E a rth a n d ... 363 E lsev ier S c ie n c e ...

E arth a n d P la n e ta ry Scien ce L etters, 83 (1 9 8 7 ) 3 6 3 -3 7 5
E lsevier S cien ce P u blish ers B .V ., A m sterd a m - P rinted in T h e N eth erla n d s
Nankai Trough, Japan Trench and Kuril Trench:
geochemistry of Fluids sampled by submersible “ Nautile”
J a c q u e s B o u lè g u e \ J e a n T . Iiy a m a 2, J e a n -L u c C h a rlo u 3 a n d J a c q u e s J e d w a b 4
' L a b o ra to ire d e G éoch im ie e l M éta llo g én ie ( C N R S VA 196), U n iversité P ierre et M a rie Curie,
4 p la c e Jussieu, 75252 P a ris C éd ex 05 (F rance)
' F aculty o f Science, U niversity o f Tokyo, B unkyo-ku, T o k y o 113 (Japan )
' D 3 G M , I F R E M E R , C en tre d e B rest, B .P . 337, 29273 B rest C éd e x (F rance)
4 L a b o ra to ire d e G éochim ie, U n iversité L ib re d e B ruxelles, 5 0 A v. F.D. R oosevelt, B -I 0 5 0 B ru xelles (B elgique)
R ev ised version accep ted February 5, 1987
D eep -w a ter sa m p les co lle cted du rin g the K a ik o project are o ften asso cia ted w ith b io lo g ica l c o m m u n itie s lo ca ted on
g eo lo g ica l structures favorable to fluid ventin g. T h e evid e n c e o f fluid v en tin g are the tem p eratu re a n om alies, the
decrease in su lfa te co n c en tr a tio n s, the c o n te n t in m eth an e and the lo w C j / f C j + C ,) ratio o f light hyd rocarbons.
B ecause o f large d ilu tio n by a m b ia n t seaw ater du rin g sa m p lin g it is d iffic u lt to c o m p u te the c o m p o sitio n o f the
a d v ectcd en d -m em b er pore fluid. Part o f this flu id sh ou ld origin ate in the “ p etroleu m w in d ow " , i.e. at tem perature
a b ou t 60 ° C. M o d e lin g the upw ard flo w o f w ater, tak ing in to acco u n t the a n o m a lie s o f tem p eratu re m easured on the
sea flo o r and th e g e o ch em ic a l a n o m a lies, leads to n o n -stea d y -sta te a d vection o f th e pore flu id. T h e o ccu rren ce o f a d eep
co m p o n en t in th e flu id has im p lica tio n s for the g eo lo g ic a l and tecto n ic m o d els o f the su b d u ctio n z o n es o f f Japan.
1. Introduction
2. Sampling and analyses
T he discovery of clam com m unities associated
with fluid venting in the accretionary com plex of
the O regon subduction zone has opened the dis­
cussion on the origin and the source depth of the
advecting fluids [1,2], O ne of the purposes of the
K aiko program was to test the possibility of a d ­
vection of fluids associated with the dew atering of
the accretionary prism in the subduction zones off
Japan. T here were some previous hints that such
advection was occurring from the D SD P results
off northern Jap an [3], It was found th at the light
hydrocarbon ratio C , / C 2 was possibly lowered by
inputs of petrogenetic hydrocarbons and that the
8n C of C H 4 was correlatively lowered [3]. Hence,
the bottom w ater sam pling during the K aiko legs
was oriented tow ards u n derstanding these possible
processes at locations where the presence of b io ­
logical com m unities indicate th at fluids m ay be
venting. T he capability of handling tools from the
subm ersible “ N au tile” was extensively used to
fulfill this program .
The w ater sam ples were collected either with
titanium syringes handled directly above the sedi­
m ent interface or with an alum inium box corer
and core sam plers lined w ith plastic by w ithdraw ­
ing the su pernatant water. T he use of syringes to
attem p t to sam ple as m uch as possible of venting
fluid associated with the biological com m unities
posed constraints on the representativeness of the
samples. The flux of fluid passing through the
com m unities is sm all (see later) and so as not to
dam age the instrum ent it was deployed in the free
bottom w ater w ithout touching the solid sub­
strate; hence the fluid sam ples were a m ixture of
m ostly bottom w ater with a small percentage of
fluid advecting from the sedim ent.
T he location of the sam ples is listed in Table 1
and shown in Figs. 1 and 2. The sam ples that are
clearly associated with biological com m unities are
HY4, 7, 8, 9, 10, 14, 15, 16. Fig. 2 shows the exact
locations of the sam ples from the clam com m unity
in the deep-sea fan of Tenryu C anyon together
with the m easured tem peratures. Sam ple HY16
0012-82 1 X /8 7 / S 0 3 . 5 0
© 1987 E lsevier Scien ce P u blishers B.V.
;h v i o
h y b
H Y ?.
water= 1.19°C
HYI 4'
Fig. 2. D etailed sa m p lin g sch em e in the b iological com m u n ity
o f T enryu d eep sea fan H Y = w ater sam ple: 5T L 5 N C and
5BC" are sed im ent sam ples. M easured tem peratures are also
given ( ° C ) . A fter a d raw ing by A . Taira. T he black d ots
corresp on d to sam plin g lo c a tio n s and tem perature m easure­
m en ts (see |28]J.
Fig. 1. G eneral lo ca tio n m ap o f d eep -w a ter sa m p les from
s u b d u ctio n zo n e s o f Japan.
was taken 2 m from sa m p le H Y I 5 so as to check
for the effects of m ixing o f advecting fluids with
seaw ater over sho rt distances. Som e of the o th e r
sam ples m ay also be associated with fluid venting
because of favorable geological structure. T his is
the case for H Y 3 located in a region of small
ledges, in the vicinity o f a thrust in the Z enisu
Basin. In the vicinity of T en ry u C a n y o n , samples
H Y I , 2 an d 5 will be considered as reference
sam ples of local deep-sea water.
