Figure 32 – uploaded by Vincent Gaffney

See full PDF downloadDownload figure

Figure 4.10 GIS Layers display within a fully 3 Dimensional environment - A cube of seismic data (displayed as a voxel volume) and a 3D surface & 3D GIS polygons (interpretation layer) — Figure 4 10 GIS Layers display within a fully 3 Dimensional environment - A cube of seismic data (displayed as a voxel volume) and a 3D surface & 3D GIS polygons (interpretation layer)

Related Figures (90)

Vincent Gaffney, Kenneth Thomson and Simon Fitch

Figure 1.1 Hypothetical maximum extent of Doggerland (redrawn from Coles 1998)

Figure 1.2 Early Holocene Doggerland (redrawn from Coles 1998,

Figure 1.3 Holocene shorelines (after Jelgersma, 1979) capable of providing detailed images or Late Pleistocene and Holocene features that have a bathymetric expression. Whilst bathymetry pro- vides a reasonable approximation for the land sur- face for the area it can rarely consider, or attempt to resolve, burial of features that may have oc- curred during or after submersion (Cameron et al., 1992). Consequently, the technique is unsuit- able for areas including the southern North Sea and the Irish Sea, where deposition has buried most of the Quaternary and Holocene. The scale of this problem was clearly stated by Dix et al. (2004, 89); “although modern bathymetry can

(this volume) that demonstrate the detail of specific geo- morphological structures, including the nature of internal features identified within the Outer Silver Pit. A full description of the technologies utilised to explore and integrate the available seismic data is provided in the papers by Thomson and Gaffney and Fitch et al. (the atlas, this volume). It is enough to note here that the signifi- cance of the first results of this work was rapidly appreci- ated, and that mapping of the area proceeded apace, as is demonstrated in the atlas paper by Fitch et al. This is com- plemented by papers from Briggs et al. and Holford et al. It is true, however, that as the project proceeded there was an increasing concern within the team concerning not so much the extent of available supporting data (including 2D seismics lines, cores etc) but the quality, or even availabil-

Figure 2.1 BGS survey coverage on the UK continental shelf as displayed by GeoIndex, October 2006. (a) shows geophysical profile lines, (b) shows shallow sediment sample stations A variety of geophysical survey techniques have been util- ised as part of this process. In order to map the shape of the sea floor all areas have been explored using an echo sounder and some by more sophisticated sonar techniques. The majority have been surveyed using a potential field instrument: a gravimeter or magnetometer. These give an indication of a variation in bulk density and magnetisation of seabed material laterally. However, the most useful technique for visualising the vertical stratigraphy of sea- bed sediments is sub-bottom seismic profiling. Most of the BGS geophysical survey lines have been covered by a

Figure 2.2 Distribution of 3D seismic surveys acquired by the UKOOA

Figure 2.3 Three study areas around the English coast The Mapping European Seabed Habitats (MESH — http://www.searchmesh.net/default.aspx) is an interna- tional marine habitat-mapping programme involving a consortium of twelve partners across the UK, Ireland, the Netherlands, Belgium and France. The aim is to produce standardised seabed habitat maps for northwest Europe. The MESH website hosts a basic GIS. A much larger metadata archive of all forms of marine surveying within specified regions can also be queried elsewhere on the website.

Figure 2.4 Areal a) Industrial threats to seabed archaeology within the Bristol Channel b) Distribution of previous BGS sampling and 2D surveys c) UKOOA surveys d) Archaeology and previous ALSF survey conducted by the BGS within Area I. Somerset coast in August 1869 (NMR Number SS 94 NE 11, Record 36754). Mesolithic blades and flakes, and Neolithic flints and scrapers, just offshore from Minehead on the Northwest

Figure 2.7 Intense use of space within the Southern North Sea

Figure 3.2 Seismic resolution of a layer of varying thickness. The blue lines show the top and base of a layer. Using « 400Hz Ricker wavelet, and assuming a sediment velocity of 1600ms", the reflections from the top and base can be show to constructively interfere at thicknesses less than YA (1 metre) with the maximum amplitude at 44. Furthermore, the toy and the base of the layer do not perfectly align with peak and troughs. At thicknesses of /2A (2 metres) and greater the to] and the base of the layer are separable with peak and troughs aligning perfectly with the position of the top and bas recsnertiyely As the vertical resolution is dependent upon the wavelength it is therefore dependent on the velocity of the medium and the frequency of the seismic source/reflected wave. Ideally, a high frequency source (>100Hz) would be used in all circumstances. However, as the geology progressively dampens high frequency seismic signals with depth, the seismic source needs to be chosen with consideration to the required depth of penetration. The

Figure 3.1 Typical marine seismic reflection acquisition. The vessel travels through the water and regularly fires seismic source. The sound wave travels through the water column (and underlying sediments) and is partially reflectec acoustic impendence contrasts. Receivers within the long towed cable astern of the vessel detect the reflected wave, wh is then transmitted to the vessel and recorded.

