Hydraulic testing and modelling of a low-angle fracture zone at Finnsjön, Sweden

Journal o f Hydrology, 126 (1991) 45-77

45

Elsevier Science Publishers B.V., A m s t e r d a m

[1]

Hydraulic testing and modelling of a low-angle fracture zone at Finnsj6n, Sweden J.-E. Andersson ~, L. Ekman~, R. Nordqvist ~ and A. Winberg b~ "Swedish Geological Company, Box 1424, S-751 44 Uppsala, Sweden bSwedish Geological Company, Pusterviksgatan 2, S-413 Ol Gfteborg, Sweden (Received 3 June 1990: accepted 10 October 1990)

ABSTRACT Andersson, J.-E., Ekman, L., Nordqvist, R. and Winberg, A., 1991. Hydraulic testing and modelling of a low-angle fracture zone at Finnsj6n, Sweden. J. Hydrol., 126: 45-77. The hydraulic characteristics of a gently dipping fracture zone in crystalline rock at the Finnsj6n test site in Sweden are described. The information is derived from single-hole injection tests of different packer spacings, including very detailed tests in 0.11-m sections in a few borehole segments. In addition, detailed hydraulic interference tests were performed by pumping in different parts of the zone and monitoring pressure changes in multiple-section observation boreholes. The results of the hydraulic testing show that the fracture zone is composed of two to five thin, highly conductive subzones in the boreholes: the uppermost subzone is consistently located close to the upper boundary of the major zone and can be correlated over several hundreds of metres. There is evidence of both lateral and vertical hydraulic anisotropy along the zone. The interference tests show that the fracture zone is delimited by semi-permeable boundaries, probably constituted by steeply dipping fracture zones. The results of the hydraulic testing are used to simulate the pressure responses obtained in the observation boreholes by numerical modelling.

INTRODUCTION

The hydraulic characteristics of a low-angle fracture zone (Zone 2) are

described, together with its hydraulic interaction with the adjacent rock. The results of an analysis of the geology and tectonics (Tir6n, 1991, this issue) together with hydrogeologic interpretations are integrated into a numerical model of the Finnsj6n study site and the Br~indan area in particular. Figure 1 shows a location map of the Finnsj6n area, including the Br~indan area. A schematic cross-section through the Br~indan area is presented in Fig. 2. Hydrogeologic interpretation is mainly based on results from single-hole tests and interference testing, including hydraulic head measurements. The ~Present address: C o n t e r r a AB, Klentagfirdsvagen 34, 5-433 32 Partille, Sweden.

0022-1694/91/$03.50

© 1991 - - Elsevier Science Publishers B.V.

46

J.-E. ANDERSSON ET AL.

4t

~

~"

NE t

HFI

\

', ,, ' " - - ~ ~

~#

~ .........//',~

"ONE,~

%

/

,~- ~'~//

\~ ~~ --

t~

LAKE ~ FIN NSJON

_j, f

#,x

~

,: iI

",, ~_ ~

\\

/,

\

,..,., t :

~,

,¢r I K ..~E:-....F I O 8 .,~ '~

(~KFIO2'~ : ,,

/

\

~K,,O,

~

_...i.'.~

/

~".t ",'

.L

~

~',

,t x\,,,,

.,~* w.

\\

,-,

OSTERSYBRUK lgkm

LESENO CORE BOREHOLE

~E~

PERCUSSION BOREHOLE

:*

FINNSJON

SITE

Map sheet

ROAD

:::::

MINOR ROAD

.....

TRACK

...........

WATER DIVIDE

--

MAJOR FRACTURE ZONES MINOR FRACTURE ZONES

::;:;:;:;:;:T:;: v,~.,~

O,

--

.

.

12 I O s t h a m m a r

.

.

SOOm

1989 ~

SWEDISH GEOLOGICAL CO

Fig. 1. Location map of the Finnsj6n test site.

objectives of the hydraulic tests performed within the Fracture Zone Project (Ahlbom and Smellie, 1991, this issue) have been to characterize the hydraulic properties of Zone 2 in detail, both in a lateral and vertical direction, and its hydraulic interaction with the adjacent rock. Furthermore, the outer boundary conditions of Zone 2 and the long-term drawdown behaviour have

47

H Y D R A U L I C TESTING AND M O D E L L I N G A LOW-ANGLE F R A C T U R E ZONE

5&0

,,..

~SO0

I(X~ Blel01 KFI~ fl,cl0Z

KEI0"/

KFI@9

~ HF[03

2100m

KF[0'5 KFIt~

":~::':::'::":'~

'

.:.:.:.:.:.:.:.:.:.~+:.:.:~.:.:.~:.:.~:

!i~i~i .~i::i::i::i::::i

FRACTURE ZONES: Mojor fmcfure zones Zone 1 and Zone 2 Hinor frodvre zones

~

~ ~

ANO

80REHOLES : Locofed in front of projected profile Locoted behind projected profile

SALINE NON SALINE WATER: Non-soline woter k~J

FLOW LINES : Major fl0w lines

Tronsitiona[ woter

~

(1500-5000 mg/t £I-) within the major zones

~

Minor flow tines Recharging groundwofer

/ // /

--

Soline woter (ov.5SO0mg/[ El-) UPPER CONDUCTIVEZONE: Zone within large waterflow toy. 1500mg/IEI ) within the mojor frctcture zones

HYDRAULIC HEAD: Direction of mojor

grodient

Direction of minor gradient

Fig. 2. Cross-section (ALA ') through the Br~indan area, Finnsj6n test site.

been addressed. To meet these objectives, hydraulic single-hole tests with different packer spacings, and hydraulic interference tests using multiplesection observation boreholes were performed within Zone 2. The results of the hydraulic tests were used in conjunction with results from tracer and dilution tests performed in boreholes transecting Zone 2. The results from the latter tests were used together with the hydraulic head conditions to estimate the natural groundwater flow within Zone 2 (Gustafsson and Andersson, 1991, this issue). SINGLE-HOLE TESTING

Single-hole water injection tests have been carried out during different phases since 1977 at the Finnsj6n study site. Table 1 lists details of the boreholes used, section lengths, intervals measured and the testing period for the single-hole tests. The test sequences prior to 1985, although not specifically performed for the Fracture Zone Project, have been used for comparison with

48

J.-E. A N D E R S S O N ET AL.

TABLE 1 Compilation of single-hole tests performed in the Fracture Zone Project at the Finnsj6n study site Borehole

KFI05 KFI06 K FI07 KFI09 K FI 10

KFI11 HFI01

Section length (m)

Interval measured (m)

Test type

Year of testing

Remarks

3 2

50-743 142-318

SS SS

1979 1986/1987

Zone 2

3 2

61-679 192-292

SS SS

1979 1987

Zone 2

3 2

18-543 263-364

SS SS

1979 1987

Zone 2

20 2

10-354 109-263

T SS

1985 1986

Zone 2

T T SS

1985 1985 1986

Zone 2

T SS

1986 1986/1987

Zone 2

SS SS

1985 1985

(Zone 2)

30-450 220-448

T SS

1987 1987

10-200 200-284 201.89-206.07 211.89-214.09 257.89-262.18

T SS SS SS SS

1987 1988 1987 1987 1987

201 51 2 20 2 10 2

BFI01

20 2

BFI02

20 2 0.11 0.11 0.11

10-230 70-105, 205-235 61-225 6-386 210-360 4-124 34-44, 104-114

Zone Zone Zone Zone

2 2 2 2

~The tests were repeated at three different occasions for comparative purposes. SS, steady-state test; T, transient test.

the recent tests. The table shows that, in general, single-hole tests have been performed using both long and short test sections (20 m and 2 or 3 m, respectively). Within Zone 2, single-hole tests with a packer spacing of 2 m have been carried out in all boreholes. In borehole BFI02 a few 2-m sections were further tested in detail along 0.11-m sections. Results from single-hole tests within the Fracture Zone Project are reported by A h l b o m et al. (1986, 1987), Andersson et al. (1988a, b) and E k m a n et al. (1988).

