Geochemistry 3 Geophysics Geosystemsforth/people/Hirokuni/Yamazaki03.pdf · Spreading process of the northern Mariana Trough: Rifting-spreading transition at ... the northern Mariana

Spreading process of the northern Mariana Trough:Rifting-spreading transition at 22�N

Toshitsugu YamazakiInstitute for Marine Resources and Environment, Geological Survey of Japan, AIST, Higashi, Tsukuba 305-8567,Japan ([email protected])

Nobukazu SeamaResearch Center for Inland Seas, Kobe University, 1-1 Rokkodai, Nada, Kobe 657-8501, Japan

Kyoko OkinoOcean Research Institute, University of Tokyo, Minamidai, Nakano, Tokyo 164-8639, Japan

Kazuya KitadaGraduate School of Science and Technology, Kobe University, Nada, Kobe 657-8501, Japan

Masato Joshima and Hirokuni OdaInstitute for Marine Resources and Environment, Geological Survey of Japan, AIST, Higashi, Tsukuba 305-8567,Japan

Jiro NakaIFREE, Japan Marine Science and Technology Center, 2-15 Natsushima, Yokosuka 237-0061, Japan

[1] We have conducted a geophysical survey of the northern Mariana Trough from 19�N to 24�N. Thetrough evolves southward from incipient rifting to seafloor spreading within this region. This study aims to

clarify the location and time of the rifting-to-spreading transition, which was controversial previously, and

processes of seafloor spreading after the transition. The new data set includes swath bathymetry with side-

scan images and magnetic vector anomaly. The mantle Bouguer gravity anomaly (MBA) was calculated

using the free-air gravity anomaly from satellite altimetry. The rifting-to-spreading transition occurs at

about 22�N, which is proved by seafloor-spreading fabric in the bathymetry, clear magnetic lineations, and

the bull’s-eye pattern in MBA. Four ridge segments separated by three nontransform discontinuities are

recognized between 19�N and 22�N. The northernmost segment has relatively abundant magma supply

compared with the other segments, which is estimated from a larger segment length, shallower axial depths

with no rift valley, and lower MBA. The next segment to the south is, on the other hand, a magma-starved

segment with a prominent rift valley. Two anomalously deep grabens (called the Central Grabens) formed

by amagmatic extension occur near the segment ends. The succession of magma-rich, magma-starved, and

normal segments with increasing distance from the volcanic arc is the same as the observation in the Lau

Basin reported by Martinez and Taylor [2002]. The magnetic anomaly revealed the detailed history of the

spreading. The seafloor spreading between 19�N and 20�N began prior to 5 Ma, and that between 20�Nand 21�300N began at about 4 Ma. Spreading half-rates in the western side of the spreading center were 2

to 3 cm/year before 2.58 Ma south of 21�300N and during the Matuyama Chron north of 21�300N, but anaverage during the Brunhes Chron is 1 cm/year or less. Orientations of the ridge axes, which range from

�20� to 0� at present, have rotated about 20� clockwise since the start of the spreading. These changes inrate and direction might be associated with changes in the motion of the Philippine Sea plate. Spreading

has been asymmetric in the northern Mariana Trough. The spreading rates of the western side of the

G3G3GeochemistryGeophysics

Geosystems

Published by AGU and the Geochemical Society

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES

GeochemistryGeophysics

Geosystems

Article

Volume 4, Number 9

12 September 2003

1075, doi:10.1029/2002GC000492

ISSN: 1525-2027

Copyright 2003 by the American Geophysical Union 1 of 18

spreading center have been significantly larger than the eastern counterpart in general. The asymmetry may

have been caused by an interaction of mantle upwelling systems under the volcanic front and the backarc

spreading center and would be a characteristic of backarc spreading.

Components: 8375 words, 10 figures, 1 table.

Keywords: Mariana Trough; spreading; rifting; magnetic anomaly; backarc basin; tectonics.

Index Terms: 3035 Marine Geology and Geophysics: Midocean ridge processes; 3005 Marine Geology and Geophysics:

Geomagnetism (1550); 3045 Marine Geology and Geophysics: Seafloor morphology and bottom photography.

Received 10 December 2002; Revised 18 June 2003; Accepted 28 July 2003; Published 12 September 2003.

Yamazaki, T., N. Seama, K. Okino, K. Kitada, M. Joshima, H. Oda, and J. Naka, Spreading process of the northern Mariana

Trough: Rifting-spreading transition at 22�N, Geochem. Geophys. Geosyst., 4(9), 1075, doi:10.1029/2002GC000492, 2003.

1. Introduction

[2] The Mariana Trough (Figure 1) is typical of

active backarc basins in the world. It is crescent-

shaped with a maximum width of about 250 km

near 18�N, located between the Mariana Ridge

(active arc) and the West Marina Ridge (remnant

arc). The Mariana trench-arc-backarc system along

the eastern margin of the Philippine Sea plate has

recently attracted a special interest as a target for

studying the operation of the Subduction Factory,

partly because the system is relatively simple; it has

grown on an oceanic crust, and sediments covering

the subducting Pacific plate are lithologically uni-

form and subducts completely. Please refer to Fryer

[1995, 1996] and Stern et al. [2003] for thorough

reviews of the Mariana trench-arc-backarc system.

[3] The northern Mariana Trough evolves south-

ward from an incipient rifting stage to a mature

backarc-spreading stage. Thus the northern Mariana

Trough provides us a rare opportunity to study

evolution processes of backarc basins. The trough

between 22�N and 24�N is in a rifting stage, and

structurally and thermally asymmetric [Yamazaki

and Murakami, 1998]. The location of the rifting-

to-spreading transition was controversial. Yamazaki

et al. [1993] inferred from magnetic total-force

anomalies and seismic profiles that seafloor spread-

ing occurs to the south of 22�N. Conversely,

Martinez et al. [1995] and Baker et al. [1996]

proposed that the Mariana Trough north of 20�Nis composed of arc lithosphere intruded by arc and

backarc-basin magmas, and defined the volcano-

tectonic zone (VTZ) where rifting and volcanism

concentrates. A series of deep half-grabens, about

5700 m at maximum in water depth, is located near

the center of the trough between 20�N and 21�N.They were considered to have been formed by

amagmatic extension during rifting, and called the

Central Grabens [Martinez et al., 1995; Stern et al.,

1996]. Martinez et al. [1995] concluded that initi-

ation of seafloor spreading occurs to the south of

the Central Grabens at 20�N. Thus prior work on

the rifting-to-spreading transition in the northern

Mariana Trough has led to different interpretations

as to its position: either 20�N or 22�N. The main

purpose of this study, then, is to clarify the position

of this transition, its relationship to the volcanic arc,

and the evolution of the back-arc region over the

past few million years using new data.

