Two and three-dimensional inversions of magnetic anomalies in the MARK area (Mid-Atlantic Ridge 23° N

Two- and Three-Dimensional Inversions of Magnetic Anomalies in the MARK Area (Mid-Atlantic Ridge 23 ~ N)

NORBERT J. SCHULZ, ROBERT S. DETRICK

Graduate School of Oceanography, University of Rhode Island, Kingston, RI 02881, U.S.A.

and

STEPHEN P. MILLER

Department of Geological Sciences and Marine Science Institute, University of California at Santa Barbara, Santa Barbara, CA 93106, U.S.A.

(Received 24 September, 1987; revised April, 1988)

Key words: Mid-Atlantic Ridge, seafloor spreading, magnetic anomalies.

Abstract. Magnetic data collected in conjunction with a Sea Beam bathymetric survey of the Mid-Atlantic Ridge south of the Kane Fracture Zone are used to constrain the spreading history of this area over the past 3 Ma. Two-dimensional forward modeling and inversion techniques are carried out, as well as a full three-dimensional inversion of the anomaly field along a 90-km- long section of the rift valley. Our results indicate that this portion of the Mid-Atlantic Ridge, known as the MARK area, consists of two distinct spreading cells separated by a small, zero-offset transform or discordant zone near 23~ The youngest crust in the median valley is characterized by a series of distinct magnetization highs which coalesce to form two NNE- trending bands of high magnetization, one on the northern ridge segment which coincides with a large constructional volcanic ridge, and one along the southern ridge segment that is associated with a string of small axial volcanos. These two magnetization highs overlap between 23~ and 23~ forming a non-transform offset that may be a slow spreading ridge analogue of the small ridge axis discontinuities found on the East Pacific Rise. The crustal magnetizations in this overlap zone are generally low, although an anomalous, ESE-trending magnetization high of unknown origin is also present in this area. The present-day segmentation of spreading in the MARK area was inherited from an earlier ridge-transform-ridge geometry through a series of small ( ~ 10 km) eastward ridge jumps. These small ridge jumps were caused by a relocation of the neovolcanic zone within the median valley and have resulted in an overall pattern of asymmetric spreading with faster rates to the west (14mmyr -I) than to the east (11 mmyr-1). Although the de- tailed magnetic survey described in this paper extends out to only 3 Ma old crust, a regional compilation of magnetic data from this area by Schouten et al. (1985) indicates that the relative positions and dimensions of the spreading cells, and the pattern of asymmetric spreading seen in the MARK area during the past 3 Ma, have characterized this part of the Mid-Atlantic Ridge for at least the past 36 Ma.

Marine Geophysical Researches 10: 41-57, 1988. �9 1988 Kluwer Academic Publishers. Printed in the Netherlands.

1. Introduction

The complexity and three-dimensional nature of the crustal accretion process at slow-spreading ridges have been well-documented in recent studies of the MARK (Mid-Atlantic Ridge at Kane) area on the Mid-Atlantic Ridge near 23 ~ N (Figure 1). This area has been the site of numerous marine geological and geophysical experiments dating back to the late 1960s (Miyashiro et al., 1969, 1970; van Andel et al., 1968; Fox, 1972; Purdy et al., 1978; Detrick and Purdy, 1980; Bryan et al., 1981; Louden and Forsyth, 1982; Karson and Dick, 1983; Purdy and Detrick, 1986). Recently, in preparation for drilling in this area by the Ocean Drilling Program (ODP), high-resolution Sea Beam bathymetry maps were obtained of the Kane Transform (Pockalny et al.,

1988) and a 100-km-long portion of the Mid-Atlantic ridge rift valley immediately to the south (Detrick et al., 1984; Kong et al., this issue). As part of this pre-drilling site survey, nearly complete Sea MARC I side scan sonar imagery was also obtained of the inner rift valley south of the Kane Fracture Zone (Mayer et al., 1985; Kong et al., this volume). These surveys, a subsequent ALVIN/ANGUS investigation (Karson et al., 1987) and the drilling on ODP Legs 106 and 109 (Detrick, Honnorez, Bryan, Juteau et aL, 1988) have made the MARK area one of the best known portions of the Mid-Atlantic Ridge.

A synthesis of results from all of these studies indicates that spreading along this/portion of the Mid-Atlantic Ridge is far more complicated than

42 NORBERT J. SCHULZ ET AL.

1 2 0 ~ 8 0 ~ 4 0 ~ 0 ~

800 60 o 40 o 20 ~ 0 ~ 20 ~

60 ~

400

200

0 o

Fig. 1. Map of the North Atlantic showing the location of the MARK area, a well-studied portion of the Mid-Atlantic Ridge south of the Kane Fracture Zone.

previously suspected, with dramatic changes along- strike in the morphology and crustal structure of the rift valley, the development of the neovolcanic zone and the style and magnitude of tectonic extension (Karson et al., 1987). These results suggest that the median valley in the M A R K area can be divided into at least two distinct spreading center segments or cells (Schouten et aL, 1985). The northern segment consists of a magmatically active spreading cell about 40 km in length that extends from the ridge- transform intersection south to about 23~ N. The rift valley in this area is dominated by a large constructional volcanic ridge that is young and associated with high-temperature hydrothermal activity (Leg 106 Scientific Party, 1986; Detrick, Honnorez et

al., 1988). A much older, magrnatically inactive segment is present in the southern part of the M A RK area (south of 23o10 ' N). The floor of the rift valley in this area is dotted with numerous small conical volcanos that are built upon relatively old, fissured and sediment-covered lavas, and in some cases are themselves fissured and faulted (Mayer et al., 1985; Kong et al., this volume). This cell appears to be in a stage of tectonic extension with only small isolated eruptions. These two segments are separated by an anomalous zone between 23~ ' and 23~ that

lacks a well-developed rift valley or a well-defined neovolcanic zone. The rift valley floor in this zone has been intensely fissured and faulted, and is associated with a large outcrop of serpentinized peridotite that was drilled on ODP Leg 109 (Leg 109 Scientific Party, 1986; Bryan, Juteau et al., 1988).

These observations raise important questions about the tectonic evolution of the Mid-Atlantic Ridge in the M A R K area. How stable are these cells in terms of their relative positions and dimensions? Is the tectonic evolution of adjacent segments linked, or do they behave largely independently of one another? What is the nature of the boundary separating these segments? Some of these questions can be addressed by an analysis of the magnetic anomaly patterns preserved in the older crust flanking the rift valley and the magnetization distribution of the very young crust within the rift valley itself. In this paper we present the results of two- and three-dimensional inversions of surface magnetic data collected in conjunction with the Sea Beam bathymetry survey (De- trick et al., 1984). Our results are consistent with a segmentation of the M A R K rift valley into two distinct spreading cells separated by a small, non- transform offset that may be a slow spreading ridge analogue of the small ridge axis discontinuities found

TWO- AND THREE-DIMENSIONAL INVERSIONS OF MAGNETIC ANOMALIES IN THE MARK AREA 43

on the East Pacific Rise. This segmentation was inherited from an earlier ridge-transform-ridge geometry through a series of small eastward jumps of the axis of accretion resulting in an overall pattern of asymmetric spreading. The relative positions and dimensions of the spreading cells in the MARK area have been stable for several tens of million of years.

