Eos, Vol. 82, No.

6, February 6, 2001

VOLUME 82 NUMBER 6

FEBRUARY 6, 2001

EOS, TRANSACTIONS, AMERICAN GEOPHYSICAL UNION PAGES 65-80

enriched in plagioclase—for example, Sr and

Evidence Suggests Slab Melting Eu—typically do not find "homes" in garnet
eclogite and go into the melt.This causes an

in Arc Magmas unusual-looking product from a g e o c h e m i c a l

perspective. Slab melts are enriched in Sr,
Na,and Eu but depleted i n Y and the HREE,
which is exactly what adakites show: geo­
PAGES 6 5 , 6 8 - 6 9 of adakite research over the past decade, chemical features atypical of volcanic arcs.
emphasizing the significance of slab melts The fact that adakites are associated with
Most recent geology textbooks state that as metasomatizing agents in active arcs, and the subduction of young crust bolstered the
subduction-related volcanism is due to the support the contention that the Archean slab-melt hypothesis.
melting of the down-going lithosphere. How­ continental crust was derived from the melt­ Adakites typically have S i 0 > 56 wt. %, A 1 0
2 2 3

ever, for the last 30 years, few in the field have ing of the subducted crust. > 15 wt. %, and N a 0 > 3.5 wt. % and are char­
2

seriously believed that the subducting slab is acterized petrographically by the presence
the source of arc basalts.The accepted hypothe­ of abundant plagioclase and amphibole
Defining Adakites
sis involves melting of the mantle wedge above phenocrysts.Trace elements worth noting
the slab via hydrous fluids produced during By the mid-1980s, Kay [1978] and others are Sr ( > 400 ppm),Y ( 1 8 ppm), HREE (for
the transition of the subducting basalt from had worked out the major and trace element example,Yb 1.9 ppm),positive Eu anomalies
amphibolite to eclogite.The parental basalts dif­ compositions of slab melts based on geochemi- and positive Sr anomalies, and Sr/Y and
ferentiate primarily through crystal fractiona­ cal modeling, and their work was eventually La/Yb ratios of 40 and 20, respectively [Defant
tion, magma mixing, and differentiation at the verified experimentally [Rapp et al, 1999, and and Drummond, 1990]. Adakites also contain
Mohorovicic discontinuity into andesites and references therein] .The combined work clearly the typical arc signature—depletion of Nb
dacites as they ascend; the basalts are too showed that garnet was stable in the source at and Ta—and, in most cases, MORB-like Sr
dense to rise through the lower continental the pressure and temperature of melting; a and Nd isotope signatures.
crust. This explains the relative abundance of garnet eclogite is the restite. Garnet preferen­ One of the best examples of the juxtaposition
differentiated rocks in arcs. tially incorporates the heavy rare earth elements of adakites with normal arc basalts, andesites,
Geochemically distinct andesitic lavas (HREE) and Y Furthermore, as amphibolite is dacites, and rhyolites (BADR) o c c u r s along
o c c u r at Adak Island, Aleutian arc, and Kay transformed into garnet eclogite during sub­ the coast of Kamchatka, Russia. In southern
[1978] proposed that they were derived from duction, plagioclase breaks down. Elements Kamchatka, subduction of extremely old
the melting of the subducted slab. Martin
[1999; and references therein] subsequently
2.6
noted that Archean tonalites were similar in
composition to Chilean adakitic volcanics, 2.4
suggesting that the siliceous Archean crust Backening xenoMhs, Kamchatka, Russia
was derived from the melting of subducted
o c e a n i c lithosphere.
At the beginning of the last decade, Defant
and Drummond [1990] and Drummond and
Defant [1990] proposed that the subducting
slab can melt when the crust is relatively young
and therefore elevated in temperature.They
surmised, based on experimental and g e o -
thermal models (see Peacock et ai, 1994, and
references therein; Rapp et ai [1999]), that
the products would b e andesitic and dacitic
in composition with rather unusual geochemi-
cal imprints.They called these rocks adakites,
noting that Kay [1978] had b e e n the first to
attribute a slab melt scenario to similar rocks 0.4
at Adak Island.They identified approximately
ten locations around the world where the 0.2 K Depleted Mantle
subduction of young crust is associated with
o.o
adakites.They also predicted several other 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
localities that had not been studied where they
Ta
thought adakites would be found; for example,
in southwestern Japan based on the subduc­ Fig. 1. MgO versus Si0 in adakites compared
2 with the fields for experimental melts and Archean
tion of young crust. Here we review the progress trondhjemites, tonalites, and dacites (TTD) and glass inclusions in sub-arc mantle xenoliths.
 Eos,Vol. 82, No. 6, February 6, 2001
enriched in Mg and Piip-type adakites are
generated (that is, HMA). Boninites, the ultra­
high Mg andesites, appear to have the least
adakite c o m p o n e n t involved and probably
reflect more hydrous conditions than during
the generation of Piip-type magmas.
Sigmarsson et al. [1998] found that young
adakites from the southern Austral Andean
Volcanic Zone fall to the left of the Th-U dis­
equilibrium equiline on a graph of 23077^2327^ _
v e r