[4] Magnetic anomalies are a basic tool for study-

ing seafloor spreading. In the Mariana Trough,

however, conventional magnetic total-force anoma-

lies are not very useful because of small anomaly

amplitudes, and detailed history of the spreading

was not clarified previously. The small amplitudes

are due to the geometrical problem of the spreading

system in theMariana Trough: approximately north-

south orientation of the spreading axis in low

magnetic latitudes. Magnetic vector measurements

can overcome this problem, and high-amplitude

signals can be obtained regardless of ridge and

geomagnetic field geometries [Isezaki, 1986]. Mag-

netic vector anomalies are hence expected to be

quite useful in the Mariana Trough. In the central

Mariana Trough, Bibee et al. [1980] succeeded to

GeochemistryGeophysicsGeosystems G3G3

yamazaki et al.: mariana trough spreading 10.1029/2002GC000492

2 of 18

correlate only a few magnetic total-force anomaly

profiles between 17�300N to 18�N, and estimated a

spreading half-rate of approximately 1.5 cm/year

within 50 km (3 Ma) of the spreading axis. Recent

magnetic vector anomaly measurements around

18�N have confirmed that seafloor spreading

started at about 6 Ma [Yamazaki and Stern,

1997; Seama and Yamazaki, 1998; Iwamoto et

al., 2002]. This is consistent with the results of

the Deep Sea Drilling Project (DSDP) Leg 60: it

was estimated that the rifting began in the latest

Miocene, ca. 6 Ma, with a spreading half-rate of

2.15 cm/year [Hussong and Uyeda, 1982].

[5] In this paper, we present swath bathymetry,

magnetic vector anomalies, and gravity anomalies

in the northern Mariana Trough from 19�N to

24�N. Previously only limited areas of the Mariana

Trough, mainly along the spreading center, were

covered with swath bathymetry. It is desirable to

survey the entire trough to clarify detailed history

of evolution because spreading in small backarc

basins is usually unstable and spreading rates and

directions frequently vary with time. We have

mapped most of the trough north of 19�N. On

the basis of the new data set, we show that the

transition from rifting to seafloor spreading occurs

near 22�N, and discuss temporal and spatial

changes of spreading in ridge segments between

19�N and 22�N.

2. Data Collection and AnalysisMethods

[6] Bathymetric and magnetic data were collected

during the KR97-11, KR98-12, KR00-03, and

KR02-01 cruises of R/V Kairei, and the Y96-13

cruise of R/V Yokosuka of the Japan Marine

Figure 1. Shaded-relief topographic map of the Mariana trench-arc-backarc system and adjacent regions. A digitalgrid of global predicted seafloor topography [Smith and Sandwell, 1997] was used. Illumination from the west.


yamazaki et al.: mariana trough spreading 10.1029/2002GC000492yamazaki et al.: mariana trough spreading 10.1029/2002GC000492

3 of 18

Science and Technology Center. Swath bathymetry

data of the R/V Kairei cruises were collected using

a SeaBeam 2112 multinarrow-beam echo sounder,

and those of the R/V Yokosuka cruise were by a

Furuno HS-10 system. A shaded-relief topographic

map of the northern Mariana Trough between

18�300N and 24�N is displayed in Figure 2. Side-

scan images (Figure 3) were also obtained using

the SeaBeam 2112 system. Post processing of the

side-scan data, which includes grazing angle cor-

rection, low-pass filtering for destripe, on-track

pixels cut, and inverse filtering for smoothing,

was carried out using the MB-system software

[Caress and Chayes, 1995].

[7] A shipboard three-component magnetometer

system aboard each vessel was used for magnetic

vector anomaly measurements, and a proton

precession magnetometer was simultaneously

towed for magnetic total-force measurements.

Figure 2. Shaded relief map of bathymetry in the northern Mariana Trough. Most part of the trough is covered withswath bathymetry. A digital grid of global predicted seafloor topography [Smith and Sandwell, 1997] was used for theareas not covered by swath bathymetry. White lines show location of ridge axis, and a white broken line indicates therifting-spreading boundary (see text). Illumination from the east. Grid interval about 370 m.



4 of 18

Please refer to Isezaki [1986] for the theory of

the onboard magnetic vector measurement and

Yamazaki and Stern [1997] and Yamazaki et al.

[1999] for the instruments and measurement

procedures on R/Vs Yokosuka and Kairei and

initial data processing. Magnetic vector anomalies

enabled us to identify magnetic lineations more

objectively than total-force anomalies; the mag-

netic polarities and locations of polarity bound-

aries were determined for each survey line from

variation patterns of both vertical and horizontal

components and the intensity of spatial differen-

tial vectors (ISDV) of Seama et al. [1993]. The

ISDV exhibits a peak at a magnetic boundary as

an example shown in Figure 4. The amplitudes of

the vertical and horizontal components are signif-

icantly larger than those of total-force anomalies

as expected.

[8] The mantle Bouguer gravity anomaly (MBA),

which is assumed to reflect variations in crustal

thickness and the temperature and density structure

of the mantle, was calculated using the free-air

gravity anomaly from satellite altimetry [Sandwell

and Smith, 1997] and the multinarrow-beam ba-

thymetry. Following the procedures of Parker

[1972] and Kuo and Forsyth [1988], we subtracted

attractions of seafloor topography and relief on the

crust-mantle interface. We assumed that the crust

has constant thickness of 6 km and that densities of

the seawater, crust, and mantle are 1030, 2700, and

3300 kg/m3, respectively. We did not correct for

Figure 3. Side-scan image of the northern Mariana Trough and interpretation of ridge segmentation. Higherbackscatter is plotted darker.



5 of 18

the effect of density anomalies associated with

lithospheric cooling.

3. Results and Interpretation3.1. Seafloor Topography and Side-ScanImage

[9] The seafloor topography shows contrasts in

depth and seafloor texture between north and south

of a boundary at 22�N near the Shoyo Seamount

(a white broken line in Figure 2). The seafloor to the

north of the boundary is much shallower than the

south of it, and characterized by fault-controlled

volcanic edifices such as the one at 22�400N,142�300E. A row of rift basins occurs along the

eastern margin of the trough, which includes the

basins northwest and southeast of the Nikko Sea-

mount at 23�N. The trough north of the boundary

consists of rifted and subsided island arc crust

[Martinez et al., 1995; Baker et al., 1996; Yamazaki

and Murakami, 1998]. To the south of the boundary,

the trough is dominated by the seafloor-spreading

fabric, narrow ridges and troughs, of nearly NNW-

SSE trend. The two deeps centered at 21�000N,143�250E and 20�050N, 144�050E are called the

Central Grabens [Martinez et al., 1995; Stern at

al., 1996]. Their water depths are about 5300 m and

5700 m at maximum, respectively. The latter is the

deepest in the Mariana Trough.

