An integrated study of a Precambrian granite aquifer, Hvaler, Southeastern Norway.

DAVID BANKS, ERIK ROHR-TORP & HELGE SKARPHAGEN.

Banks, D" Rohr-Torp, E. & Skarphagen, H. 1992: An integrated study of a Precambrian granite aquifer. riveter. Southeastern Norway. Nor. geol. unders. Bull 422. 47-66.

The Geological Survey of Norway (NGU) is performing an integrated study of the groundwater resources of the Precambrian Iddefjord Granite of Hvaler municipality in southeastern Norway. Linear fracture zones are identifiable from topographic maps, aerial photos, field inspection and geophysics. The two consistently most successful geophysical methods for identification of such zones have been total magnetic field and VLF measurements. However, investigations in a newly constructed subsea road tunnel and test-pumping of borehoies on land indicate that a topographic or geophysical anomaly is no guarantee of a substantially transmissive fracture zone. The permea­ bility of the Iddefjord granite appears rather low; a background value of around 10-' m/s has been calculated from test-pumping and from leakage into the Hvaler tunnel. The top 12 m or so of the granite appear to have an average permeability c.2-3 orders of magnitude higher.

The groundwaters can be divided into 4 hydrochemical types, based on the degree of rock-water interaction and saltwater mixing. Saline groundwaters appear to be derived from fossil or current seawaters. Bicarbonate buffering, anion .exchange and calcitelfluorite saturation appear to be important processes controlling pH, bicarbonate, fluoride and calcium concentrations.

David Banks. Norges Geologiske Undersokelse (Geological Survey of Norway). Postbox 3006­ tsae, 7002-Tronobelm, Norway. Erik Rohr-Torp & Helge Skarphagen. Norges Geologiske Undersokelse, Postbox 3813 - UI/evAI Hageby. 0805-0510, Norway.

Introduction The occurrence and tlowot groun'dwater in processes which determine the water-yielding igneous and high-grade 'metamorphic rocks is capacity of fracture systems; e.g. earlier and poorly understood. Most detailed studies have current stress-fields, secondary mineralisation, focussed on single localities, often in very low neotectonic (post-glacial) fault movement and permeability terrain, and usually in connection fracture development, overlying drift deposits. with proposals for disposal of hazardous or (c) identification of hydrochemically distinct radioactive waste or for nuclear/hydroelectric groundwater types, and their chemical evoluti­ power. Relatively few studies have attempted on. to rationalise the occurrence and flow of (d) evaluation of the use of hydraulic fractu­ groundwater in fractured aquifers on a coar­ ring, explosives, acids or dispersing chemicals ser regional scale, from a practical, water­ as methods for increasing the capacity of a resources point of view. No really reliable borehole. guidelines for the location of boreholes in (e) development of standard methods for test such aquifers exist. pumping boreholes in fractured aquifers, and The aim of the Geological Survey of Nor­ a standard programme for chemical analysis. way's (NGU) Hvaler project is to carry out an integrated study of the groundwater resources of a hard rock aquifer, encompassing the follo­ Geology of the Hvaler Area wing: The Hvaler municipality consists of a group of islands (Hvalereyene) in the mouth of Oslo­ (a) evaluation of methods for detecting trans­ fjord in south-east Norway (Fig.1 and 2). The missive fractures and fracture zones - aerial dominant lithology is the Precambrian ldde­ photography, topographical maps, field sur­ fjord Granite, described by Oxaal (1916). The veys and geophysical methods. granite consists of 13 separate plutons (peder­ (b) evaluation of most important geological sen & Maalee 1990), some of the youngest 48 David Banks, Erik Rohr-Torp & Hefge Skarphagen NGU- BULL. 422, 1992

Iy infilled by Quaternary deposits, rendering ':§/ the surface outcrops of the fracture zones ~"l /:) c;:;4 __.) unexaminable. The linear channels between the islands of the Hvaler group, such as the Leksvik / ~ two straits between Vesterey and Asrnaley I and the channel between Asmaloy and Kirke­ Trondheim (Fig.2), are also believed to have arisen I 0Y by such a process. The origin of the fracture zones themselves is uncertain. It is likely, however, that they date from an early period (J) of the granite's history, as a result of regional Rangebu\ tectonic stresses or stresses related to empla­ \~ cement and cooling of the granite. The fractu­ re pattern is likely to have been reactivated or modified several times during its history; OSL~ for example, during the Permian opening of ) the Oslo rift, post-rifting strike-slip movements ",I along the Oslo graben boundary fault (Ster­ I I mer 1935 - see Fig.2), and possibly even by ) glacial and post-glacial stresses. .J The islands have undergone substantial Hvaler post-glacial isostatic uplift in the past 10,000 years or so. The highest marine limit is c. 170 m above current sea-level (Selmer-Olsen Fig. 1. Map of Norway showing location of Hvaler area, 1964). The islands have therefore only emer­ and other sites named in text (after Banks et al. 1992). ged from the sea within the last several thou­ sand years. The. hydrogeological environment of the rocks encountered onshore is thus of which yield a Rb/Sr age of 918 ± 7 million only likely to have differed significantly from years, corresponding to the end of the Sveco­ those inthe subseatunnel during that period. norwegian orogeny. Quartz, microcline and The Hvaler islands' Quaternary deposits are plagioclase are the dominant minerals in the to a large extent limited to the lineament­ granite. Accessory minerals include biotite, controlled valleys, and consist mainly of shal­ hornblende, muscovite, iron-oxides, chlorite, low marine (or littoral) sands and silts (Olsen apatite, titanite, zircon (Pedersen & Maaloe & Serensen 1990), Limited deposits of peat, 1990) and occasionally garnet. The granite wind-blown sand, and coarser gravelly/pebbly commonly includes basic clots, pegmatites and beach deposits can be found on the southern xenoliths of gneissic host-rock. In some areas part of Kirkeey. The massive areas between the xenolith content may be extremely high; the lineament valleys consist of bare bedrock in the new Hvaler tunnel the gneiss content or bedrock with a thin covering of humus. reached some 55 % (Larsen 1990, Banks et al. 1992), Ramberg & Smithson (1971) descri­ be the Iddefjord granite as a tabular intrusion on the basis of geophysical evidence. Permeability of fractured aquifers In common with most high latitude areas, It is a common assumption that the most pro­ the Hvaler area has no regional development nounced fracture zones identifiable in a crystal­ of a heavily degraded layer of weathered grani­ line rock terrain are those that will yield the te. Relatively fresh bedrock outcrops over lar­ most water .. Such fracture zones are typically ge areas of the islands, often showing signs located by their topographic expression, by of glacial scouring, or other sub-glacial featu­ use of remote sensing (Ronge 1988, Ericsson res such as potholes. The Iddefjord granite 1988) and by various geophysical techniques, area is dissected by a pattern of linear valleys such as electromagnetic induction, VLF profi­ resulting, at least in part, from preferential ling, seismic refraction, magnetic anomaly glacial erosion along zones of fractured and detection, resistivity profiling and georadar crushed rock. These valleys are usually partial- (Mullern 1980, HenkeJ & Eriksson 1980, Davis NGU· BULL. 422, 1992 An integratedstudyofa Precambrian granite aquifer 49

+ + + + + + + 10 km

N

FAULT BOUI'!DARY

NATIONAL BOUNDARY GRANITE CONTACT

Fig. 2. The Iddefjord Granite area, with rose diagrams showing lineaments (granite, number of lineaments (N) = 341, 8 diagrams: gneiss, N = 92, 2 diagrams) identified on 1:50,000 topographic maps. Right hand side shows total lineaments, left hand side shows total length.

