GEOPHYSICAL STUDIES OF SALINE FLUIDS IN THE DEEP CRUST by
Guy M arquis
B .S c., U niversite du Quebec a Chicoutim i, 1986 M .S c., University o f V ictoria, 1988
A C C E P T E D A Dissertation Subm itted in P auial Fulfilm ent o f the FACULTY OF GRADUATE S I UQI ES Requirem ents for the D egree o f
D O CTO R OF PH ILO SO PH Y
the School o f Earth and Ocean Sciences
W e accept this thesis as conform ing to the required standard
Dr. Roy D. Hyndrnan, C o-Supervisor (School o f Earth &. Ocean Sciences) Dr. G eorge D. Spfehce, CcJ-Supervisoi (School o f Earth & Ocean Sciences)
Dr. C h r i s t o ^ r R T l l a r n e s , D epartm ental M em ber (School o f Earth & O cean Sciences)
Dr. H arry W. D osso, Outside M em ber (D epartm ent o f Physics and A stronom y)
Dr. Alan G. J^nes?sfexternai Exam iner (Geological Survey o f Canada)
«“ GUY M AR Q U IS, 1992
University o f V ictoria DATE.
DEAN
in
All rights reserved. This thesis may not be reproduced in w hole o r in part, >y
Supervisors: D rs Roy D . Hyndm an and G eorge D . Spence A B S T R A C T
G eophysical studies have shown that the low er continental crust, especially in Phanerozoic areas, is comm only m ore conductive and reflective and has a low er seism ic velocity than what is expected from the com]X)sition o f xenoliths and exposed low er crustal terrains. T hese anomalous in situ properties can be explained by the presence o f sma!1- am ounts o f free aqueous fluids in the intergranular space. A com pilation o f low er crustal geophysical data shows a correlation between electrical resistivity and seismic velocity, in agreem ent with physical properties models o f porous rocks, as well as a general decrease o f inferred porosity with geological age. C orrelations w ith geotherm al data also show that the reflective and conductive layers usually have their tops near the 400-450°C isotherm s suggesting an association with the brittle-ductile transition. T he rheology a t depth m ight have an effect on the trapping o f the fluids that a re in textural equilibrium pores in the ductile crust. A lternatively m etam orphic reactions m ay constrain free fluids to below this depth. M odel for the effects o f porosity in textural equilibrium pores on seismic and electrical properties o f rocks h iv e been developed, and are also in good agreem ent w ith the data com pilation. Re-processed L ITH O PR O B E South C ordillera m agnetotelluric and seismic reflection data in the Interm ontane Belt support a coincidence between the top o f low resistivity and high reflectivity o f the crust at depths o f about 20 km in the w est, and about 15 km in the east o f the Belt, corresponding to tem peratures around 450°C. Tw o m odels fo r reconciling the low vertical perm eab lity required for m aintaining the porosity at depth with the interconnection required to reduce the electrical resistivity are presented: one involves the deform ation o f equilibrium pores by sm all deviatoric stresses that pinch o ff the vertical interconnection, the other the flattening and alignm ent o f pores by low er crustal shear processes. A difficulty is recognized in reconciling free aqueous fluids in the lower cru st with the expected retrograde m etam orphism that should take up any free water. T h is process can be avoided if the fluids are o f high salinity. H igh-salinity fluids
are in liquid phase in the low er crust, not in supercritical phase as often thought.
Exam iners:
--- ■ ■ r —--- ---Dr. Roy D. H yndm an, C o-Supervisor (School o f Earth & Ocean Sciences)
D r. G eorge D . Spence, do-S upervisor (School o f Earth & Ocean Sciences)
Dr. Chfistej^hSTRTBarnes, D epartm ental M em ber (School o f Earth & Ocean Sciences)
D r. H arry W . Dosso, Outside M em ber (D epartm ent o f Physics and Astronom y)
iv
T A B L E O F C O N T E N T S
ABSTRACT ii
TA BLE O F C O N TEN TS iv
LIST O F TA BLES vi
LIST O F FIG U R ES vii
A C K N O W LED G EM EN TS ix
D E D IC A TIO N x
C H A PTER I IN TR O D U C TIO N 1
i. Geological inform ation on the deep crust 1
1.1 Xenoliths 2
1.2 Exposed cross-sections 3
ii. G eophysical inform ation on the deep crust 4
iii. How can they be reconciled? 6
CH A PT ER II M O D ELLIN G O F PH Y SIC A L PR O PER TIES 10 i. Effects o f porosity on elastic properties 10 ii. M odelling o f elastic param eters for equilibrium po re geom etries 19 iii. Effects o f porosity on electrical properties 26
iv. V elocity-Resistivity relations 28
CH A PT ER III W O R LD -W ID E C O M PIL A TIO N O F D EEP C R U ST A L
G E O PH Y SIC A L DA TA 31
i. D ata sources and selection criteria 31
ii. Relations o f porosity with age 39
iii. Shea- wave constraints 39
CH A PTER IV D ETA ILED IN TER PRETA TIO N : IN T E R M O N T A N E BELT,
BRITISH CO LUM BIA 44
i. M agnetotelluric survey 44
ii. East-W est profile 53
iii. N orth-South profile 61
iv. M ultichannel seismic processing and interpretation 64
vi. Refraction velocity data
vii. Joint interpretation o f geophysical data
v
77 78
C H A PT ER V CO N D ITIO N S FOR TR A PPIN G FLU ID S A T D EPTH 80 i. Tem peratures at the tops o f the porous layers 80 ii. Pressure conditions and physical properties in the ductile crust 82
iii. Equilibrium pore geom etries 84
iv. Retention o f pore fluids at low er crustal depths 87
v. Stress control o f pore geom etries 91
vi. Low er crustal shear processes %
C H A PT ER VI O R IG IN A N D NA TURE O F LO W ER CR U ST A L FLU ID S 99
i. Sources o f fluids for the deep crust 99
ii. Com position o f conducting fluids 102
C H A PT ER VII A LTERN A TIV ES TO T H E FL U ID H Y PO TH ESIS 107
i. G raphite 107
ii. U nderplating 111
iii. Shear zones 112
C H A PT E R VIII C O N C LU SIO N S 113
v i LIST OF TABLES
1. Xenolith data com pilation 3
2. Param eters for equilibrium pore geom etry m odels 21 3. Compilation o f lower crustal resistivity, velocity and inferred tem peratures 35
4. Lower crustal Poisson’s ratio 43
vii
LIST o r FIGURES
1. Exam ples o f deep crustal reflectivity 5
2. Exam ples o f deep crustal low-velocity zones 7
3. Exam ples o f deep crusta! low-resistivity zones 8
4. V elocity-Porosity relations 12
.5. Pore aspect ratio distribution 14
6. V elocity-Porosity relations for distributed pore aspect ratios 15
7. Poisson’s ratio-Porosity relations 17
8. Poisson’s ratio Porosity relations for distributed pore aspect ratios 18 9. Equilibrium pore geom etry m odels for num erical m odelling o f elastic
properties 20
10. Poisson’s ratio-Porosity relations for equilibrium pore geom etry
m odels 22
11. Shear m odulus and shear velocity vs porosity profiles for equilibrium
pore geom etry models 24
12. V elocity-porosity relation for equilibrium pore geom etry m odels 25 13. R esistivity-Porosity relations for 0.5M NaCl pore fluid 27 14. Resisti’'ity-Porositv relations for 5M NaCl pore fluid 29
15. Velocity-Resistivity relations 30
16. Velocity-Resistivity relations and values from the com pilation 40
17. Poisson’s ratio values from the com pilation 41
18. General m ap o f the C ordillera and location o f LITH O PRO BE Line 88-10 45 19. Detail o f LITHO PRO BE Line 88-10 and locations o f the m agnetotelluric
stations 46
'< 3. Skew values for selected stations 48
21. Raw M T data, East-W est profile 50
22. Raw M T data, North-South profile 51
23. E-Polarization phase pseudosection, East-W est profile
24. One-D inversion o f raw data, East W est profile 55 25. East-W est profile data corrected for static shift 56
