12. SEISMIC VELOCITIES OF LEG 37 ROCKS AND THEIR GEOPHYSICAL IMPLICATIONS

Nikolas I. Christensen, Department of Geological Sciences and Graduate Program in Geophysics,University of Washington, Seattle, Washington

INTRODUCTION

The interpretation of seismic refraction velocities inthe oceanic crust is important in understanding theconstitution of the crust and the history of oceanbasins. Of critical importance for such interpretationsare laboratory measurements at elevated pressures andtemperatures on oceanic samples obtained by dredgingand drilling. Recent ultrasonic studies of compressionaland shear wave velocities in oceanic rocks dredgedfrom fracture zones and ridge crests have provenvaluable in understanding the composition of theoceanic crust (e.g., Barrett and Aumento, 1970;Christensen, 1970, 1972; Christensen and Shaw, 1970;Fox et al., 1973; Christensen and Salisbury, 1975). Alarge number of additional velocity measurements ofigneous rocks, largely basalt, are given in severalvolumes of the Initial Reports of the Deep Sea DrillingProject. Of significance, especially for comparisonswith velocities of Leg 37 samples, are the high pressuremeasurements of velocities in metagabbro, meta-diabase, gabbro, and anorthositic gabbro obtainedfrom Site 293 of Leg 31 (Christensen et al., 1975).

In this study velocities are reported to hydrostaticpressures of 10 kbar at ambient temperature, for 13samples obtained from Leg 37 drilling. Emphasis isplaced on the elastic properties of the gabbroic andultramafic rocks from Site 334, with the objective ofunderstanding the geologic processes which wereresponsible for the emplacement of this rock suite atsuch shallow crustal depths.

VELOCITY DATA AND ELASTIC CONSTANTS

Velocities as a function of pressure and bulk densitiesare tabulated in Table 1. The travel times of thecompressional and shear waves were measured by thepulse transmission technique using a mercury delay line(Birch, 1960). The velocities are estimated to beaccurate to better than 1% (Birch, 1960; Christensenand Shaw, 1970). All samples were water saturatedprior to the measurements, and pore pressures weremaintained at values lower than external pressures byplacing a 100-mesh screen between the samples andcopper jackets. The screening requires considerabletime in the process of sample preparation; however, it isessential that water be allowed to drain from the porespaces during the application of pressure.

Several samples studied to 10 kbar were identical tothose used for ship velocity measurements (Table 1).The core ends of Samples 332B-3-3, 100-103 cm, 333A-3-1, 129-131 cm, 334-22-1, 69-71 cm, 334-24-1, 63-65

cm, and 335-11-1, 76-78 cm were relapped prior to ourmeasurements, because they were found to deviatesignificantly from parallelism. In order to detectpossible anisotropy, velocities were measured in twocores cut perpendicular to one another for Sample 334-23-1, 135-144 cm.

In addition to velocities, several parameters areuseful in describing elastic properties and interpretingcomposition from seismic refraction studies (e.g., Birch1961; Christensen, 1972). These include the shearmodulus or rigidity (µ), Lames constant (λ), the bulkmodulus (Ar), Young's modulus (£), compressibility(ß), Poisson's ratio (σ), and the seismic parameter (0),defined as K/pπ, where p is density. Although there areseveral techniques for obtaining these constants, all canbe calculated conveniently from Vp, Vs, and p (Birch,1961). Values of these parameters at selected pressuresare given in Table 2. Velocities and densities used in thecalculations were corrected for compression using aniterative routine and the dynamically determinedcompressibilities.

VELOCITY-DENSITY RELATIONS

Tendencies for compressional and shear velocities toincrease with density have been amply demonstrated(e.g., Birch, 1960; Simmons, 1964) and simplecorrelations often have been presented for oceanicbasalts (e.g., Christensen and Shaw, 1970; Christensenand Salisbury, 1973). With more extensive studies of awider variety of oceanic rocks now available, it is clearthat proportionality holds only for specific rock types(Christensen and Salisbury, 1975). For example,compressional wave velocities of serpentinized oceanicrocks have been found to be approximately 8% higherthan basalts of similar density, whereas shear velocitiesfor both rocks types are generally similar for a givendensity. Thus compressional wave velocity-densityrelations and many of the elastic constants ofserpentinized rocks differ from those of basalts. Inparticular, the ratios of Vp to Vs and hence Poisson'sratios of serpentinized oceanic rocks are relatively high(Christensen, 1972). Poisson's ratios of 0.35 (Table 2)for the two partially serpentinized rocks included in thisstudy (334-23-1, 38-44 cm and 334-26-1, 18-23 cm)agree well with these earlier findings.