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

49

Test methods and interpretation The single-hole tests with short packer spacing were generally performed as constant head, steady-state injection tests. An injection head (H 0 = 200 kPa) was applied for a period of approximately 15 min, after which the pressure recovery was registered (approximately 15min). However, in the most conductive test sections, rather low injection heads (a few kPa) were achieved depending on the flow capacity of the actual equipment system used. In test sections of particular interest, e.g. highly conductive intervals in Zone 2, transient injection tests of longer duration were employed to obtain more information on the hydraulic conditions and more reliable values of the hydraulic parameters. In these tests the injection and recovery periods had a duration of 2 h. In one of the large diameter boreholes, BFI02, single-hole injection tests were also performed in very short test sections within some of the hydraulically most conductive 2-m sections previously tested within Zone 2. To facilitate these measurements, a new single-hole testing tool was developed (Ekman et al., 1988). This tool has an overall length of 3.4m with a total inflation length of 3 m (Fig. 3). Centrally located is a perforated stainless steel sleeve, the perforation serving as orifices for the pipe string. The length of the test section is in this case governed by the packer inflation pressure, the elasticity of the rubber and the difference in diameter between the borehole and the deflated packer. The diameter of the packer is 150 mm. Together with the geometry of BFI01, the resulting effective section length becomes 0.11 m. Downhole testing was preceded by a surface and subsurface testing programme where the geometrical and mechanical character of the test equipment, including packer creep and head losses in the tool owing to friction, were established. Packer inflation and deflation periods during the very detailed tests were approximately 15 min, whereas an injection period of 5 min was used. The steady-state tests, constituting the majority of the 2 m and 0.11 m tests, were normally evaluated according to the following expression (Zeigler, 1976):

Kss =

QP " C

HoL

(1)

where Kss is the steady-state hydraulic conductivity (m s- I ), Qp is the flow rate during final stage of injection (m 3s-~ ), L is the length of test section (m), and H 0 is the injection head (m).

50

J.-E. ANDERSSON ET AL

155mm

65mm BOmm

~pipe

connectionfor injection wafer

~ [ [ 9 ~ connectionsfor packer inflation ~ ]-,I---compressed packer sleeve i

i:i:i;i'i;i:i: :::::::::::::: :::::::::::::: 15OOmm

~SOOmm

/packer rubber

/againsf /

pack.

bore-

hotewott

i NiiiNb-l / wnter inje~ ~ i i i i ! ]

p,pe

: ,':::":::::::1

~

~

I

/

1020 mm Scale

BOmm" 65ram:

mmt 2O

~

J~

'150ram' 50 150mm

Scare Fig. 3. Description of the test equipment used for the injection tests in 0.11-m test sections.

The constant C which takes into account the flow geometry of the test is defined by Moye (1967) as

C

=

2- ~

(2)

where rw is the radius of the borehole in metres. The definition of the constant C assumes flow in a homogeneous porous medium. Thus, the validity of the constant C may be questionable for tests in fractured crystalline rock with isolated flow paths (Doe and Remer, 1982). For section lengths of 20, 2 and 0.11 m in a borehole of 56 mm diameter, the corresponding values of the constant C are 1.09, 0.73 and 0.27, respectively. As it is known that only a limited number of fractures dominate the total flow rate in each section in crystalline rock, the apparent differences in results using the calculated C value for the 0.11 m tests are considered as non-representative

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

51

(cf. below). This is also confirmed by comparisons between transient and steady-state flow evaluation of these tests. Therefore, the same C value was used for the 0.11 m tests as for the 2 m tests. The C values for the 2 m and 20 m tests were calculated from eqn. (2). The transient tests were evaluated in accordance with the actual flow geometry of the tests (Andersson and Persson, 1985). The dominating flow geometry (linear, radial or spherical flow) was identified from the different test graphs. The theory and methodology of the transient tests is described by Alm~n et al. (1986b).

Results The results of the 2 m tests in boreholes within the Brfindan area (Fig. 1) reveal large contrasts in hydraulic conductivity within Zone 2, ranging from minimum values of 10-~°ms -t (lower measurement limit) up to about 10-3m s -1 . Between two and five highly conductive subzones were identified in each borehole; two of the subzones were consistently located along the upper and lower boundaries of Zone 2. The results from the 20 and 2 m sections in borehole BFI02 are presented in Fig. 4(a). To facilitate comparison of the tests, the transmissivity (T = KL) of the sections is plotted in the figure. The results indicate an apparent, rather smooth transmissivity profile from the 20-m sections above the zone. This pattern is altered significantly when the resolution in the section length is increased one order of magnitude within the zone. Two interpreted subzones in borehole BFI02 are indicated with (steadystate) transmissivities of about 8 x 10-4mZs -1 . This value is considered as the upper measurement limit of the actual equipment system used for the 2 m tests. Detailed measurements along 0.11-m test sections incorporating some of the subzones were performed in segments over a total length of 8 m in borehole BFI02 (Table 1). In Fig. 4(b) the results of the 0.11 m measurements in the two successive 2-m sections incorporating the uppermost subzone (sections 202204 m and 204-206 m) are presented in terms of transmissivity. The figure shows several interesting features, for example, the sum of transmissivities calculated from the two 2 m tests is in good agreement with the cumulative transmissivity obtained from the 0.11 m tests over the same 4 m interval (202-206 m). In fact, the uppermost subzone only consists of a few 0.11-m sections which govern the total transmissivity of the entire 4 m interval (and a major part of Zone 2). The increase of resolution (scale) of the testing to 0.11-m sections thus provided the possibility to define the 'true hydraulic thickness' of the uppermost subzone more accurately, which was estimated at about 0.4 m. The

~

Borehote teng'lh(m)

~" ...]

~"

'"'tn~l

~ ~-

. . . . . .

5. r<

-

~

IJa

v

~,

.7~

O

~o

) Zone 2

~

.~.~ t.~

3

c~

=.

C~

~ ~

Borehote teng~h (m)

I

g~

,.,

=.~

,,.

~

~

,-

~

m

Lo~er mmsurerner~ tlm,t

~1~

r ~"

,

,~"

t~

c~

~"

"qV .L~I NOSS~I3C1NV "H-'f

E~

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

53

main flow path within the subzone may be even narrower. The transmissivity values show a gradual increase towards the central part of the subzone. In addition, the fact that the subzone was located at the limit of two 2-m test sections thus tended to obscure the thinness of the subzone in the 2 m tests. If a longer section length had been used for testing (e.g. 20 m), an even more misleading impression of the true thickness of the subzone might have resulted. The single-hole tests in the other boreholes within the Br~indan area show the same general pattern for Zone 2, i.e. comprising a few highly conductive intervals (subzones). The very detailed tests in BFI02 revealed that the thickness of the subzones is of the order of decimetres. This is also consistent with the core mapping which indicates a limited number of narrow zones with increased fracturing and deformation (Tir~n, 1991, this issue). In general, between two and five subzones are present within the major extent of Zone 2; the total thickness of Zone 2 is more than 100 m. In all boreholes the hydraulic conductivity within the upper boundary of Zone 2 contrasts markedly with the overlying rock (two to three orders of magnitude). In most boreholes a large hydraulic conductivity contrast is also present between the lower boundary of Zone 2 and the underlying rock (Ahlbom et al., 1987). As shown by the subsequent interference tests, these features are laterally consistent over several hundreds of metres. Possible vertical interconnections between the subzones were not investigated during the single-hole testing campaign but instead during the interference testing (see below). In summary, the single-hole tests showed that the hydrogeological character of Zone 2 mainly consists of long intervals of rock with low to intermediate conductivity, separated by a few highly conductive, narrow subzones. INTERFERENCE TESTING

During the drilling of borehole BFI01 the groundwater head was registered in isolated sections of the other boreholes within the Br~indan area (cf. the hydraulic head measurements described by Gustafsson and Andersson, 1991, this issue). As water was flushed out of borehole BFI01 by compressed air at a fairly constant rate during the drilling periods, these periods could be regarded as (preliminary) interference tests (Ahlbom et al., 1987). Between the drilling periods the groundwater head in the boreholes was allowed to recover. Subsequently, more sophisticated interference tests were carried out by pumping from different parts of Zone 2 in borehole BFI02 and monitoring the head changes in isolated multiple sections of boreholes within Zone 2, and in a few open boreholes outside the zone (Andersson et al., 1989). The latter interference tests are described below.