2. Magnetic anomaly data

Total intensity magnetic measurements were made aboard the R/V Conrad in conjunction with a Sea Beam bathymetric survey of the MARK area in late 1984 (Detrick et al., 1984). Magnetic track coverage is shown in Figure 2. Line spacing was typically 3-4 km with ridge-normal lines extending out to the peaks of the flanking rift mountains (crust ,-~ 1.8 Ma old). Six longer lines, separated by about 20 km, extend out to 3 Ma old crust on both sides of the

ridge. Navigation was by a combination of Global Positioning System (GPS) satellites, transit satellites and dead reckoning. The final ship navigation was determined by constraining the transit satellite-navigated Sea Beam swaths to fit with the GPS-navigated swaths where they intersected or overlapped. The magnetic data were reduced to anomaly form by subtracting the International Geomagnetic Reference Field (Peddie, 1982) from each data point.

Figure 3 shows the magnetic anomaly map of the MARK area constructed from these data. The most striking feature of this map is the complexity of the anomaly field. Anomalies are typically lineated for distances of only 10-20 km along the ridge axis, are frequently asymmetric, and vary significantly in shape and amplitude along the rise axis. The MARK area is clearly not characterized by the simple, two- dimensional magnetic anomalies generally associated with the Vine-Matthews hypothesis. Although this

2fz, o'

0 �9

- ~ _ 3 O 0

_ 2 2 o : . . 5 0 ,

Fig. 2. Magnetic track lines form Conrad 25-11 superimposed on a Sea Beam bathymetry map (contour interval 1000 m). Profiles are shown in Figure 4.

44 N O R B E R T J. S C H U L Z ET AL.

23o30 '

2 3 o 0 0 '

45030 , 4 5 * 0 0 ' 44o30 '

Fig. 3. Contour map of the anomalous magnetic field in the MARK area. Contour interval 50 nT. Note the complexity of the anomaly field and the along-strike segmentation of the central anomaly.

complexity partly reflects the rugged topography at the Mid-Atlantic Ridge, we will demonstrate that it primarily indicates the three-dimensional nature of accretionary processes in the M A R K area.

The most prominent magnetic anomaly in this area is the large positive anomaly associated with the

Brunhes normal polarity interval. Thi s anomaly is segmented into a series of elongate highs which trend between 015 ~ and 020 ~ , slightly oblique to the overall trend of the median valley. In the southern portion of the rift valley the central anomaly consists of a single peak skewed toward the east side of the valley over a deep basin that lies at the foot of the eastern rift valley wall. The anomaly amplitudes ( > 350 nT) are the largest found anywhere within the M A R K area. To the north, the central anomaly narrows to

less than 10 km near 22055 , N, then broadens into a wide ( > 20 km) double-peaked anomaly that is centered over the shallowest part of the rift valley about 70 km south of the Kane Fracture Zone. Farther north, the central anomaly nearly disappears around 23~ in the topographically disturbed zone that Karson et al. (1987) interpret to be the boundary between two spreading cells. North of this discordant zone, the central anomaly consists of three distinct highs. Two of these highs are centered over the median volcanic ridge in the northern part of the rift valley, while the third high is located on the eastern wall of the rift valley, about 20 km southeast of the ridge-transform intersection.

The complexity of the central anomaly in the M A R K area is also reflected in the older, flanking


anomalies. The negative anomaly lying between the central anomaly and anomaly 2 is well-developed on the western ridge flank south of 23 ~ N, but much more poorly defined east of the ridge and nearly absent around the 23o15 , N discontinuity. The Brunhes-Matuyama boundary is relatively straight and linear on the eastern ridge flank (south of 23010 , N), but extremely convoluted west of the ridge axis.

The magnetic anomaly field in the MARK area suggests this ridge segment has had a complex tectonic evolution over the past 3 Ma. Our approach in the following analysis will be to use simple two- dimensional block models to identify magnetic anomalies and estimate spreading rates. The longer profiles are then interpreted using a two-dimensional inversion technique to remove the distortions in the

anomaly field caused by sea floor topography. Fol- lowing this two-dimensional analysis, the observed bathymetry and magnetic anomaly field are used to carry out a full three-dimensional inversion to deter- mine the distribution of magnetization in the MARK area.

3. Two-Dimensional Forward Modeling and Inversion

Figure 4 shows a compilation of thirty-six magnetic anomaly profiles across the MARK rift valley between 23~ and 22~ (profile locations are shown in Figure 2). The profiles have been projected onto an azimuth of 100 ~ normal to the trend of the median valley, and are aligned over the center of the topographically-defined rift valley. A model profile is shown for comparison based on spreading rates

CA

R.5 ~ 3

R.IO

2' 2

c A ,

CA 2' 3 2 ~ MODEL

mB3BI1 I T[]i ]! 1 . 4 1 ~ ] ~ 1.13 HALF RATE (cm/yr)

Fig. 4. Magnetic anomaly profiles across the M A R K rift valley. The profiles are arranged from north ( top left) to south (bot tom right) and have been projected along an azimuth of 100 ~ (profile locations are shown in Figure 2). The model profile was computed assuming asymmetric spreading at 14.1 m m y r -~ to the west and 11.3 m m y r -~ to the east using the magnetic reversal time scale of Kent and Gradstein (1986). Other model parameters are: d e p t h = 3 . 5 k m ; source layer t h i c k n e s s = 0 . 5 k m ; magne t iza t ion= 1 5 A m - ~ ; present inc l ina t ion=43.5 ~ d e c l i n a t i o n = - 1 8 ~ remanent inc l ina t ion=40.3 ~ dec l ina t ion=0 ~ The black blocks are normally magnetized.

CA = central anomaly, J = Jaramillo anomaly.

46 N O R B E R T J. $ C H U L Z E T AL.

estimated for profiles R.22-R.27, near the middle of the survey area. In calculating this profile we used the magnetic reversal time scale of Kent and Gradstein (1986) and assumed a 500 m thick magnetic source layer located at 3.5 km depth and a uniform magnetization of 15 A m- 1.

