Feasibility Report of a Deep Geothermal Single Well, Aberdeen Exhibition and Conference Centre

Geology at The Site

General Geological Setting

A review of the available drift and solid British Geological Survey (BGS) geological maps has been undertaken. The site falls within drift geology map NJ81/91 (1:25,000)^[4] and solid geology map and NJ81SE (1:10,560)^[5].

Superficial Geological Map

The superficial geology map^[1] indicates the site to be predominantly underlain by till described as:

"Mainly yellowish brown sandy diamicton with clasts of Dalradian metamorphic and Caledonian igneous rocks, derived locally or from a short distance to the west, Matrix weather in part and containing much decomposed (grussified) rock."

Some areas in the central and southern part of the site are shown to be underlain by alluvium described as: Gravel and sand, capped by silty and clayey floodplain deposits. Based on the fact that the location of these areas is in close alignment with the existing water courses on the site it is thought that the indicated alluvium is associated with material carried by the Green Burn, Gough Burn and the East Craibstone Burn. Some localised areas between the current Rowett Institute and the south-western boundary of the site are indicated to be underlain by made ground. The key indicates the made ground to be: Mainly spoil associated with sand and gravel workings.

There are localised areas in the southern part of the site are shown to be underlain by glacio-fluvial ice contact deposits - moundy deposits of sand and gravel. One of these areas is indicated to underlay a large area of the Rowett Institute.

Solid Geological Map

The solid geological map^[2] for this area shows that the solid geology underlying the entire site comprises of Vein Complexes and Plutonic Igneous Rocks of the Aberdeen and Kemnay Plutons with foliated biotite-muscovite-granite. An exposure approximately 1km to the north-west of the site describes the granite to be "well foliated medium grained granite - sometimes biotitic schlieren".

The depth of the underlying rock is not indicated. However, the superficial geology map indicates bedrock to be near or at the surface in areas approximately 0.7km to north, 0.5km to south east and 1.0km to east. The bedrock at or near the surface is described to be: Locally weathered to sand or micaceous silty sand (gruss) to depths >2m, especially in the west (of the area covered in the geology map). A solid geological boundary with Metamorphic Psammite of the Argyll Group is identified approximately 1km to the west of the site.

Hyrdology and Hydrogeology

A number of surface watercourses cross the site. These include the Green Burn, Gough Burn and East Craibstone Burn. These watercourses flow eastwards into the River Don, which is located approximately 600m to the east of the site and is identified by SEPA to be of moderate quality. The most recent groundwater vulnerability map produced by SEPA^[6] shows that the area in which the site is located is classified as vulnerability class 2, but classifications 1 and 3 are also within close vicinity of the site. Using the accompanying report "Vulnerability of Groundwater in the uppermost aquifer"^[7], the site can be characterised as "Vulnerable to some pollutants, but only when they are continuously discharged/leached".

SEPA have also produced maps of the superficial and bedrock aquifer Productivity^[4]. The superficial aquifer productivity map indicates that the site may have areas of very low to high productivity and that the flow is inter-granular. The solid aquifer productivity map shows that the bedrock has a very low productivity with fracture flow. An Indicative River & Coastal Flood Map, produced by SEPA^[8], classifies areas along the Green Burn as "area at risk of flooding from rivers". The current groundwater vulnerability map has been reviewed and ^[9] represents the vertical pathway of groundwater through strata overlaying an aquifer. Groundwater vulnerability is defined as the tendency and likelihood for general contaminants to reach the water table after introduction at the ground surface. The most recent vulnerability map indicates that the site is generally located in an area classed as moderately permeable, with a geological classification as follows: Minor or Moderately Permeable Aquifer - Fractured or potentially fractured rocks which do not have a high primary permeability or other formations of variable permeability.

Previous Engineering Ground Investigations (AECC Preliminary Ground Investigation 2014)

In 2014, Arup were appointed to perform a preliminary geotechnical ground investigation at the proposed AECC site in Aberdeen^[10]. The intrusive investigations performed by Raeburn Drilling & Geotechnical Ltd. encountered the following ground conditions (Table 1) at the site included rotary cored boreholes.