Adakites 238 232

sus U/ Th, which suggests a partial melting
origin from the subducted slab and not a dif­
ferentiation process from basalts derived from
a hydrous metasomatized arc mantle.They
found that the data were compatible with 2 0 %
Experimental data field ^ 7 partial melting of the subducted slab and noted
(Rappetal., 1999)
and Archean TTD field
that garnet in the residual eclogitic source
during partial melting preferentially partitions U
relative to Th. Mount St. Helens adakites fall to
the left of the equiline, which also supports the
Arc xenolith glass inclusion field
contention that they are derived from the slab.
Similar results have b e e n recorded using
50 60
Li isotopes from volcanism in western Panama,
Si0 (wt. % )
2 which can be divided into an old group (6-13
Ma) typical of normal arc volcanism and a
young group ( < 5 Ma) representing adakites.
Fig. 2. Variations in Ta and La in mantle wedge xenoliths from the Kamchatka arc. La is typically
The subduction parameters changed in the
carried by hydrous fluids while Ta is not. The fact that they correlate suggests that hydrous
region about 5 Ma from an old slab to a
metasomatism is not the only process at work in the arc mantle.
young and hot slab.The older group has rela­
7
tively high B/Be and (5 Li isotopic ratios
between +4.7 to +5.6 % o while the young
7
o c e a n i c lithosphere is associated with typical as opposed to the deeper melting that is group—the adakites—has low B/Be and <5 Li
arc volcanism believed to b e derived from producing magmatism at Mount Adams. isotopic ratios between +1.4 and +4.2 % .01d
0

