Reservoir architectures of interlava systems – a 3D photogrammetric study of Eocene cliff sections, Faroe Islands

Introduction

The Faroe-Shetland Basin in the NE Atlantic Ocean has proven Paleocene-Eocene hydrocarbon discoveries (Cooper et al. 1999; Leach et al. 1999), but exploration is hampered along its north-western margin by a blanket of predominantly subaerial basalt lava flows (Passey & Hitchen 2011). The Paleogene volcanic succession has an areal extent of at least 120000 km² and is locally up to 5 km thick. The absorption and scattering of acoustic signals by the volcanic facies has generally led to poor intra- and sub-volcanic seismic imaging hindering exploration. Despite these challenges exploration has moved within the feather-edge of the volcanic region and has led to the intra- and supra-volcanic Rosebank and Cambo discoveries, respectively (Helland-Hansen 2009;Fielding et al. in press). The reservoir units primarily belong to the Colsay and Hildsay sandstones of the Paleocene-Eocene Flett Formation. Uncertainties still remain concerning reservoir architectures and migration pathways within intra-volcanic plays and this study presents the results of a 3D photogrammetric study of seismic-scale cliff sections of the Faroe Islands to address these issues. As the islands are part of the same volcanic succession and are situated ~160-190 km NW of the Rosebank and Cambo discoveries the sea cliffs should be regarded as analogous sections.

Geological Setting

The Faroe Islands are an exposed remnant of the Faroe Islands Basalt Group (FIBG) that locally to the islands rarely exceeds 5 km in thickness (Fig.1; Passey & Jolley 2009; Passey & Varming 2010). The FIBG is predominantly composed of subaerial tholeiitic basalt lava flows commonly separated by minor volcaniclastic units of pyroclastic and/or sedimentary origin. The volcanism is associated with the arrival of the putative Iceland plume that led to widespread volcanic activity during the Paleocene and Early Eocene and resulted in the eventual continental break-up between Greenland and Eurasia, culminating with the onset of seafloor spreading in magnetochron 24r (Saunders et al. 1997; Jolley & Bell 2002).

The Faroe Islands were dissected by glacial action during the Quaternary resulting in an archipelago composed of 18 main islands with a distinctive mountainous terrain. The islands are typically elongated with a strong NWSE trend and are separated by long, narrow fjords (Fig. 1). The islands are continually battered by the North Atlantic Current resulting in the undercutting of primarily western coastlines leading to the formation of near vertical sea cliffs hundreds of metres high. These inaccessible cliffs provide excellent vertical exposure through the lava field, but also laterally over many kilometres. The opposing sides of the cliffs are typically grass covered and continuous exposure is limited to stream sections.

Renewed mapping, in part of the stream sections, has produced a revised lithostratigraphic subdivision of the islands and 7 formations are recognised (Fig. 1; Passey & Jolley 2009). Only the upper ~3.4 km of the FIBG is exposed on the islands and is chiefly composed of the Beinisvørð, Malinstindur and Enni formations. The Beinisvørð Formation is characterised by laterally extensive simple lava flows decametres thick, whereas the Malinstindur Formation consists of compound lava flows of similar thicknesses, but composed of numerous overlapping and anastomosing lava flow lobes centimetres to several metres in thickness (Passey & Bell 2007). The Enni Formation is characterised by both simple and compound lava flows erupted from fissure systems and low-angle shield volcanoes, respectively (Passey & Bell 2007).

The lava flow-dominated formations are generally separated by laterally extensive volcaniclastic sedimentary sequences, commonly composed of reworked pyroclastic material and deposited by various fluvial and mass flow processes (e.g. Hvannhagi and Sneis formations; Fig. 1). Other minor lacustrine, fluvial and mass flow deposits are found throughout the volcanic pile and some of these are also laterally extensive over many tens of kilometres. These include, for example, the bedded volcaniclastic sandstones of the Kvívík Beds and the Argir Beds (Fig. 1; Passey 2009; Passey & Jolley 2009). Coal seams formed in lakes and swamps are preserved in, for example, the <15 m thick Prestfjall Formation (Fig. 1).

Methodology

The 3D photogrammetry method used in this study was developed at the Geological Survey of Denmark and Greenland (GEUS) and partly builds upon earlier work by Dueholm et al. (1993) and Dueholm & Olsen (1993). The method allows the acquisition of geological data from vertical and oblique aerial photographs, with a threedimensional overview of the outcrops, by using a 3D stereo-plotter coupled with stereo-mirror technology (Fig. 2).

