Introduction
During the long history of mineral exploration on the island of Spitsbergen, in Svalbard, most ventures were unsuccessful. Blomstrandhalvøya in Kongsfjorden (Fig. 1), for example, is the scene of an abandoned British marble quarry and mining camp. Discovered in 1906, the Northern Exploration Company developed the territorial claim known to them as Marble Island between 1911 and 1913, but it already lay deserted in 1914 prior to the voluntary British absence from the archipelago during World War I. A brief and final revival of activities occurred in 1919 and 1920. The pre-war desertion of the site only served to fuel rumours that ‘the marble apparently disintegrated when taken to a mild climate, and was valueless.’ (Barne Reference Barne1914; Erskine Reference Erskine1994). This assumption was discredited by Siggerud (Reference Siggerud1963), but it is stubbornly persistent in current guide books and tourist blogs, which all too easily dismiss so-called Ny London as a ‘failure’. Siggerud's (Reference Siggerud1963) hypothesis that tectonic disturbances of Tertiary or more recent age may be the cause of the unsolid state of the marble has not previously been tested. Yet, to further our understanding of Svalbard's mining history and indeed the competition for Arctic mineral resources, it is pivotal to know if the reasons for quarry abandonment were primarily of a geological nature or if they must be sought elsewhere.
Overview map of (A) Svalbard in the European Arctic as well as (B) the Kongfjorden area including (C) the investigation area on Blomstrand (modified after Thiedig and Manby Reference Thiedig and Manby1992; Dallmann Reference Dallmann2015). Red boxes with individual numbers refer to further detailed maps. Legend illustrates geographical, geological and general tectonic setting. Map projection and all illustrated coordinates are within UTM zone 33X.
The project ‘Marble at fault’, funded by the Svalbard Environmental Protection Fund and supported by the University of Groningen, was conceived to investigate why historical attempts at a large-scale development of marble on Marble Island, now Blomstrand, were unproductive. The project title is an intentional play on words as a certain type of structural feature is known in geology as a fault. Faulting may have caused the marble to fracture, thereby undermining its workability and profitability, that is the ease at which good-quality rock could be extracted and its ability to make a profit on the market. To prevent foregone conclusions, however, the project combined interdisciplinary fieldwork with archival research with the aim of identifying the most pressing reasons, real, imagined, or both, for the discontinuation of the development.
This paper focuses on the fieldwork component of ‘Marble at fault’. Grounded in industrial archaeology, the fieldwork engaged geological expertise to conduct non-intrusive and therefore non-destructive geophysical survey at carefully chosen locations on Blomstrand. More specifically, ground penetrating radar (GPR), was used in the immediate vicinity of known historical quarries and boreholes at Peirsonhamna and Grottevika (Fig. 2) with the objective of detecting the presence of structural features that may have driven the work to a halt. Broadly speaking, the fieldwork validated the use of GPR under Arctic conditions and generated new knowledge about this particular marble quarry, which in all likelihood is unique in the circumpolar world.
Detail maps of investigation areas on Blomstrand (modified after Kruse Reference Kruse2013; Dallmann Reference Dallmann2015) indicating the locations of (A) GPR site 1 to 5 at Peirsonhamna and (B) GPR site 6 and 7 at Grottevika. Map projection and all illustrated coordinates are within UTM zone 33X.
General overview of the study area and previous investigations
Geological background
The fieldwork concentrated on known historical quarries and boreholes at Peirsonhamna and Grottevika in the southwestern quadrant of Blomstrand (compare Figs. 1 and 2). As the former English place-name of Marble Island suggests, the island was thought to be made entirely of marble in 1906, and a wealth of material remains attesting to its subsequent development is well-preserved. It is important, therefore, to gain an understanding of the underlying geology of those areas that merit further investigation as well as the historical-archaeological research done to date, in particular any questions currently left unanswered.
Svalbard represents the emergent northwestern part of the Barents Shelf, which uplifted during the late Mesozoic and Cenozoic (for example Johnson and others Reference Johnson, Mørk and Dypvik2001; Dallmann Reference Dallmann, Sigmond and Roberts2007). The uplift of the archipelago was most extensive in its north and west, leaving gradually older rocks in those directions (Johnson and others Reference Johnson, Mørk and Dypvik2001). The geological record of Svalbard ranges from possible Archaean age to the recent past and shows a multi-orogenic development with prominent tectonic events at several stages (for example Dallmann Reference Dallmann, Sigmond and Roberts2007; Elvevold and others Reference Elvevold, Blomeier and Dallmann2007). The geological record can broadly be separated into three divisions: the basement of igneous and metamorphic rocks formed during Precambrian to Silurian times, which has witnessed several episodes of folding and alteration; unaltered late Palaeozoic to Cainozoic sedimentary rocks, which form a trough-shaped structure (Central Tertiary Spitsbergen Basin) from the Isfjord area (central Spitsbergen) to the south; and unconsolidated Quaternary surficial deposits formed during and after the last ice age (for example Dallmann Reference Dallmann, Sigmond and Roberts2007; Elvevold and others Reference Elvevold, Blomeier and Dallmann2007).
