Progress since 2009

Sector developments

To ensure sustainable management of the submerged prehistoric cultural resource it is important to understand current use and anticipate changes in the use of the seabed. This section outlines sector-wide developments since 2009 and an overview of legislation and associated guidance which support characterisation and mitigation of the prehistoric palaeolandscape and archaeological resource.

The economic importance of the North Sea and English Channel has long been recognised and they are key areas for fisheries, offshore energy, marine aggregates, and cables and pipelines. Over the last decade there has been an overall reduction in the area of the seabed dredged for marine aggregates but the total tonnage of sands and gravels dredged under Crown Estate licence in England and Wales increased from 18 million tonnes in 2020 to a total of 21 million tonnes of sand and gravel in 2021 (The Crown Estate and British Marine Aggregate Producers Association 2021).

Over the last decade there has been staggering growth in the UK offshore wind sector and the associated energy infrastructure needed to support these developments. There has been a particular concentration of offshore wind farm developments across the southern North Sea due to shallow conditions favouring construction and access to terrestrial power grid connections; these are also same areas likely to contain submerged/buried prehistoric landscapes.

In 2009, operational wind farms were in shallow waters close to the coast with a capacity of 688 MW installed by the end of the year (The Crown Estate and Offshore Renewable Energy Catapult 2019). With the announcement of Round 3 leasing zones in 2008, leasing areas became much larger and were situated in much deeper water. By the end of 2020, the UK’s capacity was 10,412 MW which is equivalent to 33% of the global offshore wind capacity (The Crown Estate and Offshore Renewable Energy Catapult 2019). Looking forward, Round 4 offshore wind leasing, and the UK’s drive towards a net zero clean energy future means the sector will continue to grow, and a technological shift towards floating wind turbines will present new challenges for marine cultural heritage.

The NSPRMF 2009 provided an overview of the legislation and regulation of all the countries adjoining the North Sea. This document focusses on the legislation of England, and since 2009 there have been significant changes.

In January 2009, the United Nations Educational, Scientific and Cultural Organisation (UNESCO)’s Convention on the Protection of the Underwater Cultural Heritage entered into force, having been signed or ratified by 20 member states. Although the UK abstained in the vote on the convention, it has stated that it has adopted the Annex of the Convention, which provides examples of best practices concerning Underwater Cultural Heritage, including prehistoric archaeological material and submerged landscapes.

In 2009, the Marine and Coastal Areas Act (MCAA) also came into force. The Act enabled Her Majesty, by Order in Council, to designate an area as one within which the rights to which the Act applies and are exercisable, an ‘Exclusive Economic Zone’ (EEZ). The Act also created the Marine Management Organisation (MMO). The MCAA 2009 is the primary legislation relevant to marine development plans, and the MMO is responsible for marine licensing in English Waters. Activities that may need a marine licence include construction (including development of and subsequent maintenance and/or alteration of structures as well as laying cables), dredging, and the deposition or removal of any substance or object. Historic England provides advice to the MMO as part of the marine licensing process.

Under the MCAA 2009, marine plans consistent with the UK Marine Policy Statement (MPS; HM Government 2011) have been developed for marine planning regions around the UK, with the NSPRMF covered by the North East Inshore, North East Offshore, East Inshore and East Offshore Marine Plan areas (Table 1). The Marine Plans set out a policy framework which help inform decision-making on what activities take place in the marine environment, and how the marine environment will be developed, protected and improved over the next 20 years.

The majority of legislation in England focusses on shipwreck and aircraft remains, through acts such as the Protection of Wrecks Act 1973, the Protection of Military Remains Act 1986, and the Merchant Shipping Act 1995, and recently intertidal shipwreck sites have also been protected under the Ancient Monuments and Archaeological Areas Act 1979. By contrast, submerged landscapes are generally managed through identification and mapping of the seabed and shallow subsurface to identify deposits or features that have a potential to preserve archaeological material. When artefactual material is recovered within an area impacted by development activity, management is delivered through the implementation of Archaeological Exclusion Zones (AEZ) for specific locations on the seabed, such as the one within Aggregate Licence Area 240 in the East Anglian region (Wessex Archaeology 2011a, Fjordr 2016). If artefactual material is discovered in areas not impacted by development, there is no requirement for an AEZ (e.g. Bouldnor Cliff). However, consideration must be given to local processes that may impact the preservation of this material (e.g. erosion).

Following publication of the MCAA in 2009, a number of Marine Conservation Zones (MCZs) have been identified and designated, with a number of designations given to locations of geomorphic/geological interest. Those within the NSPRMF are outlined in Table 5 and include geomorphology associated with glacial tunnel valleys and the breaching of Dover Straits. Bouldnor Cliff is designated according to its geological importance despite it preserving the only confirmed in-situ submerged archaeological site across the NSPRMF regions.   

Table 5 MCZs designated for geology/geomorphology within the NSPRMF regions

There are a number of heritage guidance documents which are available to seabed developers. Some are generic (such as: JNAPC 2008; Historic England 2008), others pertain to the development type, for example: for marine aggregate dredging (BMAPA and English Heritage 2003, 2005), offshore renewables (DECC 2011; Historic England 2021; Oxford Archaeology and George Lambrick Archaeology and Heritage Consultancy 2008; The Crown Estate 2014; The Crown Estate and Wessex Archaeology 2021; Wessex Archaeology 2007), and ports and harbours (Historic England 2016; PIANC 2015). The guidance documents provide information about assessing, evaluating, mitigating, and monitoring potential effects from development on the submerged prehistoric resource, and in some cases provide frameworks for further investigations and protocols for unexpected archaeological discoveries. There is also specific guidance for archaeological assessments of geophysical and geoarchaeological data (English Heritage 2015; Plets et al. 2013), and for terrestrial Palaeolithic remains (English Heritage 2008) and lithic scatters (English Heritage 2000).

