could also represent older inherited features.

Buried peat and e stuarine carseland sediments
The humin and humic acid RC dates, obtained from the peat horizon at Locality 5 (1.38 mOD), represent a period of lower relative sea level at ~9-9.3 ka. Sea level remained below this elevation for a duration of time long enough to enable the formation of a terrestrial peat. The occurrence of a peat, directly underlying a grey/blue clay is reported at numerous locations in southeast Scotland e.g. western Forth Valley (Sissons, 1970) and at Flisk in north Fife (The SCAPE Trust, 2021). The overlying fine, grey clay represents a relative sea level rise, and return to estuarine conditions. This is known locally as the carse clay and marks the beginning of the Postglacial Marine Transgression (e.g. Smith et al., 1999).
The presence of a submerged forest at Largo Bay has also previously been suggested by Rev. Brown, who reported in the New Statistical Account the occurrence of a peat bed in Largo Bay, containing species of birch, hazel and alder as well as preserved trunks a number of feet in length (Gordon, 1845). Based on the closest RSL model constructed for the Tay region by Shennan et al. (2018; Fig. 1), sea level was approximately 1 m above present-day sea level at 9 ka, however there is a spread of underlying data points at elevations ranging from approximately +1 m to -5m, showing the potential for significant variations in sea level across the region.

Cocklemill Burn Holocene deposits
The raised deposits along the Cocklemill Burn record small-scale fluctuations in sea level within an overall marine transgression, with the depositional environment staying in close proximity to the shore throughout. The presence of fragments of coal identified within the section reflect the location of the site in relation to the underlying bedrock geology of Carboniferous coal series.
Scenario 1: OSL chronology. OSL dating results in ages of ~7.5-6 ka in the core and the lower part of the section. The samples in the upper marine sands were dated to ~5.5-4.5 ka and above this the upper sand layers have been interpreted as windblown sands, with the uppermost sample dated to <300 years. Magnetic susceptibility results for the Cocklemill Burn section also point to a potential change in provenance or deposition history at ~4.5 m depth in the section. This could mark a change in accumulation between the lower sediments of 7.5-6 ka, to a second accumulation of sands from 5.5-4.5 ka, as sea level was continuing to rise through the early to mid-Holocene. Smith et al. (2007) have described that much of the Scottish coastline falls within a convergence zone between the older identified Main Postglacial Shoreline and younger Blairdrummond shoreline, and therefore the two shorelines overlap one another.
Potential uranium mobility. Within the core at Locality 4, the OSL age of 6.02 ± 0.53 ka is younger than the radiocarbon date stratigraphically above it. The younger OSL age may be in part due to the proximity to the river and the sediments being waterlogged. A higher concentration of uranium may point to uranium solubility in the waterlogged sediments sampled within the core, with late deposition of uranium, resulting in an elevated environmental dose rate, and therefore a younger age (Olley, 1996). Assuming that the uranium concentrations would have originally been consistent with the ratio of Th: U in the sediments above, then the dose rate to this sample can be modelled. Assumed water content and effective dose rates for each dating sample are presented in Table 1 and radionuclide concentrations for each dosimetry sample are presented within the supplementary material. Within the Cocklemill Burn section, the average ratio of Th: U is 3.19 (Fig. 16). Using this ratio to revise the estimate of Uranium (ppm) in the CERSA473 dosimetry samples, results in a modelled age of 6.26 ± 0.53 ka, bringing the age closer to concordance with the radiocarbon date stratigraphically above it, although a discrepancy between the two dating techniques is still apparent. With either estimate, the potential for the soluble radionuclides to be mobile, suggests that this age needs to be interpreted as a minimum age. Using this modelled age, the new combined age of the lowest three OSL ages within this section (LL20, LL22 and CERSA473) is 6.23 ± 0.48 ka.
Scenario 2: Radiocarbon chronology. The radiocarbon date from a shell fragment (1.77 mOD elevation) within the sands stratigraphically above the carse clay yielded an age of 9181-8679 cal BP. A shell fragment collected at an elevation of 1.85 mOD from the core at Locality 4, dated to 8766-8343 cal BP, therefore dates from both of the cores at Cocklemill Burn suggest that the high energy sands were being deposited on top of the carse clay in the early Holocene. The final radiocarbon date collected within the Cocklemill Burn section was selected from the upper most sea level horizon identified (7.78 mOD) and dated to 5961-5585 cal BP. This radiocarbon age suggests that the raised marine deposits of Cocklemill Burn range from the early Holocene after the occurrence of the Postglacial Marine Transgression through to the mid Holocene.
Previous studies at Cocklemill Burn. Previous studies along the Cocklemill Burn have yielded different age profiles (Fig. 17). In these studies, the OSL ages in general are younger than the radiocarbon ages at similar horizons. Previous OSL dating of the Cocklemill Burn sediments obtained as part of an undergraduate dissertation by Overshott (2004; key results published in Dawson and Dawson, 2007), interpreted the onset of postglacial marine transgression at 8130 ± 195 cal BP, with transgression continuing until ~ 6 ka. Two unconformities were identified within the stratigraphy, the first at 6 ka and the second between 4.5 and 2.5 ka. The stratigraphically highest date obtained within the Overshott (2004) study was 2.58 ± 0.10 ka, which is younger than the dates identified by Tooley and Smith (2005) or in this current study at a similar elevation, however, the Cocklemill Burn section studied within the current study was situated approximately 80 m further inland. Tooley and Smith (2005) report five radiocarbon dates through the raised marine deposits exposed along the Cocklemill Burn. They conclude that the marine sediments range in age from 9000 to 5280 cal BP (as calibrated in 2005) and represent the two raised shorelines of the Main Postglacial Shoreline and Blairdrummond Shoreline. They also equate the high energy gravels at the base of the Cocklemill section to the Storegga Slide event (2.26-4.03 mOD), which is recorded at numerous field sites around the Scottish coastline, although usually as a sand layer within estuarine clays. The tsunami event has recently been dated within the Montrose Basin, Angus, using OSL to 8.10 ± 0.25 ka . Tooley and Smith (2005) dated a marine shell between 2.49-2.44 mOD within the Cocklemill raised marine deposits to 8180-8360 cal BP (calibrated in 2005) which they describe as redeposited and therefore the Storegga event occurred after this time. Overshott (2004) discusses the possibility of a thin clast and sand horizon, located directly above the carse clay at ~0.8 mOD, of being related to the Storegga event, however the corresponding OSL age obtained at this elevation is younger at 6.86 ± 0.3 ka (dates published in Dawson and Dawson, 2007). OSL dates from both the Overshott (2004) and the current study would place the sand and gravel deposits in the section as younger than the Storegga event. Within the current study, the lowest OSL age from the sands of the Cocklemill core is also younger than the age of the Storegga event, however this could be underestimated due to the proximity to the Cocklemill Burn river and the associated effect on the environmental dose. If this lowest OSL age is discounted, and instead only the radiocarbon age at 1.85 mOD of 8343-8766 cal BP is considered, then the high energy gravel deposits at the base of the section which are stratigraphically above this dated horizon could have been deposited within the age range of the Storegga Slide tsunami event. The high energy gravel deposits have not been directly dated within this study.