26. One-D inversion o f E-Polarization data, East-W est profile
VUl
58 27. Tw o-D model after 800 iterations, East-W est profile 59 28. Com parison o f modelled and observed M T data, East-W est profile 62 29. E-Polarization phase pseudosection, N orth-South profile 63 30. One-D inversion o f raw data. N orth-South profile 65 31. North-South profile data corrected for static shift 66 32. One-D inversion o f E-Polarization data, N orth-South profile 67
33. P eferred model, North-South profile 68
34. Com parison o f modelled and observed M'T data, North-South profile 69 35. Stack section o f L fiH O P R O B E Line 88-10 71
36. Line 88-10 reflectivity energy histogram s 74
37. Tem perature profile from one-dim ensional inversion o f heat flow data 76 38. Reflectivity, resistivity, velocity and tem perature for Line 88-10 79 39. Histogram o f tem peratures o f reflective and conductive zones 81 40. Effective pressure depth profile and its effect on physical properties 83
41. Equilibrium pore geometries 85
42. Strain rate-Tem perature relations 89
43. Influerce o f deviat ic stresses on equilibrium pore curvature 93
44. Stress depth profile 94
45. Lower crustal shear zone model 97
46. Detail o f pore deform ation by shear 98
47. W ater content vs m etamorphic grade 101
48. Fluids stable under lower crustal conditions 103
49. Effect o f salinity on the critical tem perature o f aqueous fluids 105 50. Velocity-concentration relations for graphitic inclusions 108 51. Resistivity-concentration relations for graphitic inclusions 109
ACKNOWLEDGEMENTS
ix
First and forem ost, I want to thank my research supervisor, Dr. Roy Hyndm an. Roy has shown great patience in dealing with my somewhat unusual working m ethods and was alw ays available to discuss any idea I might have, as well as providing the seeds for a lot o f the w ork presented here. 1 just hope that putting up w ith me for three years will not deter him frorr. taking another P h.D . student... I have received trem endous help from a lot o f people as I worked my way into that project: seismic data utilities from G eorge Spence, seismic processing support at Lithoprobe from Kris Vasudevan and R olf M iie r, M T data and processing advice from Alan Jones, M T inversion program s from J.T . W eaver and A .K . Agarw al, and help with ANSYS from M inh Ly and Subu Ram am ani. 1 benefited greatly from discussions with T ark H am ilton, Lawrie Law, T rev o r Lew is, and Leonid Vanyan. Still, the best thing about Grad School rem ains the fellow students: thanks to Tianson Yuan for general geophysical consulting, Chinese tutoring, and support in our daily battles against com puter equipm ent, to Chris Spindler for com puter help and radio savvy, to Stew art Langton for discussions/argum ents on any topic under the sun, and to Steve Fallow s for long discussions on the philosophical im plications o f alternative rock lyrics. I also wish to thank two other students from ultra-R ockies, John Varsek and M ike Burianyk, for data, results, suggestions, and interest in my w ork.
Really huge thank you to Richard Baldwin for com puter help, rides to PG C, and for his understanding and friendship.
O f course, thanks to NSERC and SCBC for fellowships that put food on my table for all those years and for NSERC grants to R v that paid for various travel needs and research expenses. Thanks also to the Geological Survey o f Canada that provided a variety o f services (com puter tim e, office space, supplies, drafting) at no charge.
X
D E D IC A T IO N
A Gilles et G eraldine, qui m ’ont toujours encourage faire ce qui m e plaisait, meme si les bend flees sont parfois longs a v en ir...
"... besides, 1 expect nothing better front geophysicists anyway. ”
Anonym ous petrologist, reviewing M arquis and Hyndman (1992)
CHAPTER I INTRODUCTION
The lower continental crust is the least understood part o f the lithosphere. Its instrum ental role in most tectonic processes such as crustal shortening, extensional zones and rifts, and thrust and normal fault decollem ents has Ion" Kpf>n recognized, but the lack o f direct geological data prevented geologists from developing well constrained com positional models. M ost early models w ere based on such limited evidence that alm ost any com position could be fit to the data.
W ith the advent o f more m odern geophysical and geological m easurem ent and analysis techniques in the past two decades, ;t becam e possible to obtain m ore detailed inform ation on the deeper part o f the crust. Reflection and refraction seismology, electrom agnetic m ethods, better heat-flow and heat generation data, isotope studies, fluid inclusions, geotherm om etry and geobarom etry (especially on xenoliths and in exposed cross-sections), and many other techniques provide im portant constraints around which new geological models have been built. These new data have spurred interest in the low er crust and there have been special volum es focusing on the geophysics and the geology o f the lower crust (Dawson et a l., 1986; M ereu et a l., 1989). In this first chapter, I briefly review the geological and geophysical inform ation on the nature and the physical properties o f deep-crustal rocks, and how these results can be reconciled in a single model.
i. Geological information on the deep crust
xenoliths and high-grade metamorphic '« lin s. T he latter are interpreted as com ing Irom the low er crust since their m ineralogy can only be produced under high pressure- tem perature conditions, i.e. 5-10 kbar, 600-900°C (Taylor and M cLennan, 1984). Kay and Kay (1986) have presented an overview o f data from both o f these sources. G eochemical results o f relevance for this thesis topic are sum m arized here.
i .l Xenoliths
Xenoliths are found in volcanic pipes and in kim berlites. They a re entrained and brought to the surface by m agm as and fluidized solids, usually o f m antle origin, traversing the crust. T he rapid transport c f xenoliths should ensure that they have not been subjected to retrograde m etam orphism by long exposures to mid and upper crustal tem peratures and pressures. They may have been near the surface fo r long periods o f tim e but the slow reaction kinetics at such low tem peratures prevents them from retrograding. For these reasons "they are probably the least equivocal sam ples o f the lower crust to which w e have access" (T aylor and M cLennan, 1984). They provide a good aerial sampling o f the deep crust since they can be found in many different tectonic environm ents: converging plate m argins, rift valleys, continental intraplate regions, and volcanic plateaux. The main problem s in interpreting xenolith data are that they a re small, and hence it is difficult to obtain large-scale correlations, and the vertical resolution can be quite poor, as xenoliths from different horizons have been scram bled. There is also uncertainty in correlating the observed m ineralogies to actual depths o f origin. Xenoliths are prim arily produced in high tem perature volcanic terrains so they may be at higher m etam orphic facies than typical rocks o f the present low er crust, i.e . most xenoliths being in granulite facies does not im ply that the low er crust everyw here is in granulite facies at present tim e. In m ost crustal areas present-day tem peratures in the low er crust are well below the onset o f granulite facies m etamo ohism .