& Annan 1989). Some hydrogeologists, howe­ many recent studies have cast doubt upon the ver, have gone beyond the «biggest is best» general applicability of such a rule. Recently hypothesis for fracture zone transmissivity, drilled boreholes in topographically prominent and have examined the influence of tectonic fracture zones in a variety of Precambrian and stress. Some workers (Larsson 1972, Huntoon Palaeozoic bedrock lithologies on the island 1986, Rohr-Torp 1987) have identified a regio­ of Hitra, and in geophysically prominent zones nal correlation between the past tectonic stres­ in Palaeozoic schists in Leksvik municipality ses which created or reactivated a fracture (Bueslatten et al. 1984a & b) have yielded pattern and the permeability of the constituent spectacularly little water. Furthermore, during fractures/fracture zones, while others (Olsson the course of tunnel excavation in Norway, 1979, Selmer-Olsen 1981, Carlsson & Christi­ particularly of subsea tunnels, it has been ansson 1987) have found a correlation betwe­ noted that the largest fracture-zones crossed en permeability and the current stress field by tunnels often give rise to very few water within the rock. leakage problems. The majority of large water While the common assumption that major leakages tend to arise from smaller fracture fracture zones are significantly transmissive zones, or individual fractures/groups of fractu­ has been shown to be true in some cases res in relatively massive bedrock. Examples (e.g. Carlsson & Olsson 1977, Skjeseth 1981), (see Fig.1) are described from Alesund (Olsen 50 David Banks, Erik Rohr-Torp & He/ge Skarphagen NGU·BULL422,l992

& Blindheim 1987), Ulla-F0rre (Bertelsen 1981), might be expected to be rather efficient at tigh­ Lysaker-Slemmestad, near Oslo (Leset 1981), tening fractures. Alteration processes would Flekkerey (Gulbrandsen 1989) and Karmsund be particularly intense along major fracture (Kluver 1983) and are summarised by Nilsen zones due to (a) their presumably high pre­ (1988, 1990) and Banks et al. (1992). This alteration permeability and (b) the high speci­ phenomenon is ascribed, in many cases, to fic surface area of the gouges and breccias the largest fracture zones being 'tightened' within the zones. Water inflows to the Hvafer by secondary clays resulting from weathering tunnel tended to occur through lesser fractu­ or hydrothermal activity. It is also noted that res or fracture groups, in most cases not the most 'leaky' subsea tunnels have included detected by preliminary investigations (Fig.3). the GOd0y (Storas 1988) and Frierfjord (Kluver Calculations from Lugeon testing (method, 1983) tunnels, where, significantly, there was e.g. Moye 1969) and total inflow to the tunnel very little clay mineralisation and, in the case indicated a 'background' permeability of 10" of the GOd0y tunnel, no major fault zone was - 10" m/so The permeability in the vicinity of crossed. the major leakages is estimated to be 100 to A similar phenomenon is observed in tropi­ 1000 times higher, around 10.7 -10" rn/s The cal areas. Studies of the weathered (saprolite) case study is detailed in Larsen (1990) and layer overlying fresh bedrock (e.g. Acworth Banks et al. (1992). . 1987) indicate that while a low degree of weat­ It appears, therefore, that the identification hering can be effective at destroying the bon­ of major fracture zones by geophysical and ding between mineral grains to give a gravelly remote sensing techniques may not be a satis­ texture with enhanced permeability, a higher factory method of locating groundwater resour­ grade of weathering results in extensive altera­ ces in hard rock aquifers. No current geophysi­ tion to clay minerals and a substantial decrea­ cal method can adequately distinguish betwe­ se in permeability. Although the alteration pro­ en water-transmissive and clay-filled fracture cesses involved at Hvaler (presumably low­ zones. temperature hydrothermal alteration or deposi­ tion - Storey & Lintern 1981) are somewhat different to those in tropical weathering, obser­ vations from the recently completed Hvaler tunnel suggest that clay-alteration may also have a substantial effect on the permeability of fracture-zones at some depth within a bed­ Hydrogeology and borehole yield rock aquifer (Banks et al. 1992). Borehole statistics NGU maintains an archive of data on ground­ The Hvaler tunnel water boreholes in bedrock in Norway. Statisti­ In Hvaler, a 4 km subsea road tunnel was cal data for the Hvaler municipality, and the constructed in 1988-89 to link the islands of entire area of the Norwegian Iddefjord Granite Asrnaley and Klrkeoy (Figs.2,14). Prior to ex­ within four map sheets 1913 I-IV (roughly the cavation, major fracture zones were located area of Fig.2), is presented in FigA and Table by the use of aerial photos, acoustic profiling 1. It is worth noting that the distribution of and seismic refraction (Tauqbet & 0verland borehole depths is roughly symmetrical, the 1987, Larsen 1990). On encountering these mean coinciding with the median. The distributi­ zones during tunnel construction, the majority on of borehole yields is, however, highly ske­ were found to be of low transmissivity and wed towards low yield. The mode is conside­ filled with clay minerals. These clay minerals rably lower than the median, in turn considerab­ contained 50 - 100 % smectite with extreme ly lower than the mean. The mean value is swelling capacities on contact with water (up forced up by the existence of a few boreholes to 400 % free swelling). Smectite fracture fil­ with very high yields. Such a phenomenon has lings have been found in many Norwegian also been recorded from the Bergen area by hard-rock lithologies and areas (Selmer-Olsen Ellingsen (1978), from the 0stfold area (Bryn 1964). They are probably the result of low­ 1961) and from the Drebak area (Rohr-Torp temperature hydrothermal alteration by Mg­ 1987).The median yield thus has greatest signi­ and Ca-rich fluids (Storey & Lintern 1981), and ficance for planning groundwater abstractions. NGU-BULL.422 , 1992 An integrated study ofa Precambrian granite aquifer 51

ASS UMED " ZONES OF WEAKN ESS " ( FROM PRELIM INARY INV ESTIG ATI ONS I ~ Kirk eoy -._~D...... e~l h Asma lo)' I

11 - 0 I 11 11 11 II :: : : \ (1 - 50 I

i ~kr j:::======n4==r=:=:::::::---'j - 100 1 I I I , , 2000 2500 3000 3 500 4000 4 50 0 5000 5500m

FRACTURE ZONES ENCOUNTERED DURING EXCAVATION

INJECTI ON GROUTING OF WATER LEAK AGES

[EITI..---,-----,-i IL-...,-I ------ri ------ri- ,.-JI 2000 2500 3000 3500 4 000 4 50 0 500 0 5500 m

WATER LEAKAGES ENCOUNT ERED DURING DRILLlNG/ EXCAVATION I/min 3 0 0

'00

30 I '0 , :I I I

i 11 ; 11 11 3 Id ! I III II 2000 250 0 30 0 0 3500 400 0 4 50 0 5000 55 0 0 m Chainag e Tot.l l••"'ag. Ihrou g h t, l.l ho l•• l ••kalle I. g ill e re d in tun ne l lollowing ••c.....llo n Le.keg ...... hich C• • l e d a rt.r .. , hort t ime

Fig. 3. Correspondance between presumed (from geophysics and aerial photos) and actual fracture zones , and water leaka­ ges into the Hvalar tunnel (after Larsen 1990, and Bank s et al. 1992). Leakages are shown as the total inflow from all probe holes at a specific chainage (note logarithmic scale).

If one drills a bore hole on the expectation of figure is a little less than Bryn 's (1961) analy­ a 50 % chance of achieving the m ean yield, sis of the gra nites of 0stfold (dominated by one will be seve rely disappointe d. The mean the Iddefjord granite), but is considerably less on ly beg ins to have significance wh en planning than that reported for other Scand inav ian gran i­ a co mbined abstraction co nsist ing of several tes , where the med ian yield is often quoted boreholes. as around 1000 I/hr (Persson et al. 1979, The mean yield for the Iddefjord Gran ite 1985a ,b). The mean and median yields for the appears to be in the reg ion of 1100 I/hr, and Hvaler area are less than those for the Idde­ the median yield aro und 500 I/hr . This median fjord gran ite as a whole. The re is no observab- 52 David Banks. Erik Rohr-Torp & He/ge Skarphagen GU - BULL. 422. 1992

Table 1. Corre lation be ween depth and borehote yield . All borenoles In Idde fJord grani te With Yield Information ( 310). quar tile Intervals according to Yield. IH vale r

Depth (m) Yield (I hr) 'It ({ w (Wi··.. A ,.. ts t quartile ean 57.5 81.4 . u Interval ax 130 200 Iddefjord Granite In 10 0

2nd quartile Mea n 55.4 333 .2 Interva l Max 140 500 MIn 7 200 - _ "' ODIro ~ " ~ U ~ & " I "" II _ ...... _ 1 ...... u 'wu _""'LUl S 'VtJ, .. [ utS 3rd quar ile Mea n 56.1 802.4 V /\ .. fOO U,t(, _ IO'lI O'lOiL f ' l O Interval Max 128 1400 f<::::=z -.o-EMQl,.f DL.'M Min 10 500

