On the Origins of Adakites: Evidence from B-Be-Li Systematics in Young Lavas from Panama
A COMPARATIVE GEOCHEMICAL STUDY OF META-MAFIC ROCKS OF THE SOUTHERN BLUE RIDGE
Tectonic Setting, Mineralogy and Petrology of Alkaline Rocks from the Russian Far East
Chemical Transfer Events in Subduction Zones as a Function of Depth
GEOCHEMICAL BASELINE STUDY OF UPPER FLORIDAN AQUIFER HOST ROCKS, SW FLORIDA
J.G. Ryan (Dept. Geology, Univ. South Florida), G. Bebout (Dept Earth Envir. Sci, Lehigh Univ.), A. Hochstaedter (Dept. Earth Sci. UC-Santa Cruz), J. Morris (Dept. Earth & Planet. Sci.,Washington Univ., St. Louis)
Studies of transects across subduction zones have shown that the mechanisms of slab-mantle chemical exchange change with changing P-T conditions. B and Cs contents decline during prograde metamor-phism of sediments from the Catalina Schist subduction complex (est. slab depths: 15-45 km). These declines correlate with changes in N and H2O, while K, Ba, Be and LREE contents show little change with grade. Serpentinites from seamounts in the Marianas forearc (20-40 km slab depth) are enriched in B, Sr, Li and Cs, but not in Ba, K, or Be. This combination of results suggest the selective transport of elements in forearcs via predominantly hydrous fluids. Volcanic arc lavas show different patterns. B, Sb, As and Cs en-richments decline as the slab deepens across the volcanic zone of the Kurile arc. As these elements are clearly mobile in the forearc, their patterns in Kurile lavas reflect inputs from a slab depleted in H2O-soluble species. The rearmost arc volcanoes approach MORB/OIB enrichment levels for these elements, suggesting the inventories of soluble species added to the slab via surface processes are nearly gone at these depths. Conversely, K, Ba and 10Be are strongly enriched in all Kurile lavas and show little across-arc change. As a slab phase moves all these species into arc source regions, this phase must differ compositionally from forearc fluids. Deeper slab fluids may be richer in SiO2, due perhaps to slab sediment melting, though isotope and Sr/Nd constraints entail significant inputs from subducted basalt. Alkaline (AL) and calc-alkaline (CA) lavas from the volcanic front of the Mexican Volcanic Belt (MVB) are similar to Kuriles front- and rear-arc lavas. CA lavas are B and Cs enriched, while AL lavas have MORB-like B and Cs; both types are enriched in Ba, Sr, and K. MVB-AL lavas may reflect long-term mantle contamination by slab additions: species like K and Ba, that mobilize late in the subduction process, must also have at least modest affinities for mantle wedge mineral phases. By contrast, B and Cs are incompatible in mantle phases, and wash through quickly in fluids and melts.[Back to top]
On the Origins of Adakites: Evidence from B-Be-Li Systematics in Young Lavas from Panama
S. E. Norrell, J.G. Ryan, and M.J. Defant (Dept. of Geology, Univ. South Florida, 4202 East Fowler Ave., Tampa, FL 33620: (813) 974-1598; email C/O RYAN@CHUMA.CAS.USF.EDU)
Adakites form in arcs where young downgoing plates (<25 Ma)slow subduction produce "hot" environments in which partial melting of the basaltic slab occurs at the amphibolite-eclogite phase transition. Adakites can be distinguished from normal arc lavas based on elevated Sr/Y ratios (>60), resulting from low degree partial melts of amphibolites or eclogites with garnet in the residue. B-Be-Li systematics are widely used to trace slab effects in the genesis of normal arc lavas, so we have conducted a study of B-Be-Li variations on a series of lavas from Panama, which reflect a progressive change in magmatic style: from mantle melting in old group (OG) lavas, to slab melting and the eruption of classic adakites in young group (YG) lavas. Our goal is to evaluate the adakite hypothesis, and place constraints on the sources and melting characteristics of these lavas.
All the Panama lavas examined range from 2-9 ppm B, but B is lower in YG lavas than in the OG. B/Be ratios for YG are <7, while OG lavas reach B/Be of 15. This low B/Be suggest that the slab beneath Panama is hot and relatively B-depleted, producing B-poor lavas like those of the Mexican Volcanic Belt or Cascades.