L ocatio n s and d ep th s o f w ater sa m p les o b ta in ed du rin g K aik o project
S am ple
L oca tio n
D e p th (m )
Sam pler a
C om m en ts
H Y 11
H Y 12
3 8 ° 1 6 '4 2 N , 137 ° 2 4 '1 5 E
N an k ai T rough
sam e as H Y I 1
1IY 13
3 3 ° 1 6 '5 3 N , 1 3 4 ° 2 4 '0 6 E
42 3 4
o live-grey sem i-c o n so lid a te d clayston e; N a n k a i T rough
olive-grey sem i-c o n so lid a te d cla y sto n e
3 3 ° 3 6 '6 0 N . 1 3 7 ° 3 2 '6 0 E
T enryu C an y o n d eep -sea fan
3 3 ° 3 6 '0 9 N , 1 3 7 ° 3 2 '6 0 E
T enryu C an y o n d eep -sea fan; N -S ou tcro p
3 3 ° 3 7 '6 2 N , 1 3 7 ° 3 2 '1 0 E
top o f an ticlin e; sid e o f T enryu C an y o n
3 3 ° 3 6 '3 5 N . 1 3 7 ° 3 2 '2 5 E
clam field , N W -SF . fault N 5 0 ° F . flexure; T en ryu C an yon
sam e as H Y 4
see Fig. 2
sam e as H Y 4
sec Fig. 2
sa m e as H Y 4
B .C
see Fig. 2
H Y 10
sam e as H Y 4
T .C .
see Fig. 2
3 3 ° 3 6 '2 1 N , 1 3 7 ° 3 2 '0 3 E
3 3 ° 11 '4 8 N , 1 3 7 ° 5 3 '0 1 E
T.C .
T.C .
Z en isu Basin; coarse sa n d y s em ico n so lid a te d m ud:
H Y 18
H Y 19
33 °4 5 'O O N , 1 4 2 ° 2 9 '5 6 E
flank K ashim a S eam ou nt; calcareou s ou tcrop
T .C .
H Y 14
35 ° 4 8 '3 2 N , 1 4 2 ° 2 5 '7 2 H
35 ° 5 4 '6 2 N , 1 4 2 ° 3 0 '9 0 F .
olive-green mud; K ashim a, innerw all o f Japan T rench
c lam field; N 2 8 0 ° E flexure; innerw al o f
Japan T rench, K ashim a
H Y 15
4 1 ° 1 8 '0 5 N , 1 4 4 ° 4 4 '0 7 E
clam field; sm all led ge on subvertical
H Y 16
sam e as H Y 1 5
2 m o ff H Y 15
o live-grey m ud 300 m o ff clam field; T en ryu ca n y o n
fault indurated mud
“ T .S . = titanium syringe; B.C. = box corer; T.O . = tube corer.
The syringes were subsam pled on board ship
for 3H e / 4 H e and other rare gases m easurem ents,
for light hydrocarbons and for 180 / 160 and D / H
determ inations. Sam ples for n itrate were frozen
upon recovery. T he determ inations of p H and
alkalinity were done on board ship using Tris
buffer as reference. Dissolved sulfide was de­
term ined by potentiom etry w ith an A g /A g 2S elec­
trode. Sam ples for trace elem ent chemistry' back at
the shore-based laboratory were filtered (N ucleopore 0.1 jam) and acidified with u ltrapure nitric
acid. The filters were stored for m icroscopic ex­
am ination. T he tem perature of the sam ples and
the am bient seaw ater were determ ined in situ with
the tem perature probe ( ± 0.005 °C ) attached to
one of the arm s of “ N au tile” . D uring shipboard
subsam pling and handling the tem perature was
generally less than 7 ° C. G as overpressure was
also found in several sam ples (H Y 3. 5, 7, 8, 11,
18). In the shore-based laboratory 180 / 160 and
D / H were m easured by mass spectrom etry and
light hydrocarbons by gas chrom atography. M eth­
ods for o th er elem ents are given in Table 2 to ­
gether with accuracies. T he filters were exam ined
by SEM coupled with T R A C O R to identify
m inerals. Some particles were also exam ined by
X-ray m icrodiffraction.
Results are given in Tables 3, 4 and 5. T otal
dissolved sulfides were always less than 10 7 M.
T he m ajor cations (L i*, N a 1, K / , R b / M g 2 t ,
C a 2*, S r2 + ) and anions (C l- , F - ) of seaw ater
A n a ly tica l m eth o d s
E lem ent
M eth o d
A ccu racy
N a , K, Ca
sea w a ter a
seaw ater
Li. R b, Sr
seaw ater
seaw ater
Fe, M n. C u, C o 1
M o. Ba. Pb. C d ƒ
A g* T
seaw ater
N O ,. SO„
seaw ater
seaw ater
A ik
seaw ater
F A A = flam e ato m ic a b so rp tio n ; Z C F A A = Z eem a n corrected
fla m eless a to m ic ab so rp tio n ; A g + T = titration w ith H g + u sing
an A g -A g 2S electrod e; P = p o ten tio m e tr y w ith sp ecific e le c ­
trode; IC = ion ch rom atograp h y; M C = m o ly b d a te c o lo r im e ­
try; H + T = a cid titration.
" Seaw ater w ith a d d itio n o f elem ent.
showed no significant variation. The only excep­
tion is a 1-2% decrease of M g2 + in HY4, 7 and 8,
accom panied by the decrease of the M g /C l ratio.
C onsidering that elem ents such as Ca, Mg, Li, Sr
show significant variations in the pore w aters of
nearby sedim ents [4,5], this signifies that the frac­
tions of advected pore fluid collected in these
sam ples are less than 10%. H ence the possibility of
assessing the characteristics of such fluid will be
lim ited. This point is discussed further below.
3. Geochemical evidence for advection of pore
W e will first discuss the evidence for possible
advection of fluids at the locations of sampling,
then the characteristics of the fluids and possible
controls on their com positions and last their origin
and the rate and m ode of advection of pore fluid.
T he sam ples show no significant variations in
the com position of the m ajor dissolved cations
and chloride which would have enabled us to
place lim its on the contribution of pore w ater to
the total samples. However, in the absence of such
clear diagnostic features several other im portant
param eters and species show sufficient variations
from which we can deduce the participation of
pore fluids.
O f these the m ost straightforw ard evidence is
the tem perature anom alies m easured in situ. We
have found tem perature anom alies in the range
0 .1 -0 .4 5 ° C in the im m ediate vicinity of biological
co m m u n itie s an d an o m alies in the range
0 .01-0.03° C in the overlying bo tto m waters. Even
if some may be due to biological activity, it is
doubtful that significant tem perature differences
can be m aintained in open w aters if there is no
input of fluids from the sedim ent. Similar tem per­
ature anom alies were also observed in the vicinity
of fluid vents in the subduction zones off the coast
of O regon [1],
The small M g 2+ decrease in HY4, 7 and 8 can
also be related to mixing of advected pore fluids
and to the tem perature anom alies in the same site.
T he occurrences of large concentrations of light
hydrocarbons (about ten times the expected level
for m ethane and m uch m ore for C 2 and C 3) are
also related to advection of pore fluids (Table 5).