A C C E P T E D M A
N U S C R I P T

Ruddons Point Holocene deposits
Short OSL chronology. OSL dating of the raised marine deposits at Ruddons Point yields ages in the range of approximately 2.0-3.9 ka. The darker, organic rich sands at the base of the section above the till were not dated through OSL, due to the low luminescence signal noted during initial screening using the portable OSL reader. All of the Ruddons Point OSL samples display high σd of ~30-50 %. High rates of σd can be due to multiple reasons such as sediment mixing through processes such as bioturbation, incomplete bleaching at the point of deposition and variable environmental dosimetry (Galbraith and Roberts, 2012). At Ruddons Point, the presence of modern roots through the section could have contributed to sediment mixing, and the high energy nature of the deposition of the gravels could have mixed older incompletely bleached sediments with sediments bleached upon deposition. The presence of the two gravel units and smaller gravel clasts through the sands could also suggest small changes in environmental dosimetry, all of which could be contributing to a high σd. Complimentary microfossil analysis conducted at Ruddons Point picked out a distinct environmental change between the darker inhospitable sands at the base of the section with more hospitable sands above.
RSLs are thought to have been falling during this time and this is potentially why these younger deposits are constrained further seaward, potentially onlapping older sediments at the same elevation, and not noted at the Cocklemill Burn location further inland. Modelled RSL curves for the Forth and Tay regions (e.g. Shennan et al. 2018, Fig. 1) show a gradual fall in RSL over the Late Holocene however the formation of the Ruddons Point raised deposits may suggest a more punctuated change in sea level to allow the formation of this shoreline. The difference between the smoothed GIA models and the empirical evidence is also noted further inland along the Firth of Forth by Smith et al. (2010) who describe a shoreline with a mean radiocarbon age of 3600 BP. This is of similar age to the upper Ruddons Point sediments, however the different uplift rates between the inner Forth sites, closer to the uplift centre and the field site at Ruddons Point, approximately 60 km further east, make it difficult to correlate.
Long OSL and radiocarbon chronology. The radiocarbon sampling focused on dating marine shell and shell fragments and provides an extended chronology beyond the sand units dated through OSL. The shell fragments from the base of the section were dated to 7884-8266 cal BP and 7931-8300 cal BP. The timing of the Storegga Slide event falls within the age range of these dated shells, however further detailed investigation into confirming the presence of tsunamiite deposits at Ruddons Point was not carried out within this study. Within the lower gravel/cobble unit, the dated marine shell yielded an age of 7447-7899 cal BP, suggesting that there is an age difference of ~4.6 ka between this gravel unit and the upper gravel which was dated to between 2738-3171 cal BP. These results suggest that the Ruddons Point raised marine deposits record deposition from the end of the Early Holocene right through to the Late Holocene, whereas dating of only the sand units through OSL yielded ages of the Late Holocene only. A summary schematic of the field site is shown in figure 18.

