RECOMMENDATIONS OF THE IBM ARC SYSTEM WORKSHOP

1. TECTONIC HISTORY OF THE IBM ARC SYSTEM
   • Initiation of subduction
     • We concur with the 'Magmatic Evolution Working Group', that understanding processes of subduction initiation is a first-order scientific problem that is likely to be resolved most readily for the IBM system. Top priority should be given to understanding the tectonic setting that existed in what is now the Philippine Sea region and modeling the stresses responsible for subduction initiation.
     • Drilling of areas where there is a high likelihood of obtaining pre-IBM crust (i.e., the East Amami triangle area just west of the northernmost Palau-Kyushu Ridge and the Philippine Sea floor west of Palau) should be a high priority. An objective of this drilling is to obtain paleomagnetic data to constrain reconstructions of the Philippine Sea Plate, which suggest large amounts of rotation of the plate since IBM arc initiation.
   • The 'Ophiolite problem'
     • Characterize forearc crustal structure to test the hypothesis that an oceanic crustal section exists in the IBM forearc, and establish the timescales and processes by which the IBM forearc formed.
   • Kinematic evolution and plate reconstructions
     • Confident reconstruction of the IBM arc system will require a concerted effort to understand its motions through time.
     • GPS campaigns
     • 3 component magnetometer studies (needed because of proximity to equator)
2. MODERN PROCESSES OF THE IBM ARC SYSTEM
   • Forearc deformation
     • We concur with the 'Fluxes' working group, that it is very important to understand the interaction between sediments, crust, and seamounts of the Pacific Plate and how this material interacts with the IBM forearc as it enters the trench and begins to be subducted. These Interactions include a wide range of tectonic processes, including tectonic 'erosion', accretion to the forearc, underplating beneath the forearc, and subduction to greater depths.
     • What are causes some segments of the forearc to be uplifted (Bonin Islands, Guam-Saipan)?
     • We recommend that special effort be placed on understanding ongoing processes of collision of the north end of the IBM arc with Japan. Emphasis should be placed on 1) Understanding when and why did subduction begin along the Nankai Trough 2) Resolving uplift rates and mechanisms of shortening leading to progressive migration of deformation fronts to the south; and 3) Understanding the seismic risk this collision has for the Tokyo-Yokohama area.
     • What are the processes responsible for formation of forearc basins in the IBM arc system?
   • Tectonic controls on igneous activity
     • How does the distribution of volcanic vents and magmatic centers relate to the stress field of the IBM convergent margin? Specific issues center on what controls the location of volcanic cross-chains, active arc volcanoes, and forearc sills?
     • How has collision at the north end of the IBM arc affected volcanism in this region?
   • Arc rifting and transition to spreading
     • While there is general agreement that mature backarc basins are sites of seafloor spreading and that the earliest stages of extension reflect rifting processes, controversy persists about the nature of rifting and how the transition to seafloor spreading occurs. Results from DSDP Leg 60 suggested that 'spread' crust extended across the Mariana Trough backarc basin, but more recent ODP drilling in the Lau Basin suggests that a substantial part of this and other backarc basins are floored by stretched arc crust. Arc magmas may be diverted to the rift axis during early arc rifting and may contribute to the formation of a hybrid crust, formed by processes more like spreading but from melt compositions indistinguishable from those of the magmatic arc, with an additional component of stretched, older arc crust.
     • What are the mechanics of extension in rifts? We know too little about the rheology of mafic arc and backarc basin crust under tension. A wide range of theoretical and experimental studies should be undertaken on this subject, including the nature of faulting and ductile flow, and the interactions that these have on development of magmatic conduits.
     • What are the important controls of backarc basin spreading? While we appreciate that backarc basin spreading is very similar to that of normal mid- ocean ridges, there are significant differences. These include: 1) Hydrous as opposed to anhydrous melting, and 2) Asymmetric spreading.