Table 1: Stratigraphy of the site

Stratum Description	Depth to top of stratum (m)	Depth to base of stratum (m)	Level of the top of stratum (mAOD)	Level of the base of stratum (mAOD)	Proven thickness (m)
Topsoil	0	0.1 - 0.5	48.64 - 77.17	48.34 - 76.87	0.1 - 0.5
Made Grounda	0 - 0.5	0.7 - 3.4b	52.73 - 62.56	51.98 - 61.96	0.6 - 3.3
Possible Alluviumc	0.1-0.3	0.7-0.75	51.20 - 58.43	57.98 - 50.70	0.45 - 0.6
Glacial Deposits	0.1 - 2.6	7.7 - 21.1d	48.37 - 76.87	35.62 - 53.52d	7.25 - 20.6d
Bedrock	7.25 - 20.6	Proven to between 12.2 and 30.1mbgl	35.62 - 53.49	Proven to between 26.76 and 49.42mOD	Proven to between 4.1 and 9.7m

Note: Hand dug trial pits have not been considered in the above assessment.

^a Made ground was only encountered in exploratory holes BH07, BH08, TP13-TP16,TP19-TP20 as well as a number of the hand dug trial pits.

^b The thickness of the made ground was not proven in TP14 and TP20, where made ground was recorded to the base of pits, 3.0 and 3.4mbgl respectively.

^c Possible alluvium was only encountered in BH08A, TP12 and TT02.

^d Base of the glacial deposits were only proven in boreholes BH02, BH02A, BH05, BH05A, BH06C, BH08 and BH09. Only these boreholes have been used in the assessment of the glacial deposit base and thickness.

Description of Bedrock Geology

Bedrock was encountered in all of the exploratory positions where rotary drilling was undertaken. The bedrock comprised of granite in all of the exploratory locations. The granite was typically described as distinctly weathered and often recovered as non-intact (sand and gravel).

Rotary coring of the granite was undertaken using triple tube techniques, however; in a number of boreholes the recovery of the granite proved poor (particularly the upper few metres) likely due to the heavily weathered nature of the rock. Coring between 4.1 and 9.7m of the granite bedrock was attempted in the boreholes. The weathering of the bedrock was noted to decrease with depth in some of the boreholes.

Typical descriptions of the granite included:

"Very weak massive coarse grained brownish orange GRANITE. Granite is distinctively weathered with a penetrative staining ~3mm throughout rock, clay on fracture surfaces and loss of strength. Rock is weathered to a coarse gravel in places. There are 2 fracture sets. #1 closely spaced, sub-horizontal, rough, planar. #2 closely spaced, dipping 70°-90°, rough, planar" (BH02A)

"Very weak and weak massive speckled yellowish off white GRANITE. Distinctly weathered evident as clay smears on fracture surfaces, yellow discolouration and loss of strength throughout. There are 3 fracture sets. #1 medium spaced, sub-horizontal, planar and rough. #2 close and medium spaced, dipping at ~45°, planar and rough. #3 closely spaced, dipping at ~10°, stepped and rough." (BH05)

"Very weak to strong massive speckled black and white GRANITE. Distinctly weathered evident as yellowish brown discolouration penetrating up to 5mm, clay smears on fracture surfaces and loss of strength throughout. There are 3 fracture sets. #1 close and medium spaced, dipping at ~45°, planar and rough. #2 medium spaced, dipping at ~70°-90°, stepped and rough. #3 very close to medium spaced, dipping at ~ 50°, planar and rough. Locally recovered non-intact where fracture sets meet. Weathering decreases with depth" (BH09).

Engineering tests have been performed on samples of the cores obtained from the 2014 investigations. A review of the laboratory strength test results indicate that the corrected point load index, Is(50), values typically range between 0.02kN/m² and 3.04kN/m². A calibration exercise was carried out to select an appropriate factor for the Is(50) to unconfined compressive strength (UCS) correlation factor and 10 was found to be an appropriate value. The results would indicate the strength of the granite to be between 0.2kN/m² and 30.4kN/m²; thus indicating the granite to be below 'extremely weak' to 'medium strong'. Average Is(50) values indicate that the granite is typically 'weak'.

UCS tests were also undertaken on samples of granite and indicate granite to vary between 'very weak' to 'medium strong' and to be 'weak' on average. There appears to be some correlation with the strength of the granite and depth, however, extremely weak granite was also encountered well below the rockhead.

