Search

Late Miocene to Pleistocene volcanism within the Vardar zone (North Macedonia) covers a large area, has a wide range in composition and it is largely connected to the tectonic evolution of the South Balkan extensional system, the northern part of the Aegean extensional regime. The scattered potassic to ultrapotassic volcanism developed south from the Scutari-Peć fault zone since 6.57 Ma [1]. The focus of this study is on three volcanic centers located on deep structures or thrust faults along the western part of the Vardar zone, for which there is none to very little geochronological and geochemical data available. Pakoševo and Debrište localities are represented as small remnants of lava flows cropping out at the southern edge of Skopje basin and at the western edge of Tikveš basin, respectively. Šumovit Greben center is considered as part of the Kožuf-Kozjak/Voras massif (6.5-1.8 Ma [1]), and it is located on its westernmost side, at the southern edge of Mariovo basin, which is largely comprised of volcanoclastic sediments. Here we present new eruption ages applying the unspiked Cassignol-Gillot K-Ar technique on groundmass, petrological and geochemical data, supplemented with Sr and Nd isotopes to complement and better understand the Neogene-Quaternary volcanism in the region. Obtaining the eruption ages of these volcanic centers could also help to better constrain the evolution of the sedimentary basins. All of the three centers belong to the shoshonitic series based on their elevated K-content. The oldest center amongst these three localities, as well as other Late Miocene centers within the region, is the trachyandesitic Debrište, which formed at ca. 8.1 Ma, and exhibits the highest Nd isotopic ratios (0.512441-0.512535). The trachybasaltic Pakoševo center formed at ca. 3.8 Ma and, based on its Nd isotopic ratio (0.512260), represents the strongest sign of crustal contamination. The rhyolitic Šumovit Greben center is a composite volcanic structure formed at ca. 3.0-2.7 Ma. Its youngest eruption unit has a slightly larger Nd isotopic ratio (0.512382), representing a less evolved magma at the end of its activity.This research was funded by the GINOP-2.3.2-15-2016-00009 ‘ICER’ project, the French-Hungarian Cooperation Program TÉT-FR-2018-00018 and the HORIZON 2020 grant N 676564.References:[1] Yanev et al., 2008 – Mineralogy and Petrology, 94(1-2), 45-60.

The chemical and physical processes operating during subduction-zonemetamorphismcan profoundly influence the cycling of elements on Earth. Deep-Earth carbon (C) cycling and mobility in subduction zones has been of particular recent interest to the scientific community. Here, we present textural and geochemical data (CO, Sr isotopes and bulk and in-situ trace element concentrations) for a suite of ophicarbonate rocks (carbonate-bearing serpentinites) metamorphosed over a range of peak pressure-temperature (P-T) conditions together representing a progradesubduction zoneP-Tpath. These rocks, in order of increasing peakP-Tconditions, are the Internal Liguride ophicarbonates (from the Bracco unit, N. Apennines), pumpellyite- and blueschist-facies ophicarbonates from the Sestri-Voltaggio zone (W. Ligurian Alps) and the Queyras (W. Alps), respectively, and eclogite-facies ophicarbonates from the Voltri Massif. The Bracco oceanic ophicarbonates retain breccia-like textures associated with their seafloor hydrothermal and sedimentary origins. Their trace element concentrations and δ18OVSMOW(+15.6 to +18.2‰), δ13CVPDB(+1.1 to +2.5‰) and their87Sr/86Sr (0.7058 to 0.7068), appear to reflect equilibration during Jurassic seawater-rock interactions. Intense shear deformation characterizes the more deeply subducted ophicarbonates, in which prominentcalciterecrystallization and carbonation ofserpentiniteclasts occurred. Theisotopic compositionsof the pumpellyite-facies ophicarbonates overlap those of their oceanic equivalents whereas the most deformed blueschist-facies sample shows enrichments in radiogenic Sr (87Sr/86Sr = 0.7075) and depletion in13C (with δ13C as low as −2.0‰). These differing textural and geochemical features for the two suites reflect interaction with fluids in closed and open systems, respectively. The higher-P-metamorphosed ophicarbonates show strong shear textures, with coexistingantigoriteand dolomite, carbonate veins crosscutting progradeantigoritefoliation and, in some cases, relics of magnesite-nodules enclosed in the foliation. These rocks are characterized by lower δ18O (+10.3 to 13.0‰), enrichment in radiogenic Sr (87Sr/86Sr up to 0.7096) and enrichment in incompatible and fluid-mobile element (FME; e.g., As, Sb, Pb). These data seemingly reflect interaction with externally-derived metamorphic fluids and the infiltrating fluids likely were derived from dehydrating serpentinites with hybrid serpentinite-sediment compositions. The interaction between these two lithologies could have occurred prior to or after dehydration of the serpentinites elsewhere. We suggest thatdecarbonationand dissolution/precipitation processes operating in ancient subduction zones, and resulting in the mobilization of C, are best traced by a combination of detailed field and petrographic observations, C, O and Sr isotope systematics (i.e., 3D isotopes), and FME inventories. Demonstration of such processes is key to advancing our understanding of the influence ofsubduction zonemetamorphismon the mobilization of C in subducting reservoirs and the efficiency of delivery of this C to depths beneathvolcanic arcsand into the deeper mantle.