A dominant north northwest (NNW)-south southeast (SSE) trending structural grain on Spitsbergen comprises several similarly aligned episodes of tectonic deformation (for example Johnson and others Reference Johnson, Mørk and Dypvik2001; Pirli and others Reference Pirli, Schweitzer and Paulsen2013). Svalbard's tectonic regions may be divided into: Precambrian to early Silurian rocks (Pre-Old Red basement) in the north-eastern part of Spitsbergen and Nordaustlandet; Devonian cover rocks and grabens on northern Spitsbergen; the Central Tertiary Spitsbergen Basin; the Triassic and partly Cretaceous platform areas in the eastern parts of Spitsbergen, on Barentsøya, and on Edgeøya; and the Tertiary fold-and-thrust belt along the west coast of Spitsbergen, which has possibly reactivated older structures (e.g. Johnson and others Reference Johnson, Mørk and Dypvik2001, Dallmann Reference Dallmann, Sigmond and Roberts2007, Elvevold and others Reference Elvevold, Blomeier and Dallmann2007).
This paper concentrates on the Kongsfjord area in north-west Spitsbergen, which is characterised by Mesoproterozoic rocks to the north of the fjord, and Carboniferous and Permian as well as Palaeogene and Neogene rocks to its south. The NNW-SSE trending faults of the Tertiary fold-and-thrust belt dominate the younger tectonic regime here. Furthermore, deglaciation of Kongsfjorden is reported since the Late Weichselian (e.g. Lehman and Forman Reference Lehman and Forman1992, Lauritzen Reference Lauritzen2006, Ingólfsson and Landvik Reference Ingólfsson and Landvik2013) and is linked to sea level change and tectonic uplift, which still affect the region.
Since its discovery in 1906, Blomstrand has been shown to indeed consist mainly of medium- to high-grade metamorphic Caledonian marbles as well as pelitic schists and gneisses (for example Thiedig and Manby Reference Thiedig and Manby1992; Lauritzen Reference Lauritzen2006). These authors describe the Caledonian rocks as striking north-south and having been subjected to two synmetamorphic, co-linear phases of isoclinal folding as well as refolding by post-metamorphic crenulation to kink-folds, which are often associated with west-directed thrust faulting and imbrication.
Additionally, some patches of unmetamorphosed sediments of possibly Devonian age are widely scattered on Blomstrand (Lauritzen Reference Lauritzen2006). The sediments are either preserved in wedges associated with faulting and brecciation of the underlying marbles, or they occur in fan-shaped bodies filling channels and tubes in the marbles. Furthermore, these sediments are often injected into a mesh of calcitic veins which penetrate parts of the marbles (Thiedig and Manby Reference Thiedig and Manby1992; Lauritzen Reference Lauritzen2006). The uplifted coastline of the island also comprises some karstic caves which formed during the last thousands of years (Lauritzen Reference Lauritzen2006).
Recent geological and tectonic maps (for example Buggish and others Reference Buggisch, Piepjohn and Thiedig1992; Thiedig and Manby Reference Thiedig and Manby1992; Thiedig and others Reference Thiedig, Saalmann and Piepjohn1997; Dallmann Reference Dallmann2015) indicate substantial roughly north-south trending thrust folds and faults on Blomstrand (Fig. 1). Two such folds have been drawn to end in the waters of Peirsonhamna and possibly give rise to the conspicuous shape of the natural harbour. Their impact on the former marble workings immediately to the west is, however, unclear. Similarly, a major fold approaches Grottevika from the north but has been drawn to terminate in a series of marine caves, which mask the relationship with a small marble quarry at this location.
A question which geologists often do not stop to ask, probably because of the chasm between geological and ‘human’ scales, but which is crucial to this study, is how much of this geological knowledge was available to the historical actors arriving on the scene a century ago. In fact, the chemistry of marble and its geological formation did not need to be a mystery to anyone who cared to inform himself properly. Renwick had published Marble and marble workings in 1909, which treated the dimension stone on a global scale. Yet, local knowledge of the resource and how to work it under Arctic conditions needed to be generated on site. Incidentally, the Northern Exploration Company employed Renwick as an advisor in the field in 1912 and again in 1913.
Historical-archaeological setting
The illustrious English prospector Ernest Mansfield, having had some prior experience of marble quarrying in a cold climate (Barr and others Reference Barr, Newman and Nesteroff2012), discovered and named Marble Island, measuring 17 km2, in summer 1906. Under Mansfield's charge, the Northern Exploration Company began to develop the claim in spring 1911. The site lay deserted in summer 1914 and during World War I. Development briefly resumed in 1919 and 1920 but without success. The marble was never worked again.
Under Arctic conditions, material remains in Svalbard are generally well-preserved, and the exceptional industrial heritage of Blomstrand is now protected, with many but unfortunately not all of the features listed in Norway's Askeladden database. In 2008, the LASHIPA (‘Large-scale historical exploitation of polar areas’) project carried out a comprehensive industrial archaeological survey of the former British marble quarries and associated mining camp at Peirsonhamna (UTM 33X 436791.5 m E 8767608 m N), commonly but anachronistically known as Ny London (Aalders and others Reference Aalders, Avango and Gustafsson2009). Moreover, researchers of the Netherlands Arctic Station in nearby Ny Ålesund rediscovered and recorded a small marble quarry at Grottevika (UTM 33X 435696 m E 8768720 m N) in 2011. This find remains unreported and currently does not have a corresponding Askeladden entry.