G riffin and O ’Reilly (1987) have surveyed low er crustal xenolith data from all over the w orld. M ost xenoliths have been collected from basalt outcrops in Phanerozoic terrains or from kim berlites in Precambriari cratons. Precam brian xenoliths appear to be som ew hat less mafic in com position titan Phanerozoic ones. In both environm ents how ever, the overw helm ing m ajority o f the xenoliths surveyed have m ineralogies typical
3
o f m afic granulites (see Table 1). G riffin and O ’Reilly (1987) conclude that most o f the crustal m aterial added since the Precam brian is mafic, and that the low er crust in Phanerozoic areas is especially mafic. They also note that the low er crustal composition inferred from xenoliths is more mafic than suggested by the seism ic velocity data (see C hapter III).
M afic Felsic G ranulite Eclogite O ther
Precam brian 13 3 13 2 1
Phanerozoic 44 13 40 5 12
T a b le 1 Xenolith data com pilation from G riffin and O ’Reilly (1987)
Caution should be exercised in interpreting the results above as evidence that most o f the low er crust is in granulite facies. As stated earlier, xenoliths are usually produced in high tem perature volcanic environm ents, so their m etam orphic grade may be higher than that o f typical lower crust.
i.2 Exposed cross-sections
Even though the inform ation from xenoliths is usually adm itted to be more reliable sam pling o f the low er crust (Kay and K ay, 1986), the m ost accessible samples o f large volum es o f low er crustal rocks are the exposed granulite terrains, whose m ineralogy requires that they have been form ed in the deep crust. These terrains all show increasing m etam orphic grades at present deeper crustal levels (Fountain and Salisbury, 1981), regardless o f chemical com position. One objection to the extrapolation o f such sections to the deep crust as a whole is that m ost have been exposed by overthrusting at continental collision zones, which is certainly not the typical history o f m ost crustal sections (Kay and Kay, 1986). In addition, the pressure-tem perature conditions at w hich these rocks have been m etam orphosed, as well as structural relationships, indicate that they may not com e from the low erm ost crust, but often frorn
4
the m iddle crust, at their peak m etam orphic conditions. This has been shown to be the case for the Kapuskasing structural zone, until recently considered o f very deep crustal origin. Structural (Percivai and C ard, 1985) and seism ic (Percival et a l., 1989) evidence now indicates that it actually originates from the m iddle crust . G ranulite facies rocks as a whole range from mafic to silicic in com position. In the Ivrea Zone (Italy) and in Australia, lower crustal rocks are predom inantly m afic, w hile in M anitoba (Pikw itonei subprovince) they are interm ediate to silicic. H ow ever, even in the tatter section there are m afic-ultram afic bodies present in the deeper levels.
The most fam ous sample o f the deep crust is the Lew isian gneiss com plex in Scotland. It com prises a block o f granulite facies rocks surrounded by am phibolite-facies gneisses. A com parison o f granulite vs am phibolite gneiss com position (W eaver and T am ey, 1984) shows that the inferred low er crustal rocks in the com plex are m ore mafic ( ~ 12% Fe, Mg m inerals) than the m iddle crustal rocks ( - 8 % ) . In com parison, xenoliths in general are even m ore mafic. This indicates that the deeper portions o f the crust are more mafic than the com position o f the exposed granulite terrains (T aylor and M cLennan, 1984), and therefore also m ore mafic than suggested by the seismic velocity data.
ii. Geophysical information on the deep crust
Over the past few decades, a large num ber o f geophysical m easurem ents have provided inform ation on the low er crust. The main contributors a re the large national seismic reflection program s such as CO C O R P, BIRPS, L ITH O PR O B E, and m any others that have produced huge am ounts o f geophysical data from alm ost all kinds o f geological environm ents. Yet despite the diversity o f those environm ents, three surprising results have been com m only obtained:
(1) seism ic reflection profiles com m only show a transparent upper and m iddle crust, but a highly reflective low er crust, particularly in Phanerozoic areas. Exam ples
of such reflective patterns have been presented by M atthews (1986) for the British Isles (see "typical Birp" in Figure la ), where reflectivity is increased betw een 5 to 10 s two- way traveltim e, by A llm endinger et al. (1987) for the Basin and R ange P rovince (Figure
S
E
C
O
N
D
S
A T Y P I C A L B I R P
10 -\ \\
---\\
(a)
(b)
(c)
;; . -- : -- “• ? • f*''A " \ O T y ~ V _ -■- ■'.- 'ts ~ :---“ •/'- - •’ "■=; r ~ .'’---_ ==' -L ~ •*■-*■ ^ ■X P f i m r (
, ■•"&.•' t u # ;
1 5 0 WC" . _---Figure 1 Exam ples o f deep crustal reflectivity: (a) "typical Birp" for the British Isles, (b) Basin and Range Province, (c) southw est Germ any.
T w o -w a y ti m e ( s )
6
lb ), and by Liischen et al. (19P7) for southwest G erm any (F igure lc ). F o r other exam ples o f enhanced reflectivity in the deep crust, the reader is referred to the volum es on Seism ic Reflection Profiling o f the Lithosphere (Barazangi and Brown (eds.), 1986a, b; M atthews (ed.), Geophysical Jourmd c fih e Royal Astronomical Society, 89, 1987; Leven et al. (eds.), Tectonophysics, 173, 1990; M eissner et al. (eds.), 1991). W hile more recent data and analyses have shown the division o f a non-reflective upper crust and a reflective low er crust to be rather too sim ple, a higher reflectivity in the low er crust is usually observed.
(2) seismic refraction profiles have defined low er crustal velocities com m only less than 7 .0 km /s (e.g. M eissner, 1986), low er than expected fo r the otherw ise inferred dom inant mafic com position from xenoliths and exposed terrains as discussed above (e.g. Hyndm an and Klem perer, 1989). A few exam ples are shown in F ig u re 2: Basin and Range, U .S .A . (Benz et a l., 1991; Figure 2a), Black Forest, G erm any (G ajew ski and Prodehl, 1987; Figure 2b), North Sea (Barton and W ood, 1984; F ig u re 2c).
(3) m agnetotelluric and controlled-source surveys have revealed that the low er crust generally has very low electrical resistivity com pared to rocks thought to be comm on in the low er crust as measured dry in the laboratory (e.g . Shankland and Ander, 1Q83) Some exam ples o f deep crustal low -resistivity zones are shown in Figure 3: Scotland (Hutton et a l., 1980; Figure 3a), Basin and R ange (L ienert and Bennett, 1976; Figure 3b), and the Canadian C ordillera (Jones et a l., 1992; F ig u re 3c). M any reviews and com pilations on the resistivity o f the deep crust have been published over the years: Jones (1981), Gregori and Lanzerotti (1982), Shankland and A nder (1983), Haak and Hutton (1986), Adam (1987), Hyndm an and Shearer (1989). T here is little doubt that the lower crust, especially in Phanerozoic areas, has very low resistivity.
iii. How can they be reconciled?