The Pleistocene (~460–265 ka) San Venanzo volcanic complex belongs to the IAP (Intra-Apennine Province) in central Italy, which comprises at least four small Pleistocene monogenetic volcanoes plus several unrootedpyroclastic depositswith peculiar mineralogical and whole-rock chemical compositions. San Venanzo products are strongly SiO2-undersaturated, CaO- and MgO-rich and show ultrapotassic serial character. The relatively common occurrence ofcalcitein the pyroclastic rocks and the overall high CaO content are interpreted in literature as primary mineral. The main rockfaciesat San Venanzo are calcite-rich scoria and lapilli tuffs, with minor massivelava flows, and a rare pegmatoid variant (melilitolitic pockets). All the San Venanzo rocks are feldspar-free, with a typicalparagenesisof forsteritic olivine, non-stoichiometric Ca-richdiopside,melilite, leucite,kalsilite, opaque minerals,nepheline,phlogopite,calcite,apatite, cuspidine,wollastonite, kirschsteinite-monticellite s.s. ± glass and other minor and very rare minerals typical of agpaitic melts. Based on petrographic analyses, the studied rocks can be classified as olivine melilitites, olivine leucite melilitites, venanzites (a local variant of kamafugites), calcite leucite melilitolites and Ca-rich olivine leucitemelilititetuffs. Mass balance calculations indicate a direct genetic link between the lava bodies and the pegmatoid melilitolitic pocket through afractional crystallizationprocess characterized by the removal of ~74% of a melilite-bearing uganditic cumulate made up ofmelilite, leucite, olivine,kalsiliteandchromite. Primitive mantle-normalized patterns of the lavas and tuffs are rather spiked and share negative anomalies for Ba, Nb, Ta, P and Ti resembling typicalmagmasgenerated by supra-subduction mantle wedge. These compositions are very different from the only two other kamafugite localities outside Italy (Toro Ankole and Virunga in the East Africa Rift and Alto ParanaibaIgneous Provincein SE Brazil). The melilitolite sample is more incompatible element-enriched than the other San Venanzo volcanic rocks, coherently with its evolved liquid composition proposed here. Major and trace element contents indicate a general depletion proportional to the amount of CaO content. The negative trends in Harker-type diagrams with CaO as abscissa are compatible with a process of variable interaction between a silicate magma with sedimentary marly carbonates/limestones. The presence of Mg-rich (Fo97–92) and rim-ward CaO-enriched (up to 1.72wt%) euhedral olivine, as well as the presence of thin kirschsteinite rim around olivine crystals agree with a process of crustal carbonate assimilation by an originally strongly SiO2-undersaturated silicate magma. On the other hand, the lack of feldspars even in the rocks with the highest SiO2, the high CaO content, and the extreme SiO2-undersaturation of San Venanzo rocks exclude their derivation from a simple peridotitic source. In order to generate these peculiar compositions, the presence of a SiO2-K2O-CaO-rich H2O-bearing component, identified in a carbonatedphlogopiteperidotiteis required. The results of different isotopic systematics (Sr-Nd-Pb-He-Ne-Ar) presented here are compatible with a process ofcrustal contaminationboth atmantle sourcelevels (to explain the general N-S isotopic trends recorded in Quaternary volcanic rocks of Italianpeninsulaand Sicily) and with interaction of ultrabasic melts with limestones at shallow crustal depths.