Within the framework of LASHIPA, Kruse (Reference Kruse2013) investigated the far-reaching activities of the Northern Exploration Company on Spitsbergen. She has been able to conclude that despite being a well-known and popular tourist attraction today, the history of the company's marble quarries remains poorly understood. The documents she has consulted to date are vague about the workability and the profitability of the marble. From the start, unknown actors have been quick to pinpoint the effect of permafrost, which supposedly thawed upon transportation to warmer temperatures and caused the blocks to crumble, to be the reason behind the company's inability to exploit the building stone. This has been repeated uncritically in numerous guide books and tourist blogs. At best, the quality of the marble is being described as low due to ‘many cracks’ (Stange Reference Stange2012: 301). Since in geology, ‘cracks’ can have different origins, this statement allows for analysis and reinterpretation. In fact, the notion that permafrost was to blame has already been tested and discarded by Siggerud (Reference Siggerud1963). He instead proposes unfavourable structural geology such as subsurface fracturing, which would cause the marble to fail on a medium to large scale. In this paper, we test his hypothesis.
Materials and methods
The project ‘Marble at fault’ envisaged interdisciplinary fieldwork on Blomstrand in Kongsfjorden, which took place between 12 and 16 July 2014. Owing to the island's relative isolation, there were limits to the tasks that could be carried out within the financial and time constraints. Most conveniently, researchers fly via Svalbard's capital of Longyearbyen to the international scientific village of Ny Ålesund, from where Peirsonhamna is located a 4.5 km boat-ride in northern direction across the fjord. They commonly ship any heavy or bulky material well in advance of the field season. The commercial GPR equipment, however, could not be spared until the authors’ actual date of departure, so we were forced to reduce the size and weight of our equipment, including all personal belongings, to a total of approximately 60 kg. This still exceeded the very strict allowance on flights to and from Ny Ålesund. Having a very narrow fieldwork window, we could not risk any of our equipment not reaching Blomstrand. As a result, we worked with only one GPR antenna instead of a range of antennae with different frequencies.
In 2008, members of LASHIPA observed a substantial fracture zone ~3 m wide in the Winter Quarry of 1919–1920 in Peirsonhamna (Fig. 3). At the time, they could not determine if this fracture was the reason behind the quarry's abandonment. In 2014, we therefore carried out a GPR survey to investigate this and to delineate other possible cracks, joints, faults, and breaks in a number of outcrop subsurfaces. Due to the logistical restrictions mentioned above, only one GPR antenna, a medium-frequency antenna, promised to be the best compromise between the spatial resolution of any fractures and the achievable penetration depth. We used a GSSI 400 MHz antenna with a survey wheel, a SIR-3000 data acquisition system, and a handheld GPS. Trace increment was set at 0.02 m in order to achieve a high data density; the range was set at 90 ns TWT (two way travel time) with a sample rate of 1024. Using the 400 MHz GPR antenna, a GPR target depth of ~4 m BGL (below ground level) could be aimed for.
The Winter Quarry of 1919–1920 in Peirsonhamna on Blomstrand exposed a substantial fracture zone. B. Koster is standing above the southern fault plane with the northern fault plane ~3 m to the right, where it meets a slip surface at right angles (photo: F. Kruse).
The number of GPR survey grids that could be completed was dependent on the time spent on site, which in turn depended on weather and ice conditions as well as transport arrangements between Ny Ålesund and Blomstrand. The positioning of the survey grids was guided by Kruse's (Reference Kruse2013) inquiry into historical quarries and boreholes. Ultimately, seven grids were laid out in such a way as to allow for the best comparison between historical information and subsequent GPR data (compare Figs. 1, 2 and 4). A profile spacing of 1 m was chosen, with which 125 GPR profiles were completed. Using Surfer 11 (Golden Software), it was subsequently possible to reconstruct pseudo-3D subsurface models to visualise fracture zones in the marble (for example Grandjean and Gourry Reference Grandjean and Gourry1996; Godio and others Reference Godio, Sambuelli and Ferrero2003). General GPR data processing was then carried out with ReflexW V7.5 (Sandmeier Scientific Software) and included static correction, background removal, gain adjustments, and velocity adaption for time-depth conversions. The latter was based on a hyperbola analysis where such features were present. Additionally, we used characteristic velocities based on the subsurface material, which in this study is of course marble (Grandjean and Gourry Reference Grandjean and Gourry1996; Neal and Roberts Reference Neal, Roberts, Pye and Allen2000; Neal Reference Neal2004; Sambuelli and Calzoni Reference Sambuelli and Calzoni2010).
The positions of GPR survey grids 1 to 7 (orange outlines with correspondingly numbered corners) in relation to historical quarries (solid grey areas) and boreholes (red rhombi). Fractures that were observed on the surface are marked as dashed lines. The blue arrows denote the start of the survey as well as the main direction for processing. The black arrows indicate the zigzagging profiles as well as the last profile in each grid. Blue arrows indicate starting area within individual GPR grids and main orientation for further pseudo-3D modelling (figure: B. Koster).
Subsurface features were detected due to hyperbolic features and abruptly shifted continuous reflectors in the data (for example Grandjean and Gourry Reference Grandjean and Gourry1996; Godio and others Reference Godio, Sambuelli and Ferrero2003; Sambuelli and Calzoni Reference Sambuelli and Calzoni2010). For some grids, the data could partly be correlated with known historical quarries and boreholes sunk by the Northern Exploration Company in the twentieth century, which served to verify and trace the detected features.
Simultaneous with the GPR survey, photo documentation, study area description, and the levelling of study areas were carried out in the field. Historical boreholes were calibrated by handheld GPS in order to compare borehole logs, where these existed, with the GPR data from these locations.