   • IBM arc system sedimentary systems
     • What are the interactions between tectonic and thermal subsidence controls on the architecture of IBM forearc and sedimentary basins?
     • We need to refine models of volcaniclastic sedimentation and facies evolution for the IBM arc system.
3. GEOPHYSICS
   • One of the highest priority objectives is to determine detailed arc crustal and upper mantle structure that can only be obtained by large-scale reflection/refraction seismologic and tomographic experiments. There is little question that one of the most exciting developments in our understanding of the IBM arc system over the past decade has been the new view of IBM crustal structure revealed by the efforts of ORI scientists in conducting detailed (OBS controlled) seismic imaging of the IBM arc. The ORI group plans to image Palau-Kyushu Ridge crust in fall 1996 (IBM remnant arc). These results will help to constrain the evolution and extent of the mid-crustal "tonalitic" layer identified from the Izu-Bonin and provide a nearly complete IBM- PKR crustal structure transect. We applaud these efforts and urge that a similar profiles be made across the IBM arc system across an arc segment with mature, spreading back-arc basin and associated remnant arc, such as that of the Marianas about (18°N).
   • Confident interpretation of crustal images requires a program of co- ordinated drilling. Important new results could come from either deepening of ODP sites or drilling , or initiating a new program of deep drilling from frontal arc islands
   • Study of exposed mid-crustal sections resulting from arc-arc collision in the Fossa Magna, Japan)
     • What is the structure of the upper mantle beneath the forearc, active arc, and back-arc regions?
     • What is the nature of the boundary between static mantle (lithosphere) and convecting mantle (asthenosphere)?
     • How is serpentinization distributed across the of the sub-forearc mantle?
     • What controls the lithosphere/asthenosphere boundary beneath the active arc and beneath the back-arc basin?
   • Answering these questions requires tomographic studies. We encourage research programs such as that planned by Mariana planned by ORI scientists.
     • What controls seismicity beneath the forearc, active arc, and backarc?
     • We need to advance our understanding of the many dynamic problems of the IBM mantle wedge, including:
     • What is the convective regime of mantle wedge asthenosphere, and how is this different between segments with and without actively spreading backarc basins?
     • We need a more refined understanding of mantle downwelling beneath the arc and controls on ascent (counterflow) of partially molten diapirs feeding the arc.
     • What is the thermal structure of the mantle wedge, and how does this control the location of melt generation?
4. IBM ARC SYSTEM TECTONICS & GEOPHYSICS DATA NEEDS:
   • A critical starting point for understanding these IBM processes is comprehensive bathymetric mapping. Although many researchers have collected bathymetric and backscatter data over portions of the IBM, comprehensive data sets do exist (US military SASS; Japan Hydrographic Department multibeam). However, these data have not been made generally available for IBM research. We urge the complete release of as much of these data as possible. This will provide valuable "new" data for IBM research and prevent researchers from wasting time remapping previously mapped areas.
   • Crucial new geophysical data is required for significant new advances in our understanding of the IBM arc system to occur. Such new data sets include:
   • MCS reflection and refraction investigations of crustal structure
   • Seismic tomography
   • 3-component magnetic data
   • Acoustic imagery
   • Comprehensive 3-D reconstructions of seismicity

Magmatic Evolution

1. LONG TERM MAGMATIC EVOLUTION
   • Nature and composition of the Philippine Sea region prior to and during arc inception
     • It is important to know the isotopic and trace element signature of pre- IBM lithosphere and asthenosphere so that isotopic and chemical modifications resulting from subduction-related metasomatism can be confidently identified. An effort has been made through several DSDP and ODP drilling legs (Legs 59, 60, 125, and 126) to sample the pre-IBM arc basement. The question remains whether the IBM system was built on an early episode of new lithosphere formation associated with subduction initiation at 40-48 Ma or whether Eocene lavas comprise 'frosting' on an older, pre-IBM lithospheric 'cake'. Deepening ODP Site 786B would be a superb test of the 'forearc ophiolite' model.