4th quartile Mea n 53.7 326 1 Interval Max 126 10000 MIn 10 1440 Fig. 4. Statistical distribu tion of Yield and dep h of boreno­ les registered on GU's arch ive. Bars represen he srnat­ 0 Mean Yield ~ 1124 Ilhr les interval encornpassmq 67 0 of borenotes. Median Yield = 500 Ilhr le correlation between borehole depth and the same area has been carried out (Fig.6). yield (Table 1), perhaps due to a driller continu­ With increasingly detailed stud ies (i.e. maps ing to cons iderable depths if no goo d supply to aerial photos to field measur ements) one is found , but stopping if he meets a good obtains increasingly complex resu lts. the varia­ supp ly at relatively shallow depth. tion in fracture direct ion increases. and a gene­ The statistical figures must be used with ral W direction beco mes increasingly promi- caution. The statistics are likely to be overesti­ nent over the El E direct ion. A all scales. mated because: however. he El E direction can be iden ifi- ed in most sub-areas. as can he lac of E­ (a) many dr illers use cru de methods of asses­ oriented fractures. sing a borenoles yield, or at best a short term Despite the com bined resul s of the field pumping/recovery test. Long term capac ity survey giving an apparently complex result. may be con siderably less. at each individual locality a well-defined frac u­ (b) the archive includes some (although by no re pattern consisting of two or three fracture means a major ity) boreholes with capac ities sets (typically steeply dipping) could com mon­ artificially increased by exp losives or hydraulic ly be identified. A large amount of variation fracturing. betwee n localities did, however . occur . (c) some unsuccessful, low -yield boreholes On the NW peninsula of Kirkeoy (left half may not be reported to NGU. of Fig.5), the largest. aerial-photo-identifiable fracture zones have four major direct ions, (NW, NNW, NE-NN E and ENE), dividing the terra in into a mosa ic of smaller blocks . Such Fracture mapping orthoqonal, or double-orthogo nal patterns ha­ Fract ure mapping in the Hvaler area has been ve frequently been observed in ancien grani­ carr ied out on three scales (Banks & Rohr­ tic terrains (e.g. Tiren & Beckholmen 1989). Torp 1991). Lineam ents have been identified Within each block, tecto nic stresses wou ld be and measur ed from 1:50,000 topographical significantly mod ified by the presence of he maps covering a large portion of the lddefjord major fracture zones. and by the interaction granite area and some of the adjoining gne iss of adjacent blocks. thus giving a plausible area (Fig.2). The results show a very dom i­ exp lanation for the increasingly comp licated nant NNE or NE direction over the entire gra ni­ fracture pattern at smaller scales. te area, with subsidiary NNW and direct ions It appears that two types of frac ure zone in some sub- areas. Lineaments on the well­ can be distinguished: those whose componen exposed northern half of the island of Kirkeo y fractures are approxima ely parallel 0 he have also been mapped using aerial photos zone and those comprised of fractur es lying (Fig.S). and a field survey of fractures within oblique to the main trend of the zone. This NGU - BULL. 422, 1992 Anintegratedstudy of a Precambriangraniteaquifer 53

Fig. 5. The northern part of Kirkeoy, showing lineaments identified from areal photos . Rose diagrams show lineaments identified from the pho tos (N = 1637). Right hand side shows to­ tal lineaments, left hand side shows total length.

Fig. 6. The northern part of Kirkeoy, show ing lineaments identified from aerial photos. Rose diagrams show fracture str ikes identified from field measurements (only fractures with dip S 45° are included). N = 1167. 31 locali ties. 54 David Banks. Erik Rohr-Torp & Helge Skarphagen GU · BULL. 422. 1992

the test areas have been Investigated by geo­ physical methods including electrical resistivity profiling. total magn etic .field measur ements. georadar and very low freq uency electromag­ netic induction (VLF). All method s, except geo­ radar, showed significant anomalies at the major topographical lineaments under some circum stances (Lauritsen & Ronning 1992). However, the VLF measurements were oft en disturbed by 'noise' from power-transmission cables. Total magnetic field measurem ents appeared to be the most consistently reliable of the various geophysical methods. relying on the oxidation of the granite's magnetite conte nt to haematite along fracture zones.

Drilling programme Drilling is planned at the selected test sites . o and has been completed at Tests ite 1, SE of Pulservik (Fig.7). Here. four boreholes. each ~ Conto ur int eryal S rn 1...... -t Boreholes c. 73 m deep have been drilled. Details are e W Dug Well given in Table 2 and Banks et al. (1991). Two ~ Geophysical prOfile with Vl F/ magnetic anoma ly holes. numbers 1 and 2, have been drilled - - Fracture zones into each of the two intersecting fracture zo­ -- Topographic contours nes at 73° from the horizont al. They were expected to cross the fracture zones at ca. 50 Fig. 7. Map of testsite numbe r 1. SE of Pulservik. - 60 m dept h. assuming the fracture zones to be approximately vertical. The other two holes (3 and 4) were drilled at 60° from the hor izon­ latter type are anot her pos sioie reason tor tne tal into relatively massive granite (i.e. away discrepancies between Figs. 5 & 6. from the fracture zone s). Obser vations during Fracture minerals obse rved during the fractu­ drilling (November 1990) were carefully recor ­ re survey have included quartz, chlorite, epido­ ded and the holes were geophysically logged te, calcite, fluor ite, lepidolite and pyrite. Of (electical resistivity, self-potential, fluid resisti­ these , epidote has a tendency to occur prefe­ vity and fluid temperature; point measure­ rent ially on fract ures with strike 20-40° and ments at half-meter intervals) in Septe mber c.130°, and fluor ite (usually in association with 1991 (Fig.8). calc ite. and often with epidote ) has only been Hole number 1 encountered several minor recorded on frac ture s with str ike 23-40° along water-bearing fractures . as well as some 'dry' the Korshavn-Urdal fracture zone (Banks & fractures. The main fracture zone, char acteri­ Rohr -Torp 1991). sed by fast drilling, reddish cuttings and a Most fractures surveyed in the field were powerfu l anoma ly in the resistivity log, was steep ly dipping. The occ urrence of near-hor i­ encountered, as expected, between 54 and zonta l unload ing fract ures appears to be very 62 m. It is interpreted as a substantial cru sh­ variable. In some locations, such as road cut­ zone. The major ity of the fracture zone appea­ tings on the small islands north of Vesterey. red to be of rather low permeability. a major hor izontal unloading joints are well-developed inflow only being met at c. 62 m (i.e. at the to a depth of several metres, but at other loca­ very base of the zone), and a very minor one tions, the joint set appears very poorl y develo­ at c.55'/, m. The location of the main zone ped. of water flow at the boundary of the fracture zone with relatively massive rock is a feature Geophysical investigations also observed by Ahlbom & Smellie (1989) in Severa l potential test-drilling areas have been Sweden. The inflow position is conf irmed by identified on the NW peninsula of Kirkeoy. All the fluid resistivity log. run in September 1991, NGU- BULL. 422, 1992 An integrated studyofa Precambrian granite aquifer 55

FEATURES OBSERVED DURING DRILLING 3 4 po. 9 \09 po. e;, \0 \09 pt

KEY Geophysical logs

Fracture / weathered rock Casing Electrical resistivity of rock (16" normal) ohm - m Reddish rock Most prominent inflow In each hole ...... Electrical resistivity of s Rapid drilling water ohm-m Minor water inflow ~ Crush zone ~

Fig. 8. Features observed during boring of testholes 1-4. Geophysical logs. Depth scale in m. Pa = apparent rock resistivity.pr = fluid resistivity. after the pumping tests of May 1991. The log in the electrical resistivity log at cA7 m and shows clearly that 62 m is the lowest major 58 m, together with minima in the self-potenti­ flow horizon, and has yielded relatively fresh al log at c. 46 and 53 m, There was no sign water with a resistivity about 30 ohm-m (c.330 of any major fracture zone as encountered in IlS/cm) at in-situ temperature. Below that le­ hole 1. Thus, either the fracture zone had a vel, the borehole contains a more saline 'resi­ substantial deviation from the vertical (unlike­ dual' water with a resistivity of 7 ohm-m (1400 ly, given the zones' direct cross-cutting of the IlS/cm). The temperature of the fluid column topography, the near-vertical nature of the in borehole 1 ranges from a minimum of c. zone in hole 1, and of the zones encountered 7.0°C at 15-19 m up to 8.2°C at 71 m, corre­ in the Hvaler tunnel) or it died out at a relative­ sponding to a vertical gradient of 0.022°C/m. ly shallow depth. The only significant water The temperature shows a very small 'step' inflows were from shallow, near-surface fractu­ by the 62 m level, adjacent to the inflow horizon. res. In hole 2, the only signs of a possible fractu­ Holes 3 and 4, as expected, encountered re zone at the expected depth (46-63 m) were no significant fracture zones. Hole 3 appeared slightly elevated boring rates and small minima to meet a substantial water inflow at c.24 m, 56 David Banks, Erik Rohr- Torp & Helge Skarphagen NGU - SULL. 422. 1992

Table 2: Borehole details, testsite 1

Borehole no. Depth (m) Azimuth Fall Rest water level Yield (l/hr) with (m under well top) Pumping water level = 50m 25/5/91

