OAMSTEC ≫ffi≪K 55 15^

Bathymetry, Geomagnetic and Gravity Anomalies of

the Mid-Atlantic Ridge between 14°N and 16°N

Toshiya FUJIWARA *1 Takeshi MATSUMOTO *1 Peter B. KELEMEN *2

Masato JOSH1MA * 3 John F. CASEY *4 Akira TAKEUCHI'5

Georges M. CEULENEER * 6 Michael G. BRAUN * 2 Satoru KANDA *'

The Mid-Atlantic Ridge between J d*N and 168N is supposed to be in a magma-starved portion of a slow spreading ridge, where igneous crust is

virtually absent And roeks of the lower crust and upper mantle arc disuibutcd on the seafloorovcrabout 100 km along the strike. The melting process,

melt migration mechanisms, crystal deformation mechanisms during the magm*- ≪tar>ed tccionie scafloor spreading' these are fundamental questions

for the understanding of global mid*oeean ridge systems. A geophysical survey was conducted In June and July 1993 aboard the fc/V Yokosuka to

characterize geophysieally the on and olT*axis ridge flanks. 1 he survey iracks covered a distance from 60 km north to 1^0 km south ofibc l$*20'N

Kracture Zone. The tracks reached lo 70 km off the ridge axis on both sides, the crusial age of which is about 5 m y old. Low amplitude geomagnetic

anomalies and magneu?ation variation, compared with more magmatically robust parts of the Mid-Atlantic Rtdge. were observed. This is consistent

with the exposure of peridoine due lo limited basakic magma supply over a period of several million years. Low amplitude "butt's eye" mantle

Bouguer anomalies were observed on ihe ridge segments Hanking the active transform fault This is probably due to a limited magma supply, thus mis

region has not generated thick igneous oceanic enst. The gravity data support ihc conclusion that the magina≪stajvcd naiure of the region is a ion-

lived phenomenon. Small ampliiuda! variations of mantle Bouguer anomalies across ihe ridge axis may suggest that the density structure or cmstal

thickness vajies with time over about 2 m y. <Jws the tectonic activity may be cyclic The largest mamlc Bouguer gravity anomalks are lows along ihc

ridge axis at the northernmost and southernmost limits of our survey area. The simplest interpretation of these data seems to be that the Urge gravity

iows represent the centers of unusually targe magmatie segments at approximately 14°N and 16

°^. probably as a result of 3-dimenskmal focusing of

melt gctteraiion processes m the mantle.

Keywords: Mid ≫Ai1wtic Ridge. I5°20'N fracture Zot＼c. Binhymci/y.Or ≫vnyAoom*ly.O ≪om≪|nciic Anomaly