hydrous metasomatism of the overlying mantle Two types of high-Mg andesites o c c u r in the group arc volcanism is associated with the
wedge. But in Kamchatka north of the Aleutian western Aleutian Komandorsky Island region hydrous metasomatism of the mantle wedge
7
islands, the crust off the coast is young and of Alaska and Russia: an Adak and a Piip type (high B/Be and (5 Li) while young and hot
the volcanism changes to adakites. [e.g., Yogodzinski et al, 1995,and references slabs tend to devolatilize their hydrous compo­
Bebout et al. [1999; see references therein] therein].The high-Mg found in s o m e adakites nent prior to melting. Although adakites are
found evidence for the involvement of partial results from limited interaction of adakites with hydrous, they are lower in elements typically
melts of the subducted slab and sediments peridotite in the mantle wedge before they carried by hydrous fluids, which explains the
7
associated with the hydrous c o m p o n e n t in are extruded; this is the Adak type.The fact low B/Be and (5 Li found in the young group.
the Catalina schist that represents an exposed that adakites have higher MgO and Cr and Ni
subduction zone. Pegmatites with the chemi­ concentrations compared to experimental
Additional Ways to Produce Adakites
cal composition of adakites were probably melt compositions and adakitic inclusions
derived from the partial melting of the slab. in sub-arc mantle xenoliths supports this Research over the last d e c a d e has shown
The slab exists in the Catalina schist as garnet contention (Figure l).In contrast,the Piip-type decidedly that the subduction of young and
eclogitic restites. Based on their work, there is magnesium andesites are in equilibrium hot slab is not the only way to produce adakites.
little doubt that adakitic melts can be products with olivine, which suggests that they formed Six tectonic processes also appear to b e able
of slab melting within volcanic arcs. from direct melting of the mantle wedge after to produce adakites.
A number of new adakite localities associated metasomatism by slab melts. The Setouchi • Remnant slabs: Several authors have pos­
with the subduction of young crust have belt of southwest Japan is also believed to tulated that adakites c a n b e generated from a
b e e n reported over the last d e c a d e , includ­ be a Piip-type high-Mg andesite. remnant of previously subducted slab incor­
ing Cerro Pampa in the Andes; Talamancas, Several studies have suggested that both an porated in the mantle that b e c o m e s elevated
Costa Rica; the Mexican Trans-Volcanic Belt; adakite and hydrous slab fluid are involved in in temperature by the surrounding mantle; for
and southwest Japan. the formation of boninites—extremely high example, Evia, G r e e c e and Isla San Esteban,
The Cascade arc is associated with o n e of Mg andesites that are largely partial melts of Gulf of California.
the hottest subducting slabs in the world. hot and shallow mantle wedge—which attests • Oblique or fast ( 8 - 1 0 cm/yr) subduction:
We suggest that the Cascade arc shows abun­ to the involvement of the subduction of young Within s o m e tectonic environments, older
dant evidence of slab melting. One example, crust.The chemical variations in pristine adakites, o c e a n i c lithosphere subducts obliquely or at
Mount St. Helens, erupts predominantly adakites Piip and Adak-type high-Mg andesites, and elevated rates which leads to adakite produc­
and sits in front of the Cascade arc [Defant boninites suggest that pristine adakites only tion. Shear stress related to fast subduction or
and Drummond, 1993]. In contrast, Mount Adams reach the surface if they follow pathways oblique subduction c a n increase the tempera­
erupts regular arc basalts, andesites, and dacites, through the mantle that have already b e e n ture of the upper part of the subducting slab.
but is located along the strike of the arc 50 extensively reacted with by previous adakite The Aleutians and Komondorsky Island and
km behind Mount St. Helens. Experimental interaction.This way no further reactions c a n eastern Mindanao,The Philippines, are proba­
work has shown that slab-melting processes affect the adakite that reaches the surface. bly the best examples.
should o c c u r at shallower depths than the Otherwise, the magmas primarily pick up Mg • Arc-arc collision: Adakites have b e e n
normal dehydration melting of the mantle when they react with the mantle forming found in two major arc collisional zones—the