T he m ethane contents are in the same range as
found in sam ples from the vent areas in the sub-
W ater sa m p les. M ain param eters: tem perature, p H and alkalinity (A ik ) and co m p o sitio n o f characteristic d isso lv ed elem en ts. The
tem perature read ings are from the “ N a u tile ” p ro b e and the tem p eratu re d ifferen ces ( à T ) w ere th ose o b served du rin g sa m p lin g with
reference to a m b ia n t d eep o cea n w aters
Sam p le
H Y ll
H Y 12
Sam pler
H Y 13
T .C .
T .S.
T .S.
AT( ° C)
A ik b
SC O ,
SO 2
NO, c
E ed
M nd
Cu d
Mo d
0 .7 2
Sam ple
H Y 10
Sam pler
T .C .
T .C .
T( ° )
A T (°C )
A ik b
0 .5 2
H Y 19
H Y 14
H Y 15
T .C
T .S.
T .C .
T .S.
n.m .
Cu d
M od
0 .9 4
Ba d
Fe d
M nd
0.0 9 7
N O ,“ c
sol a
S iO /
T .S.
T .S.
n.m .
0 .8 6
H Y 16
< 0.2
C on cen tra tio n scales: a 10 “ 2M; b 10
3 M ; c 10 5 M; d 10 7 M.
Pb and Cd c o n cen tra tio n s w ere alway:s less than 10 * M , ex cep t H Y 9 w ith 3 x IO “ 8 M. n.m . — not m easured. For sam pler, see Table
180 / u’0 and D / H o f w ater sa m p les (exp ressed in S r*c units versus SM O W standard)
S 's O
- 0 .2 4
-0 .3 9
- 0 .3 2
- 0 .3 0
0.9 6
- 1.31
- 0 .1 8
t 1.94
H Y 18
H Y 14
H Y 15
H Y 16
- 0 .2 5
-0 .9 7
-0 .3 0
- 0.37
- 6 .2 1
M eth an e ( C ,
eth a n e (C 2 ) and propan e ( C , ) co n c en tr a tio n s in
w ater sam ples expressed in n l/1
Sam p le
< 1
< 1
C ,/ ( C , + C ,)
H Y 14
H Y 15
< 1
< 1
< 1
duction area off O regon [1], The C , / ( C 2 + C ,)
ratios are very low. which may be indicative of:
(1) extrem e oxidation of a microbiologically p ro ­
duced hydrocarbon pool, m ostly of m ethane, in
the upper 50 m of the sedimens [6,7]; or (2) input
from therm ogenic hydrocarbons. In this case they
should have been produced at tem peratures higher
than 60° C. Therm ogenic hydrocarbons may also
have been partially oxidized in the u pper p a rt of
the sedim ent. A m ixture of both sources of h y d ro ­
carbons is also possible.
The decrease in concentrations of sulfate found
in several sam ples collected above the seafloor
(H Y 4, 6. 7, 8, 9, 3, 14, 15) is also characteristic of
the advection of pore fluids. It is difficult to
ascribe the decrease in sulfate contents to the
sulfate-reducing bacteria, since the variations are
not accom panied by an increase in dissolved
nutrien ts an d total dissolved carbonates. In the
area off Jap an such an expected correlation be­
tween S 0 4, nutrien ts and dissolved carbonates is
found in sedim ents from the N ankai area collected
during the D S D P program [5], The decrease in
sulfate contents can be due to a contrib u tio n of
advected pore w ater in the vicinity of biological
com m unities since sim ilar sedim ents collected in
the Tenryu C anyon (H Y I 2, 13) do not present this
decrease of sulfate content. T he decrease in sulfate
in the open w aters show th at the percentage of
advected pore fluid th at was collected is at least
5% for HY8, 14 and 15, 8% for HY3 and 4 and
m ore than 10% for HY7.
T h u s te m p e ra tu re anom alies an d several
anom alies of dissolved constituents point out that
the w ater sam ples collected above biological com ­
m unities (H Y 4, 7, 8, 14, 15) or above geological
structure (H Y 3) m ay represent a m ixture of n o r­
m al oceanic seaw ater and advected pore waters.
In the N ankai T rough area the results obtained
d uring the D SD P program [5,7] have shown that
diagenesis has induced only lim ited changes in the
pore w aters in the upp er hundred m eters of the
sedim ent colum n. T he mixing of such pore w ater
with deep-sea w ater should therefore not be ex­
pected to yield large deviations in concentrations
of the m ajor elem ents. Also, the m etabolism of the
Calyptogena may have induced changes in the
advected pore w aters such as oxidation of light
hydrocarbons and uptake of carbon dioxide, hy­
drogen sulfide [8], and of m etals such as Fe, Mn,
C d, Mo [9]. H ence it is difficult to assess the exact
end-m em ber com position of the advecting fluid.
different from th at of seaw ater collected at the
sedim ent interface.
The alkalinity (Aik) versus total dissolved
carbonate ( S [ C 0 2]) diagram , (Fig. 3) shows that
m ost of the free bottom w ater sam ples are in the
range expected for deep Pacific Ocean w ater [10].
T he expected variation ( A A lk /A 2 [ C 0 2]) for
Pacific deep w aters was found for HY5 and H Y I
in the Tenryu C anyon deep-sea fan. The super­
n a ta n t w aters (H Y 6, 9, 10, 12, 13) also show a
sim ilar relation characteristic of carbonate dissolu­
tion and degradation of organic m atter [10], It is
possible th a t m ostly carbonate dissolution is im ­
p o rta n t in sam ples HY2, 3, 11, 18 [10]. The AlkS [ C 0 2] values for the sam ples where advection of
pore w ater and changes due to clam m etabolism
are expected do not generally follow the above
trends. From the A lk -£ [C 0 2] relation and Alk-pH
relation it can be deduced th at sam ples HY4, 7, 8,
14, 15. 16, associated with biological com m unities
have higher alkalinity and lower S [ C 0 2] than
expected. These may be due to advected pore
w ater input and dissolved carbonate uptake by the
Fig. 3. A lk a lin ity (A ik ) vs. total d isso lv ed ca rb o n a te ( 2 [ C 0 2])
diagram . T h e nu m b ers corresp on d to the sam p le references.