54

J.-E. ANDERSSON ET AL.

Design and performance After the drilling of borehole BFI02, three separate interference tests were performed in this borehole by pumping in different parts of Zone 2 (lower part, upper part and the entire zone). The primary objectives of the detailed interference tests were to determine the hydraulic properties of Zone 2 in more detail, both in a lateral and vertical direction, and to investigate the outer hydrogeological boundaries of Zone 2. Moreover, possible hydraulic connections in the lateral direction of the most permeable horizons of Zone 2 between boreholes, and the hydraulic interaction between the zone and the overlying and underlying rock, were to be investigated. The design, performance and results of the detailed interference tests are described by Andersson et al. (1989). By the design of the interference tests, the results of the single-hole tests and the preliminary interference tests during drilling of borehole BFI01, together with associated numerical simulations, were utilized (see below). As discussed above, the dominating hydraulic features of Zone 2 in the pumping borehole BFI02 (and the other boreholes) comprised the uppermost and lowermost parts of the zone. It was therefore decided to pump these parts individually and subsequently the entire zone. A schematic illustration of the packer configuration used in BFI02 during the different interference tests is shown in Fig. 5. In the observation boreholes KFI05, KFI06, KFI09, KFI10, K F I l l , HFI01 and BFI01, a varying number of highly conductive horizons were BFI02

Ob . . . . .

3

INTERF.TEST 1

BFI02

BFI02

,on

~

6

INTERF.TEST 2

INTERF.TESTS 3AaB

Packerpos.:245-2/,6m packerpos.:192-193m Pockerpos.: 192-193m 270-271m 217-21Bm Pumpedsecf.:246-2"/Om Pumpedsecf.:lg3-217m Pumpedsect.:193-ZB6,69m

~

Borehote wail (~)Po~:ker Pipestring for wafer (~Zone 2 -- discharge. (~)Pump C.~Subzone I

~)Subz.one 3 8~ Pocker 9~ Boreho[e bo~-l'om

Fig. 5. Schematic illustration of the equipment assembly and packer positions together with the numbering of the observation sections in the pumping borehole BFI02 during the detailed interference tests.

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

55

isolated by multiple packers. The packers were installed in rock of low conductivity encompassing the highly conductive horizons so as to minimize the risk of pressure bypass near the boreholes. In the observation boreholes nearest to the pumping borehole, double packers were used at the top and bottom of each highly conductive horizon to ensure an effective hydraulic isolation. In addition, in most of the observation boreholes one section just above Zone 2 and one section just below the zone were isolated by packers. In all packed-off boreholes the variations of the groundwater table at the top of each borehole were also monitored during the interference tests. The observation intervals in the packed-off boreholes, together with their distances (in space) to the pumped borehole intervals in BFI02 (during interference test 2), are listed in Table 2. Moreover, the vertical position of the observation sections within Zone 2, and the upper and lower boundaries of Zone 2 in the boreholes as interpreted from the single-hole tests, are included in Table 2. Note that borehole HFI01 does not fully penetrate Zone 2 and that no pressure registrations were undertaken in borehole KFI05 during interference test 2. The peripheral boreholes KFI01, KFI02, KFI04, KFI07, KFI08 and HGB02 were used as open observation boreholes during all interference tests (Fig. 1). In borehole BFI02, a submersible p u m p was installed within the pumped interval to maintain a constant discharge rate. The groundwater pressure in this interval, and that above and below, was measured by pressure transducers. The pressure within the pumped interval and below was transmitted via hydraulic tubes to the transducers. The groundwater pressure in the packed-off observation boreholes was monitored by a multi-packer system described by Alm6n et al. (1986a). The water discharge from the pumping borehole was controlled automatically by a flow-rate regulator. The groundwater flow rate, together with the electric conductivity, temperature, atmospheric pressure and chemistry of the discharged water, was monitored continuously at the well site. All the automatic measuring devices were connected to a data acquisition system. In addition to the automatically registered data, manual registrations of the groundwater table were also undertaken in all boreholes (including the peripheral boreholes). The pumped sections in BFI02, together with the discharge rates and duration of the drawdown and recovery phases of the different interference tests, are listed in Table 3. Test 3 was divided into test 3A and test 3B. At the completion of test 3A the automatic flow regulation system failed and the flow rate increased from 5001 min-~ to about 7001 min -~ . After recovery the test was restarted (3B) using a manual flow-rate control. Before the start of interference test 2, tracers were injected into specific observation sections within Zone 2 (boreholes KFI06, KFI11 and BFI01) in

56

J.-E. ANDERSSON ET AL.

TABLE 2 Observation borehole intervals during the interference tests together with their distance to the midpoint of the pumped sections in BFI02 (during interference test 2) and position within Zone 2 Borehole

Observation section (no.)

Interval (m)

Distance to BFI02 (m)

Part of Zone 2

KFI05

5 4 3 2 1

0-162 163-189 227-240 241-296 297-751

309 245 217 206 293

Above Upper Middle Lower Below

M 5 4 3 2 1

0-165 166-201 202-2271 250-259 260-279 293-691

189 189 195 200 345

Above Above Upper Middle Lower Below

M 5 4 3 2 I

0-100 101-118 119-1511 152-1881 189-230 231-376

286 288 294 302 339

Above Above Upper Middle Lower Below

5 4 3 2 1

0-75 76-134 139-1581 159-193 194-255 ~

391 331 294 272 234

Above Zone 1 Upper Middle Lower

M 5 4 3 2 1

0-135 200-216 217-2401 285-304 327-340 341-390

153 155 177 200 222

Above Above Upper Middle Lower Below

3 2 1

0 50 51-81 ~ 82-1291

367 349 338

Above Above Upper

M 5 4 3 2 1

0-218 219-238 239-2501 261-270 345-364 364-459

165 168 174 221 265

Above Above Upper Upper Lower Below

KFI06

KFI09

KF110

KFI11

HFI01

BFI01

Zone 2 interval (m)

162-290

214-269

134-214

142-210

222-338

108-?

241-356

tlntervals correspond to primary response section(s) in the boreholes during interference test 2. The upper and lower boundaries of Zone 2 in the boreholes are shown. M, m a n u a l registration.

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

57

TABLE 3 Active borehole sections in BFI02 together with discharge rates and duration of the different phases of the interference tests Test no.

Phase

Active section (m)

Discharge rate (1 min ~)

Duration (days)

1 1

Drawdown Recovery

246-270 246-270

500 -

3.7 5.2

2 2

Drawdown Recovery

193-217 193-217

500 -

7.8 10.2

3A 3A

Drawdown Recovery

193-289 193-289

500-700 -

0.9 0.8

3B 3B

Drawdown Recovery

193-289 193-289

700 -

4.0 8.2

order to investigate possible anisotropic conditions in different directions from BFI02. Furthermore, by the end of the pumping phase of test 2, an additional tracer was injected below the lower packer in BFI02 to check for possible short-circuiting of water around the packer. No short-circuiting was detected. The tracer experiments are discussed by Gustafsson and Andersson (1991, this issue).

Qualitative interpretation To assist in selecting a proper model for the quantitative interpretation of the interference tests, a qualitative interpretation of the test results based on the observed data curves was first made. Some general results, c o m m o n to all three interference tests, were initially deduced and then the differences between the tests were taken into consideration. As interference test 2 (in the uppermost part of Zone 2) had the longest duration (Table 3) it is discussed in more detail than the other tests. The spherical distances from the midpoint of the pumped interval in borehole BFI02 to the midpoints of each observation section during interference test 2 are shown in Table 2. By calculating the distances, the deviation of the boreholes was taken into consideration. Figure 1 shows that the subvertical boreholes KFI06, KFI11 and BFI01 are located at similar distances (approximately 150-190m) but in different directions from BFI02. The other observation borehole sections (within Zone 2) are located at about 250-350 m from BFI02. The observed head changes in the different observation sections of borehole KFI11 and in the more distant borehole KFI09 during interference test 2,

58

J.-E. ANDERSSON

ET AL.