during regular arc volcanism.The frontal Adak-type magmas (Figure 1). Arid Hills and northern coast of Papua New
position of Mount St. Helens may be related to If the reaction is so intense that the mantle Guinea and central Mindanao, The
shallow melting of the subducting lithosphere is actually melted, the melt b e c o m e s intensely Philippines. It is important to emphasize that
 Eos, Vol. 82, No. 6, February 6, 2001
b e c a u s e the age of the crust can no longer b e Kepezhinskas et al. [1996] found HFSE-rich mobile in hydrous fluids.They c o n t e n d e d
determined in these regions since it has been phases along the contact between the felsic that a slab melt component—preferential
subducted, we do not know the thermal veins and the mantle, and presumed that the partitioning of Th c o m p a r e d to U into the
history The adakites may have been generated reaction preferentially increased the HFSE melt—must b e involved in arcs to explain this
directly from young crust or the slab could in the Na-metasomatized mantle. Nb and Ta phenomenon.
still b e below the collisional zone, where it actually correlate with incompatible elements Studies of PGE behavior in the sub-arc
has heated to produce adakites. Alternatively, such as La and Sr (Figure 2), which were mantle show that Kamchatka xenoliths form
something specific to the process of arc-arc traditionally believed to b e transported by a continuous trend on a graph of Os versus
collision could generate adakites. More work a hydrous component. Re/Os from high-Os harzburgites through lher-
is needed, particularly in Papua New Guinea, Recent experimental determinations of zolites, pyroxenites, and adakites, which indi­
to better understand the relationship between partition coefficients between silicious melts cates potential mantle-slab melt interaction. It
tectonics and adakite production. and amphibole for Nb and Ta indicate that is difficult to envision how PGE and/or HFSE
• Initiation of subduction: Peacock et al. Nb and Ta are enriched in the mantle during can vary through hydrous metasomatism.
[1994] showed that when the cold slab first silicic melt interaction via amphibole. Subse­ S o m e authors, noting variations in the HFSE
plunges into the hot mantle, the increase in quent melting of this Na-metasomatized m in arc basalts (NEAB), called upon an OIB
the temperature of the subducting slab can antle produces basalts (or high-Mg andesites - mantle component present in arcs. But titanate
produce partial melting. Adakites are probably Piip type) with high HFSE concentrations phases stable in the source during melting are
derived by this process in eastern and south­ similar to NEAB. required to control HFSE variations. No such
ern Mindanao, The Philippines. Our current research focuses on the possi­ phases have ever b e e n discovered in arc man­
• Slab tears: A tear along the subducting slab bility of the e x i s t e n c e of a genetically related tle xenoliths, nor has there ever b e e n an OIB
below Shiveluch volcano, Kamchatka, Russia, v o l c a n i c suite of rocks that would incorpo­ mantle composition found in arc xenoliths. In
has allowed the edge of the slab to b e heated, rate both Adak and Piip adakites, perhaps addition, experimental work has shown that
producing adakites. A similar scenario o c c u r s boninites, NEAB, and the associated mantle titanate phases are not stable during hydrous
where the Lesser Antilles volcanic arc termi­ xenoliths c o m m o n l y found in these rocks. metasomatism or partial melting.
nates below Grenada. We believe that the end We propose the term adakite metasomatic OIBs probably c o m e from a deep mantle
of the Central American arc in Costa Rica may v o l c a n i c series. that has had subducted lithosphere added
reflect this process, but further work is needed. Adakites and NEAB, plus or minus mantle to it. No wonder slab melt metasomatism of
• Flat subduction: Adakites in Ecuador and xenoliths in the NEAB, have b e e n documented the arc mantle might appear to b e similar to
Peru and perhaps in other arcs may b e gener­ in several distinct localities worldwide, includ­ the source of OIBs. But what s o m e authors
ated by flat subduction. During subduction, ing The Philippines, Japan, Kamchatka and suggest is an OIB mantle c o m p o n e n t in arcs
the slab literally flattens and travels horizon­ the far western Aleutian Islands in Alaska, we believe is really a slab melt c o m p o n e n t
