Geotechnical Investigation for the Scott Base Redevelopment Project
Introduction
The Scott Base Redevelopment Project (SBR) aims to modernise the facilities at Antarctica New Zealand’s Scott Base with extensive reconstruction of most of the current infrastructure. Golder was engaged to provide a geotechnical assessment to support detailed design of the proposed redevelopment. This article summarises the findings of two phases of geotechnical investigation and several site visits completed by Golder staff since 2015.
Scott Base currently comprises seven primary buildings interconnected by all-weather corridors and located over approximately 300 m x 200 m on Pram Point, Ross Island, in the Ross Dependency of Antarctica. The original buildings at Scott Base were constructed during the late 1950s as part of the International Geophysical Year. Additional buildings have been added over the subsequent decades to accommodate expansion of the base to support further scientific programmes.
The Scott Base Redevelopment (SBR) project will be the largest project ever undertaken by Antarctica New Zealand. It is a complex multi-year project that will significantly underpin New Zealand’s research programme and strategic interests in Antarctica. The purpose of the rebuild is to fully address known risks and issues with the current infrastructure and provide a fit for purpose, environmentally sustainable research base that will support New Zealand’s physical presence in the Ross Sea region of Antarctica.
Geological Setting
Ross Island is formed from volcanic rocks of the McMurdo Volcanic Group, comprising basaltic rocks including lava flows and scoria deposits. Hut Point Peninsula is a southward extension of Ross Island and, at its southern end, is largely free of persistent thick ice. It is the site of the 1901-1904 Discovery Expedition’s hut and is the site of McMurdo Station (USA) as well as Scott Base. Most of the volcanic cones on Hut Point Peninsula are centered on a prominent lineation that strikes NNE along the western side of the peninsula. The volcanic rocks are predominantly Quaternary age basalts, with minor flow banded trachyte at Observation Hill and autoclastic and pyroclastic breccias at Castle Rock and Boulder Cones (Cole et al. 1971) (Figure 1). Scott Base is located on the south eastern flanks of Crater Hill, a parasitic volcano on the south side of Hut Point Peninsula. Less than 40 km from the site is Mt Erebus, which is the largest mountain on Ross Island at 3,794 m above sea level and is one of the few continuously active volcanoes on Earth. The volcanic rocks near Scott Base are geologically young, with the youngest estimated to be approximately 50,000 years old (Cox et al. 2012).
Figure 1: Geological Map of the Hut Point Peninsula adapted after Cole et al. 1971.
Scott Base is constructed on a gently sloping site 15 m to 30 m above sea level. The adjacent coastline is approximately 60 m from the existing main buildings and 30 m from the nearest out-buildings. The sloping surface on which Scott Base is located is interpreted to be the uppermost of a series of basaltic lava flows. The flows slope down from Crater Hill, located 1500 m northwest of Scott Base, which reaches an elevation of about 240 m above sea level. The stratigraphy is exposed in cross section in adjacent slopes, and typically comprises basalt flows estimated to be 1 to 10 m thick. The lava flows appear to dip at a flatter angle than the general ground surface, suggesting that the upper, youngest rocks, sourced from Crater Hill, thin downslope away from the vent (Figure 2). The extent to which glacial activity has shaped the landscape is unknown. The surface of the slope has been slightly eroded by some surface water channels.
Figure 2: Scott Base geomorphic setting viewed from McMurdo Station Road west of Scott Base.
On the steeper ground above the McMurdo Station Road, permafrost deformation features (‘patterned ground’) can be seen in aerial images and on the ground. These features relate to expansion and contraction of the near surface ground due to seasonal freeze and thaw (Figure 3).
Figure 3: Permafrost deformation features around the road between Scott Base and McMurdo Station.
Geotechnical Investigations
Twenty-five drill holes were completed on site around Scott Base in October 2015 and October 2019 (Figure 4) by Webster Drilling and Exploration Limited (Webster), using a purpose built air flushed coring rig configured to recover continuous 55 mm diameter core (Figure 5).
Figure 4: Drill hole investigations for 2019 (black and white circles) and 2015 (blue circles).
Figure 5: Webster’s air flushed coring rig configured to recover continuous 55 mm diameter core located on the slope above Scott Base.