. . . . . . . . . ~: ..... ....... 4

I

/.,,

, ~ '~ . "

B

"''°°o--, o=.nQ..ea x.xx.~x~x

•

i

~;'. .

" .

"

* ........

.

.

.

.

.

.

.

.

SECTION SECTION SECTION SECTION SECTION

S above 4 3 z~ne2 2 I Below

i !

lo ~t

.

.

.

.

.

.

.

i

.

.

tO ~

.

.

i

.

.

.

tO

.

.

i

.

I

'

tO *~

'

[Time

i

(min)

]

Fig. 6. Drawdown responses in multiple sections of observation borehole KFII I during interference test 2 (uppermost subzone) in a logarithmic graph.

which are shown in logarithmic graphs in Figs. 6 and 7, represent the nearand distant-region borehole responses, respectively. Figure 6 shows that the near-region responses are markedly separated in the different borehole sections. The section exhibiting the most rapid and largest head change is

i.:':"

I O *~

.

I. ,

:

*~ ~

,

e "

i

.

i- " "

.

.

;

-

10"~

*

//!

."

/

i

i

° • ........

S E C T I O N S Ahoy, SECTION 4 ) S E C T I O N 3 zone z SECTION 2 I

:

°1

*a****.**

I

,xx..xxxx

!

......... SEETIO . . . . . . .

I

!I

,

i I i

i

tO *°

,

....

,

10

....

,

10 ~2

....

,

....

l

[ T i m e (min) ]

Fig. 7. Drawdown responses in multiple sections of observation borehole KFI09 during interference test 2 (uppermost subzone) in a logarithmic graph.

HYDRAULICTESTINGANDMODELLINGALOW-ANGLEFRACTUREZONE

59

likely to represent the primary pathway between the pumped section and the actual observation section (primary response). During interference test 2 the primary response in K F I 11 occurs in the uppermost (pumped) part of Zone 2 (section 4). The other borehole section responses in KFI11 should then represent more diffuse pathways, resulting in somewhat more delayed and attenuated (secondary) responses in the lower parts of Zone 2 (sections 2 and 3). The primary response section(s) in each observation borehole during interference test 2 is marked with an asterisk in Table 2. Figure 6 also shows that the observation section above Zone 2 responds rapidly whereas the section below the zone responds more slowly. However, it is uncertain if a sufficient hydraulic isolation was obtained between Zone 2 and the section measured above to obtain representative pressure responses in this section. Figure 7 shows that the more distant-region responses in borehole KFI09 are less separated from each other. No pronounced primary response section occurs in this borehole but all sections within Zone 2 respond similarly (and similarly to the secondary responses in borehole KFI11). This indicates that the zone responds as an equivalent porous medium beyond a certain distance from the pumping borehole. The section above (and also below) Zone 2 responds more slowly in borehole KFI09. Figures 6 and 7 indicate that the induced (primary) pressure wave propagates very rapidly from BFI02 to the observation boreholes (cf. KFI 11). This demonstrates a high transmissivity (and low storativity) in the lateral direction of Zone 2. On the other hand, the non-uniform responses between sections in the same boreholes, particularly in the near-region, indicate a decreased hydraulic conductivity in the vertical direction of Zone 2. The differences in head change between sections in the vertical direction of Zone 2 seem to be persistent in most boreholes both in time and with distance, although the differences decrease with the distance from BFI02. Despite the lower hydraulic conductivity in the vertical direction of Zone 2, Figs. 6 and 7 show that hydraulic interaction does exist in the vertical direction of the zone. This is also supported by an increase of electric conductivity of the discharged water during interference test 2, indicating a significant contribution of water from the lower, more saline parts of Zone 2 to the upper, pumped part of the zone (Smellie and Wikberg, 1991, this issue). During interference test 2 the primary responses in the observation sections almost exclusively occurred in the uppermost (pumped) part of Zone 2 (Table 2). This indicates that the highly conductive uppermost horizon of the zone is continuous in the lateral direction and can be correlated over large distances. Accordingly, during interference test 1 the primary responses generally occurred in the lowermost horizon of Zone 2, which also seems to be persistent with distance. Borehole BFI01 is an exception: in this case the

60

J.-E. A N D E R S S O N ET AL.

primary response occurred in the uppermost part of Zone 2 during both tests 1 and 2. This may indicate deviating hydrogeological properties (e.g. a fault) between this borehole and BFI02. In general, during all interference tests carried out the sections above and below Zone 2 responded much slower than the sections within the zone. The primary drawdown response (most rapid response section) in each borehole during interference test 2 is shown in semi-logarithmic graphs in Fig. 8a, b. The graphs clearly show that the shape of the primary response curves is very similar in all boreholes. In the pumping borehole BFI02 the drawdown is delayed up to about 8 min owing to technical reasons. Although the actual primary drawdown differs somewhat between the near-region observation boreholes, it is almost identical in the distant-region boreholes. The rate of drawdown is virtually the same in most of the boreholes, independent of the distance to BFI02. Thus, within the investigated area, the rate of drawdown is nearly constant at all points, i.e. the hydraulic gradient is constant. This fact indicates that Zone 2 is delimited by hydraulic boundaries (i.e. less conductive rock) in the lateral direction (Earlougher, 1977). This is further confirmed by the steep shape of the drawdown curves at intermediate times, as shown in Figs. 6 and 7, suggesting that Zone 2 is surrounded by barrier boundaries. This is consistent with the geological interpretation which describes Zone 2 as a triangular-shaped area (Tir6n, 1991, this issue). Figures 8(a) and 8(b) also show that the primary drawdown responses in the near-region observation boreholes (KFI06, K F I l l and BFI01) generally deviate somewhat from the other boreholes (independent of the distance to BFI02). This may be a reflection of local heterogeneities between these boreholes and BFI02 (Gustafsson and Andersson, 1991, this issue). By the end of interference test 2 the drawdown curves in Figs. 6 and 7 level out and an approximate steady-state is reached indicating a major (external) source of recharge to Zone 2. The location and nature of this recharge source is not known but it may derive from other fracture zones within the area. For example, water may be transmitted along Zone 1 which extends southwest towards Lake Finnsj6n (Fig. 1). Other fracture zones in the area may also be potential recharge sources to Zone 2. During interference test 3 the entire Zone 2 was pumped below a single packer. Simple balance calculations, based on the resulting electric conductivity of the discharged water during the test, indicate that approximately the same quantities were discharged from the upper and lower parts of Zone 2 during this test. As an example of a near-region response the drawdown in borehole KFI11 during test 3A is shown in Fig. 9. As the flow rates during test 2 and test 3A (up to approximately 13.5 h ) w e r e the same, Figs. 6 and 9 are directly comparable up to this time. Figure 9 shows, as expected, that

HYDRAULICTESTINGANDMODELLINGA LOW-ANGLEFRACTUREZONE . . . .

' Yi

I

i

~iO

.

,

I

. . . .

m

I

,

i

~

i

I

i

i

~z~

i ii

~

....

61

I

]

....

1

o.

y KFlll I ,, ,___.C-~'r ~"

....

. ....

O,

o i

. . . .

i

,r,,

t0

I0+*

i

. . . .

i

....

I

[Time (min) ]

iO+~

(a) . . . . •

Yi

I

I

• iO

. . . .

I

r

~

m

,

*

I

' ' ' l l

1

. . . .

I

I

. . . .

I

~

o

KFIIO, ~

~

ul o

,

• ii~i

0

~

i

#

l

t

I

/ I Olll~

tBI

o '

i0 +*

'

'

I

.

.

.

.

.

.

.

I

i0

....

I

i0 +2

....

I ....

[Time (min)

(b) Fig. 8. D r a w d o w n o f primary responses in the boreholes during interference test 2 (uppermost subzone) in a semi-logarithmic graph: (a) boreholes BFI02 (pumping borehole), BF]01, K F I 0 6 and K F I I I

(b) boreholes KFI09, KFII0 and HFI01.

the response curves in K F I l l are less separated during test 3A when the entire zone is pumped. In the distant-region observation boreholes the drawdowns are very similar during the beginning of test 2 and test 3A. The response pattern (order of response among sections) and the primary

62 lO

J.-E.