[10] The positions of ridge axes (Figure 2) are

determined from the topography, zones of high

backscatter on a side-scan image (Figure 3), and

magnetic anomalies (Figure 5). We can recognize

three significant left-stepping nontransform discon-

tinuities at about 20�350N, 19�550N, and 19�200N.The first two are clear on the side-scan image. The

last one is ambiguous on the image, but clearly

recognized by an offset in magnetic lineations.

MBAs also support the interpretation, which will

be shown later. We call Segment 1 for the ridge

between 22�100N and 20�350N, Segment 2 between

20�350N and 19�550N, Segment 3 between 19�550Nand 19�200N, and Segment 4 south 19�200N. Thesegment boundaries can be traced off axis as ridge-

normal or oblique bathymetric depressions, which

form irregular discordant zones. Characteristics of

the four segments and their boundaries are sum-

marized in Table 1.

[11] There is a possibility that the southernmost

Figure 4. An example of magnetic and topographic data along a survey line. Location of the line is indicated inFigure 5. Three components (X: north, Y: east, Z: down) and total force of magnetic anomaly (top panel), intensity ofspatial differential vector (ISDV) [Seama at al., 1993] (middle), and topography measured by the center beam(bottom) are displayed. Peaks of ISDV (arrow) indicate locations of magnetic boundaries.



6 of 18

30 km of the Segment 1, south of 20�500N, belongsto a separate segment. A zone of high backscatter

corresponding to the ridge axis does not appear on

the side-scan image between 20�500N and 20�350N;instead a series of narrow NNW-trending streak of

high backscatter occurs in a broad area between

143�250E and 143�550E (Figure 3). This suggests

that spreading may have taken place in a diffuse

manner; it may not be focused at the axis along

143�250E but occur widely in the area east of it.

This might be an undergoing eastward ridge jump.

This disorganized spreading, if it occurs, has began

recently because the topographic fabric to the west

of the axis exhibits organized spreading, and there

is no off-axis trail extending from the possible

segment boundary at 20�500N. No offset was ob-

served in the magnetic lineations either (Figure 5).

We hence consider that the discontinuity at

Figure 5. Isochrons and ridge segmentation based on magnetic anomaly interpretation. Profiles of magneticanomaly vertical (Z) component (positive downward is hatched) superimposed on the topographic relief map, andlocation of spreading center (yellow) and magnetic polarity boundaries: Brunhes-Matuyama boundary (0.78 Ma), topof the Olduvai Subchron (1.77 Ma), Matuyama-Gauss boundary (2.58 Ma), Gauss-Gilbert boundary (3.58 Ma), andtop of Chron 3n.3 (4.80 Ma). White broken lines show segment boundaries. Anomaly features annotated by purplebroken ovals are discussed in the text. Parameters used for calculation of the model profile are as follows: spreadinghalf-rate of 1.0 cm/year during the last 1.77 m.y., half-rate of 2.5 cm/year before 1.77 Ma for the western side of theaxis, ridge-axis orientation of �20�, magnetized layer thickness, depth, and intensity of 1 km, 4 km, and 5 A/m,respectively.



7 of 18

20�500N is less significant than the three major

discontinuities mentioned above, and call this por-

tion Subsegment 10.

[12] Segments 2, 3, and 4 show a typical morphol-

ogy of slow-spreading ridges like the Mid-Atlantic

Ridge [Ramberg et al., 1977; Severinghaus and

Macdonald, 1988; Sempere et al., 1990; Tucholke

and Lin, 1994, Gente et al., 1995]. The lengths of

the segments are from 60 to 75 km (Table 1). These

segments have well-developed rift valleys, and

neo-volcanic zones within the valleys are partly

observed, particularly in Segment 2 (Figure 6). The

floor of the valley is shallowest near the center of

each segment, and becomes deeper toward nodal

basins at the segment ends with increasing relief of

the rift valley. The rift valleys have asymmetric

structure: inside corners are higher elevation than

˚

˚

˚ ˚

Figure 6. Detailed topographic map of Segment 2. Topographic contours are at 100m intervals. Location of this boxis shown in Figure 2. Neovolcanic zone is recognized in the rift valley between two red arrows. IC: inside corner.

Table 1. Summary of the Features of Ridge Segmentsand Segment Boundaries in the Northern MarianaTrough

Segments

1 10 2 3 4

Segment length, km 140 30 75 65 60?Ridge-axis orientation �22�a/0�b �15� �9� �21� �17�Min. axial depth, m 2964 3936 4146 3650 3694Min. MBA, mgal �28 32 27 9 14

Segment Boundaries

1/10 10/2 2/3 3/4

Latitude, N 20�500 20�350 19�550 19�200

Max. axial depth, m 5034 4708 5038 4700Max. MBA, mgal 46 42 47 39Axial offset at 0 Ma, km 0 45 30 5Axial offset at 2.58 Ma, km 0 13 7 10

aNorth of 21�250N.

bSouth of 21�250N.



8 of 18

outside corners. For example, the western flank of

the ridge Segment 2 is shallowest near the northern

segment boundary, whereas the eastern flank is

shallowest near the southern segment boundary

(Figure 6). Water depths of the rift valley of

Segment 2, about 4100 m near the segment center,

are deeper than those of Segment 3, about 3700 m

near the center. The rift valley at the southern end

of Segment 2 (the southern Central Grabens) has

particularly large relief which exceeds 3000 m: the

water depth of the valley reaches to 5700 m,

whereas the top of the inside corner is about

2500 m (Figure 6).

[13] Segment 1, on the other hand, has no clear rift

valley except for the segment ends, which morpho-

logically resembles ridges with intermediate

spreading rates [Macdonald, 1982]. Its large seg-

ment length, about 140 km excluding Subsegment

10, is also characteristic of intermediate to fast

spreading ridges [Macdonald et al., 1991]. Water

depths of the ridge axis of this segment, about

3000 m near the segment center, are shallower than

other segments. This segment may be divided into

two by a minor discontinuity at 21�250N, which is

suggested by a change of the ridge orientation,

slightly deeper ridge axis, and an off-axis trace of

the depressions from the point. The southern part of

Segment 1 near the boundary between the Subseg-

ment 10 has a prominent rift valley (the northern

Central Graben) with a shallow inside corner, which

resembles the southern end of Segment 2.