A key aim of the site investigations was to recover the best possible samples of frozen ground including ice, frozen soil and rock. Webster’s rig used radiators to chill the compressed air as cold as possible before flushing the bore hole annulus. This method proved highly effective and recovered high quality samples with very little core loss. It was observed that melting of the ice in the recovered core increased when ambient air temperatures reached above a threshold of about -7°C. Golder supervised the drilling of all drill holes and undertook the logging onsite as the holes progressed. Thermistors were installed in three of the drill holes (BH-01, BH-02, BH-03) to measure variations in ground temperature at a range of depths throughout the year and to assist in determining the depth zone that could be affected by seasonal freeze and thaw (as discussed later in this paper).
A SeeSnake microReel video inspection system was utilized to produce high quality images of the borehole walls. The imagery identified sections of the drillholes that were highly fractured, small to large voids and sharp changes in geology (Figure 6).
Figure 6: Image from the video inspection of a borehole on the slope above Scott Base. Image shows horizontal and vertical discontinuities and a small void.
Geophysical techniques were used to investigate the subsurface, in particular investigating the potential presence of any significant ice that could affect the redevelopment. Six Multi-channel Analysis of Surface Waves (MASW) profiles were completed around the base (Figure 7). The MASW lines that were completed along the shoreline are interpreted to measure shear wave velocity of 2000 to 2300 m/s in the upper ground profile. This approximates the shear wave velocity of ice suggesting that the stratigraphy contains abundant ice. The MASW lines that were completed further from the shoreline indicate lower shear wave velocity (800 to 1400 m/s) suggesting the presence of highly fractured rock with little ice.
Figure 7: MASW layout on site
Ground penetrating radar (GPR) was also used to investigate for subsurface ice (Figure 8). The GPR
used at Scott Base was a USRadar Quantum Q4-300 that has three radar antenna with a frequency range of 220 MHz, 300 MHz and 380GHz. The interpretated GPR data identified areas where fill had been placed to develop roads or building platforms underlain by basalt.
Figure 8: GPR set up
Geotechnical Test Results
Point Load Tests
To characterise rock and frozen soil strength 391 Point Load Tests (PLT) of core samples were undertaken in the field to evaluate point load strength index (Is(50)). The results show that the ice bonded sandy gravel soils have a point load strength index value approximately equivalent to a weak rock. The basaltic scoria and basalt are predominately strong to very strong, which correlates to an approximate unconfined uniaxial compressive strength of 50 to 250 MPa. All Is(50) results are shown in Figure 9.
Figure 9: Point Load Strength Test (Is(50) results of frozen soil and rock samples from Scott Base.
Unconfined Compressive Strength
Unconfined compressive strength (UCS) tests were completed on 21 core samples. The samples were selected to represent the range of rock materials encountered on site. We selected 4 samples for each rock type; non-vesicular basalt, slightly vesicular basalt, moderately vesicular basalt, highly vesicular basalt and scoria. The non-vesicular basalt and slightly vesicular basalt returned UCS results between 60-120 MPa, moderately vesicular basalt ranged between 50-100 MPa, highly vesicular basalt and scoria ranged between 8-40 MPa. These results indicate a correlation of reducing compressive strength with increasing vesicularity. UCS was plotted against the closest PLT results taken adjacent to the UCS samples to estimate a conversion factor between UCS and Is(50) to help represent the Is(50) results as inferred UCS. Figure 10 graphically summarises the UCS and Is(50) results and illustrates that a conversion factor of 17 is suitable.
Moisture Content and Density
All the samples that underwent strength testing had moisture content measured along with bulk and dry density. The moisture content varied between 0.4 to 10.3 % with the moisture content generally increasing with vesicularity. The samples ranged from 1.9 t/m3 to 2.91 t/m3 for bulk density and 1.76 t/m3 to 2.89 t/m3 for dry density. As expected, the density in both cases decreased with increasing vesicularity.
Figure 10: Unconfined Compressive Strength and Is50 results for a range of vesicle content in basalt samples from Scott Base.
General Ground Conditions
The subsurface geology at Scott Base generally comprises fill, scoria, basalt fragments in ice matrix and basalt bedrock (Figure 12). Ice is also evident in much of the core either as visible ice or inferred from ice bonded material where the ice is not visible to the unaided eye.