1 .73.5 3° 73° 3.75 360 2 73.5 157° 73° 2.65 65 3 73 110° 60° 1.13 40 4 73 24° 60° 0.86 22 Dugwell 1.7 vertical 0.74

but subsequent test-pumping failed to detect water level and a A vary. In methods (b) and this, indicating it to be an unconne::ted fractu­ (c) a A is measured for an approximately cons­ re with limited storage capacity. Otherwise, in tant water level. holes 3 & 4, the significant inflows (although It is widely accepted that the specific capaci­ relatively small) were from rather shallow frac­ ty (F) of a borehole in bedrock is approximate­ tures - subsequently confirmed by test-pum­ ly proportional to the 'apparent, local transmis-.· ping, sivity' (T) of the fractures feeding the hole. The correlation between geophysical logs The Logan approximation (Kruseman & De and drilling observations was rather good. The Ridder 1989), the Moye (1967) and Banks correlation of reddish cuttings with fractured (1972) methods, the Krasny (1975) method and zones should also be noted (the cuttings other­ the Carlsson and Carlssted (1977) method, wise being grey in fresh granite), presumably all assume the following equation: due to the presence of oxidised Fe from prefe­ rential weathering along fracture planes. T = Flc where c is a constant. As the specitic capacity of the borehole is Test pumping merely the sum of the specific capacities of The four holes were capacity-tested in May the individual 'feeder fractures'. it should be 1991 (Banks et al. 1991), using four methods. possible to estimate the apparent transmissivi­ (a) low-rate step-test pumping using a Grundfos MP1, 1'I; ty of these fractures from a simple a A vs. diameter, submersible pump, and subsequent recovery. water-level diagram (Fig.10). Such a technique Yield and water level were measured. Due to rapid water level decline, the pump frequency, but not the yield, was (Banks 1991) appears to be most applicable constant for each step. Results are marked. (drawdown) to relatively low-capacity boreholes where the and 0 (recovery) on Fig.9a. low transmissivity of the feeder fractures is (b) rapid emptying of the borehole down to c.50 m by a large capacity 4" diameter submersible pump, and subsequ­ the controlling factor for the borehole's yield, ent pumping with a constant pumping water level of 50 rather than the storage properties of the wi­ m. The yield was measured. Suitable for higher capacity holes. Marked + on Figs.9. der aquifer. and where a 'pseudo-equilibrfurn' (c) rapid emptying ot the borehole down to 50 m, pump is established relatively rapidly. In higher capa­ switched off for a given time (c. 10 - 30 mins.) followed by up-pumping of the amount of water accumulated in the city holes. storage effects become important, hole during the interval. Suitable for low-capacity holes. and a A vs. water-level plots typically display Marked + on Figs.9. hysteresis between drawdown and recovery (d) monitoring of water level's recovery after methods (b) or (c). Marked x on Figs.9. (Banks 1991). The gradient G of the a A vs. h plots in Figs.9, at a given water level, is rela­ As the contribution (as) from the storage capa­ ted to the total specific capacity F of all the city of the borehole itself is significant in com­ fractures below the water level, by: parison with the total yield (a), the contribution . from the aquifer (aA) is calculated by: F = G/sin u

aA = a - as = a + 1tr'(oh/ot) where u is the fall of the borehote.

where r = the boranole's radius (0.07 m) As regards the constant c. this has been cal­ and oh is the change in water level in a short culated by Moye's (1967) method as around time interval ot (positive for rising water level) 1.4 for Lugeon testing of short sections of a borehole, by Logan (see Kruseman & De Rid­ Thus, one can plot aA vs. water level as in der 1989) as 0,82 (ostensibly for porous flow Figs.9. In methods (a) and (d) above, both aquifers), and by Krasny (1975) as 0.91. Carls- z Cl C BOREHOLE 1 BOREHOLE 2 1Il eo C 600 Gradient = 0.0944 m2/d ,... F = 0.0987 m2/ d Gradient =0.00948 m2/ d !,. T= 1.3xlO-6 m2/s 70 F=0.00991 m 2/ d 2 2 1'>'" 500 Gradient = 0.948 m / d T=1.3xlO-7 m / s x. 2 )l~__ F = 0.991 m / d 60 ---+ ~ T =1.3 x 10- 5 m2/s I _------x--..>X.~ Xx ..c 400 ..c 50 ...... 1 o o --:------..J 40 W 300 Ld r ;: ;: ~ Gradient= 1.207 m2/d 30 I F= 1.26 m2/d 200 I T=1.6x10-5m2/s 20 I r 100 I 10 I j ( 0-+---'---,------.----,------,------, 10 20 30 40 50 o 10 20 30 40 50 WATER LEVEL Cm) BELOW DATUM WATER LEVEL Cm) BELOW DATUM

eo BOREHOLE 3 BOREHOLE 4 30 70 Gradient = 0.00378 m 2/ d 2/d 2/d 60 Gradient=0.00876 m F= 0.00436 m F=0.0101 m2/d e 2/s T=1.3x 10-7 m2/s T= 5.6x10- m _------~:---- ...... c 50 ~20 _--- XX 0 o ll ..J ~_--~ ..J -- )< 40 _----x UJ 1 UJ ;: _.,.~_--L------:------;: ------l 30 "': 0, 10 i;;:.~.---- Gradient = 0.0877 m 2/d I 2/d 20 I "'-£ F= 0.101 m .~---~ 2/d "" T=1.3x10-6 m2/s xl Gradient=0.1013 m I F=0.117 m 2/d 10 / T=1.5xlO-e m2/s J. / I '<:f/"" 0 o+'------r------,-----r-----,------, 0 10 20 30 40 50 o 10 20 30 40 w WATER LEVEL Cm) BELOW DATUM WATER LEVEL Cm) BELOW DATUM

Fig. 9. Results of capacity testing carried out on testholes 1-4. See text for explanation of symbols. Plots show yield from aquifer (QA)vs. water level below well-top. 58 David Banks. Erik nonr-Torp & Helge Skarphagen NGU - BULL 422. 1992

Table 3 : Hydrogeological parameters calculated from Figs. 9 a-d.

Borehote 1 2 3 4 Borehole section 0-12m 12-73'I,m 0-12m 12-73 'I,m 0-12m 12-73m 0-12m 12-73 m

GradientG m'ld 0.948+ 0.218 1.21 0.00948 0.1013 0.00876 0.0877" 0.00378 Specific capacity F m'ld 0.991+ 0.228 1.26 0.00991 0.117 0.0101 0.101 0.00436

Entire borehole TransmissivityT m'ld 1.10+ 1.40 0.130 0.113 rnvs 1.3xl0" 1.6xl0" 1.5 xl 0'- 1.3xl0'·

Section-wise Transmissivity T rn'rd 0.848+ 0.253 1.39 0.0110 0.119 0.0112 0.108 0.00485 m'/s 9.8xl0'· 2.9xl0'· 1.6xl0" 1.3xl0" 1.4 xl O: 1.3xl0'· 1.2 xto- 5.6xl0·· Saturated depth m 8.25 69.75 9.24 61.5 10.83 61 11.08 61 Hydraulic rn/s 1.2 xl0·- 4.2xl0'· 1.7 xl0" 2.1 xl0" 1.3xl0'· 2.1 xl0" 1.1 xl0" 9.2xl0·'" conductivity K + shallow fracture appears to have dned up dunng testing • based on assumed gradient. Fig. 9d.

son and Carlstedt (1977) developed the met­ zones having permeabilities of 10" rn/s and hod further for non-steady-state pumping and higher, and more massive portions of bedrock contend that, for normal Swedish 110 mm 10'" rn/s or less (Hult et al. 1978, Olsson diameter holes, with pumping periods under 1985). The Iddefjord granite at the test-site 1 day, the c value will lie between 0.9 and thus lies towards the lower end of this scale. 1.1 (they then proceed to use values of around In defence of the «prominent fracture zones 0.84 to evaluate apparent T for four Swedish equals elevated borehole yields» theory, the bedrock areas!). A value of 0.9 is used here one borehole (no.t) which clearly crossed a (Banks 1991), and the calculated apparent significant fracture zone did have a substantial­ transmissivities are shown in Table 3. ly higher yield (c. 360 I/hr with pumping water It should be noted from Fig.9a that the draw­ level (PWL) at 50 m) than the other boreholes down curve indicates the presence of a shal­ 2, 3, 4 (65, 40 and 22 I/hr respectively). It low contributing fracture in borehole 1 at c. must, however, be pointed out that the yield 9-11 m (presumably that encountered during of 360 I/hr is still below the median (400 I/hr) drilling at c.11'/, m - see Fig.8). This appears and mean (745 11hr) yields for the Iddefjord to have dried up during pumping, the recove­ granite in Hvaler. This could be due to: ry curve being a straight line, dominated entire­ ly by the major deep inflow at 62 m. In pore ho­ les 2, 3 and 4 the yield is very low, and what (a) the fracture zone having a low-permeability little water there is is derived from shallow, filling of clay minerals. relatively transmissive fractures at around 5, 8 and (presumably) 4.75 m respectively (corre­ (b) the statistics from the borehole archive sponding well with drilling logs). Apparent being artificially exaggerated. permeability in the boreholes can be calculated by dividing the transmissivity by the boreho­ les' saturated length. Hole 1 gives an average apparent permeability of 4 x 10,' rn/s, relative­ ly low for a bore hole crossing a prominent fracture zone. In the other boreholes apparent Gradient F permeabilities in the region of 10" rn/s are =- 2 obtained for the bore hole sections below the transmissive near-surface fractures, agreeing Gradient = - (Fl+ F2) well with the results derived from Lugeon tes­ ting in the lower permeability sections of the Fracture Fracture R'NL Water level Hvaler tunnel (Banks et al. 1992). The average 2 1 permeability of Precambrian gneissic and gra­ nitic aquifers is reported to lie around 10" ­ Fig. 10. Diagram illustrating theory behind capacity-test analysis for a vertical borehole. F, = specific capacity of 10,' m/s, with many 'water yielding' fracture fracture 1 etc. RWL = rest water level. NGU - BULL. 422, 1992 An integrated studyofa Precambrian granite aquifer 59