Table 2: Summary of Measured Geotechnical Properties - Granite

Geotechnical Property	Unit	No. of tests	Range	Mean
Point Load Index	Is50	37 Axial 38 Diametral 12 Lump	0.02 - 3.04 0.02 - 2.61 0.05 - 0.52	0.73 0.63 0.25
Derived UCS (converted from Is50 using a factor of 10) [18]	MPa	37 Axial 38 Diametral 12 Lump	0.2 - 30.4 0.2 - 26.1 0.5 - 5.2	7.3 6.3 2.5
Uniaxial Compressive Strength	MPa	10	2.2 - 26.5	15.6
pH	-	10	7.6 - 8.5	8.1
Sulphate (2:1 extract)	g/l	10	<0.01 - 0.04	0.015*

*Tests recording '<0.01' of sulphate were assumed as 0.01g/l.

Aberdeen Granite Composition and Expected Geology to 2kms

The Aberdeen granite is a substantial pluton-shaped body approximately 16km in its longest dimension and 6km in its shortest with the long axis aligned approximately NNW-SSE (Figure 2). It was emplaced approximately 470 million years ago, based on ages obtained from monazites in samples from the Rubislaw quarry and dated by the U-Pb method (Kneller & Aftalion, 1987)^[11]. Textural relationships indicate that the granite was emplaced into Dalradian host rocks at peak temperatures of circa 550°C and a pressure of 0.50-0.55 GPa (Harte & Hudson, 1979.)^[12]. This implies that the present surface of the granite was located some 17-20 km below the surface at the time of metamorphism and magmatism. Such deep emplacement is also implied by the lack of a thermal metamorphic aureole (Munro, 1986)^[13]. The age of the granite, along with geochemical characteristics, indicate that the pluton does not belong to the younger Cairngorm suite of Newer granites of late Silurian-early Devonian age (Stephens & Halliday, 1984)^[14] that were the focus of study in the 1970s and 80s for their geothermal potential (Downing & Gray, 1986)^[15].

Figure 2: The Aberdeen Granite Pluton

Whole rock geochemistry, including stable and radiogenic isotopes strongly implicates metasedimentary material similar to the host Dalradian Supergroup rocks as the dominant source material for these granites (Halliday et al., 1979^[16], Kay, 1980^[17]). This is consistent with several other characteristics typical of S-type granites (Chappell & White, 2001^[18]).

Form of the Aberdeen Granite

The outer contacts of the granite pluton are generally obscure, largely because outcrop is poor. The available evidence indicates inter-digitation with the local biotite gneisses, implying that a plexus of granite sheets petrographically similar to the pluton was injected into metasediments adjacent to the contact. These sheets are often some 10-15m thick and tend to be relatively flat lying (Munro, 1986). Although no evidence has been found to constrain the dip of the plutonic granite-metasedimentary contact at any locality, it is relevant that, on a larger scale, the contacts cut across all lithological units and structures suggesting that the interior of the body at least has the form of a single pluton with relatively steep outer contacts inside the sheeted margins. This is supported by the relative homogeneity of the pluton interior compared with the more nebulitic margins. The Rowett site for the new AECC is closer to the centre of the granite than the outer contact of the Aberdeen granite and at depths of around 2km it is most likely that the site will be underlain by broadly similar granite belonging to the same pluton.

Some granite plutons in the eastern Grampian Highlands, including Cairngorm, Lochnagar, Glen Gairn and Mount Battock, are associated with a large negative Bouguer gravity anomaly bounded by the -40 mgal contour) stretching from Cairngorm in the west for about 80 km eastwards (BGS, 2007^[19], Rollin, 1984^[20], Wheildon et al., 1984^[21]). This substantial anomaly is modelled as a single batholith extending to depths of at least 10 km. The Bouguer anomaly tails off as it approaches Aberdeen but small kinks in the regional trend may reflect the presence of a weak anomaly associated with the Aberdeen granite. Density measurements give means of 2800 ±157 and 2740 ±99 kg m-3 (both 2 sigma) respectively for regional Dalradian and Moine rocks (Richardson & Powell, 1976^[22]). The Dalradian mean density value falls within the 2650-2950 kg m-3 range (mean=2740) previously determined for 80 specimens of Dalradian rocks in Aberdeenshire (McGregor & Wilson, 1967^[23]). The density of the Aberdeen granite falls within the same range, values varying from 2673 to 2688 (mean=2678 kg m-3, standard deviation of 7). This overlap in the density of the granite and its host rocks is consistent with the widely accepted view that similar rocks to the regional Dalradian were the source of much of the Aberdeen pluton however this lack of density contrast weakens the usefulness of gravity modelling for predicting the structure of the pluton at depth.