Boron has become the principle proxy for the release of seawater-derived fluids into arc volcanics, linked to cross-arc variations in boron content and isotopic ratio. Because all ocean floor serpentinites so far analysed are strongly enriched in boron, it is generally assumed that if the uppermost slab mantle is hydrated, it will also be enriched in boron. Here we present the first measurements of boron and boron isotopes in fast-spread oceanic gabbros in the Pacific, showing strong take-up of seawater-derived boron during alteration. We show that in one-pass hydration of the upper mantle, as proposed for bend fault serpentinisation, boron will not reach the hydrated slab mantle. Only prolonged hydrothermal circulation, for example in a long-lived transform fault, can add significant boron to the slab mantle. We conclude that hydrated mantle in subducting slabs will only rarely contribute to boron enrichment in arc volcanics, or to deep mantle recycling., Boron is one of the main proxies for seawater-derived fluids in subduction zone volcanics and it is vital to characterise the location and concentration of boron in the oceanic lithosphere. Here the authors show that boron concentration in the mantle of downgoing slabs has been overestimated, because boron is strongly decoupled from water in the hydration process.

The Elazig and Tunceli provinces in eastern Anatolia host a complex succession of Miocene-Pleistocene effusive and explosive volcanic rocks, divided into four distinct volcanic phases. The most abundant and widespread products are the calcalkaline Mazgirt volcanic rocks, characterized by wide Sr isotope variations (87Sr/86Sr ~0.7054-0.7077) and narrower 143Nd/144Nd (~0.51246-0.51260) and Pb isotopes (e.g., 206Pb/204Pb ~18.89-19.13). New 40Ar-39Ar ages indicate that Mazgirt volcanic activity occurred between ~16.3 and 15.1 Ma. The other three volcanic phases are represented by the Tunceli mildly alkaline basaltic lavas (~11.4-11.0 Ma), the Pliocene Karakocan (~4.1 Ma) and Pleistocene Elazig (~1.9-1.6 Ma) Na-alkali basaltic lavas with clear OIB-like geochemical signature.Mazgirt volcanics can be subdivided on the base of mode of emplacement into lava flows and lava domes units characterized by petrographic, chemical and isotopic differences: lava flows are calcalkaline, whereas lava domes mostly belong to a high-K calcalkaline series and are, on average, more LREE- and 87Sr-enriched. Lava domes are more porphyritic, with a phenocryst assemblage dominated by amphibole, whereas plagioclase and clinopyroxene are the most abundant phenocryst phases in lava flows, pointing out that evolution of dome magmas occurred in conditions of slightly higher pressure, favouring the crystallization of hydrous phases.The Karabakir Formation, previously reported as late Miocene- Pliocene, encloses Mazgirt volcanics and is capped by Tunceli basalts. These new age data constrain the Karabakir Formation emplacement from early to late Miocene.The evolution of this igneous activity mirrors the geodynamic framework of the region: the early-middle Miocene Mazgirt volcanics represent arc volcanism related to Eurasia-Arabia convergence. The late Miocene Tunceli basalts postdate the onset of post-collisional tectonics in Eastern Anatolia, whereas the Karakocan and Elazig volcanic rocks were emplaced after the initiation of strike-slip motion on the North Anatolian and East Anatolian Fault systems.

Leucite-bearing volcanic rocks are commonly found within and around the Mediterranean area. A specific type of this rock group are leucitites. They are found both in a hinterland position of active and fossil subduction systems as well as in foreland tectonic settings, but none have been found in the Maghreb (N Africa) and Mashreq (Middle East) areas. Here a review of the main leucitite occurrences in the circum-Mediterranean area is presented, with new whole-rock, mineral chemical and Sr-Nd-Pb isotopic ratios on key districts, with the aim of clarifying the classification and genesis of this rock type. Many of the rocks classified in literature as leucitites do not conform to the IUGS definition of leucitite (i.e., rocks with >10 vol% modal leucite and with foids/(foids + feldspars) ratio > 0.9, with leucite being the most abundant foid). Among circum-Mediterranean rocks classified as leucitites in the literature, we distinguish two types: clinopyroxene-olivine-phyric (COP) and leucite- phyric (LP) types. Only the second group can be truly classified as leucitite, being characterized by the absence or the very rare presence of feldspars, as well as by ultrapotassic composition. The COP group can be distinguished from the LP group on the basis of lower SiO2, Na2O + K2O, K2O/Na2O, Al2O3, Rb and Ba, and higher MgO, TiO2, Nb, Cr and Ni. The LP group shows multi-elemental patterns resembling magmas emplaced in subduction-related settings, while COP rocks are much more variable, showing HIMU-OIB-like to subduction-related-like incompatible element patterns. COP rocks are also characterized generally by more homogeneous isotopic compositions clustering towards low Sr and high Nd isotopic ratios, while LP leucitites plot all in the enriched Sr-Nd isotopic quadrant. LP rocks usually have lower 206Pb/204Pb and higher 207Pb/204Pb. This study shows that the geochemical signal of mantle melts does not always reflect the tectonic setting of magma emplacement, suggesting paying extreme attention in proposing geodynamic reconstructions on the basis of chemical data only.