Previous interpretations of magnetic anomalies in this area (Purdy et al., 1978) indicated a spreading half-rate of 23.5 mm yr -~ between anomaly 3' and 4' ( -,~ 5 to 8 Ma ago) with rates slowing to 14.4 mm yr- over the past 4 Ma. Our longest profiles extend to just beyond anomaly 2' so we can only constrain the spreading history during the past 3 Ma. The central anomaly can be clearly identified throughout the area, except on profiles R. 1-R.4 immediately south of the Kane ridge-transform intersection, and near the 22 ~ 15' N discontinuity (profile R. 14). The character of the central anomaly changes systematically along the median valley. In the southern rift valley it is characterized by a single high juxtaposed against the eastern rift valley wall (profile R.26-R.36). This gradually changes into a broad, double-peaked anomaly in the central part of the MARK area (profiles R.25-R.16), but farther north the central anomaly is again associated with a single peak that steps off systematically to the east as a series of short 10-15 km-long segments. Anomalies 2 and 2' are also easily recognized, especially throughout the southern part of the area (profiles R.20-R.35).

The position of the older anomalies relative to the central anomaly indicates that spreading has been asymmetric during the past 3 Ma with faster rates to the west (14.1mmyr -1) than to the east (11.3 mm yr-1). The regional compilation of magnetic anomaly data by Schouten et al. (1985) indicates that asymmetric spreading has characterized the Mid-Atlantic Ridge south of the Kane Fracture Zone since at least anomaly 13 time (~36 Ma ago). During the past 3 Ma this asymmetric spreading has resulted in part, from a series of small ( ~ 10-20 km) eastward jumps in the axis of accretion. The primary evidence for these small ridge jumps is the duplica- tion of the Jaramillo event on the western ridge flank on profiles R.5-R. 10 and R.26-R.36, and anomaly 2 on the western ridge flank on profile R.20. These ridge jumps have been modeled in Figure 5. In constructing these models the asymmetric spreading rates estimated above were assumed (14 mm yr-1 to the west, 11 mm yr 1 to the east) and the location of

the ridge axis was suddenly shifted (as indicated by the arrows in Figure 5) to explain the duplicate anomalies noted above. In addition to the forward modeling, the observed magnetic field was inverted for the magnetization distribution assuming a 1 km source layer using the two-dimensional inversion technique of Parker and Huestis (1974).

One of the best examples of a small ridge jump is profile R.6 (Figure 5a) which crosses the rift valley about 15 km south of the ridge-transform intersection near 23~ ' N. On the western half of this profile, anomaly 2 is separated from the central anomaly by two small peaks which we interpret to be duplicate Jaramillo events formed by a 15-20 km eastward jump of the axis of accretion about 0.8 Ma ago. This jump has left an abandoned rift axis in the western rift mountains and a central anomaly that is offset slightly east of the center of the rift valley (Figure 5a). The linear volcanic ridge found in the median valley in this part of the MARK area is located near the western edge of normally magnetized crust defined by the central anomaly. The magnetization distribution associated with the central anomaly has two peaks. The highest ( ~ 20 A m- 1) is associated with this median ridge and is consistent with its interpretation as the neovolcanic zone in this part of the rift valley (Karson et al., 1987; Kong et

al., this volume). However, the inversion solution suggests that young (<0.73 Ma), normally magnetized crust extends east out of the rift valley onto the arcuate ridge which forms the east wall of the northern rift valley. Karson and Dick (1983) have reported evidence from ANGUS photogeology for sediment-free pillow lavas, constructional volcanic highs, and a large fissure swarm along the crest of this ridge that are consistent with this crust being relatively young.

Profile R.27, which crosses the rift valley near 22~ shows evidence for a similar eastward jump of the accretion axis in the southern part of the MARK area at about the same time (Figure 5b). The main evidence for this jump is the extra crust present on the western ridge flank between anomaly 2 and the central anomaly and the apparent duplica- tion of the Jaramillo event. In this case, the jump is about 10 km or slightly less than the present width of the rift valley.

An older ridge jump occurring after anomaly 2 ( ,,~ 1.9 Ma) has been modeled on profile R.20 in the


FIELD (aT)

INVERSION (omplm}

MODEL(n~

BATHYMETRY(Km)

2oo A

_:o

-20

i o , j 2 j , 2 ~ C A 2 2"

km

I I R. 6 0 25

FIELD (nT}

INVERSION (emp;m|

MODEL(nT}

BATHYMETRY(Km)

zoo

20

i 200/3 / ' ~

- 2 0 0

i 0 2 ' 2 J J C A 2 2

km

I R. 2 7 25

FIELD (.T)

INVERSION (amplm}

MODEL (nT)

BATHYMETRY(Km)

200 / ~

2O

-200

km

I I R. 2 0 0 25

Fig. 5. Comparison of observed anomalies with model profiles calculated assuming asymmetric spreading (14mmyr -1 to the west, 11 mm yr- 1 to the east) and ridge jumps indicated by the arrows. Model parameters are the same as in Figure 4. Also shown for each profile are calculated magnetization distributions assuming a 1-km thick source layer using the two-dimensional inversion technique of Parker and Heustis (1974). (a) profile R.6 modeled with a 15 km eastward ridge jump ~0.8 Ma ago. (b) profile R.27 modeled with a 15 km eastward

ridge jump ~0.8 Ma ago. (c) profile R.20 modeled with a 10 km eastward ridge jump ~ 1.9 Ma ago.

48 NORBERT J. SCHULZ ET AL.

central part of the area (Figure 5c). This jump was

about 10 km and also is to the east, leaving a duplicate anomaly 2 on the western ridge flank. Although the Jaramillo event is difficult to identify on this

profile in the observed anomaly field, it is clearly seen

as a small secondary peak on the shoulders of the central magnetization high in the inversion solution.

The Jaramillo event is present on both ridge flanks

indicating this part of the ridge did not experience the ridge jump ,,~0.8 Ma ago documented on the profiles to the farther north and south.

Two-dimensional magnetic inversions, like those shown in Figure 5, and the close spacing of the

magnetic track lines, have allowed us to construct an

accurate isochron chart for the M A R K area (Figure 6). These anomalies show that over the past 3 Ma the

Mid-Atlantic Ridge in the M A R K area has consisted

of two distinct spreading cells separated by a small

left-stepping (dextral) transform fault or discontinuity 40-50 km south of the Kane Fracture Zone. The

northern cell has been characterized by strongly asymmetric spreading (faster to the west than to the east), as well as several small eastward ridge jumps.

The southern cell is at least 60 km long and has been

spreading more symmetrically except for a small eastward ridge jump after the Jaramillo event. The

complicated character of the central anomaly suggests that small shifts in the axis of accretion have con-

tinued during the past 700 000 yr, although the scale

of these jumps has been less than the width of the rift valley.