Structural Features

Structural features in the granite may facilitate heat replenishment at the base of the column by facilitating the advective transfer of fluids. The potential geothermal roles of foliations, joints and faults in the Aberdeen granite are described below.

The Aberdeen granite is variously foliated. A strong foliation, defined principally by the planar alignment of biotite and feldspar crystals, has a tendency to follow the orientation of the local country rocks at the contact but on a larger scale this foliation is discordant with the host formation. This evidence along with petrographic textures suggests that this mineral foliation is a primary magmatic feature formed during consolidation of the pluton (Munro, 1986). The dip of the primary foliation varies from 20° to vertical in outcrops relatively close to the Rowett site. Post-consolidation recrystallisation has affected the magmatic foliation as evidenced by textural features and crystallographic evidence of strain in quartz grains. This recrystallisation is not pervasive; recrystallized and unaffected material may occur in the same outcrop and even sometimes in the same thin section (Munro, 1986). In thin section neither of these foliations appear to create open pathways suitable for fluid flow but conditions may be more suitable for opening up these foliations at 2km depth, perhaps through the alteration of micas. Even so, the probability of high permeabilities created in this way is likely to be relatively low.

Joint planes are ubiquitous in granites such as that found at Aberdeen. The only site where these have been investigated formally is at Perseley Quarry in an early study of joint patterns (Cameron, 1945^[24]). Here, the principal joint set runs parallel to the primary foliation (described above) that at this locality is essentially vertical. From a quarry in the neighbouring Kemnay granite pluton, Cameron identified a strong horizontal joint set that he attributed to the erosion of overlying rocks. Descriptions of bedrock from shallow rotary drill cores at the Rowett site include the recognition of both sub-vertical and sub-horizontal fracture sets suggesting that these joint orientations may be common throughout the pluton and the vertical set may be anticipated at 2km depth. However, if the horizontal set is related to roof unburdening then it may not be well developed at depth. If such a sub-vertical joint set were well developed at 2km depth below the Rowett site these joints might facilitate vertical upward flow of warmer fluids from depth. Drilling a 2km vertical well at the site would represent an excellent opportunity to test this theory and the presence of vertical upflow and will substantially increase the thermal output of the well.

Major faults are few, similarly very few small faults have been mapped over the granite pluton. The only fault of importance is the post-Devonian Dee Fault that defines the southern border of the Aberdeen granite but this is some distance from the Rowett site.

Petrology

The principal granite facies over the whole Aberdeen pluton is grey muscovite biotite granite which occasionally appears pink or red when associated with alteration and joint planes. It is normally medium grained (1-3mm) and equigranular (Figure 3a). Variation in facies is usually related to coarser grained varieties (sometimes pegmatites) and banded schlieren (Munro, 1986). Enclaves are common, most typically these are metasedimentary xenoliths likely to be fragments of country rocks, while only occasional samples are mafic enclaves of more primitive origin.

Quartz is particularly abundant in these granites (33-36%) with plagioclase in the same range and somewhat less alkali feldspar (20-23%) while the micas amount to 6-11% with biotite predominating over muscovite (Munro, 1986). The rocks classify as true granites or granodiorites. Plagioclase shows normal and oscillatory zoning with albitic margins and the grains are often altered. Alkali feldspar usually shows patches of microcline twinning (Figure 3a) and large quartz grains usually take the form late interstitial crystals (Figure 3b) while small grains of quartz are found in myrmekites and as inclusions in feldspars. Ragged brown biotite grains that often define the primary foliation are often associated with later muscovite (Figure 3c).

Figure 3: (a), (b), (c): Petrology of the Aberdeen Granite Pluton

Accessory minerals are important in the context of the geothermal potential of this pluton as they are the principal repositories of uranium and thorium that undergo radioactive decay and generate heat as a by-product. A very early study recognised that monazite, a thorium-bearing cerium phosphate, is more abundant than zircon in the Rubislaw quarry outcrop, zircon being an uranium-bearing zirconium silicate (Mackie, 1926^[25]). Accessory phases are often trapped within biotite crystals that accumulate the damage from alpha-particle emissions in the form of dark haloes around microscopic inclusions. Both zircons and monazites display this effect and Figure 3c shows a highly birefringent zircon enclosed within a dark halo trapped in a brown biotite crystal among many other similar haloes. The abundance of these radiation-generated haloes in the Aberdeen granite is at least as great as in any other major granite pluton in Scotland.