• .

.

.

.

.

~Y:

.

[

~

....

1

,

,

1,~,,I

,

,

I

,,,,1

[glgo,.

L

,

,

I

ET

AL

,,,,r

r

, ........... 4

ANDERSSON

'

i

1 lo'°~_

o o

,;;~

==.0===== =~,~.~xx *~.4,**÷.

I0- i ~

.........

SECTIOH 4 SECTIOH 3 Z o n e Z SECTION 2 SECTION i Below

i ]

,

I0 -~

I 0 '°

IO

,

q , ,,

[Time (min)

Fig. 9. Drawdown responses in multiple sections of observation borehole KFII 1 during interference test 3A (entire Zone 2) in a logarithmic graph.

response section in each observation borehole are very similar during these tests. In the peripheral observation boreholes, significant responses were obtained in boreholes KFI04, KFI07, KFI08 and HGB02 (Fig. 1). During interference test 3B, drawdowns of approximately 0.7 m, 1.6 m, 0.8 m and 0.7 m, respectively, were measured in these (open) boreholes. The responses were, however, considerably delayed and occurred after more than about 25 h of pumping. The distances to BFI02 from these boreholes are respectively about 920, 800, 1310 and 1260m. Small drawdowns (about 0.1 m) were also measured in boreholes KFI01 and KFI02 located 1400-1500m from BFI02. The significant responses measured in most of the peripheral observation boreholes outside Zone 2 indicate a certain hydraulic communication across the steeply dipping Zone 1 (Fig. 1). The delayed responses may indicate pressure propagation via surficial parts of the bedrock caused by the relatively large drawdown of the (free) groundwater table that was observed in all boreholes within Zone 2 during the interference tests. To summarize, the qualitative interpretation shows that the near-region responses (drawdowns and response patterns) within Zone 2 differ significantly between the interference tests, whereas the distant-region responses are very similar in all three tests. Furthermore, the transmissivity in the lateral direction of Zone 2 is shown to be very high and the uppermost subzone is hydraulically connected over large distances. The drawdown differences

H Y D R A U L I C TESTING AND M O D E L L I N G A LOW-ANGLE F R A C T U R E ZONE

63

between observation sections in the same boreholes indicate a stratified nature for Zone 2 in the vertical direction. Nevertheless, the test responses clearly show that hydraulic interaction occurs within the entire zone. The qualitative interpretation also shows that Zone 2 is surrounded by hydraulic (barrier) boundaries. By the end of test 2 an approximate steady state was reached indicating major sources of inflow to Zone 2 after long pumping times.

Quantitative interpretation Quantitative interpretation of the interference tests is based on the results of the qualitative interpretation discussed above. This means that the quantitative interpretation has taken into account vertical anisotropy and the outer hydraulic boundaries of the aquifer. As both the single-hole tests and interference tests have documented the uppermost and lowermost parts of Zone 2 as being highly conductive and persistent in most of the boreholes, theoretical models for a two-aquifer system separated by a semi-permeable layer may be considered by the quantitative interpretation of tests 1 and 2. Such models have been presented by, for example, N e u m a n and Witherspoon (1972) and Hantush (1967). In the model of Hantush the storage capacity of the assumed semi-permeable layer is neglected. This seems to be a reasonable assumption in this case as the flow transfer between the different subzones of Zone 2 is likely to be controlled by discrete fractures of low storage capacity. The model of Hantush (1967) is thus mainly used for the quantitative interpretion of the interference tests. An alternative approach would be to treat the entire zone as a homogeneous aquifer with hydraulic conductivity anisotropy in the vertical direction. Using this interpretation, the uppermost and lowermost horizons of Zone 2 can be treated as highly conductive subaquifers separated by a fictive semi-permeable layer, representing the postulate decrease of hydraulic conductivity in the vertical direction of Zone 2 (Fig. 10). One of the subaquifers is assumed to be pumped while the other remains unpumped and vice versa during tests 1 and 2. For example, in test 1 the lowermost part of Zone 2 is regarded as the pumped aquifer whereas the uppermost part of the zone is treated as the u n p u m p e d aquifer. Figure 11 illustrates an example of theoretical type curves for the drawdown behaviour in such a leaky two-aquifer system of infinite areal extent. Thus, the type curves do not take the bounded nature of Zone 2 into account. The evaluation of the hydraulic parameters must therefore be based on prior conditions when effects of the boundaries have not become appreciable. Ideally, a model which takes the surrounding hydraulic boundaries into account should be used. However, the exact positions and nature of the

64

J.-E. ANDERSSON ET AL. --0.

initial piezome, ric

~

surface (21 initial piezomefric surface ( 1 )

'

51 ,T 1

f":"6 NO |

0quifer 2 .........

I S 2,T 2 "~ .I ....

-r

Fig. 10. Schematic representation of a leaky two-aquifer system (after Hantusch, 1967).

hydraulic boundaries are difficult to deduce. An example of the time-drawdown analysis using the type curves shown in Fig. 11 is presented in Fig. 12 for borehole KFI 11 from interference test 1. The figure shows that the effects of the hydraulic boundaries become appreciable after about 10 min of pumping. This is consistent with the response of a high transmissivity aquifer with outer hydraulic boundaries. W(u) + W(u,B) W(u) - W(u,B)

I

~o

I

TYPE CURVES FOR AN INFINITE LEAKY AQUIFER SYSTEH.

:0,5 10t

-.:... g" s~:,

10 10

10

10

10

t

10

1/u

Fig. 11. Examples of type curves for a leaky, infinite two-aquifer system shown in Fig. 10.

HYDRAULIC

TESTING

, I ....

10

I

YI

AND

MODELLING

I ....

m

I

A LOW-ANGLE

FRACTURE

I ~,,,I

65

ZONE

t ,,,,I

;2,'. ....

J

~

TESTI

-3 2 T2=T1 =1,3.10 m~/s

, I ~

|

,, . ~

.........

=s1 =,71o°

,

teoky aquifer system

PIl, ,J 10"*

°.

°

Sm: o,2sm

~

tm = O~Smin

~

K',b'

Brn=O,6

/ l ....

i0 *°

F .

10

.

.

.

lO *z

.

.

.

I . . . . .

[

[Time (rain) ]

Fig. 12. Example of time~Irawdown analysis by type curve matching from borehole K R I I ! during interference

test

l.

In addition to the time-drawdown analysis described above, semi-quantitative distance-drawdown analyses of the primary responses were also performed. The former analysis provided estimates of the transmissivity (and hydraulic conductivity) and storativity of the two subaquifers from the near-region observation boreholes. The distance-drawdown analysis gave rough estimates of the total transmissivity and storativity of the entire Zone 2 from the distant-region observation boreholes. The interpretation using analytical methods was accompanied by numerical predictions of observation borehole responses. These were based on the preliminary interference tests during drilling of borehole BFI01 and numerical simulations of the detailed interference tests in BFI02. The numerical predictions and simulations are presented in the following section. The transmissivity of the uppermost and lowermost parts of Zone 2 was estimated at about (1-2) × 10-3m2s -] from the time-drawdown analysis both for tests 1 and 2. These values are in good agreement with the values obtained from the single-hole tests (see previous section). The corresponding storativity values from the near-region observation boreholes differ between the tests. From test 1 the storativity of the lowermost part of Zone 2 was estimated at (4-7) × 10 6. From test 2 the storativity of the uppermost part of Zone 2 ranges from 6 x 10 -7 to 2 x 10 -6. In both tests, apparently higher storativity values were calculated from BFI01 indicating deviating hydrogeological conditions between this borehole and BFI02 (Gustafsson and