Ownership of shipwreck and aircraft material discovered on the seabed is determined by the Receiver of Wreck under The Merchant Shipping Act 1995. However, The Crown Estate retains ownership of prehistoric artefacts found on the sea, and this ownership can be passed to other repositories, as necessary.

Key projects and research

The Seabed Prehistory: Gauging the Effects of Marine Aggregate Dredging project funded by the Aggregate Levy Sustainability Fund (ALSF) led to the acquisition of high resolution geophysical and geotechnical data to support archaeological assessment and evaluation in five site-scale submerged study areas: The Humber (Wessex Archaeology 2008a), Great Yarmouth (Wessex Archaeology 2008b), Happisburgh and Pakefield (Wessex Archaeology 2008c), Eastern English Channel (Wessex Archaeology 2008d) and Palaeo-Arun (Wessex Archaeology 2008e; 2008f). Bathymetric, sidescan sonar and shallow sub-bottom profiler seismic data were acquired and ground-truthed with vibrocores (generally to depths of <6 m) and grab samples. A suite of palaeoenvironmental and chronological analyses were also undertaken and integrated with the results from geophysical mapping to understand palaeogeography and depositional environments in relation to archaeological potential.

Subsequent to the Seabed Prehistory project, a series of Regional Environmental Characterisation (REC) projects and associated projects were funded through the Marine Aggregate Levy Sustainability Fund (MALSF) for key regions: The Humber (Tappin et al. 2011), The East Coast (Limpenny et al. 2011), The Outer Thames Estuary (EMU 2009; Dix and Sturt 2011) and The South Coast (James et al. 2010). The study areas for these reports were at a regional scale and their reach was multi-disciplinary including assessments of geology, biology and archaeology. For each project, a suite of geophysical, geotechnical, palaeoenvironmental and chronological data was acquired, and the prehistoric archaeological objective was to characterise submerged landscapes and their archaeological significance.

In 2007/2008 a series of Palaeolithic artefacts including handaxes, flakes and cores, and a series of faunal remains (woolly mammoth, woolly rhino, bison, reindeer and horse) were recovered from marine aggregate licence area 240 located off the coast of Norfolk in the southern North Sea (Firth 2011; Tizzard et al. 2014). These findings led to a comprehensive programme of multi-disciplinary works in order to understand the palaeogeography and archaeology of the area and to improve the future management of the potential effects of aggregate dredging on the marine historic environment (Tizzard et al. 2014; 2015). This discovery was of international significance as it is the oldest submerged archaeological site in NW Europe, and it demonstrates there is high potential for minimally disturbed Palaeolithic material to be preserved in submerged contexts.

In 2013 an Audit of the Current State of Knowledge of Submerged Palaeolandscapes and Sites was published to collate and raise awareness of key projects and publicly available geophysical and geotechnical data relevant to submerged prehistory (Wessex Archaeology 2013a). The audit highlighted some of the limitations with existing data and approaches including; limited chronological control and coarse resolution of palaeoenvironmental analysis; variability in geophysical data quality, and; insufficient publication of key archaeological findings (artefactual, palaeoenvironmental or palaeogeographical). The audit baseline was subsequently updated and expanded to the UK and Ireland in 2015, along with reflections on progress of NSPRMF aims (Bicket and Tizzard 2015).

The lack of chronological data from submerged contexts is a problem not only for archaeologists but for the wider Quaternary science community. In 2011 the BRITICE-CHRONO project set out to acquire an extensive suite of chronological data to constrain the timing and rate of ice sheet collapse in Britain at the end of the last glacial period. The project included two research cruises to collect sediment cores from the North Sea providing new chronological data for glacial deposits in the Dogger Bank and Humber regions (Roberts et al. 2018).

Investigations associated with Offshore Windfarms have generated growing amounts of new data. There is an increasing number of academic publications utilising data acquired in support of the Dogger Bank Round 3 offshore wind farm zone. The principal focus is on stratigraphy (Cotterill et al. 2017) and glacial history (Phillips et al. 2018; Emery et al. 2019a), but high resolution mapping of sub-bottom profiler data has identified submerged palaeolandscape features including a buried barrier coastal system (Emery et al. 2019b) and an extensive network of palaeochannels (Emery et al. 2020).

Following on from earlier work on the North Sea Palaeolandscapes Project (Gaffney et al. 2007), the Lost Frontiers Project aims to study past environments, ecological change and the transition between hunter gatherer societies and farming in NW Europe (Gaffney et al. 2017). The results from the Lost Frontiers project were not available at the time of writing but will be published in a dedicated volume in late 2022/2023 (Gaffney and Fitch in press). The Lost Frontiers team are also working as part of a collaborative research consortium with project partners across Europe as part of the Deep History – Revealing the Palaeolandscapes of the Southern North Sea Project. Initial outcomes of this project include the recovery of a flint artefact through grab sampling (debitage) during a targeted offshore survey of a submerged palaeochannel offshore of the Humber (Missiaen et al. 2021) and recently published work indicates evidence of the Storegga Tsunami (Gaffney et al. 2020) and its potential impact on Mesolithic communities in Doggerland (Walker et al. 2020).