Results
GPR survey grids 1 to 7 were positioned in relation to known historical quarries and/or boreholes (Fig. 4). The GPR survey subsequently generated a total of 125 profiles with lengths between 10 m and 40 m. All GPR measurements were acquired on outcrop surfaces, that is not on any unconsolidated overburden overlying the marble bedrock, to ensure direct coupling with features at subsurface level. The profiles were mainly recorded at right angles to continuous fractures observed either at ground level or in historical quarries as indicated in Fig. 4. Due to the particularly strong reflectors visible in the raw data of Grid 7, profiling was carried out in both perpendicular and parallel directions. The weather conditions on the days prior and during the survey were arid and mostly sunny with temperatures between –8°C and +5°C. The survey grids and surrounds were free of snow, ice, or water retention.
The main hyperbolae velocities in the GPR data range between 0.09 m/ns and 0.10 m/ns, which is similar to electromagnetic wave velocities given in the literature for marble rocks (compare Grandjean and Gourry Reference Grandjean and Gourry1996; Neal and Roberts Reference Neal, Roberts, Pye and Allen2000; Neal Reference Neal2004; Sambuelli and Calzoni Reference Sambuelli and Calzoni2010).
Survey grids 1 and 2
The area under investigation in survey grids 1 and 2 lay immediately west and southwest of the Winter Quarry opened in winter 1919–1920, in which both a substantial fracture zone and a natural slip surface had been exposed (Fig. 3). This area also comprised three boreholes clearly labelled B1-68, B1-B-68, and B2-68 (Fig. 4A), the completion of which the chance find of a calendar from 1968 in a hut in Peirsonhamna dated to 27, 30 and 31 July respectively. Norwegian drillers were on site between 20 July and 22 August that year to sink a total of fifteen boreholes. Their identities and intentions remain unknown to us; we have neither discovered all fifteen boreholes in the field nor found their borehole logs. This is a line of inquiry for the future.
The two grids were surveyed in different directions due to the NW–SE trending of the observed fracture zone and with the objective of visualising additional structural features at subsurface level as effectively as possible (Fig. 5A). Throughout the whole of the GPR data for both Grid 1 and Grid 2, the observed fractures can be traced from NNW to SSE while approaching the maximum survey depth of ~4 m BGL (Fig. 5B).
The positions of surficially observed fractures at subsurface level within the survey grids 1 and 2 as well as two boreholes sunk in 1968 (red rhombi). (A) Fractures visible at both ground level and subsurface level (left, green circles) and fractures only detected at subsurface level using GPR (right, grey circles). Generally, the size of the circles increases with depth to ~4 m BGL. (B) Grids 1 and 2 visualised in a pseudo-3D model (colours as before; circle size is constant) (figure: B. Koster).
Abruptly shifted reflectors in the GPR data as well as vertically aligned hyperbolae reflections with velocities of around 0.09 m/ns are indications of fractured marble. In some parts, the fractures branched in various directions, which were not observed from above but detected by GPR. In this area, numerous vertical cracks, which were not seen on the ground, were visible in the GPR data, starting from a minimum depth of ~0.4 m BGL. These cracks are aligned and branched in several directions giving rise to chaotic fracture patterns. Within both survey grids, only small parts of the marble were free from cracks and faults.
Survey grid 3
The area under investigation in Grid 3 is located ~350 m SSW of grids 1 and 2 at the end of a former railway embankment and to the west of the historical No. 2 Quarry opened in summer 1911. It incorporates a historical borehole drilled and cemented in winter 1912–1913 (Fig. 4B). Interestingly, the borehole has been sunk into one of two linear fractures clearly visible at ground level along the eastern and western edges of the grid. Between the cracks, homogenous conditions of the marble could potentially exist. The survey direction was thus chosen to be perpendicular to these two cracks at the surface, both of which plainly show in the GPR data down to the maximum detection depth of ~4 m BGL (Fig. 6A). Along the northern edge of the survey grid, the GPR data also indicate several small vertical fractures in the subsurface.
The positions of surficially observed linear fractures at subsurface level within Grid 3 as well as a borehole sunk in winter 1912–1913 (red rhomb). (A) Fractures visible at both ground level and subsurface level (top, green circles) and fractures only detected at subsurface level using GPR (bottom, grey circles). Generally, the size of the circles increases with depth to ~4 m BGL. (B) Grid 3 visualised in a pseudo-3D model (colours as before; circle size is constant) with illustrations of multiple folds at subsurface level from slightly different perspectives (figure: B. Koster).
Unexpectedly, the GPR data also point to two blended folds with several branching fractures (Fig. 6B). The top of the upper fold is located only ~0.3 m BGL. The folds are characterised by relatively clear and continuous reflectors with sporadic linear adjacent hyperbolae reflections at velocities of ~0.09m/ns. Hence, the marble in Grid 3 is seriously fractured due to the presence of the folds and adjacent fracture zones as well as several scattered vertical cracks.
Survey grids 4 and 5
The areas under investigation in survey grids 4 and 5 are located only ~30 m NE of grids 1 and 2 and on either side of the former No. 3 Quarry opened in summer 1912 (Fig. 4C). There are also two historical boreholes: one presumably drilled in winter 1912–1913 and the other clearly labelled B5–68, thus dating from 8 August 1968. The GPR survey was carried out parallel to the direction of channelling, that is quarrying by machine, and perpendicular to a long fracture seen on the outcrop surface. The GPR data in both grids are slightly overdriven, and any fractures are challenging to delineate. In both grids, the long fracture visible at ground level can be followed into the subsurface using several small hyperbolae as well as displaced continuous reflectors; these most likely indicate a fault (Figs. 7 and 8). The depth of this feature can be traced in the GPR data to the maximum detection depth of ~4 m BGL.