66

J.-E. ANDERSSON ET A L

Andersson, 1991, this issue). From the distant-region observation boreholes the transmissivity and storativity of the entire Zone 2 were estimated at (2.5-3.5) x 10 - 3 and (1-2) × 10 -5, respectively, both from the timedrawdown and distance-drawdown analyses. From the separation of the data curves representing the pumped and unpumped aquifers, a rough estimation of the hydraulic conductivity in the vertical direction of Zone 2 was made. Assuming a thickness of the fictive low-conductivity layer of 50m, the equivalent (average) vertical hydraulic conductivity of this layer may be estimated at about (1-2) × 10-6ms -I. This estimate is consistent with rough estimates of the leakage rate between the subaquifers during tests 1 and 2, based on simple balance calculations from the change in electric conductivity (salinity) during the tests and knowing the initial concentrations in each subaquifer (Andersson et al., 1989). NUMERICAL MODELLING

Objectives and approach Numerical flow modelling was performed to obtain, if possible, an accurate description of the geological and hydrogeological conditions that govern groundwater flow in the Finnsj6n area. Two specific objectives were identified: (1) to use the groundwater flow model to verify geological interpretations and hydrologic evaluation of interference test data, and (2) to use the groundwater flow model to predict results from future tracer tests. Only the first objective has been considered here. Available geological and hydrogeological data have been used as input to the flow model so that the model results can be compared with actual field observations from the interference tests. By analysing the discrepancies between the calculated and measured results, the flow model can be updated by adjusting hydraulic parameters. By repeating this calibration process in an iterative manner, a satisfactory agreement between model results and field observations is achieved. The numerical modelling described here consists of two distinctive phases: one before and one after the interference tests. The first phase was aimed at the prediction of drawdown distributions in the investigation area based on somewhat limited data. The second was to study the discrepancies between predicted and measured interference test results, and improve the model performance by an updated calibration. The updated model was calibrated for test 3B only, and therefore only the prediction for test 3B will be discussed here.

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

67

The groundwater flow is assumed to take place in a porous saturated medium with a constant fluid density, and is governed by the equation (Freeze and Cherry, 1979)

-~x Tx~x

+-~y Ty

-Q

= S-~

(3)

where h is the hydraulic head (m), Tis the transmissivity (m: s-~ ), Q is the fluid sources or sinks (m 3s-J), and S is the storativity. Equation (3) was solved numerically by the 2-D finite-element simulation code SUTRA, version 1284-2D (Voss, 1984).

Predictive modelling prior to interference tests The predictive modelling of the hydraulic interference tests was performed in a vertical profile, assuming radial symmetry around the pumped borehole, BFI02. The computational domain and its assumed hydraulic parameter distribution is shown in Fig. 13, where Zone 2 appears as more or less permeable subzones between depths of approximately 150 and 250 m. The parameter values are primarily obtained from single-hole water injection tests. In addition, some model calibration was carried out based on very limited drawdown data (from borehole KFI09) from pumping during drilling of borehole BFI01. Fluid discharge was distributed along the left boundary in accordance with the extent of the pumped section for each test. Predicted versus measured drawdown time series (primary responses) in two of the observation holes for test 3B are compared in logarithmic diagrams in Figs. 14 and 15. Also included are predicted responses above Zone 2, shown as solid lines without symbols. Examination of Figs. 14 and 15 reveals two striking discrepancies between predicted and measured results. Firstly, the magnitude of the total drawdowns is relatively accurately predicted only for the borehole (HFI01) farthest away from the pumped hole. At shorter distances ( K F I l l ) , predicted drawdowns are significantly higher than measured. Secondly, the general shape of the predicted drawdown curves in all the boreholes does not conform to the measured shape. The fact that the total drawdowns (when approaching steady-states) are not accurately predicted at all distances, may be seen as a failure of the prediction model to describe the shape of the cone of depression created by the pumping. In this case, the actual slope of the cone was considerably less than predicted. This might be explained by the fact that the values of transmissivity within Zone 2 used for the predictive modelling were too small. The difference in shape of the drawdown curves is taken as an indication that the physical geometry of the modelled region is not sufficiently detailed.

68

J.-E. A N D E R S S O N ET AL.

Apparently, the hydrogeological boundaries of the model region are more complex than what the relatively simple model geometry accounts for. The conclusions from the predictive modelling indicate that a more detailed description of the hydrogeological boundaries is required to improve the model performance with respect to prediction of hydraulic heads during pumping tests. Furthermore, hydraulic parameters within Zone 2 need to be revised.

Model calibration after interference tests In order to obtain an updated and improved groundwater flow model for the Brfindan area, more detailed geological features were incorporated into the model (Tir6n, 1991, this issue). This resulted in a computational domain, as shown in Fig. 16, where the fracture zone is modelled as a subhorizontal plane. Zone 2 is generally surrounded by a less-transmissive rock mass and

900

0

"CO

200

300

~0

500

bOO

'700

800

900

T ,o

8 O0

-"

,,

600

?o ~

700

t/t

"Y o

?oo

6O0

5~

600 o

o

~

,,i,

o

o

~

o

,.,,

,.,,

too

lO0

100

"

SO0

~,00

~,

,,,

°

~

100

Fig. 13. Parameter distribution used for the predictive numerical mode] prior to the interference tests,

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE ,,r,I

,

YJ-

,

I

m

lO

r,~,l

,

,

[

,,,,T

r

!~FZtt

/

:TEST3B iYI:H~, H~ H~ H4 H~ 1 16eo3z6taooo

J

o

o

,

I

,,,,1

69 ,

,

I

,,~r[

°;:,~;;.

.LO*~ *

o

"I

°,,*o.*** .°=°...D= *w~x~x. +*++***** .........

!

SECTION5 Abo~e SECTION 4 } SECTIOM 3 zo°e z SECTION 2 SECTION 1 eelo*

!

tO-~" ~ I

r~

tO +°

tO

iO +a

[Time [m/rl) ]

Fig. 14. Predicted (solid line) and measured drawdown responses in observation borehole K F l l l during interference test 3B in BFI02.

DR*~OOWN TEST3E ~leeo~lst2ooo

J

i ~ ,

tO*°

I_0

......... TI ~

SECTION1 z~ 2,u~ part

i

IO*a

[Time (min) ]

Fig. 15. Predicted (solid line) and measured drawdown responses in observation borehole HF]01 during interference test 3B in BFI02.

70

J.-E. ANDERSSON ET AL.

6~VASTBO ZONE

BRANOAN ZONE NO PLOW

0

200

&00

600

CONSTANT HEAO'O

800

1000

1200

%00

NO FLOW

1600

1800

2000

2Z00

2~00

2600

o KFI0t, o KFT07

No..ow izoo

~oo

16oo

~soo

,

,

zooo

zzoo

o

\\ ,

z~oo

\

[

/

z6oo

Fig. 16. Layout and boundary conditions applied for the calibrated numerical model.

appears as a triangular shape in the middle of the figure. Two vertical fracture zones, the Br/indan and G~vastbo zones, are also represented as physical entities. Boundaries to Zone 2 are either no flow or semi-permeable. It can be noted that boundary conditions for the entire computational domain are either no flow or constant hydraulic head. Groundwater recharge is modelled as lateral inflows to the computational domain through constant head boundaries. Thus, no areal vertical leakage from the rock mass above or below Zone 2 is considered. If this term is significant, it would cause some error in the model calculations. However, it is judged that possible areal recharge does not affect calculations significantly for Zone 2 and the boreholes of interest. It was decided to model the entire Zone 2 rather than separate subzones, as the results from the interference tests indicated that 3-D effects within Zone 2 may be important when pumping in only one of the subzones. Thus, test 3B is the test actually modelled in this case. The hydraulic parameters for the updated model were determined by trial-and-error calibration of the numerical model, supported by analytical aquifer analyses of the interference tests (see previous section). The resulting parameter distribution and computed drawdown distribution are shown in Figs. 17 and 18, respectively. The transmissivity value for Zone 2 is 3.5 x 10 -3 m 2 s -1 , which is somewhat higher than was used in the previous modelling. Comparison of measured and modelled drawdown curves in observation holes HFI01 and KFI11 are presented in Figs. 19 and 20. This was considered, by visual inspection, to be the best fit attainable with this

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

G.~VASTBOZONE

0

1

200

/+00

' '

71

BR.~NOANZONE

600

a00 1000

1200

'T.Sx,0'$''

1/+00 1600

1000

T=3x1 0 " ~

.

