Quaternary environments of the central North Sea from basin-wide 3D seismic data

Transcription

1 2016 Quaternary environments of the central North Sea from basin-wide 3D seismic data A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Engineering and Physical Sciences Rachel M Lamb School of Earth, Atmospheric and Environmental Sciences

2 Contents List of Figures...2 Project Abstract.. 4 Declaration..5 Copyright statement...6 Acknowledgements. 7 Chapter 1: Introduction...8 Chapter 2: The early Quaternary North Sea Basin.41 Chapter 3: Early Quaternary sedimentary processes and palaeo-environments in the central North Sea..63 Chapter 4: The evolving Pleistocene palaeo-environments of the central North Sea: 2.58 Ma to 0.48 Ma...96 Chapter 5: Tunnel valleys of the Dogger Bank analysed using ultra-high resolution 2D seismic data Chapter 6: Summary References.178 Appendix 1: Project Datasets Appendix 2: 2 nd Author Manuscripts Word Count: 50,414 ~ 1 ~

3 List of Figures Figure 1.1 Regional dataset map 10 Figure 1.2 Summary of central North Sea Quaternary stratigraphy 14 Figure 1.3 3D seismic data coverage 1970 to 1989 and 1990 to Figure 1.4 Project workflow 20 Figure 1.5 Chronostratigraphic well tie panels 24 Figure 1.6 Chronostratigraphic calibration panel 26 Figure 1.7 Decompaction workflow 29 Figure 1.8 Seismic section, horizon slice and time-slices 32 Figure 1.9 Landform models 34 Figure 2.1 Location map and datasets. 42 Figure 2.2 Seismic section and line interpretation 45 Figure 2.3 Calibrated depth x TWT plot 47 Table 2.1 Chronostratigraphic correlation 48 Figure 2.4 TWT structure map of the 2.58 Ma surface 49 Figure 2.5 Depth converted 2.58 Ma surface 50 Figure 2.6 Simplified back-stripped sections 51 Figure 2.7 Simplified line transects of Quaternary margin 52 Figure 2.8 Seismic section of the base Quaternary sub-crop 54 Figure 2.9 Simplified map of the base Quaternary sub-crop 55 Figure 2.10 Thickness maps of Quaternary strata 56 Figure 2.11 Seismic attribute extractions of early Quaternary strata 59 Figure 2.12 Reconstructed palaeo-environmental map of the 2.58 Ma North Sea 61 Figure 3.1 Regional location map and datasets 65 Figure 3.2 Local map of figure locations 67 Figure 3.3 Regional seismic cross section (50x VE) 69 Figure 3.4 Amplitude variance horizon extraction 72 Figure 3.5 Table of representative trough cross-sectional forms 75 Figure 3.6 Comparison of two different types of troughs 77 Figure 3.7 Classification of furrowed seismic horizons 80 Figure 3.8 Long profile of trough 83 Figure 3.9 Depth maps of the furrowed interval 85 Figure 3.10 Amplitude extraction slice of F04 horizon 87 ~ 2 ~

4 Figure 3.11 Well correlation panels for wells 30/6-3 and 22/25a-3 89 Figure 3.12 Palaeo-geographic reconstruction of the early Quaternary basin 91 Figure 4.1 Location map and datasets 98 Figure 4.2 Seismic section and line interpretation 101 Figure 4.3 Series of TWT thickness maps from 2.58 Ma to 0.78 Ma 105 Figure 4.4 Chronostratigraphic calibration seismic well panels 109 Figure 4.5 Back-stripping transect panel 112 Table 4.1 Chronostratigraphic correlation table 115 Figure 4.6 Revised stratigraphy of the early Quaternary for the central North Sea 117 Figure 4.7 Seismic attribute horizon panels for geomorphic mapping 120 Figure 4.8 Maps of distribution of geomorphological features 124 Figure 4.9 Seismic cross sections of glaciotectonic thrust complexes 129 Figure 4.10 Reconstructed palaeo-environmental maps for the North Sea 131 Figure 5.1 Regional map of the Dogger Bank Development Zone 138 Figure 5.2 Summary of the Dogger Bank Stratigraphy 140 Figure 5.3 Map Tranche A tunnel valleys and figure locations 142 Table 5.1 Table of tunnel valley metrics 144 Figure 5.4 Seismic cross-sections of tunnel valley infill packages. 146 Figure 5.5 Seismic long profile section of dipping clinoforms of infill package I 148 Figure 5.6 Seismic cross-section of tunnel valley generational relationship 151 Figure 5.7 Seismic cross-section of tunnel valley and large thrust complex 153 Figure 5.8 Seismic cross sections of tunnel valleys and shallow deformation 155 Figure 5.9 Map of shallow deformation 157 Figure 5.10 Seismic cross-section of southern North Sea tunnel valley 158 Figure 5.11 Conceptual model of the change in shallow sediment deformation 160 Figure 6.1 Summary of North Sea Quaternary stratigraphy 168 Figure 6.2 Summary of seismic horizons and chronostratigraphy correlation 175 ~ 3 ~

5 Project Abstract Climate change during the last 2.5 million years is characterised by glacial-interglacial cycles of fluctuating sea level and temperature increasing in magnitude and duration towards the present day. The central North Sea preserves these glacial-interglacial cycles in an expanded sedimentary sequence creating a high resolution palaeo-climatic record. Basin-wide, low-resolution 3D seismic data, covering more than 80,000 km 2 of the central North Sea, is combined with high-resolution, broadband 3D seismic, regional 2D seismic and local ultra-high resolution seismic from the Dogger Bank windfarm development zone in order to investigate in full the sedimentary sequence. The evolution of the basin is analysed along with the preserved geomorphological landforms in order to build a framework for the development of the North Sea and its changing palaeo-environments from the inception of the Quaternary (2.58 Ma) until the extensive glacial unconformity formed during the Elsterian (0.48 Ma). At the onset of the Quaternary the structure of the North Sea was that of an elongate marine basin, rapidly infilled from the south by continued progradation of the large clinoformal deposits of the southern North Sea deltaic system. The basin rapidly decreased in extent and depth however it was not until around 1.1 Ma that the broad, shallow shelf of the present day was fully established. A revision of the current seismic stratigraphy is proposed, identifying four new Members within the Aberdeen Ground Formation taking into account the development of the basin through time. Powerful downslope gravity currents dominated the basin during much of the early Quaternary, although a well-established, anti-clockwise tidal gyre acted to gently modify the gravity currents. Iceberg scouring was nearly continual from the onset of the Quaternary until grounded ice sheets began to penetrate into the basin from 1.7 Ma, more than half a million years before any previous estimates. Effects of confluence of the British and Fennoscandian ice sheets are observed from 1.3 Ma. The tunnel valleys of the Dogger Bank represent a continuation of the North Sea tunnel valley network, interacting with both older glaciotectonic thrusting and younger glaciotectonic folded deformation. ~ 4 ~

6 Declaration The author declares that the contents of this thesis are her own work completed as part of the PhD project. None of the work referred to in this thesis has been submitted in support of an application for another degree or qualification at this, or any other, university or institute of learning. ~ 5 ~

7 Copyright Statement The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the Copyright ) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the Intellectual Property ) and any reproductions of copyright works in the thesis, for example graphs and tables ( Reproductions ), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see in any relevant Thesis restriction declarations deposited in the University Library, The University Library s regulations (see and in The University s policy on Presentation of Theses ~ 6 ~

8 Acknowledgements The author would firstly like to acknowledge NERC, BGS and the Forewind Consortium for funding of this PhD (NERC grant no: A87604X) and PGS, particularly Richard Lamb and Steve Morse; CGG; TGS and Forewind for supplying data. Thanks go to the project supervisors for their endless patience and assistance; Mads Huuse, Margaret Stewart and Simon Brocklehurst. The support of fellow students Rachel Harding, Andrew Newton, Astrid Ruiter and Carina Bendixen - all of whom collaborated on work presented in this thesis - was invaluable however the author would also like to acknowledge all those working in the Basins Research Group within the School of Earth, Atmospheric and Environmental Sciences and all those at the BGS working on the Dogger Bank Project, particularly Carol Cotterill who supervised all those working on the Dogger Bank. Thanks go to Heather Stewart, Mark Wilkinson and the rest of the proponents of the IODP GlaciStore bid for allowing the author to join the team and for finding the funding to allow a break in the PhD project to do so. Final recognition must go to the author s family for their endless support and patience. To Sue, Andrew, Sarah, Daniel, Ben and especially Hannah and Megan without you this project would not have even been started let alone finished. ~ 7 ~

9 Chapter 1: Introduction ~ 8 ~

10 The North Sea preserves a great deal of palaeo-environmental evidence from the early Pleistocene yet much of it remains poorly documented. Lack of regional-scale data, an accurate, detailed chronostratigraphy and uncertainty regarding the seismic stratigraphy has left large swathes of the early to middle Pleistocene under-studied. Using basin modelling and geological mapping techniques on basin-wide 3D seismic data this project provides palaeo-environmental reconstructions on a regional scale at a high resolution and spanning the early to middle Pleistocene. In order to capture glacial landforms resulting from grounded shelf glaciers (and to satisfy sponsor requirements) the thesis includes a study of late Pleistocene tunnel valleys in the Dogger Bank area. 1.1 Project Aims The principal aim of the project is to unravel the palaeo-environmental history of the North Sea basin during the Pleistocene with emphasis on glacial-interglacial changes and their imprint on the stratigraphy and subsurface geomorphology. 1.2 Project Outline and Objectives In order to fulfil the aims of the project a bottom-up approach to analysing the seismic data was taken by beginning with the early Pleistocene stratigraphy and moving upwards through the succession, and time, in order to analyse the evidence preserved in the basin. This was done through a series of objectives; 1. Identify and map in full three-dimensional detail the Plio-Pleistocene (i.e. base Quaternary) boundary using high resolution dating from the Dutch sector of the southern North Sea. 2. Using available absolute dates from the Dutch sector chronostratigraphy and palaeomagnetic data from BGS borehole studies from the UK sector to build a detailed chronostratigraphy for the early Pleistocene of the North Sea 3. Couple the chronostratigraphy with wireline log data to improve and refine the current stratigraphy for the early Pleistocene of the central North Sea. ~ 9 ~

11 4. Carry out regional-scale mapping of 3D seismic geomorphology through the Pleistocene stratigraphy of the central North Sea and place within the improved stratigraphy. 5. Carry out mapping and analysis of the tunnel valley system in the Dogger Bank Tranche A 2D seismic dataset 6. Interpret the combined stratigraphy and geomorphological mapping in terms of the evolving palaeo-environment of the central North Sea 1.3 Background Figure 1.1 Regional dataset map showing a) the location of seismic datasets across the North Sea and b) the location of wells and cores used in the project along with licence block boundaries (next page). Bathymetry and topography data from Ryan et al (2009). Full details of datasets used in the project can be found in Appendix 1. ~ 10 ~

12 The study of palaeo-climate and palaeo-environments during times of climatic change is of great importance to the understanding of the present and future climate (Alley 2000; Braconnot et al 2012; Rohling et al 2012b). Climate has been constantly changing over the lifetime of the planet and the evidence for it is preserved in the rock record (Veizer et al 1999; Alley 2000; Zachos et al 2001; Lisiecki & Raymo 2005). Studying the record of climate change and the influence it has had on the world can in fact help us to improve our predictive models of the future (Alley 2000; Braconnot et al 2012; Rohling et al 2012b; Martinez-Boti et al 2015). The Quaternary (2.58 Ma to present; Gibbard et al 2010) is an ideal target for studies of palaeoclimate and palaeo-environment because of the number of millennial- and centennial-scale climate records available (Voelker 2002). It has also been a period of incredibly dramatic climatic change characterised by the repetitive cycle of glacial and interglacial periods (Raymo 1994; Alley 2000; ~ 11 ~

13 Lisiecki & Raymo 2005, 2007; Steffensen et al 2008; Miller et al 2011). The repetitive glacialinterglacial cycles are observed most clearly on the δ 18 O marine isotope curve from which the Marine Isotope Stages (MIS), a standard way of denoting individual glacial and interglacial events in the Quaternary, are identified (Raymo 1994; Lisiecki & Raymo 2005, 2007). The onset of Northern Hemisphere glaciation during the Plio-Pleistocene transition (2.58 to 1.8 Ma; MIS ) as well as the Mid-Pleistocene transition (1.2 to 0.8 Ma; MIS 36-20) from 41 kyr glacialinterglacial cycles to the 100 kyr cycles of the late Pleistocene are two key features of the early Quaternary (Raymo 1994; Lisiecki & Raymo 2005, 2007; Gibbard & Lewin 2009; Miller et al 2011). It has been observed previously that mid- to high-latitude basins often preserve the largest variation in climatic trends (Mudelsee & Raymo 2005; Rohling et al 2012a) and where sedimentation rates are high often an expanded climatic record can be preserved. The North Sea then, as a mid-latitude marine basin, is perfectly positioned to have captured and preserved both the Plio-Pleistocene transition and the Mid-Pleistocene transition The North Sea The North Sea is an epicontinental sea bordered by three landmasses, the British, Scandinavian and Northern European. In the present day the North Sea basin averages at less than 200 m water depth outside of the Norwegian Trench (Figure 1.1). The present day configuration of the North Sea is a result of initial rifting during the mid-jurassic to Cretaceous associated with the unzipping of the North Atlantic followed by nearly continuous thermal subsidence and occasional basin inversion events during the early to middle Cenozoic (Ziegler 1992; White & Lovell 1997; Huuse 2002; Stoker et al 2005; Anell et al 2010; Goledowski et al 2012). The late Cenozoic has seen a decrease in tectonic events and the rapid infill of the basin (Huuse 2002; Nielsen et al 2009; Anell et al 2010; Goledowski et al 2012), primarily through the input of the proto-rhine and Baltic river systems which, in combination, drained large parts of the Scandinavian shield and Northern Europe ~ 12 ~

14 throughout the Pliocene and Pleistocene (Bijlsma 1981; Overeem et al 2001; Busschers et al 2007; Moreau & Huuse 2014). The North Sea has also been significantly shaped by the glaciations of the late Pleistocene, most notably in the form of the Norwegian Trench, a large elongate deep which runs parallel to the Norwegian coast displaying water depths of 700 m and the site of repeated ice streaming events during the mid-late Pleistocene (Sejrup et al 1996, 1998, 2003; Larsen et al 2000). The presence of glacial landforms on the sea bed such as partially buried tunnel valleys (Wingfield 1989; Ehlers & Wingfield 1991; Bradwell et al 2008) and large terminal moraine complexes (Bradwell et al 2008) influences the present day bathymetry, as do, for example, the large tidal bar complexes found most particularly in the southern North Sea (Caston & Stride 1970; McCave 1970; Terwindt 1971; Caston 1972; Otto et al 1990; Gatliff et al 1994). The Dogger Bank, a bathymetric high found between N and N, is thought to have been formed by the repeated build-up of glacial material during the last glacial maximum (LGM) resulting in a shallowing to a minimum of 18 metres water depth (Veenstra 1965; Cotterill et al 2012). In addition to latitude and longitude or Universal Transverse Mercator (UTM) coordinates there are three ways in which locations are described within the North Sea. Firstly the North Sea can be split into the northern, central and southern North Seas according to latitude and determined by the slightly different structural setting. The central North Sea, which sits over the central graben between N and N, is the focus of this project (Figure 1.1). The North Sea is also regularly classified by international boundaries according to the five bounding countries, the UK, Norway (NO), Denmark (DK), Germany (DE) and the Netherlands (NL) and traditionally much work done on the geology and stratigraphy by individual geological surveys for each country was done only to the sector boundaries with little cross-boundary correlation. This project has data which covers all of the five sectors although lies primarily in the UK and NO sectors (Figure 1.1). Finally the North Sea is also classified first into quadrants, then each quadrant into blocks following the hydrocarbon industry standard which is also used to name the wells according to ~ 13 ~

15 ~ 14 ~ Figure 1.2 Summary of central North Sea Quaternary stratigraphy including correlation to ICS, NW European and British Stages; global magnetic polarity; global sea level curve and biostratigraphic zonation

16 location. This project uses data from UK quadrants 14-16, 20-23, 28-31, and as well as NO quadrants 1-4, 6-10, 15-18, DK quadrants , , DE quadrant A and NL quadrant A (Figure 1.1b) Quaternary Stratigraphy of the central North Sea The Quaternary Period is formally defined as the period between 2.58 Ma and the present, split into the Pleistocene (2.58 to Ma; MIS 104-2) and Holocene (0.012 Ma to present; MIS 1) Epochs (Gibbard et al 2010; Gibbard & Cohen 2011; Cohen et al 2013). The Pleistocene Epoch is further split into the Early Pleistocene consisting of the Gelasian Stage (2.58 to 1.8 Ma; MIS ) and Calabrian Stage (1.8 to 0.78 Ma; MIS 61-20), the Middle Pleistocene (0.78 to Ma; MIS 19-6) and the Late Pleistocene (0.126 to Ma; MIS 5-2) (Figure 1.2; Gibbard et al 2010; Gibbard & Cohen 2011; Cohen et al 2013). The Quaternary stratigraphy of the central North Sea was defined and mapped initially during the 1970 s and 80 s by the British Geological Survey (BGS) through the interpretation of regional 2D seismic lines and shallow cores, generally less than 250 m in total length (Holmes 1977; Stoker 1983; Stoker & Bent 1985; Cameron et al 1987, 1992; Sejrup et al 1987, 1991, 1994; Knudsen & Sejrup 1993; Gatliff et al 1994). More recently, the Quaternary stratigraphy was revised by Stoker et al (2011) in order to more fully cross-correlate between the central, southern and northern North Sea stratigraphies as well as incorporate cross-border information (e.g. Gibbard et al 1991). The late Pleistocene seismic stratigraphy of the central North Sea is described as the Reaper Glaciogenic Group which comprises of multiple Formations and Members all of which have been glacially influenced (Figure 1.2) (Cameron et al 1992; Gatliff et al 1994; Stoker et al 2011). Many of the late Pleistocene Formations are locally distributed and lie uncomformably over the early Quaternary stratigraphy separated by the regional Ling Bank unconformity (Gatliff et al 1994; Stoker et al 2011). The Ling Bank unconformity and the Ling Bank formation together comprise the tunnel valley system which is spread across the entirely of the North Sea with the unconformity ~ 15 ~

17 representing the erosional base of the tunnel valleys and the formation the sediment infill (Cameron et al 1987; Wingfield 1989, 1990; Gatliff et al 1994; Praeg 2003; Stoker et al 2011; Stewart et al 2012; 2013). The Ling Bank tunnel valleys are generally considered to have a maximum age of 0.5 Ma from the Elsterian glaciation (MIS 12) but may also contain valleys carved during the Saalian (MIS 6-10) and Weichselian (MIS 2-4) glaciations although dates are extremely rare and ages are often assigned by analogy with adjacent areas which may or may not be robustly dated (Praeg 1996, 2003; Huuse & Lykke-Andersen 2000a; Lonergan et al 2006; Stewart et al 2012, 2013). The Pleistocene stratigraphy prior to the Ling Bank unconformity is grouped as the Aberdeen Ground Formation; a deltaic to shallow marine formation, which in seismic section appears as a series of large prograding clinoforms that is inclined seismic reflections representing the slope of the delta growing out into the basin. The Aberdeen Ground Formation transitions conformably from the clinoforms of the Pliocene Southern North Sea Deltaic Group. No physical boundary has been mapped to mark the basal Pleistocene across the basin (Holmes 1977; King 1983; Cameron et al 1987; Gatliff et al 1994; Stoker et al 2011). Very little evidence has been found for glacial activity within the Aberdeen Ground Formation prior to this project (Graham et al 2011) except the discovery of iceberg scouring events from the onset of global cooling at 2.7 Ma (Kuhlmann et al 2006; Stuart & Huuse, 2012; Dowdeswell & Ottesen 2013; Appendix 2). The initial mapping during the 70 s and 80 s which created the above descriptions was thorough, but limited by the wide spacing of the 2D lines and the restricted number and length of the cores, particularly in regards to the early Pleistocene. The growth of the hydrocarbon industry in the North Sea over the last few decades has led to a dramatic growth in the data available, particularly in regards to 3D seismic data, which has rapidly become an industry standard (Brzozowska et al 2003; Harvey et al 2010). Figure 1.3 demonstrates the step change in the 3D seismic coverage of the central and northern North Sea in just ten years allowing for a near continuous coverage of the central graben and areas immediately adjacent. ~ 16 ~

18 ~ 17 ~

19 Figure 1.3 (Previous page) Increase in 3D seismic data coverage comparing acquisition between 1970 to 1989 with 1990 to 1998 showing the near complete coverage of the central North Sea acquired by Reproduced from Brzozowska et al (2003). Although 3D seismic is usually acquired specifically to image the deep subsurface (below 1km) it has been proven useful for imaging the shallow section and following this increase in 3D seismic data acquisition, the late Pleistocene and its associated glaciations have been extensively studied (e.g. Praeg 1996, 2003; Graham et al 2007; Stewart & Lonergan 2011; Kristensen & Huuse 2012; Stewart et al 2012, 2013; Moreau & Huuse 2014). The early Pleistocene stratigraphy has received somewhat less attention, particularly at a regional scale, and is generally hindered by a lack of chronological control and detailed seismic stratigraphy. This project aims to use the now extensive 3D seismic data in order to begin to fill in some of the gaps in knowledge of the early Quaternary of the North Sea. The current stratigraphy will be refined and the growth of the ice sheets during the Plio-Pleistocene and Mid-Pleistocene Transition will be documented. The project aims to culminate with a palaeo-environmental framework from which further work into this time period can be conducted. 1.4 Project Funding, Datasets and Software The main PhD project was jointly funded by a National Environmental Research Council grant (no: A87604X) and the Forewind Consortium through the British Geological Survey s University Funding Initiative. Mapping of the stratigraphy (as detailed in Chapter 4) was partly undertaken in collaboration with an IODP proposal team (#852-CPP GlaciStore) through the University of Edinburgh in association with the British Geological Survey during a two month side project. The main datasets used in this project include: 3D Seismic o o PGS CNS MegaSurvey (50 m x 50 m x 4 ms) CGG BroadSeis TM Survey (12.5 m x 12.5 m x 4 ms) 2D Seismic ~ 18 ~

20 o o Forewind Round 3 Dogger Bank survey TGS North Sea Renaissance Survey Wells and Core Data o o TGS well data BGS shallow cores Figure 1.1 shows the location of the datasets used on this project and full details, including resolution and bandwidth, of all the datasets used in this project can be found in Appendix 1. Seismic resolution, as determined by seismic wavelength and bin spacing, is variable across the datasets from 8-16 m vertically and 50 m horizontally for the 3D MegaSurvey to approximately 1 m vertically and 5-10m horizontally with a 100 to 500 m line spacing for the 2D Dogger Bank survey. The majority of seismic interpretation for this project used Schlumberger s Petrel (2014 version) however the Dogger Bank survey was interpreted using IHS Kingdom (version 8.6). Specialist seismic interpretation software used for the project also includes Eliis Paleoscan (version 1.4), and ffa Geoteric (2013 version). Esri s ArcMap (version 10) was used intensively to map and interpret the observed seismic geomorphology. The TGS Facies Map Browser was used to access, view and download wireline-log data for over 1100 wells across the North Sea. 1.5 Methodology The use of deep 3D seismic data acquired for the petroleum industry for the analysis of stratigraphy and geomorphology is now well established in the Quaternary of the North Sea. Studies by Posamentier & Vail (1988), Praeg (1996, 2003), Lonergan et al (2006), Graham et al (2007), Stewart & Lonergan (2011), Kristensen & Huuse (2012), Stewart et al (2012, 2013) and Moreau & Huuse (2014) have shown how important consideration of three dimensions is for reconstructing basin evolution and glacial geomorphology. In using basin-wide 3D seismic data, supported by localised high resolution 3D, regional 2D and ultra-high resolution 2D lines and state of the art ~ 19 ~

21 Figure 1.4 Project Workflow bringing together various methodologies and processes in order to build the basin wide palaeo-environmental reconstruction ~ 20 ~

22 industry software, this project takes advantage of the step change in horizontal resolution between 2D and 3D seismic in order to build on these previous studies to achieve a much larger scale, yet detailed interpretation of the evolution of the entire basin during the Quaternary at the highest possible spatial and temporal resolution Seismic Mapping The project generally followed the workflow model outlined in Figure 1.4. Seismic data were first tied to known chronostratigraphic markers from wireline log and core data, namely palaeomagnetic data from Stoker et al (1983), and integrated studies on Dutch Wells, primarily A15-03 from Kuhlmann et al (2006). Following this, mapping of key horizons in both 2D and 3D seismic data was undertaken using standard techniques for the stratigraphical analysis of seismic data (Mitchum et al 1977a, b; Posamentier et al 2007). For 3D data this was primarily done in Petrel using a combination of manual, guided 2D, seeded 2D and seeded 3D auto-tracking. Initially a regular grid of 2D lines, extracted from the MegaSurvey 3D, was used moving outwards from the correlation points and mapping down depositional dip of the clinoforms where possible. By mapping down dip errors are minimised allowing for the greatest confidence in picked surfaces (Mitchum et al 1977a, b; Posamentier et al 2007). After completion the surface was interpolated between control points using 3D auto-tracking; problem areas were identified, erased and reinterpreted using a smaller grid of 2D lines. The work flow was repeated until a highly detailed, high accuracy horizon was achieved and quality control, in the form of independent composite sections back to the original correlation point were used to ensure this. The horizon was then converted to a surface using a minimum curvature method and minor corrections were made to the interpretation and the surface as needed. Finally a light smoothing up to a maximum of 3 iterations with a 2x2 cell filter width (filter width determines degree of smoothing per iteration and a larger filter width results in a greater degree of smoothing) was applied. ~ 21 ~

23 Mapping of the regional 2D lines was also done in Petrel, choosing a regular spacing of 10 x 15 km and using the same tools but without the 3D interpretation. The Dogger Bank ultra-high resolution 2D seismic was interpreted using the Kingdom suite seismic software. Mapping with this data set was restricted solely to the tunnel valleys and their fill meaning all interpretation was done using a manual interpretation tool. The density of the 2D lines made it impractical to interpret every line and so a regular grid workflow was once again applied, starting with a grid of 500 m in-line by 1km x-line and reducing to 100 m x 500 m where necessary, to the data to allow full coverage of the study area. Interpretation of basin-wide 3D surveys using this method to map each surface accurately is time consuming. In order to fully interpret the basin-wide data a semi-automated mapping technique was used to investigate between key, fully correlated surfaces. The Paleoscan software produced by Eliis allows for semi-automated mapping of very large 3D seismic datasets. The software first builds a geological model from the seismic data which is refined by the use of marker horizons produced either in the software or imported and edited as necessary. The geological model can be refined as much as is required to ensure high quality control on the resulting horizons. However, where strongly faulted or dipping reflections exist (e.g. elongate troughs seen in the Quaternary succession of the North Sea: see Chapter 3) the software often cannot accurately represent the reflections. Once the geological model is finalised a horizon stack can be created between any two reflections, automatically creating as many as 100 horizons. The created horizons may be structural (two-way time (TWT) only) or may have attributes such as instantaneous amplitude and Root Mean Squared (RMS) amplitude extracted with them. The horizons in the horizon stack can then be exported, individually or in groups, into other software, in this case Petrel, for interpretation. In this way 14 major horizons were mapped using the traditional method and 63 intermediate horizons were mapped using the Paleoscan software. Due to the process involved intermediate ~ 22 ~

24 horizons cover the entire study area, even where a horizon onlaps or downlaps onto others, which is taken into consideration during interpretation. Major horizons are mapped across the study area but terminate where the reflections downlap onto lower horizons. Additional minor horizons mapped across specific study areas (see Chapter 3) Chronostratigraphic correlation Chronostratigraphic correlation was taken in the first instance from robust studies of a variety of chronological proxies across the North Sea, including palaeo-magnetic data, biostratigraphic analysis, pollen analysis and sequence stratigraphy. The different sources of these data sets and their associated error ranges are identified in Figure 1.5. The principal chronostratigraphic ties are associated with three core/well sites. The A15-03 well (55º18 N, 3º48 E) in the Dutch sector of the southern North Sea has been closely studied by Kuhlmann et al (2006) and Noorbergen et al (2015) with respect to Quaternary chronology. It has also been tied directly with other studies from the southern North Sea such as Meijer et al (2006), Anell et al (2012), ten Veen et al (2013) and Thöle et al (2014) by Harding (2015) making it a robust dataset from which to establish a central North Sea chronology. The integration of multiple types of chronostratigraphic proxies and direct ties to gamma-ray logs from the well itself minimises data variability from the depth conversion, as the well has measured velocities and thus calibrated time-depth curves, as well as error margins from interpretation of single proxies such as biostratigraphy due to the overlapping of different proxies. This is particularly the case in A15-03 where the 2.58 Ma surface can be identified with a narrow error margin due to the onset of the Ericales pollen acme, the Gauss palaeo-magnetic Reversal and the last common occurrence of benthic foraminifera M. pseudotepida which all describe the degradation of the climate associated with the onset of the Quaternary (Figure 1.5; Kuhlmann et al 2006). The Josephine-1 core (56º36.11 N, 2º27.09 E) has been widely used in the central North Sea to tie horizons and wells to a chronology and the biostratigraphic data is published in Knudsen & ~ 23 ~

25 Figure 1.5 Well correlation panels for 77/02, Josephine-1 and A15-03 showing the identification of chronostratigraphic ties, the data variability and error margins associated with each data set. Depths for seismic panels are ms/m below sea level, depths for palaeomagnetic/biostratigraphic data are below seabed. Conversion derived from water depths detailed on original log reports for each well, 147 m at 77/02 and 72 m at Josephine-1. Palaeomagnetic data for 77/02 redrawn from Stoker et al (1983), biostratigraphic data for Josephine-1 redrawn from Knudsen & Asbjörnsdóttir (1991) and multi-proxy chronostratigraphic data for A15-03 redrawn from Kuhlmann et al (2006). ~ 24 ~

26 Asbjörnsdóttir (1991). Biostratigraphic data is often subject to differing interpretations, for example although in Josephine-1 the first common occurrence of benthic foraminifera is taken to be at 700 m below sea level (Figure 1.5; Knudsen & Asbjörnsdóttir 1991; Buckley 2012) however it could easily be interpreted within a ~50 m window. However by using 3D seismic data and seismic stratigraphical principles a direct tie can be made between Josephine-1 and the multi-proxy approach of A15-03 which minimises the error margin and allows for mapping between the two sites with a reasonable degree of confidence. Finally the core 77/02 ( N, E) has longest and best established palaeo-magnetic record for the Quaternary reaching down to the Jaramillo palaeo-magnetic Event (Stoker et al 1983) allowing for the shallower section of the Quaternary deposits to be dated and the correlation mapped southwards. Palaeo-magnetic data is also subject to interpreter bias, as with biostratigraphy, where the natural variance of the data can create variability in the interpretation. In the case of 77/02 while the base of the Jaramillo Event is clearly defined with little room for variability the Brunhes-Matuyama Reversal has a larger variability (Figure 1.5; Stoker et al 1983). Palaeo-magnetic data however is generally acquired on a much higher resolution than the available seismic and, in the case of 77/02, the variability of the Brunhes-Matuyama Reversal is encapsulated within only two seismic reflections allowing for mapping with a reasonable degree of confidence Gamma-log correlation Although chronostratigraphic ties were used for the mapping of some of the major horizons the confirmed dates for the early Quaternary are few and widely spaced in time. In order to build an accurate stratigraphy and palaeo-environmental reconstruction a more detailed chronostratigraphy was needed. As few scientific boreholes penetrate deep enough into the early Quaternary stratigraphy an alternative method, relying on the relationship between sediment grain size and gamma logs was tested as part of this study. This method (outlined below and used in Chapter 4) proposes a causal link between cyclicity in the gamma-ray logs in the central North Sea and global sea level; this method follows observations of the link between grain size and gamma in the southern North Sea and onshore Denmark (Noorbergen et al 2015; Harding 2015). ~ 25 ~

27 On the slope of the clinoform, at the point of maximum deposition, a peak in the gamma log generally represents a fining of grain size which has been noted, in the North Sea, to correlate to a fall in sea level. The proposed model for this relationship is that during lowstands more physical weathering due to glacial activity in the highland source areas, namely Scandinavia, coupled with long transport distances across large delta plains reduces the coarse fraction of sediment reaching the clinoform shelf and thus the main depocenter (Noorbergen et al 2015). During highstands the reduced influence from glacial influences in general and Scandinavia in particular couples with the increasing influence from the rivers draining northwest Europe and stronger bottom currents reworking sediment in the basin leads to the coarser sediment fraction Figure 1.6 Chronostratigraphic calibration panel correlating gamma-ray log cyclicity and seismic data at well NO 2/7-31 to the global sea level curve from Miller et al (2011) in order to date seismic reflections. Adapted from Chapter 4 Figure 4.4 being preferentially deposited (Noorbergen et al 2015). Seven wells were chosen specifically to coincide with the areas of maximum deposition between each of the major horizons and the gamma-ray logs compared to the global sea level curve from Miller et al (2011) as in Figure 1.6. The relationship between the gamma-ray and sea level was ~ 26 ~

28 used to systematically correlate peaks in gamma-ray to sea level lows between three absolutely dated horizons, extrapolated from chronostratigraphy in the Dutch sector. The concept appeared to be robust within the limitations of the available data the number of lowstands in the sea level curve corresponded to the number of peaks in the gamma-log between the dates and was subsequently applied to surfaces that did not have chronostratigraphic ties, thus constructing a significantly more detailed chronology than was previously available for the early Quaternary (see Chapter 4 for results) Depth Conversion All the mapped surfaces were depth converted to allow them to be used for calculations of both sedimentation rate and basin reconstruction through back-stripping. Check shots from 1122 wells were available in the Facies Map Browser to produce a depth conversion function (time-depth relationship). It was considered that while velocity modelling would be more accurate the regional scope and relatively deep (Quaternary) focus of this project and the intended uses of the depth converted surfaces meant that a more straightforward model would be acceptable. The check shot data from the wells were collated into one plot of depth versus two way time and a quadratic regression curve plotted through it with a very high R 2 correlation value of 0.99 (see Chapter 2 Figure 2.3). The quadratic rather than linear equation was used as the curve of the plot made it clear that a linear relationship although also with a high correlation value of 0.97 was less representative. Depth (m) = 81.0 TWT(s) TWT(s) 1.094(±100 m) [1] Once the regional depth conversion formula was calculated the mapped surfaces were depth converted and used for further calculations. ~ 27 ~

29 1.5.4 Sedimentation rate and Basin Reconstruction In order to back-strip the sediment infill and thus reconstruct the basin sediment packages are decompacted using a standard sediment compaction curve. In a similar way in order to calculate sedimentation rates a standard sediment compaction curve is used in order to fully compact a partially compacted sediment volume to its solid sediment volume, in which there is zero percent porosity. Calculations for both back-stripping and sedimentation rates follow two slightly different approaches to account for their different usage. In calculating sedimentation rate the partially compacted volume between each of the 63 intermediate horizons was calculated for conversion to solid sediment volume. The most accurate method of conversion would be to use porosity data from well logs to produce a specific depthporosity curve for the shallow section and then use this, along with average depths for each of the surfaces, to calculate the volumes and thus the sedimentation rate. However, while many wells were available for the project few had data in the shallow section (<500 m) beyond gamma logs. Due to lack of more detailed information, a standard depth-porosity curve from Marcussen et al (2010) was used in conjunction with calculated average depths from 15 partially compacted volumes to create an estimated porosity-twt regression curve in keeping with the North Sea sedimentary environment. The regression curve was used to recalculate new values for the solid sediment volume of the 15 partially compacted volumes, plus an additional 5 volumes as a control. The solid sediment volume from the depth-porosity curve was compared to the solid sediment volume from the regression curve in order to find the error margin for this method. An error margin of less than 5% was calculated; considered to be reasonable for calculating sedimentation rates. The regression curve was used then to find the solid sediment volume of all 63 sediment packages and thus calculate sedimentation rate. The results of these calculations can be found in Chapter 4. The back-stripping process uses a very different method for decompaction of sediment, based on the methodology of Allen & Allen (1990) and summarised in the workflow in Figure 1.7. This process involves moving the bounding surfaces of a stratigraphic layer upwards to a pre-defined ~ 28 ~

30 ~ 29 ~ Figure 1.7 Decompaction workflow adapted from Allen & Allen (1990) identifying a) decompaction equations and constants derived from measured North Sea values (Allen & Allen 1990) and b) graphical representation of the workflow in which sediment packages are progressively decompacted and removed allowing the sediment column to be adjusted for Airy Isostasy

31 height in this case present day sea level and using the decompaction equation (below) to find the new thickness of sediment at that point. y 2 y 1 = y 2 y 1 ɸ 0 c {e( cy 1 ) e ( cy 2 ) } + ɸ 0 c {e( cy 1 ) e ( cy 2 ) } [2] Where y 1 is the depth to the top of the sediment layer; y 2 is the depth to the bottom of the sediment layer; y 1 is the depth to the top of the decompacted sediment layer, y 2 is the depth to the bottom of the decompacted sediment layer; ɸ 0 is the porosity of the sediment at the surface and c is a constant that defines the curve representing the change of porosity with depth. The two constants in this equation are lithology dependent and based on standard North Sea values quoted by Allen & Allen (1990), y 1 and y 2 are determined by the seismic depth conversion, y 1 is set at zero or present day sea level, and y 2 is calculated iteratively using multiple values of y 2 until the two sides of the formula balance and the result becomes the proxy for de-compacted thickness and thus volume. The workflow indicated in Figure 1.7 requires the shallowest sediment layer to be decompacted and then the volume removed and the stratigraphy below adjusted using a standard Airy isostasy calculation. Y = (ρ m ρ s )/(ρ m ρ w ) S [3] Where Y is the adjusted height of the surface, ρ m is the bulk density of the mantle, ρ s the bulk density of the sediment, ρ w the density of water and S is the thickness of the sediment package above the surface including the water column. The above methodology was used in this study based on single data points only with multiple points along transects combined to reconstruct the basin with several restrictions: firstly it assumes homogeneity of sediment, although testing of the methodology suggests that the error margin for the lithology dependent constants is low (<10%); secondly it does not account for eustasy, referring to present day sea level only; finally it does not account for flexural effects and sediment layers are ~ 30 ~