Conclusions
This study delivers a detailed study of the complex environmental changes which have occurred in the south east of Scotland as a result of eustatic sea level changes and crustal responses to the last deglaciation of Scotland. Results from this study include: Undertaking an initial geophysical survey which identified buried features and aided in interpreting the characteristics of underlying sediments, which could be related to subsequent coring.
2) The utilization of portable OSL equipment to create high resolution luminescence stratigraphies for raised marine deposits and subsurface sediments cored at Largo Bay, which allowed hypotheses concerning depositional sequences to be generated in near real-time and guided sample positing for dating. The OSL chronology for the depositional sequences was augmented, and expanded, by the collection of shells from the raised marine deposits as well as sampling humin and humic fractions from the buried peats for RC dating. The comparison of two dating methods allowed for a critical assessment of results and the development of different depositional scenarios. 3) Building a chronology of the coastal and saltmarsh cores (Localities 1 and 2) spanning from the onset of the LGM, to the post glacial marine sediments dated to ~18 ka was obtained through OSL dating. 4) RC dating the basal peat at Locality 5 using both humin and humic acid fractions to the early Holocene at ~9.2 ka. This was a period of relative sea level fall when the sea level was lower than present day, allowing the terrestrial peat to form. The peats are overlain by carse clay, which represents the beginning of the Postglacial Marine Transgression and a subsequent period of relative sea level rise. 5) Dating of the inland raised marine sand and gravel deposits record small-scale fluctuations in sea level in the early to mid-Holocene, capped with much younger <300 yrs old windblown sands. The Cocklemill section contains sediments of Blairdrummond Shoreline age overlapping Main Postglacial Shoreline age sediments. OSL dating suggested sediments within the section and core were younger than that of the Storegga Slide tsunami, however RC dates at both the Cocklemill and Ruddons Point sections suggest that the sediments do encompass this time period, as concluded in the previous study by Tooley and Smith (2005). 6) Magnetic susceptibility measurements were employed and used to interpret depositional changes within the raised marine deposits. 7) OSL dating of the Ruddons Point raised marine sand and gravel were deposited in the late Holocene, however high σd values of 30-50 % and evidence of age inversions within the section suggests high energy deposition with significant mixing and reworking of sediments. 8) Microfossil analysis of the sediments of Ruddons Point identified an inhospitable organic rich environment at the base of the section and a hospitable environment in the sediments above. Climatic indications from the ostracods present in the lower sand unit suggest that the age of the raised beach could be early Holocene at least in part, which would support the RC ages of shells from the lower gravel and base of the section. The upper sand is much less fossiliferous which suggests a different depositional setting to the sediments below.
The field area provided an opportunity to examine the changing environment and depositional history through both the Late Devensian and Holocene. The multidisciplinary investigation has demonstrated the value of detailed local scale field study to evaluate and gain a fuller understanding of past sea level change. This method used can be utilized in future coastal change studies in Scotland or in similar coastal environments worldwide where fewer studies have been conducted or detailed RSL information is missing as the comparison of dating methods builds a fuller appreciation for complex RSL variations and coastal evolution.

Fig. 8.
Panel of field photographs of coastal clays. (a) Clays exposed beneath beach cobbles at the mouth of the Cocklemill Burn; (b) Clays exposed by the river, showing a darker upper clay cap under which is a lighter blueish grey clay with angular clasts. (c) Example of potential ice push structures on the exposed clay; (d) Position of Locality 2 within field study area. Basemap: High resolution (25 cm) vertical aerial imagery © Getmapping Plc, EDINA supplied service https://digimap.edina.ac.uk, Downloaded: 2022-05-08 17:08:48.323.  (Dietze et al. 2016) for CERSA447 where all 24 measured aliquots were accepted (plotted using ShinyPlot). Fig. 11. Luminescence profiling and OSL (ka) ages for the saltmarsh core (Locality 3).       17. Simplification of raised deposits and cores collected in this study across all five field localities of Largo Bay. Two summarised illustrations of previous studies at Cockle mill Burn also presented to compare dated sections to the current study (results from Overshott (2004), as published in Dawson and Dawson (2007); Tooley and Smith (2005). RC dates are presented as calibrated ages (cal BP) followed by the uncalibrated RC age in parenthesis.