GPR profile 69 recorded in grid 4, the red triangles indicating fractures visible at ground level: (A) processed GPR profile, (B) processes and analysed GPR profile, and (C) interpretation of GPR profile (figure: B. Koster).
GPR profile 78 recorded in grid 5, the red triangle indicating a fracture visible at ground level: (A) processed GPR profile, (B) processed and analysed GPR profile, and (C) interpretation of GPR profile (figure: B. Koster).
Furthermore, a smooth continuous reflector at ~2 m BGL was detected in the northern part of Grid 4 (Fig. 7). This feature possibly indicates an internal boundary or a dislocated surface within the marble. In some GPR profiles, smooth reflection patterns with several small hyperbolae were noted. These may indicate chaotically fractured areas in the subsurface. It is not clear whether these belong to fractures near the ground surface or internal alteration, that is weathering, of the marble. Besides the main fracture in grids 4 and 5, there are several small, near-vertical fractures. It remains challenging to interpret these due to the slightly overdriven quality of the GPR data.
Survey grid 6
The area under investigation in Grid 6 is located ~1.5 km NW of Peirsonhamna and ~200 m S of the beacon in Grottevika at the southern margin of that small cove (Figs. 2 and 4). It comprises the historical Quarry A13 opened in spring 1920 (Fig. 4D). Grid 6 was positioned along the northern edge of the quarry, and the profiles commenced at right angles to the direction of the quarry face but perpendicular to linear fractures observed in the quarry trench and on the outcrop surface.
The GPR data comprise several discontinued reflectors as well as scattered chaotic hyperbolae reflections. On the basis of these indicators, the linear fractures noted at ground level can be traced into the subsurface down to a depth of ~3.5 m BGL (Fig. 9). These indicators also give rise to some small, near-vertical cracks not visible at the surface but existent at depth. Broadly speaking, the marble in this area does not appear to be highly fractured. This result, however, must be viewed in relation to the results obtained in nearby Grid 7 below.
GPR profile 102 recorded in grid 6, the red triangles indicating fractures visible at ground level: (A) processed GPR profile, (B) processed and analysed GPR profile, and (C) interpretation of GPR profile (figure: B. Koster).
Survey grid 7
The area under investigation in Grid 7 is located ~40 m N of Quarry A13 and Grid 6. It comprises two historical boreholes drilled in summer 1920 (Fig. 4E). Leech (Reference Leech1920) describes one as a 3-inch hole (7.6 cm) and the other as 4-inch (10.2 cm) and that both have been sunk to 7 feet (2.13 m), where the marble crumbled due to a suspected subsurface fault. It was surprising, therefore, to find that these holes were only ~8 cm apart and that the dividing rock wall had not crumbled over the years. The smaller hole was in fact measured to be 2 m deep, the larger 1.9 m. The level of liquid water was recorded at 1.47 m BGL.
A linear fracture running SW–NE was observed on the outcrop surface. The GPR data reveals a strong continuous reflector in both the perpendicular and the parallel profiles. This reflector is accompanied by further hyperbolae at a velocity of 0.09 m/ns that are chaotically distributed alongside it. Pseudo-3D visualisation of selected GPR data (Fig. 10) indicates a fold with its axis very near ground level, i.e. at only 0.2 m BGL. This axis is most likely the reason for the linear crack observed at the surface. It is clear that the marble in Grid 7 is fractured due to the presence of the fold with its surrounding fracture zone of ~0.5 m across. The detected fold meets the borehole at ~2 m which corresponds with the maximum depth of the borehole itself.
The position and orientation of a fold within the subsurface of grid 7 as well as the position of two historical boreholes sunk in summer 1920 (red rhomb): (A) Contour map of the fold with indicative data points (green circles; constant size at different depths). (B) Pseudo-3D model of grid 7: (top) indicative data points at constant size and (bottom) 3D contour model (figure: B. Koster).
Additional observations
During an extensive walkover, we were able to see for ourselves that Blomstrandbreen, the glacier that once formed a link between Blomstrandhalvøya and the opposite northern shore as well as a physical barrier to mariners, had since the time of the historical development melted back to the extent of proving that the halvøy (peninsula) of Blomstrand is in fact an øy (island). In light of this impressive example of glacial melt, it was not difficult to imagine that the processes of isostasy are still ongoing in Svalbard and on Blomstrand.
In Peirsonhamna, we noted a fresh N–S trending fracture (Fig. 11), which may be a result of uplift in the area. This and similar fractures, presumably also at subsurface level, may in turn have had and still have significant effects on the workability and profitability of the marble here. Observed fractures were not only limited to N–S trending; they occurred in various orientations.
A recent fracture probably caused by tectonic uplift which still affects Blomstrand, looking south towards Ny-Ålesund (photo: F. Kruse).
Discussion
This paper examines to what extent unfavourable bedrock geology can be blamed for the historical quarry abandonment on Blomstrand in Kongsfjorden. Early rumours that the presence of permafrost, which melted upon transportation to milder climates, had caused the marble to crumble have been investigated and disproved by Siggerud (Reference Siggerud1963). We therefore tested Siggerud's own hypothesis that tectonic processes created detrimental structural features which led the marble to fail on a medium to large scale. We envisaged that a GPR survey would be an appropriate tool to reveal the competence of the marble at subsurface level in the vicinity of known historical quarries and boreholes in Peirsonhamna and Grottevika. We completed seven survey grids and a total of 125 profiles, the result of which are the subject of further discussion here. Beside the main N–S striking faults, we were able to detect further (sub-) surface structures with different trending respectively striking directions.