tally and c a n heat enough to melt during this Mount St. Helens, and Panama-Costa Rica. and that OIBs have an arc component; that
horizontal motion. Several other localities associated with ridge is, the deeply subducted slab.
or oblique subduction or slab tears may also
contain the adakite metasomatic volcanic
Na Metasomatism and the Arc Mantle Archean Adakites and Adakites
series; namely, Adak and Kanaga Islands
as a Prospecting Tool
Adakites are sometimes associated with Nb- (Aleutians), Grenada (Lesser Antilles), Baja
enriched arc basalts (NEAB).These basalts California and the Trans-Mexican Volcanic The Archean continental crust consists of
might b e derived from a mantle source that Belt (Mexico), and Simberi Island (Papua about 80% trondhjemitic rocks with adakitic
was metasomatized by adakites. Kepzhinskas New Guinea). characteristics. Martin [1999, and references
et al. [1996] looked carefully at NEAB associ­ Evidence in recent years has begun to sug­ therein] showed that these trondhjemites
ated with adakites from Mount St. Helens, gest that slab melts may play a minor role in have g e o c h e m i c a l characteristics much
northern Kamchatka, Panama and Costa Rica, the metasomatism of all arc mantle. S c h i a n o different from modern arc suites. These
and the southern Philippines.These basalts et al. [1995] found that melt inclusions in rocks appear to b e derived from the melting
have Nb/La (HFSE/La) ratios higher than typi­ mantle xenoliths from Batan Island, northern of the subducting slab primarily b e c a u s e the
cal island-arc basalts, and even though NEAB Luzon arc, are similar in composition to subducting lithosphere in the Archean was,
have high absolute HFSE concentrations, they adakites, which suggests that slab melts must on average, much younger than it is today
also have typical arc depletions in HFSE rela­ be a c o m p o n e n t during metasomatism of the Why? B e c a u s e the Archean mantle was much
tive to K and La. NEAB are also alkaline or arc mantle below the Luzon arc. The entire hotter than the mantle is today and probably
transitional alkaline compared with typical northern Luzon arc may have b e e n metasom­ generated more mid-ocean ridges, which led
calc-alkaline and tholeiitic arc basalts. atized by a small slab-melt component, given to the subduction of much younger crust.
It was within a group of these NEAB in that temporal variations along the strike of the Drummond and Defant [1990] showed that
northern Kamchatka that Kepezhinskas et al. arc c a n only b e explained through bulk mix­ adakites increase in the rock record from the
[1996, and references therein] found rare ing of a sediment and slab component. present to the Archean.They pointed out that
arc-mantle xenoliths that they believed repre­ Is a slab-melt c o m p o n e n t limited only to the today's adakite productivity is probably lower
sented the source of NEAB. S o m e ultramafic Luzon arc? We do not think so. We have stud­ than at any time in the past b e c a u s e the man­
xenoliths contain high-Na and Al glasses with ied xenoliths from the southern Kamchatka tle is much c o o l e r than it was in the Archean.
adakitic trace element concentrations. In arc where the subducting slab is old and Martin [1999] has also noted that the Mg
addition, the mineral phases in several of the cold. Ion-probe analyses of melt inclusions concentrations in Archean trondhjemites are
xenoliths—for example, the garnet-bearing with compositions similar to those of adakites lower than in current-day adakites. He postu­
pyroxenites—are remarkably similar to those s e e m to permeate the mantle below this arc. lated that the Archean adakites probably did
phases obtained during experimental runs High-precision Nb and Ta data analyzed not interact with the mantle as much and
when felsic melts (trondhjemites) are reacted from several island-arc volcanic samples indi­ envisioned more adakite production and sub­
with ultramafic material representative of cate that Nb/Ta ratios vary in basalts along sequently more adakite metasomatism of the
the arc mantle. Kepezhinskas et al. [1996] the arcs. To explain these changes, variations Archean mantle preventing adakites that reach
c o n c l u d e d that the xenoliths were the prod­ in a slab-melt metasomatizing c o m p o n e n t the surface from extensive interaction.The
ucts of a newly discovered type of metasoma­ were proposed. adakites produced today are much higher in
tism which they referred to as Na metasomatism. Several researchers have noted that arc MgO than not only the Archean adakites, but