32 back-stripped at points which do not influence one another. The methodology used does reasonably reconstruct the basin depth through time within 100 m, the results of which can be seen in Chapters 2 and Geomorphological Mapping The geomorphological mapping of features preserved in the seismic stratigraphy was dependent on their expression within the seismic data. Some features such as the elongate troughs detailed in Chapter 3 and the tunnel valleys detailed in Chapter 5 have a good vertical expression in crosssection allowing for them to be mapped using manual interpretation tools in the seismic interpretation software using 3D time-slices, where available, to ensure that mapping followed the planform geometry as well as the cross section. However many features have vertical expressions at or below seismic resolution, such as channel systems and ice berg scour marks, and even those with vertical expression, such as glaciotectonics, can be difficult to identify in large datasets. While horizontal time-slices have been used successfully to image and map glacial landforms in relatively flat lying strata within 3D seismic data (e.g. Praeg 1996, 2003; Graham et al 2007; Stewart et al 2012, 2013), in areas with strong clinoformal geometries this can lead to errors. As clinoform surfaces are considered to be time-synchronous (Mitchum et al 1977a, b; Posamentier et al 2007; Harding 2015) time-slice imaging in 3D seismic data cuts directly through several clinoforms at once creating an image that, though showing aspects of the plan-form geometry, is not temporally synchronous (Figure 1.8). For this reason horizon slices are used, such that they follow the clinoform geometry and thus represent a single moment in time (Figure 1.8), the duration of which depends on the seismic resolution and sedimentation rate. Suites of features from the same time but different depositional settings can be considered as one palaeo-environmental setting for example terrestrial channels on topsets linked directly to along-slope contourites. Geomorphological features observed in the 3D seismic data were mapped largely in the ArcMap software. In this project, high resolution images of horizons with either instantaneous amplitude or ~ 31 ~

33 Figure 1.8 a) Seismic section comparing the truncation of clinoforms by time slices compared to a horizon slice. b) Horizon slice plan view with instantaneous amplitude extraction showing relationship between iceberg scours and along slope features. c) time slice at 700 ms showing truncated clinoforms and iceberg scours. d) time slice at 824 ms showing truncated clinoforms and along-slope features. ~ 32 ~

34 RMS amplitude attribute extractions were created using screenshots and combined in an image composite editor. These high resolution images were then imported from Petrel into ArcMap and fully georeferenced. Mapping and interpretation of geomorphological features could then be conducted swiftly and efficiently. Features were mapped first according to their classification then the horizon upon which they occur. Where features persisted across multiple horizons the occurrence was noted for later correlation to the chronology Palaeo-environmental Reconstruction The different aspects of the methodology were brought together in order to reconstruct the palaeoenvironments of the North Sea. Although all outputs from the workflow including facies analysis from well-log data, chronostratigraphy, basin reconstruction and sedimentation rate were considered the major input into the reconstructions was the geomorphological mapping. Landform models (e.g. Stow & Mayall 2000; Phillips et al 2008; Ottesen & Dowdeswell 2009; Stuart & Huuse 2012) were used in order to understand how suites of different geomorphological features formed at approximately the same time can be linked together to build a model of the palaeoenvironment. Figure 1.9 demonstrates several landform models used to understand the relationships between different features and thus the processes involved in forming them. For example downslope processes, such as those described in Figure 1.8a, are observed in the early Quaternary (Chapters 3 & 4) however the model for the mud dominated shelf suggests that downslope processes would create a system of terminal lobes at the end of the valley canyons (Figure 1.9a). This is not observed in the North Sea, only valley canyons are observed as elongate troughs (described in Chapter 3). Landform models then are useful in determining palaeogeographies but often represent idealised palaeo-environments. The relationship between iceberg scours and along-slope contourite sands shown in Figure 1.9b is often observed within the Quaternary of the North Sea (Figure 1.8; Chapter 4), indicating a ~ 33 ~

35 ~ 34 ~ Figure 1.9 Landform models used to assess formation of geomorphic features and used for palaeoenvironmental reconstructions a) mud-dominated shelf system redrawn from Stow & Mayall (2000) b) current influenced glaciated shelf redrawn from Stuart & Huuse (2012) c) ice-marginal to pro-glacial deformation adapted from Phillips et al (2008) and d) inter-icestream glacial retreat assemblage adapted from Ottesen & Dowdeswell (2009)

36 persistent glaciomarine palaeo-environment. Figures 1.9c and 1.9d both represent subglacial conditions, the first based on observed glaciotectonic deformation and the second based on streamlining of subglacial bedforms in the direction of ice flow indicating ice streaming behaviour. The aim of the palaeo-environmental reconstructions is to provide an initial framework from which further studies can expand or add detail in order to more fully understand the climatic and environmental evolution of the early Quaternary. 1.6 Summary of Chapters The main thesis consists of four first author papers organised by chronostratigraphic age outlining the Pleistocene of the central North Sea from the initiation of the Quaternary as a period of dramatic global cooling through the early part of the Quaternary North Sea as a deep marine basin, until infill and the impact of ice sheets. The four papers are summarised below along with individual author contributions. The final chapter of the thesis consists of a synthesis of the work completed and summarises the conclusions of the project Chapter 2: The early Quaternary North Sea Basin A combination of full 3D coverage and a strong tie to recent integrated chronostratigraphies from the Dutch sector of the southern North Sea allows for a new interpretation of the base Quaternary to be identified across the southern and central North Sea. This basal Quaternary surface is presented in Chapter 2. The new interpretation suggests the base Quaternary surface as a marine basin 600 km long, >100 km in width and with a maximum water depth of ~350 m. This interpretation reveals a sediment package around 1.2 km thick, adding upwards of 600 metres of stratigraphy to the early Quaternary, with sedimentation rates estimated as >2.5 mm yr -1 during a period of dramatic global climate change, thus preserving an expanded climate record comparable to any other for this period worldwide. This redefinition of the early Quaternary has strong implications for both climate studies and basin evolution. ~ 35 ~

37 This manuscript has been submitted to Quaternary Science Reviews Chapter 3: Early Quaternary sedimentary processes and palaeo-environments in the central North Sea In Chapter 3 an interpretation of large, elongate trough-like features found in the earliest Quaternary is presented. These features have been studied previously however each of the different studies has inferred a different age and/or processes involved in their formation (Cartwright 1995; Knutz 2010; Kilhams et al 2011; Buckley 2012). Based on the new basal Quaternary as presented in chapter 2, the trough features are considered to be earliest Quaternary in age. In order to understand the relevant processes in the early Quaternary, the larger basin-wide dataset is examined to understand more fully the palaeo-environmental setting. Chapter 3 therefore maps every identified trough, collating trough length, width, orientation, fill and shape in order to critically examine the previously proposed models of formation. 380 troughs were mapped in total ranging from 0.1 to 50 km in length and 0.1 to 3.5 km in width. The troughs are separated into types 1 and 2 in which type 1 troughs are found to be smaller and uniformly orientated perpendicular to the strike of the clinoform slope. Type 2 are wider and deeper than type 1 and are observed to be closer to parallel with the strike of the clinoform slope. Troughs can also be separated by cross-sectional form and by age. The formation of the troughs is interpreted to most likely be related to downslope processes on the clinoform slope but it is proposed, due to the change in orientation and size, that type 2 troughs are further modified by shallow along-slope currents. This manuscript has been submitted to a special publication of Journal of Quaternary Science and is under revision. ~ 36 ~

38 1.6.3 Chapter 4: The evolving palaeo-environments of the Pleistocene of the central North Sea: 2.58 Ma to 0.78 Ma Chapter 4 uses the chronostratigraphic ties identified in Chapter 2 in order to build up a more detailed stratigraphy of the early and middle Pleistocene by correlation to preserved cyclicity in the gamma logs identified in Chapter 3. This study suggests that evidence from well logs in the areas of highest deposition allow for correlation to every Marine Isotope Stage in the Quaternary, thus allowing for a relatively precise dating of every seismic reflection. The detailed stratigraphy constructed here is then used to date a wide variety of seismic geomorphological features preserved in the subsurface in order to build-up a picture of the changing palaeo-environments of the early to middle Pleistocene. Geomorphological features are mapped, identified and interpreted across 63 basin scale seismic horizons correlated to the constructed chronostratigraphy. The suites of geomorphological features, along with calculations of average sedimentation rate and basin evolution through back-stripping models, are used to produce maps of the evolving palaeo-environment. The palaeo-environmental maps reveal a deep marine basin strongly influenced by currents with marine-terminating ice sheets from the very earliest Quaternary, followed by the shallowing of the basin through time and subsequent weakening of marine currents and the advancement ice sheets further into the basin. Evidence for large ice sheets grounded in the central North Sea in the form of mega-scale glacial lineations is observed from 1.7 Ma becoming common by 1.3 Ma Chapter 5: Tunnel valleys of the Dogger Bank from ultra-high resolution 2D seismic data The use of middle Pleistocene tunnel valleys in order to identify the progression of ice margins and ice flow direction is now well established thanks to extensive regional mapping. The North Sea comprises one of the most intensive networks of tunnel valleys from the middle to late Pleistocene anywhere in the world, but questions remain about the full extent of the valleys, their formation method and fill. Chapter 5 uses new ultra-high resolution seismic data (~ 1m vertical resolution) ~ 37 ~

39 acquired for windfarm development to fill in one of the holes in the map of North Sea tunnel valleys located outside the area of 3D seismic coverage on the Dogger Bank bathymetric high. 31 tunnel valleys have been identified and mapped and the geometry of the valleys and the fills analysed. The tunnel valleys of the Dogger Bank are generally representative of tunnel valleys across the North Sea, and are seen to cross-cut one another, displaying a minimum of seven crosscutting generations, similar to other tunnel valley studies (Kristensen et al 2007; Stewart & Lonergan 2011; Stewart et al 2013). The infill of the Dogger Bank tunnel valleys consists of three distinct packages, one of which is clinoformal in nature in at least three valleys, which agrees with the structure of other tunnel valley studies but is far more homogenous in the Dogger Bank tunnel valleys. However the Dogger Bank valleys do vary from the North Sea average in two ways. Firstly they are generally shorter than most North Sea tunnel valleys, appearing in many cases to represent fragments of tunnel valleys which have subsequently been destroyed or eroded by later ice advances. Secondly the Dogger Bank tunnel valleys are observed to interact with glaciotectonic features. Large scale thrusts are seen to influence the distribution of tunnel valleys from an older ice sheet advance. In the heavily deformed shallow section, above tunnel valley infills, the style of small scale deformation changes, proposed to be caused by changes in subglacial drainage Author Contributions Summarised below are individual author contributions for each research chapter. In chapter 2 the first author; Correlated the basal Quaternary and other dated horizons from the A15-03 well northwards to the Josephine-1 well and 77/02 core, as well as mapped the horizons across the central North Sea portion of the study area. Identified and mapped the basal Quaternary sub crop. Performed depth conversion, decompaction & back-stripping calculations. Was primary contributor to writing of the manuscript. Other authors; ~ 38 ~

40 R.H. correlated the basal Quaternary and other dated horizons from the A15-03 well with other southern North Sea chronostratigraphic studies and mapped horizons across the southern North Sea portion of the study area R.H., M.H., M.S., and S.H.B. aided in the conceptualization of the manuscript, quality checking of mapping and calculations and in reviewing and editing the manuscript. In chapter 3 the first author: Identified the need to reinvestigate the trough features with respect to new data availability and a wider palaeo-geographic context. Identified and mapped all troughs and horizons used to investigate the basin. Extracted all geometries including length, width, depth and orientation. Interpreted the features with respect to the palaeo-geographic context. Was primary contributor to writing the manuscript. Other authors; M.H. and M.S. were both involved in interpretation and critical examination of the proposed model. M.H. and M.S. both aided in review and editing of the manuscript. In chapter 4 the first author; Mapped major bounding horizons and produced the semi-automated geomodel from which intermediate horizons were extracted. Correlated horizons to the established chronostratigraphy for the Quaternary as far as was possible and derived the new methodology for age estimates between these dates using gamma-log correlation. Performed depth conversion, decompaction, back-stripping and sediment deposition rate calculations. Revised the existing seismic stratigraphy with respect to the new data and revised palaeogeographic context. Mapped the seismic geomorphology and interpreted it with respect to the palaeoenvironmental reconstruction. Was primary contributor to writing the manuscript. The other authors; M.H. assisted in quality checking the chronostratigraphic correlation for accuracy with respect to the seismic stratigraphy. ~ 39 ~

41 M.H., M.S. and S.H.B. were involved in the conceptualization of the manuscript as well as reviewing and editing of the manuscript. In chapter 5 the first author; Mapped in full all tunnel valleys across the data set, as well as infill packages. Extracted geometries of tunnel valleys including length, width, depth and orientation as well as calculated clinoforms dip angles. Identified distribution patterns and tunnel valley generations. Was involved in the interpretation of the tunnel valleys with respect to glaciotectonic deformation. Was primary contributor to writing of the manuscript. Other authors A.R. identified and mapped the glaciotectonic deformation packages and was involved in the interpretation of the glaciotectonic complexes with respect to the tunnel valleys. M.S. was involved in the conceptualization of the manuscript and critical examination of both results and suggested models. A.R., M.S., M.H. and S.H.B. were involved in the quality checking of results, consideration of the interpretations as well as reviewing and editing of the manuscript. ~ 40 ~

42 Chapter 2: The early Quaternary North Sea Basin Rachel M. Lamb, Rachel Harding, Mads Huuse, Margaret Stewart, Simon H. Brocklehurst Abstract: The base Quaternary (2.58 Ma) surface has been mapped in detail across the entire central and southern North Sea Basin for the first time, adding > 0.5 km thickness to the Quaternary stratigraphy of the North Sea with respect to previous maps of the central North Sea. The new base Quaternary map is based on a modern database including MegaSurvey 3D seismic datasets and regional 2D seismic data covering the entire basin and calibrated by chronostratigraphic studies from the Dutch and UK North Sea. The new map defines a deep, elongate marine basin over 600 km in length and 100 km wide orientated NNW-SSE. A shallow water connection to the northeast Atlantic is located at the northern end of the basin while early Pleistocene palaeo-water depths in the main basin are estimated at about 300 m. Extremely rapid sedimentation rates (2.6±0.5 mm yr -1 ) in the earliest Pleistocene resulted in an expanded palaeoclimate archive for the early Pleistocene. This archive preserves the record of a critical period of global climatic cooling in an important mid-high latitude epicontinental basin fed by continental scale drainage systems and connected to the global ocean. The revised early Quaternary framework has important implications for understanding basin infill, subsidence, shallow reservoirs, and hazards to hydrocarbon exploration and development. ~ 41 ~

43 2.1 Introduction The onset of cooling at the Plio-Pleistocene transition and the onset of widespread Northern Hemisphere glaciation are important markers for palaeo-climate and palaeo-environmental studies worldwide (Raymo 1994; Lisiecki & Raymo 2007). Climate records often vary strongly with latitude, and cooling trends such as the Plio-Pleistocene transition are often best preserved in midto-high latitude basins (Mudelsee & Raymo 2005; Rohling et al 2012). In the North Sea Basin, a mid-latitude epicontinental sea, understanding the nature and extent of the consequences of the early Pleistocene cooling has previously been difficult due to the poor definition and identification Figure 2.1 Location map and datasets. Data shown are the PGS Central North Sea and Southern North Sea 3D seismic MegaSurveys, the TGS North Sea Renaissance 2D lines and the locations of the Josephine-1 and A15-03 wells used for dating of the basal Pleistocene. Bathymetry and topography data from Ryan et al (2009). Location of Figure 2.2 seismic section indicated by the dashed line. ~ 42 ~

44 of the base Quaternary boundary. It has long been known that there is a considerable thickness of Quaternary deposits in the central North Sea (CNS) (Holmes 1977; Cameron et al 1987; Gatliff et al 1994). However the discrepancies among lithostratigraphic and chronostratigraphic studies in the five bordering countries, combined with spatio-temporal changes in climate and sediment supply (Huuse 2002) and a strong glacial overprint, have impeded the accurate mapping of the basal Quaternary across the basin (Cameron et al 1987). Additionally, in 2010 the global definition of the base Quaternary was changed from 1.81 (Marine Isotope Stage 65 (MIS 65)) to 2.58 Ma (MIS 104), adding 0.77 Ma of stratigraphy to the Pleistocene (Gibbard et al 2010). Regional 3D seismic data acquired for deeper petroleum exploration has been used previously for glacial geomorphological studies on local to sub-regional scales, highlighting the step change in horizontal resolution provided by these data and allowing previously hidden insights into the structure and stratigraphy of the Late Cenozoic succession (e.g. Praeg 1996, 2003; Stewart & Lonergan 2011; Kristensen & Huuse 2012; Stewart et al 2012, 2013; Moreau & Huuse 2014). This study documents the shape and evolution of the earliest Quaternary North Sea basin through integration of regional basin-scale 3D seismic data with recent biostratigraphic studies from the southern North Sea to accurately map the Plio-Pleistocene boundary across the entire central and southern North Sea (CNS, SNS) for the first time. 2.2 Regional Setting The present day North Sea basin is an epicontinental sea reaching average water depths of 100 m (outside the Norwegian Channel) and bordered by the Northwest European, Scandinavian and British landmasses. The North Sea originated during episodic extensional rifting related to the opening of the Atlantic from the Palaeozoic until the Early Cretaceous (Ziegler 1992). This was followed by continuous subsidence throughout the Late Cretaceous and Cenozoic punctuated by basin inversion episodes during the Paleogene (Ziegler 1992; White & Lovell 1997; Stoker et al 2005). The role of regional and local tectonic activity into the late Cenozoic is still debated although most evidence favours a model of passive thermal subsidence enhanced by sediment ~ 43 ~

45 loading in the CNS and marginal uplift due to denudation and unloading of the Norwegian landmass (Huuse 2002; Nielsen et al 2009; Anell et al 2010; Goledowski et al 2012). In previous studies the early Quaternary stratigraphy in the North Sea has been mapped as one unit, the Aberdeen Ground Formation, of pro-deltaic to marine sediments (Gatliff et al 1994; Stoker et al 2011), mostly sourced from the Rhine-Meuse and Baltic river systems of northern Europe and Scandinavia (Bijlsma 1981; Busschers et al 2007; Moreau & Huuse 2014;). The timing of onset of glaciation(s) in the Quaternary remains unclear, although there is growing evidence for several ice sheet advances during the Middle-Late Pleistocene, sourced from the UK and Scandinavian landmasses (Graham et al 2011) and evidence for iceberg scouring even in the SNS Basin since the onset of the Quaternary (Kuhlmann et al 2006; Stuart & Huuse 2012; Dowdeswell & Ottesen 2013). 2.3 Data and Methods This study uses seismic stratigraphical and seismic geomorphological techniques (Mitchum et al 1977a, b; Posamentier et al 2007) to analyse a basin scale 3D seismic dataset covering 128,000 km 2 of the CNS and SNS. The study uses the continuous PGS CNS and SNS 3D seismic MegaSurveys with a sub-sampled bin size of 50 m and sampling rate of 4 ms TWT to map the deepest part of the Quaternary North Sea basin (Figure 2.1). In areas not covered by the MegaSurvey, regional 2D seismic lines were used to resolve the basin shape. In the top 1.5 s TWT, the vertical resolution of the MegaSurveys is 8-16 m while the vertical resolution of the 2D lines is between 10 and 18 m with varied line spacing between 1 and 15 km. The base Quaternary stratigraphic surface was correlated to a well-defined and continuous seismic reflection using the Dutch North Sea well A15-03 (55º18 N, 3º48 E; Figure 2.1) for which very detailed bio- and magneto-stratigraphic dating has been performed (Kuhlmann et al 2006; see discussion in section 2.4). The horizon was mapped down the depositional dip of the clinoforms into the basin, which minimised correlation errors to about half the dominant ~ 44 ~

46 Figure 2.2 Seismic section and line interpretation, at 75x vertical exaggeration, showing the location of Josephine-1 and A15-03 wells plus surfaces for the mid- Miocene (c Ma), 2.58 Ma (base Quaternary), 2.35 Ma, 1.9 Ma and 1.1 Ma (base-jaramillo palaeo-magnetic Event). Key biostratigraphic events for both wells highlighted after Knudsen & Asbjörnsdóttir (1991) and Kuhlmann et al (2006) with error range from depth conversion. Location in Figure 2.1. Data courtesy of PGS. ~ 45 ~

47 wavelength (20-35 m), and the correlation compared to the biostratigraphic record of the Josephine-1 well (Knudsen & Asbjörnsdóttir 1991), which was commonly used as a correlation point for studies of the UK Continental Shelf (UKCS) Quaternary (Figure 2.2). Calibrated TWT depth data from 1122 wells from the UK and Norwegian North Sea were plotted to find a simple but robust depth conversion equation (Equation 1) with defined data variability (Figure 2.3). Depth (m) = TWT(s) TWT(s) (±100 m) [1] Several transects representing the overall structure of the basin were identified for calculating approximate palaeo-water depths from the heights of clinoforms infilling the Quaternary basin. In order to provide realistic water depth estimates from buried clinoforms, they need to be backrotated so their topsets are approximately horizontal and their heights de-compacted (Pekar et al 2000; Patruno et al 2014). In this study the infill of the basin was first split into four seismic stratigraphic packages bounded by dated surfaces. The youngest/shallowest package was decompacted first along the chosen transects using Allen & Allen s (1990) decompaction equation: y 2 y 1 = y 2 y 1 ɸ 0 c {e( cy 1 ) e ( cy 2 ) } + ɸ 0 c {e( cy 1 ) e ( cy 2 ) } [2] Where y 1 is the depth to the top of the sediment layer; y 2 is the depth to the bottom of the sediment layer; y 1 is the depth to the top of the decompacted sediment layer, y 2 is the depth to the bottom of the decompacted sediment layer; ɸ 0 is the porosity of the sediment at the surface 0.56 for sandyshale (Allen & Allen 1990) and c is a constant that represents the change of porosity with depth 0.39 for sandy-shale (Allen & Allen 1990). The decompacted package was then back-stripped using a simple Airy isostasy model and the strata below were unloaded accordingly. The process was repeated with each of the successive packages, ~ 46 ~

48 in turn restoring the early Quaternary surfaces and finally the base Quaternary surface to its approximate structure at 2.58 Ma. This back-stripping methodology is a simplified model and does not account for complexities such as eustatic changes, flexural responses of the sediment, or heterogeneity of sediment within the packages, and should be taken as an estimate only. The strata immediately below the base Quaternary surface were interpreted with regards to the age of the sub-crop in order to identify features influencing the evolution of the earliest Quaternary basin. The mapping of the sub-crop followed a standard geological mapping process in which geological contacts between units exposed on the base Quaternary surface were mapped on the regional 2D lines and then interpreted across the study area to form a complete map. Geological Figure 2.3 Calibrated depth plotted against TWT for top 1.5 seconds TWT in 1122 wells across the UK and Norwegian sectors of the CNS MegaSurvey with regression line and R2 value for depth conversion as well as maximum data variability margins. Data courtesy of TGS. units were identified according to the seismic packages defined in Evans et al (2003). 2.4 Chronostratigraphic Calibration of the Base Quaternary In seismic reflection data from the UK sector of the CNS, the base Quaternary is usually identified through correlation to one of two markers, the crenulated reflector (Holmes 1977; Gatliff et al 1994; Stoker et al 2011), which is not regionally extensive, or the SNS regional unconformity, which transitions into a correlative conformity along the Plio-Pleistocene clinoform slope. ~ 47 ~

49 Cameron et al (1987, p.46) stated The base of the Quaternary is less easily identified in the centre of the North Sea Basin The seismic boundary which we have used to define the base of the Quaternary offshore is almost certainly diachronous. Attempts at mapping into the UK sector from the Norwegian sector by Ottesen et al (2014) identified the base Quaternary as equivalent to the base of the Naust Formation at 2.7 Ma (Ottesen et al 2014). This method correlated the two through the locally extensive unconformity at the top of the middle-upper Miocene Utsira formation, but this relationship assumes that post-utsira pre- Pleistocene deposits are absent. Towards the southern half of the basin the relationship between the base Quaternary, base-naust and Utsira unconformity breaks down as the Pliocene sediments from the SNS deposited over the top of the Miocene strata. In order to identify and constrain mapping of the base Quaternary a new marker was identified through integrated bio- and magnetostratigraphic studies on shallow gas exploration wells in the SNS by Kuhlmann et al (2006). The authors identified an event in the pollen record which correlates with the climatic degradation at the Plio- Pleistocene transition, the palaeo-magnetic Gauss-Matuyama transition, and the last occurrence of the benthic foraminifera species Monspeliensina pseudotepida (Table 2.1). This is in agreement with other SNS studies on the Cenozoic section such as Meijer et al (2006), ten Veen et al (2013) Thöle et al (2014) and Noorbergen et al (2015) which were extensively correlated by Harding (2015). Table 2.1 Chronostratigraphic correlation adapted from Kuhlmann et al (2006) and Knudsen & Asbjörnsdóttir (1991) indicating three primary surfaces mapped at 2.58 Ma, 2.35 Ma and 1.9 Ma Chronostratigraphic Marker Age (Ma) Josephine-1 (Knudsen and Asbjörnsdóttir 1991) A15-03 (Kuhlmann et al 2006) Cibicides Grossus m b.s.l 770 m b.s.l MIS m b.s.l 816 m b.s.l Monspeliensina pseudotepida m b.s.l 1071 m b.s.l The new base Quaternary horizon was mapped northward in the direction of progradation into the basin and compared to the biostratigraphic record of the Josephine-1 well (56º36.11 N, 2º27.09 E; ~ 48 ~

50 Figure 2.1) to ensure that the M. pseudotepidia event was preserved, within the depth conversion error range, before being extended across the rest of the basin (Figure 2.2). For comparison a correlation was made using the base Quaternary as identified in Knudsen & Asbjörnsdóttir (1991) and Buckley (2012) at the Josephine-1 well using the benthic foraminifera species Cibicides grossus, which Buckley (2012) correlated to the crenulated reflector. C. grossus is a deep water species and its occurrence is diachronous across the North Sea due to the clinoform infill representing a range of water depths in the early Pleistocene (Chris King. Pers. Comm. 2013). In the southern part of the Quaternary basin, where basin infill occurred earlier than in the central areas, shallow-water conditions resulted in the C. grossus event corresponding approximately to Figure 2.4 TWT structure map of the 2.58 Ma surface. Contours are every 100 ms TWT. Identifies locations of Josephine-1 and A15-03 wells and seismic section in Figure 2.2 ~ 49 ~

51 2.35 Ma, whereas in the central part of the Quaternary basin, it corresponds to an age of about 1.9 Ma (MIS 75) (Chris King. Pers. Comm 2013). Therefore the C. grossus biostratigraphic event cannot be correlated to a single chronostratigraphic horizon. As such, two surfaces were mapped for the assessment of the earliest Quaternary basin infill one corresponding to the 1.9 Ma C. grossus event in the Josephine-1 well and the second to MIS 92 at 2.35 Ma in the Netherlands North Sea well A15-03 (55º18 N, 3º48 E; Figure 2.1) based on the biostratigraphic information and comparison of the well log data from Kuhlmann et al (2006) (Table 2.1, Figure 2.2). Figure 2.5 Depth converted 2.58 Ma surface using velocity function derived in Figure 2.4. Contours are every 100 m. Identifies locations of Josephine-1 and A15-03 wells, the seismic section in Figure 2.2 and transects i to v used in the back-stripping process ~ 50 ~

52 2.5 The early Quaternary North Sea basin The base Quaternary event defined in the SNS has been mapped across the southern and central North Sea basin and depth converted according to the borehole-derived velocity function provided above (Figures 2.4 & 2.5). The base Quaternary follows the structural form of the underlying Mesozoic Central Graben with a central basin trough which is elongated NW-SE, until N where the northern portion of the trough switches to a NE-SW trend. The maximum depth of the base Quaternary is 1248 ms TWT (1230 m). The minimum depth along the basin axis is 812 ms (781 m), found in the northernmost part of the basin where a relatively shallow connection exists into Figure 2. 6 Simplified sections from transect i (location Figure 2.5) used in the backstripping process, each point along the transect was used in the calculations and positioned according to distance along the transect. (a) Current configuration of dated horizons. (b) Burial history of the base Quaternary horizon. (c) The changing back-stripped basin configuration through time, note the infill of the basin in this area was virtually completed within the first 230,000 years of the Quaternary. the northern North Sea and eventually the North Atlantic. The central basin trough has a maximum width of 130 km and axial length about 600 km, amounting to approximately 70,000 km 3 of Quaternary sediment. The shallow portions of the basin outside the central trough reach widths of > 300 km and length of > 700 km at the maximum extent of the study area. This accounts for a volume of 120,000 km 3 of Quaternary sediments. The mapped base Quaternary surface was progressively back-stripped, using the dated 2.35 and 1.9 Ma reference horizons as well as the sea floor and the horizon corresponding to the base of the ~ 51 ~

53 Jaramillo palaeo-magnetic Event at 1.1 Ma, to produce transects of the changing basin geometry and architecture through time (Figure 2.6). The calculations indicate clinoformal heights in the region of 250 m to 300 m in the southernmost part of the basin, and 200 m to 250 m in the central parts, indicating a shallowing of the basin towards the north and maximum palaeo-water depths of the order of 300 m Figure 2.7 Simplified line transects used in the back-stripping process showing structure of the margins of the earliest Quaternary basin and progradation of clinoform packages ± 50 m. The eastern margin of the base Quaternary basin comprises at least three major clinoform packages (Figure 2.7), which also incorporate the northern and southern margins, while the western margin represents the distal toesets of the NW- and westward progradational clinoforms, which onlap on to Miocene strata (Figures 2.8 & 2.9). Few slope features were preserved on the base-quaternary surface however, in the CNS, pockmarks are scattered across the top- and foreset of one of the eastern clinoforms. These features were previously identified by Kilhams et al (2011) who tentatively suggested a Miocene age based on very limited chronostratigraphic data. Our revised chronology clearly demonstrates a base Quaternary age for these features. The earliest Pleistocene sediments within the basin have a clinoformal geometry, which varies between high angle (4-5 ) sigmoidal or oblique reflections with well-defined break points to very low angle (<0.5 ), comparatively flat reflections (Figure 2.7). The palaeo-environments along the ~ 52 ~

54 clinoforms most likely ranged from floodplain and shallow shelf along the topsets through outer littoral and upper to middle bathyal environments along the slope and toeset. Each clinoform thus represents a suite of palaeo-environmental conditions with associated variations in sediment grain size distribution (Posamentier & Allen 1993; Stuart & Huuse 2012; Patruno et al 2014, 2015). The majority of Quaternary strata in the centre of the basin overlie Pliocene sediments (Figures 2.8 & 2.9) and represent a continuation of the Pliocene clinoforms (Sørensen et al 1997; Overeem et al 2001). This is particularly the case for the southernmost clinoform package (Figure 2.7, transect i). The southern clinoform package prograded towards the south-west during the earliest Quaternary and has an arcuate planform with sigmoidal slopes between 2 and 4 with a maximum backstripped clinoform height of 240 ± 50 m (Figure 2.4) although this reaches 300 ± 50 m away from the main transect. Between 2.58 Ma and 2.35 Ma the southern clinoform set deposited 590 ± 50 m of mud to fine sand as the clinoforms prograded 110 km westwards (Figure 2.10). This implies an average sedimentation rate of 2.6 ± 0.2 mm yr -1 for the 230,000 year period, with the southern part of the basin nearly filled by 2.35 Ma with remaining accommodation illustrated by back-stripped clinoform heights at this time of approximately 40 m. The eastern clinoform set has an initial backstripped clinoform height of 180 m (Figure 2.7, transect ii). Between 2.58 and 2.35 Ma 60 m of sediment was deposited (Figure 2.10) at an average rate of 0.26 mm yr -1 producing a sigmoidal geometry with a back-stripped clinoform height of 216 m. The northern clinoform set is comparatively flat, and with an initial clinoform height of 150 m dropping to 108 m by 2.35 Ma. The northern clinoforms do not appear to have actively prograded during the first 230,000 years; the minor deposition (30 m at 0.13 mm yr -1 ) was more aggradational in nature and primarily on the topsets rather than the foresets (Figure 2.7, transect iv). In the SNS, downslope channels and mass transport deposits are preserved on the slopes and toesets in the 2.58 to 2.35 Ma package (Figure 2.11a). Further north into the CNS these features are replaced by a series of elongated, near-linear features with u shaped cross sections (Figures 2.2 & ~ 53 ~

55 Figure 2.8 Seismic section showing interpretation of sub-crop beneath the 2.58 Ma surface used to produce sub-crop map (Figure 2.8). Packages separated by age after Evans et al (2003). See Figure 2.8 for location. Data courtesy of TGS ~ 54 ~

56 2.11b), previously described by Cartwright (1995), Knutz (2010), Kilhams et al (2011) and Buckley (2012). 2.6 Discussion Mapping of the current geometry of the basal Quaternary reflection coupled with back-stripping of Figure 2.9 Simplified map of the base Quaternary sub-crop indicating areas of sediment hiatus particularly on the western side of the basin. In the SNS the sub-crop becomes extremely complex due to the presence of large salt tectonic features which heavily disturb the seismic reflection correlated to the base Quaternary the clinoforms indicates a basin with water depths in the region of 300 m, in agreement with the biostratigraphy from Kuhlmann et al (2006), enclosed by the NW European landmasses, with a narrow marine connection to the north (Figure 2.12). Analysis of the sedimentation rates and ~ 55 ~

57 depocentre migration patterns suggests the primary source of basin infill was from the south. The sediments appear to have originated from the Baltic (Biljsma 1981; Sørensen et al 1997; Overeem et al 2001) and Rhine-Meuse (Busschers et al 2007) river systems resulting in the dominant clinoform progradation from south to north. The two river systems drained large areas of northern Europe during this time, including the Fennoscandian shield, the Baltic platform and large areas of NW Europe from the Alps to the present day mouth of the Rhine (Overeem et al 2001; Busschers et al 2007). The rapid northward progradation has been linked with climatic cooling and increased sediment supply due to glacial activity in the sediment source areas (Overeem et al 2001; Huuse 2002). Strontium isotope ratios indicate a high freshwater input in this section (Kuhlmann et al Figure 2.10 Thickness maps of Quaternary strata used to estimate sedimentation rates. (a) 2.58 Ma to present day sea bed. (b) 2.58 Ma to 2.35 Ma. (c) 2.58 Ma to 1.9 Ma. (c) also represents the difference between the traditional base Quaternary maps such as Holmes (1977) and the new base Quaternary. Fine grey lines represent thickness contours at every 100 m thickness. Black and grey lines indicate location of clinoform breakpoints at 2.58, 2.35 and 1.9 Ma ~ 56 ~

58 2006) suggesting the expanded stratigraphic section between 2.58 Ma and 2.35 Ma may provide key information for the northern European climate record of the earliest Quaternary. The northern end of the basin saw relatively low sedimentation rates in the earliest Quaternary, with the primary sedimentation type being aggradation rather than progradation. The reason for the low sedimentation rates is likely to be the southwards drainage of the Fennoscandian shield into the Baltic river systems proposed by Overeem et al (2001). The Baltic river systems ultimately feed into the North Sea basin as part of the southern clinoform sets bypassing the northern end of the basin. This bypass of the northern part of the basin during the earliest Quaternary means any correlation between the Norwegian and North Sea depositional systems, such as that presented by Ottesen et al (2014), is very difficult to test using the present-day limited distribution of highquality chronostratigraphic calibrations. The two depositional systems could, in fact, indicate completely separate depositional systems, one preserving solely the Scandinavian climate signal, and the other the Northern European signal, or a mixed signal routed through the southern clinoform set. Surfaces after backstripping indicate that the geometry of the basin changed dramatically through the earliest Pleistocene, with the deepest parts of the basin, in the SNS, being almost filled, while the clinoforms further north increase in height. In the SNS, the sedimentation rate therefore far outstripped the rate of subsidence due to sediment loading, while sedimentation rates further north into the CNS initially were lower than subsidence rates. However, caveats to this suggested fill pattern include the relatively simplistic calculations for the de-compaction and back-stripping used in this study. The calculations do not account for eustasy, variations in sediment composition, or flexural effects from the narrow, elongate basin being loaded principally at one end. The calculations produce a first order estimate of clinoform heights, within a 50 m error margin, but for accurate models of subsidence history much more detailed is be needed. ~ 57 ~

59 The size and geometry of the southern clinoforms coupled with the presence of mass transport deposits and downslope channels suggests a shelf system onto which multiple small deltas prograded rather than a single large delta. Correlation of gamma-ray logs to Kuhlmann et al s (2006) study indicates that the base Quaternary surface is a flooding surface associated with a rise and highstand of sea level, in agreement with the global sea level curve from Miller et al (2005) which indicates a sea-level rise at 2.58 Ma. This would suggest that large parts of the shelf would have been flooded at the base Quaternary leading to a an extensive shallow marine environment on the topset part of the clinoforms possibly as shallow as 20 m water depth dropping sharply at the clinoform breakpoint of the shelf prism into the elongate deep water basin with depths in the region of 300 m (Figure 2.12). This agrees with interpretations of onshore stratigraphy from southeastern Britain which suggests that the basal Quaternary was deposited in a shallow-marine environment (McMillan et al 2005; Rose 2009). The elongate u shaped features seen on the basal Quaternary in the central part of the basin have been interpreted as troughs formed by a strong downslope currents or contourite systems (Cartwright 1995; Knutz 2010; Kilhams et al 2011) or as part of a glacial imprint (Buckley 2012). Given the water depths proposed for the basal Quaternary and lack of other glacial geomorphic features we suggest that a current system is a more likely explanation. The source and nature of these troughs remain unclear (Chapter 3) although it would not be unreasonable to suggest that the high influx of sediment-laden water during this period had some influence on the formation of a large-scale significant current system. 2.7 Implications The newly identified depocentre of >0.5 km thick MIS (2.58 to 1.9 Ma) strata is of great importance for the study of palaeo-climate, particularly surrounding the onset of Northern Hemisphere glaciation. Analysis of the seismic geomorphology is now well established for the late Quaternary of the North Sea (Praeg 1996, 2003; Graham et al 2011; Stewart & Lonergan 2011; Buckley 2012; Kristensen & Huuse 2012; Stewart et al 2012, 2013; Moreau & Huuse 2014) and, if ~ 58 ~

60 Figure 2.11 (a) RMS amplitude extraction of horizon within the earliest (2.58 to 2.35 Ma) Quaternary package in the southern North Sea showing downslope channels and fan deposits on the clinoform slopes and toe-sets. (b) Seismic amplitude extraction across base Quaternary (2.58 Ma) surface in the central North Sea showing elongate, semi-parallel furrows linked to deep-water processes. ~ 59 ~