Table 4 : Summary of hydrochemical samples. Sample numbers correspond with figs. 11-14. Samples 1-9 analysed in connection with tunnel construction. Samples 10-35 & 6b analysed by NGU (d = bore deviated from vertical).

Sample Name Sampling date Source Type Borehole depth m Chloride mgll Water Type

1 Birkeland 23/11/87 Borehole - granite 25 11.2 111 2 Berg 23/11/87 Borehole - granite > 50 4.2 I 3 Hansen 23/11/87 Borehole - granite > 31 20 11 4 Serensen 23/11/87 Borehole - granite > 43 19 111 5 Marstrander 23/11/87 Borehole - granite 78 270 IV 6a Bombua 20/10/89 Borehole - granite ? 37 6b Bombua 23/03/90 Borehole - granite ? 22 111(11) 7a serensen 2 25/10/88 Borehole - granite 60 33 11 or III 7b Serensen 2 15/11188 Borehole - granite 60 35 11 or III 7c Serensen 2 24/02/89 Borehole - granite 60 37 11 or lll 7d Serensen 2 27/06/89 aorenote - granite 60 938 IV 8 Serensen 3 28/04/89 Borehole - granite 70d 800 IV 9 Chain.2140 28/04/89 Leakage in tunnel 56 lOa Chain.4120 23/03/90 Leakage in tunnel 11600 Modified sea lOb Chain.4120 29/05/91 Leakage in tunnel 17300 Modified sea 11 Urdal 09/05/90 Leakage from granite cliff 21 11 12 Melhuus 15/05/90 Well- Quaternary 3 22 11 13 Melhuus 15/05/90 Borehole - granite 45 57 IV 14 B01ingshavn 15/05/90 Well- granite 6 96 IV 15 B0lingshavn 15/05/90 Borehole - granite 27 202 IV 16 Daulekkene N 15/05/90 Borehole - granite ? 20 111 17 Daulekkene S 15/05/90 Borehote - granite ? 56 IV 18 Sandbrekke 16/05/90 Borehole - granite 80 35 111 19 Testsite 1 08111190 Well - Quaternary 1.7 35 11 20 Belinqshavn 08/11/90 Seawater 7500 Seawater 21a Testhole 1 25105191 Borehole - granite 73.5d 32 111' 21b Testhole 1 25/05/91 Borehole - granite 73.5d 57 IV' 22 Knausen 26105/91 Borehole - granite 90 283 IV 23 Skartlien 26105191 Well- Quaternary 3.8 31 11 24 Testhole 2 27105/91 Borehole - granite 73.5d 17 111' 25 Gran/; 28/05/91 Boreho/e - granite 101 98 IV 26 Testhole 3 28/05/91 Borehole - granite 73d 32 111' 27 Chain.3615 29/05/91 Leakage in tunnel 17000 Modified sea 28 Testhole 4 29/05/91 Borehole - granite 73d 103 IV' 29 Solhell 30/05/91 Well - Quaternary ? 15 11 30 Andresen 30/05/91 Borehole - granite 40 26 I11 31 Heyerdahl 30/05/91 Borehole - granite 65 468 IV 32 Granlie 30/05/91 Borehole - granite 70d 92 IV' 33 Dahl 30/05/91 Borehole - granite 76d 87 IV 34 Testsite 1 31/09/91 Precipitation 3.0 35 Testsite 1 31/09/91 Storm run-off from granite 42

Drilling cuttings from the boreholes were analy­ Groundwater chemistry sed by X-ray diffraction. In one sample, from A series of water samples has been taken a 'dry' fracture in borehole 3 at 10'1, m, where from boreholes in the Iddefjord Granite, both it was possible to extract individual clots of in connection with the excavation of the Hva­ clay minerals from the cuttings, smectites ler tunnel, and with NGU's project. In addition, were positively identified. samples of seawater (B0Iingshavn), rainwater The test pumping results give an overall (in an open area near testhole 3 during a picture of a 'two-layered' fractured aquifer. storm), and storm run-off (from a granitic 'mas­ The upper layer, down to approximately 12 sif' near testhole 3) have been taken, together m, contains relatively transmissive fractures with samples from water leakages in the Hva­ (10-' - 10-' m'/s), which do not, however, yield ler tunnel and from wells in Quaternary depo­ large quantities of water due to the small sits. The samples are detailed in Table 4. head gradients that are achievable along such The NGU samples have been analysed for shallow fractures. The deeper layer is typical­ cations (samples filtered by a 0.45 urn Millipo­ ly of very low permeability, c. 10-' mls, except re filter) using inductively coupled emission where a major fracture zone crosses boreho­ spectroscopy, anions by ion chromatography, le 1, having a significantly, though not dramati­ alkalinity by standard titration, and for pH and cally, elevated apparent transmissivity of 3 x electrical conductivity in the laboratory using 10-6 m~/s (corresponding to an average appa­ standard electrodes. The relationships betwe­ rent permeability for the lower section of the en the most significant parameters are detai­ borehole of 4 x 10-' m/s). led in Figs. 11, 12 & 13. Four main water ty- 60 David Banks. Erik Rohr- Torp & Helge Skarphagen NGU - BULL. 422, 1992

10..------, 100..------;;::;;:J

21a• •'nb 24 ob I' .15 10 2.:·" • 20 ::::.. 22 '0 1••16 :8 ".3' E 0- E •• ­ .3 "E E ~ U u X23 '" 0,1 X,2 x ts [01 +R 0,1 1 10 100 0,1 1 10 100 1000 Chloride meq/l Chloride meq/l

100..------, 8,------,

20 7 10 2: 6 0-

.... ~" 5 0- , • e31 E :!! " '5 22 " 4 • z5 u 3 ~ tor· <, ,.'(, '3 E 0,1 , 2 2 :0 o +R III SOL <,Q" O,Ol+----<-,.----.-----r---~---_l O+---~--~--~--__:_""=__-~ 0,01 0,1 100 1000 0,01 0,1 100 lQOO

10..------., 1000.,------A 21 S lOb 20 1Oa 100

3 1 e~ 1522 15 22 •••• :::. 10 e- •••31 e- ~la"·5 ­ 2'. ~ • , •a "E 0,1 E 24 ...... 17 7d,e '0 .' " ., ­ ~ 1.~ g :0 29l(X o 29 '2 i: h9 III o 23 0,01 G 0,1 +R e,Q" O,OOl+---~---~---'r_--~---_l O,Ol*----.---~---~--_,..:_--~. 0,01 0,1 1 10 100 1000 0,01 0,1 1 10 100 1000 Chloride meq 11 Chloride meq fI

1000..------=::-:-'7<1 0,5.,-----:::==-.,=:,,------, loa S 21' 20 0,4 100

::::. 50,3 e- .31 o 15 E "~ 10 .22 e E '0 ~ 0,2 'E 20 1 o , ,;; Cii 24 ~9 23 0,1 X2. .20 O,'+---''--,,---~---.._--~---_l O+----r---~--~--~--_< 0,01 0,1 10 100 1000 0,01 0,1 100 1000 Chloride meq 11 NGU- BULL. 422, 1992 An integrated study ofa Precambrian granite aquifer 61

5 2

p?:b 21b 0 '2'a,b ~ .m,~~ .'5 .31 4 . 5 ••22 >< 13 ~N 17 .. 3·. . .33 "0 -2 4~23 <, • 15 .~2 .!: "0 24 •••13 e 12 3 25 14 .6b 0 E :;: -4 E 5. • 26 ~18 X19 3oe,~, 29 ::! ?: .3 28 22 co -6 :§ 2 '~ Ul -; • 027.' 2 010b 'u <'" 4 -; -8 3. • +20 U -10 [0] 23 X19 0. 0 12 -12 4 5 6 7 8 9 10 0 5 10 15 pH Chloride meq/I