M any hypotheses have been presented to explain these properties o f the low er crust, but most o f them apply to only a single anom alous property. R eflective patterns have often been associated with sub-horizontal com positional boundaries, basalt sills, horizontal shear zones or mylonitic fabrics. T he low er deep crustal velocities have been
7
(a)
1 0 S 20 X X H Q. Ill Q 3 0 4 0 SO it'. •Ji u 4.T 4* iii 'i i ' iua: iSiis T.f r ' . v . - i • , S.* • ' ’A ’*, « j v 11 l l i l J ; i r X : 4 » : X v X ; T J j IHsSH: $ $ ! : :ELZTAL KANOEL FELDBERG TOOTMOOS WEHR OPPENAU S A TOOTNAU S S R H IN E H A S IA C H - S ' " V J 2 i t s __ 6.7
0
10 20 30 50 100 150 "////////'•'///S/////////////////.'//////"///**""". m. 20 MOHO MOHO 0 50 DISTANCE kmF ig u re 2 Exam ples o f deep crustal low-velocity zones: (a) Basin and Range Province, (b) Black F orest, G erm any, (c) North Sea.
O rp th I km I
8
Caledomdes Uplands
1000 fim
—10 fim
100 flm
--- 1— i---1— i— r——\— i---1--- 1---1--- 1— i--- 1
W I S T 60 50 40 30 20 10 0 10 20 30 <0 50 6 0 KM I A S T
(c)
I
i
Q
F ig u re 3 Exam ples o f deep crustal low -resistivity zones: (a) Scotland, (b) Basin and Range Province, (c) southern Canadian C ordillera.
Oi---10
>tXX)Om
^ -**30Qtn ^
j
<t)Otn
VE5:1
0 I III— t— — I 1 - ---—< 500km(b)
1 0 -* 2 0- 7 30-1 4 0 -a. IM O 50-j 60-19
interpreted as an interm ediate composition for the deep crust. N one of these mechanisms can explain the low -resistivity in the deep crust, as rocks show very little variation in electrical resistivity with com position. Low -resistivity layers have been interpreted variously as zones o f partial m elting, graphitic inclusions or film s, iron oxides, but none of these should have any noticeable effect on the seismic properties o f rocks. One hypothesis that appears to reconcile both seismic and electrical observations is the presence o f horizontally layered zones o f aqueous fluids in the low er crust (e .g ., Hyndm an and H yndm an, 1968; Shankland and A nder, 1983; H aak and H utton, 1986; G ough, 1986; H yndm an and Shearer, 1989). T he prim ary objective o f the research reported in this thesis is to investigate w hether anueous fluids are present in the lower continental crust.
In o rd er to test the hypothesis o f free aqueous fluids in the deep crust, three questions have to be answ ered. First, w hat are the effects o f aqueous fluids on the physical properties o f deep crustal rocks? T his will be discussed in C hapter II. Second, are the physical properties o f lower-crustal rocks observed in situ consistent with the porosity m odels? This will be discussed for global data in C hapter III and in Chapter IV on a sm aller local scale. Third, are there any other conditions that must be met in the deep-crustal porosity model? This will be discussed in C hapters V and VI. Some o f the possible alternatives to the aqueous fluid hypothesis for a conductive, low-velocity and reflective low er crust will be presented in C hapter VII.
10
CHAPTER II MODELLING OF PHYSICAL PROPERTIES
i. Effects of porosity on elastic properties o f rocks
The effects o f porosity on the elastic properties o f rocks under lo e r crustal conditions have been discussed by many authors. T h e elastic behaviour o f low -porosity crystalline rocks is controlled mainly by the m atrix com position, b ut :i is also affected by the am ount o f fluid present, the m agnitude o f the effect depending on the geom etry o f the pore spaces. T he first com m only used velocity-porosity relation was the time- average equation o f W yllie et al. (1956):
1 _ 1 -<f> 1
w here V porous rock velocity, Vm matrix velocity, Vf pore fluid velocity, <ti porosity (</> < 1).
It is still w idely used today in the petroleum industry for sedim entary rocks, b ut is only a first-order approxim ation, since it does not take into account the pore shapes. M ore representative models applicable to crystalline rocks have been developed fo r spherical pores (Watt et al., 1976), ellipsoidal pores (Kuster and Toksoz, 1974; O ’C onnell and Budiansky, 1977), for tubular porosity (M avko, 1980), and for com binations o f various pore shapes (Schm elling, 1985). It has been shown by Shearer (1988) that these various form alism s all give sim ilar results. The form alism o f Kuster and T oksoz (1974) has been used here, mostly for its mathematical ease and flexibility. It describes the effects o f ellipsoidal pores with varying minimum to m aximum axis ratio (aspect ratio). It has been pointed out by H olbrook et al. (1991) that this form alism is valid only fo r isolated pores, and therefore m ight not be valid for interconnected porosity. Schm elling (1985) modelled the effects o f the degree o f pore interconnection on elastic properties and found that they are very sm all, even at high degrees o f interconnection. It will be argued later that the porosity in the ductile low er crust should be in textural equilibrium pores. These
11
pores are fluid pockets at grain com ers, interconnected by very narrow channels at grain boundaries. If such a porous medium is deform ed at seismic frequencies, the fluid phase will not be displaced into the narrow channels. Therefore, the fluid phase is in an unrelaxed state as an elastic deform ation goes through the rock, which is sim ilar to the case for isolated pores. Using the form alism o f Kuster and Toksoz is thus applicable. In addition, Hyndm an and Shearer (1989) present laboratory data that show good agreem ent with this theory fcr both P-velocity (Vp) and Poisson’s ratio vs porosity.
Figure 4 shows P-velocity versus porosity relations for a mafic rock with Vp = 7.2 'a n /s containing ellipsoidal pores with a range o f aspect ratios. T he most im portant property o f these relations is that thinner pores have a much m ore pronounced etfect on velocity than m ore equidim ensional pores for the sam e porosity.
A zero-porosity velocity o f 7 .2 km /s has been chosen, assum ing a mafic com position for the low er crust, as suggested by the results m entioned in the previous section. The com pilation o f C hristensen (1982) shows that gabbros, m etagabbros, and am phibolites have laboratory velocities averaging about 7.2 km /s at lower crustal pressure conditions (10 kbar). Interm ediate com positions generally have slow er velocity, about 6.5 to 6 .7 km /s (Christensen, 1979; 1982). M ost o f the velocity inferences described below depend upon the low er crust being dom inantly mafic in com position, with a zero-porosity velocity o f about 7.2 km /s.
T o the left o f Figure 4 is a histogram o f seismic refraction velocity estim ates for the low er crustal layers com piled in Table 3. The right part o f Figure 4 illustrates the inferred porosity for a range o f pore aspect ratios. T he num ber o f sites in this com pilation is limited by the requirem ent o f nearly coincident m agnetotelluric m easurem ents (see C.iapter III), but the velocity pattern is sim ilar to o th er com pilations (e.g. Hyndm an and K lem perer, 1989).