Fill
Uncontrolled fill was encountered in topographical lows near the shoreline and adjacent to buildings or roads. The fill is typically light brown in colour and comprises basalt gravel, silt and buried snow. The fill is typically well bonded with ice. Although significant ice lenses are uncommon within the fill material itself, the base of the fill is typically marked by thin ice lenses or buried snow (Figure 11).
Figure 11: Evidence of man-made ramps, roads and pads behind Scott Base with drilling rig to the right of the image. The dozing of snow into piles and tracking on roads shows how snow and ice can be incorporated into the fill (photograph supplied by Antarctica New Zealand).
Figure 12: Core run on site showing the different geology encountered. Ground surface is at the top right of the core box which is a layer of fresh snow, the upper 0.5 m is ice underlain by a moderately weathered basalt with moderate vesicularity moving into a slightly weathered basalt. The intact piece of core is a 1.5 m fully intact core run in the slightly weathered basalt. The core is fully covered in ice along all discontinuities and defects.
Basaltic Volcanics
The natural ground at Scott Base is predominantly composed of basaltic rocks of varying vesicularity, ranging from dense basalt to scoria with varying ice content. We have divided the recovered core into five different classes of vesicularity to help characterise the engineering properties of the material based on visual assessment of the relative proportion of vesicles: non vesicular basalt (0-10 % vesicles), slightly vesicular basalt (10-20 % vesicles), moderately vesicular basalt (20-30 % vesicles), highly vesicular basalt (30-40 % vesicles) and scoria (<50 % vesicles). Each class of basalt has been characterised below with a description of the geology, the range of UCS and Is(50), the Geological Strength Index (GSI) (Hoek, 1994) and Hoek Brown criterion (Hoek & Brown, 1980) values that can be used to determine the shear strength parameters as required. This summary is intended to be used to characterise the rock for foundation design and other engineering design requirements. A summary of the engineering properties for the basaltic volcanics is present in Table 1.
Table 1: Engineering Properties for the Basaltic Volcanic Material.
Ice
Ice was encountered to some degree in all the drill holes across Scott Base. The presence of ice, when not visible, was inferred by bonding of soil grains. Ice bonding ranging from friable (i.e., easily disaggregated when scraped with the point of rock hammer) to well-bonded (i.e., requiring a firm blow of the rock hammer to break). Where visible, ice ranging from sub-millimetre sized discrete crystals disseminated in the core to lenses of ice with no other soil or rock constituents. Ice also ranged from cloudy, white and granular ice to clear, hard and glassy ice. Very thick accumulations of white, granular ice are interpreted to be buried snow underlying fill. In general, the thickest (0.8 m) accumulations of ice were observed near the shoreline.
Ground Conditions Across the Redevelopment Site
The proposed Scott Base Redevelopment comprises three large buildings that are connected via hallways. The three buildings will be accommodation (Building A), science labs, offices and event staging (Building B) and store, cargo and workshops (Building C). The proposed building locations are upslope from the current Scott Base buildings (Figure 13). The ground profile at the location of the buildings will be formed by cut and fill to level off the building platforms.
The existing Scott Base is constructed on a gently sloping site interpreted to be the uppermost of a series of basaltic lava flows sloping gently to the southeast (Figure 2). Investigations determined that the properties of a basaltic rock changed across the site. The drillholes located on the upper slopes behind the current Scott Base buildings (Figure 13) encountered thaw sensitive sandy gravel (broken fragments of scoria/basalt) and weak, slightly weathered basalt/scoria with ice present in the vesicles or in an ice matrix.
Figure 13: Proposed location of the redeveloped Scott Base buildings (supplied by Antarctica New Zealand).
Further down slope behind the current buildings the drill holes encountered similar material (weak to slightly weathered basalt/scoria with ice present), but lenses of moderately strong to very strong fine grained unweathered basalt were also present. These lenses of more competent basalt were more common further downslope towards the shoreline in varying thickness. The more competent basalt generally resulted in reduced drilling speed and improved core recovery. The ground conditions at the shoreline were found to comprise up to 1.5 m of fill consisting of sandy gravel, compacted snow and buried wood fragments underlain by natural ground (Figure 14).