The elastic properties of gabbros also differsignificantly from basalts; a difference in compressionalwave velocity of about 10% is found for oceanicgabbros and basalts of similar density (Christensen andSalisbury, 1975). This tendency for compressional wavevelocities to increase with increasing grain size in mafic

389

N.I. CHRISTENSEN

TABLE 1Compressional (P) and Shear (S) Wave Velocities

Sample(Interval in cm)

332A-31-3,79-82332-A-36-2,33-36332B-3-3,100-103332B-15-1,129-131333A-3-1,129-131334-22-1,69-71334-23-1,38-44334-23-1,135-144

mean

mean334-24-1,63-65334-25-1,84-92334-26-1,18-23335-11-1,76-78

BulkDensity

2.783

2.183

2.8302.8302.863

2.8022.8023.0083.0082.7082.7083.0062.9953.0013.0062.9953.0012.8692.8692.9552.9552.6442.6442.8302.830

Mode

P

P

P

sP

P

sP

sP

sPPP

sssP

sP

sP

sP

s

0.2

5.59

6.06

6.633.365.94

5.612.886.973.565.912.856.967.057.013.753.763.766.963.526.853.615.762.756.113.28

Velocity (km/sec)at Varying Pressures (kbar)

0.4

5.67

6.09

6.663.385.98

5.642.947.003.585.952.876.997.087.043.773.783.787.013.546.883.635.812.786.143.29

0.6

5.71

6.12

6.693.406.01

5.672.987.023.605.972.887.017.107.063.783.803.797.063.576.903.645.832.796.163.29

0.8

5.74

6.15

6.713.426.03

5.703.027.043.635.992.907.027.127.073.793.823.817.093.596.923.665.902.816.183.30

1.0

5.76

6.17

6.723.436.04

5.733.057.063.656.012.917.047.147.093.803.833.827.123.606.943.675.942.826.203.31

2.0

5.85

6.26

6.783.486.11

5.863.157.113.726.082.947.097.207.153.843.873.867.203.666.993.706.112.866.263.34

4.0

5.95

6.34

6.853.546.20

5.993.247.173.786.162.977.167.287.223.893.923.917.293.737.053.726.282.926.333.37

6.0

6.03

6.39

6.893.586.26

6.083.277.223.826.212.997.217.347.283.923.943.937.343.767.093.746.342.966.383.37

10.Q

6.18

6.46

6.953.626.33

6.263.297.323.876.313.007.317.437.373.963.953.967.433.807.153.766.403.006.463.38

rocks was first observed by Birch (1961) in acomparison of gabbro and diabase velocities. Since theinfluence of porosity on rock velocities is notcompletely eliminated until hydrostatic pressures inexcess of 10 kbar are applied (Christensen, 1974a), theexplanation of these differences is apparently related todecreasing grain boundary porosity associated withincreasing grain size.

The above relationships between velocity and densityalso hold for the Leg 37 rocks and are illustrated inFigure 1. The compressional wave velocities of gabbrosand serpentinized rocks fall predictably above thebasalt data, whereas shear velocities tend to be onlyslightly higher than oceanic basalts of similar density.Thus, as observed in many earlier studies (Hughes andMaurette, 1957; Birch, 1960, 1961; Simmons, 1964;Christensen and Salisbury, 1975), Poisson's ratios ofgabbros tend to be relatively high, usually rangingbetween 0.30 and 0.32. Recent velocity measurementsin lunar gabbroic anorthosites also show high Poisson'sratios, even though the rocks were not water saturated.Wang et al. (1974) report an average Poisson's ratio of0.31 for 16 samples of gabbroic anorthosite. It is wellknown that water saturation increases Vp much greaterthan Vs, thus ratios of Vp to Vs for these rocks wouldpresumably be even higher if velocities were measuredfor saturated specimens.