t l tocp ≪Sca Research Department Japan Marine Sviciicc and Technology Cooler

* 2 Department ofCcotogy and Geophysics. Woods I lotc Oceanographic Institution

* 3 Marine Geology Department. Geological Survey of J apan

* 1 Dcpanmcm ofOcoscicnces University of Houston

* 5 IVpamriern i>fl:;mhS<ieiWC ≫ loy.-ima University

* 6 ONRS OMI*. lou1ou≪

* ? Nippon Marine Hiicrprisv-s. Co I lil

IAMSTET I 1T,-Ittkik

13

1. Introduction

The Mid-Atlantic Ridge between 14"N and 16°N issupposed

to be in a magma-starved portion of a slow spreading ridge,

where igneous crust is virtually absent and rocks of the low ≪r

crust and upper mantle are distributed on the seaHoor over about

100 km along the strike (e g. Carina! et a!., 1997). The melting

process, melt migration mechanisms, crustal deformation

mechanism during ihc magma-starved tectonic seatloor spread-

ing: these are fundamental questions for the understanding of

global mid-ocean ridge systems. The YK98-05 Shinkai/

Yokosuka cruise (MODE '98 Leg I) was conducted in this por-

tion of the Mid-Atlantic Ridge from June 17th to July 17th

1998 (Sun Juan, Puerto Rico - Lisbon, Portugal) The aim of

the dive survey aboard the research submersible Shinkai 6500

was to characterize upper mantle geochemistry. Particular em-

phasis was placed on determining the variation of these char-

acteristics along ihe axis(Kelemenet al, 1998; Matsumoto et

al., 1998). At nighttime and on the submcrsible's maintenance

days, geophysical surveys, whose items included were swath

bathymetry, gcomagnetics and gravity, were conducted aboard

the ft/V Yokosuka. The aim of the geophysical surveys was to

provide a geophysical characterization of the on and off-axis

ridge flanks, which wifl be used to unravel tectonic evolution

and crjstal structure, both in tlie vicinity of the fraciurc zone

and in the adjacent ridge segments (Fujiwara ≪t al., 1999).

2. Data Acquisition and Processing

2.1 Swath Bathymetry

Survey trecklines were designed to orient all track lines at

least 10 °, and up jo about 25-30 °, to plate flowlincs to assure

that real morphological features, which are normally oriented

parallel and perpendicular to flowlines, can be distinguished

from artifacts caused by beam-point errors in multibcam

bathymetry (Fig. I). Track spacing near the outer edge of the

survey was about 6-7 km, and track spacing over the crest of

the nft mountains again was about 5 km to get complete baihy*

metric coverage, excepting occasional shallow ridges. The sur-

vey tracks covered a distance from 60 km north to 140 km

rig i Survey ship iracks of ≪hc YK ≪ ≫05 cruise Solid circto show ilic submersible SNakHittCO dive sues a irisnctcshcws on XBT ≪≪ion

Thick solid lines dclnicaic rrdae ;i*es and the 1S*2l)xN I rjinsform fault

south of the 15°20'N Fracture Zone. The tracks reached to 70

km off the ridge axis on boih sides, the crustal age of which is

about 5 m.y old (DeMcts ei al., 1990). In the study area, the

total length of the geophysical measurement (racks amounted

to about 3,700 nautical miles (7,000 km), and 27,000 km* areal

coverage was obtained. Differential observation of Global Po-

sitioning System (D'GPS) and World Geodetic System (WGS)

84 geographical coordinates were used to provide thegeogiaphi*

ca) position. and Greenwich Mean Time (GMT) was used for

shipboard operations throughout this survey. Bathymetric data

were collected using a H5-10 multi-narrow Mam echo sounder

surveying system. The MS-10 has 45 beams and a s^ath width

of 90 °, which covers a total of double the water depth. The

sound velocity profile in the water column for depth calcula-

tion was based on the measurement of art expendable

bathoihermograph (XBT)at 15*43.5'N. 4 6°44.$'W (Fig. I and

2)

The bathymetric data were collected with data distribution

density of 100-200 m space interval. The small number of arti-

facts included the occasional "curl-up" of the outermost beams

and a few totally spurious depth readings. About 5 % of beam

points were edited out. Positioning errors, from 22:00 on June

23 1998 to 0 00 on June 24 1998 GMT and from 10:00 on June

2-1 1998 to 22:50 on June 24 1998 GM1*. were coiTected. Cross

over errors in center beam depth within the

YK98-05 data set

have amean value of 3.5431.0 m (Tig.

3} Error? between the

common grid nodes of the YK98-05 and a bathymetry <lataset

of a previous cruise obtained

by a Frcnch research group have

a mean value of

2.7*31.8 m in the northern study area, and of

3.9±2$.0 m in the southern area (Fig. 4), The two datasets were

merged afler the correction by adding these mean depth offsets

1 Smiiul vcliKiiy |)iol11t m the sunly awa

JAMS! EC J OeeoSasfJos

of 2.7 m in the northern area and 3 9 m in the southern area,

respectively. Data were gridded at 0 )' by 0. r (179 m m longi-

tude by 185 m in latitude) cell size in order to even out the data

density while preserving the characteristic high resolution of

multit>eam data.

Fig. 1 Histogram of cross over errors of ihc ccnwr beam bwhymcery

n) (he YK98-C5 sur>c>.