The late Miocene (11–9 Ma) volcanic rocks of Kef Hahouner, ~40 km NE of Constantine (NE Algeria), are commonly classified as lamproites in literature. However, these rocks are characterized by an anhydrous paragenesis with plagioclase and Mg-rich olivine phenocrysts, set in a groundmass made up of feldspars, pyroxenes and opaque minerals. Thus, we classify the Kef Hahouner rocks as ultrapotassic shoshonites and latites, having K2O N 3 wt.%, K2O/Na2O N 2.5, MgO N 3–4 wt.%, SiO2 b 55–57 wt.% and SiO2/K2O b 15. All the investigated samples showprimitive mantle-normalizedmulti-element patterns typical of orogenic (arctype) magmas, i.e. enriched in LILE (e.g. Cs, Rb and Ba) and LREE (e.g. La/Yb=37–59) with respect to the HFSE, peaks at Pb and troughs at Nb and Ta. Initial isotopic ratios are in the range of 87Sr/86Sr = 0.70874–0.70961, 143Nd/144Nd = 0.51222–0.51223, 206Pb/204Pb = 18.54–18.60, 207Pb/204Pb = 15.62–15.70 and 208Pb/204Pb = 38.88–39.16. The Kef Hahouner volcanic rocks show multi-element patterns similar to the other circum-Mediterranean lamproites and extreme Sr, Nd and Pb isotopic compositions. Nevertheless, the abundant plagioclase, the presence of Al-rich augite coupled with high Al2O3 whole rock compositions (9.6–21.4 wt.%), and the absence of phlogopite are all at inconsistentwith the definition of lamproite.Wereviewed the rocks classified as lamproites worldwide, and found that many of these rocks, as for the Kef Hahouner samples, should be actually defined as “normal” potassic to ultrapotassic volcanic rocks. Even the grouping of lamproites into “orogenic” and “anorogenic” types appears questionable.

Isla El Tigre volcano is located in the Gulf of Fonseca (Honduras) along the Central America volcanic front, where a significant change in the strike of the volcanic chain is observed. The studied samples of this poorly investigated volcano are mainly subalkaline basic to intermediate lavas (basalts and basaltic andesites) and subordinate subalkaline/alkaline transitional basalts, both having the typical mineralogical and geochemical characteristics of arc volcanic rocks. On the basis of petrographic and geochemical features, two groups of rocks have been distinguished. Lavas from the main volcanic edifice are highly porphyritic and hy-qz normative, and have lower MgO contents (< 5 wt.%). They show significant LILE and LREE enrichments and Nb-Ta depletions, and have a strong slab signature as well as incompatible element contents similar to those of the main front of the adjacent volcanoes in El Salvador and Nicaragua (e.g., Ba/La up to 80). In contrast, lavas from the parasitic cones have higher MgO contents (> 5 wt.%), are ol-hy normative and show lower HFSE depletions relative to LILE and LREE, with lower Ba/La, Ba/Nb and Zr/Nb ratios. This suggests that mantle-derived magmas were not produced by the same process throughout the activity of the volcano. The bulk rock geochemistry and 87Sr/86Sr (0.70373-0.70382), 143Nd/144Nd (0.51298-0.51301), 206Pb/204Pb (18.55-18.58), 207Pb/204Pb (15.54-15.56) and 208Pb/204Pb (38.23-38.26) isotopic data of Isla El Tigre compared with the other volcanoes of the Gulf of Fonseca and all available literature data for Central America suggests that this stratovolcano was mainly built by mantle-derived melts driven by slab-derived fluid-flux melting, while magmas erupted through its parasitic cones have a clear signature of decompression melting with minor slab contribution. The coexistence of these two different mantle melting generation processes is likely related to the complex geodynamic setting of the Gulf of Fonseca, where the volcanic front changes direction by ca. 30° and two fundamental tectonic structures of the Chortis continental block, mainly the N-S Honduras Depression and the NE-SW Guayape Fault Zone, cross each other.