61 applied to this newly defined early Quaternary succession, could fill in large gaps in our understanding of how ice sheets grew and behaved during the 41 kyr glacial-interglacial cycles of the early Quaternary (Raymo & Nisancioglu 2003). The structural map of the earliest Quaternary basin can be combined with the back-stripping calculations to reconstruct the pre-glacial palaeobathymetry of the North Sea, which is one of the more poorly defined boundary conditions in ice sheet modelling (Peltier 1994). The thickness map reveals areas of anomalously high sedimentation rates producing an expanded section which, if drilled, could be used to produce one of the strongest chronostratigraphic calibrations for the Quaternary. If combined with drilling of the marine toesets, also identified from the thickness maps, it would provide one of the most complete and detailed mid-latitude palaeo-climate records of the earliest Quaternary. The shift of the base Quaternary > 500 m deeper into the basin fill will influence estimates of burial history, stress fields, and thermal history of the sedimentary substrate (Guidish et al 1985; Ehrenberg et al 2009). All of these estimates as well as the palaeo-water depths and palaeo-surface temperatures are used in the modelling of hydrocarbon migration, trapping and leakage (Goff 1983; Duncan et al 1998) which is of interest to petroleum exploration, carbon capture and storage (Chadwick et al 2004; Eiken et al 2011; EU 2011) and the analysis of geohazards (Haskell et al 1999). 2.8 Conclusion The base Quaternary (2.58 Ma) surface in the central North Sea has been mapped in detail for the first time, revealing an elongated semi-enclosed deep marine basin with palaeowater depths throughout the early Quaternary of up to 300 m +/- 50 m. Evidence from seismic geomorphological analysis suggests this marine basin was influenced by high sediment input from Northern Europe, resulting in a palaeo-environment dominated by turbidties, channels and fans as well as shelfmargin deltas. The presence of contourites suggests strong deep water currents in the earliest Quaternary prior to 2.3 Ma. Estimates of an average sedimentation rate of 2.6±0.5 mm yr -1 from clinoform progradation suggest that a significant expanded record of climate change in the earliest Quaternary could be preserved in the southern part of the CNS and SNS. The map of the base Quaternary and the early Pleistocene depocentre have powerful implications, academically and ~ 60 ~

62 applied, for the evolution of the North Sea basin and the palaeo-climate evolution at of the Plio- Pleistocene transition. Figure 2.12 Reconstructed palaeo-environmental map of the 2.58 Ma North Sea based on results of this study, Overeem et al (2001), McMillan et al (2005), Busschers et al (2007), Rose (2009) and Noorbergen et al (2015). Large parts of the present day North Sea would have been flooded under the highstand conditions at the onset of the Quaternary creating a very shallow shelf but were otherwise terrestrial. 2.9 Acknowledgements The authors would like to acknowledge NERC and BUFI in conjunction with Forewind for supporting this work through RL s NERC CASE studentship at the University of Manchester and the British Geological Survey (No. A87604X). We would like to thank PGS, particularly Steve Morse and Richard Lamb, for providing the CNS and SNS MegaSurveys, TGS for providing 2D ~ 61 ~

63 North Sea renaissance seismic lines and well data through their Facies Map Browser, IHS for Kingdom software, and Schlumberger for the donation of Petrel software. Thanks to Stefano Patruno for his invaluable advice on the back-stripping process and Chris King for sharing his insights on the biostratigraphy of the North Sea. ~ 62 ~

64 Chapter 3: Early Quaternary sedimentary processes and palaeo-environments in the central North Sea Rachel M. Lamb, Mads Huuse, Margaret Stewart Abstract: The early Quaternary of the central North Sea basin has not been studied in detail since the 1970s and 80s. The marine basin that existed during the early Quaternary was infilled rapidly through the progradation of large clinoform sets but the sedimentary processes involved are poorly understood. 380 elongated troughs ranging from 0.1 to 50 km in length and 0.1 to 3.5 km in width are mapped over 11,000 km 2 of the early Quaternary North Sea basin. The troughs can be split into two types based on size and orientation. Type 1 troughs are narrow and are oriented along depositional dip of the clinoform slope. Type 2 troughs are wider and deeper and formed at orientations largely parallel to clinoform slope. Considering the palaeo-geographic context, it is argued that both trough types were principally formed by downslope gravity-flow processes in a marine environment. Type 2 troughs resulted from deviation of the downslope current by interaction with an anti-clockwise tidal gyre deflecting the resulting current to the east. The troughclinoform system resulted from the interaction of a relatively deep-marine basin being rapidly filled by mud-rich clinoforms sourced by the Baltic and NW European rivers whose effluent cascaded down the clinoforms under the influence of an anti-clockwise geostrophic gyre. Having documented the seismic-scale features of the clinoform-trough system, further understanding of the early Quaternary succession, its depositional history and climate signals, can best be achieved by retrieval of continuous cores of this very high-resolution palaeo-climate archive. ~ 63 ~

65 3.1 Introduction The North Sea is a mid-to-high latitude epicontinental sea located between the British, Scandinavian and Northern European landmasses connected to the North Atlantic to the north and to the English Channel to the south. The Norwegian Trench, which runs parallel to the Norwegian coast along the eastern margin of the North Sea, reaches depths of over 700 m; however the rest of the North Sea averages close to 100 m water depth, gently increasing northward towards the Atlantic margin. The present day North Sea is dominated by strong tidal currents creating features such as the large sand banks and waves in the Southern Bight (Caston & Stride 1970; Terwindt 1971; McCave 1970; Caston 1972; Otto et al 1990; Gatliff et al 1994). Additionally a shallow anticlockwise geostrophic current system exists in the North Sea. Water from the North Atlantic inflows into the basin from the northwest, around the Shetlands, flowing southwards mixing with both a relatively minor influx of water from the Channel and freshwater input from the northern European and British river systems (Figure 3.1a; Otto et al 1990; Svendsen et al 1991; Turrell et al 1992; OSPAR 2000; Gyllencreutz et al 2006). The mixed water in the southern North Sea then flows north and eastwards mixing further with water flux from the Baltic river system in the Skagerrak region, feeding into the Norwegian channel which forms the main output of water from the North Sea (Figure 3.1a; Otto et al 1990; Svendsen et al 1991; Turrell et al 1992; OSPAR 2000; Gyllencreutz et al 2006). The Quaternary deposits in the North Sea can be subdivided across the southern, central and northern North Sea depocentres of which the central North Sea is the thickest at more than 1 km thick (Cameron et al 1987; Sejrup et al 1991; Sørenson et al 1997; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014). Within this the Quaternary stratigraphy is split into the expansive early Pleistocene marine dominated stratigraphy, which in the southern and central North Sea (SNS, CNS) is a continuation of Cenozoic deltaic deposition, and the late Pleistocene to Holocene glaciogenic formations (Cameron et al 1992; Gatliff et al 1994; Stoker et al 2011). These groups are separated by the regional glacial unconformity at the base of the expansive network of tunnel valleys which separate ~ 64 ~

66 the early and middle Pleistocene (Cameron et al 1987; Wingfield 1989; 1990; Gatliff et al 1994; Stoker et al 2011; Praeg 2003; Stewart et al 2013). Figure 3.1 a) Present day circulation of shallow residual water currents in the North Sea from OSPAR (2000). b) Regional map of the central North Sea showing the extent of the 3D seismic data used in the study, the PGS CNS MegaSurvey and CGG BroadSeisTM surveys. Highlights the locations of other studies on the trough features. Bathymetry data from Ryan et al (2009) A great deal of work has focused on reconstructing the palaeo-environments of the shallowest late Pleistocene to Holocene deposits however there is comparatively less which focuses on the early Quaternary. Most of the work on the early Quaternary was undertaken during comprehensive mapping of the North Sea stratigraphy during the 80 s and 90 s, which while thorough relied on widely spaced 2D seismic lines and shallow cores generally < 250 m in length (Holmes 1977; Stoker et al 1983; Stoker & Bent 1985; Cameron et al 1987; Sejrup et al 1987, 1991, 1994; ~ 65 ~

67 Knudsen & Sejrup 1993). In the southern North Sea where the majority of the early Quaternary deposition was on topsets of deltas the shallow cores allowed for accurate description of the pretunnel valley deposition, identifying a wide plain from Britain to Denmark which was flooded at the onset of the Quaternary (Cameron et al 1984, 1993; Funnell 1996; McMillan et al 2005; Rose 2009) but became subaerially exposed following a marine regression prior to the middle Pleistocene (Cameron et al 1984, 1989, 1993; Funnell et al 1996). In the CNS however shallow cores barely penetrated the top of the sedimentary sequence, identifying glaciomarine sediments, and the geometry of the 2D seismic identified the thick sequence and general character of the marine basin however little information on the nature and processes involved could be interpreted. In the years since this extensive mapping only a handful of studies have investigated the early Pleistocene of the CNS despite a dramatic rise in the amount of data available (Brzozowska et al 2003; Harvey et al 2010). This paper uses basin-scale, continuous 3D seismic data in order to investigate the marine processes that are preserved in the seismic stratigraphy of the early Quaternary succession. It aims to improve understanding of the early Quaternary palaeo-environments in order to aid in palaeogeographic reconstruction. 3.2 Regional Setting The North Sea originated due to a series of rifting episodes during the late Palaeozoic and Mesozoic eventually culminating in volcanism and rifting in the mid Jurassic to early Cretaceous. (Ziegler, 1992) Rifting ceased when the North Atlantic rift eventually propagated northward to the northwest of Britain, leaving an aborted rift controlling subsidence patterns in the North Sea. The Cenozoic record of the North Sea shows multiple events of basin inversion, subsidence, uplift and erosion throughout the time period (Ziegler 1992; White & Lovell 1997; Huuse 2002; Stoker et al 2005; Anell et al 2010; Goledowski et al 2012). In the Quaternary thermal subsidence has been the primary driver of basin development, magnified by rapid infill of the basin allowing for the buildup of the thick Quaternary succession (Huuse 2002; Anell et al 2010; Goledowski et al 2012). ~ 66 ~

68 At the onset of the Quaternary the North Sea was represented by elongate basin orientated along a NW-SE axis over the Central Graben extending across the CNS and well into the southern North Sea (Holmes 1977; Cameron et al 1987; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014). Away from the main basin the North Sea most likely consisted of a relatively flat plain of extremely shallow water depths or terrestrially exposed. Infill of the basin was rapid, particularly in the southern North Sea which is noted to be infilled and sub-aerially exposed at some point during the early Quaternary (Cameron et al 1984; Cameron et al 1989; Cameron et al 1993; Funnell et al 1996), building up over 1.1 km of sediment in the deepest parts of the basin (Caston 1977; Holmes 1977; Cameron et Figure 3.2 Map showing the mapped thalwegs of the troughs overlying the extent suggested by Cartwright (1995). Indicates locations of figures and wells used in the study. ~ 67 ~

69 al 1987; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014). The early Quaternary stratigraphy is represented by the Southern North Sea Deltaic (deltaic to pro-deltaic marine sediments) and Dunwich (nonmarine fluvial to inter-tidal sediments) Groups, in the SNS and the Zulu Group (deltaic to prodeltaic sediments) in the CNS (Cameron et al 1992; Gatliff et al 1994; Stoker et al 2011). The deltaic sediments were fed into the North Sea primarily from the south-east by the Rhine-Meuse and Baltic river systems (Overeem et al 2001; Busschers et al 2007). The early Quaternary saw the onset of global cooling and the initiation of northern hemisphere glaciation (Raymo 1994; Lisiecki & Raymo 2005, 2007; Miller et al 2011) however evidence for ice in the CNS has thus far been restricted to iceberg scours, which have been identified from the onset of the Pleistocene (Dowdeswell and Ottesen 2013; Ottesen et al 2014). Conclusive morphological evidence of shelf-wide, grounded glaciation in the CNS has never been provided although highly debated and fragmentary evidence has suggested that there is a possibility for glacial activity in the early Quaternary (Graham et al 2011 and references therein). This study focuses on the central third of the central North Sea between N and 58 N (Figure 3.2). In this area the early Quaternary basin is at its deepest (Figure 3.3). The entire early Quaternary section is located within the Aberdeen Ground Formation of the central North Sea Zulu Group representing marine and pro-deltaic to deltaic sediments (Figure 3; Gatliff et al 1994; Stoker et al 2011). Dating of the section comes principally from mapping of chronostratigraphic markers from Dutch sector well A15-03 identified by Kuhlmann et al (2006) at 2.58 Ma (base Quaternary), 2.35 Ma and 1.9 Ma through an integration of multiple chronological proxies including biostratigraphy, palaeo-magnetism and wireline log data (Figures 3.2, 3.3; Kuhlmann et al 2006; Noornbergen et al 2015). Palaeo-magnetic data from BGS core 77/02 was also used to provide chronological ties for the Brunhes-Matuyama reversal (0.78 Ma) and the base of the Jaramillo ~ 68 ~

70 Figure 3.3 Regional seismic cross section (50x VE) showing the regional seismic stratigraphy of the study area identifying the furrowed interval in the Aberdeen Ground Formation of the early Quaternary (based on Cameron et al 1992; Gatliff et al 1994; Stoker et al 2011). Dates of chronostratigraphic horizons are from Dutch sector well A15-03 (Kuhlmann et al 2006) and BGS borehole 77/02 (Stoker et al 1983). Inset map showing the base Quaternary TWT structure, clinoform breakpoints for 2.58, 2.35 and 1.94 Ma chronostratigraphic ties and location of section. Seismic data courtesy of PGS. ~ 69 ~

71 Event (1.1 Ma) (Figures 3.2, 3.3; Stoker et al 1983). The seismic stratigraphy of the earliest Quaternary in this part of the basin shows strong characteristic features in the form of elongate troughs, v shaped in cross section, which run approximately parallel to the basin axis along the basin floor (Figures 3.3, 3.4). These features have been observed in the past (Cartwright 1995; Knutz 2010; Kilhams et al 2011; Buckley 2012), however a consensus has not been reached over the processes which formed them. Understanding the processes involved in the formation of these features could add more detail to palaeo-environmental reconstructions of the early Quaternary North Sea. 3.3 Data and Methods Data used for this study include two 3D seismic cubes; the PGS CNS MegaSurvey (v. 2012), with an area of 88,000 km 2, and the CGG BroadSeis TM seismic cube covering an area of 2,250 km 2 (Figures 3.1, 3.2). The MegaSurvey resolves to m vertically with horizontal bin spacing of 50m while the BroadSeis TM survey resolves to 7-8 m vertically with bin spacing of 12.5 m. To complement the seismic data eleven wells were selected from the south-east of the study area in order to identify any patterns in the sediment grain size distribution. The seismic stratigraphy of the furrowed area was mapped using a semi-automated mapping methodology in order to separate the troughs in time. Two reference horizons corresponding to the chronostratigraphic ties from the southern North Sea at 2.58 Ma and 1.9 Ma were mapped carefully across the area and then imported into the Paleoscan software which builds a 3D geo-model based on the seismic and reference horizons and extracts a horizon stack automatically. Comparison of these semi-automated horizons to the original seismic can be used to quality control the result, allowing for minor editing of horizons before key horizons are selected for analysis and interpretation. Thickness maps and clinoform breakpoints for each of the horizons were made in order to reconstruct the evolution of the basin structure during furrowing. ~ 70 ~

72 A combination of time slice and cross-section interpretation was used to map each individual trough. Troughs were identified firstly in time slice using a combination of seismic amplitude and the variance attribute (Figure 3.4). Locations, orientation and extent of each furrow were noted before full mapping occurred in cross-section, using time-slices to accurately correlate between each line. In cross-section troughs were mapped from the basal truncation upwards identifying the sides of the troughs based on the curvature of the reflections and truncations (Figure 3.5). The top of any trough was generally identified by the end of aggradation and beginning of drape, signifying the end of active furrowing (Figure 3.5). Acoustic anomalies, namely velocity pull-up effects, were observed and noted. Pull-ups generally occur when a contrast in sediment type increases the velocity of the seismic wave to with respect to the background sediment, causing the underlying seismic reflections to appear mounded. They can be linked to consolidated deposits in an otherwise unconsolidated sediment. Each trough was mapped individually in both the regional and broadband seismic datasets with the boundaries, thalwegs and minimum bounding rectangles digitised as polygons, which were then exported to GIS software in order to extract length, average width, depth, sinuosity and orientation. The difference in orientation between the furrows and the clinoform slope was calculated by splitting the clinoform break point for the 1.94 Ma horizon, which corresponds to the earliest stage of trough formation, into six sections (Figure 3.6). The average strike of the clinoform slope was found for each section and then the difference between each trough and the nearest section of the clinoform was calculated. This allowed the change in orientation of the troughs relative to the clinoforms from west to east to be quantified. The available well data consists primarily of gamma log data and lithological descriptions which were converted to two-way time (TWT) using pre-calibrated time-depth curves provided by TGS. Wells are used to interpret the grain size distribution and facies. It is unclear if the lithological descriptions are from well cuttings or from interpretation of the logs however the descriptions are assumed to be best available descriptions of the sediment package. ~ 71 ~

73 Figure 3.4 Extraction of amplitude variance from a 10 ms TWT window either side of the F04 horizon demonstrating the full extent of the troughs and highlighting the individual troughs used in mapping. Location in figure 3.2. ~ 72 ~

74 3.4 Results A total of 380 troughs were mapped across the study area revealing a lens-shaped furrowed interval (Figures 3.7, 3.9) covering a total area of approximately 12,000 km 2 contained within an up to 0.5 km thick succession (Figures 3.3, 3.7). The troughs are elongated and slightly sinuous in planform and vary between u and v shaped in cross section with sizes from 0.1 to 50 km in length, 0.1 to 3 km in width and 30 to 275 ms TWT deep. Each trough was separated by plan-form type, cross-sectional form and horizon in order to compare features and unlock clues to potential modes of formation. The results presented here first discuss the characteristics of the troughs in planform (3.4.1), cross-section (3.4.2) and longitudinal profile (3.4.3) before looking at the separation of troughs by horizon (3.4.4) and the analysis of the background sedimentation through well logs (3.4.5) Plan Form In plan form the troughs can be divided into two distinct types separated by differences in orientation relative to the clinoforms of the basin infill and size. Figure 3.6 shows the width/depth ratio and the angular difference between the axis of the troughs and the strike of the clinoform slope highlighting this difference. The orientations of type 1 troughs are found to be almost perfectly perpendicular to the strike of the clinoforms (Figure 3.6b). Type 2 troughs have a wider range of orientations but are, on average, found to be approximately parallel to the clinoform strike (Figure 3.6c). In terms of size there is significant overlap between type 1 and type 2 troughs however type 2 troughs tend to be both wider and deeper compared to type 1 troughs (Figure 3.6a). The two types are geographically separate and each covers approximately half of the total furrowed area (Figure 3.6d). There are 266 type 1 troughs (70% of the total) to the west of the furrowed area while type 2 troughs dominate the eastern half with 114 in total (30%). Type 1 troughs have a higher density in distribution and distances between individual troughs lie between 0 and 5 km but most commonly fall between 1 to 2.5 km (Figure 3.6d, e). Type 2 troughs show distances of 1.5 to 8 km between them, but on average are more than 2.5 km apart (Figure 3.6d, e). Type 1 troughs ~ 73 ~

75 exhibit sinuosity ranges between 1 and 1.05 with an average of 1.01 while type 2 show sinuosities from 1 to 1.09 averaging at Type 1 troughs are the only furrows that exhibit signs of lateral migration limited to troughs located on the most western margin of the basin (Figures 3.5, 3.8) Cross-Sectional Class The cross-sectional class of the troughs is defined by basal truncation, infill character, stacking pattern and size. In cross-section the troughs are remarkably uniform with five different crosssectional classes displayed and the majority (<80%) exhibiting one of two classes and frequently switching between the two along the axial length of the trough. The cross-sectional form does not appear to relate to the different planform types; both types of trough show all cross-sectional classes. Figure 3.5 summarises the five cross-sectional classes. Cross-sectional class a is the characteristic form of the troughs (Figure 3.5a). Troughs with a class tend to truncate lower stratigraphy at the base and evidence for velocity pull-up effects is common but rarely consistent. Infill of a class troughs consists of mid-to-high amplitude reflections from horizontal fill to drape. There was periodic incision, evidenced by truncations within the fill itself. Form a troughs have a purely aggradational stacking pattern and are generally symmetrical with active furrowing persisting through three or more horizons. Cross-sectional class b is the second most common cross-sectional form (Figure 3.5b). Similar to a form troughs b forms have basal truncations and pull-up effects though b forms tend to be better defined and more consistent. Infill of b forms is purely drape and do not show any aggradation in the stacking pattern nor further truncations within the fill. A number of troughs transition from a to b forms for a few kilometres along the length before returning to the dominant a form, suggesting that b class troughs are a sub-form of a class troughs. ~ 74 ~

76 Figure 3.5 A) Sketch section indicating key features used in mapping of troughs. a-e) Table of representative cross-sectional forms of troughs. ~ 75 ~

77 Cross-sectional class c is found in a geographically limited area on the westernmost margin of the furrowed area (Figure 3.9) and solely with type 1 plan-forms. Form c differs from form a only in stacking pattern which shows westwards lateral migration, which is the up-slope direction (Figure 3.5c), and is thus considered to be a sub-form of class a troughs. Cross-sectional class d troughs do not generally have well-defined basal truncations (Figure 3.5d). The infill of d troughs is drape and they do not have a stacking pattern as they generally do not persist to be expressed on the next reflection. Class d troughs are, on average, much smaller than other troughs and the majority are <5 km long and <1 km wide and may be a form of prototrough. The final cross-sectional class e has the strongest basal truncation, exhibiting smooth erosive base within the main part of the trough which becomes increasingly undulating on the sides of the trough (Figure 3.5a). The fill of class e troughs is low amplitude to near transparent horizontal fill with no stacking pattern. In terms of size form e troughs tend to be significantly wider and deeper with more shallowing dipping sides than all other classes; although class e troughs are not as long. Form e troughs are much younger and contrast strongly with the other cross-sectional forms suggesting a separate mode of formation Longitudinal profile All troughs persist from initial formation and, with the exception d and e cross-sectional forms, are expressed on numerous horizons before fully infilled. If a single horizon is taken to be timesynchronous (Mitchum et al 1977a, b) then a relative timeline can be formed of the length of time each trough is active for. Troughs are expressed on up to three horizons subsequent to their formation and, although type 2 troughs have generally thicker deposits to type 1 (Figure 3.6) there is no significant difference between the lengths of time individual troughs were active for. ~ 76 ~

78 Figure 3.6 a) Graph of average width versus maximum depth of troughs by trough type indicating the trend towards type 2 troughs being wider and deeper than type 1. b & c) Rose diagrams indicating the bidirectional orientation of types 1 and 2 troughs as well as the angular difference between the orientation of the troughs and the strike of the closest section of the clinoform slope. d) Map showing position of type 1 and 2 troughs relative to the 1.94 Ma clinoform break point as well as slope direction indicators for the active clinoform. e) (Next page) Seismic section across the axis of the basin showing the distribution of type 1 and type 2 troughs in cross-section. Note the high density of the type 1 troughs and the thicker deposits of the type 2. ~ 77 ~

79 ~ 78 ~

80 Individual trough fills shows thickness variations along the length of the trough. Long profiles of the troughs (Figure 3.8) reveal undulating thalwegs as well as a pattern of basal truncation. The deepest parts of the troughs are seen to truncate underlying reflections whilst, as the trough shallows, the reflections become increasingly conformable before further truncation occurs as the trough deepens again. The pattern is repeated throughout the life-time of the trough with between one and three other occurrences of renewed incision into the fill of the trough. The frequency of the deepening and truncation along the trough axis is variable, although commonly occur every 3.5 to 6 km along the furrow. This undulation, combined with the repeated instance of truncation and drape in the cross-sectional forms, yields a cut-and-fill geometry to the troughs Horizon Mapping A total of ten horizons, F01 to F10, were chosen from the semi-automated mapping in order to define the furrowed interval as well as the chronostratigraphic ties (Figure 3.7). Troughs incised from nine of the ten horizons, F01 to F09. The horizons show the present day TWT-structure however thickness maps between horizons can be used as a proxy for the palaeo-depocentre (Figure 3.9) and thus basin and slope contours. The basal-quaternary surface can be identified at the base of the furrowed interval and the TWTstructure reveals a basin elongated along a north-west south-east axis which roughly follows the underlying Mesozoic Central Graben until approximately N when the basin begins to deviate eastward from the graben trend eventually resulting in a north-easterly orientation. The furrowed area spans the width of the basin at an angle almost perpendicular to the overall NW-SE basin axis covering ~12,000 km 2 in the deepest part of the basin. No troughs originate from the base-quaternary horizon, instead furrowing initiated almost immediately above with some troughs truncating the base-quaternary surface from a higher level. The onset of furrowing coincides with the F01 surface. Initially most troughs are small, generally with cross-sectional class d or, occasionally, b but quickly increase in size and frequency up to a ~ 79 ~

81 Figure 3.7 a) Seismic section identifying the mapped horizons and clinoform progradation. b) Map indicating approximate progression of the clinoform break points through time. C) High resolution seismic section showing incision of troughs from different horizons. Location on figure 3.2. Seismic data courtesy of PGS and CGG. ~ 80 ~

82 peak of trough activity at horizon F04 before a steady decay in the number and spread of new troughs (Figure 3.8). From the earliest Quaternary onwards the majority of sediments came from the south-east as a large prograding clinoform set moves northwards to fill in the basin (Figures 3.3, 3.7, 3.8). Clinoform heights can be used as a proxy of water depth at any given time. Although clinoform heights vary between F01 and F09 the range consistently falls within 300 to 400 ms TWT (Figure 3.3, 3.7) which approximates to between 270 and 360 metres water depth (assuming a standard seismic velocity between 1800 and 2000 ms -1 ). All horizons show some input from the southern clinoform system however in the F06-F07 and F08-F09 intervals there is a secondary input from the north-west. By the time represented by the F09 surface the clinoform breakpoint has prograded over 200 km resulting in a dramatic change in basin shape with the axis of the basin lying on a new NE-SW axis (Figure 3.7b). The basin south of N is almost completely filled in by F09 however clinoform heights do not drop below ~250m (Figure 3.7a) thus maintaining significant water depths in the migrating basin centre. By F09 the furrowed area no longer straddles the deepest part of the basin but lies on the southern margin of the basin incising into the clinoform slope up to just downslope of the breakpoint. No channels are observed on the topsets of the clinoforms in this section of the seismic stratigraphy however both ice-berg scours and small slopeparallel mounded features are observed (Figure 3.10) Well Correlation Eleven wells with gamma-ray logs and lithological descriptions were used to investigate the background sedimentation and trough infill. Wells 22/25a-3 and 30/6-3 were found to be representative of the section and so best describe both trough infill and background sediments (Figure 3.11) with well 22/25a-3 penetrating a type 2 trough and well 30/6-3 representing an undisturbed section through the background sediments. ~ 81 ~

83 ~ 82 ~

84 Figure 3.8 A-A (Previous page) Long profile of trough showing repeated pulses of active furrowing as well as highlighting periodicity of truncations along base of active furrow. a & b) Cross-sections through long profile showing multiple incisions within fill with different colours representing fill packages based on truncation of older fill. Location in figure 3.2. Seismic data courtesy of PGS. The principle background sedimentation in the furrowed section was found to be silty-claystone as seen from the lithological descriptions of well 30/6-3. Towards the base of the furrowed section and thus the base-quaternary, 30/6-3 shows the presence of thin silt layers, unfortunately not seen in the gamma-ray log due to a gap in the data. The gamma-ray log shows numerous peaks but maintains the relatively high response expected from claystone throughout the furrowed section. The exception to this is a drop in the gamma response between horizons F05 and F06 which corresponds to the presence of a thin (< 20 m) bed of silty-sandstone in the lithological description. This bed corresponds to the topset of a prograding clinoform which onlaps the basin margin to the north-west. In the shallower section above the furrowed interval there is a switch from dominantly silty-claystone to silty sandstone. Well 22/25a-3 shows the same dominant silty-claystone lithology is consistent throughout the whole section. The trough that 22/25a-3 penetrates does not appear, from the lithological ~ 83 ~

85 descriptions, to have any effect on the sediment of the fill. In comparison the gamma-ray log shows some distinction between the trough and the background sedimentation showing a higher and more variable gamma response, though still within the typical responses of claystones ( API; Rider 2002). The higher response and isolated peaks suggests a fining of the material within the fill of the trough. 3.5 Discussion Previous studies of these troughs have inferred variable ages and models of formation without coming to a consensus. Models suggested by Cartwright (1995), Knutz (2010) and Kilhams et al (2011) invoke a variety of marine processes, from solely downslope processes to bottom-currents, while Buckley (2012) inferred a subglacial origin. Thus in order to understand the processes involved in the formation of the troughs the palaeo-geographic setting must be considered Palaeo-geographic Setting Mapping of the basin at the onset of the Quaternary and the onset of the furrowing reveals an epeiric marine basin with water depths up to 400 m (Chapter 2). The Aberdeen Ground Formation is described as marine in origin and though it has not been described from core data to the depths observed in this study the lithological descriptions and gamma-ray logs from the well tie indicate background sediments are fine silts and clays (Figure 3.11) in agreement with descriptions from the shallower portions of the Aberdeen Ground Formation (Stoker & Bent 1985; Cameron et al 1987; Knudsen & Sejrup 1993). Well 30/6-3 shows a coarsening upwards sequence with input of silty sandstones towards the top of the section (Figure 3.11), characteristic of a shallowing marine basin due to the infill by the clinoforms. The presence of iceberg scours on the clinoforms confirms a glaciomarine environment. The North Sea is only considered to have been fully glaciated since the Elsterian, after the basin was completely filled (<0.5 Ma; Graham et al 2011) and the subglacial processes inferred to have carved out the troughs (i.e. tunnel valley formation; Buckley 2012) would rely on pressurized ~ 84 ~

86 Figure 3.9 Series of depth maps showing the structural trends of the basin during the furrowed interval. Each map shows depth to named horizon and 50 m contours (black). Grey area is depocentre, based on thickness maps, from named horizon to horizon below with exception of F01 which represents depocentre from base-quaternary to F01 horizon. Grey dashed contours represent the thickness contours of the depocentre every 30 m. Maps are overlain by the troughs which incise from that horizon. ~ 85 ~

87 meltwater flow beneath grounded ice sheets (Huuse & Lykke-Andersen 2000). Tunnel valleys are noted to usually incise into near-horizontal sediments (Huuse & Lykke-Andersen 2000). As the troughs considered here formed approximately Ma on clinoform surfaces in water depths up to 300 m, a subglacial origin is thus unlikely for the troughs considered here. Given the palaeogeographic setting on submarine clinoforms and the palaeo-glaciological history of the basin, we infer that a submarine origin of the troughs is most likely, echoing previous interpretations (Cartwright 1995; Knutz 2010; Kilhams et al 2011). The marine basin itself is elongate and both narrows and shallows towards the north leaving a small, shallow lip over which water could flow into and out of the basin connecting to the North Atlantic (Figure 3.3; Cameron et al 1987; Ottesen et al 2014). The shape and geometry of the basin is different from the present day marine basin limiting any comparisons to the present day tidally driven circulation system. In terms of overall geometry of the basin the earliest Quaternary basin is more similar to the early Cenozoic North Sea due to its overall alignment of greater water depths with the axis of the Mesozoic Central Graben (Fyfe et al 2003; Rasmussen et al 2005; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012). Bottom current scours have been noted in several stratigraphy intervals, including the Eocene/Oligocene boundary, the late Oligocene and middle Miocene (Huuse & Clausen 2001; Galloway 2002; Rundberg & Eidvin 2005; Stuart & Huuse 2012). The current system in the earliest Quaternary of the central North Sea is most likely to be a hybrid of the present day and early Cenozoic and thus unique. Inflow and outflow of water would occur primarily over the narrow sill between the northern North Sea and North Atlantic however the inflow of fresh water from the Baltic and NW European rivers as well as partially glaciated catchments in Britain and Norway would likely be more dominant than for the present day system. This implies that outflow of water would far exceed inflow over the sill. The primary infill of the basin, and thus inflow of fresh water, comes from the south creating the large clinoforms observed in the seismic data (Figure 3.12). This infill is likely to have been sourced primarily from the Rhine-Meuse and Baltic river systems which, during the earliest part of ~ 86 ~

88 Figure 3.10 Instantaneous amplitude extraction slice of F04 horizon with sketch interpretation. Identifying clinoform breakpoint and slope direction indicators, mounded, sub-parallel, slope features and iceberg scours marks. ~ 87 ~

89 the Quaternary, fed into the southern North Sea (Bijlsma 1981; Overeem et al 2001; Busschers et al 2007). The rapid (200 km in approximately 1 Ma) northward progradation of the clinoforms northwards (Figures 3.3, 3.7) is suggestive of a strong northwards flow of water and sediment over the topsets. The northern input of sediment is comparatively sparse, as demonstrated by the thickness maps (Figure 3.9) and thus implies a much weaker contribution in terms of water inflow Trough Geometry and Infill The geometry of type 1 troughs orientated perpendicular to clinoform strike, with low sinuosity and with repeated cycles of basal truncation and infill is strongly indicative of downslope processes. The size, geometry and repeated instances of cut-and-fill observed in the repeated truncations are all consistent with slope gullies on continental shelves (Shepard 1965; Field et al 1999; Stow & Mayall 2000; Spinelli & Field 2001). Although several different cross-sectional classes exist, the transitions between the dominant types a and b suggests that they are simply variations dependant on the intensity of the strong downslope flow relative to the sediment grade and deposition rate. The occurrence of the smaller cross-sectional class d suggests the infrequent occurrence of weaker, relatively short-lived currents. The majority of troughs bearing cross-sectional class d are found on the lower horizons F01 to F03 during which time the distance between the furrowed area and the clinoform slope was large, of the order of 100 km (Figures 3.3, 3.7) and presumably associated with relatively weak gravity currents. The observed pattern of basal truncation and infill, as well as the repeated cycles of erosion over the lifetime of a single trough seem consistent with examples of slope gullies and larger submarine canyons globally and have previously been interpreted to be connected to changes in sea level (Shepard 1981; Praston et al 1994; Field et al 1999; Spinelli & Field 2001) which may also explain the pattern of changing distribution and the diachronous formation of the troughs. As flow varies with time due to the rise and fall of sea level, currents may switch, leading to deactivation of one ~ 88 ~

90 Figure 3.11 Well correlation panels for wells 30/6-3 and 22/25a-3 correlating gamma log response, lithological descriptions and seismic for the background sedimentation within the furrowed interval (30/6-3) and the sedimentation within the infill of the trough (22/25a-3). Location on figure 2. Seismic data courtesy of PGS, well data courtesy of TGS.