The gabbros included in this study (334-22-1, 69-71cm; 334-23-1, 135-144 cm; 334-24-1, 63-65 cm; 334-25-1, 84-92 cm) have Poisson's ratios at oceanic crustalpressures ranging from 0.30 to 0.33 (Table 2), whichagree well with previous values for similar rocks. The

highest Poisson's ratio (0.33 for Sample 334-24-1, 63-65cm) is a consequence of the relatively high (~65%)feldspar content of this rock. Feldspathic rocks havehigh Poisson's ratios (Birch, 1961), which show atendency to increase with increasing feldspar content(Christensen and Salisbury, 1975).

OCEANIC CRUSTAL STRUCTUREAND COMPOSITION

Early studies of oceanic crustal seismic structure byHill (1957) and a summary of numerous refractionprofiles by Raitt (1963) proposed a three-layer modelfor the oceanic crust which is still generally acceptedtoday. This model consists of a 0.5 to 1 km layer ofsediments overlying a second layer with a relatively welldefined thickness (1.0 to 2.5 km) and a highly variablecompressional wave velocity (3.5 to 6.5 km/sec).Drilling and dredging have shown that layer 2 is usuallycomposed of basalt; its variable velocity most likely isrelated to weathering, fracturing, intrusion by sills andfeeder dikes, and intercalated sediment. Beneath layer 2and immediately overlying the mantle is layer 3, a unitof variable thickness (usually 3.4 to 6.3 km) and uni-form velocity (usually 6.4 to 7.0 km/sec). Recentsonobuoy studies are generally consistent with theearlier studies and often show more detail than theearlier refraction profiles (e.g., Talwani et al., 1971;Sutton et al., 1971). In particular, layer 2 often is foundto have a thin, low velocity cap (2.5 to 3.8 km/sec),while layer 3 often contains a basal layer with velocitiesranging from 7.1 to 7.7 km/sec.

390

SEISMIC VELOCITIES

TABLE 2Elastic Constants, Leg 37

Sample(Interval in cm)

332B-3-3,100-103

333A-3-T,129-131

334-22-1,69-71

334-23-1,38-44

334-23-1,135-144

334-24-1,63-65

334-25-1,84-92

334-26-1,18-23

335-11-1,76-78

Pressure(kbar)

0.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.00.41.02.06.0

10.0

vp/vs

1.971.961.951.921.921.921.881.861.861.901.961.931.911.891.892.072.072.072.082.101.861.861.851.851.861.981.981.961.951.961.891.891.891.891.902.092.102.142.142.131.871.871.871.891.91

σ

0.330.320.320.320.310.310.300.300.300.310.320.320.310.310.310.350.350.350.350.350.300.300.290.290.300.330.330.330.320.320.310.310.310.310.310.350.350.360.360.360.300.300.300.310.31

Φ(km/sec)2

29.1129.4529.7730.2530.6020.2820.4121.0622.5724.5231.9032.0632.0632.5433.4024.4124.8025.3926.5027.5730.5030.7931.2132.2733.1932.4233.3933.9334.8935.7229.7630.1830.5631.4832.0523.4424.6526.3828.3628.7123.2623.8124.2725.4226.27

K(Mb)

0.820.830.840.860.880.570.570.590.640.700.960.970.970.981.010.660.670.690.720.760.920.920.940.971.010.930.960.981.011.030.880.890.910.940.960.620.650.700.760.770.660.670.690.730.75

ß Λ

(Mb"1)

1.211.201.181.161.141.761.751.691.571.431.041.041.031.020.991.511.491.451.381.321.091.081.071.030.991.071.041.030.990.971.141.121.101.071.041.611.531.431.321.301.521.481.451.381.33

(Mb)

0.320.330.340.360.370.240.260.280.300.300.390.400.420.440.450.220.230.230.240.240.430.440.450.460.470.360.370.380.410.420.390.400.400.41ú.420.200.210.220.230.240.310.310.320.320.32

E(Mb)

0.860.880.910.960.980.630.680.720.780.801.021.061.091.151.180.600.620.630.660.661.111.141.161.201.220.960.991.021.071.101.021.041.061.081.100.550.570.590.630.650.800.810.820.840.85

λ

(Mb)

0.610.610.620.620.630.410.400.410.440.490.700.700.690.690.710.510.520.530.560.590.630.630.640.660.690.690.710.720.740.760.620.630.640.660.680.480.510.560.600.610.450.470.480.510.54