Fig 4 II i$lo$ran> or*<Ji flcrcnw* hcmvcn the YK9$ ≪05 and ilw l:r≪nch

survey (;i) ninth of <≪c IJ°2Q'N 1 ruciuw/onc (bisowtliof

ihc I5*?0*n rnicfur: /oik*

2.2 Ceomagetin

Ocomagnclic total force data were obtained by using a sur*

fecc≫towcd proton pre cession magnetometer STCiO (Kawasaki

Geol. Eng. Co) Tke sensor was towed 350 m behind the ship.

duced The method uses the anomalies and differences at track

crossovers to produce an acceptable diurnal variation curve

(Buchanan et al.. 1996). Corrected anomalies reduce cross over

errors of up to -2.7± ＼ 3.5 nT (Fig. 5)

Victor nagnciic Held data were collcctcd using a shipboard

ihrce-componcnt magnetometer, 'llerra Tecnica SFG ≫ J 212, with

a Tokimec GS-110 gyrocompass an<t a vertical reference unit

ofllic I IS-10 muIti-narrow beam echo sounder. The data sam-

pling rate was 8 Hz The vector gcomagnciic field was derived

from the observed magnetic field using the formulation de-

scribed in Appendix (fsczaki, 1986). "rigure-8 tum"s (a ship

runs along an 8-shaped track consisting of two circles) were

made three times for calibration of the ship's magnetic effect

(Table I). The vector geomagnetic anomaly was also calcu-

lated by subtracting the IGRF 1995 model from the observed

field. The resultant anomalies suffered from strong bias. DC

fields were removed from the observed field with cach track

unii

30

20-

cg

o

10

0

Location

/AMSTEC

Pier-3, Sun Juan, Puerto Rico

Frontier Base Pier, Sun Juan

Capital Dome. Sun Juan

Ptcr-3,Sun Juan

Pier St. Apolonia, Lisbon, Portugal

Pier St. Apolonia, Lisbon

Doca Da Alcantara, Lisbon

Eastern End of Doca Da Alcantara, Lisbon

Pier St. Apolonia, Lisbon

>20 ･10 0

Cross Over Erro; (mgaij

TO 20

correction tak i ng i mo account the sensor cable length, ihe ge-o-

magncttc total force anomaly was calculated by subtracting (he

International Geomagnetic Reference Field (IGRF) 1995 model

(1AGA, 1995) as the rcfercncc field and by offsetting the DC

Held of *60 nT from the observed field The data yielded cross

over errors of 3.4622.3 nT. Diurnal variation was estimated by

□sing the geomagnetic field observations themselves and rc-

Tabk ) Log of Figure.8 turn"*.

Si Datc(GMT) Location

20/6/98 18; 10-18:22

01/7/98 21:43-21:55

08/7/98 17:57-18:25

15* 43.7'N, 46 °46.2*W

14* 55.8'N, 45 ° 00.2'W

14* 08 5'N. 44R 23.9'W

St. Date(LocalTime) G*1039Gravity

1 19/5/9815:202 16/6/98 09; 163 16/6/98 09;4i4 16/6/98 11:245 16/6/98 12; 176 17/7/98 14:547 17/7/98 16:208 17/7/98 18.069 17/7/98 IS 2510 17/7/98 18.56

ISO

120

SO

GO

30

o'g

33*0.90

2247.542249.752245.062247.55

3655.293655.133665.39

3663 103655.34

fable 2 On-land gravity measurements.

Absolute S ≫63

Gravity Cravuy

979758.1 10871.4

978664,5

978666.7

978662.0

978664,5 9779.4

11201.0

980079.8

980089.3

980087.9

11200.9

-60 -40 -90 0 20 40 f

Cross CWgr Sfror (nT)

5 I Iwioyfiicn of crovs over error* of the jjPom:igifcl≫c loifll

MI> iufimaiic^

.lahd^TCr I Oat n<l!l<l)

1*ig 7 Dtiihymcrituiiuipol ihc Mui-AUanuc lOd&c Cornours ore at 100 m mtciv.ili (n> norlli of the 15*20'N Fraciurc Xone (hi south

Of I lie 15'JO'N' I'Mcture/one

JA.MSTEC J Deep S≪e Rc$.. 15 <l-i|i|≫>

I?