Whole Rock Geochemistry

The weathered and fractured material in the surrounding area and from the on-site shallow boreholes is unsuitable for petrological and geochemical investigations. Consequently, fresh granite samples from surrounding quarries are used. Four samples were obtained from the Dancingcairns and four samples from the Dyce quarries some 2.4km SE and 3.7km NNW of the Rowett site respectively. Six further samples were taken from the best known location of the Aberdeen granite at the abandoned Rubislaw quarries located 5.8km to the SSE. Examination of these 14 samples taken from a wide area of the pluton confirms their broad similarity and supports their use as unaltered representatives of the granite body underlying the Rowett site.

Deep Temperature Profile

An important factor in determining the target depth of the single well to be drilled at the AECC site is the expected temperature of the rock and fluid. As with all geothermal projects, the temperature cannot be known for certain until drilling to the target depth has been completed. Without sub-surface data, the only way to estimate the temperature is to model the change in temperature from surface to target depth using thermal modelling methods based on measurements and estimates of the major parameters that influence temperature.

Following the energy crisis of the 1970s, a detailed evaluation of the UK's geothermal potential was undertaken and some deep sources of heat in granite bodies were identified, including a number of granites in NE Scotland exemplified by the Cairngorm granite (Downing & Gray, 1986). Four boreholes were drilled into granites at Cairngorm, Ballater, Bennachie and Mount Battock to depths of approximately 300m and data obtained from cores and surface outcrops were combined with data from the literature to form the basis of one-dimensional geothermal modelling studies (Lee, 1986^[26], Wheildon & Rollin, 1986^[27]). The modelling approach used in these reports is adopted in this study with the important difference of taking into account glacial cooling effects on estimates of temperature at depth. As the results of thermal modelling are very dependent on the values given to various parameters, the reasons for the values chosen is reviewed in the sections below.

Surface Heat Flow (Q₀)

Surface heat flow is estimated from the thermal gradient and thermal conductivity obtained from boreholes. In the case of NE Scotland, these boreholes are all shallow (≤300m) and subject to thermal perturbations due to the last ice sheet. No direct estimate of q₀ is available for the granite area of Aberdeen, although q₀ in Moine and Dalradian rocks hosting the granite are estimated to range from 40 to 55 mWm^-2 (Lee, 1986). A large swathe of granites stretching from near Aberdeen to Cairngorm has much higher heat flow with q₀ often exceeding 70 mWm^-2 (Lee, 1986) while a very low value of 29 mWm^-2 was obtained at Tilleydesk (Ellon) about 25 km north of the AECC site (Burley et al., 1984^[28]). A review of the heat flow data for Scotland as a whole concluded that the Tilleydesk estimate was unreliable and a mean value for Scotland of 57 mWm^-2 was derived using only data considered reliable (Gillespie et al., 2013^[29]). Based on this range of data a value of 50 mWm^-2 is chosen here to represent the apparent surface heat flow q₀ in the Aberdeen granite.

It has long been accepted that predicting temperatures at depth using surface heat flow is unreliable in regions affected by relatively recent glaciation. Thermal measurements in shallow boreholes may not yet have reached a steady state leading to anomalously low thermal gradients and thus anomalously low estimates of heat flow (Beardsmore & Cull, 2001^[30]). In the 1986 review of geothermal potential in the UK, q₀ was thought to have been underestimated in E Scotland by about 5-10 mWm^-2 (a value of 7.5 is used for this modelling) whereas a more recent European-wide study calculated the deficit in q₀ at Aberdeen (as interpolated from their maps) to be about 15 mWm^-2 (Majorowicz & Wybraniec, 2011^[31]). More recent modelling for the UK puts the estimate for NE Scotland at 18 mWm^-2 (Westaway & Younger, 2013^[32]). Palaeoclimate corrections for individual plutons in the East Grampians have recently been estimated with ranges form 16.8 to 21.7 mWm^-2 with a mean value of 19.3 mWm^-2 (Busby et al., 2015^[33]). In selecting the most appropriate figure for the deficit in q₀ it should be noted that Aberdeen was located very close to the ice front for around 5,000 years until about 15,000 years before present as the British-Irish Ice Sheet retreated (Clark et al., 2012^[34]). Locations just in front of major ice sheets tend to suffer deep chilling and permafrost due to katabatic winds (extremely cold dense volumes of air descending from the ice caps), whereas locations under the ice sheet are relatively insulated from these effects. There is abundant evidence of deeply frozen ground in the area of interest around Aberdeen in ice wedge structures that form when soil contracts and cracks at temperatures below -15 to -20°C, although their age is not yet well constrained (Gemmell & Ralston, 1984^[35]). Given the range of suggested corrections and their large effect on the predicted temperature the full range of climatic correction estimates is included in the modelling.