91 trough and the formation of a new one adjacent to the abandoned trough (Cartwright 1995; Field et al 1999; Spinelli & Field 2001). Finally the similarity between the background sedimentation and the infill of the troughs observed in the well data suggests that the trough beds and infills are formed of the same material. Although the gamma response in well 22/25a-3 does suggest a more variable and slight fining of the sediment it does not indicate the addition of large quantities of coarse grained material which could be expected from other sedimentary processes but rather a reworking of the background sediment into giant channelized bedforms discussed herein. It is concluded then that the downslope-orientated type 1 troughs are formed by downslope gravity processes in agreement with Cartwright s (1995) model (Figure 3.12). Their overall geomorphology both in cross-section and long-profile, orientation, and infill agree with this model of formation. The along-slope orientated scours, whilst sharing many characteristics with type 1 troughs, require an additional process to explain their along-slope aspect Modification of Downslope Processes The increase in trough size and the change in orientation towards the eastern side of the basin are not easily explained by downslope processes alone. Although lateral variations in downslope features do exist for example along the margin of Equatorial Guinea (Jobe et al 2011) and the margin of New Jersey (Praston et al 1994) these are often controlled by sediment supply or the shape of the basin (Shepard 1965; Field et al 1999; Stow & Mayall 2000; Spinelli & Field 2001; Jobe et al 2011). In the early Quaternary basin the depocentre is consistently found towards the centre or western side of the basin suggesting that this area had maximum sediment input, as suggested by the higher density of troughs towards the western side of the basin. However if the supply of water and sediment was greater towards the west, why then are the type 2 furrows larger? Although a stronger current would allow for more erosion or non-deposition in the type 1 troughs the lower sedimentation rates away from the depocentre should likewise limit the depths of the type

92 2 troughs. In virtually all examples of slope gullies and downslope canyons changes in the orientation of the features follows the shape of the basin maintaining the features perpendicular to the strike of the slope (Praston et al 1994; Field et al 1999; Spinelli & Field 2001; Harris & Figure 3.12 Palaeogeographic reconstruction of the early Quaternary basin and the marine current system identifying anticlockwise circulation of shallow currents and strong downslope gravity currents. Based on results of this study and Cameron et al (1987), Huuse (2002) and Stuart & Huuse (2012). Whiteway 2011). Thus, if the type 2 troughs were controlled mainly by downslope currents, they should be deflected westwards down the slope of the eastern margin rather than eastwards along it (Figure 3.6). Other possible explanations such as structural control from the subsidence of the basin or salt tectonics or the presence of strong bottom or tidal currents, such as along the Atlantic margin ~ 91 ~

93 (Stoker et al 2005a) or in the Bay of Fundy (Shaw et al 2012), could be inferred to explain the type 2 troughs. However it would be expected that with these models there would be further evidence either in a changed infill of the type 2 troughs, which is not seen, or in the preservation of other features such as contourite drift deposits, which is also not seen. The mechanism by which the type 2 troughs are formed is likely, then, to be an addition to the existing downslope processes causing modification of the troughs rather than a completely separate process. It is proposed that, although downslope gravity currents from the southern clinoforms remained dominant, the type 2 troughs were modified by along-slope currents. It is evident that along-slope currents did exist in the Cenozoic basin due to the presence of mounded along-slope features in the Miocene to Pliocene and the earliest Quaternary (Figure 3.10; Andresen et al 2008; Stuart & Huuse 2012). These features are an order of magnitude smaller than the troughs which suggest that the currents that formed them were significantly weaker than the downslope currents, reworking less material and showing no indication of large scale erosion. However the current system which formed the Miocene to Pliocene along-slope features is inferred to be an anti-clockwise gyre (Andresen et al 2008; Stuart & Huuse 2012) similar to the present day current system in the North Sea which flows (Figure 3.1a; Otto et al 1990; OSPAR 2000). Considering the relatively close temporal proximity and the similarities in basin shape between the Pliocene and early Pleistocene, we infer that the north-eastward deviation of troughs in the eastern part of the early Pleistocene depocentre was linked with a geostrophic gyre forcing shallow water currents in the earliest Quaternary to flow northwards along the eastern basin margin deflecting gravity currents eastward, eventually leading to contour-parallel troughs. The larger-scale of the eastward-deflected bedforms (Figure 3.12) may be due to a combination of greater current strength relative to sedimentation rate and persistence of the tidal gyre system in times of low sediment input. Thus it is concluded that, due to their greater size and the change in orientation relative to the clinoform slope, type 2 troughs were formed by downslope currents deflected by along-slope currents. ~ 92 ~

94 3.5.4 Limitations of the Model Although the model presented in this study is considered to be the best possible model fitting to the evidence there remain several unanswered questions which are not explained: Up-slope migration is not seen in the type 2 troughs. As the downslope gravity current and the shallow along-slope current are seen to interact to magnify the deposition of the troughs it is arguable that up-slope migration would be more likely. The Coriolis force is the dominant cause of up-slope migration of current-formed features which, in the northern hemisphere, would deflect northwards flowing currents to the east and thus type 2 troughs should migrate up-slope. In comparison on the western margin the asymmetric c crosssectional class does migrate up-slope which suggests that the southwards flowing tidal gyre, which would be deflected to the west, is determining the migration of the troughs rather than the dominant northwards downslope currents. The complexity of inferring past current systems from preserved seismic-scale bedforms without any access to core-scale bedforms does not allow an unequivocal model to be established, but the overall framework of a northward directed gravity current of variable intensity interacting with a tidal gyre seems to explain most of the variation described herein. Feeder systems for the downslope system are not observed on any of the topsets and the delta system for the rivers has previously been identified well into the southern North Sea. Although topsets could have been eroded, the preservation of iceberg scours does not suggest this is the case and does not change the fact that the clinoforms are primarily mud dominated with no evidence for the delta to have reached so far north. Along a similar vein the first initiation of the troughs on F01 occurs when the slope itself is 150 km to the south. While it is possible for downslope-orientated features to form at a significant distance from the slope it does raise the question as to why it does so in this case. Why are there no troughs further south than N? How is the strength of the downslope current maintained over the large distances from sediment input points? ~ 93 ~

95 The role of type e class troughs is not clear. These cross-sectional forms only occur at the very end of active furrowing and following their formation furrowing ends very abruptly. Clinoform progradation continues for some time after the end of furrowing and although the clinoforms themselves are shallower the slope gradient does not appear to change. There is a gradual increase in coarser grained material which may be linked to the class e troughs. It is likely that the dramatic decrease in the size of the basin between the onset and the end of furrowing controls the decrease in trough formation and the troughs with cross-sectional class e may well be representative of the threshold at which the downslope system switches off. 3.6 Conclusions 380 individually mapped trough features hosted within a large-scale progradational succession were investigated in order to further understand sedimentary processes and palaeo-environments in the early Quaternary. The troughs were found to be distributed unevenly across the basin and could be split into two types based on size and orientation relative to the strike of the clinoform slope. Type 1 troughs are smaller and orientated perpendicular to clinoform strike and were found to be distributed on the western side of the basin, type 2 are larger and orientated almost parallel to clinoform strike on the eastern side of the basin. The palaeo-geographic setting and the geometry of the troughs strongly supports a model of incision and deposition by downslope gravity currents. However the differences between type 1 and type 2 troughs cannot be solely explained by lateral variations in downslope processes. A model of modification of type 2 troughs is presented in which the downslope current is amplified and deflected by along-slope currents caused by an anti-clockwise circulation of shallow water in a tidal gyre similar to the present day. Questions remain regarding the specific sedimentary processes leading to the giant bedforms of the clinoform-trough system and these can best be resolved by the retrieval of core material from the clinoforms and the trough fills. As the clinoforms record rapid infill of the basin during the early Quaternary by Baltic and NW-European rivers it constitutes an ~ 94 ~

96 important high-resolution record of early Quaternary palaeo-climate for mid-high latitudes of the Northern Hemisphere. 3.7 Acknowledgements The authors would like to acknowledge NERC studentship and BUFI in conjunction with Forewind for supporting this work, PGS for providing the CNS MegaSurveys, CGG for providing the BroadSeis TM survey, and TGS for providing well data through their Facies Map Browser. Thanks to Schlumberger for the donation of the Petrel software and Eliis for the Paleoscan software. We would like to thank Martyn Stoker for a very thorough and helpful review which ultimately led to a greatly improved paper. ~ 95 ~

97 Chapter 4: The evolving palaeo-environments of the Pleistocene of the central North Sea: 2.58 to 0.48 Ma Rachel M. Lamb, Mads Huuse, Margaret Stewart, Simon H. Brocklehurst Abstract: The early Quaternary represents a period of dramatic climatic change with the onset of intensification Northern Hemisphere glaciation and the Mid-Pleistocene transition from 41 kyr glacial-interglacial cycles to 100 kyr cycles. In the North Sea the early Quaternary has not been studied in detail since initial mapping work during the 70 s and 80 s despite a step change in the coverage and resolution of available seismic data. This study uses semi-automated seismic mapping and basin reconstruction techniques to produce a new detailed chronology and a revised seismic stratigraphy. The Aberdeen Ground Formation, which previously incorporated the early Quaternary stratigraphy from 2.58 to 0.48 Ma, is split into four member units representing changes to the marine basin through time. Intensive geomorphological mapping then allows the palaeoenvironmental reconstruction of the North Sea basin from the Gelasian (2.58 to 1.78 Ma; MIS ) to the regional Ling Bank unconformity (0.48 Ma; MIS 12). The results suggest the development of the basin is controlled by local sea level change due to the infill of the basin as well as the global cooling trend that occurs throughout the early Quaternary. The study presents the first evidence for isolated lowland glaciation in the central North Sea at 1.72 Ma (MIS 60), significantly prior to any previous study. The onset of widespread, repeated lowland glaciation is revised to approximately 1.3 Ma (MIS 40) rather than the previously assumed Elsterian (0.48 Ma; MIS 12). ~ 96 ~

98 4.1 Introduction The study of the palaeo-environmental evolution of the Quaternary Period is the study of rapid and dramatic climate change, of glacial and interglacial cycles (Alley 2000; Lisiecki & Raymo 2007; Steffensen et al 2008). Understanding the character of rapid changes in climate in the past is helping scientists to refine predictive models for the magnitude and effects of future anthropological climate change (Braconnot et al 2012; Martinez-Boti et al 2015). The early Quaternary has two key features which make it an interesting period for the study of rapid climate change: firstly it is during this period, from the Plio-Pleistocene transition (2.7 to 1.8 Ma), when Northern Hemisphere glaciation intensified (Raymo 1994; Lisiecki & Raymo 2005, 2007; Miller et al 2011); and secondly the periodicity of the glacial-interglacial cycles switched from 41 kyr to 100 kyr during the Mid-Pleistocene transition (MPT; 1.2 to 0.8 Ma) (Raymo & Nisancioglu 2003; Lisiecki & Raymo 2005; Gibbard & Lewin 2009; Miller et al 2011). A number of potential causes and explanations exist for these two phenomena (Raymo & Ruddiman 1992; Raymo & Nisancioglu 2003; Lisiecki & Raymo 2007). The effects of global cooling trends, such as during the Plio-Pleistocene transition in the early Quaternary, have been observed previously to be preserved most strongly in mid-to-high latitude basins (Mudelsee & Raymo 2005; Rohling et al 2012). The North Sea is one such mid-latitude basin, and during this period was the depo-centre for the Rhine-Meuse and Baltic river systems (Bijlsma 1981; Sørensen et al 1997; Overeem et al 2001; Busschers et al 2007; Moreau & Huuse 2014). The river systems drained a large portion of Northern Europe and Scandinavia (Overeem et al 2001) resulting in a high rate of continual of sedimentation creating an expanded record for the early Quaternary of up to 1.2 km of sediment (Holmes 1977; Cameron et al 1987; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014; Chapter 1). A great deal of work was done on the early Quaternary of the North Sea during the 70 s and 80 s however this was limited by the type and quantity of data available, primarily widely-spaced regional 2D seismic and shallow core <250 m deep (Holmes 1977; Stoker et al 1983; Stoker & Bent 1985; Cameron et al 1987; Sejrup et al 1987, 1991; ~ 97 ~

99 Knudsen & Sejrup 1993; Sejrup et al 1994). Absolute dating of the section to establish a detailed chronology was equally limited by the data available and much of the early dating relied on palaeomagnetic data from shallow cores (e.g. Stoker et al 1983) or on interpretation of biostratigraphic zones (e.g. King 1983), both of which have been significantly revised in recent years (e.g. Kuhlmann et al 2006; Noorbergen et al 2015). The increase in coverage by continuous 3D seismic data (Brzozowska et al 2003; Harvey et al 2010) now allows for the conclusions of this work to be revisited and revised according to the more expansive data available. Figure 4.1 Location map and datasets showing 3D seismic surveys used for the study (CNS MegaSurvey & CGG BroadSeisTM survey), key wells used in study including Josephine-1 and A15-03, other wells used to supplement the study, the location of the seismic section in figure 2 and the locations of points along the transect used for back-stripping in figure 4. Bathymetry and topography data from Ryan et al (2009). ~ 98 ~

100 This study examines the 1.2 km package of Quaternary sediment across the central North Sea (CNS), between N and N, attempting to interpret a more comprehensive stratigraphy correlated to a detailed chronostratigraphy from well data. The study uses seismic geomorphological mapping to demonstrate the evolving palaeo-environments of the early Quaternary building on the history of expansion and retreat of the British and Fennoscandian Ice Sheets (BIS, FIS) prior to the onset of 100 kyr glacial cycles of the Middle and Late Pleistocene. 4.2 Regional Setting The present day North Sea is an epicontinental sea bordered on three sides by the Scandinavian, European and British landmasses, with average water depths of around 200 m (outside the Norwegian Channel/Trench) (Figure 4.1). Throughout the Cenozoic, the North Sea has undergone continuous thermal subsidence with periods of basin inversion (Ziegler 1992; White & Lovell 1997; Stoker et al 2005). At the beginning of the Quaternary (2.58 Ma) the central and southern North Sea were joined in a single elongate basin over 600 km long and approximately 100 km wide orientated NW-SE over the older structural feature of the central graben (Holmes 1977; Cameron et al 1987; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014; Chapter 2). Water depths in the earliest Quaternary (2.58 to 2.35 Ma) were 300 ± 50 m but sedimentation rates of <2.5 mm yr -1 in the southern end of the basin led to infill of the basin from the south (Cameron et al 1987; Rasmussen et al 2005; Ottesen et al 2014; Chapter 2). The infill of the southern half of the basin was the result of constant sedimentation derived primarily from the Rhine-Meuse and Baltic river systems building clinoformal geometries in the seismic stratigraphy on the older Pliocene deltaic stratigraphy (Overeem et al 2001; Huuse 2002; Busschers et al 2007; Moreau & Huuse 2014). In the CNS the same clinoform set progrades northwards during the Quaternary contributing the largest and most significant amount of sediment to the infill of the CNS basin (Figure 4.2). The secondary influx of sediment consists of a clinoform set which progrades southwards from the north-western margin of the basin, most likely sourced ~ 99 ~

101 from Sottish river systems (Figure 4.2). Minor influx from the east, sourced from the Baltic Sea and southern Scandinavia (Anell et al 2012), and west, sourced from southern English river systems such as the Thames (Gibbard 1979; Gibbard & Lewin 2003), is probable but does not form significant clinoformal geometries. In the formal CNS classification, the stratigraphy from the onset on the Quaternary (2.58 Ma) to the regional Ling Bank Unconformity (approximately MIS 12; 0.48 Ma) has been grouped into one formation, the Aberdeen Ground Formation (AGF) (Stoker et al 1985, 2011; Cameron et al 1987; Sejrup et al 1991; Gatliff et al 1994). The AGF is described as a thick sequence of pro-deltaic to shallow-marine sediments in the central and northern North Sea (Stoker et al 2011, p. 8) and has been noted in the shallow section to contain evidence for cold conditions and thus a glaciomarine palaeo-environment (Stoker & Bent 1985; Cameron et al 1987; Sejrup et al 1987, 1991; Knudsen & Sejrup 1993). The descriptions of the AGF from core data as glaciomarine are supported by evidence for repeated instances of iceberg scouring from the onset of the Quaternary (Kuhlmann & Wong 2008; Dowdeswell & Ottesen 2013; Ottesen et al 2014; Appendix 2) in agreement with the timing of the onset of Northern Hemisphere glaciation. The widespread presence of glacial landforms, such as tunnel valleys (Praeg 1996, 2003; Stewart & Lonergan 2011; Kristensen & Huuse 2012; Stewart et al 2012, 2013; Moreau & Huuse 2014), mega-scale glacial lineations (Sejrup et al 2003; Graham et al 2007), glaciotectonics on micro- and macro-scales (Huuse & Lykke-Andersen 2004; Andersen et al 2005; Carr et al 2006; Buckley 2012) and moraine features (Bradwell et al 2008) across the North Sea provides direct evidence for extensive periods of glaciation during the 100 kyr glacial-interglacial cycles of the late-middle and Late Pleistocene. However evidence for repeated lowland glaciation is limited in the CNS prior to the Elsterian glaciation (MIS 12; 0.48 Ma) and almost non-existent before 1.1 Ma and the MPT. It would not be unreasonable for some glacial activity prior to the 100 kyr cycle, though it would be expected to be smaller in both scale and impact on the basin due to the shorter 41 kyr cycle. ~ 100 ~

102 Figure 4.2 Seismic section and line interpretation X-X at 75x vertical exaggeration, showing the structure of the two principal clinoforms sets, identifying the key horizons sh1 to sh9 and nh1 to nh2 as well as the 2.58, 2.35, 1.9, 1.1 and 0.78 Ma dated surfaces. Indicates location of well 22/18-3 and distinguishes four AGF member units A-D as described in this study. 2b inset (next page: 15x VE) shows close view of 22/18-3 showing AGF member units and identifying minor horizons produced from semiautomated mapping techniques used for geomorphic mapping and sedimentation rate calculations.

103 However the lack of data coverage and poor chronological control has prevented the description of any preserved evidence. 4.3 Datasets and Methods This study utilised two continuous 3D seismic datasets (Figure 4.1): the PGS central North Sea MegaSurvey, which resolves to 10 to 15 m vertically with a horizontal bin spacing of 50 m (effectively providing horizontal resolution); and the CGG BroadSeis TM dataset with a vertical resolution of 7 to 8 m and horizontal bin spacing of 12.5m. Together they cover an area of approximately 88,000 km 2. In addition to the seismic data, 108 wells from the TGS Facies Map Browser have been used to classify and interpret lithology. Mapping of major horizons followed standard seismic stratigraphical technique (Mitchum et al 1977a, b; Posamentier et al 2007), correlating chronostratigraphic data from cores (e.g. palaeomagnetic data from Stoker et al 1983) and wells (e.g. biostratigraphic correlations from Kuhlmann et al 2006) to the seismic data in order to build an accurate stratigraphy and illustrate the evolving basin shape. In total, eleven major horizons were mapped corresponding to widespread prominent reflections in addition to the base Quaternary surface (2.58 Ma), base-jaramillo magnetic reversal equivalent (1.08 Ma) and Brunhes-Matuyama magnetic reversal equivalent (0.78 Ma), each one representing a major progradation maxima and therefore a highstand sequence boundary (Mitchum 1977a), split across two major clinoform packages. The first clinoform package in the south is ~ 102 ~

104 divided into horizons from sh1 to sh9, the second northern clinoform set has two major horizons, nh1 & nh2, which intersect with the southern clinoforms (Figures 4.2 & 4.3). Intermediate horizons for geomorphological mapping and chronostratigraphic calibration were extracted between the conventionally interpreted horizons using semi-automated techniques with the Paleoscan software in which a geological model is built from the seismic data and control surfaces, and a horizon stack is extracted, producing horizons for every seismic reflection. 63 horizons were then manually chosen from the stack such that the entire package was covered by regularly spaced horizons with minimal noise and maximum resolution of notable features, to give a consistent, high standard coverage of the entire basin. Dating of the horizons, major and minor, was done using absolute dates from integrated chronostratigraphic studies where possible, however the lack of chronostratigraphic control between the few available key dates makes a robust interpretation difficult. The study uses an alternative method in which the gamma response of the sediment from seven wells (Norwegian sector well 2/7-31 and UK sector wells 31/26a-12, 30/18-4, 29/9a-1, 22/24b-7, 22/18-3 and 21/20b- 3) were compared to the global sea level curve from Miller et al (2011) (Figure 4.4). This method is explained in full in section 4.4. To further understand the evolution of the basin thickness maps were produced between major horizons (Figure 4.3) in order to identify the migrating depo-centre and find clinoform breakpoints. Average sedimentation rates were calculated from the thickness between each dated intermediate horizon. The horizons were depth converted (after Chapter 2), the thickness between each horizon found and from that the volume of the depocentre calculated. This volume was then converted to a solid volume to account for sediment compaction, using the average depth of burial for each surface and a depth-porosity curve determined from North Sea porosity data (Marcussen et al 2010). Using the minor horizons for these calculations gives an average resolution of ~30,000 years for the entire package from 2.58 to 1.08 Ma. In a similar way, to understand the configuration of ~ 103 ~

105 the basin through time, decompaction and backstripping of major horizons was performed using a simplified methodology after Allen & Allen (1990) detailed in Chapter 2, in which packages separated by horizons are progressively decompacted along transects, beginning with the youngest/shallowest, and then back-stripped using an Airy isostasy model in order to restore surfaces to their initial position at time of deposition (Figure 4.5). Mapping of the geomorphology was done through mapping of features across horizon slices. Traditionally, 3D seismic geomorphological mapping has taken place using amplitudes extracted along time slices (e.g. Praeg 1996, 2003; Andersen et al 2005; Stewart et al 2012, 2013), allowing for the 3D geometry to be fully resolved. However, time slices are limited where clinoforms are inclined as they cut through multiple clinoforms and thus multiple timelines (Mitchum 1977a; Posamentier et al 2007; Harding 2015). By using horizon slices to image geomorphic features it can be assured that all mapped features are contemporaneous (Posamentier et al 2007). Instantaneous amplitude and RMS amplitude was extracted from a 5 ms window around the horizon (11 major and 63 intermediate) in order to image preserved geomorphic features. The mapped features were then used to interpret the palaeo-environments based on their formation. For example: iceberg scours tend to occur within a limited depth range between free-floating and grounded ice and are used to determine the extent of the shelf; terrestrial channels determine the minimum position of the coastline; glaciotectonic structures form at ice margins and determine maximum ice sheet extent; mega-scale glacial lineations are formed under ice streams and represent minimum ice sheet extent. 4.4 Chronostratigraphic correlation Dating of the major horizons in order to understand the temporal evolution of the North Sea basin relies in the first instance on the integrated chronostratigraphy from the SNS. Horizons dated by Kuhlmann et al (2006), Noorbergen et al 2015 and Harding (2015) in the Dutch sector, specifically on well A15-03, can be mapped northwards with a high degree of confidence. These datings have ~ 104 ~

106 ~ 105 ~ Figure 4.3 Series of TWT thickness maps between surfaces from 2.58 Ma to 0.78 Ma showing location of key wells used for chronostratigraphic calibration and approximated clinoforms breakpoints derived from seismic crosssection, TWT structure maps and thickness maps. Solid black lines refer to clinoform breakpoint at the beginning of each package and dotted lines to the clinoform break point at the end of the package

107 ~ 106 ~

108 been found by Harding (2015) to correlate well with other SNS studies such as Meijer et al (2006), ten Veen et al (2013) and Thöle et al (2014). A combination of pollen analysis, palaeo-magnetic data, biostratigraphy, well logs and sequence stratigraphy identifies several confidently dated events which correlate to the major horizons used in this study; for example sh1 correlates to 2.44 Ma, sh2 to 2.35 Ma and sh3 to 1.94 Ma. The date for the Réunion palaeo-magnetic Event at ~2.15 Ma (Kuhlmann et al 2006) has been used independently to test chronological control; however this event does not correlate to any of the major horizons. In conjunction, palaeo-magnetic data published on BGS core 77/02 and others by Stoker et al (1983) allows for mapping of the base of the Jaramillo palaeo-magnetic Event at 1.08 Ma and the Brunhes-Matuyama reversal at 0.78 Ma (Table 4.1). In the absence of core or other data to assist in identifying and dating stratigraphy between these dates proves difficult. In the SNS the structure of the clinoforms, as well as the preservation of slope systems, can be compared to the global sea level curve to identify sequences of highstands and lowstands, creating a relative chronostratigraphy in the absence of more accurate proxies (Sørensen et al 1997; Overeem et al 2001; Kuhlmann & Wong 2008; Harding 2015). In comparison, the CNS has very few slope systems such as channels, mass transport deposits and forced regression wedges preserved. Figure 4.2 shows many of the CNS clinoforms decreasing in slope angle to the point where differentiating between topsets and slope and pinpointing the breakpoint is difficult (Figure 4.2, horizons sh3-6). Thus, sequence stratigraphy cannot be confidently used to derive relative age (Harding 2015). An alternative method is applied here by utilising the apparent relationship between sea level and the gamma response of the sediments in order to correlate reflections to the global sea level curve. Gamma logs from hydrocarbon wells are widely available and usually have extremely high vertical resolution in comparison to the seismic data, allowing for very small fluctuations in lithology to be resolved. In basins dominated by clastic sedimentation such as the North Sea, gamma logs reflect the grain size distribution of the sediments a relative high gamma count indicates a relative ~ 107 ~

109 decrease in grain size while an increase in the gamma response suggests a fining of the sediment (Rider 2002; Noorbergen et al 2015). A strong relationship between the gamma response of sediments and sea level has previously been observed in the SNS (Noorbergen et al 2015). During lowstands when cooler temperatures dominate Scandinavian ice advances, physical weathering increases and the width of the delta and shelf increase (Noorbergen et al 2015). The coarse sediment fraction therefore is decreased in proportion due to the increased weathering and is less likely to reach the main depocenter due to long transport distances (Kuhlmann & Wong 2008; Noorbergen et al 2015). During highstands the Scandinavian ice retreats and there is an increase in the amount of less mature material from north-west Europe in proportion to the more-mature sediments from Scandinavia across shorter (Noorbergen et al 2015). The coarse fraction increases in the sediment and short transport distances across the shelf due to the highstand means more of the coarse fraction reaches the main depocentre (Kuhlmann & Wong 2008; Noorbergen et al 2015). The relationship between gamma response and sea level is best preserved on the slope, where maximum sedimentation occurs even where standard sequence stratigraphical relationships are not easily observed. In the CNS, thickness maps between the major horizons were examined and wells were selected according to shortest distance to the point of maximum thickness for each map and thus the best preservation of the gamma response. For example the package between the base Quaternary and sh2 shows an elongate depocentre in the south of the study area to which well 2/7-3 is closest (Figure 4.3a). The base-quaternary and sh2 are both correlated to absolute dates, 2.58 and 2.35 Ma respectively, and so the gamma log between the two absolutely dated horizons were compared to the applicable section of the global sea level curve (Miller et al 2011). Five peaks in the gamma log are observed in this section as well as five lows (Figure 4.4a) and although sequence stratigraphical relationships are difficult to observe the correlation between coarsening and sea level fall is in agreement with the evidence from the Danish sector (Kuhlmann et al 2006; Kuhlmann & Wong 2008; Noorbergen et al 2015; Harding 2015). The next package, from 2.35 to 1.94 Ma (sh2 to sh3; Figure 4.4b), was examined in a similar way to test the methodology using well 31/26a-12. In this ~ 108 ~

110 ~ 109 ~ Figure 4.4 Series of seismic well panels used to produce chronostratigraphic calibration. Each panel demonstrates seismic well tie, well gamma log and global sea level curve from Miller et al (2011) from either a-b) 2.7 to 1.8 Ma c-g) 2.0 to 0.8 Ma. Correlation lines are between peaks in gamma and sea level lows. Note the condensed section from 2.35 to ~2.2 Ma in panel b

111 package a condensed section is observed where the clinoforms pinch out towards 31/26a-12 and is absent at the well itself (Figure 4.4b). This condensed section has been dated through correlation to the Réunion Event (Kuhlmann et al2006) in the SNS to between 2.35 and 2.15 Ma. The correlation between gamma response, seismic stratigraphy and global sea level is strong in the section of the package between 2.15 and 1.94 Ma despite the hiatus and so the methodology was then applied to the remaining packages. As sh3 is correlated to 1.94 Ma the package from sh3 to sh4 was used to find the age of sh4 by correlating the gamma log to the sea level curve in the same fashion as the previous packages. The same method was applied to the sh4 to sh5, sh5 to sh6, sh6 to sh7 and sh7 to Jaramillo Event packages. By making correlations this way every lowstand within the base Quaternary to Jaramillo package could be correlated to a gamma peak. Following this, each intermediate horizon from the base Quaternary to the Jaramillo Event was then dated by correlating the horizon to the gamma log and identifying the sea level event, and thus date, associated with it. The above method provides age estimates only. Without further chronostratigraphic calibrations to test the packages between sh3 and sh7, the possibility of another condensed section cannot be dismissed. This method also does not account for other controls on the gamma log such as changes in sediment composition such as clay mineralogy which may give the appearance of fining or coarsening of the sediment without the associated sea level change. More accurate dating would require detailed integration of multiple proxies such as palaeo-magnetics, pollen analysis and biostratigraphy on core data. We provide estimate of ages and a relative chronology from which a stratigraphic framework can be produced, to be revised upon availability of further chronostratigraphic work. In this manner horizons sh4 to sh9 were dated to 1.72, 1.57, 1.44, 1.17, 1.16 and 1.13 Ma respectively (Figure 4.4; Table 4.1). Horizons nh1 and nh2 were dated to 1.94 and Ma respectively. Once this was completed, minor horizons could also be dated to allow for relative positioning of preserved features once mapped. ~ 110 ~

112 4.5 Basin Evolution The initial marine basin at the onset of the Quaternary has previously been mapped and described in detail revealing a deep elongate basin largely located above the older structural lineament of the Central Graben (Holmes 1977; Cameron et al 1987; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014; Chapter 2). This basin is more than 600 km long and 100 km wide at its maximum, and is oriented NNW- SSE, except north of N where axis orientation changes to NE-SW. Estimates of sea level from backstripping (Chapter 2; Figure 4.5) suggest maximum water depths in the region of 300±50 m at the onset of the Quaternary. From the 2.58 to 1.08 Ma horizons, two principal depocentres are observed, corresponding with two sets of clinoforms (Figure 4.3). The southern clinoform set is a continuation of the prograding clinoforms of the Pliocene and progrades firstly to the south-west, as evidenced by both migration of the depocentre and the clinoform breakpoint, until the width of the basin is almost infilled (Figure 4.3a-d). At approximately 1.94 Ma, Figure 4.3d, the progradation direction switched towards the north-west and the position of the depocentre adjusted accordingly (Figure 4.3d). The progradation of the southern clinoforms determined basin shape and by 1.57 Ma the breakpoint indicates the orientation of the basin axis has moved through from NNW-SSE to NE-SW (Figure 4.3e) with the basin narrowing further until 1.16 Ma at which point clinoform progradation ended and the breakpoint became indistinguishable (Figure 4.3j). The northern clinoform sets are smaller in terms of volume, migration of the depocentre and progradation of the breakpoint. During some periods, for example 1.94 to 1.72 Ma (Figure 4.3c), deposition is so low from the north that at the basin scale it appears to be non-existent. Overall, this clinoform set progrades approximately 60 km between 2.58 and 1.16 Ma compared to the >200 km of the southern clinoform set. After 1.16 Ma, the depocentres for the two clinoform sets met and merged and progradation appears to have ceased (Figure 4.3k). Deposition is focused in a shallow, narrow basin until 1.07 ~ 111 ~

113 ~ 112 ~ Figure 4.5 Backstripping transect following the seismic section in figure 2. 5a) Simplified present seismic stratigraphy 5b) Structure of backstripped basin from present day to 1.44 Ma. 5c) Structure of back-stripped basin from 1.57 to 2.58 Ma. Backstripping is estimate only and hills in results for 1.44, 1.72 and 2.58 Ma surfaces most likely caused by lack of modelling of flexural effects.

114 Ma and the base-jaramillo equivalent. After this point, deposition becomes widespread and relatively equable across the entire study area until the Brunhes-Matuyama reversal at 0.78 Ma (Figure 4.3l). Decompaction and backstripping of the basin broadly agrees with observations made from the seismic data (Figure 4.5). At the onset of the Quaternary a marine basin of >300 m existed with the deepest part of the basin located towards the south. Through the early Quaternary the basin shallowed differentially with more rapid filling of the basin in the south so that by 1.57 Ma average water depths in the southern half of the basin were <100 m. The northern part of the basin maintains evidence for deeper conditions until 1.07 Ma, after which water depths do not exceed approximately 150 m anywhere in the basin and the basin became relatively flat, similar to the present day North Sea. The backstripping calculations must be treated with caution as they are estimates only and do not account for eustatic sea level changes or flexural effects between individual points on the transect (Allen & Allen 1990; Pekar et al 2000; Patruno et al 2014). This is evidenced by the presence of several anomalous small hills namely in the 1.44, 1.72 and 2.58 Ma surfaces as well as depressions 1.17 Ma surface (Figure 4.5). Likely the backstripping results represent the pattern of basin evolution only rather than precise depths and more accurate calculations are needed. The combination of basin infill and the backstripping calculations however allow for some conclusions about the basin development to be made and for the stratigraphy to be revised. The evidence above suggests that progradation into the early Quaternary marine basin was continuous from 2.58 Ma until approximately 1.17 Ma. After 1.17 Ma the basin became too shallow to allow progradation to continue, and was most susceptible to sea level change related to glacial-interglacial cycles, with the possibility of widespread sub-aerial depositional environments. A relatively shallow basin was maintained until post the Jaramillo palaeo-magnetic Event at 1.07 Ma at which point two events occur. Firstly, the basin was completely infilled and all reflections ~ 113 ~

115 after that point, baring the truncations from the later tunnel valleys, are near horizontal and parallel, reflecting continuous sedimentation across a wide, flat plain; secondly the 100 kyr glacialinterglacial cycles begin to dominate and the magnitude of global sea level change increased, indicating that while the North Sea could have flooded during this time, environmental conditions may have been terrestrial for significant periods of time. In this way the AGF has been split into four member units, A B C and D (Figure 4.6). Member A is the largest unit within the AGF and is defined by the clinoformal progradation sequence in the south of the basin. Member A is a continuation of the Pliocene Southern North Sea Deltaic Group (Stoker et al 2011) consisting of low to medium amplitude seismic horizons with a clinoformal geometry and the sediments are sourced from Northern Europe through the major river systems. Based on gamma logs and well cutting descriptions from the investigated wells, Member A most likely comprises primarily of muds and silts. Member B is the northern equivalent to member A and consists of similar seismic facies, but these are sourced from the British landmass. Between Member B and the older stratigraphy below there is a significant hiatus; deposition did not commence until approximately 1.94 Ma (Figure 4.6). Member B shares many characteristics with member A with no significant difference in either seismic facies or well log data in the bulk of the package suggesting a primary composition of muds and silts. Towards the top of member B however, gamma logs suggest a gradational increase in the proportion of sand and seismic sections reveal several high amplitude, possibly sandy reflections which, from the mapped geometry, are likely sourced from the north. Members A and B interfinger in the centre of the basin (Figure 4.2). Member C is interpreted to comprise primarily shallow marine to terrestrial deposits from 1.17 to 1.0 Ma based upon the gradual coarsening implied by gamma logs and the shallowing of the basin from the backstripping results. This member is the final infill of the basin after progradation ends, likely due to the reduced accommodation space available as the basin shallows. Member C continues the increase in sandy content with at least two localised high amplitude deposits sourced from the north and can be considered a transitional member. ~ 114 ~

116 Finally, member D (1.0 to 0.48 Ma) is primarily glacial terrestrial to glacio-marine in nature. In seismic character these sediments are near horizontal and parallel except where truncated by the Ling Bank Unconformity which caps the formation. The reflections maintain the same mid to low amplitude as the lower stratigraphy with intermittent high amplitude reflections. Member D is the only member to be fully investigated by shallow core data from the original BGS studies of the AGF (Holmes 1977; Stoker 1983; Stoker & Bent 1985; Cameron et al 1987; Sejrup et al 1987, 1991, 1994; Knudsen & Sejrup 1993). Sediments are therefore interpreted to be dominantly muddy-silt with intermittent sands based on the seismic character and core descriptions. 4.6 Geomorphological Mapping Geomorphological analysis incorporated the mapping of features such as iceberg scours, alongslope contourites, downslope troughs, slope channels, mega-scale glacial lineations (MSGLs), glaciotectonics, tunnel valleys, terrestrial channels and beach deposits. The preserved features are grouped into four classifications: marine-glacial; marine-non glacial; terrestrial-glacial; and terrestrial-non glacial. Figure 4.7 shows a series of panels identifying each feature on a horizon slice using either an amplitude extraction or RMS extraction. Table 4.1 Chronostratigraphic correlation table identifying the dates of major horizons used in this study. Also including chronostratigraphic markers used to date each horizon Horizon Date (Ma) Chronostratigraphic Marker Base Quaternary 2.58 A15-03 (Kuhlmann et al 2006) Southern Clinoforms sh A15-03 (Kuhlmann et al 2006) sh A15-03 (Kuhlmann et al 2006) sh A15-03 (Kuhlmann et al 2006) sh Gamma-sea level correlation (this study) sh Gamma-sea level correlation (this study) sh Gamma-sea level correlation (this study) sh Gamma-sea level correlation (this study) sh Gamma-sea level correlation (this study) sh Gamma-sea level correlation (this study) nh A15-03 (Kuhlmann et al 2006) nh Gamma-sea level correlation (this study) Northern Clinoforms Base-Jaramillo Event /08 (Stoker et al 1983) Brunhes-Matuyama Reversal /08 (Stoker et al 1983) ~ 115 ~

117 Figure 4.7a-d represents the marine non-glacial features and can be further classified into slope and current features. 4.7a identifies long arcuate forms on horizon slice which, in cross-section, appear as small-scale clinoforms imprinted on the larger basin-scale clinoforms. Features such as these are common in the Pliocene section of the SNS (Harding 2015) and are referred to as forced-regressive wedges of intra-slope deltas, representative of local scale sea level fluctuations within the higher order systems (Harding 2015). In this study the forced-regressive wedges are limited in extent, existing in the study area only in the southern most parts of the basin prior to 2.35 Ma. Figure 4.7b shows meandering and dendritic channels which run parallel to the strike of the palaeo-slope, usually initiating halfway down the slope and terminating at the basin floor, often with a fanshaped change in seismic facies at the termination point. These features are interpreted as slope channels which deliver material from the marine topsets to the basin floor. Similar to the forced regressive wedges few sloped channels are preserved in the study area and all prior to 2.35 Ma. In contrast, the features seen in Figure 4.7c, which are elongated to tear-dropped shaped in horizon slice and overlapping, mounded features in cross section, are seen regularly from 2.58 Ma onwards. They form subparallel to palaeo-slope contours with the teardrop shape consistently orientated in the same direction. These features are interpreted as along-slope contourites. They appear very similar in size and form to along-slope contourites seen in the SNS throughout the Cenozoic (Andresen et al 2008; Stuart & Huuse 2012). The elongate, semi-sinuous troughs shown in 7d have previously been described and interpreted in detail most recently in Chapter 3 but also Cartwright (1995), Knutz (2010), and Kilhams et al (2011). They are believed to be formed by strong downslope currents flowing northwards perpendicular to the strike of the clinoform slope (Cartwright 1995; Chapter 3). Figure 4.7e represents the single type marine glacial features observed. The multidirectional, linear and curvilinear features are characteristic of iceberg scour marks and have been identified on 3D ~ 116 ~

118 ~ 117 ~ Figure 4.6 Revised stratigraphy of the early Quaternary for the central North Sea indicating the stratigraphy, new member units for the AGF, horizons used in this study, mapped geomorphology, calculated average sedimentation rate and comparison to the Naust Formation sequences, global sea level and MIS stages.