Based on the assumption that the direction of working was downward, the former quarry floors would have been obvious targets to survey. This, however, was not possible due to the quarries being too small in addition to their irregular floors being covered by rubble or submerged in meltwater. The best alternative, and in view of the planned development probably also being outward, was to survey the surrounding area of the quarries and to establish a relationship between the workings and the survey grids. Due to the maximum detection depth of ~4 m BGL with the 400 MHz GPR antenna, the marble subsurface was barely investigated to depths exceeding the already exposed quarry floors, which themselves scarcely surpassed the weathered zone or the active layer of the permafrost. Nonetheless, the survey provided a wealth of information.
The presence of the obvious fracture zone in the Winter Quarry of 1919–1920 could be confirmed at subsurface level in grids 1 and 2, thereby validating our choice of method. The question is did the relevant people present at the time deem this fracture, which would have affected the marble's workability and profitability in its vicinity, a reason to abandon the trench. Fieldwork alone cannot provide the answer, but a fortunate circumstance in historical archaeology is the existence of historical documents that may. Thus, marble manager Leech (Reference Leech1920) writes,
[. . .] for the production this season, the winter quarry would be quite useless, owing to the fact that the trench had been driven along a natural slip. This was visible on the surface and I could not understand why Mr. Sant had chosen this spot. Explosives had been used, and there did not exist a quarry floor or face, owing to the manner in which the rock had been shattered. It would have been necessary to remove the rock on the left hand side to the bottom of the quarry trench (some fifteen feet) then to channel a cut some five feet on right hand side and end of trench to get beyond the area damaged by explosives – thus making a quarry face and floor. The fault fifteen feet wide running through the present trench would all be waste. Until the above has been accomplished this quarry will not be in a producing stage. In my opinion, with No. 3 quarry able to produce almost the same variety of marble some 50 feet away, when the quarry floor has been cleared and the channelers capable of work again, it would be a waste of time and money to continue with the winter quarry (Leech Reference Leech1920: 6).
The relevant people at the time did therefore not deem the fracture alone a reason for abandonment. This example demonstrates the limitations of doing fieldwork only and relying solely on geological results; it does not provide the whole picture. In combination with archival sources, any features encountered on site can be evaluated according to people's knowledge and perceptions. It is commonly a mixture with seemingly tangential factors that influences the decision to reject one quarry, in this case in favour of another location. Regarding the Winter Quarry, Leech (Reference Leech1920) points to managerial incompetence and pressure on finite resources as the most pressing reasons. The fault is not crucial, although Leech (Reference Leech1920) recognises that it would lead to the production of much waste rock. His statement implies that while some waste rock is acceptable, there is a limit to the proportion between waste and marketable product before a quarry is no longer lucrative. On that note, we may also assume that due to high production and labour costs only the highest quality of competent Arctic marble blocks could find a market; anything less would probably spell a quick end for the venture.
Besides geological information provided on the outcrop surface and in the Winter Quarry, how much could the Northern Exploration Company have known about the marble's workability and profitability in light of the technological capabilities of the day? They may have sunk a borehole into the outcrop, but we could not assign one of the many holes in that area with any certainty. Neither do we have the borehole logs from 1968. The company will not have had access to geophysics, and it is valid to ask at this point how the generation of data to which our historical personalities will not have had access has anything to add to the subject of mining history. In fact, the newly acquired GPR data of areas that were never worked in history serve to dispel persistent myths and help to clarify the above-mentioned historical process of more or less informed decision-making by raising pivotal questions about all the options available. Leech will have known about large faults negatively affecting production. If he was able to discern a general fault pattern at the surface, he will have wanted to know if the marble between the large faults was homogenous and intact. The GPR data in grids 1 and 2 reveal numerous fractures at subsurface level that branch in multiple directions. Only small parts of the grids are seemingly free from the emerging chaotic fracture pattern, thus leading us to conclude that further development of the Winter Quarry would have produced more broken rock and would most probably have been uneconomical. Had the Northern Exploration Company continued at this location, it would only have been a matter of time before they had discovered this themselves.
On the subject of borehole data, a borehole had been drilled into the outcrop investigated in Grid 3, its edge cemented to prevent it from breaking off. We are certain to have the corresponding log in a borehole that was sunk by the Scot David Booth at the ‘old Quarry place’, that is No. 2 Quarry, between 20 January and 6 March 1913. The completion of the borehole was delayed because some of the winterers joined in the search for two members of the lost Schröder-Stranz expedition believed to be in Wijdefjorden. On completion, Booth reached 15.27 m BGL, but the core was very broken with the best piece retrieved only being 24 cm long. He never stopped to question his technique, while the discovery of a nail down another hole led him to suspect sabotage. It is curious that he had placed this borehole on one of two fractures clearly visible on the outcrop surface. This may perhaps be explained in terms of a light snow cover obscuring his vision or a lack of basic geological knowledge. It is more curious still that Leech (Reference Leech1920) should report that,
many of the cores obtained were polished and sent to the London Office, and appear to be sound, attractive and good class marbles. I presume there is a report in the London Office by Mr. Miller, the Manager of this party, and this probably contains valuable information as to the quality and variety of marble at various depths. There is a report by Mr. Booth, the Mechanic, but this gives no detailed information of the marble at various depths (Leech Reference Leech1920: 3).