Stratigraphic and chronological frameworks

Stratigraphic frameworks

At a national and regional scale, stratigraphic frameworks still rely heavily on the informal BGS stratigraphical framework outlined in various BGS offshore regional reports (e.g. Cameron et al. 1992; Hamblin et al. 1992). A new, formal, unified, lithostratigraphic framework for the entire UK continental shelf was published by the BGS (Stoker et al. 2011) that outlines Formations as principal mapping units that can be correlated with the Quaternary deposits of Great Britain onshore. The resulting hierarchical lithostratigraphic framework identifies 99 Formations across the entire UK continental shelf that can be broadly categorised into 1) Lower to Middle Pleistocene (pre-Anglian), predominantly non-glacial sediments, and 2) Middle Pleistocene (Anglian) to Holocene glacially dominated sediments. In the context of the NSPRMF, the Quaternary deposits of the English Channel remain ‘undivided’.

Despite publishing an updated lithostratigraphic framework, the BGS have not revised or updated Quaternary deposit maps (paper or digital) since the 1990s. A series of BGS Geological Factor Maps including a Quaternary Geology Summary Lithologies map are available digitally at a 1:1M scale (BGS Geoindex Offshore[1]) but these provide a broad lithological overview (e.g. firm to hard mud) and are not formally subdivided according to lithostratigraphy. New products will be available from the BGS in due course and include maps of seabed geomorphological features that will provide some information on Quaternary deposit extent at the seabed (Dove 2021).

At a regional scale, more detailed stratigraphic frameworks have been published, e.g. the revised stratigraphic framework for the Dogger Bank of Cotterill et al. (2017) but there has not been a systematic review or update of Quaternary stratigraphy. Revision largely depends on the locality and availability of, and access to, developer-funded geophysical and geotechnical data. The resulting stratigraphic frameworks typically represent small-scale localised changes in stratigraphy and are located in leased areas of seabed, although given the size and number of offshore wind developments in the North Sea, data coverage is expanding year-on-year and there is potential to correlate frameworks across leased area boundaries to provide regional to continental-scale updates to stratigraphic frameworks.

There are examples where archaeological and geological interpretation of developer-funded geophysical and geotechnical survey data has revealed there is considerably more complexity and sub-divisions within BGS lithostratigraphic Formations (Eaton et al. 2020). Within a single major marine development (e.g. Round 3 or 4 offshore wind farm), multiple frameworks may be developed independently depending on the data being interpreted (e.g. seismo-stratigraphic vs lithostratigraphic) and the objectives of the interpretation (e.g. engineering vs archaeological) (Wessex Archaeology 2019a; 2019b).

There are even notable differences in stratigraphic frameworks developed for developer-funded projects and academic research which in part may reflect differences in the time available for interpretation (commercial = months; academia = years) (Gaffney et al. 2007; Phillips et al. 2018; Eaton et al. 2020; Emery et al. 2019a; 2019b; 2020) and considerable efforts need to be made to correlate frameworks to enable archaeologists, geologists, engineers and planners to interpret outputs (Wessex Archaeology 2019a; 2019b).

The NSPRMF 2009 highlighted the limitations in using static lithostratigraphic frameworks in applied research and expectations should be managed to allow for continued updates and revision of frameworks as additional sedimentological, palaeoenvironmental or chronological information is acquired.

Chronological frameworks

Improved chronological anchoring was a key priority in the NSPRMF 2009 and since its publication, a considerable effort has been made to acquire absolute chronological data through scientific dating and develop chronological frameworks for the Quaternary succession offshore (Tappin et al. 2011, Limpenny et al. 2011; Dix and Sturt et al. 2011; Tizzard et al. 2015; Gearey et al. 2017; Brown et al. 2018; Roberts et al. 2018; Wessex Archaeology 2018c; 2018f; Gaffney et al. 2020). Technical developments in radiocarbon and OSL dating have also allowed age ranges to be extended or provided higher precision date ranges. Radiocarbon calibration curves have also been updated (Reimer et al. 2020; Heaton et al. 2020) together with associated databases (e.g. marine offset database) that can be used for the calibration of marine data (Reimer and Reimer 2017). Furthermore, the use of other dating techniques such as palaeo-secular variation (PSV) (Dix and Sturt 2011), amino acid racemization (AAR) (Penkman 2009) and luminescence dating of K-feldspar is providing independent chronological control and allowing deposits or material not suitable for radiocarbon or OSL dating to be dated. Finally, the use of Bayesian chronological modelling is not only improving the precision of age ranges but is also providing a means to estimate the age of events that cannot be directly dated (e.g. erosion), wiggle match and manage large uncertainties (e.g. Bayliss 2009; Bronk Ramsey 2001; 2009a; 2009b; Bronk Ramsey and Lee 2013).

Recognising the need for improved chronological information in relation to submerged prehistory, a database of intertidal, nearshore and offshore peat deposits around the English Coast was compiled (Hazell 2008). Over 300 records were included, of which 20 are located in the North Sea or English Channel, often in shallow water nearshore settings. The majority of dated offshore peat formed between 11,000 – 9,000 cal. BP but this may reflect a bias towards the relatively shallow water depths from which the peats have been recorded (Ward et al. 2006; Wessex Archaeology 2013a; Waller and Kirby 2020).