0 Kl 32 ·3' ~5 32 ~3 >< 6b ~615.·.13 • .'5 .22 K2 22 .. 6a,b "0 •• re .5" 31',4· - 1 .4 • C" • :21a,b c - 28 ••18 ..E 30 .2 24 ea iii .,6 :; X29 :2 0,1 .7d -2 (; 1; X12 .r, Ul " ~ 23 u: -3 'S<19 (; 7:.7a .16 0 u:" 0 7e. X12 -4' X29 X,9 @jfa,b 23 -5 6 7 8 9 10 o 5 10 15 pH Chloride meq/I

Fig. 12. Fluoride and alkafinity plotted against pH. Symbols Fig. 13. Estimated calcite and fluorite saturation indices as for Fig. 11. (SI) plotted against chloride. Calcite index estimated from the simpfified equifibrium : K=(Ca++)(HCO,-)/(H+); log K = 2.17 at rc (L1oyd & Heathcote 1985). For fluorite saturati­ pes, and one subsidiary type, can be identified on index, K, shows SI relative to log K5 = -10.4 (Krau­ skopf 1977); K. shows SI relative to Nordstrom & Jenne's on the basis of sodium, chloride and silicon (1977) value of -10.96 (both at 25°C). K values are tempera­ content, as follows. ture-adjusted to 7°C using the Van't Hoft isotherm, and activities are calculated using the Debye-Huckel formula. Type I has sodium and chloride concentrati­ Speciation modelling of water has not been included in ons very similar to those in rainwater, and is calculations. Symbols as for Fig. 11. only recorded in one borehole (number 2). This indicates little interaction in the soil zone or rock, and probably implies very rapid rechar­ NGU, and the borehole is no longer available ge/pipe flow. The possibility of direct surface for examination. run-off into a badly constucted borehole can­ Type 11 waters show significantly (typically not be excluded. Unfortunately, sample collecti­ 4 to 10 times, agreeing well with Jacks 1973) on was not, in this instance, performed by higher sodium and chloride concentrations

Fig. 11. Various hydrochemical constituents plotted against chloride concentration. Numbers refer to well numbers in Fig.14 -and Table 4. • = borehole in bedrock 0 = fracture at Urdal (nr.11) x = dug well in drift 0 = freshwater leakage In tunnel (no.9) • = seawater sample 0 = saltwater leakage m tunnel 0= storm run-oft (no.35) + R = rainfall s.am~le (no.34) SOL =seawater dilution line * =groups III ,IV [0] = alkalinity calculated from ion-balance (unreliable) .5= average seawater after Home (1969) & L10yd and Heathcote (1985) +G = precipitation at Geteborq (further south along Skaggerak coast) after Jacks (1973) 62 David Banks, Erik Rohr-Torp & Helge Skarphagen NGU-BULL422,l992

than precipitation, although similar Na/CI rati­ granite, may be derived from fluid inclusion os. This is indicative of evapotranspirative leakage. They also note, however, that such up-concentration on vegetation and in the 'exotic' saline groundwaters tend to have ele­ soil zone (together with dry fallout of salt), vated bromide concentrations relative to sea but relatively little interaction with geological water. The saline waters from Hvaler do not materials (Iow Si concentrations). Such waters exhibit this, but lie approximately on, or a litt­ are typical of wells in drift deposits. Few bed­ le below, the seawater dilution line, with re­ rock waters fall in this group. spect to bromide. Edmunds & Savage (1991) Type III waters include the majority of wa­ believe that chloride can also be derived from ters from bedrock boreholes. They show eleva­ the weathering of some minerals such as bioti­ ted Na/CI ratios, elevated Si levels but similar te. They also note, however, that in such ca­ Cl concentrations with respect to type 11 wa­ ses, there is usually a surplus of CI- ions over ters. This is indicative of mineral-water interac­ Na-: the opposite to the situation observed tion (leading to elevated Si and 1':3), but with on Hvaler. Less exotic derivations for the sali­ very little salt-water mixing. These tend to be nity of Type IV water must therefore be accep­ sodium (or occasionally calcium) bicarbonate ted, and can include: intrusion of present-day dominated waters. 'seawater' (the water surrounding Kirkeey is Type IV waters reflect some degree of mix­ in fact significantly less saline than ocean ing between Type III water and saline water water, the island lying just south of the estua­ (elevated Cl concentrations relative to type Ill, ry o~ the Glama, Norway's largest river); pum­ but reduced Si concentrations, due to seawa­ ping-induced upconing of salt water from be­ ter's relatively low Si content). low the freshwater lens; residual pre-emergen­ The four water types represent an attempt ce seawater in 'blind' or low-permeablity fractu­ to systematise the groundwater chemistry. In re systems; downward leaching of residual reality there are no sharp boundaries between saline pore water from the island's marine the water types, and the progression from Quaternary deposits (this last explanation is groups I to IV represents the evolution of unlikely, shallow wells in Quaternary deposits waters with an increasing degree of water­ exhibiting no 1ncreased salinity). rock interaction (probably reflecting retention The dug wells in Quaternary deposits (ex­ time in the aquifer) and mixing with salt water. cept for Solhell) typically yield relatively low Type lll' and IV· waters form subsets of pH, low alkalinity water. This indicates short Types III and IV. They have elevated concentra­ residence time and/or little buffer capacity in tions of most cationic species and very high the Quaternary aquifers against rather acidic contents of silicon, aluminium, iron and titani­ recharge waters. Although recharge waters um. These samples come from newly drilled are often acidic for other reasons than acidity boreholes (Testholes 1 to 4, drilled Nov.1990 in precipitation (pH of rainfall in SE Norway = and sampled during test-pumping May 1991; c. 4.3, Henriksen et al. 1987), the effects 01 and Granlie's new hole, sampled in May 1991 contaminated 'acid' rain are revealed in the day after drilling completion). It is believed groundwater sulphate concentrations. Most that interaction between the water and the of the waters lie approximately on the seawa­ newly exposed drilling cuttings, with high speci­ ter - precipitation mixing line, indicating non­ fic surface area, has led to these elevated geological sources for the majority of the sulp­ concentrations, despite filtering of the samp­ hate. The rainfall, run-off and lower water ty­ les. The anions appear relatively unaffected. pes lie above the seawater dilution line, howe­ The origin of saline waters in bedrock aqui­ ver, due to precipitation being contaminated fers is a subject much debated. While seawa­ by anthropogenic sulphur oxides. There is ter intrusion is the obvious source for an is­ reported an average of cA mg/l sulphate in land situation such as Hvaler, it must be noted the Hvaler area's precipitation (Soveri 1982). that relatively saline waters occur in the cent­ This value lies between the single point preci­ re of Kirkeoy (Fig.14), in boreholes which, at pitation sample from Hvaler (sample 34) and least according to the Ghyben-Herzberg relati­ the reported values for Goteborg (Jacks 1973) on, are not deep enough to approach the further south along the Skaggerak coast. The base of the presumed freshwater lens. Nordst­ smell of hydrogen sulphide has been detected rom et al. (1989) propose that the deep saline in at least one of the boreholes, number 18, waters in the Stripa mine, also in Precambrian indicating some degree of sulphate reduction NGU- BULL.422,1992 An integrated stUdy ofa Precambrian granite aquifer 63

Fig. 14. Map of Klrkeey and Asmaley showing location of sampled groundwaters. Small figures show sample number. Large figures show Cl-concentratlon in mg/l.