This work was supported as part of Patrick Sugden's PhD studentship funded through the Leeds-York SPHERES NERC doctoral training partnership (NE/L002754/1). Part of the fieldwork and research was funded by the University of Leeds, the Carnegie Institution of Washington, the ERASMUS exchange programme (for IS) and the Science Committee of the Armenian Ministry of Education and Science (project #18T-1E368). The majority of the B isotope analytical work was supported by IGG-CNR funds P1600514. We present the first boron abundance and δ11B data for young (1.5-0 Ma) volcanic rocks formed in an active continent-continent collision zone. The δ11B of post-collisional volcanic rocks (−5 to +2‰) from the Armenian sector of the Arabia-Eurasia collision zone are heavier than mid-ocean ridge basalts (MORB), confirming trace element and isotope evidence for their derivation from a subduction-modified mantle source. Based on the low B/Nb (0.03-0.25 vs 0.2-90 in arc magmas), as well as low Ba/Th and Pb/Ce, this source records a subduction signature which is presently fluid-mobile element depleted relative to most arc settings. The heavier than MORB δ11B of post-collision volcanic rocks argues against derivation of their subduction signature from a stalled slab, which would be expected to produce a component with a lighter than MORB B, due to previous fluid depletion. Instead, the similarity of δ11B in Plio-Pleistocene post-collision to 41 Ma alkaline igneous rocks also from Armenia (and also presented in this study), suggests that the subduction signature is inherited from Mesozoic-Paleogene subduction of Neotethys oceanic slabs. The slab component is then stored in the mantle lithosphere in amphibole, which is consistent with the low [B] in both Armenian volcanic rocks and metasomatic amphibole in mantle xenoliths. Based on trace element and radiogenic isotope systematics, this slab component is thought to be dominated by sediment melts (or supercritical fluids). Previously published δ11B of metasediments suggests a sediment-derived metasomatic agent could produce the B isotope composition observed in Armenian volcanic rocks. The lack of evidence for aqueous fluids preserved over the 40 Myr since initial collision supports observations that this latter component is transitory, while the lifetime of sediment melts/supercritical fluids can be extended to >40 Myr. Publisher PDF

Recent geochemical work shows that subduction-zone serpentinites are repositories for fluid-mobile elements absorbed during interaction with sediment-derived fluids. Unraveling the geochemical fingerprint of these rocks helps to define timing of tectonic accretion of sediments along the subduction interface and the role ofserpentinitein element recycling tovolcanic arcs. Here we present the trace element andisotopic composition(B–O–H, Sr, Pb) of high-pressure serpentinites from the Voltri Massif (Ligurian Western Alps, Italy), to discuss their role as incompatible element carriers and their contribution to recycling of sediment-derived components in subduction zones. The serpentinites presented here record metamorphic olivine growth during eclogite-faciesmetamorphismand show undeformed and mylonitic textures. Field relations show that undeformed rocks are enclosed in deformed ones and that nometasedimentary rocksare present nearby. Undeformed serpentinite has very high δ11BSRM951(from +26‰ to +30‰), low Sr and Pbisotope ratios(87Sr/86Sr=0.7053–0.7069;206Pb/204Pb=18.131–18.205) and low As and Sb contents (0.1 and 0.01μg/g, respectively). Oxygen and hydrogen isotope compositions are +4.5‰ and −67‰, respectively. In contrast, mylonitic serpentinite shows lower δ11B (from +22‰ to +17‰), significant enrichment in radiogenic Sr and Pb isotopes (87Sr/86Sr up to 0.7105;206Pb/204Pb up to 18.725), and enrichment in As and Sb (1.3 and 0.39μg/g, respectively). δ18O of the mylonitic serpentinites reaches values of +5.9‰, whereas δD is comparable with that of undeformed rocks (approximately −70‰). In mylonitic serpentinites, the B and Sr isotopic values and the fluid-mobile element (FME) concentrations are near those for the Voltri metasedimentary rocks (calc- and mica-schists). Pb systematics also reveal influx of a crust-derived component. Our dataset shows that undeformed serpentinite still preserves an oceanic geochemical fingerprint, whereas mylonitic serpentinite is reset in its concentrations of FME and its B, Sr and Pb isotope compositions, due to interaction with sediment- and crust-derived fluids. The environment of this interaction is either compatible with (i) an outer-rise zone setting, withpercolationof seawater-derived fluids enriched in sedimentary components into bending-related fault structures, or with (ii) subduction channel domains, where ascending sediment-derived slab fluids infiltrate slices of former oceanic serpentinite accreted to the plate interface domain. Influx of sediment-derived subduction fluids along major deformation zones in serpentinite modifies the element budget of the rocks, with important implications for element recycling and the tectonic history of serpentinite. The B, Sr and Pb isotopic systematics, coupled with FME concentration in serpentinites are particularly helpful geochemical tracers of interaction between different reservoirs in subduction-interface environments, and are more sensitive than the traditionally applied stable oxygen and hydrogen isotope compositions.