     • It is clear from the cross-cutting relations of the Palau-Kyushu Ridge with respect to the Amami Plateau-Daito Ridge 'fault', that the early IBM arc was not simply constructed on an older transform system, and that the Palau- Kyushu Ridge was built on some pre-existing basement (?Cretaceous - Paleocene), at least in the area east of the Amami Plateau. As a primary constraint, we need to know the geochemical character of the igneous basement of the West Philippine Basin, both here and south of the Central Basin Ridge, near Palau. Drilling to basement at these sites will also provide paleomagnetic constraints on the location of the Philippine Sea Plate at the time of IBM arc initiation. We need to know what the basement is both here and further south, in the vicinity of the junction of the Central Basin Ridge with the Palau-Kyushu Ridge, and further south towards Palau.
     • The fact that the CBR continued to spread for more than 10 million years after IBM inception may have important implications for understanding the petrogenesis of boninites, which are common in IBM but uncommon in coeval or otherwise similar 'infant arc' systems such as the Tonga- Kermadec system during the Eocene. In general, the tectonic and thermal setting of subduction of cold lithosphere beneath an active spreading center is unique and merits study, particularly the composition of CBR basalts and contemporaneous volcanics of the Palau-Kyushu Ridge. The CBR has been a target of Shinkai 6500 diving, and dredge sampling is planned by the Hydrographic Survey of Japan. Based on the sequences recovered at DSDP Sites 447 and 290 and DSDP Site 448, sampling of the volcanic succession on the Palau-Kyushu ridge is best achieved by drilling through the volcaniclastic apron just north or south of the CBR.
   • Subduction Initiation
     • We know very little about the sequence of events actually responsible for initiation of the IBM subduction zone, although there is a very strong hint that the ultimate cause lay in the excess density of old Pacific seafloor relative to buoyant lithosphere of the Philippine Sea. A concerted effort is needed to constrain the unique processes responsible for subduction initiation in the IBM arc system. For this effort, it is critical to have the input of theoretical and experimental geodynamicists on how subduction zones begin. This will have important implications for our understanding of the driving forces responsible for plate tectonics, and is likely to be a fruitful area of research for the next several decades.
   • Boninites
     • These unusual igneous rocks are common in the IBM forearc, where they formed during the arc initiation phase. What is the relationship between subduction initiation and the generation of boninites?
   • Stabilization of the Magmatic Front and Change or Constancy of Magmatic Compositions with Time
     • The transition from subduction initiation, characterized by ultra-depleted boninites and primitive arc tholeiites forming in a broad zone of extension now occupied by the forearc, to a stable arc, characterized by a well-defined magmatic front behind a cold, stable forearc, is the single most important event in arc evolution after subduction begins. Once established, the position of the magmatic front is relatively stable over time. In the case of the IBM arc system, the magmatic front was established about 35 Ma. The location of the magmatic front and its long-term stability provide powerful constraints for our understanding of subduction zone magmagenetic controls. For example, does the position of the magmatic front reflect an isobaric dehydration reaction, or does it follow a high-angle boundary between hot, fertile, and convecting asthenospheric mantle beneath the arc and back-arc and cold, depleted, lithospheric mantle beneath the forearc?