4. Three-Dimensional Inversion

The relatively complicated tectonic development of

the M A R K area over the last 3 Ma illustrates the

4s 4s~ ' 4 4o'

2~301

; 25o00 '

22~

4-5 20' 45 00' 44 40'

Fig. 6. Isochron chart of the MARK area based on the anomaly identifications shown in Figure 4 superimposed on Sea Beam bathymetry (contour interval 500 m). The position of the central anomaly (CA) is based on magnetization highs (see Figure 7c). Average interval spreading rates are shown in mm yr 1. A small left-lateral transform or non-transform discontinuity has existed 40-50 km south of the

Kane Fracture Zone for the past 3 Ma.

TWO- A N D THREE-DIMENSIONAL INVERSIONS OF MAGNETIC ANOMALIES IN THE M A R K AREA 49

three-dimensional nature of the crustal accretion process at slow-spreading ridges and emphasizes the need for a fully three-dimensional inversion of the magnetic field. A true three-dimensional inversion will yield maps of crustal magnetization that remove the distortions in the magnetic anomaly field caused by the skewness effects of the earth's main field and the complex bathymetric relief associated with the rift valley. The distribution of magnetization within the median valley porentially can place important con- straints on the tectonic development of this portion of the Mid-Atlantic Ridge over the past 700 000 yr.

4.1. METHOD

We have applied the three-dimensional Fourier inversion technique initially used by Macdonald et al. (1980) and described in more detail by Miller and Hey (1986). The inversion technique works in the Fourier domain and thus requires that the data be sampled at equal spacing and bordered to insure smooth replication at the boundaries. Both the Sea Beam bathymetry and observed field were gridded using a biharmonic cubic spline fitting technique. The spacing of the grid points was 1000 m in both dimensions. The grid covered an area 56 km wide and 89 km long centered on the rift valley (see Figures 7a and 7b). This is a subset of the survey area which w/Is chosen to avoid instability problems with the inversion caused by the extreme topographic relief associated with the Kane ridge-transform intersection and the very high rift mountains on the southeastern part of the area.

The inversion was performed assuming a magnetic source layer thickness of 1 km. To avoid instability problems at short wavelengths a 4 km spatial filter was applied to the bathymetry and magnetics data at each step of the inversion. A magnetization distribution, known as the annihilator, which produces zero external field was computed from the bathymetry (Parker and Huestis, 1974). Any amount of the annihilator may be added to the inversion solution without affecting the computed field. The initial inversion solution produced a geologically reasonable bal- ance between positive and negative magnetizations across the Brunhes/Matuyama boundary so no annihilator was added to the final solution.

4.2. I N V E R S I O N R E S U L T S

The magnetization distribution determined from the

inversion is shown in Figures 7c and 7d in plan view and with a three-dimensional perspective. Expanded plots of the inversion solution and the observed bathymetry in the vicinity of ODP Sites 648 and 670 in the central part of the MARK area (Detrick, Honnorez, Bryan, Juteau et al., 1988), and around the Snake Pit hydrothermal vent area at ODP Site 649 (Leg 106 Scientific Party, 1986; Detrick, Honnorez et al., 1988), are shown in Figures 8 and 9.

The calculated magnetizations range from + 17 A m- 1 to - 10 A m - i. This compares with a mean NRM intensity of 11.8A m- l for the basalts recovered from Site 648 (Detrick, Honnorez, Bryan, Juteau et al., 1988). The along-strike segmentation of the central magnetic anomaly, apparent in the observed field (Figure 7b), is also clearly seen in the inversion solution. The central anomaly consists of a series of distinct magnetization highs, separated by 20-40 km, which coalesce to form two NNE-trending bands of high magnetization, one on the northern part of the rift valley which is associated with the median volcanic ridge, and one in the southern rift valley which coincides with the string of small axial volcanos that can be traced on Sea MARC I records from 22~ N to the eastern rift valley wall near 23~ ' N (Kong et al., this volume). These two bands of high magnetization overlap near 23~ N where the lowest magnetizations in the median valley ( <4 A m -l) are found.

Several different factors could be responsible for these variations in crustal magnetization: (1) Recently extruded pillow basalts have very high magnetizations which are subsequently reduced by rapid oxidation of the outer variolitic zone of the pillows (Marshall and Cox, 1972). The along-strike variations in crustal magnetization within the median valley could thus be an indication of the relative ages of the lavas covering the valley floor. (2) The thickness of the magnetic source layer may vary reflecting differences in magmatic budget along the ridge axis. (3) Chemical variations in the compositions o f the erupting basalts due to shallow- level, crystal fractionation can affect crustal magnetizations (Vogt and Johnson, 1973). Highly magnetized, FeTi basalts are associated with high- amplitude magnetic anomalies along the Galapagos Spreading Center (Vogt and DeBoer, 1976), at ridge-transform intersections (Detrick and Lynn, 1975), at propagating rifts (Sinton et al., 1983) and

50

4aoo' 4~5d t i

N O R B E R T J. S C H U L Z E T A L .

4~0o' I

- 2 5 * 2 0 '

44~ ~

2 5 * 2 0 '

-25"10 ' 23"10'

- 25*00 ' 25 *00 '

- 2 2 " 5 0 ' 2 2 * 5 0 '

BATHYMETRY (Km)

45o00 ~ o K 4 4 50 i i

- 2 5 ~ '

FIELD (nT)

0 20 [ _ _ i

km

- 2 5 O l 0 '

- 2 3 * 0 0 '

- 2 2 * 5 0 '

MAGNETIZATION (omp/m) MAGNETIZATION

Fig. 7. Three-dimensional inversion of the magnetic anomaly field in the MARK area (Color Plate 1). (a) The Sea Beam bathymetry interpolated onto a l-km grid. Contour interval 250 m. (b) The observed magnetic anomaly field interpolated onto a 1-km grid. Contour interval 50 nT. (c) The magnetization distribution derived from the inversion solution. Contour interval 2 A m - 1. (d) A three-dimensional

view (looking north) of the inversion solution showing the two overlapping bands of high magnetization within the median valley.

at overlapping spreading centers (Sempere et al., 1984). Their occurrence in the median valley could reflect the existence of non steady-state magma chambers at these slow spreading rates. (4) Bodies of serpentinized peridotite with high magnetic susceptibilities and induced magnetizations may be present. Serpentinites can have induced magnetizations comparable to fresh basalt (Fox and Opdyke, 1973) and are known to be present in the MARK area (Karson et aL, 1987; Leg 109 Scientific Party, 1986).

In the following sections we use the calculated magnetization distribution within the central mag-

netic anomaly and available geological observations (Karson et al., 1987; Kong et al., this volume) to constrain the tectonic development of the rift valley during the past 700 000 yr.