Z000 .

.

2200

2~00

.

2600 /

T=3 x 1 0 " ~

T=lxlO"~

•

\

\

1200 1/,O0

1600

1800

2000 2200

2/,00

2600

Fig. 17. Parameter distribution used for the calibrated numerical model.

particular model. For most boreholes, computed and measured responses are practically identical, the main exception being borehole BFI01, where modelled steady-state drawdowns are somewhat larger than measured. This cannot be modelled without significantly altering the geological description of the fracture zone, or introducing some local heterogeneities which may or may not be artificial. Some attempts to improve model performance by assuming a general horizontal anisotropy in Zone 2 were made, but without conclusive results (Andersson et al., 1989). ~00

400

600

BOO

1000

IEO0

400 ~

1400

1600

1800

I

~000

E200

2~00

EEO0

EBO0

06

400

200

0

200

eoo

40o

soo

800

1ooo

leoo

14oo

16oo

18oo

eooo

e~oo

e~oo

esoo

egoo

0

Fig. 18. Contour map of the drawdown distribution near steady state according to the calibrated numerical model. Equidistance = 0 . 2 0 m .

72

J.-E. ANDERSSON ET AL.

.... ::;:::'.:o

~ m m D a a a

SECTION 4

i t2

SETION,,

J

t

~o'°

~0'~

to

Trim; i&;i

Fig. 19. Predicted (solid |ine) and measured drawdown responses in observation borehole KFI! 1 during interference test 3B according to the calibrated numerical model.

It should be pointed out that the calibration described here is not verified, that is, tested during a different hydraulic event (for example pumping in a different borehole). Thus, there is significant uncertainty as to the uniqueness of the hydraulic parameters and flow geometry obtained. An additional factor .~,,I

m'

,

i ....

I

,

,

I . . . . .

I ....

x

• ~i÷.÷÷.

.........

tO *°

/

j iO

,

,

I ....

SECTION 3 SECTION 2 SECTION I

A~

z~. 2,,p~..,~rt

I tO "a

I_

x

~xx~xxx

I

l

i [Time imin}

]

Fig. 20. Predicted (solid line) and measured drawdown responses m observation borehole HFI01 during interference test 3B according to the calibrated numerical model.

H Y D R A U L I C TESTING A N D M O D E L L I N G A LOW-ANGLE FRACTURE ZONE

'73

of uncertainty is that spatial data are somewhat sparse. However, the flow modelling has proved to be extremely valuable for understanding the flow conditions in the Br~,ndan area, as well as for verification of geological and hydrogeological interpretations. SUMMARY AND CONCLUSIONS

The studies in the Br~ndan area have provided basic data with regard to the geological and hydrogeological disposition of the studied fracture zone in particular, and fracture zones in general. The detailed hydraulic testing provided data for a conceptual model used in the evaluation of the interference tests and tracer tests. The results clearly demonstrate the heterogeneous nature of the fracture zone studied. Thin, highly conductive conduits occur intermittently, separated by wider intervals of more competent rock of low to intermediate conductivity. The uppermost and lowermost parts of the zone consistently showed increased hydraulic conductivity in all boreholes. The single-hole tests also indicated a large contrast in hydraulic conductivity between the zone and the overlying and underlying rock. The very detailed hydraulic testing clearly demonstrates that the decrease in section length by about two orders of magnitude, from 20 m to 2 and 0.11 m, increases the resolution in the estimation of the effective hydraulic width of identified hydraulic features accordingly. The detailed interference tests in BFI02 showed that different borehole response patterns were generated in the near and distant regions from the pumping borehole, in particular when the uppermost and lowermost parts of the zone were pumped separately. When the entire zone was pumped the multiple response curves obtained from the boreholes were less separated from each other. In the near region the flow pattern was dominated by the primary responses, corresponding to the most rapid flow path between the pumping borehole and the actual observation borehole. These flow paths may be strongly influenced by local heterogeneities between the boreholes, e.g. restricted flow channels, stratification etc. The secondary responses in the boreholes (in the near region) are in general more delayed and attenuated, representing more average flow conditions. In the distant region, no dominating (primary) response occurs in the boreholes but instead the responses in all borehole sections within Zone 2 almost coincide. The secondary responses in the near and distant regions were very similar in all tests. These responses are believed to represent the averaged (long-term) hydraulic properties of the entire zone, whereas the primary responses in the near region should represent the short-term behaviour of the

74

J.-E. ANDERSSON ET AL.

system. This indicates that Zone 2 may be treated as an equivalent porous medium beyond a certain distance from the pumping borehole. This interpretation is supported by the fact that the long-term drawdown behaviour in all boreholes was very similar between the different tests, whereas the short-term behaviour in boreholes close to the pumping borehole varied considerably between tests. The interference tests have confirmed that the upper and lower subzones of Zone 2 have a high hydraulic diffusivity (high transmissivity and low storativity). The transmissivity of each subzone was estimated at about (1-2) x 10-3m2s -~ from the time-drawdown analysis. The calculated storativity values corresponding to the primary responses in the subzones are generally small, ranging from about 6 × l 0 - 7 to 7 x 10 - 6 in the near-region boreholes KFI11 and KFI06 during tests 1 and 2. The lowest storativity values were obtained from test 2 indicating particularly low-porosity flow paths in the uppermost subzone close to BFI02. F r o m the distance-drawdown analysis for the most distant observation boreholes and the numerical simulations, the representative (long-term) transmissivity of the entire Zone 2 was estimated at about 3 x 10-3mZs -I and the storativity at about 2 x 10 -5. F r o m the observed drawdown differences in sections in the same boreholes, the leakage coefficient was estimated at about (1-5) x 10-8s I. Assuming a thickness of about 50m for the fictive semipermeable layer between the subzones, these values correspond to an equivalent hydraulic conductivity in the vertical direction of Zone 2 of about (1-2) x 10 6 m s- 1. However, the actual flow transfer within Zone 2 is likely to be controlled by discrete fractures which locally may have much higher hydraulic conductivity. The interference test responses clearly show that hydrogeologically Zone 2 is bounded by outer delimitations. These boundaries are probably constituted by other fracture zones as suggested by Tir6n (1991, this issue). According to this interpretation the geometry of Zone 2 may be characterized by a triangular-shaped area. In addition, interference test 2 also showed that a steady state was reached by the end of the test. This indicates major inflows to Zone 2 by the end of the test, possible from adjacent fracture zones across the boundaries, e.g. the Brfindan Zone (Zone 1), or other more distant zones. The responses of the detailed interference tests were compared with those predicted from a previous model, and with simulations based on an updated model taking into account the results of the present interference tests. The relatively simple predictive model, based on the single-hole injection tests and limited pumping test data, did not prove to yield satisfactory predictions of hydraulic heads during the transient pumping tests. This was attributed to a lack of detailed hydrogeological description of the fracture zone, mainly the