     • It is important to understand first-order changes in magmatic compositions with time. This provides important constraints for understanding competing processes of mantle depletion by melting, mantle replenishment by convection, and the progressive effects of subduction- related metasomatism. For example, if the now-discredited model of arc temporal evolution from arc tholeiitic to calc-alakalic to shoshonitic were true, this might indicate systematic decreases in the degree of melting with time reflecting cooling mantle or thickening crust. Stratigraphically- controlled tephra recovered from DSDP and ODP cores provide an unparalleled record of the magmatic evolution of this arc system and suggest that there may be a subtle, long-term increase in potassium contents. However, the larger signal is a much stronger increase in potassium contents at about 10 Ma. Also apparent in the tephra record are indications of a dramatic increase in tephra abundance during the middle Miocene, indicating an increase in explosive volcanic activity at this time, with possible consequences for understanding global climate change. Continuing efforts to reconstruct the temporal evolution of the arc should include ion-probe analyses of tephra for the purpose of understanding whether the observed variations in potassium manifest greater subduction- related fluxes or lower degrees of melting, and concerted efforts should be made to use these data to build a consistent model for the IBM subduction zone. In conjunction with the above, efforts should be made to identify periods of significant felsic volcanism and their relationship with the emplacement of felsic plutonic rocks like that inferred for the mid-crustal felsic layer of the northern IBM arc. Ages and compositions of tonalitic rocks dredged from the northern Palau-Kyushu Ridge and elsewhere should be integrated with similar information for felsic volcanics to determine the relationship between volcanism, subduction fluxes, and crustal growth.
2. SHORT TERM MAGMATIC EVOLUTION
   • Magmatic vs. Lava Compositions
     • Before all else, we must deal with the problems associated with inferring magmatic evolution from studying porphyritic arc lavas. Many IBM arc lavas are coarsely porphyritic and while analyzing these may be appropriate for extracting isotopic and incompatible element ratios, the use of major and compatible trace element data from such samples may mislead petrogenetic models. Recent studies focusing on tephra glass or glass inclusions in phenocrysts demonstrate this point convincingly. For example, while high- alumina compositions are characteristic of Mariana arc basalts, analyses of tephra or inclusion glass fails to find such compositions. Instead, Mariana arc mafic glasses are characterized by high Fe contents. Fe-rich melts are likely to lead to stably zoned magma chambers and suspension of plagioclase at the top of the mafic melt zone. In contrast with studies of Mariana arc lavas which indicate a unimodal distribution of silica contents consistent with low-P fractionation, Mariana glasses show a bimodal distribution with mafic (53% silica) and felsic (70% silica) modes. Such felsic compositions are not reported from subaerial Mariana volcanoes, possibly because such material is found as altered and poorly consolidated ash deposits. These results indicate that we must critically re-evaluate our understanding of IBM arc magmatic compositions that, in the past, have been based on analyses of whole-rock lavas, and concentrate in the future on analyzing materials likely to improve our understanding of melt evolution.
   • IBM arc Parental Magmas
     • Glasses recovered from subaerial volcanoes along the Mariana arc magmatic front are invariably fractionated, with the most primitive glasses having Mg# <60. This contrasts with glasses from cross-chain volcanoes where primitive glasses having Mg# of 65-70 are common. The reasons for these differences are not obvious, but rapid crystallization accompanying loss of magmatic water along the magmatic front is suspected. Understanding the mantle-to-crust magmatic flux depends on identifying the parental magma(s) of the IBM arc system, and inferring petrogenetic regimes responsible for the generation of such melts will provide crucial constraints on the thermal structure of the sub-IBM arc mantle. Similarly, reconstructing arc crustal and lithopheric structure depends on understanding the processes responsible for the ubiquitous and extensive fractionation reflected in IBM lavas.
   • Generation of IBM Felsic Magmas
     • Doubtless one of the most important new results coming out of the last decade's study of the IBM arc is the identification by geophysicists of a ca. 10 km-thick mid-crustal layer beneath the northern IBM arc that appears to be composed of felsic plutons. This interpretation is gathering increased acceptance as data for Poisson's ratio for the same crustal section has been interpreted as most consistent with felsic rocks, and as geologists identify correlative exposures in Japan along the IBM collision zone and as xenoliths in northern IBM arc volcanoes. This discovery has the potential for revolutionizing models for the evolution of the continental crust and so has implications far beyond the IBM arc system. This comes at about the same time that we are beginning to appreciate that felsic melts are much more important in the Mariana arc than previously thought. It is thus critical that petrologists turn their attention towards the difficult question of how such felsic melts form, whether by fractionation, anatexis, or liquid immiscibility/zoned magma chamber.