4.2.1. Northern Ridge Segment

In the northern part of the rift valley the highest magnetizations (10--14 A m -1 ) are associated with the 600-m high, linear volcanic ridge located near the axis of the median valley (Figure 9). This ridge, the largest single volcanic constructional feature ever found within the Mid-Atlantic rift valley (Karson et al., 1987), is very young and still

TWO- AND THREE-DIMENSIONAL INVERSIONS OF MAGNETIC ANOMALIES IN THE MARK AREA

45000 , 44050 '

51

23010 '

23~

MAGNETIZATION (omp/m)

I J

0 I0 km 45000 , 44050 "

2 2~ 0 ~

2.3~ O"

2 3 0 0 0 ,

22050 ,

BATHYMETRY (Kin) Fig. 8. Detail of the bathymetry (bot tom) and magnetization distribution (top) in the vicinity of ODP Sites 648 and 670, near the center

of the M A R K area. Contour interval 1 A m-1 (magnetization), 100 m (bathymetry).

52 N O R B E R T J. S C H U L Z E T AL.

0 /

4 5 0 0 4 4 0 5 0 ,

~i:~i~,i~i~#~;'i::',;:6 : . : . : " i ........... '~i:ii~i'~;@N~ ': .li~i:i~:~ ' iiiiii~i~ii:~i!;2:ii~i~i::!::i::!:;:::;;::: �9 . . . . . . ...... ...:::.:.:.::.:.: ................

2 5" 25 '

o /

2 5 20

0 �9

2 5 15

25 ~ I 0 /

MAGNETIZATION (omplm)

I I 0 I0

km

4 5 0 0 0 , 4 4 0 5 0 ,

-./@ iiiiiiil)i ............... . . . . '" ~..V~ ~':s' ..... 6 4 9 : ....... : : "" 5::::::"

!'I.!.!.... iiiiiiiii,k.liF:..::i~i~,~,~, I~;~'~' .... 2 3 ~ 2 5 '

2 5 o 2 0 ,

23o15 ,

230 I 0 '

B A T H Y M E T R Y (Km) Fig. 9. Detail of the bathymetry (bottom) and magnetization distribution (top) in the vicinity of ODP Site 649 and the Snake Pit

hydrothermal area. Contour interval 1 A m-] (magnetization), 100 m (bathymetry).

hydrothermally active (the Snake Pit hydrothermal area is located on the crest of this ridge at Site 649). Basalts dredged from this feature are normal mid- ocean ridge basalts without significant FeTi enrich- ment (C. Langmuir, pers. comm., 1987). Thus the

high crustal magnetizations associated with this ridge

are almost certainly due to the freshness of the basalts and the large volume of lava that has been erupted to form this feature.

The high magnetizations can be traced along the


southern extension of this ridge to about 23~ where they are juxtaposed against the western rift valley wall (Figure 7c). The peak magnetizations generally decrease from north to south suggesting that the southern part of this ridge, which appears to be in the early stages of uplift into the western rift mountains, is older than the ridge to the north. This interpretation is consistent with submersible observations which indicate that south of 23~ the ridge becomes more subdued, faulted and covered by pelagic sediment (Karson et al., 1987). These relative age relationships, and the continuity of the magnetic signature, suggest that this 50-km- long volcanic ridge has formed by fissure eruptions along a single NNE-trending rift that has propa- gated from south to north within the median valley.

The Snake Pit hydrothermal area at ODP Site 649 is associated with somewhat lower average crustal magnetizations (6-8 Am -1 ) than the volcanic ridge immediately to the north or south (Fig- ure 9). The Snake Pit area is a large, active vent field consisting of more than 12 separate black- smoker chimneys and thick hydrothermal deposits (Log 106 Scientific Party, 1986; Sulanowska et al.,

1986). The fluid compositions and temperatures (up to 350~ of the vent waters strongly suggest the presence of a magma chamber at shallow depths in the crust (Edmond et al., 1986). The high temperatures associated with the presence of magma in the crust and/or extensive hydrothermal alteration of the crust could be responsible for the reduced crustal magnetizations in this area.

4.2.2. Discordant Zone

The Mid-Atlantic Ridge between 23~ and 23~ has been the boundary between two separate spreading cells for at least the past 3 Ma (Figure 6). Surprisingly, the magnetic expression of this boundary within the present median valley is not very pronounced (Figure 7c). As noted above, the high in crustal magnetization associated with the volcanic ridge in the northern median valley can be traced southward through this zone to about 23 ~ N. In this area it clearly overlaps the magnetization high associated with the southern ridge segment which extends at least as far north as 23~ , N (possibly to 23~ , N) along the eastern side of the rift valley (Figure 8). The boundary between the northern and southern spreading cells thus appears to be a rela-

tively wide (10-20 km) zone characterized by two overlapping rifts.

One unusual feature of the magnetization distribution in this area is the ESE orientation of the magnetization high near 23o10 ' N, just east of ODP Site 670 (Figure 8). This feature corresponds to a small westward shift in the central magnetic anomaly seen between profiles R.16 and R.14 (Figure 4). Its trend is approximately parallel to the Kane Fracture Zone, 40 km to the north, and nearly orthogonal to the predominantly NNE trends of topographic and structural lineations in the rift valley at this latitude (Kong et al., this issue). Karson et al. (1987) report this area is intensely fissured and faulted with no well-defined neovolcanic zone, and propose that this part of the median valley has been magma-starved during the evolution of the two adjacent ridge segments.

The high magnetization and unusual orientation of this feature are difficult to explain. Several out- crops of serpentinite were found by submersible along the western edge of the median valley in this area (Karson et al., 1987), and subsequent drilling by R/V JOIDES RESOLUTION (Leg 109 Scientific Party, 1986; Bryan, Juteau et al., 1988) penetrated about 100 m of serpentinite at Site 670, near the western edge of this zone of anomalous magnetization. The serpentinites recovered at this site have mean volume susceptibilities of 4.99 x 10 -2 and NRM intensities comparable to the fresh basalts recovered from Hole 648B (Detrick, Honnorez, Bryan, Juteau et al., 1988), so they clearly are capa- ble of explaining the magnitude of the observed anomaly. However, the extent of serpentinite within this part of the median valley is still poorly known. Alternative explanations for this anomaly include FeTi basalts, which are found at overlapping spreading centers on the East Pacific Rise (Sempere et al.,

1984), or edge effects due to the juxtaposition of rotated, crustal blocks.

4.2.3. Southern Ridge Segment

The southern ridge segment is associated with a 40 km-long, NNE-trending band of high magnetizations located along the eastern side of the median valley (Figures 7c and 7d). These high magnetizations coincide with a NNE-trending string of small (50-100 m high) axial volcanos that can be traced on Sea MARC I records from 22~ to the eastern


edge of the rift valley near 23~ (Kong et al., this volume). The peak magnetizations decrease systematically from south ( > 1 4 A m -1) to north (<8 A m -]) and the width of the band of high magnetizations narrows from more than 15 km at 22~ N to less than 5 km at 23 ~ N.