119 seismic globally but are described in detail in the North Sea by Kuhlmann & Wong (2008), Dowdeswell & Ottesen (2013), Ottesen et al (2014) and Newton et al (Appendix 2). Figure 4.7f and 4.7g represent the terrestrial non-glacial forms. 4.7f describes a series of interlocking arcuate features which in horizon slice build up to form meandering, dendritic channels primarily along two major axes. These features appear in cross-section to be small constructional bars and are located only on the topsets of the clinoforms. The features are interpreted to be preserved bar deposits from meandering terrestrial channels. Figure 4.7g shows an elongate arc following the contours of the palaeo-slope for more than 100 km close to the base of the slope with a maximum width of 2 km. The entire length of the arc has anomalously high seismic amplitude but otherwise has no vertical expression in cross-section. Small overlapping breaks in the arc can be observed towards the south and a change in the seismic facies on either side is prevalent towards the south. Given the likely depositional setting and comparison to other observed features it is considered the most likely interpretation is a shoreline or beach deposit, although for a more certain interpretation a closer examination or higher resolution seismic data would be required. Figure 4.7h-j represents the terrestrial glacial features which mostly occur during the latter part of the stratigraphy. 4.7h describes two forms of elongate curvilinear scours, often tens of kilometres in length. The two forms are distinguished by size only, the first are narrow (<2 km) and have no vertical expression on the seismic cross section, and the second are wider (<5 km) with a simple v shaped cross-section (<20 ms TWT). The scours form in sets which are generally semi-parallel but may diverge slightly towards the south, creating a series of lineaments across the basin. Some of the features are associated with seismic facies changes in the surrounding sediment. These features are interpreted as mega-scale glacial lineations after Ottesen et al (2002), Sejrup et al (2003), Stoker & Bradwell (2005) and Graham et al (2007). Figure 4.7i describes a large arcuate complex made up of smaller lineaments which may be semiparallel or convergent. In cross-section at the higher resolution of the BroadSeis dataset small thrust faulting can be observed directly connected to the small lineaments. Though the thrusts ~ 118 ~

120 cannot be directly observed in the MegaSurvey the location of the lineaments can be directly compared to those observed by Buckley (2012), who demonstrated a series of thrusts in high resolution 2D seismic in the same location. Figure 4.7i is considered to show the planform texture of large glacio-tectonic thrust complexes. Figure 4.7j describes the large erosional channels of the Ling Bank Unconformity which separates the AGF from the Late Pleistocene stratigraphy. These features, well established as tunnel valleys, are Middle to Late Pleistocene features (Praeg 1996, 2003; Stewart & Lonergan 2011; Kristensen & Huuse 2012; Stewart et al 2012, 2013; Moreau & Huuse 2014). Due to their complex fill tunnel valleys also often associated with strong velocity effects in the underlying seismic which may penetrate well into the early Quaternary stratigraphy (Graham et al 2007; Stewart et al 2012), truncating or hiding features that otherwise may have been preserved there. This problem is particularly strong in the very upper part of the AGF, making it particularly difficult to map channel features which may, on horizon slice, appear superficially identical to tunnel valleys. The mapped features were dated according to the established chronostratigraphy (Figure 4.6) and grouped to produce distribution maps of features seen in each of four time periods, the Gelasian (2.58 to 1.78 Ma; Figure 4.8a), the early Calabrian (1.78 to 1.4 Ma; Figure 4.8b), the late Calabrian (1.4 to 1.0 Ma; Figure 4.8c) and post Jaramillo (1.0 to 0.48 Ma; Figure 4.8d). Figure 4.8 shows these maps along with the progradation of the clinoforms during each period. The Gelasian (Figure 4.8a) is characterised primarily by the decline in slope channels, of which only two are observed in the CNS, and the onset of the strong downslope currents. Iceberg scouring is extremely extensive during this period, particularly during highstands, however few terrestrial channels are observed. Well data suggests surrounding sediments are primarily muddy sediment with infrequent silt and little evidence for sands. ~ 119 ~

121 Figure 4.7 Horizon slice panels used for identifying, mapping and describing geomorphology. Separated by palaeo-environmental setting a-d) marine non-glacial e) marine glacial f-g) (next page) terrestrial nonglacial h-i) (next page) terrestrial glacial. 7a, b & e represent RMS amplitude extractions across horizons. 7b-d & f-k represent exact amplitude extractions. ~ 120 ~

122 ~ 121 ~

123 The early Calabrian (Figure 4.8b) shows terrestrial channels extending much further north while the number of iceberg scours decline. During early Calabrian the downslope troughs continue to form and an intensification of the along-slope current features, both in terms of number of events and areal extent of these events is observed. Most of the along-slope contourites form on the slope of the southern clinoform set, however a minority form repeatedly along the northern clinoform set throughout the early Calabrian. This suggests that during the early Calabrian highstands the basin is more strongly dominated by marine currents. The early Calabrian also shows the first evidence of MSGLs, at around 1.72 Ma (MIS 60), and corresponding to a spike in the sedimentation rate (Figure 4.6). Three sets of MSGLs are observed, the first and most extensive lie on the northwestern slope of the basin, overprinting the along-slope contourites, where the MSGLs curve almost subparallel to the contours of the basin slope, but diverge slightly towards the south. The second and third sets are oriented ENE-WSW to the north and E-W towards the centre of the basin, and are found to be closer to 1.52 Ma (MIS 52) in age (see Figures 4.6 & 4.8). The late Calabrian (Figure 4.8c) sees the end of both the along-slope contourites and the downslope troughs in the centre of the basin and a further decrease in the number of preserved iceberg scours. However undeniable evidence for extensive and repeated grounded glaciations is observed in the form of numerous suites of MSGLs. All the MSGLs are curvilinear in nature and are found in the northern half of the study area orientated either north-west to south or north-east to south. The MSGLs are often associated with high amplitude deposits, which correlate with well logs (e.g. 21/20b-3; Figure 4.4g) to localised sand-dominated facies. The MSGLs of the late Calabrian can be split into two principal groups, one on each side of the basin, which are orientated from the northeast and north-north-west into the centre of the basin towards the south (Figure 4.8). At the very end of the late Calabrian period a large high-amplitude arc is seen orientated along the palaeo-slope and is interpreted to be a gas charged beach deposit. Few horizons were mapped in the post-jaramillo period (Figures 4.6 & 4.8d) as at this point the influence of velocity effects from the tunnel valleys above makes interpretation of individual ~ 122 ~

124 features extremely difficult. While iceberg scours can be seen, they are relatively uncommon and the extent of the scouring is often limited by the tunnel valleys. Terrestrial channels become increasingly difficult to differentiate from tunnel valleys. The main features observed are two large glaciotectonic complexes. The older and larger of the two complexes spans a total length of more than 280 km, reaching tens of kilometres in width. The complex can be directly correlated to the glaciotectonic thrusts identified and presented by Buckley (2012) and appear primarily to comprise of thrust faults (Figure 4.9; Buckley 2012). Deformation of the sediment is restricted to just above the base-jaramillo Event with significant deformation-free sediment between the thrusts and the Brunhes-Matuyama Event (Figure 4.9a) and thus is suggestive of MIS 22 glaciation (~0.87 Ma) for the age for their formation. The younger of the two glaciotectonic complexes is seen solely in the BroadSeis TM dataset and is significantly smaller in extent. The age of this complex is correlated to a probable MIS 16 date (~0.62 Ma) based on the position above the Brunhes-Matuyama reversal but with undeformed sediment between the thrusts and the tunnel valleys of the Ling-Bank unconformity (MIS 12; ~0.48 Ma) (Figure 4.9). 4.7 Palaeo-environmental reconstructions The palaeo-environment was reconstructed by combining basin reconstructions, facies interpretations and geomorphological mapping. The reconstructions of the basin through time are presented in figure 4.9 depicting the Gelasian, early Calabrian, late Calabrian and the post- Jaramillo palaeo-environments. Outside the study area ice extents and river locations are only estimated based on the assumption of increasing ice volume through the early Quaternary and the presence of the Baltic river systems from Overeem et al (2001) and Noorbergen et al (2015) Gelasian (2.58 to 1.78 Ma; Figure 9a) During the Gelasian the North Sea basin was an elongate basin of approximately 300m water depth (Figure 4.5) with a wide shallow shelf influenced significantly by a large number of iceberg scours as well as along-slope and downslope features (Figure 4.8a). There is a lack of evidence for iceberg scouring on the shallow sill between the North Sea and the North Atlantic (Figures 4.8a & 4.9) and ~ 123 ~

125 Figure 4.8 Maps indicating the distribution of mapped geomorphological features across the basin for set time periods a) Gelasian (2.58 Ma to 1.78 Ma) b) Early Calabrian (1.78 to 1.4 Ma) c) Late Calabrian (1.4 to 1.0 Ma) [next page] d) Post-Jaramillo (1.0 to 0.48 Ma) [next page] ~ 124 ~

126 ~ 125 ~

127 on the western margin of the basin. This suggests the basin shelf was probably flooded for the majority of the time and glaciation was restricted to the surrounding landmasses with no penetration into the basin. Currents dominated marine processes during the Gelasian. The configuration of the currents likely comprises strong northwards flowing downslope gravity currents and a comparatively weaker shallow along-slope flow (Chapter 2). The shallow flow is the process by which icebergs were distributed across the basin Early Calabrian (1.78 to 1.4 Ma; Figure 4.9b) The penetration of terrestrial channels northwards, an increase in the frequency and size of alongslope contourites at the same time as a gradual reduction in the number of downslope troughs and the first MSGLs are key features of the early Calabrian (Figure 4.8b). The terrestrial channels representing branches of the Rhine-Meuse and Baltic rivers suggest that more of the shelf was spending time subaerially exposed and can likely be linked directly to the gradual decrease in global sea level (Figure 4.6) and the shallowing of the basin as shown by the backstripping results (Figure 4.5). This is further supported by the presence of the MSGLs, which also identify an increase in the extent of ice sheets in line with the continued trend of global cooling throughout the Calabrian (Lisiecki & Raymo 2007). The decrease in the number of downslope troughs at the same time as an increase in the number of along-slope contourites (Figures 4.6 & 4.8b) suggests a gradual weakening of the downslope but a relative strengthening of the along-slope current system. This is also suggestive of a shallowing of the basin as the downslope processes are dependent on gradient of the clinoform slope which reduces as the basin shallows Late Calabrian (1.4 to 1.0 Ma; Figure 4.9c) In the late Calabrian the marine basin is close to being completely filled in (Figure 4.5) while both sets of current features, along-slope and downslope, disappear entirely (Figure 4.8c). Infrequent iceberg scouring can still be observed and a large coastline deposit is seen (Figures 4.7 & 4.8c). There is a significant increase in the number of MSGL events in the late Calabrian with individual MSGLs orientated towards north-west and north-east respectively, with sets on either side of the ~ 126 ~

128 basin and converging towards the south (Figures 4.7 & 4.8c). This suggests that where the marine basin remained conditions were too shallow to maintain any large scale current system and large portions of the North Sea were likely to be subaerially exposed. The coastline deposit is subparallel the final mappable position of the northern clinoform break point suggesting a gradual transition from clinoform deposits to the shallow shelf. The multiple sets of MSGLs identify repeated glaciation into the basin. Their orientation suggests either a southern source for the ice or two ice streams flowing from the north and converging. A southern source of ice would be unusual given that glaciations from the Elsterian to the Last Glacial Maximum identify the BIS and FIS as the principal sources of ice flow, both of which are to the north of the study area where the highland areas are located, rather than the lowland areas of southern Britain and northern Europe. Similarly models of confluence of the BIS and FIS have been inferred for the Elsterian, Saalian and Weichselian glaciations to explain the build-up of ice in the North Sea. Therefore it would appear that a confluence would more accurately fit the geometries of the MSGLs Post Jaramillo (1.0 to 0.48 Ma; Figure 4.9d) The character of seismic reflections and the backstripping results indicate that post the Jaramillo palaeo-magnetic reversal the North Sea consists of a wide flat plain similar to the modern day North Sea (Figures 4.2 & 4.5). The geomorphology preserved in the stratigraphy is limited primarily to the glaciotectonic complexes and isolated iceberg scours (Figure 4.6). The glaciotectonics identify two ice margins (Figure 4.8d) and the direction of the thrusting is from east to west in the older set and north-east to south-west in the younger. Well data and core descriptions from BGS core 77/02 (Stoker et al 1983; Cameron et al 1987; Sejrup et al 1994) identifies the sediment as primarily marine in origin. The glaciotectonic complexes indicate the advance and retreat of large terrestrial ice sheets while the core data suggest widespread flooding of the shelf. The contradiction likely arises from the change from low magnitude 41 kyr glacial cycles to the larger 100 kyr glacial cycles which begin to dominate during this period (Figure 4.6). Long periods of terrestrial exposure ~ 127 ~

129 occur during glacial events contrasting with flooding of the sea during interglacials. The direction of thrusting in the older glaciotectonic complex, suggests FIS source while the younger complex ~ 128 ~

130 Figure 4.9 Seismic cross sections and base map of glaciotectonic thrust complexes in the youngest part of the stratigraphy. A-A represents older glaciotectonic complex correlated to a probable MIS 22 date and has previously been observed by Buckley (2012). B-B represents younger glaciotectonic complex correlated to a probable MIS 16 date ~ 129 ~

131 could conceivably be due to the FIS or another confluence, however there is a lack of evidence to argue for one or the other. 4.8 Discussion The palaeo-environmental evolution of the North Sea basin was controlled by two major factors: firstly the infill of the marine basin through time, with local variations in sea level, namely an overall shallowing of the basin, overprinting the global signal; and secondly, the extent of ice sheets from the British and Scandinavian landmasses. These two controlling factors are interlinked the ice sheets cannot expand far into the deep marine basin of the Gelasian, but strongly influence the shallower basin of the late Calabrian. In the Gelasian water depths in the basin are relatively deep and although global cooling and thus the onset of Northern Hemisphere glaciation began from 2.7 Ma it is likely that ice sheets remain relatively small, reaching marine termination only in areas adjacent to the highland regions. However while the duration and magnitude of glacial events with respect to the δo 18 ratio and thus ice volume proxy prior to the MPT was low, it should also be remembered that the interglacial events were equally low in terms of magnitude. During the Gelasian the magnitude of the interglacial-glacial cycles remains relatively constant but the average global temperature cools (Lisiecki & Raymo 2005) and thus by the late Gelasian the interglacial climate maxima were notably cooler than those at the onset of the Quaternary or those that occur after the MPT (Figure 4.6; Lisiecki & Raymo 2005; 2007). This would lead to a reduction of ice melt during interglacial periods and a positive mass-balance allowing for a consistent increase in ice volume throughout the Gelasian into the early Calabrian. The geomorphological evidence in the early Calabrian of isolated MSGLs at 1.72 Ma (MIS 60) and 1.52 Ma (MIS 52) significantly pushes back the date of grounded, lowland glaciation in the CNS. As the ice volumes increased due to global climatic change the local sea level was fluctuating significantly during the same period. Infill of the southern half of the basin to < 100±50 m water ~ 130 ~

132 Figure 4.10 Palaeo-environmental reconstructions of North Sea a) Gelasian (2.58 to 1.78 Ma) b) Early Calabrian (1.78 to 1.4 Ma) c) Late Calabrian (1.4 to 1.0)[next page] d) Post-Jaramillo (1.0 to 0.48 Ma)[next page]. Based on geomorphic evidence and basin reconstructions from this study. Positions of Northern European Rivers estimated from Overeem et al 2001, Busschers et al 2007 and Noorbergen et al Locations of ice sheet extents confidently mapped only within study area and estimated otherwise on the assumption of increasing ice cover during the Quaternary. ~ 131 ~

133 ~ 132 ~

134 depth was complete while overall depth of the basin has reduced to a maximum of 200±50 m. Yet sea level may well have been shallower due to eustatic change which is estimated to be in the region of 60 m below present day sea level during MIS 60 and 70 m during MIS 52 (Figure 4.6; Miller et al 2011). The shallowing of the basin, coupled with an expanding ice volume, allowed for the ice sheets to expand much further than previously and begin forming ice streams. The reason this only occurred during two of the glacial stages during the early Calabrian however is not clear and repeated occurrences of MSGLs across several glacial-interglacial cycles did not occur until the late Calabrian. By the late Calabrian the depth of the basin was extremely shallow and it was during this period that clinoform progradation could no longer be supported. At the same time the magnitude of the glacial-interglacial cycles began to increase as the MPT approached (Figure 4.6; Lisiecki & Raymo 2005; Miller et al 2011). Six distinct episodes of MSGL formation are inferred for MIS stages 40 to 34 (Figure 4.6). The MSGLs indicate ice streaming behaviour at the confluence of the BIS and FIS in which the ice sheets meet and are each deflected southwards. The palaeo-ice streams were maintained through to the end of the Calabrian, with each successive glacial episode seeing the advancing ice sheet reoccupying the same pathway. However the ice streaming behaviour ends shortly after the Jaramillo palaeo-magnetic Event. Post the Jaramillo Event (1.08 to 1.0 Ma) there was a change in the nature of the subglacial processes and large scale deformation takes precedence building large scale glaciotectonic complexes. The restructuring of the subglacial palaeoenvironment may be related to the final transition from 41 kyr to 100 kyr cycles as glacial episodes become longer and colder although the mechanism driving the change from ice streaming to deformation is unclear. 4.9 Summary During the early Quaternary the central North Sea basin underwent significant change. The elongate basin that existed at the onset of the Gelasian was rapidly infilled, primarily from the ~ 133 ~

135 south through the Rhine-Meuse and Baltic river systems, leaving behind a vast flat shelf which was periodically flooded by the end of the Calabrian. The basin itself shortened and shallowed shifting through time from a NNW-SSE axis orientation to a NE-SW orientation before finally being infilled completely leaving a relatively flat shelf similar to today s North Sea. The stratigraphy of the early Quaternary of the central North Sea has previously been categorised into one single formation, the Aberdeen Ground Formation, but seismic investigation of the basin infill and the well data available suggests a new classification with at least four member units, named in this study as A to D, which have very different depositional settings and provenances. Aberdeen Ground A and B are progradational basin infill packages, with internal facies interpreted as deep marine to terrestrial. Member A was the more expansive of the two and sourced from northern Europe via two major river systems. Member B was sourced from mainland Britain, particularly Scotland, though the rivers delivering this sediment are not well preserved. Mapping and interpretation of geomorphological features allows for additional detail to be added to the palaeo-environmental setting. In the earliest part of the Quaternary, during the Gelasian, the North Sea was dominated by a large, elongate and deep marine basin rapidly infilling from the south and surrounded by a shallow shelf. Iceberg scouring was common, indicating marine terminating ice, while the basin itself was dominated by a strong current regime. As time progressed global temperatures and sea level dropped and the North Sea basin shallowed, with infill increasing the areas of the shelf that were sub-aerially exposed. The first evidence for grounded glaciation in the form of MSGLs is found at 1.72 Ma but does not become common until 1.3 Ma, at which point the basin was almost filled in and a terrestrial depositional environment dominated. By the Elsterian, grounded glaciation during the glacial cycles was well established in the central North Sea. ~ 134 ~

136 4.10 Acknowledgements The authors would like to acknowledge NERC and BUFI in conjunction with Forewind for supporting this work through RL s NERC CASE studentship at the University of Manchester and BGS (No. A87604X). They would also like to acknowledge the BGS and University of Edinburgh for funding part of this work as a part of the GlaciStore IODP bid. Thanks go to PGS for providing the CNS MegaSurvey, CGG for the BroadSeis TM dataset and TGS for the well data through their Facies Map Browser also Schlumberger for donation of the Petrel software, IHS for the Kingdom Suite and Eliis for providing Paleoscan. Final thanks go to Stefano Patruno for his advice on the back-stripping process. ~ 135 ~

137 Chapter 5: Tunnel valleys of the Dogger Bank analysed using ultra-high resolution 2D seismic data Rachel M. Lamb, Astrid Ruiter, Margaret Stewart, Mads Huuse, Simon. H. Brocklehurst Abstract: The Dogger Bank, a bathymetric high in the North Sea, lies along boundary of the southern extent of several advanced ice margins during the Middle to Late Pleistocene. New ultrahigh resolution 2D seismic data shot for United Kingdom windfarm development zone allows the tunnel valleys of Dogger Bank to be mapped in detail to further our understanding of this southern ice margin. 31 tunnel valleys are identified and mapped across the Dogger Bank high. Orientations of the Dogger Bank tunnel valleys are consistent with reconstructions of ice flow south to southeastwards from the British ice sheet. The infill of the Dogger Bank tunnel valleys is remarkably uniform and formed of a chaotic cut-and-fill package, a thin, high-amplitude package and a final low-amplitude package which, in three valleys, is observed to have clinoformal geometries that dip gently to the north. We observe a strong relationship between the tunnel valleys of the Dogger Bank and a series of glaciotectonic structures. A large thrust complex formed prior to the formation of the tunnel valleys is proposed to have influenced the distribution of the tunnel valleys preferentially towards the hanging wall of the thrusts. Shallow, small-scale deformation in posttunnel valley sediments is observed to change from chaotic to folded in the region above tunnel valleys. A model is proposed in which changes in meltwater drainage at the base of the ice sheet due to the differential infill of the tunnel valley caused this change in shallow deformation style. ~ 136 ~

138 5.1 Introduction The North Sea is a shallow epicontinental marine basin bordered on three sides by the Scandinavian, British and Northern European landmasses. Due to its position between Britain and Scandinavia the North Sea has been strongly influenced throughout the Pleistocene period by the growth and decay of both the British and Fennoscandian ice sheets (Larsen et al 2000; Bradwell et al 2008; Graham et al 2011). Water depths average 100 m across the North Sea, with the exception of the Norwegian Trench which runs along the Norwegian coast reaching water depths of up to 700 m (Figure 5.1). The Norwegian Trench was carved by ice and occupied repeatedly by an ice stream (Sejrup et al 1996, 1998, 2003; Larsen et al 2000). Geomorphic evidence for ice sheets during the Middle and Late Pleistocene is extensive in the North Sea both on the present day sea floor and in the shallow seismic section. Geomorphologic features have been mapped extensively across the North Sea and used to recreate past ice margins. This includes moraine deposits (Bradwell et al 2008) to mega-scale glacial lineations (Graham et al 2007), glaciotectonics on macro and microscales (Andersen et al 2005; Carr et al 2006; Buckley 2012) and subglacial features such as tunnel valleys (Praeg 1996, 2003; Stewart et al 2012, 2013). Tunnel valleys large, elongate erosional channels formed by the movement of meltwater under the ice are of particular use in identifying ice margins due to their tendency to form close to the ice front and parallel to ice flow direction (Stewart et al 2013). Hundreds of tunnel valleys have been mapped across the North Sea (Van der Vegt 2012 and references therein), and used to reconstruct past glacial margins, as well as investigate the dynamics of the British and Scandinavian ice sheets during the Middle to Late Pleistocene (Graham et al 2011; Stewart et al 2013). The Dogger Bank is a present day topographic high within the North Sea close to the southern extent of several ice sheet advances during the Late Pleistocene (Figure 5.1). Sitting 125 km off the coast of Yorkshire the Dogger Bank is outside of the central graben and principal hydrocarbon fields and thus outside most glacial geomorphological studies which use data from the hydrocarbon ~ 137 ~

139 industry (Harvey et al 2010; Van der Vegt 2012; UKCS 2015). Mapping of the Quaternary succession of the Dogger Bank area was carried out by the British Geological Survey (BGS) and completed by the 1990s with a number of buried and surface tunnel valleys identified, largely based on BGS 2D seismic data (BGS 1986; 1991; BGS & Rijks Geologische Dienst 1986; 1991). The licensing of the Dogger Bank as one of the UK round 3 windfarm development zones means that new data has been acquired which can allow for more detailed mapping of the Dogger Bank (Cotterill et al 2012), including the Middle to Late Pleistocene tunnel valleys. Figure 5.1 Regional map of the Dogger Bank Development Zone and the Tranche A dataset. Bathymetry data from Ryan et al (2009). ~ 138 ~

140 5.2 Regional Background The North Sea originated during rifting in the late Jurassic and early Cretaceous relating to the opening of the North Atlantic (Ziegler, 1992) and since then has been subject to multiple inversion events and near-continuous thermal subsidence throughout the Cenozoic (Ziegler 1992; White & Lovell 1997; Huuse 2002; Stoker et al 2005; Anell et al 2010; Goledowski et al 2012). The Dogger Bank comprises a bathymetric high with water depths between 18 and 63 m and is considered to have been formed during the last glaciation between 30,000 and 15,000 BP (Veenstra 1965; Fitch et al 2005; Cotterill et al 2012). Part of a larger landmass referred to as Doggerland the Dogger Bank was terrestrial during periods of low sea level and formed a land bridge between Britain and mainland Europe prior to its final marine inundation between 6-10,000 BP (Veenstra 1965; Fitch et al 2005; Cotterill et al 2012). The stratigraphy of the Middle to Late Pleistocene strata for both the North Sea in general and Dogger Bank in particular is made up of multiple packages of glacial and interglacial sediments which lie unconformably over the more homogenous Aberdeen Ground Formation of the Early Pleistocene (Figure 5.2) (Veenstra 1965; Cameron et al 1992; Gatliff et al 1994; Stoker et al 2011). The unconformity is regional in nature and generally referred to as either the Ling Bank, in the central North Sea, or the Swarte Bank, in the southern North Sea, unconformity (Sejrup et al 1991; Cameron et al 1992; Gatliff et al 1994; Stoker et al 2011). The unconformity is erosional in nature, and formed by a complex and widespread network of elongate, u shaped cross-cutting channels filled in by a strongly heterogeneous fill; features identified as subglacial tunnel valleys (Cameron et al 1987; Wingfield 1989; 1990; Gatliff et al, 1994; Stoker et al, 2011; Praeg 2003; Stewart et al 2013). These buried tunnel valleys which incise the Aberdeen Ground Formation have recently been extensively mapped across the central and southern North Sea, aided by the acquisition of regional 3D seismic datasets such as the PGS Central North Sea MegaSurvey used by Stewart et al (2013). Buried tunnel valleys have not so far been mapped on the Dogger Bank High for two reasons: Firstly the lack of high quality seismic data either in the form of tightly spaced 2D or full 3D (Harvey et al 2010) over the area, largely due to the lack of hydrocarbon prospectively in the ~ 139 ~

141 area; Secondly the location of the Dogger Bank on the boundary between the southern and central North Seas (Cotterill et al 2012). The position of the Dogger Bank has meant that previously mapping of the seismic-stratigraphy has taken place across two completely separate BGS mapping surveys and cross-correlation between the two was poor (Cotterill et al 2012). Figure 5.2 Summary of the Dogger Bank Stratigraphy correlated to the global sea level curve (Miller et al 2011), δ18o marine isotope stages (Lisiecki & Raymo) and confirmed Middle and Late Pleistocene glacial events (Graham et al 2011) Tunnel valleys, owing to their presumed model of formation, are considered to be one of the key features to understand and reconstruct past ice sheet extents and dynamics. Tunnel valleys are formed under subglacial conditions, close to the ice margin and parallel to ice flow direction however models for their formation are varied and remain highly debated (Ó Cofaigh 1996; Huuse & Lykke-Andersen 2000a; Kehew et al 2012; Van der Vegt et al 2012; Stewart et al 2013). Three principal models of steady-state, time-transgressive and catastrophic flow have been presented and debated intensely with arguably stronger evidence, in the North Sea at least, for a combination of slow steady-state flow in small channels underneath the ice sheet which build up into the larger complex and a more rapid time-transgressive model underneath a retreating ice sheet (Boulton & Hindmarsh 1987; Ó Cofaigh 1996; Van Dijke & Veldkamp 1996; Huuse & Lykke-Andersen 2000a; Kehew 2012; Van der Vegt 2012; Stewart et al 2013). Both models rely on the formation of channels in a situation where meltwater at the base of a warm-based ice sheets exceeds the capacity of the permeable substrate to remove water through groundwater flow (Kehew et al 2012 and references therein). In the steady-state model small channels are formed from meltwater which ~ 140 ~

142 steadily grow larger through lateral migration and channel formation is controlled by the pressure gradient within the ice sheet (Kehew et al 2012; Van der Vegt 2012). The time-transgressive model infers formation and fill of channels occurs solely during retreat in which excess meltwater ponds under the ice and is released in small channelized outbursts, as opposed to the large extreme flood events of the catastrophic model (Kehew et al 2012; Van der Vegt 2012). Traditionally most tunnel valleys in the North Sea are thought to date from one of three major glaciations, the Elsterian (MIS 12; ~0.48 Ma), the Saalian (MIS 10-6: ~0.38 Ma) and the Weichselian (MIS 4-2: ~0.12 Ma) of the Middle to Late Pleistocene (Figure 5.2) (Praeg 1996; Huuse & Lykke-Andersen 2000a; Graham et al 2011). This was based on a combination of palaeomagnetic data, namely the Brunhes-Matuyama magnetic reversal dating to 0.78 Ma (Stoker et al 1983) and correlation of the tunnel valley infill to dated interglacial facies (e.g. Scourse et al 1998) or onshore facies (e.g. Kluiving et al 2003). However accurate and direct dating is rarely possible due to the erosive nature of the tunnel valleys. Newer techniques based on mapping and separation of buried tunnel valleys through their cross-cutting geometries allows construction of a relative timeline through truncation of one tunnel valley by another (Kristensen et al 2007; Graham et al 2011; Stewart & Lonergan 2011; Stewart et al 2013). In this way between four and seven generations of buried tunnel valleys have been identified, appearing after the B-M reversal and prior to the MIS 4-2 glaciations (Kristensen et al 2007; Stewart & Lonergan 2011; Stewart et al 2013). In this way the three-stage model is considered to no longer fit the evidence presented and is thus outdated. 5.3 Data and Methods The Dogger Bank is the largest of the nine UK round 3 windfarm development zones covering an area of 8660 km 2 (Figure 5.1a). Starting in 2010 the area was surveyed with the goal of mapping four tranches with the potential to develop a large windfarm across the Dogger Bank High. The 2D seismic data for Tranche A which was used in this study was shot in 2011 with an extremely intense grid of 100 metre in-line spacing and 500 m crossline spacing across the entire 2000 km 2 ~ 141 ~

143 area (Figure 5.1b; 5.3). The dominant frequency of the seismic is 400 Hz and the sampling rate 0.5 ms resulting in a vertical resolution of close to one metre. Figure 5.3 Map of 31 Tranche A tunnel valleys classified by generation, determined by crosscutting relationships and geometry. Note orientation and spacing of 2D seismic lines, location of thrust complex, clinoformal fill dip directions and figure locations. Mapping of valleys was undertaken on a grid basis, manually interpreting the basal truncation of each valley first every five in-lines then every other cross-line to map the overall geometry of valleys. A finer grid of every in-line and cross-line was then used where necessary, particularly in areas of cross-cutting and over-printing of different tunnel valleys, and the major fill packages of the tunnel valleys were mapped. Surfaces were created from the mapped horizons using a minimum curvature calculation. Statistics for the valley geometry in the form of length, width, depth and orientation were calculated for each valley. Total length of each valley was determined by measuring along the valley thalweg. Widths were measured from valley shoulder to shoulder at intervals of 100 to 500 ~ 142 ~

144 m perpendicular to the thalweg orientation. Depth was measured first in ms TWT from valley shoulder to thalweg along selected valley transects to achieve a good coverage of valleys. Depths were then converted to metres by assuming a seismic velocity of 1800 ms -1 in agreement with other studies performed on North Sea tunnel valleys (Huuse & Lykke-Andersen 2000a; Praeg 2003; Lonergan et al 2006; Kristensen & Huuse 2012; Stewart et al 2012, 2013). Finally orientations of the valleys were measured following Stewart et al (2013) by splitting polygons produced for thalweg length into a series of individual straight line segments; each segment had its bidirectional orientation calculated and the distribution of orientations were collated firstly for each tunnel valley and then for each generation to create rose diagrams. 5.4 Results A total of 31 individual valleys were mapped across tranche A with a north to south naming convention (Figure 5.3). Valleys generally have a curvilinear planform with no indications of anastomosing planforms and only one case of branching; 11 valleys extend past the data coverage limits (Figure 5.3). Valleys range from 1 to 42 km in length, with a mean length of 7.6 km, and 300 m to 3.6 km in width, with a mean of 1.2 km (Table 5.1). Valleys stop and start abruptly, often shallowing from maximum depth and the point where they are absent on the seismic data within a few hundred metres. Valley depths vary between 140 and 320ms TWT (approximately m) with a mean of 218 ms (197 m). The orientation of the valleys shows two dominant directions of N-S and NW-SE (Table 5.1) Fill The fill of the valleys on Dogger Bank is relatively homogenous between different valleys with three clear fill seismic facies identified as discreet fill packages across the majority of valleys (Figure 5.4). Package I consists of medium to high amplitude facies usually chaotic in nature which generally forms in discreet v shaped units below the main u shaped erosional surface. Package I truncates ~ 143 ~

145 Table 5.1 Table of measured tunnel valley length, average width, maximum thalweg depth and orientation separated by generation. Where the length measurements have + the tunnel valleys extend past the data limits. All Gen 1 Gen 2 Gen 3 Gen 4 Gen 5 Gen 6 Gen 7 No of Valleys Orientation Length (km) Width (m) Depth (ms TWT/m) Max Min Mean Max Min Mean Max 320 (288) Min 140 (126) Mean 218 (197) ~ 144 ~

146 the stratigraphy below the valley and individual units within the package may truncate one another suggesting multiple phases of formation. Package I also appears inconsistently along the length of the valley, often appearing for only a few hundred metres before disappearing. Some valleys do not appear to contain package I at all. The majority of instances of package I are in the form of discreet units with average thicknesses of approximately 60 ms and maximum widths of 500 m (Figure 5.4a). However in some instances Package I is seen to form a single large unit spanning the width of the valley, up to 1 km wide with a maximum thickness of up to 0.1 ms (Figure 5.4b). Package II consists of a high amplitude package of variable thickness (although rarely more than 20 ms thick). Seismic reflections in this package are generally parallel to one another, but are curved in nature (Figure 5.4a), following the smooth u shape of the valley cross-section. The lower boundary of Package II is erosional, truncating the stratigraphy into which the tunnel valley is cut as well as package I (Figure 5.4b). Package II is present in all valleys and generally defines the sides of the valley through truncation of the background sedimentation and onlap of the infill strata. The high amplitude nature of Package II is used to allow mapping of the valleys. Package III consists of a low to medium amplitude seismic facies which vary from sub-horizontal, sub-parallel reflections (Figure 5.4b) to semi-chaotic in nature (Figure 5.4a). The lower boundary is erosional creating a smooth u shape between package I and package II (Figure 5.4). The top of Package III generally has a conformable relationship with the background sedimentation (Figure 5.4b) although erosional tops do occur infrequently (Figure 5.4a). Package III is present in every tunnel valley in some form and consists of up to 90% of the total valley infill. In three of the tunnel valleys (09, 10 and 19) the sub-parallel reflections of Package III are clinoformal in nature and dip gently towards the north along the long profile of the valley (Figures 5.3, 5.5). The dips of these clinoforms are measured to be between 2 and 5 (Figure 5.5). ~ 145 ~

147 Figure 5.4 Seismic cross-section and sketch interpretation (~2x V.E.) of a) V10 and b) V09 identifying valley structure and three mapped infill packages. ~ 146 ~

148 5.4.2 Generations In this study valleys were assigned to different generations in order to form a relative time-line of formation according to the generational framework outlined by Stewart et al (2013) in which relative age assignments were made according to cross-cutting relationships between valleys. Cross-cutting relationships between the valleys in the Dogger Bank Tranche A reveal a minimum of seven valley generations (Figure 5.3; 5.6). Geomorphological statistics for each generation is shown in Table 5.1. Generation 1 contains the most number of valleys however it is also the shortest in length excepting the single valley in generation 7, although every generation contains at least one valley which extends past the data limits and thus must be longer. Valleys in generation 1 appear to be more fragmentary in nature, regularly being subsumed beneath younger, larger valleys (e.g. valley 13, Figure 5.3) or terminating after only a short distance (e.g. valley 16; Figure 5.3). Generation 1 displays a N-S trend with a comparatively low range of orientations. Generations 2 and 3 both have average lengths in the region of 5 km and average widths of 780 and 850 m respectively; well below the overall average of 1.1 km. Generation 2 shows a narrower range of orientations, almost purely E-W, while generation 3 shows a much wider spread around a NW-SE trend. There are strong similarities in character between generations 2 and 3 but these are distinguished by clear cross-cutting relationships. Generation 4 consists of only two valleys (valley 08 and valley 12) within the Tranche A area, one of which (valley 12) is almost completely destroyed by a later phase of tunnel valley formation, generation 5, leaving only a fragment behind. Valley 08 on the other hand is extensive with a strong NNE-SSW orientation and shows the closest length, more than 12 km, and width, 2.35 km, to the overall average (Figure 5.3; Table 5.1). Generation 5 contains the singular branching valley, 10, as well as valley 19 which is the most strongly affected by large scale deformation in the east of Tranche A. Generations 5 and 6 contain ~ 147 ~

149 Figure 5.5 Seismic long profile section and sketch interpretation (~ 2x V.E.) of V19 showing northwest-dipping clinoforms of infill package I. ~ 148 ~

150 the largest tunnel valleys with average lengths of 13.6 km and 17.6 km and average widths over 1.5 km. Valley 10 in generation 5 and valley 09 in generation 6 are the only valleys of comparable size to others mapped in the North Sea as detailed in Van der Vegt s (2012) comprehensive review. Valleys 09 and 10 have the best resolved fills allowing for clinoforms to be observed in fill package III. Generation 5 shows a dual N-S and NW-SE distribution whereas generation 6 retains a WNW-ESE orientation. Generation 7 consists of a single valley (valley 21) 2.8 km in length which extends out of the study area (Figure 5.3) and thus measurements are considered to be minimum values for the valley. The identification of a minimum of seven generations of tunnel valleys on the Dogger Bank is in strong agreement with other studies which have identified the multi-generational nature of tunnel valley formation (Kristensen et al 2007; Stewart & Lonergan 2011; Stewart et al 2013) Deformation Seismic data from the eastern margin of Tranche A depicts a large thrust complex with a y shaped planform which has been interpreted as a glaciogenic thrust complex (Figure 5.3). The thrust complex interacts with several of the valleys in the Dogger Bank dataset and the combination of deformation and acoustic effects makes mapping difficult. Valley 19 is the most strongly affected both by the deformation and by a large acoustically blank zone which persists on the data on the hanging wall of the faults resulting in a section of the valley which is completely un-mappable (Figure 5.7). However valley 19 also shows the best insights into the relationship between the valleys and the glaciotectonic faulting. Mapping of the faults and the walls of valley 19 indicate that the faults initiate from a décollement surface in the sediments below the tunnel valley and that the faults themselves appear to be truncated closer to the surface by the valley wall (Figure 5.7). This provides strong evidence to suggest that the initial glaciotectonic event occurred prior to the formation of the tunnel valleys. ~ 149 ~

151 This is borne out by valley 20 which bisects the smaller of the two planform branches (Figure 5.3). However beyond valley 19 the thrust faults extend to the seabed deforming the shallower and younger sediments overlying the valley. This implies a second thrust deformation event or a reactivation of the older thrusts (Figure 5.7). The valleys associated with the thrust faulting show a curious orientation. Valleys 08, 19 and 20 are all found on the hanging wall side of the faults orientated approximately along the strike of the faults (Figure 5.3). However in contrast there are no valleys found on the footwall side of the complex. Valley 20 terminates abruptly after moving past the extent of the thrust complex (Figure 5.3). This is suggestive of a second order control on the location of valleys. Other valleys in close proximity to the faults, namely valleys 06, 21 and 22, (Figure 5.3) have an unclear relationship with the thrust faulting for differing reasons. Valleys 21 and 22, in generation 6 and 7 respectively, both appear to terminate in the acoustically blank zone (Figure 5.3) leaving the relationship unclear. However as the thrust complex appears to be older than valley 19, in generation 5, it is probable that they share a similar relationship to the thrusts as valley 19 does. Valley 06, in generation 3, is subsumed below the generation 4 valley 08, and so the relationship with the thrust faults is not clear. The more probable relationship is that the glaciotectonics also predate valley 06, although it cannot be ruled out that the thrust faulting occurred between generation 3 and 4. While the thrust complex in the eastern margin of Tranche A is the most spectacular it is not the only deformation in the Dogger Bank seismic stratigraphy. Shallow (to a maximum depth of 0.08 ms), small scale deformation can be seen in the package above the valleys across the entire data coverage area (Figures 5.4, 5.5, 5.6, 5.7 and 5.8). This small scale deformation is observed to change in nature in specific localities above valleys from a chaotic deformation domain to a more regular, folded domain (Figures 5.4a, 5.5, 5.6, and 5.8). Although other shallow deformation styles exist across the Dogger Bank (Figures 5.4b, 5.7; Ruiter & Dove 2014) the specific change from a chaotic deformation to regular folding is only observed in relation to the valleys (Figure 5.9). A ~ 150 ~

152 ~ 151 ~ Figure 5.6 Seismic cross-section and sketch interpretation (2x V.E.) of V07 and V05 showing the generational relationship between different tunnel valleys and the truncation of the infill packages of the older tunnel valley V07