In fact, none of Booth's borehole logs speak of ‘good class marbles’. We would like to believe that this is a simple misunderstanding, but unless the cores were collected during a drilling campaign entirely unknown to us, it suggests fraud while at the same time demanding extreme caution when dealing with historical documents. If the Northern Exploration Company is guilty of falsifying their records, Booth's day book becomes all the more intriguing as containing information that the public by and large was never meant to know (Booth Reference Booth1912–1913). Further inquiry along these lines, however, is outside the scope of this paper.
Besides serving as an instructive example of the use of fieldwork data in combination with historical sources, Grid 3 showed GPR survey to be capable of detecting not only fracturing but also folding at a shallow depth and at a small scale within the subsurface. We could not discern whether movement had recently taken place along the fold and whether this was a result of continuing tectonic uplift on Blomstrand either reactivating older or generating new structures. In terms of intactness of the marble, however, the presence of the fold had possibly led to an adjacent fracture zone and scattered vertical cracks. At the same time, significant areas in the grid remained unaffected. The fracturing would have undermined the outcrop's overall productivity, but if the proportion of waste rock to marketable product fell into the acceptable limit, the outcrop may yet have been worked profitably.
Grids 4 and 5 lay on either side of No. 3 Quarry opened in 1912, which even Leech, ostensibly a capable marble manager, continued to be enthusiastic about in 1920. The GPR data from the grids had been slightly overdriven, yet they disclosed a smooth continuous reflector, probably a dislocated surface, at a depth of ~2 m BGL, which may or may not have been disruptive during working. Although the results remain inconclusive, they remind us not only to be looking for fractures in the search for competent rock. We tend, in fact, to call the absence of indicative fractures and folds inconclusive, questioning the data, instead of concluding homogenous, intact, and therefore workable conditions.
Grid 6 lay adjacent to Quarry A13 opened in Grottevika in spring 1920. This small forgotten quarry covers an area of ~600 m2 and only reaches a maximum depth of ~1.4 m BGL. About the quality of its marble, Leech (Reference Leech1920) relates.
The broken nature of the marble on the surface almost entirely disappears, the fractures were closely cemented together and on testing with the hammer in situ rang soundly.
I regret to say, however, that other disturbing features appeared. This marble has a greyish blue ground with white bands running through; these bands are approximately 12 ins. wide interspersed with greyish blue in places. Thus the grey is far too prominent to make it valuable marble. However, this information regarding the soundness of the marble at slight depth under so shattered a surface, is a very reassuring factor.
As the strata dips at an angle of about 25° it would be far too expensive to cut out the white bands, and leave so large proportions worthless. Besides the cost of extraction, it is by no means up to the samples of Norden Lys now in the London Office, and I am very disappointed with it.
Also a series of beds of similar marble protruding through what is at present the quarry floor is a very disturbing feature.
The quarry floor runs practically in a line due south east to north west. The intruding layers run north and south. This would mean, in working the quarry, the break up of the whole quarry face, as far as good blocks are concerned, and would also lead to extensive waste (Leech Reference Leech1920: 7)
Broken surface rock has always been a point of contention in the ‘failure’ discussion in the past and at present, but Leech indicates that the weathered zone, that is the zone affected by freeze-thaw action, was not necessarily deep and marble at shallow depths could be sound. His statement further highlights a number of factors that influenced the suitable positioning and successful working of a quarry. Our GPR data merely revealed some surface cracks which ceased at shallow depths. Rather than look for problems with the marble where there may not be any, we are inclined to say that the rock at this location is the soundest yet, but grid 7 lay uncomfortably close by.
Grid 7 comprised two boreholes, which were sunk under Leech's supervision in 1920 to enable him to form an opinion of the strata underlying those exposed in Quarry A13. He reports that, ‘Although both 3 in. and 4 in. holes were drilled to a depth of 7 ft. each, the marble crumbled, and shewed that we had struck a fault not visible on the surface. This should not be taken as evidence of unsoundness of the marble at depth, but was due solely to a fault’ (Leech Reference Leech1920: 8). Following the drilling, Leech abandoned Grottevika owing to a shortage of water to continue exploration, wanting to drill at a different location, and thinking it unwise to spend time here when No. 3 Quarry should soon produce an output to test on the market. The abandonment was to be temporary, but after 1920, neither the area nor the island were explored let alone worked again.
During the fieldwork, the said boreholes were found to be separated by a mere 8 cm, and we cannot think of a plausible reason as to why two explorative holes should be placed so close together. Perhaps the driller wanted to make absolutely certain that there was nothing amiss with the technique, a faulty technique leading to shattered rock, and that there was in fact a fracture at the reported depth. Most impressively, the narrow rock wall between the boreholes had not crumbled over time, which, despite Leech's assertion to the contrary, indicates that the marble could be very sound indeed. Our GPR results, however, show very clearly that Leech was right about a fracture at ~2 m depth.
Grid 7 comprises a substantial anticlinal fold at relatively shallow depth with a ~50 cm fracture zone around the limbs. It is not known how, if at all, this fold affects Grid 6 and Quarry A13. It may not even undermine the workability of the outcrop greatly, although it would most likely lead to some amount of waste rock. It may be possible to reduce the waste rock by carefully positioning the quarry. To make the most of the GPR results, they will in the near future be discussed with marble quarry specialists.