2.3 Gravity

Gravicy data along the ship tracks were collected using the

LaCostc & Romberg shipboard gravimcter S≫63. The daia were

recorded every ) 0 seconds. Sh i pboard gravity data were tied to

absolute gravity values at some reference points measured us-

ing LaCostc & Romberg gravimeter G-1093 (Table 2) After

Eotvfo correction and sensor drift correction with ihe rate of

0.12 mgaVday, free-air gravity anomaly was calculaicd by sub*

if acting the theoretical gravity formula of the Geodetic Refer-

ence System 196? from the observed data. A total of203 track

crossings for estimation of crossover errors were obtained dur-

ing th i s cnjisc. Crossover errors have a mean value of -0 5 mgal

ami a standard deviation of 3.4 mgal (fig. 6).

3. Summary of Observational Results

3.! Swath Bathymetry

The median val Icy of the ridge axis i n the north of the 15°20'N

freclurc Zone trends in the N-S direction (Figure 7(a)). Mor-

phological trend of about N10 °E is found on the ridge flanks

and in the median valley as well as the N-S trending. The di-

rection ofNIO^E corresponds to the normal direction of the

flowlinc of plate spreading estimated by a global analysis

(DcMetsct a I.. 1990). Along axis discontinuity, represented by

traccd in the western flank from the rldgc axis at 15'50'N in

the strike of N60*W. The ridge flanks from the fraciure zone lo

15 °S0'N show asymmetrical morphology (Figures 7(a) and

8(a)). Dome-shaped topographic highs arc dominant in ihc west*

(Casey ctal., 1998). A dome-shaped topographic lugh located

30 km west of the northern segment, termed a "megamullion>*

was identified. Oabbro and pcridolite were collected by the

Shinkai 6500 along a transect up the southern slope of this fea-

ture and upper mantle rocks on the seafloor on very long-lived*

low angle faults due to tectonic seafloor spreading In the ab-

sence of magmatism In contrast, topographic depressions

deeper than 3,000 m are dominant except inside comcr highs

in the eastern flank.

The 15 °20%N Fracture Zone is lineatcd In a direction of about

NIOO'F. From 46°40'W to45 ≫$5*W. the fraciurc wne (trans-

form fault) has another (incation in a direction of about N9066.

The ridge axis Is running in the strike ofaboutN106E between

the eastern ridge-transform intersection and 14 °40'N (Figure

7(b)). In these off-ridge flanks, two topographic lineations of

N10*E and N≪S coex i st. The ridge nan ks show asym mctry: the

western flank is shallower than the eastern flank except inside

comcr highs (figures 7(b) and 8(b)) Seamount-hkc topographic

(b)

/' '"-1 f ' /T

J . -A/-

v＼/ Vv^.y ＼ /≫.

/' ,> X-'-A A" 4 /-/x' xN'n-,.W ＼

,,-w.

＼V

.. '* w/?J-

*/･' v

･d&e ･*$ ≪ '4^.4 .45.? '44.6 >44.6 -4*.4 ≪≪2

LorigmKJc f)

JAMSTEC J. Deep See Res . 15 < ifWGI

(a)

tysiy

iw

･sur

is***

iS'i/r

': / - ･"*･'

Lcigiiude O

M

ojxa -jc 5CK) m apan Dashed line iab<;l> cfntcspondj; in

3 000 m depth for ihc profile <*) norih of the 15°20'N I rac≪

Jure/otic III south of Ihc J5*20'N Fractuic/one

･2000

･3000

e,

i

a>O

･4000

･5000

･6000

(a) -2000

･3000

£

jz

"5.cQ

･4000

･5000

･600015 2 15.4

Along Axis Depth Profile

15.6 15.8 16.0 16.2 16.4

13.8 14.0 14.2 14.4 14.S 14.8 15.0 15.2

Latitude (')