3.2. Magnetic Vector Anomaly

[14] Interpretation of magnetic polarity boundaries

is based on all three components and the ISDV,

although only vertical component anomalies are

shown in Figure 5. The anomalies are somewhat

skewed, and thus positive vertical-component

anomaly does not exactly correspond to normal-

polarity magnetization, as shown by a forward

model in Figure 5. This is because azimuths of

the ridge axis more or less deflect westward from

the north. The magnetic anomaly patterns can be

correlated to the polarity reversal sequence from

the Brunhes Chron (the last 0.78 m.y., hereafter

based on the timescale of Cande and Kent [1995])

to the Gilbert Chron (3.58 to 5.89 Ma). Even

shorter polarity subchrons in the timescale fit the

anomaly patterns well, although only isochrons

that correspond to the major polarity boundaries

are displayed in Figure 5. For example, the positive

anomalies in the vertical component between the

isochrons of 2.58 and 3.58 Ma (the end and

beginning of the Gauss Chron) consist of two

peaks (Figure 5), which reflect the presence of

short reversed subchrons (C2An.1r and C2An.2r)

in the middle of the Gauss Chron. Further, the two

minor peaks between the isochrons of 3.58 Ma

(Gauss/Gilbert boundary) and 4.80 Ma (top of

C3n.3n) in Segment 3 would correspond to the

two normal subchrons (C3n.1n and C3n.2n) in the

Gilbert Chron. The variation patterns are identical

to those near 18�N [Yamazaki and Stern, 1997;

Iwamoto et al., 2002], where organized seafloor

spreading is believed to take place.

[15] Locations of ridge-axis discontinuities at pres-

ent and their changes in the past can be known from

offsets of the magnetic anomaly isochrons, and such

information was used for identifying the ridge seg-

ments mentioned in the previous section. No dis-

placement inmagnetic lineations is recognized at the

boundary between Segment 1 and Subsegment 10.

In the zone to the east of the ridge axis of the

Subsegment 10, magnetic lineations older than the

Brunhes Chron are obscure. This is consistent with

the idea of diffuse spreading there. The length of

Segment 3 has decreased with time; it was 110 km at

4.8 Ma but now 65 km. Magnetic lineations are

disrupted at a block in the western margin of the

trough between 19�500N and 20�200N. The block

has small volcanoes and no spreading fabric. We

estimate that it consists of rifted arc crust.

[16] The identification of magnetic lineations

revealed spatial and temporal variations of spread-

ing rates, directions, and initiation of seafloor

spreading. Seafloor spreading of Segment 3, be-

tween 19�N and 20�N, started a little prior to 5 Ma,

which was extrapolated from the oldest magnetic

isochron identified (4.8 Ma) and its distance from

the West Mariana Ridge. Segments 1 and 2 began

spreading at about 4 Ma, more than 1 m.y. after the

initiation of Segment 3. It is estimated that the two

segments between 20�00N and 21�250N started

spreading simultaneously without northward prop-



9 of 18

agation of the spreading, because the 3.58 Ma

isochron is almost parallel to the West Mariana

Ridge. In the northern part of Segment 1, north of

21�250N, the spreading started during the Gauss

Chron. The age is younger than the southern part of

this segment, and hence northward propagation is

suggested here. Recent magnetic vector measure-

ments have confirmed that the central Mariana

Trough around 18�N began seafloor spreading at

about 6 Ma [Seama and Yamazaki, 1998; Iwamoto

et al., 2002]. These observations indicate that the

start of the seafloor spreading becomes later

according as the trough narrows northward. The

northward propagation of the spreading was step-

wise, but not continuous in general.

[17] The magnetic lineations show that the spread-

ing has been asymmetric in the northern Mariana

Trough. Spreading rates of the western side of the

spreading center are much larger than the eastern

counterpart in general (Figure 7), although mag-

netic lineations of the eastern side are not as clear

as the western side partly because of volcanic

overprints such as the Kasuga Seamount Chain

(Figure 2). An exception is the spreading during

the Brunhes Chron in Segment 2, where the eastern

side is faster. The asymmetry of the spreading in

the Mariana Trough is also derived from the fact

that the location of the spreading center is not at the

center of the trough but biased toward the volcanic

arc. The asymmetry is observed also in the central

and southern part of the Mariana Trough [Iwamoto

et al., 2002; Seama et al., 2002].

[18] Spreading rates have decreased with time

(Figure 8). The western side of the spreading center

shows clearer evidence for the spreading rate

change as follows. The half-rate was 2 to 3 cm/

year in average before 2.58 Ma (Gauss/Matuyama

boundary) south of 21�300N. In the northern part of

Segment 1, north of 21�300N, where the initiation

of spreading was later, the half-rate of 2 to 3 cm/

year continued during the Matuyama Chron. These

rates were reduced to be about 1 cm/year or less

during the Brunhes Chron. An exception is part of

Segment 1 between 21�150N and 21�450N: the rateduring the Brunhes Chron is a little higher, 1.5 to

2.0 cm/year. This part corresponds to the ridge

without axial rift valley. Changes in spreading half-

rate in the eastern side of the spreading center are

not clear, but the rates are slower than the western

side in general as mentioned above, and there is no

evidence for any increase associated with the

decrease in the western side. This implies a de-

crease of spreading full-rate.

[19] Spreading directions can be estimated from

strikes of magnetic anomaly isochrons. The trends

of the isochrons agree with those of topographic

Figure 7. Comparison of average spreading ratesbetween the eastern (open squares) and western (solid)sides of the spreading center based on the identificationof magnetic lineations in Figure 5.



10 of 18

fabric in general (Figure 5). Spreading directions

have changed with age. In each segment, this can be

recognized visually from a difference in trend

between the oldest isochron and the present

spreading center in Figure 5. Orientations of the

spreading center, which range from �20� to 0� atpresent (Table 1), have rotated about 20� clockwisesince the start of the spreading (Figure 8). This

clockwise rotation resulted in increases of ridge-

axis offset with time at the segment boundaries. The

magnetic isochrons indicate that the offset at the

boundary between Subsegment 10 and Segment 2

was about 13 km at 2.58 Ma but is about 45 km now

(Table 1). The offset at the boundary between

Segments 2 and 3 increased from 7 to 30 km within

the same period of time.

[20] No magnetic lineation of seafloor-spreading

origin is observed north of the boundary at 22�N(Figure 5). The magnetic anomaly profiles there are

correlative neither with each other nor with the

magnetic polarity reversal sequence back from the

Brunhes Chron. These observation is consistent

with the idea that the trough north of 22�N consists

of rifted arc crust [Yamazaki et al., 1993; Martinez

et al., 1995; Yamazaki and Murakami, 1998].

Dominant positive anomalies in the vertical com-

ponent north of 22�N (Figure 5) are associated with

volcanoes in general: for example, a row from

22�300N, 142�350E to 22�45�N, 142�250E (eastern

purple broken oval in Figure 5). A pair of positive

and negative anomaly bands in the western margin

of the trough between 22�250N and 22�450N (west-

ern purple broken oval in Figure 5) is probably

caused by an intrusive body.

3.3. Mantle-Bouguer Gravity Anomaly

[21] By a first look of the MBA map (Figure 9), we

can recognize a contrast between lows along the

Mariana and West Mariana Ridges and highs in the

trough south of 22�N. Steep gradients in MBA

Figure 8. Orientation of magnetic polarity boundaries (left), and average spreading half-rates between polarityboundaries (right) along the western side of spreading center.