4. Geochem ical characteristics of the fluids col­
lected during Kaiko
A part from the characteristics discussed above
the chem istry of the sam ples is generally not very
T h e square su rrou n d in g the sam p le nu m ber corresp on d s app roxim atively to the a n alytical un certainty. T h e d ash ed lin es
corresp on d to the relation ( A A l k / A S [ C 0 2 ]) = 0.93 exp ected
for d eep P acific w aters (see [10]). T h e low er lin e corresp on d s to
th e sup ern atan t w ater sam ples. T h e d o tte d lin e corresp on d s to
th e relation du rin g carb on ate d isso lu tio n .
clams. Samples HY15 and HY16 were collected
exactly at the sam e site, with a large dilution by
seaw ater for HY16 (as shown by the values of AT,
S 0 4 and C H 4). The Aik and 2 [ C 0 2] values imply
that the advected pore w ater has in this case
higher alkalinity and total dissolved carbonate.
This is in agreem ent with pore waters characteris­
tics found during the D SD P [4,5]. Thus the al­
k a lin ity -to ta l dissolved carbonate relations show
that the advected pore w ater has indeed higher
concentrations than local seawater. However, the
uptake of carbonate and ions (such as C a 2*) for
m etabolism purpose by the clams [8.9] makes it
difficult to deduce m ore accurately the properties
of the advected pore water.
The n itrate and silica concentrations are varia­
ble and some are not in the ranges expected for
deep-sea w ater [10]. Sam ples H Y 2 and 11 w'here
the degradation of organic m atter was lim ited on
the basis of the A lk -2 [C 0 2] relation have low
n itrate and low silica concentrations; sam ples HYI
and 5, where this relation is m ore extensive, have
higher n itrate and silica concentration, as ex­
pected. The sam ples associated with the biological
com m unities have very variable n itrate concentra­
tions and higher silica concentrations (except
HY7). Since we have discovered that some silicate
form ation was associated with the clam com m uni­
ties of the Tenryu C anyon it is, however, difficult
to interpret this d ata (see further section). O ne can
only argue th at the advection of pore w ater should
bring high silica concentrations, which will favor
silicate neogenesis.
The redox potential properties of the waters
associated with clams can be com puted from the
two redox couples C 0 2/ C H 4 and S 0 4 “ / H S .
T he first one gives Eh = —0.260 V and the second
is in the range Eh = —0.250 V to - 0 .2 8 0 V. The
N 2/ N H 4 couple cannot be assessed due to lack of
N H 4 data. However, with N H 4 in the 10 5 M
range and N 2 in equilibrium with the air the
corresponding Eh should be about - 0.260 V. This
is an indication that the N 2/ N H 4 may also be
active in the vicinity of the clams, which would be
in agreem ent with N 2 uptake by the clam s [8].
The trace elem ent geochem istry is difficult to
decipher. Several causes may affect their con­
centrations and they are very sensitive to the
biological activity [8,9]. Also pollution due to the
sam pling tools and to the subm ersible is affecting
num erous metals. We have only given the results
for the trace metals for which this last factor is
negligible. As expected near the sedim ent-w ater
interface, the concentrations of iron, copper and
m anganese are m uch higher than in open ocean
water. There is, however, no clear trend appearing
for these metals. In the samples where sulfate
depletion was observed and where hydrogen
sulfide was produced, the iron concentrations are
in the range expected for equilibrium with greigite
(F e 3S4) or pyrite (FeS2) [11], In the same samples
the concentrations of copper should be controlled
by sulfide form ation as well. This is in agreem ent
with the finding of m etal sulfides on the corre­
sponding filters (see further section below).
The concentrations of m olybdenum are in the
range expected for ocean w ater near the sea bottom -w ater interface [12], At the m easured con­
centrations of hydrogen sulfide ( ^ 1 0 7 M) and
Eh conditions discussed above, the range of
m olybdenum concentrations correspond to equi­
librium between M o 0 4 “ and M oS2.
Barium concentrations are in the range found
in deep-ocean w aters from near the sedim ent in­
terface [13], The ion activity product (I.A.P.) of
barium sulfate o f the w ater samples, calculated
with correction as given by C hurch [13], is quite
variable: from 1.8 X 10“ 11 to 10.5 X 10“ " . Since
the sulfate concentration is alm ost constant in the
sam ples where no activity from vent organism s
was noticed, the variations in I.A.P. correspond to
local variations of the barium concentration. Thus
we find that the barium concentrations in the free
bottom waters are decreasing from the northern
K urile w aters tow ards the south. These points are
illustrated in Fig. 4, in a plot of I.A.P. versus Ba.
The afore-m entioned “ non-biological” sam ples are
linearly correlated in the I.A.P.-Ba plot (Fig. 4).
T he sam ples where sulfate depletion was noticed
are well off this correlation. This can be due to
several causes discussed in the following. T heir
positions suggest that deposition of B aS04 may
have occurred in several cases. A simple mixing
model between local seaw ater and an unknow n
advected pore w ater can be com puted for both Ba
and S 0 4. The com putations can be done by con­
sidering the constraints im posed on the mixing
ratio of an advected pore w ater with local deep
free w ater due to the decrease of sulfate, which
lead to lim iting conditions for Ba, S 0 4, and I.A.P.
/ « / \ 3,4
900 B a [ n M )
Fig. 5. Ion activity prod u ct (l.A .P .) for B a S 0 4 versus barium
co n c en tr a tio n for p o ssib le ad vected p ore w ater end-m em ber.
T h e nu m b ered stars are the K aik o sam p les. SW T is the c o n d i­
Ba (nM)
tion co rresp o n d in g to th e T en ryu C a n y o n area, S W K is that
c o rresp o n d in g to the K ash im a S ea m o u n t area, an d S W K T is
F ig. 4. Ion activ ity p rod u ct (I.A .P .) for B a S 0 4 versus barium
co n c en tr a tio n in K a ik o sam ples. T h e I.A .P . sca le is m u ltip lied
that co rresp o n d in g to the K urile T rench. T h e nu m bered areas
c orresp on d to the c o m p u ted I.A .P . and Ba c o n d itio n s for the
b y IO 11. T h e d a sh ed lin e corresp o n d s to th e relation b e tw een
p ore w ater en d -m em b ers. T h e d ash ed lin e corresp on d s to the
th e “ n o n -b io lo g ic a l” sa m p les (H Y I , 2. 5, 11, 12, 13, 18, 19).
I.A .P .-B a relation for th e “ n o n -b io lo g ic a l” K aik o sam p les (see
T h e rectan gle w ith the sa m p le nu m ber corresp o n d s to the
Fig. 4).
un certainty. T h e black rectan gles are th o se o f the o p e n ocean
b o tto m w ater sam ples.
for the advected pore w ater end-m em ber. The
results of these calculations are given in Fig. 5.