On a plot of B/Be vs Sr/Y, OG and YG lavas form distinct arrays. B/Be in OG lavas correlates inversely with Sr/Y, suggesting a 2 step process where mantle melting was triggered by B-rich slab fluids, typical of arcs. However, in YG lavas B/Be correlates positively with Sr/Y. A positive array can only be generated by a single stage melting event which fractionates B slightly from Be, and Sr strongly from Y. Such a model is consistent with the partial melting of eclogitic source materials to form adakites.
Li data provide insight into residual phases to adakite melting. K/Li ratios in YG lavas reach 3000, entailing either small degrees of melting or extensive crystal fractionation. These high K/Li values preclude amphibole fractionation, which would deplete K. Li/Y ratios span a large range (1-9) in YG lavas, but are < 1 in OG lavas. Garnet is capable of fractionating Li from both K and Y. Our new data confirm that residual garnet on the slab controls the trace element systematics in adakites. [Back to top]
J.G. Ryan, E. Tenthorey, A. Welty (Dept. Geology, Univ. South Florida, 4202 E. Fowler Ave., Tampa FL 33620; (813) 974-1598; email RYAN@CHUMA.CAS.USF.EDU) M.R. Perfit (Dept. Geology, Univ. Florida) C.H. Langmuir (L-DEO/Dept. Geol. Sci., Columbia Univ.)
The Woodlark Basin (SW of the New Georgia Group, Solomon Is.) is a region of complex tectonics: a fast converging (10 cm/yr), very steeply dipping (90) subduction zone has produced abundant forearc volcanism, with centers that straddle the trench. Subduction polarity changed in the Miocene with the collision of the Ontong-Java plateau to the north. Also, the active Woodlark spreading center has been subducting since the Miocene, generating magmas near the ridge-trench intersection that suggest exchanges between ridgecrest and sub-arc mantle sources. B, Be and Li have been extremely useful in other arcs as tracers of fluid/sediment inputs from the downgoing plate. As such, we have chosen to examine the B-Be-Li systematics along this arc in a suite of fresh Woodlark Basin whole rock lavas, in order to assess how slab-mantle chemical interactions change in proximity to a subducting ridge, and to try to infer an origin for "arc-like" chemical signatures in some Woodlark spreading center lavas.
Mafic lavas from the New Georgia forearc (dredges near Kavachi, Rendova, and Simbo Ridge) all show the elevated B typical of arcs: up to 25 ppm B, and B/Be ratios from 7-80. B/Be becomes lower on average near the ridge-trench intersection, but each volcanic center reflects a range in B/Be. B enrichments in all the forearc lavas correlate inversely with measures of extent of melting (Ba/Y, K/Li and La/Yb), suggesting that here, as in other arcs, melting occurs in response to B-rich fluid inputs from the slab. B enrichments have short mantle residence times, due to the affinities of B for fluid and melt phases. Thus, varying B/Be along New Georgia probably relates to slab inputs from recent subduction, and not to pre-existing mantle enrichments due to earlier subduction from the north.
A sample from the Ghizo Ridge on the downgoing plate shows arc-like elemental patterns and elevated B/Be (20), while lavas from RD 29 in the Woodlark spreading center are closer to MORB values (7). The high B/Be may relate to ridge/arc chemical exchanges, but the role of cryptic B contamination from seawater in these low B, ocean ridge samples is under investigation. [Back to top]
J.G. Ryan, S.E. Norrell, P.K. Kepezhinskas, M.J. Defant (Dept. Geology, University of South Florida)
While similar in terms of many physical parameters of subduction, the Kuriles and Kamchatka segments of the Kuril-Kamchatka arc system differ in the thickness and composition of their crustal foundations. In the Kuriles, where magmas extrude through relatively thin (20 km) oceanic crust, lava geochemistry is dominated by the slab. Across-arc volcanic transects in the Kuriles show strong declines in B/Be, Cs/Th, Sb/Be, As/Be and Pb/Ce ratios. These declines mirror declining abundance levels of B,Cs,Sb,N and H2O in "subduction complex" metamorphic associations, suggesting that diminished slab signatures result from the progressive devolatilization of the subducting plate. Relatively constant enrichment factors for K, Ba, La, and Th may indicate a role for a second, compositionally distinct flux (slab sediment melt?) and/or the buffering effect of hydrated phases in the mantle wedge. Correlated B, Cs, and chalcophile declines argue for the importance of subducted sediments as a source of the slab flux, though Pb results from other arcs have been seen to indicate altered ocean crust inputs (Miller et al. 1994).