11 of 18

occur along the topographic and magnetic bound-

ary near 22�N, as well as in the northernmost part

of the trough near 23�N and along flanks of the

Mariana and West Mariana Ridges facing the

trough. This indicates a large contrast in crustal

thickness at the boundary near 22�N: rather abruptthinning southward. This supports the idea that the

rifting-to-spreading transition takes place here.

[22] South of 22�N, the MBA is lowest near the

center of each segment, and increases toward the

segment ends (Figures 9 and 10). It is anti-correlated

with the water depth of the spreading center

(Figure 10). This variation pattern, the so-called

‘‘bull’s-eye,’’ is known to occur commonly along

slow-spreading ridges like the Mid-Atlantic Ridge

(MAR) [Kuo and Forsyth, 1988; Lin et al., 1990],

and is interpreted to represent thicker crust near the

segment center than at the ends caused by focused

magmatic accretion. The ridge axis of Segment 1 has

lower MBAs and shallower water depths than the

other segments. The center of Segment 1 is located

on the slope ofMBAdecreasing toward theMariana

Ridge, but not on an enclosed low like the other

segments. A drop in MBA with a corresponding

peak of axial depth at about 20�450N (Figure 10)

suggests that the southern part of Segment 1

between 20�500N and 20�300N probably belong

Figure 9. Mantle Bouguer gravity anomaly map calculated using the Free-air gravity anomaly from satellitealtimetry. Contours are at 10 mgal intervals. Thick lines show location of ridge axis. Grid interval about 3700 m.



12 of 18

to a separate short segment (Subsegment 10), as

inferred from the topography and side-scan image

mentioned above. Segments 2, 3, and 4 have axial

MBA amplitudes of 20 to 30 mgal and axial

depth relief of about 1000 m (Figure 10 and

Table 1), which agrees with the relationship

between the two observed along the MAR be-

tween 15�N and 40�N [Thibaud et al., 1998].

[23] The low MBAs along the West Mariana Ridge

protrude into the trough between 19�500N to

20�200N, which corresponds to the remnant arc

crust estimated from the topography and magnetic

lineations. High MBAs are observed along the

western margin of the trough between 21�N and

22�N. Similar tendency, although weak, is recog-

nized south of 19�500N. This could represent the

thermal effect of lithospheric cooling. On the other

hand, such highs in MBA are absent in the eastern

margin of the trough, probably because of a ther-

mal effect of the hot volcanic arc.

4. Discussion

[24] We have presented the occurrence of spreading

fabric in the swath bathymetry, clear lineations in

the magnetic vector anomalies, and the bull’s-eye

pattern in the MBAs associated with ridge segmen-

tation to the south of the boundary near 22�N. We

thus conclude that seafloor spreading has already

started south of 22�N. We consider that it is not

necessary to define a rifting-to-spreading transition

zone in the south of 22�N like the Volcano-Tectonic

Zone (VTZ) of Martinez et al. [1995]. In contrast,

such spreading features are not observed north of

the boundary, which indicates that the trough north

of 22�N is in a rifting stage. The VTZ model of

focused rifting and volcanism would be applicable

here, which can explain the occurrence of the

magnetic anomaly bands in the western margin of

the trough between 22�250N and 22�45’N.

[25] A major reason for the conclusion of Martinez

et al. [1995] that the trough between 20�N and

22�N is underlain by rifted arc crust was that

broad, positive and negative magnetization bands

were commonly observed north and south of 22�Non their results of a three-dimensional inversion of

magnetic total-force anomaly data. A possible

cause of the different conclusion from ours is that

fine features, which are useful for the identification

of magnetic lineations, seem to have been reduced

Figure 10. Along-axis variations of mantle Bouguer gravity anomaly (MBA) and depth in the northern MarianaTrough between 19�N and 22�N.



13 of 18

in their inversion solution, because they applied a

band-pass filter which suppresses wavelengths less

than 12 km. Short-wavelength signals that corre-

spond to the Olduvai Subchron and two subchrons

within the Gauss Chron may have been affected by

the filtering. It can also be pointed out that our

results from the magnetic vector anomalies would

have higher resolution than theirs on the basis of

the total-force anomalies. The latter has inevitably

small anomaly amplitudes in the Mariana Trough

as mentioned before. Moreover, their inversion

assumption may not be valid in some points.

They assumed the source layer thickness of 1 km

for their inversion. This assumption is commonly

used for magnetic anomaly inversions of seafloor-

spreading origin [e.g., Macdonald et al., 1980;

Grindlay et al., 1992; Pariso et al., 1996], but

may not be applicable for arc crust north of 22�N.The upper part of arc crust would consist of thick

volcaniclastics, which contributes little to the

magnetic anomalies. Instead, middle and lower

crusts may have strong magnetization associated

with intrusion and underplating of magmas. Their

inversion includes the Mariana and West Mariana

Ridges, where deep-seated plutonic bodies may

contribute significantly to the magnetic anomalies.

The importance of such anomaly sources beneath

oceanic island arcs is suggested by the occurrence

of long-wavelength magnetic anomalies over the

Izu-Ogasawara (Bonin) Arc and the Kyushu-Palau

Ridge [Yamazaki and Yuasa, 1998]. Thus the

inversion which included altogether the Mariana

and West Mariana Ridges and the trough from

20�N to 24�N but assumed the constant 1 km

source-layer may not be valid; they have different

crustal structures and hence should have different

source-layer depths and thicknesses. Furthermore,

the assumption is not consistent with their model.

If the trough south of 22�N is underlain by rifted

arc crust with volcanic intrusions as they postu-

lated, the sources of the anomalies could be

deeper and thicker than surface 1 km, in particular

in the western part of the trough where the crust

should be cooler and have experienced relatively

small extension.

[26] The gravity data are not consistent with the

continuation of rifting southward beyond the

boundary near 22�N. A large change in MBA

occurs only between 22�N and 22�200N, and no

tendency indicating progressive crustal thinning is

observed south of it (Figure 9). A model calcula-

tion by Ishihara and Yamazaki [1991] using ship-

board free-air gravity data suggested that the

crustal thickness decreases from about 15 km near

24�N to about 6 km south of 22�N where water

depth is 4500 m. The thickness of the latter is

typical of oceanic crust. If the trough south of 22�Nconsists of rifted arc crust and the thinning is

localized along the VTZ with focused extension

[Martinez et al., 1995], the crust should be thinnest

along the VTZ and become thicker toward the

margins of the trough. A low in MBA is hence

expected for the western margin of the trough from

this model, but this is contrary to the observation.