153 notable thickness of up to ~ 20m of undeformed sediment is observed between the valleys and the shallow deformation (Figures 5.4a, 5.5, 5.6, and 5.8). 5.5 Discussion The seismic character, geometry and infill of the valleys strongly agree with a tunnel valley interpretation adding the valleys of the Dogger Bank to the extensive network of tunnel valleys across the North Sea. 31 tunnel valleys have thus been mapped across the Dogger Bank high Geometry The tunnel valleys of Dogger Bank are very similar to the rest of the North Sea tunnel valleys in terms of geometry and structure with a singular exception; they tend to be significantly shorter. The statistics calculated by Stewart et al (2013) for buried tunnel valleys across the central North Sea suggest average widths in the region of 700 m to 2.5 km which is comparable to the 800 m to 1500 km in this study. Average lengths in Stewart et al (2013) are all above 10 km, in comparison to the 4.5 to 17.5 km in this study. Some of the disparity can be considered to be due to the lack of areal coverage of the dataset, as a third of the Dogger Bank tunnel valleys extend past the limits of the dataset and thus may be significantly longer than measurements suggest. The tunnel valleys in Tranche A appear to be fragmentary compared to the tunnel valleys seen elsewhere in the North Sea. It is unclear if this is a real character or an artefact of the seismic data, as several of the smallest tunnel valleys appear to fade into the seismic with no clear termination (e.g. valleys 16-18; Figure 5.3) while others are clearly subsumed below older valleys (e.g. valley 13, Figure 5.3). There are two principal differences between the mapping on the Dogger Bank and that done in the MegaSurvey by Stewart et al (2013). Firstly the Dogger Bank survey provided a grid of 2D seismic data, and although the density of the 2D lines gives a pseudo-3d data coverage it does not allow for useful 3D mapping techniques such as time slicing which may aid in the interpretation and mapping of valleys. Early mapping of tunnel valleys on more widely spaced 2D data seemed to suggest branching of tunnel valleys was common (e.g. Wingfield 1989, 1990) and it was not until full 3D coverage was available that it became clear that valleys were more accurately described as ~ 152 ~

154 Figure 5.7 Seismic cross-section and sketch interpretation (~2x V.E.) of V19 and large thrust fault complex including acoustic blanking zone. Indications of truncation of the thrusts and the basal décollement by the tunnel valley suggests an older age of formation for the thrusting ~ 153 ~

155 singular features which cross-cut one another in a dense network (Stewart & Lonergan 2011; Stewart et al 2012, 2013). The second difference between the Dogger Bank survey and the MegaSurvey is the resolution. Although the MegaSurvey data is in full 3D the maximum vertical resolution is 8 m and while the horizontal bin spacing reaches a minimum of 12.5 m it is more commonly 50 m (Stewart 2008, 2013). It seems possible that, particularly in cases where fragmentary tunnel valleys are overprinted or crosscut by younger, more complete valleys, the MegaSurvey does not fully resolve the features where the ultra-high 1 m resolution of the Dogger Bank survey does. If this is the case then it can be assumed that a higher resolution 3D seismic dataset may uncover a great deal more information than has previously been observed for North Sea tunnel valleys. The overall orientations of the tunnel valleys can be considered to be consistent with tunnel valleys across the North Sea when one considers the structure of the ice sheet at the time and likely flow directions. Models of ice flow during the Weichselian, particularly Boulton & Hagdorn s (2006) model, identify that dominant ice flow directions on the Dogger Bank would have probably been south to south-eastwards which is consistent with the idea that tunnel valleys form during deglaciation parallel to flow direction where no other controls on formation exist. The tunnel valleys most likely were sourced under the British Ice Sheet or on the southern edge of the confluence between the British and Scandinavian ice sheets (Boulton & Hagdorn 2006; Stewart et al 2013) Fill Although the fill of tunnel valleys across the North Sea has often been described as heterogeneous and variable (Wingfield 1989; Ó Cofaigh 1996; Jorgensen & Sandersen 2006; Kristensen & Huuse 2012; Hepp et al 2012; Stewart et al 2012; Van der Vegt 2012) the overall structure of the infill has been observed to follow a two to four package structure (Wingfield 1990; Ehlers & Wingfield 1991; Huuse & Lykke-Andersen 2000a; Kluiving et al 2003; Praeg 2003; Lonergan et al 2006; Kristensen et al 2007; Kristensen & Huuse 2012; Stewart et al 2012). However, the relative ~ 154 ~

156 Figure 5. 8 Seismic cross section and sketch interpretation of V10 and relationship to the change in deformation style in the shallow section above the tunnel valley. ~ 155 ~

157 homogeneity of the fill of the Dogger Bank tunnel valleys is unusual. The fill of the Dogger Bank tunnel valleys appears to strongly resemble the infill structure observed in Hepp et al s (2012) high resolution study of a single tunnel valley in the southern North Sea. Particularly Hepp et al s (2012) site 5 can be overlain with the same three fill packages observed in the Dogger Bank (Figure 5.10). There are also similarities between the fill of the Dogger Bank tunnel valleys and the fill architecture presented by Huuse & Lykke-Andersen (2000a), Praeg (2003), Kluiving et al (2003), Lonergan et al (2006), Kristensen et al (2007), Moreau & Huuse (2012) and Stewart et al (2012). Package I does not appear to be observed on studies of lower resolution seismic data (e.g. Lonergan et al 2006; Kristensen et al 2007 and Stewart et al 2012) however it correlates to Unit VI in Hepp et al s (2012) study. Hepp et al s (2012) Unit VI is only cored on the flanks of the tunnel valley but appears to consist of well sorted sands with interbedded medium to coarse gravel layers and reworked till remnants. The majority of package I on the Dogger Bank is much smaller than Unit VI in Hepp et al (2012), although some valleys do show a thick sequence. The discontinuous nature of package I as well as the v shape appears suggestive of an initial cut and fill facies, similar to those observed by Jorgensen & Sandersen (2006), which was later strongly altered either by direct ice erosion or by reoccupation of the valley. Initial erosion by meltwater formed the elongate v shaped channel which partially infills with more permeable sands creating a preferential route which later meltwater events follow. Given sufficient flow it is possible for part or even the entire original channel to be eroded. The seismic facies of Package II most closely correlates to Hepp et al s (2012) Unit Vb and arguably the chaotic basal package of Stewart et al (2012). Although the chaotic basal package, which has also been observed by Huuse & Lykke-Andersen (2000a), Kluiving et al (2003) and Lonergan et al (2006), is similar in seismic character to package II it is generally thicker and tends to be discontinuous along the thalweg unlike package II. Hepp et al (2012) identified this package as strongly laminated stiff clay with periodic centimetre scale sand layers, probably lacustrine in origin. ~ 156 ~

158 Figure 5.9 Map showing the areas where the style of deformation in the shallow section changes in relation to the tunnel valleys. Note the majority of the changes are located on the downstream sides of the tunnel valleys particularly where the tunnel valley, or parts of the tunnel valley, is observed to be oblique to the inferred direction of ice flow. Ice flow direction inferred from the orientation of deformation structures. Package III is consistently similar with low amplitude, sub-horizontal reflectors which, in at least three valleys, dip gently north to north-westwards. This agrees with Hepp et al s (2012) unit III of rhythmically laminated silty-clay with occasional sands of probably lacustrine to marine origin and also correlates well with Stewart et al s (2012) middle/dipping unit of sub-horizontal, gentlydipping reflections. The presence of northwards dipping clinoforms within the fill package is consistent with observations by Praeg (2003), Lonergan et al (2006), Kristensen et al (2007) and Moreau & Huuse (2012) which identified the clinoformal fill of many tunnel valleys across the North Sea. This is in contradiction with Stewart et al (2012) who did not observe clinoformal packages. Several different models have been inferred for the source of the clinoforms including a backfilling hypothesis in which clinoforms represent the progressive infill of tunnel valleys while tunnel valleys were being formed (Praeg 1996). Moreau & Huuse (2012) suggest that the clinoforms represent deltaic interglacial stratigraphy from the Northern European river systems ~ 157 ~

159 with rivers preferentially infilling partially filled tunnel valleys. There is not enough evidence available from the Dogger Bank tunnel valleys to conclusively say which model fits best Deformation Although studies exist which show both tunnel valleys and glaciotectonics in close association (e.g. Huuse & Lykke-Andersen 2000a, b; Høyer et al 2013) very few have studied the relationship between them further than the relative age. The curious distribution of tunnel valleys only on the hanging wall of the glaciotectonic thrusts of the Dogger Bank is a novel observation. Other studies have attempted to observe a pattern for distribution of tunnel valleys in regards to controls other than ice sheet flow direction varying success. For example Praeg (1996) and Huuse & Lykke- Andersen (2000a) both observed a localised relationship with older structural lineaments and tectonically displaced salt in the southern North Sea. Stewart (2008) in comparison could find no similar relationship further north in the central North Sea. Figure 5.10 a) Seismic cross-section from Hepp et al (2012) and b) sketch interpretation identifying the same infill packages observed in Dogger Bank tunnel valleys within the southern North Sea tunnel valley. Used to assess possible and likely facies characteristics of the Dogger Bank infill packages. ~ 158 ~

160 Here we suggest a model which proposes excavation on the hanging wall side of the thrust faults creating palaeo-topography which created a minor diversion in ice flow and thus meltwater flow. This would have the same effect as the cut and fill package I in causing focusing of meltwater flow along the hanging wall which would be topographically lower; eventually leading to the formation of tunnel valleys. The thrust faults, and thus the topographic expression, lie north-south, which agrees with the orientation of the valleys on the handing wall. Given the bisection of the more NE-SW orientated secondary faults by valley 20 it suggests that if topographic control is influencing the distribution of tunnel valleys it is not a strong control and easily overridden by the dominant ice flow direction. Another possible explanation suggests lower sediment strength in the deformed sediments of the thrusts and hanging wall compared to the less deformed footwall making the latter easier to erode. This could explain the abrupt termination of valley 20 once beyond the extent of the thrust faults (Figure 5.3). Most likely the distribution of tunnel valleys in relation to the glaciotectonic thrust features is a combination of several minor controls on top of the primary ice flow direction. Although the smaller, younger glacial deformation observed cannot have been formed at the same time as the tunnel valleys (due to the presence of significant undeformed sediment between the two packages; Figures 5.4, 5.5, 5.6 & 5.8) it would seem that the close relationship between the two is not a coincidence. A model is proposed here which links subglacial drainage to basal coupling of the ice sheet, as summarised in Figure Initially, a tunnel valley was formed parallel to ice flow (Figure 5.11a). After ice sheet retreat, the tunnel valley in infilled and then overlain by a significant thickness of sediment during the subsequent interglacial period and the onset of the next major ice advance (Figure 5.11b). As the ice advances again from a different direction it advances on a sliding bed lubricated by significant subglacial meltwater. As the ice reaches the position of the tunnel valley the differential infill between the tunnel valley and the background sedimentation creates a path for drainage of the subglacial meltwater away from the base of the ice (Figure 5.11c). ~ 159 ~

161 As water drains away the ice sheet is locally coupled to its bed, increasing the strain on the sediments and causing them to begin to buckle and fold (Figure 5.11c). This explains the strong distribution trend in which the change in deformation style is observed principally to the south and west of the tunnel valleys (figures 5.4a, 5.5, 5.6, 5.8 and 5.9) mainly where tunnel valleys are oriented obliquely to models of ice flow direction, based on evidence from the deformation style, and on the downstream side of the valleys (Figure 5.9). Figure 5.11 Conceptual model of the change in shallow sediment deformation in relation to tunnel valleys in the subsurface. A Ice advance 1 causing the formation of tunnel valleys (ice flow parallel with tunnel valleys, outwards in this figure) with fill during and after glaciation. B Deposition of nonglacial and distal-glacial sediments during retreat and the onset of the second advance. C Ice advance 2 in which the re-advance of the ice causes deformation of overlying sediments. After Høyer et al (2013). ~ 160 ~

162 The second episode of thrusting, or the reactivation of thrusting, in the western margin of Tranche A to allow thrusts to reach the seabed (Figure 5.7) has no clear model of formation or timing. Thrusts of this size are observed to form ahead of the ice sheet due to gravity spreading (Huuse & Lykke-Andersen 2000b; Andersen et al 2005; Phillips et al 2008), requiring excavation of material and the formation of a depression closer to the ice margin. Due to the orientation of the thrusts this would be west of Tranche A and outside of the data area. The thrusts may be related to the same glacial advance that formed the shallow deformation or may be a separate event. 5.6 Summary 31 previously unknown buried tunnel valleys have been mapped across the Dogger Bank. These tunnel valleys are very similar in form to tunnel valleys described within the surrounding North Sea area, albeit notably shorter. The Dogger Bank tunnel valleys display a minimum of seven crosscutting generations, in agreement with other studies of North Sea tunnel valleys, and a remarkably uniform three-tiered fill. The infill packages observed are interpreted to represent initial cut and fill facies, followed by stiff clays of probable lacustrine origin and then by silty-clays of lacustrine to marine origin. The tunnel valleys of Dogger Bank show a relationship to glaciotectonised sediments in the Dogger Bank; with large thrust faulting formed prior to the tunnel valleys and with minor deformation subsequent to erosion and infill of the valleys. The former is observed to have a weak influence on the deposition of the tunnel valleys, though not strong enough to override the principal pressure control, the cause of which is not completely clear. The later minor deformation is suggested to be a direct result of changes in subglacial meltwater flow due to the changes in sediment properties between the tunnel valley fill and the background sedimentation. The Dogger Bank tunnel valleys are observed to have an overall bi-modal orientation of N-S and NE-SW which fit with reconstructions of British Ice Sheet retreat across the North Sea. ~ 161 ~

163 ~ 162 ~

164 Chapter 6: Summary ~ 163 ~

165 The principal aim of the project was to investigate the influence of the changing climate of the Quaternary on the North Sea. Over the course of the early Quaternary the global climate cooled dramatically and average global sea level lowered by more than 50 m (Miller et al 2011), and both trends can be observed in the seismic stratigraphy of the central North Sea. However, of far greater influence were the increasing magnitude and length of the glacial-interglacial cycles during the Mid-Pleistocene transition as well as the infill of the marine basin completely changing the structure of the North Sea between the earliest Quaternary and the middle Pleistocene. What follows is a summary of the evolving story of the North Sea Quaternary, the key insights learned from the project and the limitations of it. It also addresses collaborations undertaken using data and interpretations from the project and future work which may be undertaken to address the remaining questions. 6.1 Summary of Previous Understanding Prior to this project the central North Sea basin during the early Quaternary was identified as principally shallow marine, with facies ranging from fine muds to sands (Stoker et al 1985; Cameron et al 1987; Sejrup et al 1991; Gatliff et al 1994; Stoker et al 2011). The shallowest section was noted to be glaciomarine in nature from core samples (Stoker & Bent 1985; Cameron et al 1987; Sejrup et al 1987, 1991; Knudsen & Sejrup 1993) with isolated evidence for subglacial conditions (Stoker & Bent 1985; Carr et al 2006). Infill of the basin principally from the south via the large clinoforms of the Cenozoic was recognised (Holmes 1977; Cameron et al 1987; Rasmussen et al 2005; Nielsen et al 2008; Fyfe et al 2003; Knox et al 2010; Anell et al 2012; Goledowksi et al 2012; Ottesen et al 2014). The glacial influence on the North Sea was limited to limited iceberg scouring from the onset of the Quaternary (Kuhlmann & Wong 2008; Dowdeswell & Ottesen 2013; Ottesen et al 2014;) with the earliest date for grounded glaciation at 1.1 Ma (MIS 30) (Sejrup et al 2000; Graham et al 2011), but not becoming extensive until 0.48 Ma (MIS 12) (Cameron et al 1987; Graham et al 2011). The official stratigraphy for the central North Sea listed the entire stratigraphic section from the Pliocene to the Ling Bank Unconformity at 0.48 Ma (MIS ~ 164 ~

166 12) as Aberdeen Ground Formation, a pro-deltaic to marine mud facies (Gatliff et al 1994; Stoker et al 2011). 6.2 The Evolution of the Quaternary North Sea Basin At the onset of the Quaternary, the North Sea consisted of an elongate basin centred over the Central Graben with water depths exceeding 300 m, which had little resemblance to the shallow, flat shelf that exists in the present day (Chapter 2). There was a strong similarity between the depositional setting of the Pliocene North Sea (as described by Harding 2015 and others) and the depositional setting in the earliest Quaternary (MIS 104 to 98). Infill of the basin was primarily from the south with large, shelf-scale clinoforms prograding first westwards and then switching to a north-westwards progradation (Chapter 2). The clinoforms show evidence for feeder systems in the form of channels and slope fans in a very similar manner to those observed in the Pliocene. However there was a rapid change around MIS 96 as channels and slope fans rapidly disappeared (Chapter 4) and were instead replaced by powerful downslope currents. These currents formed a series of elongate troughs with cut-and-fill architectures that dominated the basin and clinoforms of the North Sea during most of the Early Pleistocene (MIS 96-34) (Chapters 3, 4). They were modified by a weaker, shallow system of anti-clockwise flow of water into and out of the North Sea basin which also formed along-slope orientated mounded deposits (Chapters 3, 4). During the period when current systems were dominant (MIS 96-34) the basin shrank by several hundred kilometres and shallowed dramatically due to the progradation of clinoforms into the basin, infilling it and building up a significant stratigraphy. Two principal clinoform packages in the Early Pleistocene prograded from the south and north respectively (Chapters 2, 4). The sediment that built the southern clinoform set was sourced from the Northern European river systems and this package has been classified as Member A of the Aberdeen Ground Formation (Chapter 4; Figure 6.1). In a similar way the northern clinoform set, sourced most likely from Scotland, is classified as Member B (Chapter 4; Figure 6.1). Members A & B are both ~ 165 ~

167 identified as pro-deltaic to marine muds and silts and deposition of the two members persisted from 2.58 Ma (MIS 104) to 1.2 Ma (MIS 36) (Chapter 4; Figure 6.1). The infill and shallowing of the basin is considered to be the primary cause for the cessation of large scale current systems in the North Sea, leaving the basin unable to sustain strong downslope currents and ending the formation of the large downslope troughs (Chapters 3, 4). The period between 1.2 and 1.0 Ma (MIS 36-26) was transitionary as progradation into the basin came to a halt despite a rise in average sedimentation rates (Chapter 4; Figure 6.1). The transitionary unit is identified as Member C of the Aberdeen Ground Formation, marked by primarily a muddy facies with intermittent sands of a shallow marine basin (Chapter 4; Figure 6.1). Member C transitions then into Member D, muddy-silts and sands, which existed from 1.0 to 0.48 Ma (MIS 26-12) and is equivalent to the shallow glacio-marine and intermittently subglacial facies observed in shallow core (Chapter 4; Figure 6.1). The onset of Member D coincided with a change to near-horizontal, parallel reflectors and represents the final infill of the marine basin. From 1.0 Ma onwards the North Sea had the same shelf-like configuration of the present day sea (Chapter 4). Evidence for glacial activity is limited to iceberg scours during the early part of the Quaternary, from 2.58 Ma to 1.7 Ma (MIS ) (Chapter 4, Appendix 2: Newton et al). During this period ice sheets were likely to be relatively small and marine-terminating, due to the shorter, smaller magnitude glacial-interglacial cycles and the warmer average global temperature (Chapter 4). As global climate cooled the interglacial climatic maxima also cooled restricting the degree of ice melt between glacial intervals and allowing ice sheets to begin building significant volume. The first evidence for grounded glaciation comes in the form of mega-scale glacial lineations combined with a peak in average sedimentation rates within the basin at approximately 1.72 Ma (MIS 60), but the incursion was small and limited (Chapter 4). The second grounded glaciation into the basin occurred at 1.56 Ma (MIS 52) accompanied by a more modest increase in sedimentation rate (Chapter 4). MSGLs and thus grounded glaciation did not become common until 1.3 Ma (MIS 40). ~ 166 ~

168 It is considered that there were two controls on the increase in grounded glaciation at 1.3 Ma, firstly the ~ 167 ~

169 ~ 168 ~ Figure 6.1 Summary of North Sea Quaternary stratigraphy using the conclusions of this project. Identifying four Aberdeen Ground Formation members (A to D) as well as a minimum of seven additional grounded glacial events on top of those proposed by Graham et al (2011).

170 infill of the basin causing large-scale shallowing and increased potential for sub-aerial exposure during glacial lowstands; and secondly the magnitude of glacial cycles increasing as the Mid Pleistocene transition approached (Chapter 4). After the Jaramillo palaeo-magnetic Event the MPT began in earnest, causing glacial events to last longer and reach much cooler temperatures, and the subglacial processes of the ice sheets to change. The dominance of MSGLs switched to large scale glaciotectonic complexes (Chapter 4) before changing once again and culminating in the widespread glacial unconformity at the base of the North Sea tunnel valleys (Chapter 4, Chapter 5). The regional glacial unconformity at the base of the tunnel valleys is most usually dated to 0.48 Ma (MIS 12) and covered the entire North Sea basin. Multiple generations of tunnel valleys exist, representing advance and retreat of the British and Scandinavian ice sheets after the Mid- Pleistocene transition, adding a great deal of uncertainty to this MIS 12 date. The Dogger Bank tunnel valleys are typical of North Sea tunnel valleys (Chapter 5). The infill of the Dogger Bank tunnel valleys follows the same broad structure of tunnel valleys across the North Sea but is remarkably homogenous (Chapter 5). The Dogger Bank tunnel valleys are seen to have a strong relationship with glaciotectonic features; truncating a significant glaciotectonic thrust complex and influencing the style of deformation in young, shallow scale deformation complexes by acting as an efficient drainage system for the subglacial meltwater system. The older thrust complex may be of the same age as the MIS 16 glaciotectonic thrusts observed in the North Sea (Chapter 4), however with no dating or correlation available this can only be supposition. 6.3 Key Insights and Conclusions The basal Quaternary surface for the central and southern North Sea has been mapped in full for the first time and can be used in any future work on the stratigraphy of the North Sea either for industry or academic interest. The early Quaternary was dominated by strong downslope currents which carved the basin and slope of the central North Sea. There is evidence however that the strong currents were ~ 169 ~

171 modified by a shallow circulation system driving significant change in the downslope troughs. The seismic stratigraphy of the early Quaternary has been revised and updated in line with the new data available. The Aberdeen Ground Formation has been split into four new members to aid in the classification and understanding of the palaeo-geography. Marine-terminating ice was present from the onset of the Quaternary in the North Sea, in line with the onset of Northern Hemisphere Glaciation. The first evidence for grounded ice sheets in the North Sea basin is identified at 1.72 Ma. Grounded glaciation becomes common in the North Sea from 1.3 Ma and the onset of the mid-pleistocene Transition. A minimum of seven additional grounded glacial events are identified in the central North Sea between 2.58 Ma (MIS 104) and 0.48 Ma (MIS 12) on top of those previously identified by Graham et al (2011) (Figure 6.1). Two new methodologies have been developed; one for the building of a detailed chronostratigraphy in the absence of reliable chronological ties, and one for the interpretation of palaeo-environments at a basin scale using 3D seismic data. The methodology for chronostratigraphic calibration requires further testing for robustness however the workflow for palaeo-environmental reconstruction can be applied to many other marine basins. The current maps of tunnel valleys across the North Sea are not yet complete, as evidenced by the new mapping of the Dogger Bank tunnel valleys, and other examples of tunnel valleys outside of traditionally data heavy areas may yet be discovered. A relationship between tunnel valleys, subglacial drainage and subglacial deformation has been observed in the Dogger Bank region of the North Sea. Tunnel valleys are observed to both be influenced by large scale thrust sheets and in turn influence shallow small scale deformation, which has not been noted previously. ~ 170 ~

172 6.4 Limitations of the Work Limitations of specific methods and results have been discussed in each chapter of the project however there are several limitations which are applicable to the project as a whole and are discussed here. Most limits of the project are concerned with the data available. Seismic and well data are established as tools for imaging subsurface however it is good practise to keep the limitations of the data in mind when drawing conclusions. Seismic resolution. The seismic resolution of any dataset determines the level of detail that can be extracted about the subsurface. The seismic datasets in this project have several degrees of resolution from the low-resolution PGS MegaSurvey which resolves vertically to between 8 and 16 m and horizontally to 50 m to the ultra-high resolution Dogger Bank survey with 1 m vertical resolution and a 100 m line spacing. The low resolution MegaSurvey can be particularly noted to have issues with seismic resolution many of the geomorphic features, such as channels and iceberg scours mapped in Chapter 4, have no vertical expression on seismic data due to them falling below seismic resolution. Although the 3D coverage does aid with this, it is likely that many features preserved in the subsurface are not imaged at all by the seismic data. Even the 1 m-scale Dogger Bank survey may miss a great deal of heterogeneity within the subsurface geology. In the case of the tunnel valleys mapped on the Dogger Bank survey the infill of some of the valleys is observed to contain clinoforms however any facies changes that could be expected from clinoformal geometries are not resolved. Well log data. Well logs provide a wealth of data for interpreting the characteristics of the subsurface sediments, however it must be remembered that well log data are geared towards the classification of reservoirs and all tools are designed to measure properties of the subsurface which are used as proxies for sediment composition, texture and structure. Thus trends in well log data may have multiple interpretations. Most well logs used in this study were drilled specifically for the hydrocarbon industry and so in the shallow section it is rare to have more data than gamma-ray logs and basic lithological descriptions. The data provided by TGS do not specify whether the lithological descriptions are from well ~ 171 ~

173 cuttings, and thus actual samples of the sediment, or from interpretation of the logs. As such well data has been used with caution only to supplement evidence from the seismic data and lithological descriptions from shallow core. Scale. In the particular case of the MegaSurvey, an area of over 80,000 km 2 is covered by seismic data and when combined with all other sources of data an area well over 100,000 km 2 is covered down to a depth of ~1.5 km. It would be impossible, even over the course of a 3 year project, to completely map the seismic stratigraphy as well as map every single feature preserved in the subsurface. Much of the issue of scale is aided by the semiautomated mapping of the reflections by the Paleoscan software which allows for much more complete coverage of the dataset. However even with this the level of detail achievable must be considered on a basin scale and as such much of the fine detail is, by necessity, lost. The conclusions on the evolving palaeo-environment are a framework only and may well be revised by future work. Semi-automated mapping. The process of semi-automated mapping of the seismic stratigraphy is a tool which has the potential to permanently change the workflow of basin analysis. However the model from which the horizon stack is produced must be considered carefully as there are limitations to the ability of the model to adapt to the seismic data. The presence of strongly dipping horizons, such as faults or the troughs detailed in Chapter 3 can disrupt the model. Similarly while control horizons can and should be used to refine the model any mistakes in the control horizons are propagated through the model and into the produced horizon stack. Careful checking and quality control must be done to ensure the best model possible is built such that the horizon stack produced accurately depicts the subsurface stratigraphy. 6.5 Collaborations During the course of this PhD several related collaborative projects have arisen in which work from this project was used to underpin the work of other researchers and these studies are summarised ~ 172 ~

174 below along with contribution by the author. Where available copies of the manuscripts for these collaborations have been included in Appendix IODP proposal #852-CPP GlaciStore proponents The GlaciStore proposal aims to drill the central North Sea in order to further the understanding of glaciation and basin processes in the Quaternary in order to inform the development of secure large scale offshore CO 2 storage. The establishment of the depositional and chronological framework of the Quaternary glaciations is one of the key scientific objectives of the project. A series of 13 sites have been identified, 4 primary and 9 secondary, across the northern portion of the study area for scientific drilling, with several aiming to penetrate to the basal Quaternary surface as presented in Chapter 2 and most aiming to quantify and investigate the depositional setting of the sediment. The stratigraphic analysis presented in Chapter 4 was undertaken in conjunction with the GlaciStore proposal during a two month placement leading to the identification of a proposed drill site to better capture the early Quaternary stratigraphy Evidence for repeated mid-latitude marine-terminating ice sheets in a 41 kyr early Quaternary world A. Newton & R. Harding. Iceberg scours have been noted in the North Sea and surrounding areas from 2.7 Ma. The number and distribution of iceberg scouring events have not previously been investigated in any detail nor have the implications of such events been discussed. In this study iceberg scours from the basal Quaternary surface (presented in Chapter 2) to 1.8 Ma have been mapped across the southern and central North Sea, and the occurrence of events has been compared to the δ 18 O curve and records of ice rafted debris from the Norwegian margin to identify 22 iceberg calving events between 2.58 and 1.8 Ma in the North Sea. The study uses the basal Quaternary from Chapter 2 as well as major and intermediate horizons from Chapter 4 in order to map the individual iceberg scouring events across the North Sea. The ~ 173 ~

175 chronology established in Chapter 4 was used to refine the dates of the iceberg scouring events in conjunction with dating from the southern North Sea MIS 16 glaciotectonic complex in the BroadSeis TM dataset C. Bendixen A glaciotectonic complex in the central North Sea prior to the tunnel valley unconformity has been observed in the BroadSeis TM dataset (Chapter 4). This study is a detailed case study which maps the complex, examining the patterns of thrust sheets and deformation and attempts to reconstruct the ice sheet dynamics involved during the formation of the complex. The study details the chronostratigraphic calibration which suggests a likely MIS 16 date for the thrusting. The study uses the chronology established in Chapters 2 and 4 in order to date the glaciotectonic complex. The case study by Bendixen was completed under the supervision of Lamb and Huuse during Bendixen s research visit in spring 2015, using this project s datasets. 6.6 Further Work This project sees a significant step forward in our understanding of the early Quaternary of the North Sea and the influence that the changing climate of the Quaternary had on the evolution of the basin. However there is still scope for further work to refine our understanding and to build accurate models of the evolving palaeo-climate. Chronological control. The methodology proposed in chapter 4 in order to date sediments by their relationship to the global sea level curve is a novel approach. Further work is necessary in order to fully test this methodology and thus refine our chronostratigraphy. Should the methodology prove to be robust it means that the chronostratigraphy of the North Sea could become one of the strongest globally for the Quaternary allowing models to reach a temporal resolution of tens of thousands of years. If there are errors in the methodology it will call for a much more detailed study of the cyclicity observed in the gamma-ray logs in an attempt to understand the cause and controls on these cycles and thus improve our understanding of depositional history. ~ 174 ~

176 Palaeo-climate archive. It is apparent from this project that the North Sea can be considered as a significant palaeo-climate archive for the early Quaternary. Scientific Figure 6.2 Summary of seismic horizons and chronostratigraphic correlation used in this project including both major and intermediate seismic horizons used for regional assessment of basin evolution and geomorphological mapping (chapter 4) and furrowed seismic horizons used to map, date and interpret downslope trough features (chapter 3). Also including correlation to the Dogger Bank tunnel valley generations (chapter 5) and Kuhlmann & Wong s ~ 175 (2008) ~ SNS seismic units.

177 drilling, such as the GlaciStore proposal, to target the expanded early Quaternary sediments could be combined with the framework presented in this project in order to create a detailed and high resolution record of mid-latitude climate change and environmental evolution. Basin evolution. The workflow used in this project to reconstruct the environmental evolution of the basin could easily be applied to other basins globally. This workflow would be of particular use in other areas with large 3D seismic datasets, to make mapping more efficient and to give more time for accurate and detailed interpretations. Early Quaternary of the North Sea. This project marks a huge step forward in our understanding of the early Quaternary in the North Sea, however it is likely that there is much more still to be discovered. One example of a focused study would be of the seismic sequence stratigraphy to further refine chronological control as well as note changes in the preserved geomorphology between highstands and lowstands, adding more detail to the models of depositional systems during the early Quaternary. Similarly a more detailed study on the period between the Jaramillo palaeo-magnetic Event at 1.1 Ma and the Ling Bank Unconformity at 0.48 Ma could provide evidence for the changing subglacial dynamics of the ice sheets which led to a decline in the number and size of MSGLs and the onset of glaciotectonic deformation. It is probable that this project can and will be used as a starting point for more detailed case studies on the early Quaternary, both with the extensive MegaSurvey and with more detailed, high resolution surveys such as the BroadSeis TM survey and other broadband (high resolution) 3D seismic datasets. Figure 6.2 summarises the horizons used in this project and identifies the dates found from the chronostratigraphic calibration. The horizons are also correlated to the Dogger Bank tunnel valleys for a more complete picture of the Quaternary stratigraphy and the SNS seismic units described by Kuhlmann & Wong (2008) so that the horizons can be compared to other studies on the stratigraphy of the central and southern North Sea. This will allow the horizons to be used in further studies on the early Quaternary of the North Sea. ~ 176 ~

178 6.7 Final Conclusions The evolution of the North Sea during the Early Pleistocene is determined by the changing global climate and the infill of the basin. A global cooling trend overlain on glacial-interglacial cycles allowed ice sheets to expand into the basin by 1.7 Ma and for confluence of the British Ice sheet and Fennoscandian Ice Sheet from 1.3 Ma. The infill of the basin by significant sediment input changed the structure of the North Sea significantly from a narrow, elongate basin over 300 m deep at 2.58 Ma to the shallow, near-flat shelf that forms the present day North Sea by 1.0 Ma. This project has significantly revised the current understanding of the early Quaternary in the North Sea. The project has implications for our understanding of basin and palaeo-environmental evolution during periods of dramatic climate change for the North Sea and surrounding landmasses. It also has scope to extend that understanding to a global scale through both new models of palaeoenvironmental evolution and new methodologies that can be applied to other basins. ~ 177 ~

179 References ~ 178 ~

180 Allen, P., Allen, J., 1990 Basin Analysis: Principles and Applications Blackwell Science, pp. 451 Alley, R.B., 2000, Two-Mile Time Machine: Ice Cores, Abrupt Climate Change and Our Future Princeton University Press, Princeton, New Jersey, pp. 229 Andersen, L.T., Hansen, D.L., Huuse, M., 2005 Numerical modelling of thrust structures in unconsolidated sediments: implications for glaciotectonic deformation Journal of Structural Geology 27, Andresen KJ, Huuse, M, Clausen, OR Morphology and distribution of Oligocene and Miocene pockmarks in the Danish North Sea implications for bottom current activity and fluid migration. Basin Research 20, Anell, I., Thybo, H., Stratford, W., 2010 Relating Cenozoic North Sea sediments to topography in southern Norway: The interplay between tectonics and climate Earth and Planetary Science Letters, 300, Anell I, Thybo H, Rasmussen E A synthesis of Cenozoic sedimentation in the North Sea. Basin Research 24, Bijlsma, S., 1981 Fluvial sedimentation from the fennoscandian area into the North-west European basin during the late Cenozoic Geologie en Mijnbouw, Boulton, G., Hagdorn, M., 2006, Glaciology of the British Ice Sheet during the last glacial cycle: form, flow, streams and lobes Quaternary Science Reviews, 25, ~ 179 ~

181 Boulton, GS., Hindmarsh, RCA., 1987, Sediment deformation beneath glaciers: rheology and geological consequences, Journal of Geophysical Research, 92(B9), Braconnot, P., Harrison, S.P., Kageyama, M., Bartlein, P.J., Masson-Delmotte, V., Abe-Ouchi, A., Otto-Bliesner, B., Zhao, Y., 2012 Evaluation of climate models using palaeoclimatic data, Nature Climate Change, 2, Bradwell, T., Stoker, M.S., Golledge, N.R., Wilson, C.K., Merritt, J.W., Long, D., Everest, J.D., Hestvik, O.B., Stevenson, A.G., Hubbard, A.L., Finlayson, A.G., Mathers, H.E., 2008, The northern sector of the last British Ice Sheet: Maximum extent and demise, Earth-Science Reviews, 88, British Geological Survey (BGS), 1986, 54 N 0 California Quaternary Geology, 1:250,000 Map Series, Keyworth, UK: BGS BGS, 1991, 55 N 0 Swallow Hole Quaternary Geology, 1:250,000 Map Series, Keyworth, UK: BGS BGS & Rijks Geologische Dienst, 1986, 54 N 2 E Indefatigable Quaternary Geology, 1:250,000 Map Series, Keyworth, UK: BGS BGS & Rijks Geologische Dienst, 1991, 55 N 2 E Dogger Quaternary Geology, 1:250,000 Map Series, Keyworth, UK: BGS Brzozowska, J, Eriksen, S., Holm, L., Olden, S Exploration History. In: Evans, D., Graham, S., Armour, A., Bathurst, P. (Eds) The Millennium Atlas: petroleum geology of the central and northern North Sea. London: The Geological Society of London, ~ 180 ~

182 Buckley, F.A., 2012, An Early Pleistocene grounded ice sheet in the Central North Sea, In: Huuse, M., Redfern, J., Le Heron, D.P., Dixon, R.J., Moscariello, A., Craig, J. (Eds) Glaciogenic reservoirs and Hydrocarbon Systems Geological Society, London, Special Publications, 368, Busschers, F.S., Kasse, C., van Balen, R.T., Vandenberghe, J., Cohen, K.M., Weerts, H.J.T., Wallinga, J., Johns, C., Cleveringa, P., Bunnik, F.P.M., 2007, Late Pleistocene evolution of the Rhine-Meuse system in the southern North Sea basin: imprints of climate change, sea-level oscillation and glacio-isostasy, Quaternary Science Reviews 26, Cameron, TDJ., Bonny, AP., Gregory DM., Harland, R Lower Pleistocene dinoflagellate cyst, foraminiferal and pollen assemblages in four boreholes in the Southern North Sea. Geological Magazine 121(2), Cameron TDJ., Bulat, J., Mesdag, CS., High resolution seismic profile through a Late Cenozoic delta complex in the southern North Sea. Marine and Petroleum Geology 10, Cameron TDJ., Crosby, A., Balson PS., Jeffery, DH., Lott, GK., Bulat, J., Harrison, DJ United Kingdom offshore regional report: The geology of the southern North Sea. London: HMSO for the British Geological Survey, pp. 152 Cameron, TDJ., Laban, C., Schüttenhelm, RTE., Upper Pliocene and Lower Pleistocene stratigraphy in the Southern Bight of the North Sea. In: Henriet, JP. De Moor, D., (Eds) The Quaternary and Tertiary Geology of the Southern Bight, North Sea Ghent: Ministry of Economic Affairs Belgian Geological Survey, Cameron, T.D.J., Stoker, M.S., Long, D., 1987, The history of Quaternary sedimentation in the UK sector of the North Sea Basin, Journal of the Geological Society 144, ~ 181 ~