The m odels presented above are dependent upon tw o assum ptions; the first, as already m entioned, is that o f a m afic com position; the second is that o f a uniform pore geom etry. Pore shapes are how ever never uniform , but are distributed depending on grain sizes and shapes, which are determ ined by localised, m icroscopic igneous crystallization and m etam orphic conditions (e.g. Jurew irz and Jurew icz, 1986). Because
12 15 1 0 5 0
No. Sites
Velocity-Porosity Relations
Aspect
Ratio
7 .00.1
6 .5 Oh6.0
.0.01
0.01
0 .0 30
1
2
3 4 5Porosity (%)
Figure 4 V elocity-porosity relations for a m afic rock containing ellipsoidal pores o f selected aspect ratios. To the left, histogram o f velocity values from Table 3.
13
there is undoubtedly a distribution of pore geom etries present, elastic param eters have also been calculated for various distributions o f p o .e aspect ratio (Figure 5). Lognormal distributions with standard deviation o f up to 0 .5 are realistic (Jurew icz and Jurew icz, 1986), since it is unlikely that more than two orders o f m agnitude o f pore aspect ratio a re present in significant am ounts in the low er crust, because the expected tendency tow ards equilibrium pore geom etries will lim it the range o f effective aspect ratios. Elastic param eters have been .ecalculated using a weighted average o f the effects for each value o f pore aspect ratio. The theory o f Kuster and Toksoz (1974) is still used here. Figure 6 chows how P-velocity is influenced by various distributions centred on selected values o f mean pore aspect ratio: thicker mean pores show a greater variation in effects, from m oderate ( s = 0 ) to large ( s = l ) distributions. A wide distribution o f aspect ratios alw ays has a larger effect on the elastic param eters because even a small fraction o f very thin pores affects them dram atically.
A nother im portant param eter that is sensitive to pore geom etry and which is receiving increasing attention in deep crustal surveys is Poisson’s Ratio (related to Vp/V s). A critical factor in the interpretation o f deep crustal Poisson’s ratio from seismic data in term s o f porosity and pore geom etry is its zero-porosity value. T his value depends on the assum ed composition and on the extrapolation o f the available laboratory data to zero porosity. Most laboratory m easurem ents on rocks that are likely candidates for the low er crust, at high confining pressure and low tem perature, give Poisson’s ratios ranging from 0 .2 7 to 0.32. Low- to medium grade mafic rocks (e.g. gabbro, m etagabbro) have Poisson’s ratios generally in the higher part o f the range, 0 .2 9 to 0.32, w hile mafic granulites’ values are in the low er part o f the range from 0 .2 7 to 0 .2 9 (see sum m ary in G oodw in and M cCarthy, 1990). These Poisson’s ratio values must be corrected to zero oorosity. T here does not seem to be any laboratory study o f the effects o f porosity on P oisson’s ratio for gabbros, m etagabbros o r mafic granulites that allow us to extrapolate with confidence to zero porosity. The porosity o f the samples is usually not given, but the few reported range from 0 .2 to 1%. Laboratory m easurem ents for basalts are given in Hyndm an and Shearer (1989). T he relatively small range in Poisson’s ratios for a probable factor o f 5 in porosity for low er crustal type rocks
P
e
rc
e
n
ta
g
e
of
P
o
re
s
14100
0.0
0 .0 0 1 0 .0 1 0 .1A spect Ratio
0 .0 0 1 0 .0 1 0 .0 0 1 0 .0 1 0 .1 140
SO20
10
0 P
,
,
,
0 .0 0 1 0 .0 1 0 .1 1Figure 5 Pore aspect ratio distribution for calculating P-velocity and Poisson’s ratio values in Figures 6 and 8. s= sta n d a rd deviation o f lognorm al distribution
F ig u re 6 .S ' C 7 . 2
Mean
=0.01
6.8
6.4
6.0
7.2
Mean
=0.03
6.8
6.0
15 7 .26.8
6.4
6.0
Mean = 0.1
0
1
2
3
4
5
P orosity (% )
Velocity-Porosity relations for distributed pore aspect ratius. The num bers refer to the lognormal standard deviations as in Figure 5.
16
suggests that under low tem perature and the high confining pressure laboratory conditions that close the thinner pores (above 0.5 kbars or 50 M Pa), the effects o f porosity are small, i.e. the effective aspect ratios are m ore equidim ensional than 0 .0 3 . T he effect o f tem perature on Poisson’s ratio is negligible for the lithologies o f interest (variation o f less than 1 % for tem perature increase o f 500°C using data from C hristensen, 1979). A zero-porosity Poisson’s ratio o f 0 .2 9 is thus a reasonable average value fo r low er crustal conditions, with a range o f 0.27 to 0.31.
Figure 7b shows the variation in Do isson’s ratio against porosity fo r selected pore aspect ratios, using the form alism o f K uster and Toksoz (1974). T he figure illustrates that for pores o f small aspect ratios P oisson’s ratio increases rapidly w ith increasing porosity, while pores o f larger m ore equidim ensional aspect ratio it tends to decrease with increasing porosity (see also Shearer, 1988). It is im portant to note that quite thin pores o r cracks are required for Poisson’s ratio to increase strongly w ith increasing porosity at the porosities o f interest in the lower crust (i.e ., aspect ratios < 0.01 required). Thus the absence of high Poisson’s ratio cannot be taken as evidence against porosity in the lower crust (H yndm an, Lew is and M arquis, 1991). T h e effect o f a distribution o f p oie aspect ratios are shown in Figure 8. It is seen that fo r thin average aspect ratio pore (mean 0.01) a substantial spread in pore geom etries results in a m ore rapid increase in Poisson’s ratio with increasing porosity (top o f Figure 8). F o r m ore equidim ensional mean pore aspect ratio (e.g . 0.1) there is a sm aller effect on P oisson’s ratio; the change with increasing porosity rem ains negative except fo r quite w ide distributions. Again w ide distribution o f pore aspect ratio have dram atic e f.e c t on Poisson’s ratio because o f the large contribution o f the few very thin pores.
Even though the ellipsoidal pore m odels used above are in close agreem ent with experim ental results for near-surface porous rocks, pore geom etries m ay be fundam entally different for rocks in the ductile low er crust. At these depths, pore fluids, if present, are expected to be in textural equilibrium w ith the rock m atrix. Equilibrium pore geom etries differ greatly from ellipsoidal pore shapes. In o rd er to obtain a velocity- resistivity relation for equilibrium pore geom etries, the effect o f such pores on the elastic properties o f a rock have been m odelled num erically .
0.05
porosity
1 %
b
0.00
<
2 %
5%
-0.05
0.001
0.01
0.1
Aspect ratio
0.05
aspect ratio
(b)0.01
0.3
b
0.00
<
0.1
0.03
-0.05
0 12
3
4
5
Porosity (%)
F ig u re 7 (a) Influence of pore aspect ratio on P oisson’s ratio for 1,2 and 5% porosity, (b) Poisson’s ratio vs porosity relations for selected pore aspect ratios.
18
0.05
✓Mean
=0.01
-0 .0 5 0 .0 5Mean
=0.03
0.75
0.000.5
-0 .0 5 0 .0 5Mean = 0.1
0.00
- 0 .0 5 0 1 2 3 4 5P orosity (% )
F ig u re 8 P oisson’s ratio-Porosity relations for distributed pore aspect ratios. T he num bers refer to the lognormal standard deviations as in Figure 5.