[27] The two deeps near the center of the trough

between 20�N and 21�N, the Central Grabens,

have been formed by amagmatic extension at the

segment ends. The southern one near 20�N occurs

at the southern end of Segment 2, and the northern

one near 21�N is located at the southern end of

Segment 1 if we introduce Subsegment 10. They

were previously thought to have been formed by

concentration of rifting [Martinez et al., 1995] or

be enigmatic [Stern et al., 1996], but can well be

explained in the context of seafloor spreading. The

topographic asymmetry of the deeps, the prominent

eastern scarps vs. the gentle western slopes, can be

explained by tectonic thinning of the inside corners

by low-angle detachment faults [Tucholke and Lin,

1994; Tucholke et al., 1998]. Gabbros and ultra-

mafic rocks were recovered from the deep at 20�N[Stern et al., 1996, 1998; Ohara et al., 2002].

Exposure of lower crust and upper mantle at inside

corners is common in the MAR [Tucholke and Lin,

1994; Cannat et al., 1995; Gracia et al., 1997].

The great depths of the Central Grabens were

emphasized previously and thought to be anoma-

lous compared with the MAR, where maximum

depths of some second-order segment boundaries

are close to 5000 m. But the maximum depth of

5700 m of the Central Grabens are not quite

anomalous, considering that depths of backarc

basins in the Philippine Sea plate including the

Mariana Trough is about 700 m deeper in average



14 of 18

than expected from the age-depth curve of major

oceans [Park et al., 1990].

[28] Segment 1 shows the characteristics of the

ridges with intermediate spreading rates, those are

the larger segment length and the shallower depths

of the ridge axis without rift valley, even though

the spreading rate during the Brunhes Chron is

only slightly larger than the southern segments.

Furthermore, this segment has much lower MBAs

than the other segments in the bull’s-eye pattern,

suggesting thicker crust and/or hotter mantle. Since

this segment is closest to the volcanic arc and the

MBA low of the segment center continues to the

Mariana Ridge, these observations lead us an

interpretation that this segment receives extra

magma delivery from the volcanic front. Similar

changes of topography and gravity are observed in

segments of the MAR approaching to the Azores

hot spot [Detrick et al., 1995; Thibaud et al., 1998]

and the Iceland hot spot [Bell and Buck, 1992;

Searle et al., 1998]. On the basis of chemical

compositions of basalts obtained from the northern

part of the trough between 21�N and 22�N, Stern etal. [1990] estimated a mixing of 50–90% Mariana

arc component with a mid-ocean ridge basalt

(MORB) component. Subsegment 10 and Segment 2

are, on the contrary, cold, magma-starved seg-

ments. The centers of the segments are deeper than

Segment 3 as well as Segment 1, and they accom-

pany the deep grabens near the segment ends.

Segments 3 and 4 show normal morphological

features for slow-spreading ridges. The succession

of magma-rich, magma-starved, and normal seg-

ments with increasing distances from the volcanic

arc is the same as the observation in the Lau Basin,

from the Valu Fa Ridge to the Central Lau spread-

ing center, as proposed by Martinez and Taylor

[2002]. Presence of the volcanic arc would be a

fundamental difference of backarc spreading from

mid-ocean ridges. The asymmetry in spreading

rates also suggests that an interaction between

mantle upwellings under the backarc spreading

center and the volcanic front may control the

distance between the two.

[29] Seafloor spreading in small backarc basins is

considered to be unstable in general compared

with mid-oceanic ridges in major oceans because

the former may be affected by kinematics of

surrounding plates more easily, and changes in

spreading direction and rate may be frequent.

Recent GPS measurements revealed that both

the Philippine Sea plate and the Mariana Ridge

are moving westward relative to the Eurasia

[Kato et al., 1998; Sella et al., 2002], and the

opening of the Mariana Trough occurs because

the movement of the former is faster than the

latter. This observation is not consistent with

the opening model assuming a trench rollback

[Martinez and Fryer, 2000]. Spreading of

the Mariana Trough predicted from the GPS

measurements is almost east-west with a full-rate

of about 4.7 cm/year, which was measured in its

southern part (Guam). It is not inconsistent with

those in the northern Mariana Trough estimated

from the magnetic anomalies of Brunhes age,

considering that the rates would decrease north-

ward with decreasing width of the trough. The

reducing spreading rate and the clockwise rota-

tion of ridge-axis strike observed in the northern

Mariana Trough during the last ca. 1 to 2 m.y.

should reflect a change in the motion of the

Philippine Sea plate. Some evidences suggest

that the convergence along the Nankai and

Sagami Trough subduction zones in the northern

margin of the Philippine Sea plate changed from

the north to northwest at ca. 1 Ma [Nakamura et

al., 1984; Yamazaki and Okamura, 1989]. It is,

however, difficult to construct a plausible model

of the past Philippine Sea plate motion only from

these constraints, and further accumulation of

data concerning a detailed spreading history in

the central and southern part of the Mariana

Trough, internal deformation of the Mariana

arc, and convergence along margins of the Phil-

ippine Sea plate are required.

5. Conclusions

[30] This study revealed following characteristics

of the spreading in the northern Mariana Trough

between 19�N and 24�N.

[31] (1) Transition from rifting to seafloor spread-

ing occurs near 22�N. Four ridge segments sepa-



15 of 18

rated by three nontransform discontinuities are

identified between 19�N and 22�N. Seafloor

spreading between 19�N and 20�N began prior

to 5 Ma, and that between 20�N and 21�300Nbegan at about 4 Ma. The northward propagation

of the spreading was stepwise, but not continuous.

The bull’s-eye pattern of MBA, which is a char-

acteristic of slow-spreading ridges, occurs along

the ridge.

[32] (2) The northernmost segment is morpholog-

ically similar to ridges with intermediate spreading

rates: a large segment length and shallow water

depths without rift valley. The MBAs near the

segment center are lower than the other segments.

This suggests an abundant magma supply to this

segment. The next segment is, on the contrary, a

magma-starved segment with a prominent rift val-

ley. Two anomalously deep grabens (the Central

Grabens) occur near the segment ends between

20�N and 21�N, and were formed by amagmatic

extension. The succession of magma-rich, magma-

starved, and normal segments with increasing dis-

tances from the volcanic arc is the same as the

observation in the Lau Basin reported by Martinez

and Taylor [2002].

[33] (3) Spreading rate has decreased with time.

The spreading half-rates in the western side of

the spreading center were 2 to 3 cm/year before

2.58Ma south of 21�300N and during theMatuyama

Chron north of 21�300N, but an average during the

Brunhes Chron is 1 cm/year or less. Direction of

spreading has also changed with time. Orientations

of spreading ridges, which range from�20� to 0� atpresent, have rotated about 20� clockwise since

the start of the spreading. These changes might

be associated with changes in the motion of the

Philippine Sea plate.