Since publication of the database, additional peat deposits have been recorded and radiocarbon-dated from offshore settings (Tappin et al. 2011, Limpenny et al. 2011; Dix and Sturt et al. 2011; Tizzard et al. 2015; Gearey et al. 2017; Brown et al. 2018; Roberts et al. 2018; Gaffney et al. 2020). A total of 86 records of peat deposits were recorded through the Offshore Renewable Protocol for Archaeological Discoveries (ORPAD) (The Crown Estate 2014) between 2011 and 2020. Peats reported through ORPAD typically have little contextual information and are often disturbed as they have been dredged up or recovered from the legs of jack-up platforms. As these deposits are not in-situ, they are only dated in rare cases (e.g Russell and Stevens 2014). Further chronological data is being acquired through development-led research being undertaken to support the marine consenting process (Brown et al 2018; Brown and Russell 2019; Maritime Archaeology 2013; 2016; 2017; Wessex Archaeology 2016; 2018c; 2019a; 2019b). However, lag times between data acquisition and public release of the grey-literature means some data is unavailable for the time being.

A systematic review and database update of additional radiocarbon-dated offshore, and nearshore peat deposits has not been undertaken but those records published or publicly available indicate there is an increasing number of dates from the Late Pleistocene (Brown et al. 2018; Maritime Archaeology 2013; 2016;  2017; Wessex Archaeology 2016; 2018c; 2019a; 2019b). They also show a wide variety of material is being dated including seeds and wood fragments, bulk samples and marine shell (Tizzard et al. 2014; Brown et al. 2018). The reliability of these dates is highly dependent on the formation processes of the deposits as well as the nature of the dated material, and there is often a severely restricted choice of dateable material due to preservation factors. While peat deposits are the primary source of organic material for radiocarbon dating, on rare occasions faunal remains (e.g. mammoth tusk) have been dated (e.g. Allen et al. 2008) but this approach depends on the recovery of suitable material and there is often very little information on context and the radiocarbon dating is used simply to corroborate biostratigraphy. Across the wider North Sea more than 200 fossil bone samples have been dated by radiocarbon methods, with localised hot spots in the Brown Bank and Eurogeul areas of the North Sea (van der Plicht and Kuitems 2022). The results of such a comprehensive dating programme highlight issues in understanding context but also the potential for contamination of the fossilised material by younger carbon producing anomalous ages (van der Plicht and Kuitems 2022)

OSL dating is typically used for deposits not suitable for radiocarbon dating and over the last decade it has been used routinely to resolve chronological frameworks in offshore contexts (Limpenny et al. 2011; Tappin et al. 2011; Dix and Sturt 2011; Tizzard et al. 2014; 2015; Roberts et al. 2018; Wessex Archaeology 2018c; 2018f) and has provided ages for deposits from which significant archaeological material has been recovered (e.g. Limpenny et al. 2011; Tizzard et al. 2014; Tizzard et al. 2015). The level of reporting associated with OSL dates from development-led research is variable and there are also differences between laboratories in the number and size of aliquots used. Collectively this makes it difficult to assess the reliability or precision of the dates produced.

The OSL dating strategy for the Thames REC (Dix and Sturt 2011) included the use of Rangefinder Ages (Roberts et al. 2009) to provide a rapid approximate age estimate and identify samples most suitable for conventional OSL dating. Producing Rangefinder Ages is rapid, but they do not have the same accuracy or precision as conventional OSL ages and may produce uncertainties of ~30% (Roberts et al. 2009). This approach may be suitable in development-led projects where the aim of the dating may be to broadly estimate age (i.e. within a Marine Isotope Stage) or to support the design of a targeted sampling strategy. However, the collection of ‘good data points’ was highlighted as a priority in the NSPRMF 2009 and continues to be of great importance and in contexts that have associated archaeological remains, acquiring precise and accurate ages are key.

OSL dating relies on quartz minerals that are ubiquitous in sediments and act as relatively stable dosimeters. Infrared Stimulated Luminescence (IRSL) is an established technique that uses feldspar instead of quartz. However, feldspar minerals are less stable than quartz leading to an effect called anomalous fading which commonly underestimates ages. Recent developments in infrared stimulated luminescence (IRSL) dating of K-Feldspar using post-IR IRSL (Zhang and Li 2020) have allowed a more stable feldspar signal to be isolated providing more accurate age estimates. This is significant, as the feldspar signal saturates at a higher dose than quartz, which means it can be used to date deposits older than the age range of OSL (~200 ka) (Roberts et al. 2015). Furthermore, paired OSL-IRSL dating of the same deposit can provide an independent chronological control as a further test of the reliability of ages (Colarossi et al. 2015).

One of the biggest developments in chronological frameworks since the NSPRMF 2009 is the use of Bayesian modelling which incorporates all chronological information, including radiocarbon and other absolute dates, alongside other relative age information (e.g. stratigraphy) to construct a statistically robust chronological framework (Bronk Ramsey 2009a). This approach is almost routinely applied to sequences of radiocarbon dates in archaeology (Bayliss 2015) but is becoming more widely used when integrating a range of scientific dating techniques (Chiverrell et al. 2013) which means independent chronological data can be tested and validated.

Bayesian chronological modelling is being used in development-led projects where multiple dates are being obtained in a single sequence (Wessex Archaeology  2018c; 2018f) and it has also been applied retrospectively in Marine Aggregate Licence Area 240 (Tizzard et al. 2014; 2015) to refine the age of the deposit containing significant archaeological material, from MIS 9-MIS 6 to MIS 7-MIS 6 (Marshall et al. 2020).