Studies of the structure, petrology, and physicalproperties of rocks from ophiolites (Salisbury, 1974;Christensen and Salisbury, 1975) and velocities of rocksdredged from Fracture zones (Christensen, 1974b;Christensen and Salisbury, 1975) have shown that thelower oceanic crust consists of a metamorphicassemblage containing hornblende and plagioclase,which overlies a sequence of late differentiates, which,in turn, grade downward into cumulate gabbro. Thehornblende metagabbro occurs within sheeted dikesand has elastic properties similar to those observed forlayer 3 from seismic refraction studies (Vp = 6.5-6.9km/sec, σ = 0.26-0.29). The boundary between layers 2and 3 does not separate intrusive from extrusive crustalrocks, but instead marks the boundary of ametamorphic front which separates an upper section ofchloritized sheeted dikes from similar underlying rockswhich contain abundant amphibole (Salisbury, 1974).

Underlying the sheeted dikes are late quartz-richdifferentiates which produce velocity inversions withinlayer 3. The underlying thick section of cumulategabbros has compressional wave velocities generallyranging between 6.9 and 7.3 km/sec and Poisson'sratios of 0.30 to 0.33.

The velocities and Poisson's ratios of the gabbrosrecovered from Leg 37 agree well with previousmeasurements on similar rocks and observations oflower crustal seismic structure. Thus it is possible thatthe gabbros from Site 334 represent samples of thelower portions of the cumulate section of layer 3.However, it should be emphasized that the agreementbetween the laboratory velocity measurements andlower crustal velocities does not prove that thesegabbros are samples of the lower oceanic crust. Ofsignificance, it should be noted that between 1 and 3 kmof layer 3 are missing at Site 334, if the gabbros are

391

N.I. CHRISTENSEN

7.0

6.0

5.0

, km

/sec

VE

LOC

ITY

b

3.0

2.0

1 1

-

-

-

• • a

1 '

•

I i i

i

o

1

o β

1 'Δ

ΔΔ

V

-

-

-

12.0 2.2 2.4 2.6 2.8 3.0

DENSITY, gm/cc

Figure 1. Velocities at 0.5 kbarfor previously studied DSDPbasalts (solid dots) and Leg 37 basalts (open circles), ser-pentinized rocks (diamonds), and gabbros (triangles).

lower crustal cumulates. This would require extremelycomplex tectonic activity to remove thick sections ofmetagabbro and late differentiates.

Alternate possibilities come to mind to explain thepresence of gabbro at shallow crustal depths. Since Site334 is located near a major fault scarp, the coarse-grained igneous rocks may have been emplaced inupper crustal levels by a mechanism similar to thatproposed by van Andel et al. (1969), in which new crustis formed within "leaky" fracture zones. Similargabbros, which have been dredged from several majorfracture zones (e.g., Bonatti et al., 1970; Thompson andMelson, 1972), apparently have originated in thismanner. Recently Christensen and Salisbury (1975)have proposed that intermittent off-ridge intrusion fedfrom underlying anomalous mantle is responsible forthe seismic observations of increasing layer 3 thicknessaway from ridge crests. Even though these gabbroicintrusions usually solidify within the lower levels oflayer 3, it is probable that some reach the upper levelsof the oceanic crust and small gabbroic bodies formwithin layer 2.

ACKNOWLEDGMENTS

I would like to thank R. Hyndman who provided many ofthe specimens used in this study. The velocity measurementswere made by M. Brown and R.L. Carson. J. Hull helpedwith the data reduction and tabulation. The research was

supported by the Office of Naval Research Contract N-00014-67-A-0103-0014.

REFERENCES

Barrett, D.L. and Aumento, F., 1970. The mid-Atlantic ridgenear 45°N. XI. Seismic velocity, density and layering ofthe crust: Canadian J. Earth Sci., v. 70, p. 1117-1124.

Birch, F., 1960. The velocity of compressional waves in rocksto 10 kilobars, 1: J. Geophys. Res., v. 65, p. 1083-1102.

, 1961. The velocity of compressional waves in rocksto 10 kilobars, 2: J. Geophys. Res., v. 66, p. 2199-2224.