[34] (4) The seafloor spreading has been asymmet-

ric. The spreading rates of the western side of the

spreading center have been significantly larger than

the eastern counterpart in general. We estimate that

diffuse spreading occurs in the eastern side of the

axis between 20�350N and 20�500N, which may be

an undergoing eastward ridge jump. The asymme-

try may have been caused by an interaction of

upwelling systems under the volcanic front and the

backarc spreading center, and would be a charac-

teristic of backarc spreading.

Acknowledgments

[35] We thank Naoki Wakabayashi, Yuichi Hasegawa, and

Hisanori Iwamoto for their help of the measurements and data

processing, and Robert J. Stern, Makoto Arima, Makoto

Yuasa, and Takemi Ishihara for discussion. We also thank all

onboard scientists, officers and crew of the cruises of R/V

Kairei and R/V Yokosuka for cooperation, and Hajimu

Kinoshita, Kiyoshi Suyehiro, and other JAMSTEC personnel

related to the cruises for their support. The manuscript was

greatly improved by comments from two anonymous

reviewers and an associate editor, Bruce Anderson. The

GMT software [Wessel and Smith, 1995] was used for pro-

ducing the figures. This study was partly supported by the

Grant-in-Aid for Scientific Research ((B)(2) No. 12440116) of

the Japan Society for the Promotion of Science.

References

Baker, N., P. Fryer, F. Martinez, and T. Yamazaki, Rifting

history of the northern Mariana Trough: SeaMARC II and

seismic reflection surveys, J. Geophys. Res., 101, 11,427–

11,455, 1996.

Bell, R. E., and W. R. Buck, Crustal control of ridge segmen-

tation inferred from observations of the Reykjanes Ridge,

Nature, 357, 583–586, 1992.

Bibee, L. D., G. G. Shor Jr., and R. S. Lu, Intra-arc spreading

in the Mariana Trough, Mar. Geol., 35, 183–197, 1980.

Cande, S. C., and D. V. Kent, Revised calibration of the geo-

magnetic polarity timescale for the Late Cretaceous and Cen-

ozoic, J. Geophys. Res., 100, 6093–6095, 1995.

Cannat, M., et al., Thin crust, ultramafic exposures, and rugged

faulting patterns at the Mid-Atlantic Ridge (22�–24�N),Geology, 23, 49–52, 1995.

Caress, D. W., and D. N. Chayes, New software for processing

sidescan data from sidescan-capable multibeam sonars, paper

presented at IEEE Oceans 95 Conference, Inst. of Electr. and

Electron. Eng., San Diego, Calif., 1995.

Detrick, R. S., H. D. Needham, and V. Renard, Gravity anoma-

lies and crustal thickness variations along the Mid-Atlantic

Ridge between 33�N and 40�N, J. Geophys. Res., 100,

3767–3787, 1995.

Fryer, P., Geology of the Mariana Trough, in Backarc Basins:

Tectonics and Magmatism, edited by B. Taylor, pp. 237–

279, Plenum, New York, 1995.

Fryer, P., Evolution of the Mariana convergent plate margin

system, Rev. Geophys., 34, 89–125, 1996.

Gente, P., R. A. Pockalny, C. Durand, C. Deplus, M. Maia,

G. Ceuleneer, C. Mevel, M. Cannat, and C. Laverne,

Characteristics and evolution of the segmentation of the

Mid-Atlantic Ridge between 20�N and 24�N during the last

10 million years, Earth Planet. Sci. Lett., 129, 55–71, 1995.



16 of 18

Gracia, E., D. Bideau, R. Hekinian, Y. Lagabrielle, and L. M.

Parson, Along-axis magmatic oscillations and exposure of

ultramafic rocks in a second-order segment of the Mid-

Atlantic Ridge (33�430N to 34�070N), Geology, 25, 1059–1062, 1997.

Grindlay, N. R., P. J. Fox, and P. R. Vogt, Morphology and

tectonics of the Mid-Atlantic Ridge (25�–27�300S) from Sea

Beam and magnetic data, J. Geophys. Res., 97, 6983–7010,

1992.

Hussong, D. M., and S. Uyeda, Tectonic processes and the

history of the Mariana Arc: A synthesis of the results of

deep-sea drilling project Leg 60, Initial Rep. Deep Sea Drill.

Proj., 60, 909–929, 1982.

Isezaki, N., A new shipboard three-component magnetometer,

Geophysics, 51, 1992–1998, 1986.

Ishihara, T., and T. Yamazaki, Gravity anomalies over the Izu-

Ogasawara (Bonin) and northern Mariana Arcs, Bull. Geol.

Surv. Jpn., 42, 687–701, 1991.

Iwamoto, H., M. Yamamoto, N. Seama, K. Kitada, T. Matsuno,

Y. Nogi, T. Goto, T. Fujiwara, K. Suyehiro, and T. Yamazaki,

Tectonic evolution of the central Mariana Trough, Eos

Trans. AGU, 83(47), Fall Meet. Suppl., abstract T72A-1235,

2002.

Kato, T., et al., Initial results from WING, the continuous GPS

network in the western Pacific area, Geophys. Res. Lett., 25,

369–372, 1998.

Kuo, B.-Y., and D. W. Forsyth, Gravity anomalies of the ridge-

transform system in the South Atlantic between 31� and 34.

5�S: Upwelling centers and variations in crustal thickness,

Mar. Geophys. Res., 10, 205–232, 1988.

Lin, J., G. M. Purdy, H. Schouten, J.-C. Sempere, and

C. Zervas, Evidence from gravity data for focused magmatic

accretion along the Mid-Atlantic Ridge, Nature, 344, 627–

632, 1990.

Macdonald, K. C., Mid-ocean ridges: Fine scale tectonic, vol-

canic and hydrothermal processes within the plate boundary

zone, Annu. Rev. Earth Planet. Sci., 10, 155–190, 1982.

Macdonald, K. C., S. P. Miller, S. P. Huestis, and F. N. Spiess,

Three-dimensional modeling of a magnetic reversal bound-

ary from inversion of deep-tow measurements, J. Geophys.

Res., 85, 3670–3680, 1980.

Macdonald, K. C., D. S. Scheirer, and S. M. Carbotte, Mid-

ocean ridges: Discontinuities, segments and giant cracks,

Science, 253, 986–994, 1991.

Martinez, F., and P. Fryer, Geophysical characteristics of the

Mariana Trough, 11�500N–13�400N, J. Geophys. Res., 105,16,591–16,607, 2000.

Martinez, F., and B. Taylor, Mantle wedge control on back-arc

crust accretion, Nature, 416, 417–420, 2002.

Martinez, F., P. Fryer, N. A. Baker, and T. Yamazaki, Evolution

of backarc rifting: Mariana Trough, 20�–24�N, J. Geophys.Res., 100, 3807–3827, 1995.

Nakamura, K., K. Shimazaki, and N. Yonekura, Subduction,

bending and eduction: Present and Quaternary tectonics of

the northern border of the Philippine Sea plate, Bull. Soc.