F o r sam ples HY3, 4, 6, 9 from the T enryu C anyon,
H Y 14 from the K ashim a Seam ount and HY15
from the K urile T rench the pore w ater end-m em ­
ber have S 0 4 in the range 0 .3 -0 .6 X I O '2 M, Ba
in the range 300-700 nM in the Tenryu and
700-1000 nM in the K ashim a and K urile areas.
T he low sulfate values are in agreem ent w ith pore
w ater com positions found in the D SD P sam ples
from the sam e areas [4,5]. If we consider this
end-m em ber it im plies that the passage of the
advected w aters through the clam colonies in the
Tenryu C anyon have led to B aS 0 4 deposition
(seen from sam ples HY7, 8). This is shown by the
lower Ba contents in HY7 and HY8 as com pared
to local deep w aters (H Y 2 and H Y5) and the
com puted com position of the deep com ponent
(Fig. 5). In sam ple H Y10 redissolution of B aS 0 4
m ay have taken place next to the clam colonies. In
the case of the K urile T rench there are two possi­
bilities: (1) redissolution of barium sulfate in the
vicinity of the colonies (seen from H Y I6) or (2)
deposition of B aS 0 4 linked to clam m etabolism
(seen in this case from HY15). T hus one should
expect a barium anom aly in connection with the
biological com m unities. This is w hat we have
found upon exam ination of the biological sam ples
and the sedim ents [9],
A last discussed feature of the deep-w ater sam ­
ples is their oxygen isotope and D / H characteris­
tics (Fig. 6). We will only discuss the N ankai area
where several sam ples are available. T he ô 180 are
in the range expected for the Pacific deep bottom
waters. If mixing had occurred with w aters de­
rived from the upper part of the sedim entary
colum n one should expect a slight shift tow ards
m ore heavy oxygen isotope values as found during
D S D P Leg 87 [14], In the N ankai T rough area,
w here eight sam ples were obtained, the shift is
rath er tow ards m ore light oxygen values which
m ay reflect mixing with w aters originated from
deeper in the sedim ent colum n as expected from
the light hydrocarbons data. A puzzling effect is
an enrichm ent in deuterium correlated with the
decrease of 0 180 (called “ N ankai area mixing
line” in Fig. 6). Since this shift is larger in sam ples
a n d of the biological activity o f the vent c o m m u n ­
ity. Some o f the conclusions are s u p p o r te d by
authigenic m in erals observed on the filters from
the syringes a n d the s u p e rn a ta n t waters.
í; u«
5. M inerals found on the filters
T h e overall results from exam ining the filters
by optical m icroscopy, S E M coupled energy
an a lyse r ( J E O L / T R A C O R ) an d X-ray m ic ro d if­
fraction. are sum m arize d in T ab le 6. M ost of the
m inerals are those which can also be identified in
the se dim e nts [9], a n d they are derived from the
w e a thering of the J a p a n e se islands an d from b io ­
logical se d im e ntatio n. Sulfides are associated with
the sam ples taken directly above the biological
c o m m u n ities (H Y 4 . 8, 15). T h e sulfides identified
are mostly pyrite, iron sulfides with som e small
perc en ta g e of Z n a n d C u (Fig. 7a. b), p y rrhotite
(Fig. 7c); a single crystal of M oS , has also been
identified. T hese findings are in agre em e nt with
the w ater chemistry. In places som e oxide ov er­
gro w th s are found (Fig. 7d)). C a lc ium c a rb o n ate s
with very different m orphologies were also id e n ti­
fied. Several co rresp o n d to spicules from c o m ­
p o u n d ascidians [15J. A scidians were indeed id e n ­
tified on the shells of clanis [8] a n d u p o n e x a m in a ­
tion of p h o to g r a p h s taken from the “ N a u tile ”
[16]. C a lc ium c a r b o n a te s fra m b o id s were also
fo u n d (Fig. 7e). they are alm ost p u re calcium
c a r b o n a te a n d a thin organic m a tte r veil is left
Fig. 6. 5 D - 6 1!iO relationship in fluid sa m p les (tita n iu m syringe).
D a ta in perm il versus SM O W . T h e m eteoric w ater lin e is given
as w ell as the “ m ixing line" for the N a n k a i T rough sam ples.
T h e area o f the rectangle for each sa m p le corresp o n d s to the
m easurem en t un certainties. T h e sa m p le nu m ber is given near
the rectangle. O n the 6 I80 scale “ P .D .B .W ." corresp o n d s to
the range for P acific d eep b o tto m w aters and “ D S D P 8 7 " to
the range fou nd in the upper sed im en ts from the N an k ai
T rough investigated du rin g D S D P Leg 87 [14).
w here m e th a n e was in larger c o n c e n tra tio n (H Y 4 )
this m ay be characteristic of the p o re fluid. It m ay
also be a sa m pling artefact, however, n o t ex p la in a ­
Thus, the geochem istry of the w ater sam ples
also illustrates the in p u t of ad vected po re w'aters
M inerals observed by m icroscopic exam ination o f the filters
Sam ple
P rism atic rutile
Black op aq u es (oxides)
W h ite m ica
G reen o r black m ica
G reen silicates
Q u artz fragm ents
D iato m (com plete o r fragm ents)
G rap h ite and grap h ito id s
Iro n oxyhydroxides
W h ite clay m ineral
M etal fragm ents
(from subm arine
o r sam pling devices)
+ + +
+ +
H Y 10
H Y 18
+ +
- = absent; + = frequent; + + = a b u n d an t; + + + = very ab u n d an t.
+ +
+ +
H Y 16
F ig. 7. SHM p h o to g ra p h s o f particles fo u n d o n the filter c o rresp o n d in g to H Y 4. T h e scale bar is given in |im in the low er right angle
o f each view , (a) C luster o f sm all pyrite crystals, (b ) C u b e o f pyrite covered by a Z n -F e su lfid e, (c ) P yrrhotite crystal, (d ) Sm all
m agn etite crystal (w ith o u t T i) overgrow th s partially c o v erin g a C a -F e -M g alu m in o -silic a te crystal, (e) C alciu m ca rb o n a te (w ith ou t
M g, M n. F e) fram b oid . (f) C a lciu m ca rb o n a te cry sta ls sta ck ed togeth er (traces o f F e, M g and M n).
upon acid dissolution. They are probably of bio­
logical origin, however, not identified. Stacks of
calcium carbonate crystals, containing traces of
Fe, Mg. M n were also found (Fig. 7f). They are
probably of inorganic origin. X -ray m icrodiffrac­
tion revealed that vaterite, calcium carbonate
m onohydrate as well as aragonite and calcite are
present. Calcium carbonate hydrate suggests that
it may be derived from m ethanehydrate yielding
C a C 0 3, 6 H 20 (ikaite) which then decom poses at
low er pressure [17], Such u nstable calcium
carbonates have been identified in the deform ed
sedim ents from the N ankai T rough during D SD P
Leg 87 [18] and in the vents o ff O regon coast [1]
and are the precursors of pseudom orphes in an ­
cient subduction zones of Jap an and Oregon. Thus
the presence of calcium carbonate hydrates seems
to be related to the advection of m ethane.