They envisioned that this process o c c u r s volcanics fall along the equiline on a graph also those produced experimentally and
when slab melts percolate and interact with of 2soTh/232Th versus 238^/23270 rather than to the adakitic melt inclusions within mantle xeno­
the sub-arc mantle. 238 232
right toward higher U/ Th ratios. U is more liths (Figure l ) . I n fact, we would argue that
 Eos,Vol. 82, No. 6, February 6, 2001
pristine adakites are rare at the surface in people have helped contribute to the adakite growth via slab melting, J Geophys. Res., 95,
present arc systems, if they exist at all. story but could not be referenced in the paper 21,503-21,521,1990.
So, is the adakite metasomatic volcanic due to Eos reference limitations: J. Ayer, Kay R.W, Aleutian magnesian andesites: Melts from
subducted Pacific oceanic crust,J.Volcan. Geo-
series present in the Archean? Until recently M. Carroll, J. de Boer, J. Devaney D. Desonie,
therm.Res., 4,117-132,1978.
no NEABs had been reported from the Archean. T Dunn, F Hochstaedter, S. Kay, J. Luhr,
Kepezhinskas, P, M. J. Defant, and M. S. Drummond,
But in several new papers, authors have Y Marquez, R. Maury, P Morris, R. Nesbitt, Progressive enrichment of island arc mantle by
reported NEABs from a number of Archean J. Pearce, G. Pe-Piper, J. Piper, V Ramos, melt-peridotite interaction inferred from Kamchat­
terrains noting their similarities to those found T. Rushmer, J. Ryan, F Rye rson, D. Sajona, C. Sen, ka xenoliths, Geochim. Cosmochim.Acta, 60,
in present-day arcs. C. Stern, A. Stolz, R.Taylor, A.Thompson, 1217-1229,1996.
The Archean was a time of immense ore M.Tiepolo, PTomascak, R.Vannucci, P Vidal, Martin, H., Adakitic magmas: Modern analogues of
production. Numerous immense gold deposits B.Watson, E.Widom, PWyllie, and D.Wyman. Archaean granitoids, Lithos, 46,411-429,1999.
greater than 2 - 2 . 5 billion years old are found Peacock, S. M.,T. Rushmer, and A. B.Thompson, Par­
tial melting of subducting oceanic crust, Earth
in Archean rocks. There may b e a genetic Authors Planet. Sci. Lett, 121,227-244,1994.
relationship between slab melting and precious
Rapp, R. P, N. Shimizu, M. D. Norman, and G. S.
and base metal mineralization. In several Marc J. Defant and Pavel Kepezhinskas
Applegate, Reaction between slab-derived
cases, adakites have been linked to epithermal Department of Geology, University of South melts and peridotite in the mantle wedge:
gold and c o p p e r porphyry deposits [e.g., Florida, Tampa, USA Experimental constraints at 3.8 Gpa, Chem.
Sajona and Maury, 1998]. Geol., 760,335-356,1999.
We have also noticed the extensive gold References Sajona, EG., and R.C.Maury Association of adakites
and porphyry copper production associated with gold and copper mineralization in the Philip­
with NEABs and adakites, particularly along Bebout, G.,J. G. Ryan,W PLeeman, and A. E. Bebout, pines, Cr.Acad. Sci II,A326,27-34,1998.
Fractionation of trace elements during Schiano, P, R. Clocciatti, N. Shimizu, R. C. Maury K.-R
the northern coast of Papua New Guinea,
subduction-zone metamorphism—effect of con­ Jochum, and A. W Hofmann, Hydrous, silica-rich
Lihir, Bouganvilla, and several localities in
vergent margin thermal evolution, Earth Planet. melts in the sub-arc mantle and their relationship
western Panama. We suggest that NEABs and Sci.Lett, 777,63-81,1999. with erupted \avas,Nature,377,595-600,1995.
adakites could b e used for gold and c o p p e r Defant, M. J., and M. S. Drummond, Derivation of Sigmarsson, 0., H. Martin, and J. Knowles, Melting of
exploration the way kimberlites are used to some modern arc magmas by melting of young a subducting oceanic crust from U-Th disequilib-
find diamonds. 7
subducted lithosphere, Nature, ^z ,662-665,1990. ria in the austral Andean lavas, Nature, 394,
Defant, M. J., and M. S. Drummond, Mount St. Helens: 566-569,1998.
Acknowledgments Potential example of the partial melting of the Yogodzinski, G. M.,R.W Kay, O. N.Volynets, A.Y
subducted lithosphere in a volcanic arc, Geol, 21, Koloskov, and S. M. Kay, Magnesian andesite in the
We thank J o h a n Varekamp for editorial han­ 547-550,1993. western Aleutian Komandorsky region: Implica­
dling and insightful comments and Suzanne Drummond,M.S.,and M.J.Defant,A model for tions for slab melting and processes in the mantle
Kay for a careful and helpful review. Many trondhjemite-tonalite-dacite genesis and crustal wedge, Geol. Soc.Am. Bull, 107,509-519,1995.