The Estancia Glencross Area (EGA) volcanic rocks form a series of five isolated buttes located at the southern end (∼52°S) of the discontinuous belt of Cenozoic basaltic lava formations occurring in the extra-Andean Patagonia. EGA volcanics are subalkalinebasaltsand basalticandesiteserupted at 8.0–8.5 Ma in a region closely behind the Andean Cordillera. EGAvolcanismpredated by about 4–5 my the onset of the volcanism in the nearby Pali Aike Volcanic Field, which produced highly primitive, alkaline lavas. Incompatible trace-element distributions and Sr–Nd isotope compositions of EGA rocks are those typical of within-plate OIB-type basalts and are indicative of minimal interaction of sub-lithosphericmagmaswith enriched reservoirs. The geochemical characteristics of EGA volcanics, as well as their age and location are consistent with a model of slab window opening beneath this region. The high silica content and thegarnetsignature of the estimated EGA primary magma are explained by a two-stage process involving the initial production of melts from a garnetlherzolitesource followed by the reaction of these melts withharzburgitecountry rocks during their ascent through the mantle lithosphere. The melt/harzburgite reaction, favoured by a slow melt ascent rate, as well as the low magma production at EGA, are likely related to the dominantly compressive stress regime operating in this area during LateMiocene.

The Cenozoic igneous activity of Sardinia is essentially concentrated in the 38-0.1 Myr time range. On the basis of volcanological, petrographic, mineralogical, geochemical and isotopic considerations, two main rock types can be defined. The first group, here defined SR (Subduction-Related) comprises Late Eocene-Middle Miocene (~ 38-15 Ma) igneous rocks, essentially developed along the Sardinian Trough, a N-S oriented graben developed during the Late Oligocene-Middle Miocene. The climax of magmatism is recorded during the Early Miocene (~ 23-18 Ma) with minor activity before and after this time range. Major and trace element indicators, as well as Sr-Nd-Pb-Hf-Os-O isotope systematic indicate complex petrogenetic processes including subduction-related metasomatism, variable degrees of crustal contamination at shallow depths, fractional crystallization and basic rock partial melting. Hybridization processes between mantle and crustal melts and between pure mantle and crustally contaminated mantle melts increased the isotopic and elemental variability of the composition of the evolved (intermediate to acid) melts. The earliest igneous activity, pre-dating the Early Miocene magmatic climax, is related to the pushing effects exerted by the Alpine Tethys over the Hercynian or older lower crust, rather than to dehydration processes of the oceanic plate itself. The second group comprises volcanic rocks emplaced from ~ 12 to ~ 0.1 Ma. The major and, partially, trace element content of these rocks roughly resemble magmas emplaced in within-plate tectonic settings. From a Sr-Nd-Pb-Hf-Os isotopic point of view, it is possible to subdivide these rocks in two subgroups. The first, defined RPV (Radiogenic Pb Volcanic) group comprises the oldest and very rare products (~ 12-4.4 Ma) occurring only in the southern sectors of Sardinia. The second group, defined UPV (Unradiogenic Pb Volcanic), comprises rocks emplaced in the remaining central and northern sectors during the ~ 4.8-0.1 Ma time range. The origin of the RPV rocks remains quite enigmatic, since they formed just a few Myr after the end of a subduction-related igneous activity but do not show any evidence of slab-derived metasomatic effects. In contrast, the complex origin of the mafic UPV rocks, characterized by low 206Pb/204Pb (17.4-18.1), low 143Nd/144Nd (0.51232-0.51264), low 176Hf/177Hf (0.28258-0.28280), mildly radiogenic 87Sr/86Sr (~ 0.7044) and radiogenic 187Os/188Os ratios (0.125-0.160) can be explained with a mantle source modified after interaction with ancient delaminated lower crustal lithologies. The strong isotopic difference between the RPV and UPV magmas and the absence of lower crustal-related features in the SR and RPV remain aspects to be solved.