   • Tectonic controls on arc magmatic compositions
     • Earlier studies have indicated that arc magmatic compositions are controlled by factors including depth to the Benioff Zone, degree of melting, and age of the arc. Detailed petrologic and geochemical studies are needed to assess these possibilities and interpret their significance for our understanding of how the sources of arc magmas change through time and space. Another model ascribes greater depletion in arc magmas where these are associated with actively spreading back-arc basins. The fact that the IBM arc system has backarc spreading in the south and not in the north allows a unique opportunity to test and refine this model.
     • A related issue concerns the transition from unrifted arc to rifted arc to back-arc basin spreading. A controversy exists regarding when true back- arc basin basalts (BABB) are erupted from evolving back-arc rifts. One group contends that BABB may erupt from the earliest stages of extension, while another group contends that BABB cannot form until back-arc seafloor spreading is established. This is a very important test of models for subarc mantle convective patterns. For example, if BABB may erupt immediately upon rifting, then melting is distinct from that of normal MORB where adiabatic decompression controls basalt petrogenesis. Similarly, because arc melts and BABB are associated with distinct mantle convective patterns (BABB is associated with regional upwelling, arc lavas reflect mantle counterflow, that is, arc diapirs ascend against a regionally downwelling mantle), understanding the transition from arc lavas to BABB offers important insights into mantle flow patterns. The IBM arc system has examples of each of the important tectono-magmatic settings (unrifted arc, rifted arc, BAB seafloor spreading) and is in fact the only arc system on earth which such a diversity of tectono-magmatic environments.
   • Across-Arc magmatic variations
     • Compositions and magma volumes change markedly from the magmatic front rearwards. So-called 'cross-chains' are relatively unusual in the southern IBM but common in the north. What is responsible for these variations? What do these enriched, relatively low-degree melts tell us about sources and melt processes at convergent margins? Primitive lavas (Mg#>65) are unheard of along the magmatic front but are common in these cross-chains - why?
   • Along-Arc Variations
     • One of the most interesting features of the IBM arc is an important variation in lava compositions along strike, from medium-K in the south, through shoshonitic and alkalic in the center, to low-K in the north. Are these compositional features long-lived or are they transient responses to changing tectonic regimes? What is responsible for these variations - relative source depletions, fluid flux from the slab, or degree of melting? Can these variations be related to changes in subduction rate from south to north?
3. NEED FOR AN IBM ARC SYSTEM DATABASE
   • Many questions in magmatic evolution would be more readily answered if a comprehensive database were available. We recommend as a high priority item that an IBM database be developed with information on location, age, petrography, mineral compositions, and chemical and isotopic compositions be listed. Because much of the data exists only in Japanese-language journals, it is imperative that this database be developed as a collaborative effort between U.S. (and other western) and Japanese scientists. We recommend that this database be designed for flexibility and ease of access, based on EXCEL or a similar spreadsheet, with an effort made to build on formats developed by the RIDGE and INTERIDGE programs for MORB or BABB.

Subduction Fluxes in the IBM Arc

• Continental growth
• Mantle chemical and isotopic evolution
• Forearc evolution
• Controls on seismogenic zone
• Melting of the mantle at convergent margins
• Controls on explosive volcanism (and associated climate modification)
• Volatile (H2O, CO2) fluxes and the evolution of the hydrosphere
• Formation of ore deposits

1. MODERN FLUXES
   • Input Fluxes
     • Quantify spatial heterogeneities in sediment input, and the relationship that these variations have to output from the arc. At present, this can be done with existing data for the Mariana segment; additional analytical studies of existing samples and perhaps additional sampling will be required for the Izu-Ogasawara sections.
     • Develop methods for quantifying spatial heterogeneities in off-ridge (mostly Cretaceous) volcanism and volcaniclastic sedimentation, as well as a model for the alteration and hydration of such materials.