people, property and local natural environment

Dead Sea Gives Life to a Unique (fully confirmed after the shots). However, in
Seismic Calibration Experiment order to verify the assumptions upon which
the safety analysis is based, Gil began the
experiment by detonating 0.5-ton and
PAGES 6 5 , 7 0 (Gil) in collaboration with several organizations. 2-ton explosions prior to the main 5-ton shot.
The experiment, realized in the autumn of Based on a series of small-scale underwater
An unusual seismic and hydro-acoustic experi­ 1999, is part of ongoing research on improve­ shots in 1993 [Gitterman et al, 1998], it was
ment involving large underwater c h e m i c a l ments in monitoring the CTBT in the Middle estimated that a 5-ton explosion would corre­
explosions has b e e n conducted in the Dead East by the Israel Seismic Network (ISN). It is spond to an event of local magnitude of about
Sea.The purpose of this creative, cost-efficient funded by the U.S. Defense Threat Reduction 4.0. Conducting large explosions in this politi­
venture was to improve monitoring and verifi­ Agency (DTRA) to provide calibration sources cally sensitive area, just a few kilometers from
cation of the Comprehensive Test Ban Treaty now lacking in the region. the Israel-Jordan border, required serious,
(CTBT) in the Middle East. Apart from the well-known e n h a n c e m e n t long-term diplomatic efforts.The agreement of
The goals of the experiment were to calibrate effect of underwater explosions, the "deadness" the Jordanian authorities was obtained after
the regional travel times of the seismic waves of the Dead Sea was o n e of the main reasons contacts at the highest government levels. A
from the explosions; to provide data for source for this site choice. (The seismological com­ number of international observers from the
characterization; and to improve detection, munity may remember the failure to conduct United States (DTRA), Russia (its embassy in
location, and discrimination capabilities of the similar calibration shots in the Atlantic O c e a n Israel), Jordan (the Royal Engineering Corps),
International Monitoring System (IMS) stations. near Portugal in 1993, caused mainly by "green" the CTBT Organization in Vienna (CTBTO),
From a seismologist's point of view, the experi­ protesters). Moreover, the unique properties of and the Palestinian Authority (An-Najah Uni­
ment's main 5-ton explosion exceeded expec- hyper-saline Dead Sea water—with a density of versity) were invited to participate in the
tations.The blast was recorded at 8 IMS stations 3
1.236 gr/cm and acoustic velocity of 1770.6 experiment and inspect the weighing and
(including PDYAR in Russia, more than 5000 m/sec [Anati, 1997]; much higher than in the preparation of charges before the explosions.
km away), with the estimated location only o c e a n — c a n further e n h a n c e the seismic Explosives and measurement equipment
2.3 km from ground truth, and an event mag­ effect of an underwater explosion. were delivered to the detonation point by an
nitude of mb=3.9 [Center for Monitoring Although there is no aquatic life in the Dead iron raft (Figure 1-top) also bearing a crane
Research, 1999].These results compare to the Sea, it has many natural p h e n o m e n a such as for downloading the largest charge (Figure 1-
predicted local magnitude of about 4 and the mineral springs, desert oasis, and wildlife bottom).The experiment was videotaped. A
maximal observable range of about 2600 km preserves nearby. The explosion point was h o m e camera was placed on the raft at a
[Gitterman, 1998] .A lack of a detectable signal located a few kilometers from local spas, distance of several hundred meters from the
on s o m e nearer stations will likely reveal settlements, and an important engineering shot site.The video and audio tracks provided
important path effects. installation, the Dead S e a Works dam. Conse­ observations and approximate time measure­
The three explosions, ranging in size from quently, all safety conditions were analyzed ments of interesting physical phenomena,
0.5 to 5 tons, were conducted by the Seismology and defined long before the experiment took such as a "cavitation hat," arrivals of shock
Division of the Geophysical Institute of Israel place and demonstrated no hazard to the waves, and bubble pulsations near the raft,