In the Tenryu C anyon biological vent com m un­
ity a white clay m ineral w ith a fluffy appearance
was also abu n d an t. A lthough difficult to isolate,
X-ray diffraction suggests that it m aybe a sm ec­
tite. This implies the presence of silica in much
larger concentration than in seaw ater and hence
its form ation may be related to advection of pore
6. Discussion
A lthough the sam ples we have collected during
the K aiko legs result from a com plex m ixing b e ­
tween am bient open ocean deep-sea w ater and
pore w ater from the sedim ent, there are good
geochemical d ata which show th at p art of these
fluids have com ponents which may have a deep
origin. The association of the biological com m uni­
ties with these geochem ical anom alies and with
deep geological structures (from seismic data) also
suggest a local inp u t of chem icals and n utrients by
fluid venting. H ere we wish to discuss further the
possibility of a deep com ponent input to surface
sedim ents and to assess the characteristics of pore
fluid m igration with special em phasis on the
N ankai Trough area where d ata from D SD P are
The m ethane concentrations, and the C , / ( C 2 +
C \) ratios, of the samples raise several questions:
(1) w hat is the origin of the hydrocarbons?; (2)
w hat is the background concentration of m ethane
in deep-sea w ater from subduction areas? This last
question is difficult to answer straightforw ardly
since what we thought was a background sam ple
(H Y 3) may be associated with a structure favora­
ble to fluid venting according to the bottom p h o ­
tographs and its sulfate depletion. M ethane con­
centration in the same range as HY3 (50-70 nl/1)
were found in w ater sam ples taken several miles
from the subduction of the Oregon m argin [1], It
is possible th at background concentration in sub­
duction area are several tim es larger than the
expected 10 nl/1. In HY3 the C , / ( C 2 + C 3) ratio
is very low and suggests that the hydrocarbon
concentrations are influenced by therm ogenic
m ethane input [19]. T he concentration of light
hydrocarbons introduced by advection of pore
fluid may, however, have been reduced by oxida­
tion in the upper p art of the sedim ent colum n
[8,20], W ith the help of the constraint due to
sulfate depletions, the m ethane concentration that
can be deduced for the advected com ponent in
HY3 is about 600 nl/1, reflecting the low con­
centration due to oxidation. In the pore w aters of
sedim ents where advection of therm ogenic hydro­
carbons was found, however, not from the Oregon
m argin, concentrations in the range 10 100 x IO1
nl/1 were obtained [10], In the deep-w ater sam ples
clearly associated with biological activity (HY4. 7.
14, 15, 16) the m ethane concentrations are in the
range of HY3 or larger. One should expect this
since dilution of advected fluid and increase of
consum ption of light hydrocarbons have in­
fluenced this factor. T he C l/ ( C 2 + C 3) ratio is
even lower than in HY3, showing clearly the input
of a therm ogenic com ponent. If HY3 represents a
background sam ple for the hydrocarbons we can
solve a sim ple mixing model with an advected
end-m em ber for C j, C 2 and C 3. In the case of
HY4, 15, 16, the corresponding advected endm em ber should be C , = 4 0 0 nl/1 and C , / ( C 2 +
C 3) = 2. These low concentrations are due to the
oxidation o f the light hydrocarbons. If we take the
pore w ater value of C , = 10-100 X IO3 nl/1, the
am ount oxidized is larger than 95-99.5% of the
Part of the m ethane encountered in the Kaiko
sam ple can be derived from biogenic activity.
However, in this case it is difficult to explain the
C 2 and C 3 concentrations. The presence of ther­
mogenic hydrocarbons in the sam ples should
rather be used as an argum ent related to the
plum bing of the advection of pore water. Since the
clam colonies are located above thrusting struc­
ture, the therm ogenic hydrocarbons are evidence
for the deep p enetration of these structures.
The light hydrocarbons are m ost probably of
therm ogenic origin. T he biochem ical studies have
shown that the isotopic com position of the
m ethane should be 0 13C = —50 to —30%o which
is also characteristic of therm ogenic m ethane [8],
T herm ogenic hydrocarbons are produced at tem ­
peratures higher than 5 0 -6 0 ° C. T he therm al
gradient found in the Jap an T rench is in the range
of 1 5 - 3 6 ° C /k m [21,22]. This yields depths in the
range 1 .5 -3 km for the generation of the light
hydrocarbons. This d ep th range com pares favor­
ably w ith th at deduced from the m e th a n e /e th a n e
ratio versus d ep th diagram obtained during D SD P
Hole 440. A ccording to that, the location of the
source of therm ogenic hydrocarbons should be
around 2.5 km [3], It is also in the sam e depth
range as the vertical projection of the thrust faults
found in the N ankai accretionary prism [23], and
for dew atering in accretionary prism [24].
The processes of m igration of w ater through
the prism during dew atering are not well known.
The location of vents in close association with
structure related to thru st faults as at the base of
the Tenryu C anyon deep-sea fan suggests it occurs
as shear dew atering, i.e. due to fracture perm eabil­
ity. T he location of vents where erosional features
allow the th ru st faults to reach the sedim ent-w ater
interface also suggests th at the advection can be
easily stopped o r diverted if buried under sedi­
m ent slump.
A sim ple m odelization of the rate of advection
of the fluid using a tw o-dim ension steady-state
m odel has been attem p ted [25]. It was found that
it is not possible to sustain steady-state advection
and th at the rate of advection should be at least
5000 c m /y r [25], In this case the tem perature
anom alies observed can be o btained if cooling of
the expelled fluid is effected by a lateral flux of
seaw ater through the sedim ent near the vent loca­
tion [9,25]. T he geochem ical anom alies, such as
sulfate depletions, are also difficult to explain if
we have low advection rates. T he flux in the
subduction areas off Jap an is probably non-stable.
T his is in con trast w ith the stable steady-state
fluxes com puted from the geochemical anom alies
obtained off the coast of O regon [26] and the
corresponding low rates of the pore fluid (1 -1 0
c m /y r) for the case of O regon [26]. Sim ilar low
rates have also been obtained to explain the th er­
mal behavior of the subduction in the northern
Jap an T rench [27].