HYDRAULIC TESTING AND MODELLING A LOW-ANGLE FRACTURE ZONE

75

outer boundary conditions, and inaccurate hydraulic parameter description for Zone 2. A considerably improved description of the flow conditions in the fracture zone was obtained using the updated calibrated model. The fact that the shapes of the measured drawdown curves are accurately reproduced, indicate that the flow geometry is described correctly. The calibrated model failed to reproduce exactly the drawdowns in the boreholes adjacent to the pumped hole. This is interpreted to be an effect of local heterogeneities in the rock material. The numerical flow-modelling work has proved to be of considerable value in assisting analytical aquifer analyses and gaining a general understanding of the factors governing flow in the fracture zone. The calibrated model will be used to predict future tracer tests and, ideally, also serve as a verification tool for the model. In summary, the present series of interference tests have demonstrated the usefulness of conducting such tests by pumping and recording pressure changes in multiple borehole sections. This provided detailed studies of the propagation of the pressure waves created in the pumping borehole and the associated flow pattern within Zone 2. The results of the present interference tests may provide a basis for future developments of models for flow in fractured crystalline rock. Although Zone 2 within the Br~indan area may appear unique in a geological and hydrogeological sense, fracture zones with similar properties are likely to be encountered in other areas. Combined use of hydraulic (interference) tests and tracer tests is recommended in future investigations. Finally, the question arises whether Zone 2 may be regarded as representative for other fracture zones in crystalline rock with a similar geological and hydrological character. Several low-angle fracture zones in crystalline rock have been encountered in other study sites in Sweden. However, only one of these zones has been subjected to hydraulic interference testing. For this zone, located in the Forsmark area in eastern central Sweden, the transmissivity was estimated at about 1 x 10 -5 m 2 s - I , i.e. about two orders of magnitude lower than that calculated for Zone 2 at Finnsj6n (Carlsson et al., 1986). Single-hole tests have, however, indicated higher transmissivities for lowangle fracture zones at other study sites, e.g. Klipper~s at a depth of about 800m (Gentzschein, 1986). Hydraulic information from low-angle fracture zones is also available from the Canadian investigation programme at Chalk River, Ontario. A thin subhorizontal fracture zone (Zone No. 1) in monzonitic gneiss recorded an estimated transmissivity of 4 x 10-6m2s -I (Raven, 1986); the anisotropy ratio between horizontal and vertical hydraulic conductivity was estimated at 10-170 from hydraulic interference tests (Raven, 1986).

76

J.-E. ANDERSSON ET A L

In the Finnish site investigation programme, several subhorizontal fractured and weathered zones were encountered in massive Rapakivi granite. Increased hydraulic conductivity in the upper and lower fractured zones showed these to be the most hydraulically significant. F r o m single-hole tests the arithmetic mean of hydraulic conductivity was estimated at about 1 x 10-Sms -t and 5 x 10-6m s - I for the upper and lower zones, respectively (Anttila, 1986). The thickness of the upper fracture zone varies between 3 and 6 m in the boreholes, whereas the lower fracture zone varies between 4 and 17 m (Anttila, 1986). The mean transmissivity of the upper and lower fracture zones is thus (3-6) x lO-5m2s i and (2-8.5) × 10 Sm2s-t, respectively. ACKNOWLEDGEMENTS

This work was carried out on behalf of the Swedish Nuclear Fuel and Waste Management Company (SKB). Special thanks are expressed to Professor S.P. Neuman, University of Arizona, Tucson, AZ, for reading and providing constructive criticism of the original manuscript. REFERENCES Ahlbom, K. and Smellie, J.A.T., 1991. Overview of the Fracture Zone Project at Finnsj6n Sweden. J. Hydrol., 126: 1-15. Ahlbom, K., Andersson, P., Ekman, L., Gustafsson, E., Smellie, J.A.T. and Tullborg, E.-L., 1986. Preliminary investigations of fracture zones in the Brfindan area, Finnsj6n study site. SKB Tech. Rep. TR 86-05, Stockholm. Ahlbom, K., Andersson, P., Ekman, L. and Tir6n, S., 1987. Characterization of fracture zones within the Br~indan area, Finnsj6n study site. SKB Status Rep. AR 88-09, Stockholm. Alm6n, K.-E., Andersson, O., Fridh, B., Johansson, B.-E., Sehlstedt, M., Hansson, K., Olsson, O., Nilsson, G. and Wikberg, P., 1986a. Site investigation - - equipment for geological, geophysical, hydrogeological and hydrochemical characterization. SKB Tech. Rep. TR 86-16, Stockholm. Alm6n, K.-E., Andersson, J.-E., Carlsson, L., Hansson, K. and Larsson, N.-A., 1986b. Hydraulic testing in crystalline rock. A comparative study of single hole test methods. SKB Tech. Rep. TR 86-27, Stockholm. Andersson, J.-E., and Persson, O., 1985. Evaluation of single hole hydraulic tests in fractured crystalline rock by steady-state and transient methods. SKB Tech. Rep. TR 85-19, Stockholm. Andersson, J.-E., Ekman, L. and Winberg, A., 1988a. Detailed investigations of fracture zones in the Brfindan area, Finnsj6n study site. Single hole water injection tests in detailed sections. Analysis of the conductive fracture frequency. SKB Status Rep. AR 88-08, Stockholm. Andersson, J.-E., Ekman, L. and Winberg, A., 1988b. Detailed hydraulic characterization of a fracture zone in the Brfindan area, Finnsj6n, Sweden. Paper presented at the 4th Canadian/American Conference on Hydrogeology: Fluid Flow, Heat Transfer and Mass Transport in Fractured Rocks, Banff, Alta., 21-24 June 1988, Alberta Research Council and the Association of Groundwater Scientists and Engineers (Div. NWWA), Alberta.

H Y D R A U L I C TESTING AND M O D E L L I N G A LOW-ANGLE F R A C T U R E ZONE

77

Andersson, J.-E., Ekman, L., Gustafsson, E., Nordqvist, R. and Tirbn, S., 1989. Hydraulic interference tests and tracer tests within the Brfindan area, Finnsj6n study site. SKB Tech. Rep. TR 89-12, Stockholm. Anttila, P., 1986. Geological and hydrogeological conditions of the island Hfistholmen, Loviisa. Nucl. Waste Comm., Finn. Power Co., Rep. YJT 86-05, Helsinki. Carlsson, L. Winberg, A. and Arnefors, J., 1986. Hydraulic modelling of the final repository for reactor waste (SFR). Compilation and conceptualization of available geological and hydrogeological data. SKB/SFR Prog. Rep. SFR, 86-03, Stockholm. Doe, T. and Remer, J., 1982. Analysis of constant-head well tests in nonporous fractured rock. In: T.W. Doe and W.J. Schwarz (Editors), Proceedings of the 3rd International Well-testing Symposium, 26-28 March 1980, Berkeley, CA, Earth Sciences Division, Berkeley, CA, 170 pp. Earlougher, R.C., 1977. Advances in well test analysis. Society of Petroleum Engineers of AIME, Henry L. Doherty Series, Dallas, TX Vol. 5, 264 pp. Ekman L., Andersson, J.-E., Andersson, P., Carlsten, S., Erikssom C.-O.~ Gustafssom E., Hansson, K. and Stenberg, L., 1988. Documentation of borehole BFI02 within the Brfindan area, Finnsj6n study site. SKB Status Rep. AR 89-21, Stockholm. Freeze, R. A. and Cherry, A., 1979. Groundwater. Prentice-Hall, Englewood Cliffs, N J, 604 pp. Gentzschein, B., 1986. Hydrogeological investigations at the Klipper~s study site. SKB Tech. Rep. TR 86-08, Stockholm. Gustafsson, E. and Andersson, P., 1991. Groundwater flow conditions in a low-angle fracture zone at Finnsj6n, Sweden. J. Hydrol., 126:79-111. Hantush, M.S., 1967. Flow to wells in aquifers separated by a semipervious layer. J. Geophys. Res., 72(6): 1709-1720. Moye, D.G., 1967. Diamond drilling for foundation exploration. Civ. Eng. Trans. Inst. Eng. Aust., April, pp. 95-100. Neuman, S.P. and Witherspoon, P.A., 1972. Field determination of the hydraulic properties of leaky multiple aquifer systems. Water Resour. Res., 8(5): 1284-1298. Raven, K.G., 1986. Hydraulic characterization of a small groundwater flow system in fractured monzonitic gneiss. Atomic Energy of Canada Limited, Rep. AECL-9066. Smellie, J.A.T. and Wikberg, P., 1991. Hydrochemical investigations at Finnsj6n, Sweden. J. Hydrol., 126: 129-158. Tir6n, S., 1991, Geological setting and deformation history of a low-angle fracture zone at Finnsj6n Sweden. J. Hydrol., 1-26: 17-43. Voss, C., 1984, A finite-element simulation model for saturated-unsaturated fluid-densitydependent groundwater flow with energy transport or chemically reactive single species solute transport. U.S. Geol. Surv., Water Resour. Invest. Rep. 84-4369. Zeigler, T.W., 1976. Determination of rock mass permeability. U.S. Army Eng. Waterways Exp. Stn., Vicksburg, MS, Tech. Rep. S-76-2.