Across-arc geochemical patterns in Kamchatka are less clear. While Klyuchevskoi Group volcanoes are sited over deeper slab segments and typically have lower B/Be ratios than lavas of the Southern Volcanic Front, studies of a volcanic transect at Petropavlosk (Kozelsky, Avachinsky, Koryaksky, Aag, Arik, Kupol and Bakening) demonstrate no clear across-arc changes in B contents or in B/Be ratios (mean B/Be: 25-30 across the arc). In contrast, data for Avachinsky indicates considerable geochemical diversity in recent lavas (B/Be range: 6-83). All the volcanoes in this transect have erupted crustal and mantle xenoliths, suggesting pervasive chemical and physical interactions between rising lavas and the crust and lithosphere. High and uniform B/Be ratios along the Petropavlosk transect may thus reflect the mean B/Be ratio of the volcanically modified crust beneath Kamchatka, while within-volcano variations derive from chemical interactions (or lack thereof) of specific magma bodies in the crustal plumbing system.[Back to top]
A COMPARATIVE GEOCHEMICAL STUDY OF META-MAFIC ROCKS OF THE SOUTHERN BLUE RIDGE
WILLSE, Keith R., RYAN, Jeffrey G.; Dept. of Geology, University of South Florida, 4202 East Fowler Ave., Tampa, FL 33620
Amphibolites, greenstones and other metamorphosed mafic rocks are common in Precambrian to Ordovician-age strata in the Blue Ridge geologic province. These rocks represent basalts, gabbros, and shallow intrusives related to 1) Late Precambrian rift-related magmatism; 2) ophiolitic assemblages emplaced during Taconic or Acadian collision, or 3) potentially, mafic magmatism associated with pre-collision subduction. We have collected new major and trace element data for mafic granulites from Winding Stair gap (NC), amphibolites from the Buck Creek mafic-ultramafic complex (NC), and greenstones from the Mount Rogers Formation (VA) and Grandfather Mountain Formations (NC), which we compare with existing data for Appalachian mafic rocks (Bakersville gabbro and Catoctin greenstones) as well as data from modern mafic volcanic settings, both to discover the igneous protoliths of these rocks, and to gain insights into the tectonic evolution of the Blue Ridge.
Amphibolites from the Buck Creek complex are relatively low in SiO2 (avg 46% wt.) and moderately high in MgO (9-10% wt.). All show >200 ppm Ni and >300 ppm Cr, which supports the contention of MacElhaney and McSween (1983) that these rocks represent gabbroic cumulates. For gabbros, the Chunky Gal samples have high contents of Ti (>0.8 % wt.) and Zr, both incompatible elements which are not remobilized during metamorphism. Assuming the gabbros reflect 10-30% crystallization of a mafic melt, primary magma TiO2 contents range from 2.4-2.8% wt., similar to Mt. Rogers and Catoctin formation greenstones, the Bakersville gabbro, and modern basalts from continental rifts.
Winding Stair Gap pyroxenites are higher in MgO, Cr, and Ni than the Chunky Gal amphibolites, suggesting a more olivine-rich gabbroic protolith, or an origin as strongly picritic basalt flows. Preliminary data suggests much lower TiO2 and other incompatible elements indicating a more chemically depleted source, possibly similar to ocean ridge source regions, or sources for volcanic arcs.