Bonatti, E., Honnorez, J., and Ferrara, G., 1970. Equatorialmid-Atlantic ridge: petrologic and Sr isotope evidence foran alpine type rock assemblage: Earth Planet. Sci. Lett.,v. 9, p. 247-256.

Christensen, N.I., 1970. Compressional wave velocities inbasalts from the Juan de Fuca Ridge: J. Geophys. Res.,v. 75, p. 2773-2775.

, 1972. The abundance of serpentinites in the oceaniccrust: J. Geol., v. 80, p. 709-719.

., 1974a. Compressional wave velocities in possiblemantle rocks to pressures of 30 kilobars: J. Geophys. Res.,v. 79, p. 407-412.

., 1974b. The petrologic nature of the lower oceaniccrust and upper mantle. In Kristjansson, L. (Ed.),Geodynamics of Iceland and the North Atlantic area:Dordrecht-Holland (D. Reidel Publishing Co.), p. 165-176.

Christensen, N.I. and Salisbury, M.H., 1973. Velocities,elastic moduli and weathering-age relations for PacificLayer 2 basalts: Earth Planet. Sci. Lett., v. 19, p. 461-470.

, 1975. Structure and constitution of the loweroceanic crust: Rev. Geophys. Space Phys., v. 13, p. 57-86.

Christensen, N.I. and Shaw, G.H., 1970. Elasticity of maficrocks from the mid-Atlantic ridge: Geophys. J. Roy.Astron. Soc, v. 20, p. 271-284.

Christensen, N.I., Carlson, R.L., Salisbury, M.H., andFountain, D.M., 1975. Elastic wave velocities in volcanicand plutonic rocks recovered on DSDP Leg 31. In Ingle,J.C., Jr., Karig, D.E., et al., Initial Reports of the DeepSea Drilling Project, Volume 31: Washington (U.S.Government Printing Office), p. 607-610.

Fox, P.J., Schreiber, E., and Peterson, J.J., 1973. The geologyof the oceanic crust: compressional wave velocities ofoceanic rocks: J. Geophys. Res., v. 78, p. 5155-5172.

Hill, M.N., 1957. Recent geophysical exploration of theocean floor: Phys. Chem. Earth, v. 2, p. 129-163.

Hughes, D.S. and Maurette, C, 1957. Variation of elasticwave velocities in basic igneous rocks with pressure andtemperature: Geophysics, v. 22, p. 23-31.

Raitt, R.W., 1963. The crustal rocks. In Hill, M.N. (Ed.), Thesea, v. 3: New York (Wiley), p. 85-102.

Salisbury, M.H., 1974. Investigation of seismic velocities inthe Bay of Islands, Newfoundland ophiolite complex forcomparison with oceanic seismic structure: Ph.D. thesis,University of Washington, Seattle.

Simmons, G., 1964. The velocity of shear waves in rocks to 10kilobars, 1: J. Geophys. Res., v. 69, p. 1123-1130.

Sutton, G.H., Maynard, G.L. and Hussong, D.M., 1971.Widespread occurrence of a high-velocity basal layer in thePacific crust found with repetitive sources and sonobuoys.In Heacock, J.G. (Ed.), The structure and physicalproperties of the earth's crust, Geophys. Monogr. Ser.,v. 14, Washington (Am. Geophys. Union), p. 193-209.

392

SEISMIC VELOCITIES

Talwani, M., Windisch, C.C., and Langseth, M.G., Jr., 1971. van Andel, T.H., Phillips, J.D., and Von Herzen, R.P., 1969.Reykjanes ridge crest: a detailed geophysical study: J. Rifting origin for the Vema Fracture in the NorthGeophys. Res., v. 76, p. 473-517. Atlantic: Earth Planet. Sci. Lett., v. 5, p. 296-300.

Thompson, G. and Melson, W.G., 1972. The petrology of Wang, H., Todd, T., Richter, D , and Simmons, G., 1974.oceanic crust across fracture zones in the Atlantic Ocean: Elastic properties of plagioclase aggregates and seismicEvidence of a new kind of sea-floor spreading: J. Geol., velocities in the moon: Fourth Lunar Sci. Conf. Proc,v. 80, p. 526-538. Geochim. Cosmochim. Acta, Suppl. 4, v. 3, p. 2663-2671.

393