Obtaining material or samples suitable for scientific dating from offshore contexts is limited when compared to terrestrial counterparts and in many cases, relative dating is the only means for providing age information. Relative age can be determined from stratigraphic position (vertical and lateral) and interpreted sequence of events (e.g. before sea-level transgression). Biostratigraphy, and in the context of prehistoric archaeology, mammalian faunal assemblages can provide key relative age information. Assessment of microfossil assemblages has proved key, particularly in Middle Pleistocene deposits where the presence of an individual taxon (e.g. Azolla) can be used to indicate relative age according to when it became extinct in Britain (Watts 1998).

Sea-level data can be used to provide “proxy” age information where elevation data relative to Mean Sea Level (MSL) is known. This approach is typically used in scenarios where there is no other chronological information (Emery et al. 2019b) but can be used alongside absolute dates as an independent chronological control (Mellett et al. 2012a; Shennan et al. 2018). However, it should be noted that relative sea-level curves derived from Glacio-Isostatic Adjustment (GIA) models for the North Sea and English Channel are poorly constrained and may have errors of 5-10 m for some time periods, particularly those that have very few Sea-Level Index Points (SLIPs) to constrain the model outputs (Shennan et al. 2018). Furthermore, the application of this technique to Pleistocene deposits or surfaces that formed before the Last Glacial Maximum is challenging and relies on global sea-level reconstructions (e.g. Waelbroek 2002) which can have much greater errors of ± 20 m (Sturt et al. 2013). These errors should be considered when using relative sea-level curves to establish a broad chronology. As part of the BRITICE-CHRONO Project, work is ongoing to improve GIA modelled sea-level curves for the North Sea which can feed into chronological and paleogeographic studies.

Landscape and palaeogeography

Building on the successes of the North Sea Palaeolandscapes Project (Gaffney et al. 2007), geomorphological mapping of submerged landscapes preserved at the seabed, or in the shallow sub-surface, has been a key focus of both academic (e.g. Emery et al. 2020; Gaffney et al. 2020; Walker et al. 2020; Dudley et al 2021) and development-led research (COWRIE 2011; Plets et al. 2013). The importance of submerged landscapes was recognised and the European Marine Observation Data Network (EMODnet) published a harmonised classification of submerged landscape features at a pan-European scale (https://www.emodnet-geology.eu/data-products/submerged-landscapes/). The British Geological Survey also published a guide on seabed geomorphological classification to support submerged landscape mapping (Dove et al. 2016). A full suite of submerged landscape features have been identified including coastlines (e.g. Mellett et al. 2012a; Emery et al. 2019b); estuaries and embayment’s (e.g. Wessex Archaeology 2019a; 2019b; Gaffney et al. 2020), peatlands and wetlands (e.g. Maritime Archaeology 2017; Eaton et al. 2020; Waller and Kirby 2020) and river networks (e.g. Emery et al. 2020). Collectively, these advances provide the tools and the means to harmonise and synthesise submerged landscape mapping across the NSPRMF study area which should be a priority for future research.

Paleogeographic reconstructions are underpinned by geological data (lithostratigraphy, Quaternary geology maps, seabed and sub-surface geomorphology) tied to chronostratigraphic frameworks and a strong understanding of sea-level history which essentially controls the timing and extent of sub-aerial exposure, but also preservation and erosion and thus the survival of any archaeological material. By synthesising these data and information, conceptual or numerical models can be constructed at various scales in space and time.

When considering sea-level research in relation to palaeogeographic reconstructions, significant efforts have been made to further understanding. A new sea-level database for Britain and Ireland has been produced, containing >2100 data points records relative sea-level change over the last 20 ka for 86 regions, including the North Sea and English Channel (Shennan et al. 2018; Cohen et al 2021). This extensive dataset has been used to test Glacial Isostatic Adjustment (GIA) models showing good agreement between sea-level index points (SLIPs) and limiting data, and GIA relative sea-level predictions (Shennan et al. 2018). In the NSPRMF study area, most of the sea-level data points correlate to the Holocene period as they are located in and around the present-day coast and there are very few SLIPs from submerged contexts to constrain Pleistocene relative sea level increasing uncertainty in model predictions for this period. There are a number of sea level limiting points for the last interglacial period (Cohen et al. 2021) but these are located on land due to the relatively higher sea-level during the Eemian period.

The sea-level data of Shennan et al. (2018) from submerged contexts is a publicly available geospatial product delivered through the EMODNet Geology Portal (https://emodnet.ec.europa.eu/en/geology). Within the updated NSPRMF study area, 38 sea-level data points are available, but most are limiting data points and only two data points are from radiocarbon-dated basal peat. A limiting data point is a dated sediment from freshwater, estuarine and marine environments that provides information on the relative position of the sea but cannot be directly linked to the past tidal frame (see Hijma et al. 2015).

Development-led geotechnical investigations of the seabed and shallow sub-surface are acquiring hundreds of new Cone Penetration Test (CPT), boreholes and shallow vibrocores from areas of proposed development. Through geoarchaeological assessments, these new data are being assessed to identify deposits of archaeological interest including peat. This is providing new transgressive sequences in areas where pre-existing data was very sparse (e.g. Maritime Archaeology 2017; 2018; Wessex Archaeology 2018c; 2018f; 2019a; 2019b) that can be dated and feed into the sea-level database if metadata is recorded following the protocol of Hijma et al. (2015). This will improve the accuracy of sea-level reconstructions for the Late Pleistocene which are particularly relevant to submerged contexts at lower elevations.