19
ii. M odelling o f elastic p ro p e rtie s f o r e q u ilib riu m p o re geom etries Pore models have been developed using the program ANSYS (Swanson Analysis System s, 1989), a package com m only used for m echanical engineering applications. ANSYS uses a matrix displacem ent m ethod o f analysis based upon a finite-element idealization o f me m odel. T he displacem ent m ethod consists o f approxim ating the solution by interpolating a function from the boundaries to the interior o f the elem ents. T he function has to satisfy the governing equations (H ooke’s Law and Equation o f C ontinuity in this case) and the pre-defm ed boundary-conditions. Pre-defined elem ents are chosen depending on the problem to be solved. T h e main advantage o f using ANSYS is the ease o f defining the system and the boundary conditions for calculations.
The study is lim ited to a two-dim ensional case, since it is unlikely that a three-dim ensional model w ould change the results by an am ount larger that the accuracy obtained from seismic surveys, especially for the deep crust. T he pore shapes of Cheadle (1989) have been discretised using two-dim ensional fluid elem ents, and have been surrounded by two-dim ensional solid elem ents (Figure 9). Both types o f elem ents have been given realistic param eters of elastic moduli and density. T he elastic moduli
have been calculated using 2
M = pV*
X = p ( V2p - 2V? )
K . X . 2 ,
w here Vs shear velocity
Vp com pressional velocity
H shear m odulus p density
X Lam e’s constant K bulk m odulus
T he param eters obtained are shown in Table 2. The elem ents are then divided by ANSYS using an autom atic griding procedure, with a higher density o f grid points at edges and boundaries: there are 48 nodes on either side o f the boundary betw een each grain pocket and the m atrix, 4 nodes inside each pocket, and 20 nodes on either side o f the channel- m atrix boundaries.
a
P O IS S O N ’S R A T IO T E S T Solid P re ss u re P re s s u rev \
F lu id N o v e rtica l d isp la c e m e n t S H E A R V E L O C IT Y T E S T S h e a r s tre s s N o d isp la c e m e n t rFigure 9 Equilibrium pore geom etry model: (a) for calculation o f P oisson’s ratio, showing the pore configuration and the appropriate boundary conditions, (b) for calculating the shear modulus.
21
T a b le 2 P a ra m e te rs f o r e q u ilib riu m p o re g eom etry m odels
j|
M atrix FluidVp (m /s) 7200
1500
V, (m /s) 4200
0 Shear M odulus n (Pa) 5 .3 10'°
0 L am p’s C onstant X (Pa) 4 .9 10‘0
2.3 10* Bulk M odulus K (Pa) 8.5 1010
2.3 109
Density (kg/m 3) 3000
1000
The resulting porous rock m odels have then been analyzed to study the effects o f uniaxial pressure (for Poisson’s Ratio, Figure 9a) and shear stress (for the shear m odulus, Figure 9b). A ppropriate boundary conditions have been applied: no displacem ent through the lower horizontal axis for the Poisson’s Ratio test (Figure 9a), and no horizontal and vertical displacem ents at the low er horizontal axis for the shear m odulus test (Figure 9b). Knowledge o f both Poisson’s Ratio and shear m odulus enable us to obtain both Vs and V p, and com parison o f the results with ellipsoidal pore models predictions gives an effective pore aspect ratio.
N um erical models o f the deform ation o f rocks with equilibrium pores have been calculated for a few values o f porosity. A special attention has been to porosity values less than 3% , because porosities in the deep crust a re expected to be low.
P oisson’s Ratio for each case has been calculated using the ratio o f vertical displacem ent (strain) to the horizontal displacem ent. These values a re output by ANSYS after each iteration. T he solution converges rapidly. In all cases one iteration was sufficient, subsequent iterations giving very sim ilar solutions.
The variation in Poisson’s Ratio (Aa) increases with porosity (Figure 10), with a second-order fit to the data of w ith a correlation coefficient (r) o f 0 .9 2 . T his suggests that the equilibrium pore geom etries have a quite low effective aspect ratio. T his in
2 2
0.05
2nd order f i t
a =0.003
a= 0 .0 1
0.00
-0.05
0 12
3
4
5
P orosity (% )
Figure 10 Poisson’s Ratio-Porosity relations obtained from equilibrium pore geom etry m odels (Figure 9a). Ellipsoidal pore models a re shown for com parison.
23
agreem ent w ith the electrical resistivity m odels for which it will be shown that the small grain-boundary channels have much m ore influence on the physical properties than the larger volum e grain-corner pockets (see next section). By com parison with the ellipsoidal pore m odels, the equilibrium pore geom etries have effective aspect ratios slightly low er than 0.01 (see Figure 10).
The horizontal displacem ent at the top o f the model volum e caused by horizontal shear stress has been used to determ ine the shear m odulus. D isnlacem ents have been obtained for the same porosity values used for determ ining Poisson’s ratio. The conversion to shear velocity is not as straightforw ard how ever. Vs is defined as:
V, = (fi/p)m
where p. Shear m odulus (i.e. Shear stress/Shear strain)
p Density.
Since the applied shear stress is the same for all porosity m odels, the norm alized shear m odulus is inversely proportional to the norm alized shear strain. All param eters (shear strain, shear stress, density, V s) have been norm alized to their values for dry rock.
Figure 11 shows the main results o f this analysis: (a) the norm alized shear m odulus and (b) the resulting V, vs porosity relation. V, vs porosity relations for ellipsoidal pores o f aspect ratio 0.01 and 0.03 are shown for com parison. A first-order regression line has been fit to the data (r= 0 .8 8 ). As for the P oisson’s ratio situation, these results show that the narrow grain-boundary channels have m ore influence on the elastic properties o f a rock than the grain-corner pockets. T he effective pore aspect ratio is slightly higher than 0.01.
Once V s and Poisson’s ratio (a) are known, the com pressional wave velocity (Vp) can be obtained from
24 Figure 11
1.0
(a)"§ 0.8
X
K 5S^
0.6
bo0.2
3
4
5 12
04 .0
3.5
3 .0
2.5
2.0
0
1
2
3
4
5
Porosity (%)
Norm alized shear modulus (a) and norm alized and true shear velocity (b) vs porosity obtained from equilibrium pore geom etry m odels (Figure 9b). Ellipsoidal pore models are shown for com parison.
1.0
(b)a —0.03
0.9
£
0.8
1st order
v ,regression
0.6
a —0.01
Vs
(k
m
/s
)
P
-V
el
o
cit
y
(k
m
/s
)
r
257.2
6.8
a = 0 .0 3
6.4
m odel
6.0
2
0 13
Porosity (%)
i! F ig u re 12 Velocity-porosity relation for equilibrium pore geom etry m odel, using results from Figures 10 and l i b . Equilibrium pore geom etries have very low effective pore aspect ratio.