183 Carr, S.J., Holmes, R., Van Der Meer, J.J.M., Rose, J., 2006, The Last Glacial Maximum in the North Sea Basin: micromorphological evidence of extensive glaciation, Journal of Quaternary Science, 21(3), Cartwright J Seismic-stratigraphical analysis of large-scale ridge-trough sedimentary structures in the Late Miocene to Early Pliocene of the central North Sea. Special Publications of the International Association of Sedimentologists 22, Caston VND Linear sand banks in the southern North Sea, Sedimentology 18, Caston VND. 1977, Quaternary deposits of the central North Sea, 2: The Quaternary deposits of the Forties field, northern North Sea, HMSO Caston VND, Stride AH Tidal sand movement between some linear sand banks in the North Sea off Norfolk. Marine Geology 9: M38-M42 Chadwick, R.A., Zweigel, P., Gregersen, U., Kirby, G.A., Holloway, S., Johannessen, P.N., 2004, Geological reservoir characterization of a CO2 storage site: the Utsira Sand, Sleipner, northern North Sea, Energy, 29, Cohen, K.M., Finney, S.C., Gibbard, P.L., Fan, J.-X., 2013 (updated), The ICS International Chronostratigraphic Chart, Episodes 36: Cohen, K.M., Gibbard, P., 2011, Global chronostratigraphic correlation table for the last 2.7 million years. Subcommision on Quaternary Stratigraphy (International Commission on Stratigraphy), Cambridge, England. ~ 182 ~

184 Cotterill, C., Dove, D., Long, D., James, L., Duffy, C., Mulley, S., Forsberg, C.F., Tjelta, TI., 2012, Dogger Bank A Geo Challenge, In: Offshore Site Investigation and Geotechnics: Integrated Technologies Present and Future, September, London, UK. Society of Underwater Technology, London, pp.8 Dowdeswell, J.A., Ottesen, D., 2013, Buried iceberg ploughmarks in the early Quaternary sediments of the central North Sea: A two-million year record of glacial influence from 3D seismic data, Marine Geology, 344, 1-9 Duncan, W.I., Green, P.F., Duddy, I.R., 1998, Source rock burial history and seal effectiveness: Key facets to understanding hydrocarbon exploration potential in the east and central Irish Sea basins, American Association of Petroleum Geologists Bulletin, 82 (7), Ehlers, J., Wingfield, R., 1991, The extension of the Late Weichselian/Late Devensian ice sheets in the North Sea Basin, Journal of Quaternary Science 6(4), Ehrenberg, S.N., Nadeau, P.H., Steen, O., 2009, Petroleum reservoir porosity versus depth: Influence of geological age, The American Association of Petroleum Geologists Bulletin, 93 (10), Eiken, O., Ringrose, P., Hermanrud, C., Nazarian, B., Torp, T.A., Høier, L., 2011, Lessons learned from 14 years of CCS operations: Sleipner, In Salah and Snøhvit, Energy Procedia, 4, EU, 2011, Implementation of Directive 2009/31/EC on the geological storage of carbon dioxide Guidance Document 2: Characterisation of the Storage Complex, CO2 stream composition, monitoring and corrective measures, ~ 183 ~

185 Evans, D., Graham, C., Armour, A., Bathurst, P., (Eds) 2003, The Millennium Atlas: Petroleum geology of the central and northern North Sea, The Geological Society of London, pp.989 Field, ME., Gardner, JV., Prior, DB Geometry and significance of stacked gullies on the northern California slope. Marine Geology 154: Fitch, S., Thomson, K., Gaffney, V., 2005, Late Pleistocene and Holocene depositional systems and the palaeogeography of the Dogger Bank, North Sea, Quaternary Research, 64, Funnell, BM Plio-Pleistocene palaeogeography of the southern North Sea basin ( Ma). Quaternary Science Reviews 15: Fyfe, JA., Gregersen, U., Jordt, H., Rundberg, Y., Eidvin, T., Evans, D., Stewart, D., Hovland, M., Andresen, P., Oligocene to Holocene. In: Evans, D., Graham, S., Armour, A., Bathurst, P. (Eds) The Millennium Atlas: petroleum geology of the central and northern North Sea. London: The Geological Society of London Galloway WE Paleogeographic setting and depositional architecture of a sand-dominated shelf depositional system, Miocene Utsira formation, North Sea Basin. Journal of Sedimentary Research 72(4): Gatliff, R.W., Richards, P.C., Smith, K, Graham, C.C., McCormac, M, Smith, N.J.P., Long, D, Cameron, T.D., Evans, D, Stevenson, A.G., Bulat, J, Ritchie, J.D United Kingdom offshore regional report: the geology of the central North Sea. London: HMSO for the British Geological Survey, pp. 118 Gibbard, P.L., 1979, Middle Pleistocene drainage in the Thames valley, Geological Magazine, 116(1), ~ 184 ~

186 Gibbard, P., 1991, Early and Early Middle Pleistocene correlations in the southern North Sea, Quaternary Science Reviews, 10, Gibbard, P.L., Head, M.J., Walker, M.J.C., 2010, Formal ratification of the Quaternary System/Period and the Pleistocene Series/Epoch with a base at 2.58 Ma, Journal of Quaternary Science 25(2), Gibbard P.L., Lewin, J., 2003, The history of the major rivers of southern Britain during the Tertiary, Journal of the Geological Society, London, 160, Gibbard, P.L., Lewin, J., 2009, River incision and terrace formation in the Late Cenozoic of Europe, Tectonophysics, 474, Goff, J.C., 1983, Hydrocarbon generation and migration from Jurassic source rocks in the E Shetland Basin and Viking Graben of the northern North Sea, Journal of the Geological Society of London, 140, Goledowski, B., Nielsen, S.B., Clausen, O.R., 2012, Patterns of Cenozoic sediment flux from western Scandinavia, Basin Research 24, Graham, A.G.C, Lonergan, L., Stoker, M.S., 2007, Evidence for Late Pleistocene ice stream activity in the Witch Ground Basin, central North Sea, from 3D seismic reflection data, Quaternary Science Reviews, 26, Graham AGC, Stoker MS, Lonergan L, Bradwell T, Stewart MA The Pleistocene Glaciations of the North Sea basin. In: Ehlers J, Gibbard PL. (Eds), Quaternary Glaciations- Extent and Chronology, 2 nd Edition Developments in Quaternary Science 15: ~ 185 ~

187 Guidish T.M., Kendall, C.G.St.C., Lerche, I., Toth, D.J., Yarzab, R.F., 1985, Basin evaluation using burial history calculations; an overview, The American Association of Petroleum Geologists Bulletin, 69 (1), Gyllencreutz, R., Backman, J., Jakobsson, M., Kissel, C., Arnold, E., Postglacial palaeoceanography in the Skagerrak. The Holocene 16(7): Harding, R., 2015, Evolution of the Giant Southern North Sea Shelf-Prism: Testing sequence stratigraphic concept and the global sea level curve with full-three dimensional control [Ph.D. Thesis]: University of Manchester: 296 p. Harris, PT., Whiteway, T., 2011, Global distribution of large submarine canyons: Geomorphic differences between active and passive continental margins, Marine Geology, 285, Harvey, A., Brauner, H-J., Breunese, JN., Hoffman, M., Jagosiak, P., Olsen, SB., Stoker, SJ., Van Orsmael, J., Pasternak, M., Conrad, N., Andersen, JH Licensing and exploration history. In: Doornenbal, JC., Stevenson, AG., (Eds) Petroleum and Geological Atlas of the Southern Permian Basin Area. Houten: EAGE publications b.v Haskell, N., Nissen, S., Hughes, M., Grindhaug, J., Dhanani, S., Heath, R., Kantorowicz, J., Antrim, L., Cubanski, M., Nataraj, R., Schilly, M., Wigger, S., 1999, Delineation of geological drilling hazards using 3D seismic attributes, The Leading Edge, 18 (3), Hepp, D.A., Hebbeln, D., Kreiter, S., Keil, H., Bathmann, C., Ehlers, J., Mörz, T., 2012, An eastwest-trending Quaternary tunnel valley in the south-eastern North Sea and its seismicsedimentological interpretation, Journal of Quaternary Science, 27(8), ~ 186 ~

188 Holmes, R., 1977, Quaternary deposits of the central North Sea 5. The Quaternary geology of the UK sector of the North Sea between 56 and 58 N, Report of Institute of Geological Sciences 77/14 50 pp Høyer, A.-S., Jørgensen, F., Piotrowski, J.A., Jakobsen, P.R., 2013, Deeply rooted glaciotectonism in western Denmark: geological composition, structural characteristics and the origin of the Varde hill-island, Journal of Quaternary Science, 28(7), Huuse, M, 2002, Cenozoic uplift and denudation of southern Norway: insights from the North Sea Basin, In: Dore, A.G., Cartwright, J.A., Stoker, M.S., Turner, J.P. White, N, (Eds) Exhumation of the North Atlantic margin: Timing, Mechanisms and Implications for Petroleum Exploration, Geological Society, London, Special Publications, 196, Huuse, M., Clausen, OR., 2001, Morphology and origin of major Cenozoic sequence boundaries in the eastern North Sea Basin: top Eocene, near-top Oligocene and the mid-miocene unconformity, Basin Research, 13, Huuse, M., Lykke-Andersen, H., 2000a, Overdeepened Quaternary valleys in the eastern Danish North Sea: morphology and origin, Quaternary Science Reviews, 19, Huuse, M., Lykke-Andersen, H., 2000b, Large-scale glaciotectonic thrust structures in the eastern Danish North Sea, In: Deformation of Glacial Materials, Maltman, A.J., Hubbard, B., Hambery, M.J. (Eds) Geological Society, London, Special Publications, 176, Jansen, E., Fronval, T., Rack, F., Channell, J.E.T., 2000, Pliocene-Pleistocene ice rafting history and cyclicity in the Nordic Seas during the last 3.5 Myr, Palaeoceanography, 15(6), Jobe, ZR., Lowe, DR., Uchytil, SJ Two fundamentally different types of submarine canyons along the continental margin of Equatorial Guinea. Marine and Petroleum Geology ~ 187 ~

189 Jørgensen, F., Sandersen, PBE., 2006, Buried and open tunnel valleys in Denmark erosion beneath multiple ice sheets, Quaternary Science Reviews, 25, Kehew, A.E., Piotrowski, J.A., Jørgensen, F., 2012, Tunnel valleys: Concepts and controversies A review, Earth Science Reviews, 113, Kilhams, B., McArthur, A., Huuse, M., Ita, E., Hartley, A., 2011, Enigmatic large-scale furrows of Miocene to Pliocene age from the central North Sea: current-scoured pockmarks? Geo-Marine Letters 31, King, C., 1983 Cainozoic micropaleontological biostratigraphy of the North Sea H.M.S.O. London pp.40 Kluiving, S.J., Alied Bosch, J.H., Ebbing, J.H.J., Mesdag, C.S., Westerhoff, R.S., 2003, Onshore and offshore seismic and lithostratigraphic analysis of a deeply incised Quaternary buried valleysystem in the Northern Netherlands Journal of Applied Geophysics, 53, Knox, RWOB, Bosch, JHA, Rasmussen, ES, Heilmann-Clausen, C., Hiss, M., De Lugt, IR., Kasińksi, J., King, C., Köthe, A., Słodkowska, B., Standke, G., Vandenberghe, N Cenozoic. In: Doornenbal, JC., Stevenson, AG., (Eds) Petroleum and Geological Atlas of the Southern Permian Basin Area. Houten: EAGE publications b.v Knudsen, K.L., Asbjörnsdóttir, L., 1991 Plio-Pleistocene foraminiferal stratigraphy and correlation in the central North Sea Marine Geology 101, Knudsen, KL., Sejrup, HP Pleistocene stratigraphy in the Devils Hole area, central North Sea: foraminiferal and amino-acid evidence. Journal of Quaternary Science. 8(1) 1-14 ~ 188 ~

190 Knutz, P.C., 2010 Channel structures formed by contour currents and fluid expulsion: significance for Late Neogene development of the central North Sea basin Petroleum Geology Conference series 7, Kristensen, T.B., Huuse, M., 2012 Multistage erosion and infill of buried Pleistocene tunnel valleys and associated seismic velocity effects From: Huuse, M., Redfern, J., Le Heron, D.P., Dixon, R.J., Moscariello, A., Craig, J. (Eds) Glaciogenic reservoirs and Hydrocarbon Systems Geological Society, London, Special Publications, 368 Kristensen, T.B., Huuse, M., Piotrowski, J.A., Clausen, O.R., 2007, A morphometric analysis of tunnel valleys in the eastern North Sea based on 3D seismic data Journal of Quaternary Science, 22(8), Kuhlmann, G., Langereis, C.G., Munsterman, D., van Leeuwen, R.-J., Verreussel, R., Meulenkamp, J.E., Wong, T.E., (2006) Integrated chronostratigraphy of the Pliocene-Pleistocene interval and its relation to the regional stratigraphical stages in the southern North Sea region Netherlands Journal of Geoscience, 85(1), Kuhlmann, G., Wong, T.E., 2008, Pliocene palaeo-environment evolution as interpreted from 3Dseismic data in the southern North Sea, Dutch offshore sector Marine and Petroleum Geology, 25, Larsen, E., Sejrup, H.P., Janocko, J., Landvik, J.Y., Stalsberg, K., Steinsund, P.I., 2000, Recurrent interaction between the Norwegian Channel Ice Stream and terrestrial-based ice across southwest Norway Boreas, 29, Lisiecki, L.E., Raymo, M.E., 2005, A Pliocene-Pleistocene stack of 57 globally distributed benthic δ 18 O records, Palaeoceanography, 20(1), PA1003, pp.17 ~ 189 ~

191 Lisiecki, L.E., Raymo, M.E., 2007, Plio-Pleistocene climate evolution: trends and transitions in glacial cycle dynamics, Quaternary Science Reviews, 26, Lonergan, L., Maidment, S.C.R., Collier, J.S., 2006, Pleistocene subglacial tunnel valleys in the central North Sea basin: 3-D morphology and evolution, Journal of Quaternary Science, 21(8), Marcussen, Ø., Faleide, J.I., Jahren, J., Bjørlykke, K., 2010, Mudstone compaction curves in basin modelling: a study of Mesozoic and Cenozoic Sediments in the northern North Sea, Basin Research, 22, Martínez-Botí, M.A., Foster, G.L., Chalk, T.B., Rohling, E.J., Sexton, P.F., Lunt, D.J., Pancost, R.D., Badger, M.P.S., Schmidt, D.N., 2015, Plio-Pleistocene climate sensitivity evaluated using high-resolution CO 2 records, Nature, 518, McCave IN Sand waves in the North Sea off the coast of Holland. Marine Geology 10: McMillan, A.A., Hamblin, R.J.O., Merritt, J.W., 2005, An overview of the lithostratigraphical framework for the Quaternary and Neogene deposits of Great Britain (onshore), British Geological Survey Research Report RR/04/04 38 p. Meijer, T., Cleveringa, P., Munsterman, DK., Verreussel, RMCG., The Early Pleistocene Praetiglian and Ludhamian pollen stages in the North Sea Basin and their relationship to the marine isotope record. Journal of Quaternary Science. 21; ~ 190 ~

192 Miller, K.G., Kominz, M.A., Browning, J.V., Wright, J.D., Mountain, G.S., Katz, M.E., Sugarman, P.J., Cramer, B.S., Christie-Blikc, N., Pekar, S.F., 2005, The Phanerozoic Record of Global Sea- Level Change, Science, 310, Miller, KG., Mountain, GS., Wright, JD., Browning, JV A 180-million-year record of sea level and ice volume variations from continental margin and deep-sea isotopic records. Oceanography 24(2) Mitchum Jr, R.M., Vail, P.R., Thompson III, S., 1977a, Seismic stratigraphy and Global Changes of Sea Level, Part 2: The Depositional Sequence as a Basic Unit for Stratigraphic Analysis, In: Payton, C.E. (Eds), Seismic stratigraphy applications to hydrocarbon exploration, American Association of Petroleum Geologists Memoir 26, p Mitchum Jr, R.M., Vail, P.R., Thompson III, S., 1977b, Seismic stratigraphy and Global Changes of Sea Level, Part 6: Stratigraphic Interpretation of Seismic Reflection Patterns in Depositional Sequences, In: Payton, C.E. (Eds), Seismic stratigraphy applications to hydrocarbon exploration, American Association of Petroleum Geologists Memoir 26, p Moreau, J., Huuse, M., 2014, Infill of tunnel valleys associated with landward-flowing ice sheets: The missing Middle Pleistocene record of the NW European rivers? Geochemistry Geophysics Geosystems 15, 1-9 Mudelsee, M., Raymo, M.E., 2005, Slow dynamics of the Northern Hemisphere glaciation, Palaeoceanography, 20, PA4022 Nielsen, T., Mathiesen, A., Bryde-Auken, M., 2008, Base Quaternary in the Danish parts of the North Sea and Skagerrak Geological Survey of Denmark and Greenland Bulletin 15, ~ 191 ~

193 Noorbergen, LJ., Lourens, LJ., Munsterman, DK., Verreussel, RMCH Stable isotope stratigraphy of the early Quaternary of borehole Noordwijk, southern North Sea. Quaternary International 386, Ó Cofaigh, 1996, Tunnel valley genesis, Progress in Physical Geography, 20(1), 1-19 OSPAR commission Quality Status Report 2000, Region II Greater North Sea. OSPAR Commission, London xiii pp. Ottesen, D., Dowdeswell, J.A., Bugge, T., 2014, Morphology, sedimentary infill and depositional environments of the early Quaternary North Sea Basin (56-62 N), Marine and Petroleum Geology 56, Ottesen, D., Dowdeswell, J.A., Rise, L., Rokoengen, K., Henriksen, S., 2002, Large-scale morphological evidence for past ice-stream flow on the mid-norwegian continental margin, In: Dowdeswell, J.A., Ó Cofaigh, C., (Eds), Glacier-influenced Sedimentation on High-Latitude Continental Margins, Geological Society, London, Special Publications, 203, Otto L, Zimmerman JTF, Furness GK, Mork M, Saetre R, Becker G Review of the physical oceanography of the North Sea. Netherlands Journal of Sea Research 26(2-4): Overeem, I, Weltje, G.J., Bishop-Kay, C., Kroonenberg, S.B., 2001, The Late Cenozoic Eridanos delta system in the Southern North Sea Basin: a climate signal in sediment supply? Basin Research 13, Partridge, T.C., 1997, Reassessment of the position of the Plio-Pleistocene boundary: is there a case for lowering it to the Gauss-Matuyama palaeomagnetic reversal? Quaternary International, 40, 5-10 ~ 192 ~

194 Patruno, S., Hampson, G.J., Jackson, C.A.-L., Whipp, P.S., 2014, Quantitative progradation dynamics and stratigraphic architecture of ancient shallow-marine clinoform sets: a new method and its application to the Upper Jurassic Sognefjord Formation, Troll Field, offshore Norway, Basin Research 1-41 Patruno, S., Hampson, G.J., Jackson, C.A-L., 2015, Quantitative characterisation of deltaic and subaqueous clinoforms, Earth-Science Reviews, 142, Pekar, S.F., Miller, K.G., Kominz, M.A., 2000, Reconstructing the stratal geometry of latest Eocene to Oligocene sequence in New Jersey; resolving a patchwork distribution into a clear pattern of progradation, Sedimentary Geology 134, Peltier, W.R., 1994, Ice Age Paleotopography, Science, 265 (5169), Phillips, E., Lee, JR., Burke, H., 2008, Progressive proglacial to subglacial deformation and syntectonic sedimentation at the margins of the Mid-Pleistocene British Ice Sheet: evidence from north Norfolk, UK. Quaternary Science Reviews, 27, Piotrowski, J.A., 1997, Subglacial hydrology in North-Western Germany during the last glaciation: groundwater flow, tunnel valleys and hydrological cycles, Quaternary Science Reviews, 16, Posamentier, H.W., Allen, G.P., 1993, Variability of the sequence stratigraphic model: effects of local basin factors, Sedimentary Geology, 86, Posamentier, H.W., Davies, R.J., Cartwright, J.A., Wood, L. 2007, Seismic geomorphology an overview, In: Davies, R.J, Posamentier, H.W., Wood, L.J., Cartwright, J.A. (Eds), Seismic ~ 193 ~

195 Geomorphology: Applications to Hydrocarbon Exploration and Production, Geological Society Special Publications, 277, 1-14 Posamentier, H.W., Vail, P.R., 1988, Eustatic controls on clastic deposition II Sequence and systems tract models, In: Wilgus, CK., Hastings, BS., Posamentier, H., Van Wagoner, J., Ross, CA., St. C. Kendall, CG. (Eds), Sea-Level Changes An Integrated Approach, SEPM Special Publication, 42, Praeg, D, 1996, Morphology, stratigraphy and genesis of buried mid-pleistocene tunnel-valleys in the southern North Sea basin, [Ph.D. thesis]: University of Edinburgh 207 p. Praeg, D., 2003, Seismic imaging of mid-pleistocene tunnel-valleys in the North Sea Basin high resolution from low frequencies, Journal of Applied Geophysics 53, Praston, LF., Ryan, WBF., Mountain, GS., Twichell, DC Submarine canyon initiation by downslope-eroding sediment flows: Evidence in the late Cenozoic strata on the New Jersey continental slope, Geological Society of American Bulletin, 106, Rasmussen ES, Vejbæk OV, Bidstrup T, Piasecki S, Dybkjær K Late Cenozoic depositional history of the Danish North Sea Basin: implications for the petroleum systems in the Kraka, Halfdan, Siri and Nini fields. In Petroleum Geology: North-West Europe and Global Perspectives Proceedings of the 6 th Petroleum Geology Conference Doré AG, Vining BA, (Eds) Raymo, M.E., 1994, The Initiation of Northern Hemisphere Glaciation, Annual Review of Earth and Planetary Sciences, 22, Raymo, M.E., Nisancioglu, K., 2003, The 41 kyr world: Milankovitch s other unsolved mystery, Palaeoceanography, 18 (1), 1011 ~ 194 ~

196 Raymo, M.E., Ruddiman, W.F., 1992, Tectonic forcing of late Cenozoic climate, Nature, 359(6391), Rider, M.H., 2002, The geological interpretation of well logs, 2 nd Ed, Rider-French Consulting, Sutherland, Scotland, 280 pp. Rise, L., Chand, S., Hjelstuen, B.O., Haflidason, H., Bøe, R., 2010, Late Cenozoic geological development of the south Vøring margin, mid-norway, Marine and Petroleum Geology, 27, Rohling, E.J., Medina-Elizalde, M., Shepherd, J.G., Siddal, M., Stanford, J.D., 2012a, Sea surface and high-latitude temperature sensitivity to radiative forcing of climate over several glacial cycles, Journal of Climate, 25, Rohling, EJ., Sluijs, A., Dijkstra, HA., Köhler, P., van de Wal, PSW., von der Heydt, AS., et al. 2012b, Making sense of palaeoclimate sensitivity, Nature, 491, Rose, J., 2009, Early and Middle Pleistocene landscapes of eastern England, Proceedings of the Geologists Association, 120, 3-33 Ruiter, A., Dove, D., 2014, Glaciotectonism on the Dogger Bank; a case study from Tranche-A, Technical Report BGS-TN , Forewind Consortia-Dogger Bank Windfarm. Rundberg, Y., Eidvin, T., 2005, Controls on depositional history and architecture of the Oligocene- Miocene succession, northern North Sea Basin, Norwegian Petroleum Society Special Publications, 12, ~ 195 ~

197 Ryan, W.B.F., S.M. Carbotte, J.O. Coplan, S. O'Hara, A. Melkonian, R. Arko, R.A. Weissel, V. Ferrini, A. Goodwillie, F. Nitsche, J. Bonczkowski, and R. Zemsky (2009), Global Multi- Resolution Topography synthesis, Geochem. Geophys. Geosyst., 10, Q03014, from: Scourse, J.D., Ansari, M.H., Wingfield, R.T.R., Harland, R., Balson, P.S., 1998, A Middle Pleistocene shallow marine interglacial sequence, Inner Silver Pit, Southern North Sea: pollen and dinoflagellate cyst stratigraphy and sea-level history, Quaternary Science Reviews, 17, Sejrup, HP., Aarseth, I., Ellingsen, KL., Reither, E., Jansen, E., Løvlie, R., Bent, A., Brigham- Grette, J., Larsen, E., Stoker, M Quaternary stratigraphy of the Fladen area, central North Sea: a multidisciplinary study. Journal of Quaternary Science. 2, Sejrup, HP., Aarseth, I., Haflidason, H The Quaternary succession in the northern North Sea. Marine Geology. 101, Sejrup, HP., Haflidason, H., Aarseth, I., King, E., Forsberg, CF., Long, D., Rokoengem, K Late Weichselian glaciation history of the northern North Sea. Boreas. 23, 1-13 Sejrup, H.P., Hjelsuen, B.O., Torbjørn Dahlgren, K.I., Haflidason, H., Kuijpers, A., Nygård, A., Praeg, D., Stoker, M.S., Vorren, T.O., 2005, Pleistocene glacial history of the NW European continental margin, Marine and Petroleum Geology, 22, Sejrup, H.P., King, E.L., Aarseth, I., Haflidason, H., Elverhøi, 1996, Quaternary erosion and depositional processes: western Norwegian fjords, Norwegian Channel and North Sea Fan, Geological Society of London, Special Publications, 117, ~ 196 ~

198 Sejrup, H.P., Landvik, J.Y., Larsen, E., Janocko, J., Eiriksson, J., King, E., 1998, The Jæren area, a border zone of the Norwegian Channel Ice Stream, Quaternary Science Reviews, 17, Sejrup, H.P., Larsen, E., Haflidason, H., Berstad, I.M., Hjelstuen, B.O., Jonsdottir, H.E., King, E.L., Landvik, J., Longva, O., Nygard, A., Ottesen, D., Raunholm, S., Rise, L., Stalsberg, K., 2003, Configuration, history and impact of the Norwegian Channel Ice Stream, Boreas, 32, Sejrup, HP., Larsen, E., Landvik, J., King, EL., Haflidason, H., Nesje, A., 2000, Quaternary glaciations in southern Fennoscandia: evidence from southwestern Norway and the northern North Sea region, Quaternary Science Reviews, 19, Shaw, J. Todd, BJ. Li, MZ. Wu, Y.,2012. Anatomy of the tidal scour system at Minas Passage, Bay of Fundy, Canada. Marine Geology , Shepard. FP Types of Submarine Valleys: Geological Notes. AAPG Bulletin, 49(3), Shepard, FP Submarine Canyons: Multiple Causes and Long-Time Persistence. AAPG Bulletin. 65(6), Sørensen, J.C., Gregersen, U., Breiner, M., Michelsen, O., 1997, High-frequency sequence stratigraphy of Upper Cenozoic deposits in the central and south-eastern North Sea areas, Marine and Petroleum Geology, 14(2), Spinelli, GA., Field, ME., Evolution of continental slope gullies on the northern California margin. Journal of Sedimentary Research. 71(2), Steffensen, J.P., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl-Jansen, D., Fischer, H., Goto- Azuma, K., Hansson, M., Johnsen, S.J., Jouzel, J., Masson-Delmotte, V., Popp, T., Rasmussen, ~ 197 ~

199 S.O., Röthelisberber, R., Ruth, U., Stauffer, B., Siggard-Andersen, M-L, Sveinbjörnsdóttir, Á.E., Svensson, A., White, J.W.C., 2008, High-resolution Greenland Ice Core data show abrupt climate change happens in few years, Science, 321, Stewart, M.A., 2008, 3D Seismic Analysis of Pleistocene Tunnel Valleys in the Central North Sea, [Ph.D. Thesis]: Imperial College London, 312 p. Stewart, M.A., Lonergan, L., 2011, Seven glacial cycles in the middle-late Pleistocene of northwest Europe: Geomorphic evidence from buried tunnel valleys, Geology 39, Stewart, M., Lonergan, L., Hampson, G., 2012, 3D seismic analysis of buried tunnel valleys in the Central North Sea: tunnel valley fill sedimentary architecture In: Huuse, M., Redfern, J., Le Heron, D.P., Dixon, R.J., Moscariello, A., Craig, J. (Eds), Glaciogenic reservoirs and Hydrocarbon Systems, Geological Society, London, Special Publications, 368 Stewart, M.A., Lonergan, L., Hampson, G., 2013, 3D seismic analysis of buried tunnel valleys in the central North Sea: morphology, cross-cutting generations and glacial history, Quaternary Science Reviews, 72, 1-17 Stoker M.S., Balson, P.S., Long, D, Tappin, D.R., 2011, An overview of the lithostratigraphical framework for the Quaternary deposits on the United Kingdom continental shelf. British Geological Survey Research Report, RR/11/ p. Stoker, MS., Bent, A Middle Pleistocene glacial and glaciomarine sedimentation in the west central North Sea. Boreas 14, Stoker, M., Bradwell, T., 2005, The Minch palaeo-ice stream, NW sector of the British-Irish Ice Sheet, Journal of the Geological Society, London, 163, ~ 198 ~

200 Stoker MS, Praeg D, Hjelstuen BO, Laberg JS, Nielsen T, Shannon PM, 2005a. Neogene stratigraphy and the sedimentary and oceanographic development of the NW European Atlantic margin. Marine and Petroleum Geology 22: Stoker, M.S., Praeg, D., Shannon, P.M., Hjelstuen, B.O., Laberg, J.S., Nielsen, T., van Weering, T.C.E., Sejrup, H.P., Evans, D., 2005b, Neogene evolution of the Atlantic continental margin of NW Europe (Lofoten Islands to SW Ireland); anything but passive. In: Dore, A.G., Vining, B., (Eds) Petroleum Geology: North West Europe and Global Perspectives, Proceedings of the 6 th Petroleum Geology Conference, Stoker, MS., Skinner, AC., Fyfe, JA., Long. D Palaeomagnetic evidence for early Pleistocene in the central and northern North Sea. Nature. 304, Stow, DAV., Mayall, M Deep-water sedimentary systems: New models for the 21 st century. Marine and Petroleum Geology 17, Stuart, JY., Huuse, M., D seismic geomorphology of a large Plio-Pleistocene delta Bright spots and contourites in the Southern North Sea. Marine and Petroleum Geology 38(1), Svendsen, E., Sætre, R., Mork, M Features of the northern North Sea circulation. Continental Shelf Research 11(5), Ten Veen, JH, Verweij, H., Donders, T., Geel, K., de Bruin, G/, Munsterman, D., Verreussel, R., Daza Cajigal, V., Harding R., Cremer, H., Anatomy of the Cenozoic Eridanos Hydrocarbon System. TNO Report R10060 Terwindt JHJ Sand waves in the Southern Bight of the North Sea Marine Geology 10: ~ 199 ~

201 Thöle, H, Gaedicke, C, Kuhlmann, G, Reinhardt, L Late Cenozoic sedimentary evolution of the German North Sea? A seismic stratigraphical approach. Newsletters on Stratigraphy, 47; Turrell, WR., Henderson, EW., Slesser, G., Payne, R., Adams, RD Seasonal changes in the circulation of the northern North Sea. Continental Shelf Research. 12(2-3), UKCS, 2015, UKCS Offshore Infrastructure Map, Department of Energy and Climate, London Van der Vegt, P., Janszen, A., Moscariello, A., 2012, Tunnel valleys: current knowledge and future perspectives, Geological Society, London, Special Publications, 368, Van Dijke, J.J., Veldkamp, A., 1996, Climate-controlled glacial erosion in the unconsolidated sediments of North-western Europe, based on a genetic model for tunnel-valley formation, Earth Surface Processes and Landforms, 21, Veenstra, H.J., 1965, Geology of the Dogger Bank area, North Sea, Marine Geology, 3, Viezer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, D., Bruhn, F., Carden, GAF., Diener, A., Ebneth, S., Godderis, Y., Jasper, T., Korte, C., Pawellek, F., Podlaha, OG., Strauss, H., 1999, 87 Sr/ 86 SR, δ 13 C and δ 18 O evolution of Phanerozoic seawater, Chemical Geology, 161, Voelker, A.H.L., and workshop participants, 2002, Global distribution of centennial-scale records for Marine Isotope Stage (MIS) 3: a database, Quaternary Science Reviews, 21, White, N., Lovell, B., 1997, Measuring the pulse of a plume with the sedimentary record, Nature 387, ~ 200 ~

202 Wingfield, RTR Glacial incisions indicating Middle and Upper Pleistocene ice limits off Britain. Terra Research, 1, Wingfield, R The Origin of Major Incisions Within the Pleistocene Deposits of the North Sea. Marine Geology. 91, Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001, Trends, Rhythms and Aberrations in Global Climate 65 Ma to Present, Science, 292, Ziegler PA North Sea rift system Tectonophysics 208: ~ 201 ~

203 Appendix 1: Project Datasets ~ 202 ~

204 Name Company 3D Extent Dominant Frequencies Sampling Rate Vertical Resolution Bin Size / Line Spacing Survey Date A1.1: Seismic Datasets CNS MegaSurvey PGS Y 88,500 km Hz 4 ms 8-16 m 50 m 2012 SNS MegaSurvey PGS Y 39,500 km Hz 4 ms 8-16 m 50 m 2012 BroadSeis TM CGG Y 2, km 2 Hz 4 ms 7-8 m 12.5 m 2010 Dogger Bank Tranche A Forewind N 2,000 km Hz 5 ms 1 m 100m inline 500m x- line 2011 North Sea Renaissance TGS N - various 4 ms m 1-15 km various A1.2: List of BGS Cores Name Total Depth (m) Water Depth (m) Recovery (%) Drill Date 77/ June / / June / / Aug / Aug / Aug / / / / / / ~ 203 ~

205 A1.3: List of Wells by Chapter All well data has been obtained from TGS Facies Map Browser, with the exception of A15-03 and Josephine-1. Data includes well headers, total depth, deviation surveys, gamma log and lithological descriptions for all wells and, where available, calibrated time-depth logs and sonic logs. Chapter 2: The early Quaternary North Sea Basin NO 1/9-2 2/3-3 2/5-4 1/9-3 2/3-4 2/5-5 1/2-1 1/9-4 2/4-2 2/5-6 1/3-1 1/9-5 2/4-3 2/5-7 1/3-10 1/9-6S 2/4-4 2/5-8 1/3-11 2/1-1 2/4-5 2/5-9 1/3-2 2/1-2 2/4-6 2/5-10A 1/3-4 2/1-3 2/4-7 2/5-11 1/3-5 2/1-4 2/4-8 2/5-13 1/3-6 2/1-5 2/4-9 2/6-1 1/3-7 2/1-6 2/4-10 2/6-2 1/3-8 2/1-7 2/4-11 2/6-3 1/3-9S 2/1-8 2/4-12 2/6-4S 1/5-2 2/1-9 2/4-14 2/6-5 1/6-1 2/1-10 2/4-16 2/7-1 1/6-2 2/1-11 2/4-17 2/7-2 1/6-3 2/1-12 2/4-18R 2/7-3 1/6-4 2/2-1 2/4-19B 2/7-4 1/6-5 2/2-2 2/4-20 2/7-5 1/6-6 2/2-3 2/5-1 2/7-6 1/6-7 2/2-4 2/5-2 2/7-7 1/9-1 2/2-5 2/5-3 2/7-8 ~ 204 ~

206 ~ 205 ~ 2/7-9 2/7-10 2/7-11 2/7-12 2/7-13 2/7-14 2/7-15 2/7-16 2/7-19 2/7-20 2/7-21S 2/7-22 2/7-23S 2/7-26S 2/7-27S 2/7-28 2/7-29 2/8-2 2/8-3 2/8-4 2/8-5 2/8-6 2/8-7 2/8-8 2/8-9 2/8-10 2/8-11 2/8-12S 2/8-13 2/8-14 2/8-15 2/9-1 2/9-1 2/9-3 2/10-2 2/11-1 2/11-2 2/11-4 2/11-5 2/11-6S 2/11-7 2/11-8 2/11-9 2/12-1 3/4-1 3/5-1 3/5-2 3/6-1 3/7-2 3/7-3 3/7-4 3/7-5 3/7-6 3/7-7 6/3-1 6/3-2 7/1-1 7/4-1 7/4-2 7/7-1 7/7-2 7/7-3 7/8-1 7/8-2 7/8-3 7/8-4 7/9-1 7/11-1 7/11-2 7/11-3 7/11-4 7/11-5 7/11-6 7/11-7 7/11-8 7/11-9 7/11-10S 7/11-12S 7/12-1S 7/12-2 7/12-3 7/12-3A 7/12-4 7/12-5 7/12-6 7/12-7 7/12-8 7/12-9 7/ / /9-1 8/10-1 8/10-2 8/10-3 8/10-4S 8/11-1 8/12-1 9/1-1S 9/2-1 9/2-2 9/2-4S 9/3-1 15/3-4 15/6-9S 15/ /6-11A 15/6-11S 15/ /8-2 15/12-8A 15/ /1-3

207 ~ 206 ~ 16/1-4 16/1-5 16/1-8 16/ / /2-7 16/ /3-4 16/4-2 16/4-4 16/4-5 16/7-2 16/7-9 16/ /12-4 UK Josephine-1 14/ / / / / /15a-4 14/15b-3 14/17a-2 14/ / / / /18a-6 14/18a-8 14/18a-8Z 14/18a-9 14/18a-10 14/18a-13 14/18a-14 14/18b-7 14/18b-11 14/18b-12 14/ / / / / / / / / / / / / / / / / / / / / / / / / / / /20b-17 14/20b-18 14/20b-22 14/20b-22 14/20b-24 14/20b-29 14/ /24a-3 14/24a-4 14/24a-5 14/25a-4 14/28a-1 14/28b-2 14/28b-4 14/29a-2 14/29a-3 14/29a-4 14/ /30a-2 14/30a-4 14/30b-3 15/6-1 15/7-1 15/8-1 15/11a-4 15/12a-3 15/12b-4Y 15/ /13a-4 15/13b-5 15/13b-6 15/14a-3 15/14b-1 15/ / /16-4A 15/ / / / /16b-20 15/16b-21 15/17-1A 15/17-2

208 ~ 207 ~ 15/ / / / /17-8A 15/ / / / /17-20Z 15/ /17-22Z 15/ / / / / /18a-6 15/18a-7 15/18b-3 15/18b-10 15/ / / / / / / / /20a-5 15/20a-9 15/20a-10 15/20a-13 15/20a-14 15/20a-17 15/20b-18Z 15/ / / / / /21a-7 15/21a-8 15/21a-9 15/21a-10 15/21a-11 15/21a-12 15/21a-12A 15/21a-13 15/21a-14 15/21a-15 15/21a-20 15/21a-34 15/21a-36 15/21a-38Z 15/21a-39 15/21a-40 15/21a-43 15/21a-44 15/21a-46 15/21a-51 15/21a-52 15/21a-52Z 15/21a-54 15/21a-55 15/21b-21 15/21b-37 15/21b-41 15/21b-45 15/21b-47 15/21b-48 15/21b-50 15/21b-53 15/21b-56 15/ / / / / / / / /22-9Z 15/ / /22-12Z 15/ / / / /23-6A 15/23a-7 15/23a-9 15/23a-11A 15/23a-12 15/23b-14 15/23d-13 15/ /24a-2 15/24a-7 15/24a-9 15/24b-3 15/24b-5 15/25a-4 15/25b-1A 15/25b-3 15/ /26a-2 15/26a-6A 15/26a-6Z 15/26a-7 15/26b-5