Sea-level reconstructions are a primary input to paleogeographic models at a continental to regional scale. As stated above, relative sea-level changes in Britain for the Holocene are well constrained but there are relatively few data points from the Late Pleistocene or from earlier interglacial periods (e.g. Ipswichian). For these time periods, palaeogeographic models rely on global mean sea-level reconstructions which can have errors of ± 20 m (Shakun et al. 2015).

Sturt et al. (2013) published new models showing inundation and palaeogeography of Britain and Ireland at a 500-year interval time series for the duration of the Holocene (11.7 ka – 500 yr BP). This research provides a continental shelf-scale overview of the extent, timing, and significance of landscape change in a digital interactive format that can be used to test, and in turn refine, palaeolandscape and palaeogeographic reconstructions. Principal inputs into the palaeogeographic model were mean eustatic sea-level calculations at 500-year intervals, on a 5 km grid, extracted from the Glacial Isostatic Adjustment (GIA) model of Bradley et al. (2011) and a gridded bathymetric surface based on all publicly available data. There are limitations in this approach; bathymetry shows the elevation of the present-day seabed and due to changes in erosion and sedimentation through time, obtaining an accurate representation of palaeo-topography is challenging and prominent modern seabed features such as sand banks influence the model outputs. Despite the challenges, the models provide a baseline dataset that can help shape targeted research (e.g. Dudley et al. 2021).

Seabed bathymetry has been used as an approximation of the land surface before Holocene sea-level rise inundated the continental shelf (Bicket et al. 2017; Gaffney et al. 2020), notwithstanding the limitations outlined in Sturt et al. (2013). Land surfaces representing earlier periods of sub-aerial exposure are buried in the sub-surface and much more likely to be fragmented. Despite this, interpretation of seismic data can be used to identify and map former land surfaces and where closely spaced 2D or 3D seismic data has been collected, it is possible to produce a palaeo-digital elevation model (DEM) of this land surface (e.g. Gaffney et al. 2007). A palaeo-DEM can be used alongside sea-level reconstructions to model palaeogeography of the subsurface (Wessex Archaeology 2019a; 2019b; Emery et al. 2019b; Eaton et al. 2020; Walker et al. 2020; Dudley et al. 2021). These models should again be interpreted with caution due to changes in sedimentation through erosion and reworking but they can provide an insight into former landscapes that can be used to identify habitable areas with potential to preserve archaeological material.

Producing palaeogeographic models using numerical information in the form of elevation and sea-level data is not always possible due to sparse data resolution or a fragmentary sedimentary record. More often, conceptual paleogeographic maps are produced using expert judgment and geological principals (Hijma et al. 2012; Mellett et al. 2012a; Tizzard et al. 2014; 2015; Emery et al. 2019b). These conceptual models can be used to broadly understand landscape evolution of an area but could not be used to pinpoint or target archaeological sites.

Palaeontological and palaeoenvironmental assemblages

Since publication of the NSPRMF 2009, the following new palaeoenvironmental datasets have been acquired:

• Reassessment of historic finds of Pleistocene terrestrial vertebrate remains (Bynoe 2014, Bynoe et al. 2016);
• New terrestrial Quaternary mammal finds made through the BMAPA protocol and Operational Sampling events associated with aggregate extraction in the PalaeoYare catchment, and;
• Paleoenvironmental studies of Quaternary micro-fossils from marine vibrocores.

The NSPRMF 2009 strategic priorities identified were the provision of:

• new and expanded datasets;
• improved biostratigraphy chronologies and absolute dates, and
• application of new palaeoenvironmental techniques (e.g. aDNA and isotopic analysis).

New and expanded datasets

Two sets of data are available for submerged prehistoric terrestrial landscapes of the North Sea, these are dredged finds of terrestrial mammalian fauna and paleoenvironmental proxy data obtained from direct sampling of sediments through vibrocores.

Some significant individual finds of terrestrial vertebrate fauna have been made since 2009 and reported through the BMAPA protocol and Operational Sampling of aggregate dredging within the PalaeoYare catchment. A combination of historic datasets and finds reported since 2003 by the BMAPA protocol provide a baseline dataset for establishing areas in the southern North Sea where fossiliferous deposits are present, in particular those dating to Pleistocene. However, these are from a relatively constrained geographic area (Area 240 and the offshore zone of Happisburgh in East Anglia, and aggregate licence Area 447 in the Thames region) and there is very little contextual information meaning additional retrospective geoarchaeological and palaeoenvironmental assessment is often required to trace the source of this material.

Perhaps the most notable ‘new’ mammalian fauna dataset is renewed understanding of historic dredged Pleistocene material from the southern North Sea (Bynoe 2016; 2018). This material was recovered by the fishing industry during 19th and 20th centuries. By combining curatorial records with an assessment of historic trawl locations it has been possible to locate the fishing  grounds from which material was recovered on a regional, and in some instances more local scale, and relate this to broad chronological patterning based on biostratigraphy.