T he case of fluid venting in the Tenryu C anyon
area is well docum ented and geological as well
geochemical characteristics can be com pared with
sim ilar occurrences in the O regon margin. The
finding of fluid venting and biological com m uni­
ties associated with a thrust in the K ashim a area
(H Y 14) and in the Jap an T rench [29] can be
sim ilarly understood. T he possible occurrence of
fluid venting in the Zenisu Basin (HY 3) and the
K urile T rench scarp (HY15, 16) is m ore difficult
to understand. In the case of the Zenisu Basin
advection of fluid is suggested by the sulfate de­
pletion and the low C , / ( C 2 + C 3) ratio. The
Seabeam data, the seismic data and the possibility
of a thrust structure in the Zenisu Basin [28],
correspond to com pressive stresses which are
favorable for fluid expulsion. A small clam colony
at the apex of a fold above a thrust fault at the
southernm ost deform ation front of the Zenisu area
[28] may be related to this interpretation. In the
K urile area, the tem perature and geochemical
anom alies as well as the presence of clam colonies
m ake a clear cut case for active venting. The
geological setting, however, of the structure which
would allow such venting is less clear. We can
note that the deep biological com m unities were
aligned along N 330° which is along the m ajor
structural direction of the southw est inner wall of
the Kurile Trench. The characteristics of the fluid
venting suggest that this structure m ust have
penetrated deeply.
7. Conclusions
The deep w aters sam pled above biological com ­
m unities in the N ankai T rough and the Japan
Trench areas can be related to fluid venting in
connection with the geological setting which would
favor advection of the fluids yielded by dew ater­
ing. In the upw ard path these fluids are enriched
with molecules and elem ents of deep sedim entary
origin: light hydrocarbons, trace metals. Hydrogen
sulfide is also probably brought in the upper part
of the sedim ent as shown by the dissolved sulfate
depletion and the presence of m etal sulfides in the
samples. H ydrogen sulfide is utilized by bacteria
associated with the gills of the clams from the
biological com m unities [8]. G eochem ical and bio­
geochemical d ata also show that part of the
m ethane is also utilized for nutrient requirem ents.
T he fluid sam ple com positions result from the
m ixture of these different processes and sources.
T he light hydrocarbon d ata suggest the depth
scale of the fluid circulation: 1.5-3 km for the
N ankai area. A model of fluid circulation suggests
a non-steady-state advection rate of at least 5000
c m /y r. The relation of the fluid vents to overpres­
sure along the thrusting plane in the N ankai
T rough and the deep origin of some of the fluid
com ponents give a good base for the use of the
electro kinetic properties of the venting fluids for
prevision of earthquake in this area [30].
The results obtained in this geochemical study
have also im portant im plications concerning the
budget of elem ents in the ocean. For instance, the
budget of m ethane and light hydrocarbons will
probably have to be reassessed. Thus the q u a n tita ­
tive im portance of the prelim inary results pre­
sented here will only be possible when a m ore
com plete overview of the im portance of fluid vent­
ing in subduction areas will be obtained.
vent sites a lon g the su b d u ction z o n e o ff O regon, B iol. Soc.
W ash. Bull. 6. 4 7 5 -4 8 4 . 1986.
3 J.K . W helan and S. Sato. C , - C 5 hyd rocarbons from core
gas p ock ets. D eep Sea D rillin g Project L egs 56 and 57.
Japan T rench transect, in: Initial R eports o f the D e ep iÿea
D rillin g Project 57, pp. 1 3 3 5 -1 3 4 7 . U .S . G overn m en t Print­
in g O ffice. W ash in gton D .C .. 1980.
4 G .W . M oore and J.M . G iesk es. Interaction betw een sed i­
m ent and interstitial w ater near Japan T rench, L eg 57,
D e ep Sea D rillin g Project, in: Initial R eports o f the D eep
Sea D rillin g Project 57, p. 1 2 6 9 -1 2 7 5 . U .S . G overn m en t
Printing O ffice. W ash in gton D .C .. 1980.
5 H. K aw ahata. K. Fujioka and T. Ishikuza, S ed im en ts and
interstitial w ater at Site 582 and 584. the N an k ai T rough
and Japan T rench landw ard slope, in: Initial R eports o f the
D e ep Sea D rillin g Project 87, pp. 8 6 5 -8 7 5 , U .S . G o v ern ­
m ent Printing O ffice. W ash in gton . D C .. 1986.
6 M . S ch oell, R ecen t ad van ces in p etroleum iso to p e g e o ch em ­
istry. Org. G e o ch em . 6, 645 -6 6 3 . 1984.
7 G .E . C layp ool. A .K . V uletich and K .A . K ven vold en . I so ­
top ic co m p o sitio n o f interstitial flu id s in sed im en t o f the
N an k ai T rough, D e ep Sea D rillin g Project Leg 87. in:
Initial R ep orts o f the D e ep Sea D rillin g Project 87. pp.
8 5 7 -8 6 0 . U .S . G overn m en t Printing O ffice. W ashington
D C .. 1986.
8 J. B oulègue, b .L . B en ed en i, D . D ron , A . M a rio n i and R.
L étolle. G eo ch em ica l and b io g e o c h e m ic a l o b serv a tio n s on
the b io lo g ica l co m m u n ities associated w ith fluid ven tin g in
the N an k ai T rou gh and Japan T rench su b d u ction zones,
Farth Planet. Sei. Lett. 83. 3 4 3 -3 5 5 . 1983 (th is issue).
9 D
D ron. J. B oulègue. A . T aira and C. R angin . G e o c h e m ­
istry o f the T enryu C an yon d eep -sea fan b io lo g ica l c o m ­
m u nity (K a ik o ), Earth Planet. Sei. L ett. 83. 3 5 6 -3 6 2 . 1983
(th is issue).
10 W .S. Broecker and T .H . Peng, Tracers in the Sea. 69 0 pp.
We acknowledge the invaluable assistance of
the “ N autile” group and the help of R /V “ N ad ir”
crew members. W e thank the K aiko Planning
C om m ittee for giving access to shipboard sam ­
pling and H. Bougault, L. Floury, C. Levèque and
M. Sibuet, all from IF R E M E R , for help with
instrum entation. Part of the analytical study was
done by Dr. A.M. de K ersabiec and F. Vidot. This
study was partly supported by grants from
P IR O C E A N /C N R S and IF R E M E R . We thank
Dr. E. Suess for very helpful discussion on fluids
in subduction areas.
Hldigio Press. N e w Y ork. N Y .. 1980.
11 J. B oulègue. T race m etals (F e, C u. Z n. C d) in anoxjc
en viron m en ts, in: Trace M etals in Sea W ater, N A T O C onf.