Figure 3.3 Plots of seismic resolution as a function of burial depth and frequency. (a) The minimum resolvable vertical resolution. (b) The minimum separable vertical resolution. (c) The horizontal resolution of unmigrated seismic data. (d) The assumed velocity-depth structure.

Figure 3.4 A comparison between (a) high frequency 2D seismic reflection line and (b) low frequency 3D seismic line from the same location. Note that higher frequencies yield greater vertical detail.

Figure 3.5 (a-d) Four possible interpretations of a channel morphology based on a coarse 2D seismic grid. Each interpretation is equally valid. (e-h) Schematic illustrations of how each of the interpretations shown in a-d would appear on a timeslice from a laterally continuous, binned 3D seismic volume. This demonstrates that 3D seismic data has the potential to distinguish between the possible alternatives. This method of acquisition has two main disadvantages. Firstly, the reflected seismic energy is assumed to have originated from a point directly beneath the profile even though it could have originated from a point laterally offset from the profile. This aliasing means that the location of a feature cannot be accurately constrained, as the spacing be- tween lines is too wide to correct this error. Secondly, the spacing between lines is sufficiently wide that it can be diffi- cult to map the position of a morphological feature across the region of interest. For example, Figure 3.5 demonstrates how wide line spacing can lead to several equally valid in- terpretations. Traditional seismic reflection data is generally referred to as 2D as it is acquired as a series of discrete vertical profiles using a single streamer towed behind the vessel. This acquisition pattern results in the collection of sev- eral profiles with the spacing between profiles being several orders of magnitude greater than the trace spac- ing (i.e. the horizontal sampling interval along the pro- file).

Figure 3.6 Typical 3D marine seismic reflection acquisition. The vessel travels through the water and regularly fires the seismic source. The sound wave travels through the water column (and underlying sediments) and is partially reflected at acoustic impendence contrasts. Receivers within the towed cables astern of the vessel detect the reflected wave, which is then transmitted to the vessel and recorded. In contrast to Figure 3.1, 3D acquisition involved multiple towed cables with each cable being capable of producing a seismic profile.

Figure 3.7 (a) Geological map of Poole and Christchurch bays with the artificially illuminated plan view of the seabed reflector. (b) Artificially illuminated plan view of the seabed reflector. (c) Artificially illuminated plan view of the seabed reflector with palaeo-channels as mapped using 3D seismic. (d) Artificially illuminated plan view of the seabed reflector with palaeo-channels as mapped using 2D seismic data (Velegrakis et al. 1999). Note the poor correlation between the 2D interpretation and the actual channel locations.

Figure 3.8 A comparison of seismic images from the Dogger Bank produced using different techniques. (a) Simple seismic amplitude timeslice revealing some Holocene depositional features (channels). (b) RMS amplitude extraction for the top 200ms of the data. (c) Opacity rendering of the same seismic volume as (b). Note that both the RMS extraction and the manipulation of the opacity allow some improvement in channel definition. (d) Seismic coherence (semblance) timeslice at the same level as (a). Note that coherence provides some additional detail of the channels.

The primary goal of the North Sea Palaeolandscapes Pro- ject (NSPP) was to explore the Late Pleistocene and Holo- cene landscapes of the Southern North Sea through the use of c. 23,000 km° of contiguous 3D seismic data provided by PGS UK (www.PGS.com). Processing this large vol- ume of data required considerable investment in both hardware and software. In line with most archaeological projects utilising large amounts of digital, remote sensed data, there was a requirement to utilise specialist softwares for the processing of primary seismic datasets and standard geographic information systems to manage, manipulate and display supporting data and interpretative layers (Gaffney and Gater, 2003 Chapter 5; Chapman 2006). What is, perhaps, less usual within archaeology has been the requirement to provide access to high bandwidth net- works and storage to cope with the large volumetric data- Figure 4.1 HP VISTA infrastructure diagram