N

2k", '==='=="

LEGEND - Leakage In Hvaler tunnel Borehole /well·ln bedrock ... Seawater sample • Well In drHt deposits o Leakage from Joint in bedrock

in the aquifer. te. Assessing the importance of fluorite precipi­ The precipitation also contains a significant tation as a control on fluoride concentrations nitrate content (4.0 mg NO,'/I from the point is complicated by the uncertainty surrounding sample no.34, c. 2.6 mg NO,'/I on average fluorite's solubility product. Estimates of log according to Storre 1990). Very few groundwa­ Ks at 25°C range from -8.27 to -11.19 (Nordst­ ter samples display concentrations above 0.05 rom & Jenne 1977). Krauskopf (1979) gives a mg/l, and the surface run-off (sample 35) conta­ value of -10.4, while Nordstrom & Jenne (1977) ins only 0.22 mg/l. This indicates nitrate upta­ conclude a value of -10.96. Fig.13 shows satu­ ke or denitrification in the soil zone. rations calculated from these values (corrected A plot of alkalinity vs. pH (Fig.12) reveals a to rc using the Van't Hoff isotherm, and a typical titration-type curve. Calcium and bicar­ .1Hs of 4.7 Kcal/mole - Krauskopf 1979). It bonate concentrations are also controlled by appears that the type IV waters, at least, appe­ calcite saturation (Fig.13), many of the Type ar to approach saturation with respect to fluori­ III and IV waters apparently being saturated te, though many of the waters in lower water with respect to calcite. types do not. In the latter case it appears The granitic waters contain very high concen­ that pH is the controlling factor, fluoride con­ trations (up to 6 mg/l, compared with the SIFF centration increasing with elevated pH (Fig. (1987) drinking water limit of 1.5 mg/l) of fluori­ 12). Such a trend is noted by Englund & Myhr­ de. Possible sources for this fluoride include stad (1980), and is explained by OH' versus fluorite (observed on joint surfaces), sheet sili­ F- ion exchange on iIIites, micas, chlorites and cates (micas, chlorite), amphiboles and apati- amphiboles at high pH values. 64 David Banks, Erik Rohr-Torp & Helge Skarphagen NGU - BULL. 422, 1992

Acknowledgements Conclusions Thanks to Marit-Liv Solbjorg and Olav Schmedling of Sta­ Mapping of fractures and fracture zones in the tens Vegvesen for data from the Hvaler tunnel, and useful Iddefjord granite by means of topographical discussion thereof. Jan-Steinar Aonning. Torleif Lauritsen and Jomar Staw performed and processed the surface­ maps, aerial photos and field measurements and bore hole-geophysics. Norges almenvitenskaplige Iorsk­ has revealed an increasing degree of complexi­ ningrad (NAVF). Narske hydrologiske komite (NHK) and ty with decreasing scale. Due to the probable NGU have supported the Hvaler project financially. reactivation of the fracture pattern many times during the granite's history, it has not been possible, hitherto, to produce a simple tecto­ nic interpretation of the pattern, or its signifi­ cance for the granite's hydrogeology. Magnetic and VLF measurements have pro­ References ved the most powerful tools for identifying Acworth, A.1. 1987: The development of crystalline base­ fracture zones in the granite. Ooservations in ment aquifers in a tropical environment. Quart.JEng. ce«. 20. 265-272. the Hvaler subsea road tunnel and in connecti­ Ahlbom, K. & Smellie. J.A.T. 1989: Overview of the fractu­ on with test-pumping of NGU's boreholes on re zone project at Finnsjbn. Sweden. In Ahlbom. K. & land confirm that the presence of a geophysi­ Smellie. J.A.T. (eds.) Characterisation of fracture zone cal anomaly, or a fracture zone, does not 2, Finnsjbn study-site. SKB Technical Report 89-.19. Svensk Karnbranslehanterinq AB. Stockholm. . quarantse a highly increased transmissivity in Banks, D.C. 1972: In-situ measurements of permeability in the zone. Possible reasons for this include the granite. Proc. symp. on percolation through fissured tightening of fracture zones by secondary clay rock. Section TI-A ISRM. Stuttgart. minerals, or the non-persistence, or 'closure', Banks, D. 1991: Boring og provepumping av bydroqeolopis­ ke testhull i gronnstein - Ostrnarknesset, Trondheim. of some fracture zones with depth. Nor.geol.unders. report 91.213. Statistical analysis of 310 bore holes in the Banks. D. & Aohr-Torp. E. 1991: Hvaler-prosjekt - grunn­ Iddefjord granite reveals a distribution of yield vann i fast fJel1. Sprekkekartlegging: fddefJordgranitten. strongly skewed towards low yields. A median Nor.geol.unders. report 91.214. Banks, D.. Lauritsen, T., Skarphagen, H. & Aohr-Torp. E. yield of 500 I/hr contrasts with a mean of 1991: Hvaler-prosjekt - grunnvann i fast fjell. Baring 1124 I/hr. The median, rather than the mean, og kapasitetstesting av fire hull. ved Pulservik. Kirke­ yield should be used when assessing the pro­ oy. Nor.geol.unders. report 91.215. bable yield of a planned borehole in a bed­ Banks. D., Solbjorg. M.L. & Hohr-Torp, E. 1992: Permeabi­ lity of fracture zones in a Precambrian granite. Quart.J. rock aquifer. Eng.Geol. (accepted for publication). Test-pumping results indicate a background Bertelsen, G. 1981. Vannlekkasjer ved Ulla-Forre. Section permeability in the granite at NGU's testsite 25 in Heltzen A.M., Nilsen, B. & Mowinckel A. (eds.) of 10-' rn/s. In the top c.12 m of the granite, Fjellsprengningsteknikk, bergmekanikk, geoteknikk, 1981. Tapir farlag, Trononeim. however, transmissive fractures are more fre­ Bryn. K.0. 1961. Grunnvann est for Oslo-feltet. Nor.geol. quent, leading to apparent permeabilities of unders. 213, 5-19. 10-' - 10" m/s. Where one of the test-boreho­ Carlsson, L. & Carlstedt. A. 1977: Estimation of transrntssi­ res crossed a prominent crush-zone, a signifi­ vity and permeability in Swedish bedrock. Nordic Hyd­ rol.. 8. 103-116. cant, though not spectacular, water-inflow was Carlsson, A. & Christiansson. A. 1987. Geology and tecto­ noted at the base of the zone, and a transmis­ nics at Forsrnark. Sweden. venentsu, FUO-Rapport No. sivity of 3 x 10,' rnd was calculated. U(B) 1987/42. Statens Vattentatlsverk, Alvkarleby, Swe­ The groundwater chemistry on Hvaler is den. 91 pp. controlled by mixing between up-concentrated Carlsson. A. & Oisson, T. 1977: Water leakage in the Fors­ mark tunnel, Uppland. Sweden. Sveriges qeot.unaers.. precipitation and seawater end-members, series C. 734, Stockholm. 45 pp. coupled with water-rock interactions. Bicar­ Davis. J.L. & Annan A.P. 1989: Ground-penetrating radar bonate buffering, anion exchange and calcite/ for high resolution mapping of soil and rock stratiqrap­ fluorite saturation appear to be important pro­ hy. GeophysProspecting. 37. 531-551. Edmunds, W.M. & Savage. D. 1991: Geochemical characte­ cesses controlling pH, bicarbonate, fluoride ristics of groundwater in granites and related crystalline and calcium concentrations. No evidence of rocks. In Downing. AA & Wilkinson. W.B.: Applied cation exchange has, as yet, come to light, qroundweter hydrology, a British perspective, Clarendon the effect of seawater mixing overshadowing Press, Oxford. U.K. Ellingsen. K. 1978: Bergen. Beskrivelse til hydrogeologisk any traces of such a process. kart 1115 I - 1:50.000. Nor.geol.unders., 342. 44pp. Englund, J.-O. & Myhrstad, J.A. 1980: Groundwater charms­ try of some selected areas in Southeastern Norway. Nordic Hvdrot. 11, 33-54. NGU-BULL.422,1992 An integratedstudyofa Precambrian granite aquifer 65