     • Develop methods for constructing meaningful estimates of altered igneous oceanic crust mineralogy and composition, and use this to estimate volatile and elemental fluxes out of the subducted plate.
     • Constrain along-arc variations in input fluxes by drilling and coring one ODP hole seaward of the Izu-Bonin trench to > 300m into oceanic basement.
     • Deepen ODP 801 to >300m into oceanic basement for the purpose of understanding alteration and hydration of old oceanic crust.
     • Synthesize existing seismic data to create isopach maps of sediment and basement. These will allow detailed chemical and isotopic stratigraphy from ODP 801 and the to-be-drilled site east of the northern IBM arc to be tied together. New programs of detailed MCS profiling may be required so that meaningful input fluxes can be calculated all along the IBM arc.
     • Ultimately, we need several more drill sites seaward of the IBM trench system, to explore spatial heterogeneities in the sedimentary sequence. > Resolve the nature of the mantle input, whether MORB-like or OIB-like, and whether this has affinities to the mantle beneath the Indian Ocean or Pacific Ocean.
   • Output Fluxes
     • At present, we do not know of any sites of fluid egress from the IBM trench axis or inner trench wall, although these sites are likely to be significant and vent fluids with distinctive compositions. If present, such vents are too deep to be dived on using existing manned submersibles. These investigations are natural targets for a co-ordinated program of KAIKO ROV investigations.
     • There is very little indication that Pacific plate materials are 'scraped off' anywhere along the IBM trench. If tectonic accretion does occur, it is to be expected to accompany the arrival of large seamount chains such as the Ogasawara plateau in the central IBM arc. The inner trench wall of this region should be studied for evidence of such processes.
     • Determine time-integrated fluxes out of the fore-arc. This includes quantifying the volume and duration of serpentinization of the forearc, and modelling how this zone forms. Such studies should lead to an understanding of why serpentinite diapirs in the Mariana forarc are still active, whereas those in the Izu forearc have not been active since mid- Tertiary time. These studies will need to be supplemented by seismic imaging to constrain the sub-surface zone of serpentinization.
     • Determine real-time fluid fluxes. This will require approaches like that of CORK or other seafloor observatories.
     • Estimate compositional variations of fluids both across and along strike of the arc.
     • Constrain low-temperature water-rock reactions from fluid compositions, experimental data and subduction zone metamorphic assemblages.
     • Constrain original mantle lithosphere compositions by subtracting hydration and associated metasomatic effects.
     • We need better estimates of magma production rates along the IBM arc system. Obtaining this will be an interdisciplinary undertaking, requiring geophysical imaging of crustal structure, good estimates of rates of crustal evolution, and bathymetric and gravity mapping. Petrogenetic modelling will be needed to relate erupted lava compositions to cumulate volumes (See recommendations of "Magmatic Evolution" working group, II).
     • Volatile and trace element compositions of primary melts and melt inclusions.
     • Degassing history of magmas inferred from fumaroles, eruption clouds, etc.
     • Generate a good database for the IBM arc system. In this regard, we strongly a support recommendation III of the "Magmatic Evolution" working group. We emphasize that this data base will require new analyses as well as compilation of existing data sets for the Izu-Bonin volcanic arc.
     • Generate and test models for delivery of slab-derived hydrous fluids and melts to the zone of melt generation, and continue to refine models of melt generation and ascent.
     • Model the cause of along-strike variations in lava compositions of the IBM arc
     • Quantify volcanic and hydrothermal fluxes in the back-arc region.
     • Generate and test models for delivery of slab-derived hydrous fluids and melts to the zone of melt generation. This is particularly important with regards to the Mariana Trough, where the spreading axis for the most part does not lie above the subducted slab.