Broadly speaking, we were able to detect structural features within the subsurface of the marble on Blomstrand to a depth below the weathered zone. That the marble should comprise fractures and folds is in itself not surprising, it being metamorphic rock, in this case partially or completely recrystallised under increased pressure (as opposed to increased temperature). Whether the marble is workable and profitable depends on the degree to which any cracks, joints, or faults have been cemented, or in fact reactivated. The dimensions and orientations of the detected fractures and folds vary greatly between the survey grids. They can be said to occur on a small geological scale as opposed to a medium or large one. On this small geological scale, the frequency at which detrimental features occur may still be intermittent enough for quarrying ‘in between’ to be lucrative on a historical scale, that is a human scale. The GPR survey was not sufficiently extensive for commenting on an overall pattern across the outcrops. Regarding orientation, not all the features could be related to the major thrust faults and folds on the island.
While some fractures were observable on the outcrop surfaces, the majority of the fractures was not. The folds could not be seen at all. This is the issue at the heart of this paper. Arriving on Blomstrand just over a century ago, the actors involved in exploration and quarrying were unable to gain a comprehensive understanding of subsurface features from visual survey and boreholes alone, especially in light of a poor drilling technique being used. Modern techniques have not addressed the question of subsurface damage near historical workings, as a result of which present geological and tectonic maps do not indicate fractures and folds in areas of historical interest. The crux of the matter lies in the difference between geological and historical scales, be it in research or otherwise. It is doubtful, however, if detailed geological mapping at the former quarries and boreholes can close any of the remaining knowledge gaps in Blomstrand mining history.
In view of the good-quality data obtained, we found GPR to be an appropriate as well as cost-effective tool for the investigation of structural features in Arctic bedrock. Previous studies have validated the use of GPR in marble quarries in the Mediterranean (Grandjean and Gourry Reference Grandjean and Gourry1996; Godi and others Reference Godio, Sambuelli and Ferrero2003). We expected Arctic conditions to be very different, and our investigation may only have been possible due to the favourable Arctic summer weather between 12 and 16 July 2014. The air temperature was around 0°C. It was arid, there was no snow cover or water detention on the outcrop surface, and there was no permafrost-affected overburden. We chose an appropriate GPR setup and an investigation-matching antenna frequency, which was 400 MHz. A lower frequency antenna of 270 MHz or 200 MHz could perhaps have detected to lower depths below ground level, but the trade-off would have been a lower resolution (for example Neal Reference Neal2004), in which case smaller fractures may no longer have been revealed.
The fieldwork focussed purely on possible geological reasons behind the abandonment of historical marble quarries, yet complementary historical documents showed that fractures and folds were not the most pressing and arguably not a reason at all. In fact, a multitude of other factors needs to be considered that drove the historical process. Kruse (Reference Kruse2013) touches on some of these factors while the project ‘Marble at fault’ aims to discern more using comprehensive archival research. The combination of economic, socio-political, technological, cultural, and other driving forces that played a role on both the local and global scale will be the subject of an upcoming paper.
Conclusion
In this paper, we set out to test firstly if Siggerud's (Reference Siggerud1963) hypothesis that tectonic disturbances of Tertiary or more recent age caused the unsolid state of the marble on former Marble Island, now Blomstrand, and secondly if structural features were responsible for the British development on the island being abandoned. Using GPR, we were able to detect impressive fractures and folds in the areas surrounding historical quarries and boreholes. These occurred at shallow depths and on a small geological scale, and we conclude that tectonic processes have indeed had a negative effect on the marble. While we did not determine their age, field observations indicated that isostatic uplift is still ongoing in the region, leading to further stress and strain resulting in (re-) fracturing and (re-) folding of the bedrock. Using fieldwork alone, we could not establish if unfavourable geology directly led to quarry abandonment. Complementary historical sources, however, suggest that geology was not a crucial factor at all.
We generated knowledge about the marble that the original personnel will not have had. The strength in our research lies in dispelling the persistent myths that permafrost caused the marble to crumble and that shattered surface rock is an infallible indicator of unsound rock at depth. Persistent myths inexcusably undermine progress in historical research. Furthermore, we draw attention to the differences between geological and historical scales and foster the notion that bridging the gap leads to the necessary appreciation of the historical process. In mining history, decision-making is not based on geology alone; successful ventures are an interplay of a multitude of economic, socio-political, technological, cultural, and other driving forces on both the local and global scale.
Uncovering pivotal information about the hitherto poorly understood and probably unique material remains of the Arctic marble quarry and its associated mining camp on Blomstrand has the potential of benefitting Svalbard. This new information enhances the site's value as a source of historical and polar knowledge, as an intense visitor experience of a bygone industrial age, as well as a source of revenue through improved tourism.
Acknowledgements
The research would not have been possible without funding from the Svalbard Environmental Protection Fund and additional support from the University of Groningen. The authors also thank Winfried Dallmann and Karsten Piepjohn for their enthusiasm at the planning stage; Wieke de Neef and Burkart Ullrich for providing the missing link; the Georadar Forum for bringing us together and providing the GPR equipment; Maarten Loonen of the Netherlands Arctic Station as well as Max Schwanitz and Thomas Dupeyron of AWIPEV for support in the field, and Anna Lena Ekeblad and Sander Solnes of the Svalbard Museum for their exceptional hospitality.