Ser., M ar. Sei., pp. 5 6 3 -5 7 7 , Plen um Press. N e w Y ork,
N Y .. 1983.
12 K .K . B ertine. T h e d e p o sitio n o f m olyb d en u m in an oxic
w aters. M ar. C hem . 1. 4 3 - 5 3 , 1973.
13 T .M . C hurch. M arine Barite. 100 pp, P h .D . Thesis. U n iv e r ­
sity o f C alifornia at San D ie g o , 1970.
14 Y . M atsuhisa and R. M a tsu m o lo . O xygen iso to p e ratios o f
interstitial w aters from the N a n k a i T rough and the Japan
T rench. L eg 87, in: Initial R eports o f the D eep Sea D rillin g
Project 57, pp. 8 5 3 -8 5 6 . U .S . G overn m en t Printing O ffiee.
W ash in gton D C .. 1986.
15 B. K niprath and F. L afargue, Sp icu le form ation in the
D id em n id ae, in: T h e M ech an ism o f B iom in eralization in
A nim als and Plants, M . O m ori and N . W atab e. ed s.. T okai
U niversity Press. T o k y o . 1980.
16 L. I.aubier and S. O hta. D eep b iological c om m u n ities in the
1 L. D . K uim . F.. Suess, J.C. M oore. B.T. L ew is, S .D . R itger.
su b d u ction z o n e o f Japan from b o tto m p h otograp h s taken
D C . K adk o, T .M . T hornburg, R .W . b m b lev . W .D . R ugh.
du rin g " N a u tile " d ives in the K aik o project. Earth Planet.
G .J. M assoth. M .G . L angseth. G .R . C ochrane and R.L
S cam m an. O regon su b d u ctio n zone: ventin g, fauna and
carb on ates. Scien ce 231, 5 6 1 -5 6 6 . 1986.
2 F.. Suess. B. C arson. S.D . R itger. J.C. M oore. M .J. Jones.
L .D K uim and G .R . C ochrane. B iological c o m m u n ities at
Sei. Lett. 83. 3 2 9 -3 4 2 . 1987 (th is issue).
17 b. Suess. W Balzer, K .F. H esse. P.J. M uller. C .A . U ngerer
an d G . W efer. C alciu m carb on ate hexahyd rate from organic
sed im en ts
A n tarctic
glen d on ites. Scien ce 216. 1 1 2 8 -1 1 3 1 . 1984.
18 C .L . Stein and A .J. Sm ith, A u th ig e n ic ca rb o n a te n o d u les in
25 X . L e P ich ón , S. L allem an t and S. L allem an d , T ec to n ic
the N a n k a i T rou gh , Site 583, in: Initial R ep o rts o f the
c o n te x t o f flu id v e n tin g a lo n g Jap an ese trenches, A bstr.,
D e ep Sea D r illin g Project 87, pp. 6 5 9 -6 6 6 , U .S . G o v e r n ­
Int. K aik o C o n f. on S u b d u ction Z o n e s, N o v e m b e r 1 0 -1 5 ,
m en t P rin ting O ffice, W a sh in g to n D .C ., 1986.
T o k y o 1986, 2 - 1 , p. 24.
19 M .J. W hiticar, E. S u ess an d H. W chner, T h erm o g en ic h y ­
d ro ca rb o n s in surface sed im en ts o f the B ran sfield strait,
26 M .W . H an and E. S u ess, C h em ical evid e n c e o f pore v en tin g
in th e O regon su b d u ctio n zo n e, E O S 57, 1219, 1986.
A nta rctica P eninsu la, N a tu r e 314, 8 7 - 9 0 , 1985.
27 B .H . R eck, E ffect o f w ater flo w from d ew aterin g subd u cted
2 0 R .Y . Stanier, M . D o u d o r o ff and E .A . A d elb erg , G en eral
M icro b io lo g y , M a cM illa n , L o n d o n , 1971.
sed im en ts o n therm al grad ien ts in the N o rth ea st Japan
accretion ary prism , EO S 67, 379, 1986.
21 Sh ipboard S cie n tific Party, Sites 4 3 8 a n d 4 39: Japan D eep
28 X . Le P ich ón , T. Iiyam am a, J. B oulègue, J. C harvet, M.
Sea T errace, L eg 57, in: Initial R ep o rts o f the D e ep Sea
Faure, K. K ano, S. L allem ant, H. O kad a, C . R an gin . A.
D rillin g Project 5 6 / 5 7 pp. 2 3 - 1 9 1 . U .S . G o v ern m en t P rin t­
T aira, T. U rab e an d S. U y ed a , N a n k a i T rou gh and Z en isu
in g O ffice, W a sh in g to n D C .. 1980.
R idge: a d ee p -se a su b m ersib le survey. Earthy Planet. Sei.
22 Sh ipboard S cie n tific Party. S ites 440: Japan T ren ch m id ­
slo p e T errace. L eg 57, in: Initial R ep o rts o f the D e ep Sea
L ett. 83. 2 8 5 -2 9 9 , 1987 (th is issu e).
29 J.P. C ad et, K. K o b a y a sh i, J. A u b o u in , J. B oulègue, J.
D rillin g Project 5 6 / 5 7 , pp. 2 2 5 -3 1 7 . U .S . G o v ern m en t
D u b o is, H . H o tta , T . Ishii, L. Jolivet,
Printing O ffic e, W a sh in g to n D .C ., 1980.
L allem an d , N . N iitsu m a and H. S h im am u ra. D e ep scie n ­
23 C .J. Bray and D E. K arig, P orosity o f sed im en ts in accre­
prism s
so m e
im p lic a tio n s
d ew aterin g
p ro cesses. J. G eo p h y s. R es. 90, 7 6 8 -7 7 8 , 1985.
24 T .H . S h ip ley and G .F . M oore, S ed im en t accretion , su b d u c ­
tion and d ew a terin g at the base o f the trench slo p e o ff
C o sta R ica a seism ic reflectio n view o f th e d é c o lle m e n t. J.
G e o p h y s. R es. 91, 2 0 1 9 -2 0 2 8 . 1986.
K. K on ish i. S.
tific d iv es in Japan an d K uril T renches, Earth Planet. Sei.
83, 3 1 3 -3 2 8 , 1987 (th is issu e).
30 J. B oulègue, X . Le P ich ón and T. Iiyam a, P révision des
trem b lem en ts de terre d a n s la région d e T ok ai (Jap on ),
C .R . A cad . Sei. Paris, Ser. II 301 1 2 1 7 -1 2 1 9 , 1985.