Figure 4.2 Inspecting data at the Visual and Spatial Technology Centre (HP VISTA) The primary engines for the majority of analysis were three HP xw8200 workstations with 64bit dual processors, 16Gb of RAM, 2Tb of local storage and high-end graphics cards. Whilst the majority of primary processing was car- ried out on these machines, specialist softwares used to carry out attribute analysis on the seismic data, for exam- ple, were resident on machines elsewhere on the campus. Project data was held centrally on a secure 5Tb NAS stor- age system resident within the University Information Services’ machine room. The need to move data between these machines and to the central storage was supported by gigabit-enabled cabling between most project computers, and assisted by the availability of a dedicated fibre optic link between the NAS and the main display centre in the HP VISTA centre. The extent and complexity of the volumetric data, and the utilisation of solid modelling and opacity rendering techniques, required that the data should be viewed stereoscopically and also that displays were large enough to provide the opportunity of group viewing. This was provided within the HP VISTA centre via an HP SV7 scaleable visualisation system. This is a multi-pipe system providing high-end, high quality visualisation. A particular benefit of the system is the ability to set the level of performance, image quality and resolution inde- pendently, so a large model can be rapidly manipulated at a lower quality/higher performance setting and then a par- ticular region of interest examined at a high quality/lower performance setting, without having to interrupt the visu- alisation. The SV7 supported projection across a 4.27m by 1.8m dual channel, rear projection Fakespace Power- Wall. This provided a geometrically accurate stereoscopic display using active stereo glasses (Figure 4.2). This combination of analytical power and sophistication of dis- play, rare within archaeology, was demanded by the nature of the datasets under investigation and is comparable with specialist facilities dealing directly with petroleum geol- ogy and remote sensing data sets. To put this in context, a paper published 10 years ago and also concerned with the issues of integrating GIS and other spatial technologies within archaeological projects at Birmingham, records that the entire archaeological computing group was then ser- viced by a single server with 64 megabytes of RAM and 4 gigabytes of mass storage (Buteux et al. 1997, Gaffney and Gaffney 2000a). The NSPP's computing require- ments, and its demand for exponentially increased storage capacity, presumably indicates the scale of change that can be anticipated in future, comparable digital landscape pro- jects.

Figure 4.3 Seismic data slice illustrating a fluvial channel and estuary Thus such an approach is compatible with standard proc- essing and display methodologies used in archaeological prospection (see Watters (2006) for a comparable process used to process GPR data).

Figure 4.4 3D amplitude surface within a GIS. The green lines are bathymetric contours, and allow for the visual corre tion between the anomalies and seabed topography.

Figure 4.5 Segmentation of features of interest from a seismic volume.

Figure 4.6 Wrapping of identified features within seismic data To extract the required volumetric models, the original seismic data was directly segmented utilising picking techniques that are commonly employed in the oil and gas sector (Figure 4.5). A series of user determined lists are generated following segmentation, which contain features of interest. Fully automated selection tools can be utilised

Figure 4.7 Solid model generated by wrapping Figure 4.8 Removal of elements of the solid model within the Avizo package from Figure 4.7 permits visualisation of th internal structure of the system

igure 4.9 Exported solid model within the GIS system. The feature can be seen to be composed of a contour model and a eries of shapes representing the geophysical anomaly at the given time slice. As the volume forming the solid model has real world at- tributes in all three dimensions it becomes possible to gen- erate CAD models that can be imported into a GIS for display and spatial analysis. This mode of representation is more suited to GIS as it is composed of polygonal ele- ments, which, as they wrap the entire anomaly, can visual- ise and facilitate the user’s appreciation of the actual vol- ume and size of anomalies. The results of part of such a process can be seen in Figure 4.9. However, although export of these models from a fully three-dimensional en-

A distinct east-west trending bathymetric deep, the Outer Silver Pit (OSP), lies at approximately 54°N 2°E on the bed of the North Sea (Figure 5.1). This deep is the largest of a series of offshore depressions in the Southern North Sea (SNS) that are thought to have formed during Quater- nary glaciations, either as the product of subglacial proc- Figure 5.1 40 and 60m bathymetric contours of the Southern North Sea. The major bathymetric depressions are labelled.

Table 5.1 Position, dimensions and trends of Ridges A and B

Whilst there has been considerable research into the for- mation of the OSP, the localised geomorphology within the depression has been largely ignored despite its poten- tial to provide further insights into the processes that led to the formation of the depression. The availability of 3D seismic data to this study provided an ideal dataset for the investigation of the morphological and _ stratigraphical characteristics of the OSP. This chapter will demonstrate that 3D seismic data can be used to identify a number of distinctive geomorphological features within the OSP, the Figure 5.2 Bathymetric contours (40, 50, 60, 70 and 80m below sea level) of the Outer Silver Pit area; depicted in white are the elongate ridge features

Figure 5.3 Hilbert transform time slice at 0.06 seconds (top) depicting the two ridges. Two arbitrary lines are annotated which pass through Ridges A (X-X) and B (Y-Y).