The Foca-Karaburun and Ayvalik-Lesvos grabens (western coast of Anatolia, Turkey) are two important NW-SE-trending extensional areas generated in response to the Early Miocene-Holocene extension of the Western Anatolian region, related to the opening of the 'unconventional' back-arc basin of the Aegean Sea. The abundance of geo-structural evidence and the occurrence of volcanic rocks representing all the stages of the Aegean-Western Anatolia volcanism render the Foca-Karaburun and Ayvalik-Lesvos Grabens key localities to exemplify the petrogenetic and geodynamic evolution of the area. In this context, the Foca-Karaburun and Ayvalik-Lesvos grabens, possibly formerly a single graben, formed along an original NE-SW-trending extension, later dissected by E-W-trending transtensional faults, are investigated to constrain the petrogenetic and geodynamic evolution of the whole Aegean region. Calc-alkaline and shoshonitic volcanic rocks with scattered ultrapotassic-shoshonitic or lamproitic lavas and dykes represent the orogenic phase of the magmatic activity, while the younger K-and Na-rich alkaline basaltic rocks are the result of later magmatism characterized by an intraplate geochemical signature reflecting progressively decreasing subduction rates.

[1] A comprehensive geochemical study, including B, Pb, Sr, and Nd isotopes, has been carried out on El Salvador subduction-related lavas. The rocks have arc-type incompatible element distributions with high LILE/HFSE ratios, nearly constant 143Nd/144Nd (≈0.5130), and small differences in 207Pb/204Pb (15.53–15.57), whereas 87Sr/86Sr ranges from 0.7035 to 0.7039. Boron isotopic composition varies widely, between −2.7‰ and +6.3‰. The boron isotope signature points to involvement of fluid inputs from (1) a high-δ11B serpentinite fluid from serpentized mantle wedge dragged beneath the volcanic arc or from the subducting lithosphere and (2) a low-δ11B fluid from the progressive dehydration of subducted altered basaltic crust and/or sediments. The observed sample variability is explained with a model in which different proportions of serpentinite-derived (10–50%) and slab-derived fluids are added to an enriched-DMM source, triggering its partial melting. We suggest a model in which tectonic erosion, i.e., dragging down of slivers of serpentinized upper plate mantle, was responsible for the occurrence of serpentinite reservoir, 11B-enriched in the forearc by shallow fluids.

In El Salvador, Central America, active deformation takes the form of a major dextral strike-slip fault system, the El Salvador Fault Zone, resulting from the oblique subduction of the Cocos Plate. The fault system is laterally discontinuous, being subdivided into different major en-echelon segments that partially overlap to form pull-apart structures. Volcanic activity is spatially confined to the fault segments and absent in the intervening pull-apart basins; no significant temporal gap exists in the erupted products, at least during the Plio-Quaternary. Detailed analyses within the geothermal fields of Berlín and Ahuachapán have revealed important volcano-structural and petrologic differences between the two areas. In the Berlín area active deformation is controlled by the regional transcurrent stress field, resulting in the development of systems of right-lateral E–W-trending strike-slip faults. Conversely, the structural setting of the Ahuachapán area is more complex, reflecting an interaction among different stress fields. Berlín products exhibit a marked geochemical and isotopic homogeneity indicating the presence of a single magmatic system. At Ahuachapán, on the other hand, the rocks display significant variations in both Sr isotopes and the LILE/HFSE ratios: this area is characterized by multiple volcanic centres, fed by different magma batches that reach the surface without reciprocal interactions in shallow reservoirs. Thus, the characteristics of the volcanic products at Berlín and Ahuachapán reflect their different tectonic settings, with important implications for geothermal investigations.