Pig. 9 A long-axis dcpih vatiairon. (a) north of the I5R20'N Fracturc Zone. Dashed line show* overlapping spreading ccnicc

(b) south of the I5°20'N FraciweZonu

hncfited rift mountains arc dominant in ihe western flank. South

of]4*40'N, the topographic (mention shows only the N-5 trend-

ing, and the off-axis morphology

is symmetries. The depths of

the ridge Hanks in this portion arc the iliaHov/csi in the study

area.

Four major segments are recognized

along trie ridge axis in

the study area. North of I 5R50N, axial depth is nearly constant

M about 3.SOO m (Figure 9(a)). Prom 15°50'N to the fracture

zone, the axial depth increases gradually. Between I4 °J$'N

and 15R05'>J, th≪ ax ial depth i s constani al at>out 4.000 m (Pig-

tire 9( b≫. These is an isolated deep al 14°25 >J. The axial depth

<lcc reaves toward the soulh.

3.2 Geomagnetic Anomaly

Lincation pans rn< of geomagnetic anomalies are unctcar in

ihc ridge Oanks. even the central anomaly magnetic high in ilic

"icdian valley {Figures 10 and 11). The patterns, of gcomag-

rictic ＼orrkl force can be highly skewed bccause ihc ≫UKJy area is

dry. A uniform magnetic source layer with thickness of >00 m

was assumed. The direci ion of niagncazai ion of ihc source layer

was assumed to be oriented parallel to an gcocertiric dipolc

field {Figure ＼2).

The major isocliron was identified preliminarily using ihe

polarity lime scale of Cande and Kchi {1992). Tlie mean spread*

ingraie i$ consistent with the ftill spreading rate of 26.7 km/

rn.y. calculated from the model of global plate moiion (DcMets

ct aL 1990). The geomagnetic lincation pattern is, however,

still unclear and discontinuous after this deskew operaiion, ck-

cept the area south of14R40'N. The observed pdiclij Imcation

partem is consistent with limited basalliv magma supply over a

period of several million years. compared with more

magmaiically robust parisofihc Mid*Atlantic Ridge,

'Hie highest gcomagnelie anomaly is situated at the western

ridgC'transform intersection. The result suggests that ihe most

active volcanism is occurring at this region, litis 1$ consilient

with the submersible observation. Soulh of the fracture /one,

magnctixaiion highs are not located along the median valley.

1 he result may surest an abnormal mode of seaOoor spread*

ing m litis region. Low amplitude magnetisation variation of

the ridge flanks in the south of the fracuirv *o≫e. cornered

xv nil ihm in iKe north of the fraeturc /.one was observed

Hi

calculated from ittc gcomagnciic loial force anomalies, can show

magnetic lineauon [Wiicrns related to plaic spreading. Crusial

tfiajjncii'auon was calculated n^ing (lie method of Pariccr and

Hucsiis (i974>. which took imo account (liceffcci oi'bfiihjm-

JAMSTEC J. Deep Ses Res , 15 (10OT)

Geomagnetic Total Foice Anomaly MAR 15e20'N FZ (25nT contoured)

(a) 16'20'N

16s OO'N

15* 40 N

!5e20'Nf2

15' 20'N J-

15* OO'NAT 20'W 47# 00'W 46* 40'W 46' 20 W 46' OOW

0>) 15'20'N

no

15' OO'N

14* 40'N

14 20'N

14 ° OO'N

(nT)

300

250

200

150

100

50

0

･50

･100

･150

･200

･250

-300

45 ° 40'W 45s 20'W 45 ° OO'W 44' 40'W 44s 20'W

f-'ig 11> Oconicigatcnc (oral intensity anomnJ> t<niunir≫ arc at 25 n'C uitcrviK I (vavy black imcs bhew baihymciry wuii