The underlying natural ground was found to comprise strong basalt and weaker slightly weathered scoria. The core was filled with glassy, hard, clear to cloudy ice crystals in all vesicles and on discontinuities. We conclude that the ice content near the shoreline reflects the impact of seawater infiltrating the rock mass.
We infer from the distribution of materials that the upper slope above Scott Base may comprise local material sourced from Crater Hill and that the underlying material that outcrops closer to the shoreline may comprise a series of extensive flows, possibly from a separate source.
Figure 14: Drilling upslope of Scott Base (in the background). The material was predominately recovered as basaltic gravels with ice present (photograph supplied by Antarctica New Zealand)
Figure 15: Drilling near the shoreline and outlet pipe for Scott Base. The start of the pressure ridges are forming in the background (photograph supplied by Antarctica New Zealand).
Active Layer Affected by Permafrost Processes
Three thermistor strings were installed during the 2015/2016 drilling season at Scott Base. The thermistor strings were installed in drillholes close to the existing base. Figure 16 shows the highest and lowest temperatures measured at each thermistor depth.
Figure 16: Summary of available temperature data from Scott Base.
The data collected between 30 January 2016 and 10 February 2020, presented in Figure 16, indicates that only shallow thermistors measure temperatures approaching 0 °C. We infer that the active layer that is affected by seasonal freeze and thaw processes extends to a depth of about 0.4 m below ground level at Scott Base.
Conclusions
Golder has completed two geotechnical investigation programmes at Scott Base in 2015 and 2019 for Antarctica New Zealand. During these investigations we have completed multiple drill holes, drill hole imagery, MASW, GPR and installed thermistors.
From these investigations we determined the site is underlain by man-made fill, basaltic volcanic rocks and ice. The fill is located to areas of previous construction, i.e. ramps and roads or building platforms. Ice is present within or under the fill and interpreted as buried compacted snow, or as ice crystals on defects in the basalt. There is no evidence of significant continuous ice lenses. The basaltic volcanic rock encountered was divided into 5 sub-categories based of the vesicularity of the material, non-vesicular basalt, slightly vesicular basalt, moderately vesicular basalt, highly vesicular basalt, and scoria. Each of these categories was found to have different engineering properties, which will influence engineering behaviour.
We infer that the more vesicular the basalt present on the slope above Scott Base may be local deposits sourced from Crater Hill and the more competent basalt present below Scott Base is part of a more extensive deposit, possibly from a different volcanic source.
References
- ASTM D4083-89 (2007), Standard Practice for Description of Frozen Soils (Visual-Manual Procedure), ASTM International, West Conshohocken, PA, 2007, www.astm.org.
- Bannister, S., F. J. Davey, B. Kennett, (2003), Variations in Crustal structure across the transition from West to East Antarctica, Southern Victoria Land, International Journal of Geophysics, Volume 155, pages 870-884.
- Cole, J.W., Kyle, P.R., and Neall, V.E., 1971: Contributions to Quaternary Geology of Cape Crozier, White Island and Hut Point Peninsula, McMurdo Sound Region, Antarctica. New Zealand Journal of Geology and geophysics, vol. 14 no. 3, 528 – 546.
- Cox, S., Turnbull, I., Isaac, M., Townsend, D., Smith Lyttle, B. (2012). Geology of southern Victoria Land, Antarctica. IGNS 1:250 000 geological map 22.
- Hoek, E. (1994). Strength or rock and rock masses. International Society of Rock Mechanics and Rock Engineering (ISRM) News Journal, 2 (2): 4-16.
- Hoek, E., Brown E,T., (1980) Empirical Strength Criterion for Rock Masses. Journal of geotechnical and Geoenvironmental Engineering. Vol 109, Number GT9, Pp. 1013-1035.
- Golder Associate (NZ) Limited (2016). Scott Base – Antarctica Geotechnical Assessment Report. Report Number 1413806_7407-003-R-Rev0, dated February 2016.
- Golder Associates (NZ) Limited (2020). Scott Base Redevelopment Project Geotechnical Assessment Report. Report Number 19123189_7407-004-R-Rev0, dated July 2020.
- New Zealand Standard (NZS) 4407:2015. Methods of sampling and testing road aggregates. ISBN 978-1-77551-962-1.
- NZGS, (2005). Field Description of Soil and Rock, New Zealand Geotechnical Society, December 2005.