The Western Anatolia Miocene-to-Present Day magmatism evolved from calc-alkaline and shoshonitic rocks (21-16 Ma) to lamproites (16-14 Ma), and eventually into OIB-type magmas (2-0 Ma) represented by the Kula volcanics. In the calc-alkaline and shoshonitic association, Sr and Nd isotopic ratios and trace element variations suggest that the interaction with the crust was moderate, so that the geochemistry of these rocks is considered to reflect the heterogeneous chemical nature of their mantle source. The ultrapotassic and lamproitic rocks are characterised by a high Sr and low Nd isotopic composition and are strongly enriched in K and Rb with respect to Ba, indicating a phlogopite-bearing lithospheric source. Low Sr and high Nd isotopic compositions, together with low LILE/HFSE ratios, reveal the OIB-type nature of the Kula volcanics. Therefore, the products switch from supra-subduction orogenic suites to volcanics coming from sub-slab astenospheric mantle. The evolution is interpreted as being due to a 'horizontal' stretching of the slab (no slab pull break-off) generated by different velocities in the subduction hangingwall lithosphere. This triggered the extensional movement between Greece and Turkey and the stretching into two slabs of the NE-directed African subduction, due to the faster southwestward slab rollback of Africa underneath Greece relative to the slab segment below Cyprus and Anatolia. (c) 2005 Elsevier B.V. All rights reserved.

Scattered Late Miocene high-Mg basaltic andesites to dacites can be found in Western Anatolia. These rocks display Mg#>65, high CaO/Al2O3 ratio, low alkalies and TiO2 contents. Trace element distribution shows typical orogenic signature with higher values of Fluid Mobile Elements and lower values of HREE and HFSE with respect to Early Miocene Western Anatolia calc-alkaline rocks. The 87Sr/86Sr and 143Nd/144Nd ratios virtually overlap the values of the less evolved calc-alkaline rocks. The variations observed in this association have been attributed to an FC process combined with interaction with crustal material. A thermal anomaly affecting a depleted mantle source has been invoked for the genesis of these products. Such an anomaly was produced by the ascent of deep sub-slab mantle, which replaced the underthrust lithosphere, already thinned and stretched by extensional process.

The Pali Aike Volcanic Field (PAVF) represents the southernmost occurrence of the Cenozoic back-arc Patagonian Plateau Lavas. Its activity (Pliocene–Recent) started forming tabular lavas followed by the growth of about 470 essentially monogenetic volcanic centers (tuff-rings, maars, spatter and scoria cones). Azimuths of cone alignment, cone elongation and morphologic lineations show prevailing ENE–WSW and NW–SE trends. Erupted products consist mainly of alkaline basalt and basanite, with minor olivine basalt. PAVF rocks are quite primitive in composition (average Mg#=66, Ni=220ppm and Cr=313ppm) with relatively high TiO2(average 3.0wt.%). Ultramafic garnet- and/or spinel-bearing xenoliths are found within PAVF volcanics. Chondrite-normalized REE patterns are significantly LREE-enriched and almost rectilinear [(La/Yb)N=10.9–21.0]. Primordial mantle-normalized distributions of incompatible trace elements, as well as Sr and Nd isotope ratios (87Sr/86Sr=0.70317–0.70339,143Nd/144Nd=0.51290–0.51294), show values typical of intra-plate basalts, despite the fact that these rocks occur only 200km east of the Andean Cordillera. Primary magmas were generated from a fertile garnet-bearing asthenospheric source atP=1.9–2.9GPa andT=1420–1470°C. The data suggest a geodynamic model that implies sub-slab asthenosphere flow through a slab window, which started opening below this sector of South America 14m.y. ago as a consequence of the collision of the Chile Ridge with the Chile Trench. The trailing edge of the Nazca Plate crossed below the Pali Aike area at 9–10Ma, that is 6–5m.y. before the onset of the volcanic activity. We hypothesize that this time delay resulted from changes in the kinematics of the South America–Scotia transform plate boundary which only allowed the Pali Aike magmas to rise after about 4m.y.