Table 5.2 Average dip of the flanks of Ridges A and B 1.13° on the steeper slope of Ridge B. Given that the trends of the ridges are offset the steeper slope of Ridge A faces south east whereas the steep flank of Ridge B faces south west (Figure 5.3 and Figure 5.4). Cross sectional views of the features (Figure 5.3) also reveal that the crests of the ridges are relatively flat. In cross section, Ridge A displays an asymmetry along its major axis and both ridges are asymmetric along their minor axes (Figure 5.3 and Figure 5.4); the asym- metry of the minor axis of Ridge A decreases towards its southeasterly end. The steeper slopes of the ridges are slightly concave whereas the shallower slopes are predominantly convex (Figure 5.3). The dip of the ridges’ flanks are relatively shallow (Table 5.1 and Fig- ure 5.5), reaching an approximate maximum of only

Figure 5.4 3-D illuminated view of Ridges A and B, with a depression referred to in the text highlighted in red. The ridges are vertically exaggerated.

Figure 5.5 Dimensions of the OSP and Ridges A and B

Figure 5.7 Quaternary geology map of the eastern end of the OSP overlain with contours of the depth to the base of the Holocene sediments (black) and the bathymetry of Ridges A and B (dashed red). (Adapted from Larminie, 1989a and 1989b)

The geology map of the Holocene and seabed sediments (Figure 5.7) reveals many significant properties of the ridges. The contours of the depth to the base of the Holocene sediments in the Silverwell area shows an absence of the ridges in the topography, thus it can be concluded that the ridges are of Holocene age. The Holocene sediments of the ridges consist of the Tershel- lingerbank Member (of the Nieuw Zeeland Gronden Formation), which is defined as open marine sediments and are slightly muddy sands (mean grain size 120- Figure 5.8 An RMS amplitude map from 0.004 to 0.05 seconds (TWT) of the eastern end of the OSP. N.B. the lighter col- ours depict high amplitude response areas 300m) that have been derived from Pleistocene glacial and periglacial deposits (Larminie, 1989a). These sediments are overlain by sands at the seabed and in relation to the sur- rounding areas this is a larger sediment size than present at the seabed of the OSP depression but of the same size or smaller than the seabed sediments on the nearby higher ground. Furthermore, Figure 5.8, an RMS amplitude map of the ridge area, depicts the sand ridges as having a high am- plitude signal commonly associated with sandy sediments.

Figure 5.9 Truncated strata on the bed of the Outer Silver Pit (location depicted in the insert). The red horizon is the Holocene Base reflector (Hebbeln and Meggers, 1999); the blue horizons show the strata that have been truncated

Figure 5.10 Changes in land area with rising sea level based upon the depth to base Holocene map (Larminie, 1989b, Ridges A and B are depicted in red

Figure 5.11 Postulated land configuration at the time when the OSP sand banks were last active. The contours are the bathymetric contours from Larminie (1989a). The map is based on a number of assumptions highlighted in the text

Figure 6.1 (a) Amplitude time slice (0.1 seconds) centred on prominent salt dome with a characteristic concentric reflector pattern. (b) Amplitude time slice (0.1 seconds) with interpretation. A fluvial channel that flows sub-parallel to the salt dome is identified, and it is suggested that salt-related faults have also influenced the direction and geometry oj the fluvial channel. (c) Uninterpreted arbitrary seismic line A-A’, confirming that the concentric feature identified in (a-d) is a salt dome

Figure 6.2 (a) Amplitude time slice (0.076 seconds). (b) As (a) but with interpretation; late Pleistocene-Holocene fluvi channels signified by yellow arrows, Holocene sand waves shown in purple, with the green annotation signifying Pleistocene tunnel valley. (c) Seismic section. (d) Part of seismic section (c) with interpretation

$Figure 6.3 (a) Part of BGS sparker profile 81/03/53 that trends east-west through the OSP. (b) Sparker profile 81/03/53 with interpretation. Seafloor reflector and seafloor multiples shown in red. Tunnel valleys shaded in green, Pleistocene sediments in purple and pre-Pleistocene sediments in yellow. Reflections within the Pleistocene succession are picked ir dark purple. Within the pre-Pleistocene succession discernable reflections are marked by dark blue picks, whils: diffractions (which mark the radial scattering of incident seismic energy) are shown in light blue. (c) Close up of boxec region indicated in (b). (d) Close up of collapse graben with interpretation, showing channel development within collapse graben.$

Figure 6.3 (a) Part of BGS sparker profile 81/03/53 that trends east-west through the OSP. (b) Sparker profile 81/03/53 with interpretation. Seafloor reflector and seafloor multiples shown in red. Tunnel valleys shaded in green, Pleistocene sediments in purple and pre-Pleistocene sediments in yellow. Reflections within the Pleistocene succession are picked ir dark purple. Within the pre-Pleistocene succession discernable reflections are marked by dark blue picks, whils: diffractions (which mark the radial scattering of incident seismic energy) are shown in light blue. (c) Close up of boxec region indicated in (b). (d) Close up of collapse graben with interpretation, showing channel development within collapse graben.

Figure 32 – uploaded by Vincent Gaffney

Related Figures (90)

Connect with 287M+ leading minds in your field