   • Fluxes Into the Mantle
     • Given increasing evidence that subducted lithosphere sinks into the deep mantle beyond the region underlain by the arc and back-arc, we need models to explain what part of the subducted chemical budget is not recycled to the surface but is returned to the deep mantle. We need to know what is returned to the deep mantle, in what mineral phases it is carried, and how this flux has affected the composition of the mantle over geologic time.
     • In addition to the flux to the deep mantle, we should resolve the flux from the subducted slab to the convecting asthenosphere. This flux is predominantly accomplished by metasomatism accompanying hydration, much of which results in melting of the affected mantle and addition to the arc magmatic budget. A significant portion of affected mantle, both residual and unmelted, is likely to persist as asthenosphere.
2. ANCIENT FLUXES
   • Once relationships between subducted input and output are well understood for current tectonic settings, the composition of older arc lavas may provide constraints on models of trench rollback linked to rifting, seamount subduction, or ridge collision.
3. ADDITIONAL REQUIREMENTS
   • To incorporate the types of data discussed above into an integrated model of arc magmagenesis, additional types of information are essential, including the following:
   • Well-constrained thermal and convection models of the slab and wedge.
   • Partitioning of elements between slab and fluids or melts over a wide P-T space.
   • Experimental determination of hydrous mineral stabilities in mantle compositions.
   • The ultimate objective of these studies should be the integration of all salient processes and processes of the IBM arc system, including thermal structure, rheology, mineralogic transformations, and fluid and melt generation and transport processes into a realistic physico-chemical model of this convergent margin 'factory'.

Hydrothermal Activity and Mineralization

1. PROBLEMS TO BE SOLVED WITH RESPECT TO HYDROTHERMAL ACTIVITY IN THE IBM ARC SYSTEM
   • Fluid Circulation and Dynamics Within a Selected Submarine Arc Volcano Several parameters need to be investigated. These include:
     • Measurement of the dimensions of the convection cell (including the recharge and discharge zones) within an active submarine volcano. This should involve studying the hydrology and thermal regimes within the volcano. Mineralization within and at the surface of the volcano should also reveal processes occurring within the hydrothermal convection cell. For example, within the Izena Caldera, the radius of the convection cell is about 1 km.
     • Determination of typical water/rock ratios developed during interactions of hydrothermal fluids with the host rocks.
     • Measurement of relative heat and mass fluxes of hydrothermal flows at active vents and diffusive flows.
     • Time variability of fluid flow should also be established. This should include determination of the periodiciity of the fluid flow, its relationship to seismic and volcanic activity, the role of phase separation during the upward flow of the hydrothermal fluid as well as other parameters.
   • Relationship of Hydrothermal Activity and Composition to Tectonic Framework
     • This should include and E-W transect across the IBM arc (including, as feasible, fore-arc, volcanic front, backarc rift, backarc spreading, and off- axis volcano) in order to establish the role of tectonics and associated volcanic style on the composition of hydrothermal fluids as well as to evaluate the role of faults for the migration of hydrothermal fluids. At mid- ocean ridges, hydrothermal activity is restricted to basaltic volcanic systems. In the back-arc rift setting, the influence of felsic and mafic volcanism on the characteristics of hydrothermal fluids. The presence of submarine shoshonitic centers in the central IBM arc allows the further possibility that the influence of this unusual magmatic composition on the composition of hydrothermal fluids can be investigated.
   • Identification of Hydrothermal End Members
     Several possible sources of hydrothermal fluids in the IBM back-arc environment need to be distinguished. These include:
     • Interaction of seawater with volcanic crust and to a lesser extent with surface sediments. Bacterial activity plays a role in redox processes, isotope fractionation and precipitation within hydrothermal fluids.
     • Magmatic fluids derived from arc and backarc basin magmas. This involves contributions from the mantle wedge and subducted crust and sediments.
     • Interaction of hydrated arc crust
     • Crystallization within the magma chamber
     • Phase separation during upward flow of the hydrothermal fluid.