The low average crustal magnetizations determined for the along-axis topographic high near Site 648 are consistent with geological evidence that the rift valley in this area is floored by relatively old lavas (Karson et al., 1987; Kong et al., this volume). Although a 3-5 km wide high reflectivity zone in- dicative of young lavas is apparent in Sea MARC I records, the rift valley floor is cut by numerous valley-parallel faults and fissures indicating this part of the ridge is in a predominantly extensional phase (Kong et al., this volume). Karson et al. (1987) reported some relatively fresh lavas from small isolated eruptions in this area, but they concluded that a significant time has elapsed since the last major volcanic episode. The much higher magnetizations found farther south could indicate more recent volcanic activity along the southern part of this ridge segment. However, Toomey et al. (1985; 1988) have argued that the seismotectonics of the rift valley near 22~ indicates continued horizontal extension and block faulting, without significant volcanic activity, for at least the past 104yr. Unfortunately the surficial geology of this area is poorly known (Sea MARC coverage does note extend south of 22~ and no ALVIN dives were conducted in this area) so the origin of these high crustal magnetizations must remain speculative.

5. Discussion

Figure 10 presents a schematic model for the spreading history of the MARK area over the past 3 Ma based on the results presented in this paper. A small left-stepping transform was located 50 km south of the Kane Transform at anomaly 2' time, ,~ 3 Ma ago (Stage A). The offset in anomaly 2' is about 10 km which is comparable to the width of the present rift valley. Over the next 1.2 Ma, between anomalies 2' and 2, nearly symmetric spreading on the ridge segment south of this transform (13mmyr -1) and strongly asymmetric spreading on the short ridge segment to the north (17.9mmyr -] to the west; 10.6 mm yr -] to the east) combined to nearly elimi- nate this offset. However, shortly after the formation of anomaly 2, 1.7 Ma ago, a 20 km-long segment of the ridge axis to the south jumped east re-establish- ing a small offset at this latitude (Stage B). This jump in the axis of accretion left a duplicate anomaly 2 in the western rift mountains near 23~ that was modeled on profile R.20 (Figure 5c). A 10- 15 km left-lateral offset appears to have existed at this latitude until another eastward ridge jump, this time on the northern ridge segment, eliminated this offset about 0.8Ma ago leaving the duplicate JaramiUo anomaly modeled on profile R.6 (Figure 5a). After this jump the northern ridge segment appears to have consisted of three short en-echelon rifts which stepped off to the northeast (Stage C), a pattern which is preserved in the central anomaly in the northern rift valley (Figure 3). A similar eastward jump of the southern ridge segment eliminated

3Ma

f 1.7 Ma Present

I l

e I I

0.8 Ma

./ !

| E

/ I # / 7' r

// Km 0 2 0

A B D I ~

25*50'

23'00'

22~ '

Fig. 10. Model for the tectonic evolution of the MARK area over the past 3 Ma. See text for discussion.

TWO- AND THREE-DIMENSIONAL INVERSIONS OF MAGNETIC ANOMALIES IN THE MARK AREA 5 5

the small offset created near 23000 ' N after anomaly 2. The most recent event in the tectonic evolution of the MARK area has been the northward propaga- tion of a NNE-trending rift along the northern median valley which has re-established spreading in this area along a single through-going rift (Stage D). The development of this rift has been accompanied by voluminous fissure eruptions forming the median volcanic ridge and its associated hydrothermal activity. Meanwhile, the southern ridge segment has experienced only small isolated eruptions and is apparently in a predominantly extensional phase.

Two aspects of this model deserve further discussion: the nature of the discordant zone between 23005 ' N and 23~ N, and the importance of small ridge jumps on the tectonic evolution of rift valleys at slow spreading rates. Our results indicate that over the past 3 Ma the Mid-Atlantic Ridge in the MARK area has consisted of two spreading ridge segments separated by a small left-stepping transform 40- 50 km south of the Kane Fracture Zone. Regional compilations of magnetic data (Schouten et al.,

1985) show that this offset has been a characteristic feature of spreading in this area for at least the past 35 Ma, although the size of the offset has varied during this period. The disturbed topography of the western rift mountains between 23~ and 23015 ' N, and the absence of a well-developed rift valley at this latitude, probably reflect the presence of this small discontinuity. However, there is no evidence within the present median valley of structural features related to strike-slip faulting (Karson et al., 1987; Kong et al., this volume), nor is there evidence for the type of oblique spreading that often characterizes small spreading center offsets at slow slip rates (Macdonald, 1986; Searle, 1986).

Schouten and White (1980) originally used the term "zero-offset transform" to describe minor and variable offsets in sea floor magnetic lineations that separate strips of crust with distinctively different bathymetric and magnetic signatures. The discordant zone near 23~ ' N appears to be a good example of this type of non-transform offset and shares many similarities with the overlapping spreading centers recently described at the East Pacific Rise (Macdon- ald and Fox, 1983; Lonsdale, 1983). It consists of two overlapping rift zones- one coinciding with the median volcanic ridge in the northern rift valley, the other associated with a string of small axial

volcanos in the southern rift valley. Where these two rifts overlap near 23~ ' N, they are separated by an elongate basin which is several hundred meters deep (Figure 8). There is no evidence of strike-slip faulting that could accommodate the offset in a classical way and the overlap zone appears to be characterized by a relatively low magma supply. There are differences as well; for example, the rift zones do not curve into the overlap zone as the spreading centers do at the East Pacific Rise and it is not associated with a major undulation in the along- axis depth profile. However, the bathymetric and magnetic signature of this discordant zone suggest it may be a reasonable slow-spreading ridge analogue to the small ridge axis discontinuities, like overlapping spreading centers, found on the East Pacific Rise.

The tectonic evolution of the rift valley in the MARK area has been strongly influenced by a suc- cession of small ridge jumps that typically have been on the order of 10 km. Since the present rift valley is only 10-15 km wide, these jumps probably occurred within the median valley itself. Evidence for several of the larger ridge jumps that occurred near a magnetic reversal is preserved in the form of duplicate or missing anomalies like those modeled in Figure 5. However, there is no reason to suspect that jumps of a comparable magnitude have not occurred during uniform polarity intervals. The only evidence for these jumps will be a magnetic pattern of asymmetric spreading. The regional compilation of Schouten et

al. (1985) indicates that the MARK rift valley has been characterized by asymmetric spreading with faster rates to the west ( ~ 1 4 m m yr -1) than to the east ( ~ 11 mm yr- 1) for at least the past 36 Ma. The small, consistently eastward ridge jumps over the past 3 Ma documented in this study provide a plausi- ble explanation for this asymmetric spreading.