Geol. France, 26, 221–243, 1984.

Ohara, Y., R. J. Stern, T. Ishii, H. Yurimoto, and T. Yamazaki,

Peridotites from the Mariana Trough: First look at the mantle

beneath an active back-arc basin, Contrib. Mineral. Petrol.,

143, 1–18, 2002.

Pariso, J. E., C. Rommevaux, and J.-C. Sempere, Three-di-

mensional inversion of marine magnetic anomalies: Implica-

tions for crustal accretion along the Mid-Atlantic Ridge

(28�–31�300N), Mar. Geophys. Res., 18, 85–101, 1996.

Park, C.-H., K. Tamaki, and K. Kobayashi, Age-depth correla-

tion of the Philippine Sea back-arc basins and other marginal

basins in the world, Tectonophysics, 181, 351–371, 1990.

Parker, R. L., The rapid calculation of potential anomalies,

Geophys. J. Astron. Soc., 31, 447–455, 1972.

Ramberg, I. B., D. E. Gray, and R. G. H. Raynolds, Tectonic

evolution of the FAMOUS area of the Mid-Atlantic Ridge,

lat 35�500 to 37�200N, Geol. Soc. Am. Bull., 88, 609–620,1977.

Sandwell, D. T., and W. H. F. Smith, Marine gravity anomaly

from Geosat and ERS 1 satellite altimetry, J. Geophys. Res.,

102, 10,039–10,054, 1997.

Seama, N., and T. Yamazaki, Asymmetric sea-floor spreading

in the central Mariana Trough, Eos Trans. AGU, 79(45), Fall

Meet. Suppl., F908, 1998.

Seama, N., Y. Nogi, and N. Isezaki, A new method for precise

determination of the position and strike of magnetic bound-

aries using vector data of the geomagnetic anomaly field,

Geophys. J. Int., 113, 155–164, 1993.

Seama, N., T. Yamazaki, H. Iwamoto, K. Kitada, M. Yamamoto,

T. Fujiwara, Y. Nogi, K. Okino, and K. Suyehiro, Tectonic

features of the Mariana Trough, Eos Trans. AGU, 83(47), Fall

Meet. Suppl., abstract T71F-08, 2002.

Searle, R. C., J. A. Keeton, R. B. Owens, R. S. White,

R. Mecklenburgh, B. Parsons, and S. M. Lee, The Reykjanes

Ridge: Structure and tectonics of a hot-spot-influenced,

slow-spreading ridge, from multibeam bathymetry, gravity

and magnetic inversions, Earth Planet. Sci. Lett., 160,

463–478, 1998.

Sella, G. F., T. H. Dixon, and A. Mao, REVEL: A model for

recent plate velocities from space geodesy, J. Geophys. Res.,

107(B4), 2081, doi:10.1029/2000JB000033, 2002.

Sempere, J.-C., G. M. Purdy, and H. Schouten, Segmentation

of the Mid-Atlantic Ridge between 24�N and 30�400N,Nature, 344, 427–431, 1990.

Severinghaus, J. P., and K. C. Macdonald, High inside corners

at ridge-transform intersections,Mar. Geophys. Res., 9, 353–

367, 1988.

Smith, W. H. F., and D. T. Sandwell, Global sea floor topo-

graphy from satellite altimetry and ship depth soundings,

Science, 277, 1956–1962, 1997.

Stern, R. J., P.-N. Lin, J. D. Morris, M. C. Jackson, P. Fryer,

S. Bloomer, and E. Ito, Enriched back-arc basin basalts from

the northern Mariana Trough: Implications for the magmatic

evolution of back-arc basins, Earth Planet. Sci. Lett., 100,

210–225, 1990.

Stern, R. J., S. H. Bloomer, F. Martinez, T. Yamazaki, and

T. M. Harrison, The composition of back-arc basin lower

crust and upper mantle in the Mariana Trough: A first report,

Island Arc, 5, 354–372, 1996.

Stern, R. J., T. Yamazaki, S. Danishwar, and C.-H. Sun, Back-

arc basin lower crust and upper mantle in the northern



17 of 18

Mariana Trough studied with ‘‘Shinkai 6500,’’ JAMSTEC

J. Deep Sea Res., 13, 47–61, 1998.

Stern, R. J., M. J. Fouch, and S. L. Klemperer, An overview

of the Izu-Bonin-Mariana Subduction Factory, in Inside

the Subduction Factory, Geophys. Monogr. Ser., edited by

J. Eiler, in press, AGU, Washington, D. C., 2003.

Thibaud, R., P. Gente, and M. Maia, A systematic analysis of

the Mid-Atlantic Ridge morphology and gravity between

15�N and 40�N: Constraints of thermal structure, J. Geo-

phys. Res., 103, 24,223–24,243, 1998.

Tucholke, B. E., and J. Lin, A geological model for the struc-

ture of ridge segments in slow spreading ocean crust,

J. Geophys. Res., 99, 11,937–11,958, 1994.

Tucholke, B. E., J. Lin, and M. C. Kleinrock, Megamullions

and mullion structure defining oceanic metamorphic core

complexes on the Mid-Atlantic Ridge, J. Geophys. Res.,

103, 9857–9866, 1998.

Wessel, P., and W. H. F. Smith, New version of the generic

mapping tools released, Eos Trans. AGU, 76(33), 329, 1995.

Yamazaki, T., and F. Murakami, Asymmetric rifting of the

northern Mariana Trough, Island Arc, 7, 460–470, 1998.

Yamazaki, T., and Y. Okamura, Subducting seamounts and

deformation of overriding forearc wedges around Japan, Tec-

tonophysics, 160, 207–229, 1989.

Yamazaki, T., and R. J. Stern, Topography and magnetic vector

anomalies in the Mariana Trough, JAMSTEC J. Deep Sea

Res., 13, 31–45, 1997.

Yamazaki, T., and M. Yuasa, Possible Miocene rifting of the

Izu-Ogasawara (Bonin) arc deduced from magnetic anoma-

lies, Island Arc, 7, 374–382, 1998.

Yamazaki, T., F. Murakami, and E. Saito, Mode of seafloor

spreading in the northern Mariana Trough, Tectonophysics,

221, 207–222, 1993.

Yamazaki, T., K. Okino, Y. Hasegawa, H. Saitake, and

M. Ito, Geophysical mapping of Mariana Trough and West

Philippine Basin: A preliminary report of Kairei KR9812

cruise, JAMSTEC J. Deep Sea Res., Part II, 15, 63–72,

1999.



18 of 18

Documents

Geochemistry 3 Geophysics Geosystemsforth/people/Hirokuni/Yamazaki03.pdf · Spreading process of the northern Mariana Trough: Rifting-spreading transition at ... the northern Mariana