[Back to top]
RYAN, Jeffrey G.; Geology Dept., Univ. South Florida, 4202 E. Fowler Ave., Tampa, FL 33620: ryan@chuma.cas.usf.edu; BENTON, Laurie, Dept. Geosciences, Univ. Tulsa, OK 74104; TERA, Fouad, DTM-CIW, 5241 Broad Branch Rd, NW, Washington, DC, 20015
Combining the database for boron abundances in subduction-related rocks with new data for B isotope compositions provides new insights into the details of slab-mantle chemical exchanges. B abundance systematics tell a straightforward story: sediments and ocean crust enriched with B in the marine environment subduct, and progressive slab metamorphism releases B into the overriding plate beneath forearcs and arcs. Hydro-thermal redistribution of B during subduction creates a stratified B profile in the earth, with B/K ratios in arc lavas and surface sediments > mean continental crust >> MORB (upper mantle) > OIB (lower mantle). B isotope data for marine sediments suggests two compo-nents: a) labile B with 11B of +17 per mille, removed via compaction and diagenetic reactions in the trench, and b) "non-desorbable" B at -10 per mille, isotopically similar to continental crust (You 1995; Chaussidon and Albarede 1993). 11B in altered ocean crust ranges from +3 to +10 per mille (Smith et al. 1995; Ishikawa and Nakamura. 1992). The downgoing plate is isotopically diverse compared to arc lavas, which range from -4 to +7 per mille, and lean toward "heavy" B in the absence of crustal inputs. Both MORBs and OIBs have "light" B isotope signatures, at -1 to -8 per mille and -6 to -11 per mille, respectively (Chaussidon and Marty 1995).
Isotope-based models for B addition to arc sources have thus far presumed that no isotopic fractionation occurs when B is released from the slab. Slab B must therefore come from altered ocean crust, with little input from sediment. This model is consistent with current B isotope data, but 10/9Be-B/Be correlations in several arcs point to sediment contributions. Cross-arc B variations and data from subduction-related metamorphic rocks show that >>90% of B input to trenches is released from the slab by arc depths, and that sediment-hosted B mobilizes at low P-T, i.e., forearc depths.. However, our B isotope data for clasts and muds from an active serpentinite seamount in the Marianas forearc range from +6 to +15 per mille. These heavy values require the preferential extraction of 11B from the slab, +/- subduction of labile B in sediments. Heavy B may be stored in hydrated phases near the slab/wedge interface and released into arc sources via later dehydration reactions. The bulk slab will develop a distinctly negative 11B and low B/K ratio, which may be reflected in the sources of intraplate lavas.
[Back to top]
Tectonic Setting, Mineralogy and Petrology of Alkaline Rocks from the Russian Far East
Suzanne Norrell, Pavel Kepezhinskas, Marc Defant, Jeff Ryan (1), Ken Collerson (2), Vladimir Prikhodko, Anatoly Romashkin (3) (1) Department of Geology, University of South Florida, Tampa, FL 33620; Phone (813) 974-1698; email: snorrell@chuma.cas.usf.edu (2) Department of Geology, Univ. Queensland, Brisbane, Australia (3) Institute of Tectonics and Geophysics, Khabarovsk, Russia
The southeastern Siberian craton is surrounded by mobile belts which range in age from Late Proterozoic to late Tertiary, growing younger eastward toward the Pacific Ocean. The mobile belts include minor cratonic blocks rimmed by granite-greenstone belts, ophiolites, and arc-derived rocks. These blocks probably represent disaggregated parts of the North China craton. Diamondiferous alkaline rocks occur in both the cratonic blocks and in the surrounding mobile belts. Throughout Eastern Russia, they form volcanic pipes, dikes, sills, tuffs and minor lava flows.
The alkaline lavas of the Russian Far East are chemically and mineralogically diverse (picrites and alkaline basalts to lamproites and lamprophyres). Variations in major and trace elements point to diverse melting and crystallization histories. Sr, Nd and Pb isotope signatures indicate an enriched mantle source. Russian 'lamproite-like' samples are lower in K2O and Ba, and higher in CaO than lamproites as traditionally defined; but are similar in TiO2, Sr, Cr and Ni. Bent and kinked phlogopite occur as phenocrysts and in the groundmass. Clinopyroxene and olivine crystals are partially to completely replaced. Cr-spinel is present as well as perovskite, and carbonates are found both as secondary minerals and as a groundmass phase. These 'lamproite-like' rocks lack leucite, amphibole and sanidine. Diamonds have been found as float in association with these rocks. Russian 'lamprophyres' consist mainly of olivine and clinopyroxene with lesser phlogopite and amphibole. The 'ultramafic lamprophyres' include basaltic "clasts" (mainly groundmass CPX and glass with few CPX phenocrysts) surrounded by a green glass and carbonate veins.