Heat Production (A)

The presence of radioactive elements, principally uranium, thorium and potassium, in rocks such as granites, leads through radioactive decay to the accumulation of heat energy over time. Heat production in surface rocks (A0) can be estimated using empirical formulae based on gamma ray activity or the concentrations of the principal radioactive elements (K, U and Th) present in the rock and its density (Rybach, 1988[36]). No published data was found for gamma ray activity in the Aberdeen granite so heat production has been calculated for nine samples obtained from quarries as part of this study. K, U and Th were analysed by ICPMS on solutions prepared from rock powders fused in a flux of 80% lithium metaborate and 20% lithium tetraborate dissolved in nitric acid. Density was estimated from the mass of the sample and the mass of sample suspended in pure water. The K, U, Th and density data and the heat production values obtained using the Rybach formula are presented in Table 3. Although the variability in model heat production is quite large in these samples the mean value of 2.0 µWm-3 is close to the mean value of 2.1 µWm-3 calculated from three samples of the Aberdeen granite published in a thesis on UK granites (O'Brien, 1985[37]) and is closely comparable with a mean value from three samples of 2.2 µWm-3 quoted in the 1986 UK geothermal review (Lee, 1986). A mean value of 2.1 µWm-3 is used for A0 in the present modelling.

Surface Thermal Conductivity (λ0)

Four outcrop samples were collected from the Aberdeen granite and thermal conductivity was measured in the laboratory using portable divided bar apparatus at St Andrews University specifically designed for measuring thermal conductivity in rock samples (Antriasian, 2010[38]). Samples were cut into cubes or cuboids with a diameter of up to 65mm and no thicker than 13mm. The two faces were polished to achieve parallel surfaces for maximum contact with thermal plates. Three to seven measurements were made per sample and the summary results are presented in Table 3. The best estimate for λ0 from these samples is 2.71 Wm-1K-1 +/-0.28 (1σ). The BGS GeoReport[39] for the area provides a value for thermal conductivity of 3.27 Wm-1K-1, although it is not clear how this value was determined and whether it was obtained from local granite samples or derived from regional data (BGS, 2015). Given that the BGS GeoReport is mostly computer generated from generic data, the 2.71 Wm-1K-1 value obtained from laboratory testing of samples from known locations is used for λ0 in the modelling.

A Thermal Model for the AECC DGSW Site

This section deals with the use of thermal modelling to predict the temperature of rocks that will be encountered at depth. As there are no deep wells in this vicinity and detailed land-based deep geophysical surveys are lacking, the modelling is not ideally constrained, consequently a range of outcomes is possible. One-dimensional heat flow modelling has been undertaken (Figure 4); with no deep drilling or deep geophysical datasets available onshore in Scotland it is considered that there is very limited value in performing two or three dimensional modelling. Any modelling of this type performed without site based deep data should be considered with caution. Every effort has been made to identify the most appropriate model parameters whilst recognizing that improved accuracy of temperature prediction will only come from deep drilling in the area; as highlighted above, the same will apply to all other deep geothermal prospects in Scotland.

The model assumes that the single vertical well will remain in similar granite throughout and thus a constant vertical distribution of heat production may be appropriate. An alternative model that assumes an exponential distribution of heat production with depth yields a temperature profile that differs by less than 1°C over the interval of interest and is not developed further here. The average ground temperature is assumed to be 9.2 °C (BGS, 2015).

The aim of the single well project is to achieve a fluid temperature at the bottom of the borehole of 60-65°C at depths of between 2.0 and 2.25 km. The modelling identifies that this is a realistic expectation if the conservative assumptions (based on measured parameters and the most recent literature values) hold to be correct to the target depth.

Figure 4: Steady state temperature profiles with depth at the new AECC site near Aberdeen based on a standard heat flow model (Wheildon & Rollin, 1986) and assumptions discussed in the text. Solid lines represent q0 values corrected for palaeoclimate according to different estimates (see text for explanation) whereas the broken line represents q0 before applying a correction.

Table 3: New determinations of heat production (A0) and thermal conductivity (λ0) and density for the Aberdeen granite with their locations. See text for methods used.