I 00c inwnuni; inlcrv.il* (≫>north rvfthe 15°?0 N f;mciurcZone <b>SOtilh i>f tlic l^°20'N l-rnclurc /one

JAMSTEC J Oeeji S≪a ftos., 15 (!W

(a)

16' 20'N

16* OOtJ

15' 40'N

15* 20'N

45* WN

A7' 20 W 47* OOW AS' 40W 46' 20W 46* Oiyw

(c)

16* 20'N

IS* WN

15' 40'N

15*2014

15* OO'N'

(c)

CM

15* 2CN

15' OffN

14* JQ'N

14'20'N

14* OffN

<0

M)

15* 2CTN

15* OON

W dO'N

14* 2CfN

Id* OO'N

19* 20N

15* OCTN

14' 40'N

W 20'N

14' OO'N i

16* SON

18'00'N

15* 4QN

15' 2ffN(

IS' OffN

47' 2(KW 47' 00"W At' 40W 46' 20"W 46* CO'W

Pig. 11 Vccior §eoniag≪≫citc anomaly along shipnacks lajXinonliwariUixinipiincni anomaly nordk c>rthe IS*20 N'Kracmrc2o ≪c

<b) X(norihward) component an≪nal> 9o≪nlt of ihc 1 $R20'N Fracture /one |c> Y loastv.ard) componeni Anomaly nouli of

ihc IS°20*N Praciurc '/ijtk id) Y ioasiwaril) tfomponcni anomaly south of ■!)< 15'20'fc i-jnuirc /one <cl V (downward)

tomponeni anomaly nwihor (he 15e20'N Fracture /i>nu. (f) 7. (downward) component anomaly loolliof the

f mc≪urc ionc

JAM8T6C J Omu Sua . >5<| ≪1S!))

9.1

(a) 16 ° 20'N

16* OO'N

15* 40'N

Magnetization MAR 15'20'N FZ (2A/m contoured)

^ °20'Nfi2.

15' 20'N I―

15s OO'N47s 20 W 47' OO'W 46s 40'W 46' 20'W 46s OO'W

(b) 15e20'N

22

15* OO'N

14 ° 40 N

14 ° 20'N

14s OO'N45s 40'W 45 ° 20'W 45 ° 00'W

15'20-NFZ

44° 40'W 44* 20 W

(Mn)

24

20

16

12

8

4

0

-4

･8

12

16

20

24

!rig 12 Magnciir-tiion disiributio". Contours Arc si 2 AJm intervals. I tcavy black lines show bathymetry with 1,000 m

conjoin inicT'als <a) north of ihc 15°2Q'N f-'rjiciurc /jjitc (h) somh of the 1 S*20'N Fracture ?.onc

JAMSTEC J Deep Ses Res., 15 <Hra>

(a) (b)

(Figures 12 and 13). This result is consistent with mag-

netic measurements of cocks sampled by the Sliirikai 6500

(Joshimaet el., 1999) I Ins result suggests weathering of

basalt dlie to tectonic spreading, or else the geochemistry

of basalts is different from the north region.

3.3 Cravity Anomaly

Free-air gravity anomalies obiained in this cruise are

shown In Figure 14. The free-air gravjiy anomaly dau are

merged with Ok gravity anomaly dale derived froin satel-

lite altimetry (Sandwell and Smith. 1997) to cover ihc un-

stirvcyed area Mantle Bouguer correction was calculated

using the method of Kuo and f orsylh {) 9SS). The crusi

was assumed 10 have constant thickness of

6 km thai fol-

lows ihc scaftoor relief. The assumed density of the crustal

layer is 2.700 kg/ni'. ajid thai of Ihc underlying mantle is

3.300 kg/m ＼ The global predicted baihymetric data (Smith

and Sandw ≪|l. 1904) were employed to fill the unsurveyed

area. Manlte Bouguer anomaly was obtained by subtract-

ing the mantle Bouguer correction from Ihc frec-air grav-

ity anomaly.

)/>w amplitude "hull's eye' msintle Gougucr anomalies.