[Back to top]
Chemical Transfer Events in Subduction Zones as a Function of Depth
Jeff Ryan, University of South Florida
In order to assess the effects of subduction-related chemical recyling, one must first develop a complete picture of the slab enrichment process. Results from recent studies of subduction-related igneous and metamorphic rocks are leading to detailed chemical descriptions of the materials released from subducting plates, and insights into the timing of these releases.
While information on slab fluxes drawn from arc volcanic data is necessarily indirect, multi-element studies of arc suites in which data from related arc centers are compared (either along an arc segment (Miller et al 1992; 1994; Edwards et al. 1993) or a volcanic cross-chain) indicate the involvement of two distinguishable slab-derived "fluxes": one with enrichments of B, Cs, As, Sb, Pb, and N U and Ba that vary with depth to the slab beneath the volcanoes and/or mean extents of slab inputs, and one with more-or-less uniformly elevated contents of K, Ba, Sr, LREE's, 10Be, and like species.
Patterns of B-Cs-As-Sb enrichment in arcs mirror abundance decline observed in "subduction complex" metamorphic massifs formed at P-T conditions analogous to slabs at forearc depths (Bebout et al 1993; Bebout 1996). These patterns suggest a progressive distillation of hydrous fluids from the subducting plate, as opposed to stepwise releases due to mineral decompositions. Alkaline elements such as K and Ba show little abundance change in these rocks. Diapiric serpentinites produced during mantle-slab fluid reactions beneath forearcs show enrichments of B and like species, and no enhancements in K, Ba, or Sr. Thus, hydrous slab fluids transport a very limited menu of elements, and they do so at shallow depths, as ratios such as B/Be and Cs/Th approach MORB/OIB values in lavas from volcanoes above deep slabs.
The "second" slab flux matches aspects of published descriptions of both "sediment melt" (Plank 1992), and "slab melt" (Kepezhinskas et al. 1996; Schiano et al. 1996) inputs: rich in SiO2, high in K, Na, Ba, Sr and other alkaline elements; with elevated LREE, Th, and (sometimes) 10Be. Fluid/solid D values are very small for many of these species (see Kepler et al. 1996), pointing to a, "melt-like" medium for their transport. This flux is a high P-T phenomenon, as forearc samples exhibit little mobility for these elements, and enrichment factors for these elements are high in back arc settings.
This current picture of the slab input process places several constraints on subduction recycling: 1) boron and other species which follow H2O recycle poorly, if at all, so the chemical cycles of these "fluid-mobile" elements must include a large return flux in forearc regions to balance extensive inputs to trenches. 2) Deeply recycled "subducted material" may represent some flavor of the "melt" flux released from slabs, or the refractory and dehydrated slab itself; all of which will be markedly depleted in the "fluid mobile" elements released early in the subduction process. The tools to resolve among the possible deep recycled components are incomplete, but the geochemical systematics of OIB sources (uniformly low Pb/Ce, U/Nb, Cs/Rb (Hart and Reid, 1991), and B/K despite significant radiogenic isotope variability) is consistent to a first order with a "subducted" influence in many mantle reservoirs.
REFERENCES CITED
G. Bebout, GSA Abst. Prog., A-254, 1996.
G. Bebout, et al., Geochim. Cosmochim. Acta 57, 2227, 1993.
Edwards et al., Nature 362, 530, 1993.
S.R. Hart and M.R. Reid Geochim. Cosmochim. Acta 55, 2379, 1991
P. Kepezhinskas et al., Geochim. Cosmochim. Acta 60, 1217, 1996.
H. Kepler, Nature 380, 237, 1996.
D. Miller, et al. J. Geophys. Res. B 97, 321, 1992
D. Miller, et al., Nature 368, 514, 1994.
T. Plank, Ph.D. diss. Columbia Univ., 1992
A.J. Stolz et al., Geology 24, 587, 1996.
P.Schiano, et al., Nature 377, 595, 1996.
[Back to top]
GEOCHEMICAL BASELINE STUDY OF UPPER FLORIDAN AQUIFER HOST ROCKS, SW FLORIDA
Ron Bek, Chris Langevin, and Jeff Ryan, University of South Florida
When trying to assess groundwater quality, one needs a baseline which relates to the