Ericsson, L.O. 1988: Grunnvattenprospektering i kristallin Nordstrom, D.K., Ball, J.w., Donahoe, R.J. & Whittemore, berggrund med hjalp av satellitdata. In Grunnvatten­ 0.1989: Groundwater chemistry and water-rock interac­ terserjninq - Geohydrologi i praktiken, proc. symp. tions at Stripa. Geochim.et cosmochim. acta, 53, 1727­ 4/5/88. VIAl<. AB and SvensKa Hydr010gisKa RMet, 1740. 121- 130. Nordstrom, OK & Jenne, EA 1977: Fluorite solubility Gulbrandsen, A. 1989: Vtbytte av torundersekelse kontra equilibria in selected geothermal waters. Geochim.et undersekeiser tre stuff for undersieiske tunneler. Un­ cosmochim. acta, 41, 175-188. published thesis, Geologisk Institutt, Norges tekniske Olsen, A.B. & Blindheim, O.T. 1987: Erfaringer med stabili­ hoyskole. Trondheim. 82 pp. tetssikring og tetting av vannlekkasjer for Atesunc­ Henkel, H. & Eriksson, L. 1980: Interpretation of Iow-altitu­ Giske-tunnelene. Section 36 In Fjellsprengningsteknikk, de airborn magnetic and VLF measurements for identifi­ bergmekanikk, geoteknikk, 1987, Tapir forlag, Trond­ cation of fracture zones. In Bergman, M. (ed.) Subsurfa­ heim. ce Space, Pergamon Press, Oxford & N.Y., 913-918. Olsen, L. & Serensen, E. 1990: Fredrikstad. 1913111. 1:50000 Henriksen, A., Lien, L., Traaen, T.S. & Sevaldrud, I. 1987: kvartrergeologisk kart. Nor.geol.unders. Unpublished 1000 lake survey, 1986, Norway. Norwegian state pollu­ manus form. Due for publication 1992. tion control authority (SF. T.), Report 283/87. Olsson, T. 1979: Hydraulic properties and groundwater Horne, RA 1969: Marine Chemistry. Wiley-Interscience. balance in a soil-rock aquifer system in the Juktan area, Hult, A., Gidlung, G. & Thoregren, U. 1978: Permeabilitets­ northern Sweden. Striae, 12, Uppsala, 72 pp. bestarnninqar. KBS Teknisk Rapport No.61, Karnbrans­ Olsson, T. 1985: The Baltic Shield. Region of Precambrian tesakerhet, Stockholm, Jan.1978. crystalline basement covered by Quaternary deposits Huntoon, p.w. 1986: Incredible tale of Texasgulf well 7 and in Scandinavia. Section 4.1.1 In Persson et al. 1985b. fracture permeability, Paradox Basin, Utah. Ground Oxaal, J. 1916: Norsk granitt. Nor.geol.unders., 76. Water, 24, 5, 643-653. Pedersen S. & Maalee S. 1990: The lddetjord Granite: Jacks, G. 1973: Chemistry of some groundwaters in igne­ geology and age. Nor.geol.unders., 417, 55-64. ous rocks. Nordic Hydrol., 4, 207-236. Persson, G., Karrenberg, H., Dubertret, L., Andersen, L.J., Kluver, B.H. 1983: Undersleiske tunneler -Inqenierqeoloqis­ Bryn, K.0., De Geer, J., Kelstrup, N. & Kirkhusmo, L.A. ke erfaringer. Section 30 In Heltzen A:M., Fredriksen 1979: Explanatory notes for the hydrogeological map U., Nowacki F. & Berg K.R. (eds.) Fjellsprengningstek­ of Europe. Sheet C3, Oslo. International Association nikk, bergmekanikk, geoteknikk, 1983, Tapir forlag, of Hydrogeologists. Unesco, Paris. Trondheim. Persson, G., Kirkhusmo, L.A., Olsson, T. & Wikner, T. Krasny, J. 1975: Variation in transmissivity of crystalline 1985a: Explanatory notes for the hydrogeological map rocks in southern Bohemia. Vestnik Ustredniho ustavu of Europe. Part 1: Sheet C2, Trondneim. International geologickeho, (Prague), 50, 207-216. Association of Hydrogeologists. Unesco, Paris. Krauskopf, K.B. 1979: Introduction to geochemistry. 2nd Persson, G., De Geer, J., Egorov, S.V., Kirkhusmo, LA, edn. McGraw-Hill. Olsson, T., Panova, Z.P., rasa, M. & Wikner, T. 1985b: Kruseman, G.P. & De Ridder, N.A 1989: Analysis and eva­ Explanatory notes for the hydrogeological map of Euro­ luation of pumping test data. 2nd edn. International pe. Part 2: Sheet 02, Haparanda. International Associa­ Institute for Land Reclamation and Improvement/lLRI tion of Hydrogeologists. Unesco, Paris. Publication 47. Wageningen, Netherlands. 377 pp. Ramberg, 1.£1. & Smithson S.B. 1971: Gravity interpretation Larsen, M.L. 1990: Riksvei 108 Hvalertunnelen, geologisk of the southern Oslo graben and adjacent precambrian sluttrapport. Statens Vegvesen 0stfold, Vegkontoret, rocks, Norway. Tectonophysics, 11, 419-431. report no. 88/3744, 31 pp. Rohr-Torp, E. 1987: Beskrivelse til hydrogeologisk kart Larsson, I. 1972: Groundwater in granite rocks and tecto­ Drebak 1814 11. Nor.geol.unders. skritter, 78, 19pp. nic models. Nordic Hydrol., 3, 111-129. with map. Lauritsen, T. & Rf2Inning, J.S. 1992: Geofysiske mAlinger Ronge, B. 1988: Lokalisering av bergborrade brunnar med over mulige vannterende sprekkesoner pA Kjerk0Y, avseende pA spredtektonlska element i bergrunden och Hvaler, 0stfold. Nor.geol.unders. report 92. 173,28 pp. deras permeabilitet. In Grunnvattenff21rsf21rjning - Geohyd­ L1oyd, J.W. & Heathcote, J.A. 1985: Natural inorganic che­ rologi i praktiken, Proceedings of symposium 4/5/88. mistryin relation to groundwater, an introduction. Claren­ VIAK AB and Svenska Hydrologiska Radet, 107-115. don Press, Oxford. Huestatten, H.R., Lile, 0.£1., Veslegard, G. & Fjeld, O.C. l.eset, F. 1981: Ingenif21rgeologiskeerfaringer fra kloakktun­ 1984a: Vann i fjell prosjektet. Unoersekelsesrapport. nelen Lysaker-Slemmestad. Section 31 In Heltzen A.M., Norges tekntske hogskole report, Trondheim. Nilsen, B. & Mowinckel A. (eds.) Fjellsprengningstek­ RueslAtten, H.R., Lile, 0.£1., Veslegard, G. & Fjeld, O.C. nikk, bergmekanikk, geoteknikk, 1981, Tapir forlag, 1984b: Vann i fjell prosjektet, sluttrapport. Norges teknis­ Trondheim. kehogskole report, Trondheirn. Moye, D.G. 1967: Diamond drilling for foundation explorati­ Selmer-Olsen, R. 1964: Alminnelig geologi og ingeniorgeolo­ on. Civ.Eng.Trsns.lnst. of Engineers, Australia, April gi. Tapir forlag, Trondheim, 409 pp. 1967, 95-100. Selmer-Olsen, R. 1981: Betraktninger over store vannlekka­ Mullern, C.F. 1980: Airborne geophysical measurements sjer i dyptliggende tunneller. Section 21 In Heltzen A.M., used for hydrogeological mapping. 6th Nordiske Hydro­ Nilsen, B. & Mowinckel A. (eds.) Fjellsprengningstek­ geologiska Konferansen i Vemdalen, Ur UNGI Report nikk, bergmekanikk, geoteknikk, 1981, Tapir forlag, No. 53, 135-142. Trondhetrn. Nilsen, B. 1988: Norwegian sub-sea tunnels - A review SIFF 1987: Kvalitetsnormer for drikkevann. Statens institutt with emphasis on water leakages. In Serrano, J.M. for folkehelse (SIFF), report «Orikkevann G2". (ed.), Proc.lnt.Congr. on Tunnels and Water, Madrid, Skjeseth, S. 1981: Vassdragsregulering- konsekvenser for 12-15 June 1988, 'Oa\\<.ema, PlaneTdam, 913-918. grunn'lann og mili0. Section 20 In He\tzen A.M., Ni\sen Nilsen, B. 1990: Tunneler under fjorder og sund. Geonytt B. & Mowinckel A. (eds.) Fjellsprengningsteknikk, berg­ 3,3-5. mekanikk, geoteknikk, 1981, Tapir forlag, Trondheim. 66 David Banks. Erik Rohr- Torp & Helge Skarphagen NGU-BULL.422.1992

Storey, B.C. & Untern, B.C. 1981: Alteration, Iracture inlills Storm, G. 1990: Hydrokjemi, hydrologi og geologi i Kaldvell­ and weathering 01 the Strath Hatladale granite. Environ­ feltet. Dok.lng. thesis 1990:54, Institull for geolog; og mental protection unit, report ENPU 81-13, Institute 01 bergteknikk, Norges tekniske neqskore, Trondhairn. Geological Sciences, Harwelr, U.K. Taugbol & 0verland A/S 1987: Ingeniorgeologisk rapport; storas, I. 1988: Ingeniorgeologisk analyse av vannlekkasjer Hvalertunnelen. Report No. 2479.9, Taugbol & 0verland og injeksjon i Hvalertunnelen. Unpublished thesis, Geolo­ AIS, Oslo, 8 pp. plus diagrams. gisk Institull, Norges tekniske heyskole, Trondheim. 80 Tiren, SA & Beckholmen, M. 1989: Block faulting in southe­ pp. astern Sweden interpreted Irom digital terrain models. stcrrner. L. 1935: Contribution to the geology 01 the sout­ Geot.toreninqens i Stockholm forhandlingar. 111. Pt.2. hern part of the Oslofjord. Norsk geol.tidskr.• 15,43-114. 171-179. Soveri, J. 1982 (ed.): Acid groundwater in the Nordic Coun­ tries. Nordic Hydrological Programme NHP-report nO.3. Nordic Association for Hydrology. 29 pp.

Manuscript received December 1991; revised typescript accepted April 1992.