The resulting V p-porosity profile is shown in F igure 12. The effective po re aspect ratio is again slightly larger than 0.01, indicating that indeed the grain-boundary channels have a large influence on the elastic properties: the velocity is reduced by 10% for a porosity of about 1.5% . T herefore, very small porosities a re required to obtain the velocities reported in the com pilation, with a zero-porosity value o f 7.2 km /s.
t:i. E ffects of porosity on e le c tric a l p ro p e rtie s o f ro ck s
Although other conduction mechanisms such as graphite cannot be excluded (Frost et a l., 1989; C hapter V I), the low resistivity o f the deep crust has been assum ed to be the result of saline fluid porosity. In this fluid porosity m odel, the electrical resistivity of a rock is prim arily controlled by the fluid porosity, the resistivity o f the po re fluid, and the pore geom etry. The resistivity o f the m atrix (1 0 M 0 5 ohm m; K ariya and Shankland, 1983) bears an insignificant influence. Figure 13 shows th e influence o f porosity on electrical resistivity for various geom etries including the em pirical relation of A rchie (m odified by H erm ance, 1979):
Pf + (P s ~ P f) <£'"
where p porous rock resistivity
p „ p, matrix and pore fluid resistivity
4> porosity
m Archie exponent
In F igure 13, the pore fluid is seawater salinity (0.5M N aC l), and in F ig u re 14, it is a 5M N aC i brine, a fluid that also could be present in the deep crust. A rchie’s L aw with pore tortuosity exponents between 1.5 and 2.5 fit m any laboratory m easurem ents on crystalline rocks (e.g. Brace et a l., 1965), while an exponent o f 1.2 approxim ates closely
Resistivity-Porosity Relations
10
o
No. Sites
i
.S3 CObo
3
4
Isol 0.03
3
m=2.5
2
m=2.0
■
TubesS
1m=1.5
HS bound
0 0 12
3
4
5
Porosity (%)
F ig u re 13 Resistivity-Porosity relations for 0.5M NaCl po re fluid. Fluid configurations used are: isolated ellipsoids, A rchie’s Law (exponent m), tubes, the low er Hashin-Shtrikm an bound and 60° dihedral angle equilibrium pores. T o the left, a histogram o f lower crustal resistivity layers from the com pilation in Table 3.
28
the theoretical behaviour o f very well interconnected, tubular porosity. U nlike the elastic case, ellipsoidal pores are a poor representation for the effect o f porosity on resistivity h th e lower crust, since the critical factor is the degree o f pore interconnection. Equilibrium pore geom etries may be a m ore appropriate approxim ation at tem peratures above about 400°C because mafic rocks are ductile under these conditions (H yndm an and Shearer, 1989), but the electrical behaviour o f such geom etries is com putationally difficult: num erical solutions would be required. It can how ever be approxim ated by an abrupt resistivity variation arising from a transition from pinch-off (isolated porosity) to pore interconnection below about 1 % porosity (Cheadle, 1989; H yndm an and Shearer, 1989). The actual pinch-off porosity depends on the dihedral o r fluid-grain w etting angle of the equilibrium pores (von Bargen and W aff, 1986). Equilibrium pore geom etries w ill be discussed m ore thoroughly in C hapter V. The approxim ation used here consists o f the resistivity-porosity relation for isolated pores up to the pinch-off porosity value. T he relation for tubes is then used for larger porosity, as the equilibrium po re geom etries form a netw ork o f grain-boundary channels sim ilar in shape to th e tubular porosity m odel of G ran t and W est (1965). A histogram o f electrical resistivities from the com pilation of low er crustal layers is shown to the left in Figure 13, to illustrate w hat porosities are inferred for different pore geom etries.
iv. V elocity-R esistivity re la tio n s
If both velocity variations and resistivity variations in the deep crust a re prim arily associated with porosity, there should be a relation between velocity and resistivity. Theoretical relations between velocity and resistivity (Figure 15) can be obtained from the velocity-porosity relations (Figure 4) and from the resistivity-porosity (Figure 13) already calculated. Again, the fundam ental assum ption is that both electrical resistivity and seism ic velocity are controlled m ainly by porosity. T he effects o f variations in com position and in pore geom etry are assum ed to be second order. T hese m odel relations will be com pared with data from approxim ately coincident m agnetotelluric and seismic refraction results in the next chapter.
Lo
g
R
es
is
ti
vi
ty
(o
h
m
-m
)
R e sistiv ity -P o ro sity R elation s
4
Isol. 0.03
3
2
m = 2 .0
1m = 1.5
H S
00
1
2
3
4
5
Porosity (%)
F ig u re 14 Resistivity-Porosity relations for 5M NaCI pore fluid. Same conventions as in Figure 13.
7.0
6.8
6.6
6.4
0
.
01
,
1.2
0
.
1
,
2.0
0.03, 1.5
0.03, 2 .0
6.2
6.0
4
3
2
i 0Log Resistivity (ohm-m)
F ig u re 15 V elocity-R esistivity relations fo r selected pairs o f pore aspect ratio (a) and A rchie’s Law exponent (m ). T h e thicker solid line (E) represents a discontinuous m odel analogous to equilibrium pore geom etries, using the velocity-porosity relation o f F ig u re 12, and the resistivity-porosity relation for equilibrium pores w ith a 60° dihedral angle. T he relation show s a sharp change because o f the transition from isolated to interconnected porosity.
U> o
31
CHAPTER III GLOBAL COMPILATION OF DEEP GEOPHYSICAL DATA
i. Data sources and selection criteria
To test the validity o f the velocity-resistivity relations obtained above, a compilation o f approxim ately coincident velocities fiom seism ic refraction surveys and resistivities from m agnetotelluric and deep-probing controlled-source electrom agnetic surveys is presented here, as well as some associated data on seism ic reflection and heat flow (T a tie 3). The num ber o f reflection studies reported is sm all, because o f the coincidence lequirem ent. There is a large num ber o f coincident reflection and refraction studies (e.g. M ooney and Brocher, 1987), but since the least equivocal evidence for deep crustal fluids is the low electrical resistivity, the em phasis w as put on m agnetotelluric studies. The refraction and m agntiotelluric survey n-’i r s w ere required to be within the same tectonic unit, usually separated by less than 100 kilom etres. T h e layers reported here were also required to be in the low er half o f the crust, to justify the earlier assum ption o f a predom inantly mafic com position.
Results from a w ide range o f tectonic environm ents have been included. To study w hether the tectonic regim e has an effect on lower crustal physical properties, the sites have been divided according to the most recent ( < 300 M a) tectonic deform ation type in T able 3 as being stable, com pressional, or extensional. It is how ever difficult to determ ine w hat recent event was the m ost influential in determ ining the structure o f the present crust in some areas. F o r exam ple, the C ordillera o f western Canada-northw est U nited States has undergone com pression in the M esozoic, transcurrent dislocation in the Cretaceous-Paleocene, extension in the Eocene, and is under com pression by subduction again today. T he sites have also been divided according to age (m ost recent orogenic or therm otectonic event) to establish if there is any difference in inferred porosity with geologic age.
The electrical resistivity data have been restricted m ainly to broad frequency band m agnetotelluric soundings with modern form s o f data inversion or m odelling. A few controlled source surveys have also been included. D ata from geom agnetic depth sounding and from narrow -band m agnetotelluric surveys have been excluded, the form er
32
because o f their poor vertical resolution for both depth and resistivity, and the latter because tl . . - provide little inform ation on either the deep crust (if only shorter periods are used) o r t..c near-surface (if longer periods are used).