JAMST6C J Deep See Rc$.. 16 (I9M)

13 Acrcss'ixis vansnon of mignedzaiion. The tick marts on ihu vertical axes arc 5 A/jii span Dashed Hoc with labels corrc≫ponds

to 0 A/n* for the profile, (a) north of Oie 15R20'N Fracture Zone. (W south of (he 15e20'N Fracturc Zone.

compared with more magmaticaily robust parti of the Mid*

Atlantic Ridge, were observed on the ridge segments Hank*

ing the active transform fault (Figures 15 and 16). This i$

probably due to a limited magma supply thai has not gen*

crated thick igneous oceanic crust. The gravity data sup-

jiort the conclusion that the magma-starved nature of ihis

region is a long lived phenomenon. Gravity anomaly maps

show concentric, sctri-cireular contours surrounding these

gravity lows, becoming less negative toward the fracture

zone from the norih arid the soulh. Tltc largest groviiy

anomalies arc "bull's eye" lows along the rtdge a*is at the

northernmost and southernmost limits of our survey area,

though some of them fa ■ I to show thei r com plcte shapes

due to the limitation of the surveyed area. The ccntcrof

the largest anomaly is located at 14 °00'N and the ampli-

tude amounts 10 -60 mgals (Figure 15(b)). The simplest

intenwetation of these data seems to he that the large gr*v*

icy lows represent the centers of unusually large tnaginatic

segments ai approximately I6 °N ami !4aN. probably as a

result of 3-dimensional focusing of melt generation pro-

cesses in the mantle-dominating tectonics and

geociynamics in this region.

Free-air Gravity Anomaly MAR 15*20'N FZ (5mgal contoured)

(a) 16# 20'N

16s QQ'N

15 ° 40 N

IS'20'N'fz.

15e 20N

47e 20'W 47s OO'W 46* 40'W 46' 20'W 46 ° OO'W15* OO'N

(b) 15'20'N

15 ° OO'N

14e 20'N

14' OO'N

14* 40'N

45* 40'W

'5 20'NF2

45* 20'W 45# OO'W 44' 40'W 44# 20'W

(mgal)

100

80

60

40

20

0

･20

-40

-60

-80

100

25

(a) north of liiu 15*20 N Kmclurc Yaw <bl ^on<1aoTihc I5R20'N rr≫cUirc7onc

JAM8T6C J. Deep Sea Acs . 15 (lim>

滿 ｀

Fig 15 Mallllc Buu&u ｅralｓmaly （ ｏ`ｎ１０ｕrij靱『vahsSmga 囗lcavyblackllncsshowbalhymclrywilhl,000m

叭lnltlllr intervals Ｏ¶）norllldlhc 15 °20` NFraalrc かnt （b ）so凵lh,1rLhc15 °20` N Ffac10『c Ｚonc

(a) (b)

Ofiglhrfe (']

Pig. 16 A≪oss-axi* variation of mantis Bougucr anomaly.

The nek marks on ihc vertical axes arc 5 mj-al

apart. Dathcd lnic wiih labels

corresponds 10 0 m-gi Tor the profile, (a) north of

the 15'20'N Fracii^c Zone, (b) south of the

I5*20'N Traewre Zone.

anomaly from Gcosaiand ERS I satellite allim≫

euy, J. Geophys Res.. 102, 10.039-10.054

(1997).

Smith. W. H. F ar≫d D T Sandwell. 'Uaihymetric pccdic-

lion from dense allimetry nnd Sparse shipboard

bathymciry". J. Geophys Res, 99, 21.803-

21,824(1994).

Smith. W. H. F. and P. Weasel. "0ridding wnh continuous

curvature splines in icnsion". Geophysics, 55*

293-305(1990).

Vessel. I', and W. H. F. Smiih."Free software helps map anil

display data". EOS Trans. AGU, 72. 441.445 ≫≪M6

(1991).

(≪C≪M : 1999* 8 17 H>

JAMST6C J Deep Sea Ros.. 15 UStt≫>

lOngHu0c ()

･>7