British Geological Survey Ground Source Heat Pump (GSHP) Geo Report

A ground source heat pump report was obtained from the British Geological Survey (BGS) to provide an estimate of mean annual ground temperatures at the site.

The BGS report provides an estimate of temperatures at depths of between 1m and 200m below the ground surface that "are made from an estimate of the local heat flow and the thermal conductivity of the bedrock geology". The report estimates the thermal conductivity (W m^-1 K^-1) of the granite bedrock to be 3.27W m^-1 K^-1 and the thermal diffusivity to be 0.1284 m² day^-1 with an estimated temperature at 200mbgl of 13^oC.

The report states that temperatures at a depth of below 15m are "affected by the small amount of heat conducted upwards from the sub-surface" which in the UK "creates an increase of temperature with depth that has an average value of 2.6^oC per 100m". Based on this average temperature gradient the average temperature at 2000mbgl would be estimated at 60^oC and 73^oC at a depth of 2500mbgl which are within the ranges identified by the thermal modelling exercise undertaken for the site.

Geochemistry of The Site

General Setting

This section addresses the composition of fluids and gases that originate in deep rocks and could be channeled to the surface by the proposed DGSW with contaminating effects on the surface environment in ways potentially deleterious to ecologies and human activity. The geochemistry will also influence the choice of materials used in the heat exchanger selected for the site. For this reason it is important to assess the likely impact in advance of drilling.

The composition of surface waters is well documented but the nature of waters from depths that the DGSW will encounter at the Rowett site (approximately 2 km) may have very different origins and be quite different in composition. Addressing this issue directly is currently not possible without access to boreholes of similar depth but analogous geological settings have been investigated worldwide and general patterns emerge that can be applied to the geology of the Aberdeen granite. Another aspect of deep drilling in granite addressed in this section is the possibility of radon escaping to the surface, especially given the hazardous levels of radon detected in many dwellings to the west of the proposed site.

Site Specific Geochemistry (Including at 2km Depth)

The sources of surface waters and groundwaters at the site and the vulnerability of the groundwater aquifers to contamination have been discussed in an earlier section. These are important for agriculture as well as domestic use, public water supply (including recreation such as golf course) and the food and drink industry (including whisky). All the water in the area is young with no evidence of palaeowater and any contamination by deep waters from the DGSW is likely to introduce older waters of a very different composition. A baseline for groundwater composition is provided by recent studies of Scotland and Aberdeenshire in particular⁴⁰⁴¹. While these reports provide useful general ranges of the main compositional parameters none provides specific information on the composition of groundwaters hosted by the Aberdeen granite or its close analogue, the Kemnay granite.

Literature describing the composition of water in rocks at depths similar to the planned DGSW (approximately 2km) generally identify three components that may contribute in greater of lesser proportions, namely:

a) surface water from various forms of precipitation

b) a seawater component derived from modern or fossil seawater

c) the products of chemical reactions between water and the rocks at depth where increased temperature and long periods of time may have a major effect on water chemistry.

The possibility of seawater is included here as the eastern margin of the Aberdeen granite lies very close to the coast but as no evidence of substantial amounts of seawater was found in any of the shallow aquifers of Aberdeenshire, a major marine influence is not likely to be present at 2 km depth.

Granite at depths of 2km would normally have some degree of interconnected fracture porosity enabling some hydraulic conductivity to form. The fluids trapped in the fracture pore system are usually highly mineralised salt-rich water or brines that can react with constituents of the host rocks. The Aberdeen granite consists predominantly of the primary minerals plagioclase, alkali feldspar (orthoclase and microcline), quartz, biotite and muscovite along with minor iron and titanium oxides, zircon and the phosphate minerals apatite and monazite. Alteration of these minerals by geothermal waters at temperatures expected at the base of the DGSW (60-70°C) is likely to lead to new minerals including the clay minerals kaolinite and montmorillonite, the zeolite laumontite and amorphous analogues of gibbsite. The solution tends to increase in total dissolved solids with depth and at 2km depth may amount to a few hundred gL^-1.

Very few studies exist on the chemical nature of geothermal fluids in bodies of granite at broadly similar depths and temperature ranges to the proposed DGSW at the AECC. Those that are available, discussed in the next paragraph, are important as they might are provide some guidance on the nature of contaminants that could reach the surface during the operation of the well.