The pattern of small ridge jumps evident in the MARK area is consistent with previous suggestions that the tectonic evolution of rift valleys at slow spreading rates is controlled by distinct periods of volcanic construction superimposed on nearly con- tinuous tectonic extension (Sempere and Macdonald, 1987; Pockalny et al. 1988). During predominantly extensional phases the entire crust within the rift valley may be cooled by hydrothermal circulation so that when a new volcanic construction phase begins, rifting can occur anywhere within the 10-15 km wide


median valley. Depending on the location of this new rift a small ridge "jump" may occur. In the MARK area, as yet unknown forces result in a preference for the new rift to occur on the eastern side of the previous rift resulting in asymmetric spreading and the pattern of ridge jumps observed during the past 3 Ma. In other areas this may not occur and the spreading will be more symmetric. One of the most remarkable features of the tectonic evolution of the MARK area is that this pattern of asymmetric spreading has persisted for so long (> 36 Ma) and the relative positions and dimensions of the spreading cells have remained relatively stable throughout this entire period.

southern spreading cells is a relatively wide (10- 20km) zone near 23~ characterized by two overlapping rifts. It is not associated with normal strike-slip tectonism and may be a slow spreading ridge analogue of the overlapping spreading centers found on the East Pacific Rise. The crustal magnetizations in this overlap zone are generally low, although an anomalous ESE-trending magnetization high of unknown origin is also present in this area.

(4) The relative positions and dimensions of the spreading cells and the pattern of asymmetric spreading seen in the MARK area during the past 3 Ma have characterized this part of the Mid-Atlantic Ridge for at least the past 36 Ma.

6. Conclusions

In this paper we have described the results of two- and three-dimensional inversions of surface magnetic data from the MARK rift valley, south of the Kane Fracture Zone. The most important results to emerge from this study are:

(1) The Mid-Atlantic Ridge in the MARK area consists of two distinct spreading cells separated by a small, zero-offset transform or discordant zone between 23~ , N and 23~ ' N. This segmentation was inherited from an earlier ridge-transform-ridge geometry through a series of small ( ,-~ 10 km) eastward ridge jumps that have resulted in an overall pattern of asymmetric spreading with faster rates to the west ( 14 mm yr- 1) than to the east ( 11 mm yr- l).

(2) The youngest crust within the median valley is characterized by a series of distinct magnetization highs, separated by 20-40 km, which coalesce to form two overlapping, NNE-trending bands of high magnetization. We have interpreted this pattern in terms of two overlapping rift zones within the present median valley. The northern rift zone extends from the Kane Fracture Zone south to about 23 ~ N. This rift appears to get progressively younger from south to north and has been associated with voluminous fissure eruptions forming a large constructional volcanic ridge in the northern part of the median valley. The southern rift zone is a less well-defined band of small axial volcanos that stretch from 22~ to 23~ along the eastern side of the median valley. This rift is older and has been associated with relatively minor recent volcanism.

(3) The boundary between the northern and

Acknowledgements

This work was supported by a contract (67-84) from Joint Oceanographic Institutions Inc. to the Univer- sity of Rhode island for a predrilling site survey of the MARK area, and a post-Leg 106 USSAC science support grant. We thank the officers, crew and scientific complement of Conrad 25-11 for help collecting these data, and R. Tyce and the staff of the NECOR/ Sea Beam Development Center for the use of their computing facilities. Nancy Adams, Mike Sundvik and Rob Pockalny assisted with various aspects of this work. This manuscript was improved by thorough reviews by Richard Blakely and Vicent Courtillot.

References

Bryan, W. B., Thompson, G., and Ludden, J. N., 1981, Composi- tional Variation in Normal MORB from 22-25 ~ N; Mid- Atlantic Ridge and Kane Fracture Zone, J. Geophys. Res. 86, 11815 11836.

Bryan, W. B., Juteau, T. et al., 1988, Proc. ODP, Init. Repts. (Pt. A), 109: College Station, TX (Ocean Drilling Program), 203- 237.

Detrick, R. S. and Lynn, W. S., 1975, The Origin of High-Ampli- tude Magnetic Anomalies at the Intersection of the Juan de Fuca Ridge and Blanco Fracture Zone, Earth Planet. Sci. Letters 26, 105-113.

Detrick, R. S. and Purdy, G. M., 1980, Crustal Structure of the Kane Fracture Zone from Seismic Refraction Studies, J. Geophys. Res. 83, 3759-3777.

Detrick, R. S., Fox, P. J., Kastens, K., Ryan, W. B. F., Mayer, L., and Karson, J., 1984, A Sea Beam Survey of the Kane Fracture Zone and Adjacent Mid-Atlantic Ridge Rift Valley, EOS 65, 1006.

Detrick, R., Honnorez, J., Bryan, W. B., Juteau, T. et al., 1988, Proc. ODP, tnit. Repts. (Pt. A), 106/t09: College Station, TX (Ocean Drilling Program), 357 pp.


Detrick, R. S., Honnorez, J. et al., 1988, Proc. ODP, Init. Repts. (Pt. A), 106: College Station, TX (Ocean Drilling Program), 135-161.

Edmond, J. M., Campbei1, A. C., Palmer, M. R., and Kiinkham- mer, G. P., 1986, Preliminary Report on the Chemistry of Hydrothermal Fluids from the Mid-Atlantic Ridge, EOS 67, 1021.

Fox, P. J., 1972, The Geology of some Atlantic Fracture Zones, Caribbean Escarpments and the Nature of Oceanic Basement and Crust, Ph.D. Dissertation, Columbia University, New York.

Fox, P. J. and Opdyke, N. D., 1973, Geology of the Oceanic Crust: Magnetic Properties of Oceanic Rocks, J. Geophys. Res. 78, 5139-5154.

Karson, J. A. and Dick, H. J. B., 1983, Tectonics of Ridge-Trans- form Intersections at the Kane Fracture Zone, Mar. Geophys. Res. 6, 51-98.

Karson, J. A., Thompson, G., Hurnphris, S. E., Edmond, J. M., Bryan, W. B., Brown, J. R., Winters, A. T., Pockalny, R. A., Casey, J. F., Campbell, A. C., Klinkhammer, G., Palmer, M. R., Kinzler, R. J., and Sulanowska, M. M., 1987, Along- Axis Variations in Sea Floor Spreading in the MARK Area, Nature 328, 681~85.

Kent, D. V. and Gradstein, F. M., 1986, A Jurassic to Recent Chronology, in P. R. Vogt and B. E. Tuckholke (eds~), The Geology of North America: The Western North Atlantic Region, Geological Society of America, Boulder, CO, 45-50.

Kong, L. S. L., Detrick, R., Fox, P. J., Mayer, L. A., and Ryan, W. B. F., 1988, The Morphology and Tectonics of the MARK Area from Sea Beam and Sea MARC I Observations (Mid- Atlantic Ridge 23 ~ N), Marine Geophys. Res. 10, 59-90 (this issue).