Prior to 2009 palaeoenvironmental proxy datasets resulting from direct sampling of submerged Quaternary terrestrial deposits were rare. Since 2009 paleoenvironmental assessment of such samples taken from vibrocores in advance of proposed offshore developments (particularly windfarms) has begun to become standard practice as part of environmental impact assessments, with significant new palaeoenvironmental datasets obtained (e.g. Wessex Archaeology 2019b). This, combined with associated sedimentological and chronological data, has begun to improve understanding of submerged Pleistocene and Holocene landscapes. Continued developments in this area are key to further developing knowledge of prehistoric terrestrial landscapes, and for identifying contexts with potential to preserve prehistoric archaeology.

Improved biostratigraphy chronologies and absolute dates

The provision of improved biostratigraphic chronologies and absolute dates were identified as strategic priorities in the NSPRMF 2009. Recovery of samples taken from vibrocores and advances in dating techniques has increased the number of absolute direct dates for Pleistocene and Holocene sediments. Of particular significance in this regard has been the provision of OSL and post IR IRSL dates for inorganic sediments, expanding on both the type and chronological range of Quaternary sediments that can be dated. Nonetheless, establishing well-constrained chronologies remains a challenge, particularly for earlier Pleistocene sequences, as has been demonstrated by the dating of late Middle Pleistocene sediments from Area 240 (Tizzard et al. 2015).

Despite increases in the number of paleoenvironmental datasets, there has been little development in biostratigraphic chronologies for submerged terrestrial deposits since 2009, particularly mammalian biostratigraphy. This reflects the fact that despite an increase in the number of mammal bone finds (both rediscovery of historic material and new finds), they remain chance discoveries of large, robust elements, principally mammoth teeth, whose specific contexts are unknown. To develop improved mammalian biostratigraphies a wider range of data from known stratigraphic contexts is required, including chronologically significant small mammal material but detection of such material through the BMAPA and ORPAD protocols can be challenging due to their small size.

Application of new palaeoenvironmental techniques

Since 2009 assessment of terrestrial Quaternary deposits for an array of palaeoenvironmental proxy datasets (including plant macro-fossils, insects, pollen, diatoms, ostracods, foraminifera and molluscs) have become increasingly standard as part of environmental impact assessments for offshore developments and research programmes. Additionally, new and developing techniques, such as sedaDNA, have been trialled (Gaffney et al. 2020b; Momber et al. 2021).

There has, however, been no paleoenvironmental investigation utilising two techniques highlighted in the NSPRMF 2009, stable isotope analysis to track animal population movements and aDNA aimed at established, geographical species differentiation at the population level.

Archaeological assemblages

This section considers the advances in archaeological datasets from the North Sea in relation to strategic research priorities identified in the NSPRMF 2009. The strategic priorities identified in the previous review were the provision of:

• new and expanded datasets;
• improved chronology, and;
• development of ‘site’ prospection techniques.

New and expanded datasets

Since 2009 the number and quality of prehistoric archaeological datasets have expanded, but this expansion has not been spatially or chronologically even.

The principal expansion in prehistoric archaeological datasets has resulted from investigations following the 2008 discovery of Middle Palaeolithic archaeology from submerged deposits of the PalaeoYare, in the East Anglia region (Tizzard et al. 2015). Initial investigations provenanced these discoveries to marine aggregate Licence Area 240. Subsequent detailed work has established context, albeit to a regionally extensive sedimentary deposit, and a chronology for this archaeology (Tizzard et al. 2015 – see also Marshall et al. 2020). Reporting of prehistoric finds through the BMAPA Protocol and active archaeological Operational Sampling of aggregate dredging in the PalaeoYare catchment has resulted in the recovery of further prehistoric artefacts, both from Area 240 and other licence areas (Wessex Archaeology 2021).

Although not directly in NSPRMF study area, since 2009 there has been further work and publication relating to Lower Palaeolithic archaeology from Pakefield an Happisbugh, on the Suffolk and north Norfolk coasts (Parfitt et al. 2005; 2010; Ashton et al. 2018). These investigations how now extended into the nearshore coastal area below MLWS (Wessex Archaeology 2008c; Ashton et al. 2018) with the intentions of mapping the extension of the archaeologically important CF-bF offshore. In both cases, the CF-bF was not identified but it was noted that the spatial extent of existing geophysical data is not sufficient to detect the CF-bF if present (Ashton et al. 2018). While the CF-bF was not detected, diving off the coast of Happisburgh uncovered an ex-situ Pleistocene rhinoceros radius, although its context and relationship with the CF-bF is unknown (Ashton et al. 2018).

It is notable that since 2009, despite the implementation of the BMAPA protocol, no prehistoric archaeology has been recovered from marine aggregate licence areas outside of the PalaeoYare catchment, and that the only significant archaeological discoveries from these areas are material of late Mesolithic sites from Bouldnor Cliffs (Momber et al. 2011; 2021), where a submerged prehistoric landscape associated with archaeology had been identified prior to 2009. However, it is apparent from the significant number of finds being reported from beach-replenishment schemes at Walcott-Bacton in Norfolk (source aggregate PalaeoYare catchment) and Clacton (source aggregate Thames) that a significant amount of Palaeolithic material is preserved in particular locations in the North Sea.

This lack of expansion in datasets from other areas may at least partially reflect the archaeological potential of deposits being dredged or a bias resulting from differences in mitigation strategies depending on the significance of archaeological material being recovered. However, advances made in understanding of submerged landscapes and palaeogeographies in these regions suggests that deposits with prehistoric archaeological potential do occur. Consequently, a key lesson that may be drawn from the expansion of archaeological datasets from the PalaeoYare since 2009 is that this success can be attributed to active monitoring of aggregate dredging through Operational Sampling. Although the specific approaches to Operational Sampling are dependent on aggregate processing methods at individual wharfs, methods of Operational Sampling undertaken to date have generally involved samples of aggregate from oversize fractions being assessed by eye by archaeological contractors.