209 ~ 208 ~ 15/26b-9 15/ / / / / / / /28a-3 15/28a-8 15/28b-4 15/28c-6 15/28c-9 15/ / /29a-6 15/29a-7 15/29a-8 15/29a-10 15/29a-11 15/29b-4 15/ / / / / / / / /3a-4 16/3b-7 16/6a-2 16/6a-3 16/6a-4 16/6a-8 16/6b-5 16/6b-6 16/6b-7 16/7-3 16/7a-4A 16/7a-5 16/7a-6 16/7a-10 16/7a-11 16/7a-12 16/7a-13 16/7a-14 16/7a-15 16/7a-16 16/7a-17B 16/7a-18 16/7a-19 16/7a-22 16/7a-27 16/7a-30Z 16/7a-31 16/7a-32 16/7b-21 16/7b-23 16/7b-24 16/7b-25 16/7b-28Z 16/8-1 16/8a-4 16/8a-8 16/8a-9 16/8a-10 16/8a-11 16/8b-2 16/8b-5 16/8b-6 16/8c-13 16/ /11a-3 16/12a-2 16/12a-5 16/12a-17 16/12b-6 16/12b-10 16/12b-12 16/ /13a-2 16/13a-2Z 16/13a-3 16/13a-4 16/13a-5 16/16a-3 16/16b-1 16/16b-4 16/16b-5 16/17-2A 16/ / / /17-8A 16/ / / / / / / / / / / /21a-2 16/21a-6 16/21a-7 16/21a-8 16/21a-13 16/21a-15

210 ~ 209 ~ 16/21a-17 16/21a-17Z 16/21a-18 16/21a-19 16/21a-23 16/21b-4A 16/21b-5 16/21b-9 16/21b-21 16/21c-22 16/21d-28 16/21d-30 16/21d-31 26/21d-36 16/ / / / / / / / / /24a-1 16/ / / / /26-7Y 16/ /26-8Z 16/ / / /26-12Z 16/ / / / / / / / / /27a-2 16/27a-3 16/27b-4Z 16/27b-5 16/ / / / / / / / / / / /29a-5 16/29a-9 16/29c-7 16/29c-10 16/29c-14 19/ /1-1 20/1-2 20/1-3 20/1-4 20/1-8 20/1-9 20/2-1 20/2-2 20/2-3 20/2-6 20/2-7 20/3-1 20/3-2A 20/3-3 20/3-5 20/3a-6 30/4-1 20/4a-2 20/4a-5 20/4a-9 20/4a-10Z 20/4b-4 20/4b-6 20/5a-5 20/5b-2 20/5b-3 20/5c-4 20/5c-6 20/6-1 20/6-2 20/6-3 20/6-3Z 20/6-4 20/6-4Z 20/6-5 20/6-6 20/7-1 20/7-2 20/7a-3 20/7b-5 20/8-1 20/8-2 20/8-3 20/9-2 20/9-3 20/ /10-2

211 ~ 210 ~ 20/10a-3 20/10b-4 20/10b-5 20/ / / / / / /1-1 21/1-3 21/1a-12 21/1a-15 21/1a-20 21/1b-17 21/2-1 21/2-2 21/2-3 21/2-4 21/2-5 21/2-6 21/2-8 21/2-9 21/3-1A 21/3-2 21/3b-3 21/3b-6 21/4-1 21/4-3 21/4b-5 21/5b-1 21/6-b 21/6-2 21/6a-3 21/6b-4 21/6b-5 21/7-1 21/7-2 21/7a-3 21/8-1 21/8-2 21/8-4 21/9-2 21/9-4 21/9-6 21/ / / / / / / / / / / / / /12-2B 21/13a-3 21/13b-1A 21/14a-4 21/14b-2 21/14b-3 21/15a-1 21/15a-2 21/15a-3 21/15a-6 21/15b-5 21/ / / / / / / / /17b-5 21/18-1A 21/18-2A 21/ / / / / / / /20a-1 21/20a-2 21/20a-5 21/20b-3 21/ /23b1 21/23b-3 21/23c-5 21/ / / / / / /26-1D 21/29a-6 21/ / / /30-6A 21/ / /30-9A 21/ /30-14

212 ~ 211 ~ 21/ / / /1a-3 22/1a-4 22/1b-8 22/1b-10 22/1c-7 22/1c-9 22/2-1 22/2-2 22/2-3 22/2-4 22/2-5 22/2a-7 22/2b-9 22/2c-10 22/2c-10Z 22/3a-1 22/3a-1 22/3a-2 22/3a-3 22/4-1 22/4-2 22/4-3 22/4b-4A 22/4b-5 22/5a-1A 22/5a-6 22/5a-10 22/5b-3 22/5b-4 22/5b-5 22/5b-7 22/5b-8 22/5b-9 22/5b-12 22/5b-14 22/6-1 22/6a-8 22/6a-9 22/6a-10 22/6a-11 22/6a-12 22/6a-14 22/6b-13 22/7-1 22/7-2 22/8a-2 22/8a-3 22/9-1 22/9-2 22/9-3 22/9-4 22/9-5 22/ /10a-2 22/10a-3 22/10a-4 22/10a-5 22/10a-6 22/ / / / / /11-8Z 22/ /11b-13 22/12a-1 22/12a-2 22/12a-3 22/12a-6 22/12b-7A 22/13a-1 22/13a-2 22/13a-4 22/13b-5 22/13b-6 22/ /14a-2 22/14b-4 22/14b-5 22/ / / /16a-4 22/16b-3 22/ / / / / / /19a-3 22/19b-4 22/19b-5Z 22/19c-6 22/ /20-3Z 22/ / / / / /21-5Z 22/ / /22a-1 22/22b-4 22/23a-2 22/23b-1

213 ~ 212 ~ 22/24a-1 22/24a-2 22/24a-2 22/24a-3 22/24a-3 22/24a-6 22/24b-7 22/24b-8 22/24b-9 22/24d-10 22/25a-1 22/25a-3 22/25a-9Z 22/25ab-2 22/26a-1 22/26a-2 22/27a-1 22/27a-2 22/28a-1 22/28b-2 22/ / /30a-1 22/30a-6 22/30a-7 22/30a-9 22/30a-12 22/30b-4 22/30b-11 22/30b-15Z 22/30c-8 22/30c-10 23/6-1 23/ / / /16a-2 23/16a-4 23/16a-5 23/16a-5Y 23/16d-6 23/16f-11 23/ /21-3A 23/22a-2 23/22a-2Z 23/22a-3 23/22a-3Z 23/22b-4 23/26a-2 23/26a-3 23/26a-5 23/26a-7 23/26a-9 23/26a-10 23/26a-11 23/26a-13 23/26a-16 23/26a-18 23/26a-19 23/26a-19Y 23/26a-19Z 23/26a-21 23/26b-4 23/26b-8 23/26b-14 23/26b-15 23/ / / / / /27-8Z 23/ / /4a-3 28/5-1 28/5a-2 28/5a-4 28/ /1a-7 29/1b-2 29/1b-6 29/1c-4 29/2a-2 29/2a-3 29/2a-4 29/2b-5 29/2c-9 29/2c-11Y 29/3-1 29/3a-2 29/3a-3 29/3a-5 29/3a-5 29/3a-7 29/3b-4 29/4a-1A 29/4a-2 29/4b-3 29/5a-1 29/5a-3 29/5a-5 29/5a-7 29/5b-2 29/5b-4 29/5b-6Z 29/5b-8 29/6a-1 29/6a-3 29/6a-4 29/6b-2

214 ~ 213 ~ 29/7-1 29/7-2 29/7-3 29/7-4 29/7-5 29/7-6 29/7-7 29/8b-2 29/8b-5 29/9a-1 29/9a-5 29/9b-2 29/9b-9 29/9c-4 29/9c-7 29/9c-8 29/ / /11a-1 29/ / /13b-1 29/13b-2 29/14a-4 29/14b-1A 29/14b-2 29/14b-3 29/ / / / /19-1A 29/19a-3 29/ /20b-2 29/ / / / /1c-2A 30/1c-4 30/1c-5A 30/1c-6 30/1c-9 30/1f-8 30/2-1 30/2a-2 30/2a-5 30/2a-7 30/3a-1 30/6-2 30/6-6 30/6-6Z 30/7a-1 30/7a-2 30/7a-3 30/7a-4A 30/7a-5 30/7a-5 30/7a-6 30/7a-7 30/7a-9 30/7a-10 30/7a-11Z 30/7a-12 30/8-1 30/8-2 30/8-3 30/11a-2 30/11b-1 30/11b-3 30/11b-4 30/11b-5 30/ /12a-5 30/12b-2 30/12b-3 30/12b-4 30/12b-6 30/12b-7 30/12b-8 30/ / / / / / / / / / / / / / / / / /17a-4 30/17a-10 30/17a-12 30/17b-2 30/17b-5 30/17b-9 30/17b-16 30/ / / /19a-4 30/19a-5 30/19a-7Y 30/21-1

215 30/22b-1 30/ / /26b-18 30/23-2A 30/25a-4 31/ /26c-13 30/23a-3 30/25b-3 31/ / /23b-4 30/ /26a-5 30/ / /26a-6 DK 30/ /29a-1 31/26a-9A 30/ /29a-3 31/26a-10 A / / /26a-12 30/ / /26b-17 Chapter 3: Central North Sea contourites and scours NO UK 22/28a-1 23/ /30a-9 29/3a-3 7/8-3 22/25a-3 22/30c-13 30/7a-1 22/25b-5 23/26a-2 Chapter 4: Evolving Pleistocene palaeo-environments of the central North Sea NO 2/7-3 7/4-2 15/9-15 2/7-4 7/8-1 15/12-1 1/2-2 2/7-31 7/9-1 15/12-4 1/3-8 2/8-1 7/ /4-1 1/6-4 2/9-1 7/ /7-2 1/9-4 2/ /3-4 16/7-3 2/1-3 3/7-3 15/5-3 16/7-9 2/5-1 6/3-1 15/6-4 16/10-2 2/5-7 7/1-1 15/8-1 ~ 214 ~

216 UK 16/ /24b-7 30/11b-1 16/29a-12 22/25a-3 30/13-3 Josephine-1 20/5c-4 22/ /17a-12 14/ /6-1 23/16a-2 30/17a-13 14/24a-4 20/ /16a-5 30/17b-2 14/28b-4 20/ /22a-2 30/17b-9 15/6-1 21/3-2 23/26a-5 30/ / /7-1 23/26a-9 30/ / / /5a-2 30/ / / / /19a-4 15/ / /2a-3 30/ /23a-12 21/20a-5 29/2a-5 30/25b-5 15/ /20b-3 29/5b-4 30/ /7a-16 21/ /9a-1 31/26a-12 16/7a-6 21/ /7-1 31/26a-9A 16/ /1a-5 29/14b-3 16/21a-13 22/8a-3 30/1c-2A DK 16/ /12a-1 30/1c-6 16/24a-1 22/ /7a-9 A / / / /26-21Z 22/ /8-1 A1.4: Culture Data Culture data, specifically shapefiles and rasters used to produce maps, were generally downloaded from other sources. The following list identifies the data, figures that used the data, the source website and, where available, the full reference for the culture data. Topography & bathymetry of North Sea and NW Europe ~ 215 ~

217 o Figures: 1.1, 2.1, 2.12, 3.1, 4.1, 4.10 & 5.1 o Website: o Reference: Ryan, W.B.F., S.M. Carbotte, J.O. Coplan, S. O'Hara, A. Melkonian, R. Arko, R.A. Weissel, V. Ferrini, A. Goodwillie, F. Nitsche, J. Bonczkowski, and R. Zemsky (2009), Global Multi-Resolution Topography synthesis, Geochem. Geophys. Geosyst., 10, Q03014, Countries outline o Figures: 2.4, 2.5, 2.9, 2.10, 3.12 o Website: o Reference: UKCS Oil and gas offshore maps and GIS shapefiles North Sea Sector Boundaries o Figures: 1.1, 2.1, 3.1, 3.2, 4.1, 5.1 o Website: o Reference: UKCS Oil and gas offshore maps and GIS shapefiles North Sea Blocks o Figures: 1.1 o Website: o Reference: UKCS Oil and gas offshore maps and GIS shapefiles Core & Well locations o Figures: 1.1, 2.1, 2.4, 2.5, 3.2, 3.3, 4.1, 4.3 o From well header & core summary information Extents of past studies o Figures: 2.1, 2.2 o Digitised from georeferenced images from referenced studies ~ 216 ~

218 Appendix 2: 2 nd Author Manuscripts ~ 217 ~

219 Evidence for repeated mid latitude marine-terminating ice sheets in a 41 kyr early Quaternary world Andrew M. W. Newton 1, Rachel Harding 1, Rachel M. Lamb 1, Mads Huuse 1, and Simon H. Brocklehurst 1 Iceberg scours are a common feature on high-latitude margins where glaciers have, or once had, a marine-terminating margin. Scours are formed when the iceberg draft exceeds water depth and it ploughs a furrow in the seabed as it is moved by winds and ocean currents. Scours and ice-rafted detritus preserved in the stratigraphic record provide insights into ice sheet geometry 1 and location 2, how and when ice sheets grew and collapsed 3, and their freshwater influence on ocean circulation 4-7. However, the mapping of iceberg scours has only been performed on local 7,8, rather than basin-wide scales. Here we show the results from the first basin-wide mapping of iceberg scours in the early Quaternary record of the North Sea, based on 3D seismic reflection data. Over 8000 iceberg scours have been mapped across 36 stratigraphic horizons in the interval Ma. This period was dominated by lowamplitude 41 kyr climate cycles 9, which were progressively replaced with high-amplitude 100 kyr cycles after ~1.2 Ma 10. It was previously thought that European ice cover was restricted to small ice caps over Fennoscandia prior to 1.1 Ma 11, as a 41 kyr glacial cycle was not considered long enough for ice to advance to the continental shelf. Our results prove the presence and chronology of marine-terminating ice sheets in the early Quaternary, and show that glaciations in southern Norway reached the coast within most of the 41 kyr glacial cycles of the early Quaternary. Age control from core data 12,13 shows a minimum of 11 events where Fennoscandian ice reached the marine setting. These results raise important questions regarding assumptions and models used to simulate the evolution of ice sheets under external forcings and in particular about the efficacy of a 41 kyr glacial cycle to form large ice masses. Understanding long-term changes in ice dynamics is critical for models of sea level change 14 and ocean-climate evolution 15,16. Developing a clearer understanding of historical changes means that ~ 218 ~

220 we can better evaluate climate models used for projecting future environmental changes 17. The North Sea sedimentary succession (Fig. 1) comprises an up to 1.5 km thick archive of glacial to interglacial fluctuations through the Quaternary. The early Quaternary succession consists of glacimarine and marine-deltaic deposits 18 that are strongly related to low-amplitude 41 kyr climate cycles 12,13. After ~1.2 Ma the 41 kyr cycles were progressively replaced with high-amplitude 100 kyr cycles 10 during the mid-pleistocene transition (MPT) which eventually resulted in a switch from a marine-deltaic, to a glacial setting. By Marine Isotope Stage (MIS) 12 ( ka) the North Sea was covered by grounded ice sheets during glacials 18. The sediments deposited during the Quaternary vary from deep marine contourites to glacial tills and provide a record of changing glaciation intensity and sediment delivery patterns from the UK, Norway, and mainland Europe The record of δ 18 O fluctuations 9 (Fig. 2) is a proxy for changes in sea level 14 that are related to global glacial-interglacial cycles. Ice-rafted detritus (IRD) records from offshore cores can be used to infer the presence of marine-terminating ice margins. Low amounts of IRD recorded from the Vøring Plateau 21 off the mid-norwegian shelf suggest that northern Norway was first covered by ice ~2.8 Ma and that ice was restricted to high latitudes and altitudes during the early Quaternary. After the MPT 22 a marked increase in IRD after ~1.1 Ma 23 is attributed to the expansion of ice onto the continental shelf in the west, and into the North Sea in the south. Most work investigating the glacial geomorphology of the North Sea 18 has concentrated on ice sheets from MIS 12 and younger. Limited work has investigated older glaciations and recently it was shown that iceberg scours occur intermittently within the early Quaternary 7 in the central North Sea, although the geographic extent, chronology, and frequency of scour was uncalibrated. In order to evaluate the frequency and intensity of marine-terminating ice sheets reaching the North Sea, we mapped every well-defined reflection in a basin-scale 3D seismic dataset that covers the entire Quaternary depocentre (Fig. 3a). We use this framework to define six seismic units that start at the Plio-Pleistocene boundary and cover the early Quaternary (Fig. 2). Dating from nearby ~ 219 ~

221 drilling sites is tied to the seismic data 13, and using different seismic volume attributes (Fig. 3b) over 8000 linear and curvilinear iceberg scours between 2.58 and 1.7 Ma were mapped. We find no evidence for Late Pliocene iceberg scouring in the North Sea but numerous events through the early Quaternary. This record of seismic geomorphology shows that Quaternary glaciations were much more extensive and frequent than previously thought 24. The shape of the North Sea basin at the beginning of the Quaternary is an elongate, enclosed basin (Fig. 1a) compared to the present (Fig. 1b) 25. The evolution of the basin shape by progressive deltaic infill is an important factor in the distribution of iceberg scours as there is very limited evidence for scours beyond the palaeo-shelf break of individual clinoforms. The majority of the mapped scours are at the palaeo-shelf edge or landward of the palaeo-shelf break. In the oldest seismic unit (SU1), Ma (Fig. 4a), we found six iceberg-scoured horizons (Fig. 2). Using the relative position of each horizon within the seismic unit, seismic architecture, the global sea level curve 14, and records of ice-rafted detritus from the mid-norwegian margin 11,21 we are able to correlate scoured horizons to specific glacials between dated surfaces calibrated with borehole data 12,13 (see methods). The first scoured horizon shows iceberg scours in the southern part of the survey area and is attributed to MIS 100. Although the position of the scoured horizon is toward the base of the seismic unit and suggests a correlation with MIS 100, there is uncertainty over its exact age and it may also be attributed to MIS 98. These scours are close to the base of the unit, and indicate a glacial influence in the North Sea very soon after the beginning of the Quaternary (MIS 100, c Ma) whilst analysis of earlier clinoforms show no evidence for scour during the Pliocene. Scours in this earliest horizon in the northern area are distributed either side of the central trough in the basin. The eastern and western scours show NE-SW and NW-SE trajectories respectively. The remaining horizons in the SU1 package and the scoured horizons in the SU2 ( Ma) package (Fig. 4b) correlate with each glacial event between MIS 98 and 92 (Fig. 2). ~ 220 ~

222 The SU3 (2.35 ~2.15 Ma) package (Fig. 4c) shows what appears to be a hiatus in iceberg scouring in the North Sea. The amplitude of sea level change in this period is muted; the glacial lowstand is m higher than those before and after. The lack of scours may reflect greater water depths, such that any scouring would have taken place in shallower waters outside of the study area. Therefore, any icebergs scours that were formed may not have been preserved due to sediment bypass and reworking of the shelf during later transgressions. Alternatively, the higher sea level may simply reflect, among other global ice sheets, a smaller southern Fennoscandian ice sheet and a reduction of its marine-terminating margins. Despite the lack of geomorphological evidence for iceberg presence, IRD data on the Vøring Plateau remain steady during this time period 21 and suggest a persistent northern Fennoscandian ice sheet with marine-terminating margins 11. Necessitated by the lack of datable material, the dating of the IRD record 11,21 assumes constant sedimentation rates between magnetostratigraphically dated intervals. However, environmental transitions between glacial, interglacial and deglacial stages would mean such constant sedimentation rates are unlikely. IRD deposition also requires that any sub- or englacial material not be melted out before the iceberg gets to the borehole site on the Vøring Plateau. This IRD data from the Vøring Plateau requires careful consideration when compared to the North Sea early Quaternary history because of its distal location. However, given the lack of more proximal data, a comparison is, nonetheless, informative. The SU4 (~ Ma; Fig. 4d) package contains the two largest glacial lowstands of the study period (MIS 82 and 78). The date for the ~2.15 Ma boundary is not as well-constrained as the others but three reflections at the base of the package appear to coincide with MIS 80, although this may indeed be MIS 82 due to the small magnitude of MIS 80 and dating uncertainty at ~2.15 Ma. Three further seismic reflections across the entire survey area are correlated to MIS 78. Several reflections across the basin are correlated to the MIS 76 and 74 glacials. One of the surfaces in the central area is attributed to MIS 76 but with the inferred errors it may belong to the later glacial. The southern scouring event is, however, still suggested to have occurred during MIS 76 as the southern basin became progressively shallower. During the time period covered by the SU4 ~ 221 ~

223 package, scours become more common in the northern area of the survey and less common in the southeast as the basin becomes increasingly terrestrial during glacial lowstands (Fig. 4). In SU5 ( Ma; Fig. 4e) and above (Fig. 4f) there are numerous scouring events across all areas of the survey. Over 2000 scours are present across several surfaces and have been attributed to all glacial stages from MIS with the exception of the smaller glacial peak during MIS 68. No scours are present in the south during the latest part of the package as the basin had infilled, and subsequently the area had become fully terrestrial during glacial lowstands 13. In the upper-most part of SU5 and lower SU6 there are multiple scoured horizons (Fig. 3a) that appear to correlate to MIS 64. The surfaces are separated by the dated surface at 1.78 Ma which is halfway through the MIS 64 deglacial. Each seismic reflection shows evidence for a different iceberg scour event and there is no surface without iceberg scour separating the scoured horizons that might suggest an interglacial. It is plausible that these different scoured horizons represent several parts of the MIS 64 deglacial. If the scoured events are from one glacial cycle, the frequency of scouring suggests that sedimentation rates were sufficiently high to allow different scouring events to plough younger sediments and not destroy evidence of older scouring. The scouring events may represent different phases of the glaciation/deglaciation and may indicate that retreat was periodic rather than a dramatic collapse. However, given that the other scoured surfaces are often separated by surfaces without iceberg scour it is unusual by comparison for so many scoured events to be preserved for one glacial cycle. The constraints imposed from vertical seismic resolution (>10 m) might mean that any interglacial sediments are not reflected in the seismic and these scours may be from more than one glacial-interglacial cycle. Compaction of the scoured horizons from burial throughout the time period may also mean that the younger sections of the stratigraphy are more expanded than the older sections and this would make them more easily observable in seismic records. As such, it is difficult to know if the youngest scoured horizons in SU6 can be attributed to MIS 64. The frequency with which icebergs have scoured palaeo-seafloors in the early Quaternary North Sea has a number of implications. The most obvious implication is that these scours represent ~ 222 ~

224 several different events where icebergs drifted around the basin. early Quaternary margin progradation shows a dominant sediment source from Fennoscandia and northern Europe 19 rather than from Great Britain. Sediments in the northern area are of glacigenic origin 19 and suggest that the parent ice sheet of the icebergs was likely over Fennoscandia. Such a system would require that there was at least seasonally open water access in the North Sea 7 and that ice extended beyond the coastline onto the shelf. Whilst the early Quaternary history of the Nordic Seas is complicated by carbonate dissolution indicating cooler conditions and significant sea ice cover, northward currents were still recorded in the Norwegian Sea 26. The oceanography of the North Atlantic in the early Quaternary also shows northward-moving currents during glacials despite reduced benthic δ 13 C levels indicating reduced North Atlantic Deep Water production; the same pattern recorded for the late Quaternary 23. Even in the late Quaternary when meltwater pulses during Heinrich events heavily disrupted deep water production, the current flow in the Nordic Seas was still northwards along the Norwegian coast 6,27. Therefore it seems reasonable to suggest that the northward currents during glacial times and sea ice cover in the early Quaternary would prevent icebergs from floating into the North Sea from other glacial margins such as Greenland, Svalbard or northern Norway. At the beginning of the Quaternary, shallowing of the basin created a sill between the central and northernmost North Sea at the northeastern end of the central trough in the North Sea basin 25. With the sill at the entrance to the North Sea and the palaeocurrent information described above, it is unlikely that the icebergs were from anywhere other than the surrounding coastal regions. The majority of the scours show similar NE-SW trajectories, suggesting they may have been calved from Norway (Fig. 4). However, in the first surface investigated here a number of iceberg scours on the western side of the trough have NW-SE trajectories, suggesting they might have come from a British ice sheet. Alternatively, the icebergs may have been calved from Norway and then been carried across the deeper part of the basin and scoured as they moved to the NW onto the shelf. If these icebergs were indeed calved from a British ice sheet it would present additional geomorphological evidence to the IRD data offshore Ireland that suggests a British-Irish ice sheet since the earliest Quaternary 28. ~ 223 ~

225 It was previously suggested that ice coverage was restricted to the more mountainous regions of Norway in the early Quaternary 23 as ice masses could not grow large enough to reach the coast in a 41 kyr glacial cycle. However, the evidence presented here suggests that marine-terminating ice was much more extensive and more common than previously thought. Pre-existing regolith may have allowed the first ice sheet to thin and stretch towards the coast on a 41 kyr cycle 22. Interglacial periods were also cooler in the early Quaternary 29 and mountains were possibly higher and colder. This scenario would have allowed the ice cover to become a more permanent feature in Norway as it failed to fully melt away during cooler interglacials. Cooler interglacials would also require less accumulation during the next glacial for the ice to reach the sea. This is supported by consistent observations of IRD on the Vøring Plateau which suggests persistent glacial conditions and ice cap stability over Norway in the early Quaternary 11. Regardless of approximate correlations with different MIS stages the dating chronology presented here explicitly indicates a minimum of 11 events where ice sheets over southern Norway could grow large enough to develop marine-terminating margins. This provides compelling evidence for high-frequency oscillations of early Quaternary marine-terminating ice sheets in NW Europe. There remains extreme uncertainty as to the full chronology and dynamics of the earliest glaciations in northwest Europe. Methods The 3D seismic volumes include the PGS Southern North Sea MegaSurvey and the Central North Sea MegaSurvey covering ~128,000 km 2. Eliis Paleoscan software was used to auto-track all resolved stratal surfaces for the early Quaternary in a semi-automated model-constrained fashion across the 3D seismic volumes. Surfaces were imported into the Petrel seismic interpretation software for analysis. Amplitude, root-mean squared amplitude and variance volumes were extracted onto the auto-interpreted surfaces to identify all occurrences of iceberg scours in the data. Additional surfaces were picked so that all reflections with iceberg scours were represented. ~ 224 ~

226 Volume attribute extractions were exported as georeferenced images, and added to a GIS. Using ArcMap software all of the scours from each surface were digitised for subsequent analysis. Dates from nearby drill sites were tied to the seismic 23 (in prep, R.H., and M.H.) and scoured surfaces were relatively dated by using their position between six absolutely dated seismic horizons. To develop a more absolute date for each scoured surface a number of other techniques were implemented. Using the classic model of seismic stratigraphy 30, work has been carried out to elucidate the relationship between global sea level changes and seismic architecture in the basin (in prep, R.H., and M.H.). This allowed the surfaces to be more confidently attributed to specific glacial stages depending on their stratigraphic location and its position within regressive and transgressive cycles. In addition to this correlation, the surfaces and the abundance of scouring was qualitatively compared to records of IRD 11,21 to see which glacials had higher levels of IRD and were thus more likely to have a marine-terminating margin. Necessitated by uncertainties in biostratigraphical and magnetostratigraphical dating, and incomplete records for seismic sequence stratigraphy used to position the scoured surfaces within glacial-interglacial cycles we developed ranges where the dates were less certain. These ranges included our best estimate for dating the scoured surfaces to specific glacials based on the methods above, but also the range in which the combination of dating methods allowed us to place them in. This showed that whilst we were able to use the different methods to attribute scouring events to specific glacials, the resolution of data meant that scouring events may have been associated with a previous or subsequent glaciation. The surfaces and their spatiotemporal relationships to margin development and environmental changes through different glacial-interglacial cycles were then compared to the global sea level curve and IRD data to elucidate the early Quaternary glacial history, as described in the text. References 1 Dowdeswell, J. A. & Bamber, J. L. Keel depths of modern Antarctic icebergs and implications for sea-floor scouring in the geological record. Marine Geology 243, (2007). ~ 225 ~

227 2 De Schepper, S., Gibbard, P. L., Salzmann, U. & Ehlers, J. A global synthesis of the marine and terrestrial evidence for glaciation during the Pliocene Epoch. Earth-Science Reviews 135, (2014). 3 Jansen, E. & Sjøholm, J. Reconstruction of glaciations over the past 6 Myr from ice born deposits in the Norwegian Sea. Nature 349, (1991). 4 Kleiven, H. F., Jansen, E., Fronval, T. & Smith, T. M. Intensification of Northern Hemisphere glaciations in the circum Atlantic region ( Ma) ice-rafted detritus evidence. Palaeogeography, Palaeoclimatology, Palaeoecology 184, (2002). 5 Hill, J. C. & Condron, A. Subtropical iceberg scours and meltwater routing in the deglacial western North Atlantic. Nature Geoscience 7, (2014). 6 Bigg, G. R. et al. Sensitivity of the North Atlanctic circulation to break-up of the marine sectors of the NW European ice sheets during the last Glacial: A synthesis of modelling and palaeoceanography. Global and Planetary Change 98-99, (2012). 7 Dowdeswell, J. A. & Ottesen, D. Buried iceberg ploughmarks in the early Quaternary sediments of the central North Sea: A two-million year record of glacial influence from 3D seismic data. Marine Geology 344, 1-9 (2013). 8 Todd, B. J., Lewis, C. F. M. & Ryall, P. J. C. Comparison of trends of iceberg scour marks with iceberg trajectories and evidence of paleocurrent trends on Saglek Bank, northern Labrador Shelf. Canadian Journal of Earth Science 25, (1988). 9 Lisecki, L. E. & Raymo, M. E. A Pliocene-Pleistocene stack of 57 globally distributed benthic δ 18 O records. Palaeoceanography 20 (2005). 10 Head, M. J. & Gibbard, P. L. in Early-Middle Pleistocene Transitions: The Land-Ocean Evidence Special Publications (eds M. J. Head & P. L. Gibbard) 1-18 (Geological Society, 2005). 11 Henrich, R. & Baumann, K. H. Evolution of the Norwegian current and the Scandinavian ice sheet during the past 2.6 my: evidence from ODP Leg 104 biogenic carbonate and terrigenous records. Palaeoceanography, Palaeoclimatology, Palaeoecology 108, (1994). ~ 226 ~

228 12 Noorbergen, L. J., Lourens, L. J., Munsterman, D. K. & Verreussel, R. M. C. H. Stable isotope stratigraphy of the early Quaternary of borehole Noordwijk, southern North Sea. Quaternary International, 1-10 (2015). 13 Kuhlmann, G. et al. Chronostratigraphy of Late Neogene sediments in the southern North Sea Basin and paleoenvironmental interpretations. Palaeogeography, Palaeoclimatology, Palaeoecology 239, (2006). 14 Miller, K. G., Mountain, G. S., Wright, J. D. & Browning, J. V. A 180-million-year record of sea level and ice volume variations from continental margin and deep-sea isotopic records. Oceanography 24, 40-53, doi: /oceanog (2011). 15 Rahmstorf, S. Ocean circulation and climate during the past 120,000 years. Nature 419, (2002). 16 Lambeck, K., Esat, T. M. & Potter, E.-K. Links between climate and sea levels for the past three million years. Nature 419, (2002). 17 Braconnot, P. et al. Evaluation of climate models using palaeoclimatic data. Nature Climate Change 2, (2012). 18 Graham, A. G. C., Stoker, M. S., Lonergan, L., Bradwell, T. & Stewart, M. A. in Quaternary Glaciations - Extent and Chronology (eds J. Ehlers & P. L. Gibbard) (2011). 19 Ottesen, D., Dowdeswell, J. A. & Bugge, T. Morphology, sedimentary infill and depositional environments of the early Quaternary North Sea Basin (56 o 62 o N). Marine and Petroleum Geology 56, (2014). 20 Bradwell, T. et al. The northern sector of the last British Ice Sheet: maximum extent and demise. Earth-Science Reviews 88, (2008). 21 Krissek, L. A. in Proc. ODP. Sci. Results 104 (eds O. Eldholm, J. Thiede, E. Taylor, & et al.) (ODP, 1989). 22 Clark, P. U. & Pollard, D. Origin of the middle Pleistocene transition by ice sheet erosion of regolith Palaeoceanography 13, 1-9 (1998). ~ 227 ~

229 23 Jansen, E., Fronval, T., Rack, F. & Channell, J. E. T. Pliocene-Pleistocene ice rafting history and cyclicity in the Nordic Seas during the last 3.5 Myr. Palaeoceanography 15, (2000). 24 Ehlers, J., Gibbard, P. L. & Hughes, P. D. in Quaternary Glaciations - Extent and Chronology: A Closer Look Vol. 15 Developments in Quaternary Science (eds J. Ehlers, P. L. Gibbard, & P. D. Hughes) 1-14 (2011). 25 Lamb, R. M., Huuse, M., Harding, R., Stewart, M. A. & Brocklehurst, S. H. The Early Quaternary North Sea Basin. Quaternary Science Reviews (In review). 26 Baumann, K. H., Meggers, H. & Henrich, R. in Proc. ODP, Sci. Results Vol. 151 (eds J. Thiede et al.) (1996). 27 Levine, R. C. & Bigg, G. R. Sensitivity of the glacial ocean to Heinrich events from different iceberg sources, as modeled by a coupled atmosphere-iceberg-ocean model. Palaeoceanography 23, doi: /2008pa (2008). 28 Thierens, M. et al. Ice-rafting from the British-Irish ice sheet since the earliest Pleistocene (2.6 million years ago): implications for long-term mid-latitudinal ice-sheet growth in the North Atlantic region. Quaternary Science Reviews 44, (2012). 29 Raymo, M. E., Hodell, D. & Jansen, E. Response of deep ocean circulation to initiation of Northern Hemisphere glaciation (3-2 Ma). Palaeoceanography 7, (1992). 30 Vail, P. R., Mitchum Jr, R. M. & Thompson III, S. in Seismic Stratigraphy Applications to Hydrocarbon Exploration Vol. 26 Memoir (ed C. E. Payton) (American Association of Petroleum Geologists, 1977). Acknowledgments We thank Cairn Energy and The Natural Environment Research Council, BUFI, and TNO for funding A.M.W.N., R.M.L. and R.H. respectively. Schlumberger, Eliis, and ESRI are thanked for providing software for the analysis of seismic data. PGS, and TGS are thanked for releasing the seismic and well data used in this research. ~ 228 ~

230 Author Contributions A.M.W.N. prepared the data for digitising, mapped the iceberg scours, wrote the draft manuscript and prepared the figures. R.H. and R.M.L. provided the reflection surfaces for analysis and correlated dates to seismic surfaces. All authors contributed to the interpretation of the data and writing of the paper. ~ 229 ~

231 Figure 1: Map of the North Sea and data coverage. a, large red polygon indicates the area of 3D seismic coverage used for this study. Smaller polygons represent areas with no 3D data coverage inside the study area. Depths represent the distance between the contemporary sea floor and the base Quaternary reflector. Compaction of sediment has not been accounted for here, but the data nevertheless show the general, substantially smaller shape of the basin at 2.58 Ma. Backstripping of the base-quaternary surface suggests maximum water depths of ~300m 25. Drill sites used for dating and seismic correlation are also indicated. The white and green circles refer to the A15-03 and Josephine-1 core sites respectively. Contours are every 200 m. The north, central and southern regions referred to in Fig. 2 are shown and separated by the white stippled lines. The yellow line is the location of the seismic line in Fig. 3. b, inset map. GEBCO bathymetry for the North Sea, ~ 230 ~

232 showing the contemporary environment, using the same colour scale as (a). Red circles indicate the locations of ODP boreholes on the Vøring Plateau. ~ 231 ~

233 Figure 2: Late Cenozoic datasets. The sea level record 14 calculated from benthic δ 18 O records 9 and the ice-rafted detritus record from ODP site 642 on the Vøring Plateau 21 (Fig. 1b) for the late Pliocene and early Quaternary (Fig. 1b). IRD abundance is dated using a constant sedimentation rate between magnetically dated intervals. Using the six dated seismic units as a chronological constraint, the scoured horizons are dated by using their position within the seismic unit and by correlations with the sea level record, IRD, and seismic stratigraphy (see methods). The black circle indicates the best estimate for the age of the scoured surface and an error range is estimated from the combination of techniques described in the methods section. See Fig. 1 for the location of the north (N), central (C) and southern (S) areas used to describe iceberg scour locations. ~ 232 ~

234 Figure 3: Clinoform geometry. a, lines show the semi-automatically picked reflections from the seismic data. Blue lines show the seismic horizons that are used as a dating control. The red horizons are those with iceberg scour features on and the white horizons are those whose surface does not have glacial landforms on. These reflections were then used with different seismic volume attributes to investigate the resultant 3D surface for glacial geomorphological features. The yellow stippled line shows the stratigraphic location of the surface displayed in (b). Seismic line location is shown in Fig. 1a. The clinoform geometry highlights how time slices (horizontal planes of equal two-way-travel time) cut across different time periods represented by the clinoform stratigraphy. b, surface created from the 3D seismic with the root-mean squared amplitude attribute displayed. White arrows show examples of a curvilinear iceberg scour that were visible on the surface and digitised. ~ 233 ~

235 Figure 4: Iceberg scour distribution. Distribution of iceberg scours during deposition of each seismic unit. Location of the shelf edge in each figure is indicated for the start (red line) and the end (blue line) of each time period. This shows how the basin shape changed with migration of the shelf edge to the west and reduction in the size of the central trough. Note that in the north the blue line covers the red line. This is due to the minimal change across the full time period for the northern-most part of the study area. The southwestern basin shape is not displayed because there is no depositional system from which to pick a shelf edge and seismic resolution and postdepositional movement make interpretation difficult. Rose diagrams are plotted for iceberg scour trajectory for all the scours in each time period, showing a predominantly north to south trajectory. Stippled lines show the shelf edge position outside the 3D coverage from 2D seismic profiles. a. deposition of the SU1 unit from Ma. b. SU2 unit from Ma. c. SU3 unit from 2.35 ~2.15 Ma. d. SU4 unit from ~ Ma. e. SU5 unit from Ma. f. SU6 unit after 1.78 Ma. ~ 234 ~