Leg 106 Scientific Party, 1986, Drilling the Snake-Pit Hydrother- mal Sulfide deposit on the Mid-Atlantic Ridge, lat 23~ ' N, Geology 14, 1004-1007.

Leg 109 Scientific Party, 1986, Coring the Crust and the Mantle, Nature 323, 492-493.

Lonsdale, P., 1983, Overlapping Rift Zones at the 5.5 ~ Offset of the East Pacific Rise, J. Geophys. Res. 88, 9393-9406.

Louden, K. E. and Forsyth, D., 1982, Crustal Structure and Isostatic Compensation near the Kane Fracture Zone from Topography and Gravity Measurements- I Spectral Analysis Approach, Geophys. J. R. Astr. Soc. 68, 725-750.

Macdonald, K. C., 1986, The Crest of the Mid-Atlantic Ridge: Models for Crustal Generation Processes and Tectonics, in P. R. Vogt and B. E. Tuckolke (eds.), The Geology of North America: The Western North Atlantic Region, Geological Soci- ety of America, Boulder, CO., 51-68.

Macdonald, K. C. and Fox, P. J., 1983, Overlapping Spreading Centers: New Accretion Geometry on the East Pacific Rise, Nature 302, 55-57.

Macdonald, K. C., Miller, S. P., Huestis, S. P., and Spiess, F. N., 1980, Three-Dimensional Modeling of a Magnetic Reversal Boundary from Inversion of Deep-Tow Measurements, J. Geophys. Res. 85, 3670-3680.

Marshall, M. and Cox, A., 1972, Magnetic Changes in Pillow Basalts due to Sea Floor Weathering, J. Geophys. Res. 77, 6459-6469.

Mayer, L. A., Ryan, W. B. F., Detrick, R. S., Kong, L., and Manchester, K., 1985, Structure and Tectonics of the Mid- Atlantic Ridge South of the Kane Fracture Zone Based on Sea MARC I and Sea Beam Site Surveys, EOS 66, 1092.

Miller, S. P. and Hey, R. N., 1986, Three-Dimensional Modeling

of a Propagating Rift, Galapagos 95~ ' W, J. Geophys. Res. 91, 3395-3406.

Miyashiro, A., Shido, F., and Ewing, M., 1969, Composition and Origin of Serpentinites from the Mid-Atlantic Ridge, 24 ~ and 30 ~ N, Contrib. Mineral. Petrol 32, 38-52.

Miyashiro, A., Shido, F., and Ewing, M., 1970, Petrologic Models for the Mid-Atlantic Ridge, Deep-Sea Res. 17, 10~123.

Parker, R. L. and Huestis, S. P., 1974, The Inversion of Magnetic Anomalies in the Presence of Topography, J. Geophys. Res. 79, 1587 1593.

Peddie, N. W., 1982, International Geomagnetic Reference Field: The Third Generation, J. Geomag. Geoelec. 34, 309-326.

Pockalny, R. A., Detrick, R. S., and Fox, P. J., 1988, The Morphology and Tectonics of the Kane Fracture Zone from Sea Beam Bathymetry Data, J. Geophys. Res. 93, 3179-3194.

Purdy, G. M., Rabinowitz, P. D., and Schouten, H., 1978, The Mid-Atlantic Ridge at 23 ~ N: Bathymetry and Magnetics, in W.C. Melson et al., Initial Reports of the Deep Sea Drilling Project, 45, 119-128.

Purdy, G. M. and Detrick, R. S., 1986, The Crustal Structure of the Mid-Atlantic Ridge at 23 ~ N from Seismic Refraction Stud- ies, J. Geophys. Res. 91, 3739-3762.

Schouten, H. and White, R. S., 1980, Zero Offset Fracture Zones, Geology 8, 175-179.

Schouten, H., Klitgord, K. D., and Whitehead, J. A., 1985, Segmentation of Mid-Ocean Ridges, Nature 317, 225-229.

Schouten, H., Cande, S. C., and Klitgord, K. D., 1985, Magnetic Anomaly Profiles South (22~ to 28 ~ N), in Ocean Margin Drilling Program Regional Data Synthesis Series, Atlas I1, Mid-Atlantic Ridge between 22 ~ and 38~ Marine Science International, Woods Hole, MA.

Searle, R. C., 1986, GLORIA Investigations of Oceanic Fracture Zones Comparative Study of the Transform Fault Zone, J. Geol. Soc. Lon. 143, 743-756.

Sempere, J. C. and Macdonald, K. C., 1987, Marine Tectonics: Processes at Mid-Ocean Ridges, Rev. of Geophys. 25, 1313-1347.

Sempere, J. C., Macdonald, K. C., and Miller, S. P., 1984, Overlapping Spreading Centers: 3-D Inversion of the Magnetic Field at 9~ on the East Pacific Rise, Geophys. J. R. Astr. Soc. 79, 799-811.

Sinton, J. M., Wilson, D. S., Christie, D. M., Hey, R. N., and Delaney, J. R., 1983, Petrologic Consequences of Rift Propaga- tion on Oceanic Spreading Ridges, Earth Planet. Sci. Letters 62, 193-207.

Sulanowska, M., Humphris, S. E.,'Thompson, G., and Schroeder, B., 1986, Hydrothermal Mineralization in the MARK Area, Mid-Atlantic Ridge, 23 ~ N', EOS 67, 1214.

Toomey, D. R., Solomon, S. C., Purdy, G. M., and Murray, M. H., 1985, Microearthquakes beneath the Median Valley of the Mid-Atlantic Ridge near 23 ~ N: Hypocenters and Focal Mechanisms, J. Geophys. Res. 90, 5443-5458.

Toomey, D. R., Solomon, S. C., and Purdy, G. M., 1988, Micro- earthquakes beneath the Median Valley of the Mid-Atlantic Ridge near 23 ~ N: Tomography and Tectonics, J. Geophys. Res. 93, 9093-9112.

van Andel, T. J. and Bowin, C. O., 1968, Mid-Atlantic Ridge between 22 ~ and 23 ~ North Latitude and Tectonics of Mid- Ocean Rises, J. Geophys. Res. 73, 1279-1298.

Vogt, P. R. and Johnson, G. L., 1973, Magnetic Teleehemistry of Oceanic Crust?, Nature 245, 373-375.

Vogt, P. R. and De Boer, J., 1976, Morphology, Magnetic Anoma- lies and Basalt Magnetization at the Ends of the Galapagos High Amplitude Zone, Earth Planet. Sci. Letters 33, 145-163.

Two and three-dimensional inversions of magnetic anomalies in the MARK area (Mid-Atlantic Ridge 23° N

Documents