     • Estimation and Modeling of Hydrothermal Geochemical Fluxes to the Ocean
     • At present, only limited data on the flux of gases (e.g., CO2, CH4, H2S) and metals from submarine hydrothermal systems are available. These data do not take into account the variability of hydrothermal systems with time or with tectonic setting. This is particularly relevant since gas concentrations in hydrothermal fluids from backarc basin settings are about ten times greater than those of mid-ocean ridges. Although the length of the global backarc system is only one tenth of that of the mid-ocean ridges, the much higher gas concentrations there mean that the contribution of hydrothermal fluids from the global backarc system is comparable to that from the mid- ocean ridges.
     • Estimating and modeling of hydrothermal gas and metal fluxes needs to be based on a data set which takes into account variations in tectonic setting and variability of hydrothermal systems with time and should give an estimate of the contribution of backarc hydrothermal systems to the global output of metals and gases.
     • CO2 and CH4 data will also contribute to modeling of global climate change.
     • Establish the variability of fluid flow in relation to seismicity. This should include determination of the periodicity of fluid flow, its relationship to seismic and volcanic activity, the role of phase separation during the upward flow of the hydrothermal fluid as well as other parameters.
   • Outline of Recommended Research Programs
     • The following outline of future research programs emphasizes detailed studies of already known hydrothermal systems. This is not intended to diminish the importance that should be placed on discovering new hydrothermal areas in the IBM system, of which there may be many.
       • Phase I should aim to characterize one or more hydrothermal areas within the IBM arc system and determining the nature of convection cells. Precise sampling of fluids, precipitates and sediment by submersible, ROV and surface ship will be followed by chemical analysis of these phases. Studies of the chemistry of pore fluids in drill holes will also be carried out. Hydrothermal endmembers will be identified using indicators such as 3He, C isotopes, S isotopes, N isotopes, and concentrations of CL, Pb, Sc, Ga, Pt, and REE.
       • Phase 2 should involve monitoring the long-term variability of hydrothermal fluid circulation. Techniques to be used should include continuous thermistor measurements, use of in-situ sensors for determining hydrological parameters, gases and metals, measurements of heat and fluid flow, periodic fluid sampling and monitoring of seismic and tidal events. Reconstruction of long-term hydrothermal activity may be determined from studies of hydrothermal minerals.
       • Phase 3 should involve the drilling of an ODP hole in an active hydrothermal area of the IBM arc system. The location and depth of this hole should planned in order to better understand the nature and composition of fluids from an active hydrothermal area. The objective should be characterization of fluids in order to obtain a better understanding of the different sources contributing to the hydrothermal fluid compositions.
       • Phase 4 should focus on comparing the nature and composition of hydrothermal systems from various tectonic settings and magmatic compositions within the IBM arc system. Possible target areas include Suiyo Seamount or Kaikata Seamount in the Bonin arc. These seamounts have the highest temperature hydrothermal fluids encountered so far from the IBM arc system. They are also relatively close to Japan. For the E-W transect across the IBM arc, this needs to be at the southern end of the arc system so that both arc and back-arc basin hydrothermal systems can be studied. It is proposed to study the Topless Tower area in the Mariana Trough at 18°, Kasuga cross-chains at 22°N (rear-arc volcanoes) , Esmeralda Bank (submarine arc volcano at 15°N), Conical Seamount (forearc low-temperature fluid vent). Because many ancient economic ore deposits are associated with felsic igneous bodies and because hydrothermal activity associated with submarine felsic igneous activity is so poorly known, we think it is very important to identify a vigorous hydrothermal system associated with a submarine felsic volcanic center.
   • Exploitation of Hydrothermal Resources
     • Hydrothermal resources are an important yet poorly known asset of the region. Hydrothermal deposits may be profitably mined, and hot fluids may be used for generating electrical power. These activities would be important for development of the regional economy. We encourage the governments of the region to work together to develop these resources, within the context of minimal damage to IBM ecosystems.