The Rosemanowes borehole at Carnmenellis, Cornwall is sited in a granite that is petrologically quite similar to the muscovite biotite granite of Aberdeen. Extensive research was carried out on ground waters at the surface, in mines and in boreholes during the hot dry rock project in Cornwall^[40]. The composition of the waters reflect a mixing between surface meteoric waters and deeper brines, the latter being the products of water rock interaction over long time periods^[41]. Thermal groundwaters of over 50°C are observed in old tin mines and are generally saline, probably resulting from alteration reactions with biotite and plagioclase. An unusual feature is the high concentration of Li and sometimes Rb, Cs and F in the groundwaters which reflects a geochemical characteristic of the Cornish granites, a feature that is unlikely to be as pronounced in the Aberdeen granite. Groundwater salinities can reach up to 300 gL^-1 but this high salinity is not considered to be an indicator of a major seawater source. An experimental replication of the Carnmenellis geothermal system involved reacting drill core samples taken from a 2km in the granite with Na-HCO₃-Cl fluids at temperatures of 80°C and above at pressures of 50 MPa for up to 200 days^[42]. The dissolution of plagioclase and formation of smectite and calcite was the principal observed effect of water-rock interaction in these experiments.

Another example can be found in California where two wells at Cajon Pass were sampled over uncased intervals at around 2km in order to extract fluids from fractures within granitic rocks^[43]. Major differences were observed between the fluids recovered from different fracture systems. Salinity varied from 0.95 to 2.15 gL^-1, very low compared with Carnmenellis, and there were very large variations in Cl, Ca, Na, Fe, HCO3 and SO4. This study demonstrates that water composition may vary significantly if they have become isolated after following different evolutionary paths.

Geothermal fluids in the shield areas of North America tend to have very high salinities, exceeding 300 gL^-1. These brines are also acidic in terms of pH. As with Carnmenellis, the origin of these brines is not considered to be sea water but due rather to reactions between rather dilute aqueous solutions (i.e. waters with low total dissolved solids) and the primary minerals of a largely granitic upper continental crust. The products of these reactions include new minerals such as zeolites, clays and quartz along with an aqueous fluid carrying a high burden of total dissolved solids.

Potential for Surface Water Pollution

The analogue studies above indicate a significant degree of variation in the fluids associated with granite hosted geothermal systems. Significant variability is found in salinities, total dissolved solids and pH as well as the major cations and anions which determine the contaminants that can be transported to the surface. The closest analogue to the Aberdeen granite is Carnmenellis, which demonstrates the importance of the chemistry of the host rock in determining the composition of the fluids. As the Aberdeen granite is rather deficient in heavy metals it is very unlikely that any fluids from the DGSW will be the cause of significant environmental contamination by heavy metals.

Other possible contaminants for which special handling may be necessary include the halogens although it is unlikely that any recovered brines will be as saline as sea water. There also is a possibility that lithium (probably largely derived from the alteration of micas) will be present in the fluids and some geothermal plants are considering turning this into an asset by recovering and selling lithium although it is very unlikely to be an economic prospect at the proposed site.

The principles of water-rock interaction are now well understood and amenable to modelling. When rocks and fluids are recovered from drilling to depth at the Rowett site it will be possible to model the system and make better predictions concerning the composition of fluids.

Radon Gas

The granites of the eastern Grampian Highlands are known for high levels of radon. Radon has been implicated as a significant cause of lung cancer in non smokers and a recent Scotland-wide study measured the distribution of the radioactive gas in homes. Figure 5 shows the distribution of homes that exceed the action Level of 200 Bq m^-3 by 30% or more (deepest brown ornament) over the main area of granites in the eastern Grampians[44]. Using this measure of radon emissions it is clear that the Aberdeen granite underlies an area of very low radon emission. This is related to rather low levels of uranium in the bedrock granite compared with the biotite granites further to the west including those at Banchory [45]. An earlier smaller scale study of radon dissolved in water supplies from wells in the Aberdeen area found that this source makes only a small contribution to radon levels[46].

Figure 5: Radon level as indicated by levels of the radioactive gas measured in dwellings. Deep brown pixels represent the highest levels. Note the near absence of radon in dwellings over the area of the Aberdeen granite (delineated by a red line). Map taken from the report of Miles et al. (2011).

Summary

A review of the literature on the geochemistry of natural materials at the surface in the vicinity of the proposed DGSW in Aberdeen and on analogous geothermal wells elsewhere in the UK and beyond leads to the provisional conclusion that there is a very low probability of significant sources of pollution in fluid and gaseous form contaminating the site and its surrounds from the drilling and production of deep geothermal heat from a DGSW at the AECC site.