Approximately 940 oblique photographs were collected around 9 islands over 2 days, but only 9 cliff sections were set-up for 3D analysis, providing a cumulative distance of ~20 km. The photographs were taken out of the open door of a Bell 412 helicopter hired from Atlantic Airways in May 2010. The photographs were taken with GPS coordinates with an overlap of 60-80% using a Canon EOS-1Ds Mark III digital camera with a 36 mm x 24 mm CMOS sensor with 21 megapixels and a 35 mm Canon lens. The exposure time was set to a minimum of 1/250 seconds in order to reduce image blur due to helicopter vibration and relative movement of the camera with respect to the cliff sections. During the photographing the helicopter flew close to the cliff sections (<800 m) and at a constant altitude along straight lines approximately parallel to the cliffs. The digitization of important geological features over several kilometres and the determination of strata thickness, strike direction and dip values were obtained from working on high resolution visualisation of the cliffs (Fig. 2). Although aerial photographs were not used, because of the prohibitive cost, the relative accuracy of mapping and measuring the geological features observed on the oblique photographs was down to the decimetre scale. All the mapped features are stored in a GIS database and 3D polylines can be exported as shape files suitable for 3D modelling using, for example, Petrel reservoir engineering software. Moreover, using 3D feature databases in ArcGIS, geological cross-sections can be generated automatically to obtain real representations of outcrops where the relative accuracy is as high as the resolution on the photographs, and then projects onto 2D profiles.

Reservoir Architectures

This study concentrated on the cliff sections composed of the Enni Formation as it consists of different lava flow facies (i.e. compound v. simple flows) and as volcanism was waning during this period, the development of interlava units (e.g. volcaniclastic sandstones) was more widespread (e.g. Argir Beds). The heterogeneity of the cliff sections make tracing individual flows and interlava units more straightforward. Although the reservoir properties of the interlava units in the Enni Formation may be limited by low permeabilities, due to high proportion of volcanic material, they are still relevant as analogues to the high quality, interlava siliciclastic reservoir sandstones offshore as they are both largely governed by the same mechanisms controlling their geometries, lateral extents, compartmentalisation and interactions with the confining lava flows. Virtual videos compiled from the oblique photographs clearly demonstrate many of these features.

The lateral extent of the interlava units generally exceeds the extent of the mapped cliff sections, which range in length from 1.7 to 7 km. The Argir Beds are known to extend for more than 50 km between islands. The interlava units are generally <3 m thick, except for the Argir Beds that ranges from 1 to 6 m in thickness. The interlava units typically maintain constant thicknesses on the planar or slightly undulating relief, although commonly thickening in depressions, occasionally doubling in thickness. The Argir Beds is locally observed onlapping and wedging-out against highly eroded lava flow surfaces. The boundaries between the interlava units and the overlying lava flows is generally sharp and planar suggesting the passive and non-destructive emplacement of the lava flows.

Connectivity between interlava units is inferred where they are seen to merge laterally due to the wedging out of intervening lava flows. This is well exposed along the southern coastline of Svínoy, where 3 lava flows up to 20m thick each wedge out towards the NE, resulting in 4 interlava units merging over a distance of ~3 km. The merged interlava units are at least as thick as the sum of the constituent individual units, and consequently there are 4 times more interlava units in the SW as opposed to the NE. On Skúgvoy, the Argir Beds are separated into lower and upper parts by at least 5 laterally discontinuous flow lobes emplaced along the same horizon and presumably represent the distal fingers of an advancing lava flow.

Major channel structures up to 180 m wide by 45 m deep are observed in the cliff sections on Skúgvoy, Fugloy (Fig. 3) and Koltur. The channels vary in shape and include “box-shaped” and asymmetrical types. The channels show complex, multiple phases of development and are filled with volcaniclastic conglomerates and sandstones as well as lava flows. Channel axis orientations are estimated to be approximately NNW-SSE and WNW-ESE for the main channels on Fugloy and Skúgvoy, respectively. The channels were most likely formed during short, but intense erosional events and the conglomerates were deposited by mass flows when erosion and sediment bypass ceased.

Compartmentalisation of the interlava units is inferred from a number of different scenarios, including inclined to near-vertical dykes up to 6 m wide that form lateral barriers across units. Interestingly, however, out of the 20 km of cliff sections studied only 3 dykes were observed. The channels described above may also act as barriers or conduits for fluid flow between interlava units depending on the channel fill. A lava tube orientated NNW-SSE and with a diameter of ~50 m is also observed intersecting an interlava unit as well as having an apophysis of lava extending upwards and intersecting overlying interlava units. Lastly, a NNW-SSE trending reverse fault, dipping ~20° to the W, is observed transecting the southern coastline of Svínoy. The fault has a vertical displacement of ~6 m resulting in the disconnection of individual interlava units. The fault may have acted as a conduit for fluid migration or conversely as a seal.

Conclusion

The 3D photogrammetric study of cliff sections of the Faroe Islands has revealed and quantified the lateral extent, geometry, connectivity and compartmentalisation of interlava units as an analogue to intra-volcanic plays. The interlava units may be better developed and consist of siliciclastic facies at the leading edge of the lava field, but the relationships between the sedimentary units and the lava flows should be largely similar. The recognition of major channel complexes confirms for the first time that periods of erosion and sediment bypass occurred during the development of the Faroese lava field and sediment was most likely transported offshore towards the Faroe-Shetland Basin. Data generated during this study can be used in Petrel reservoir engineering software.