There are tw o serious difficulties in obtaining resistivities over depth ranges that correspond to velocity layers. T he first is that m agnetotelluric surveys only resolve well layer conductance o r thickness/resistivity (in one-dim ension inversions), rather than resistivity alone (e .g ., Edwards et a l., 1981). This problem has been dealt w ith in part by a normalization scheme based on the low er crustal layers being 10 km thick (approxim ately the data average). Applying this additional assum ption as a constraint gives somewhat less scatter in the velocity-resistivity plot. T h e second im portant problem is the static shift (e.g. Jiracek, 1990), caused by local near-surface galvanic effects, resulting in a shift o f the w hole resistivity-depth profile. A m ore thorough discussion o f static shift effects is presented in C hapter IV. A t only a few localities have detailed static shift corrections been made.
Resistivity anisotropy is another im portant problem . The current flow and thus the resistivity sensed : i m agnetotelluric surveys is approxim ately horizontal a t the e a rth ’s surface, but in some o f the surveys in the com pilation, the apparent resistivity curves for the E- (electric field parallel to most conductive direction - strike) and B- (m agnetic field parallel to strike) polarizations are quite different, especially at longer periods. These discrepancies may be caused either by the larger scale geological structure, o r by small grain-scale anisotropy, for exam ple in the pore geom etry. In m any cases, inversion of both E- and B-polarizations gave sim ilar results fo r the low er crust. In cases w here the results greatly differed, the E-polarization inversion has usually been presented, because the investigators judged that it provided the best resistivity average. In m any instances only the E-polarization has been inverted by the original investigators. T h e polarization used is indicated in T able 3.
For seismic refraction velocities, the most serious problem is that low -velocity layers in the deep crust have undoubtedly been m issed in older interpretations that w ere based only on first-arrivals, since such layers often generate only secondary arrivals. T hus, only velocity data that have been obtained through am plitude synthetic m odelling
33 o f first and secondary arrivals have been used in the com pilation. This m inimizes the possibility that low er crustal layers that do not produce first arrivals are missed in the interpretation (see Hyndm an and K lem perer, 1989). Refraction velocities refer prim arily to horizontal propagation, and anisotropy in different horizontal directions are frequently observed in surveys that include m ore than one direction, as w ith m agnetotelluric surveys, although the effect is usually small. In most o f the surveys presented here, only one direction o f propagation is available. Typical uncertainties for deep crustal refraction velocities when prim ary and secondary arrivals are analyzed are about 0.1 km /s. If interm ediate layers are m issed, the uncertainties can be m uch larger.
L ow er crustal seismic reflectivity has not been correlated with velocity and resistivity in detail in this study. In some areas, there may be a depth correlation betw een high reflectivity and low resistivity; exam ples o f coincidence in which the depth o f the conductive layers are defined by controlled source electrom agnetic surveys are given in Connerey et al. (1980) and Haak and Hutton (1986). An exam ple o f low- resistivity from m agnetotelluric data and high seism ic reflectivity in the lower crust for a restricted area in the southern Canadian C ordillera will be presented in the next chapter. W here seism ic reflection surveys have defined bands o f low er crustal reflectors in the areas o f the velocity-resistivity data, the depth range o f the bands has been given by the authors fo r com parison with the resistivity and velocity d ata depths (Table 3). R eflector band depths and thicknesses given only in tw o-w ay reflection tim e have been converted to depth using the seismic refraction velocities.
The depths to the top o f velocity layers cannot be readily tested against the other param eters, because the effect o f porosity on velocity is superim posed on a progression w ith depth to m ore high velocity mafic rocks. It is not practical to try to separate these effects.
The correlation between the tops o f conductive layers and reflective bands with tem perature at depth has also been investigated. It has been shown by Addm (1978), Shankland and A nder (1983), and by K lem perer (1986) respectively that both conductive and reflective layers appear to be influenced by the geotherm al regim e. T he tem peratures ranges given in the com pilation here are from either published tem perature
34
inversions when available, or from published heat flow results converted into tem peratures at depth using the generalize/ continental geotherm s o f Chapm an (1986). There is a considerable uncertainty in inverting heat flow data, especially w hen the surface heat production is not known. Heat flow m easurem ents having typical uncertainties around 10%, deep crustal tem perature uncertainties are about 50°C (Lewis e t a l . , 1992).
Table 3. Compilation of lower crustal resistivity, velocity and inferred temperatures Locality Z ' (km) re c p (fl m) P 10 (0 m) Pol v P (km/s) T ° range (°C) Ref. int. (km) References
PR ECA M B RIA N AREAS
1. C harlevoix 22-30 30-40 S 30-70 70-120 62 100 E 6.9 7.1 225-300 300-Kurtz (1982), Lyons (1980) 2. Laurentides 20-30 30-40
s
12-30 30-70 20 50 E 6.8 7 .0 200-300300-Kurtz (1982), Berry & Fuchs (1973)
3. W isconsin 12-40
s
10-100 28 B 6.5 150- D ow ling (1970)4. Tim m ins 22-30
s
270 250 - 6.6 -6 .8 200-250 Duncan et al. (1980), Boland & Ellis (1989), M isener et al. (1951) 5. Kapuskasing 20-30 30-40s
150 1000 150 1000 6 .6 -7 .0 7 .0 -7 .4 200-300300-M areschal et a l. (1989), Boland & Ellis (1989)
6. K aapval, South A frica
25-40
s
10 6 .7 - 6 .6 -7 .0 450-600 Van Zijl (1977), Bloch et al. (1969), G upta (1989) 7. Indian Shield (Choutuppal) 16-25 25-40s
35 5200 39 3467 6.5 6.8 200-300 300-500S a s try e ta l. (1 9 9 0 ),K a ila e ta l. (1979), G upta (1989)
Ui VI
8. Sm aland- 10-80 S 5 0 0 -104 V arm land, Sweden
9. N orrland, 15-35 S 20-80 Sweden
10. N orbotten, 25-50 S 50-200 Sweden
125 B 6 .5 -7 .4 200-1000 19-25 Rasmussen e ta l. (1987), C lo w e se ta l. (1987), Eriksson & M alm qvist (1979), Dahl-Jensen et al. (1987)
30 B 6.75-7.15 200-325
40 EB 6 .9 300-500
Rasmussen et al. (1987), H irschleber et al (1975), Eriksson & M alm qvist (1979)
Jones e ta l. (1983), Luosto & Korhonen (1986), Eriksson & M alm qvist (1979)
11. Baltic Shield 35-50 S 18-36 20 B 6 .5 -6 .9 380-450
PH A N ER O ZO IC AREAS
12. A ppalachians 20-25 C 10-30 40 - 6 .7 350-400
13. G eorgia 15-22 C 10 14 - 6.05 250-350 22-35 2000 1538 6 .0 5 -6 .7 350-550
Jones et al. (1983), H irschieber et al. (1975), C erm ak & H urtig (1979)
C o n n e re y e ta l. (1981), L u e tg e rte ta l. (1987), Lachenbruch & Sass (1977)
Thom pson et a l. (1983), Kean & Long (1980), Lachenbruch & Sass (1977), C ook & Oliver (1981)
Rio G rande Rift
-14. Santa F e 13-22 E 1-10 5 .5 EB 6.09-6.15 400-450 15-28 H erm ance & Pedersen (1980), Sinno et al. 22-30 50-100 100 6 .6 2 -6 .7 2 450-650 (1986), Lachenbruch & Sass (1977), Brown
et al. (1980)
-15. El Paso 25-35 E 10 10 EB 6 .6 2 -6 .7 2 550-700 sam e as above
u> ON