The requirement for Operational Sampling is assessed according to the archaeological potential of a deposit pending a comprehensive palaeolandscape assessment which is undertaken in support of aggregate licence applications or renewals. It is important that any Written Scheme of Investigation for marine aggregate licence areas considers the methodologies of individual wharfs in relation to the age and archaeological significance of the material that may be recovered and incorporates a management plan to ensure proactive, as oppose to reactive, mitigation.

Improved chronology

As with the expansion in prehistoric datasets, improvements in chronology for prehistoric archaeology since 2009 have largely focussed on the Lower/Middle Palaeolithic archaeology from Area 240.

Dating of the Palaeolithic archaeology from Area 240 has been achieved through provenancing the material to a particular stratigraphic unit (Unit 3b) which is an extensive unit widely mapped across the area likely representing multiple phases of channel incision and infill. The results of an OSL dating programme within Unit 3b suggest that it dates to the late Lower/early Middle Palaeolithic (MIS 9–MIS 7; Wessex Archaeology 2011a; Limpenny et al. 2011) although Bayesian modelling suggests a later date (MIS 7-MIS 6; Marshall et al. 2020). Dating this archaeology is a notable achievement; however, it remains unclear whether the handaxe and Levallois assemblages date to the same, or separate periods. The relatively coarse chronological resolution and differing interpretations of OSL age estimates partially reflects universal issues with providing chronology for earlier Palaeolithic sites. However, in this submerged context the issues are exacerbated by lack of detailed contextual information on from which deposits within Unit 3b particular artefacts, and groups of artefacts, originate.

Establishing detailed chronology for prehistoric archaeology from submerged contexts is a challenge. However, in the absence of organic artefacts which can be directly dated, the results of Area 240 investigations suggest that this can be achieved by establishing a stratigraphic context for the archaeology and provision of a chronological framework for that stratigraphy. The results of the Area 240 investigations also demonstrate that to provide more detailed chronology for archaeology, particularly Palaeolithic archaeology, more detailed contextual information on the provenance of archaeological finds would be required.

Prospection techniques

The NSPRMF 2009 identified improved ‘site’ prospection techniques for prehistoric archaeological ‘site’ detection and characterisation as a strategic priority. Approaches to prospection techniques have been reviewed by Sturt et al. (2016).

The expansion in archaeological datasets since 2009 has demonstrated that implementation of initiatives such as the BMAPA protocol and archaeological Operational Sampling of dredged aggregate are key tools for expanding knowledge of the prehistoric archaeological record from submerged contexts. Although the datasets this provides are contextually ‘coarse-grained’, and can at best be related to landscapes, rather than ‘sites’, they are key to establishing areas where landscapes and sites associated with prehistoric archaeology may be preserved. Analogy can be drawn between the current state of knowledge regarding prehistoric archaeology from submerged contexts and to the terrestrial Palaeolithic record in the later nineteenth and early twentieth centuries. During this period extensive collecting of Palaeolithic artefacts, particularly from aggregate extraction, identified deposits and contexts in different geographic regions preserving Palaeolithic archaeology. The patterns established, and the archaeological datasets amassed, form the basis of current more nuanced and targeted Palaeolithic investigations.

The results of investigations in Area 240 and Bouldnor Cliff demonstrate that to identify archaeological ‘sites’ (i.e. locations and archaeological assemblages reflecting human activity at an ethnographic scale), targetted archaeological prospection is required. Direct archaeological prospection through grab sampling was undertaken as part of the Seabed Prehistory Projects (Wessex Archaeology 2008a-f) but no artefacts were recovered. There are limitations in grab sampling in that the deposit of interest needs to be exposed at the seabed with no or little modern sediment cover. Even where this occurs, it is difficult to determine if a find is in-situ given the large volume of material recovered by a grab and limited contextual information. If detailed geophysical and vibrocore surveys are undertaken to complement grab sampling the find may be contextualised to some extent (e.g. Missiaen et al. 2021). The success of direct prospection for Mesolithic sites associated with the development of Yangzte Harbour at Rotterdam as part of the Maasvlakte 2 programme (Vos et al. 2010; Moree and Sier 2015; Peeters et al. 2020; Peeters and Amkreutz 2020) potentially demonstrates how approaches to targeting direct grab sampling locations can be successfully applied.

In the case of Yangtze Harbour the key to success was detailed analysis of the palaeolandscape to model palaeography, linked to assessment of landscape contexts favoured by prehistoric groups, based on terrestrial datasets. This enabled similar submerged landscape contexts to be directly targeted for sampling and which resulted in the recovery of archaeological evidence. Such an approach is applicable to all prehistoric periods. Combining this approach with advanced prospection techniques such as high-resolution and ultra-high-resolution geophysical data, vibrocore data and SedaDNA analysis (e.g. Smith et al. 2015; Gaffney et al. 2020b; Missiaen et al. 2021) complemented by palaeoenvironmental and chronological assessments, potentially provides a framework for future targeted archaeological ‘site’ prospection in the North Sea region.

Image credit: On The Way Up, Wessex Archaeology CC BY-NC 2.0

[1] https://www.bgs.ac.uk/map-viewers/geoindex-offshore/