Subduction zones are unique to Earth and fundamental in its evolution, yet we still know little on the causes and mechanisms of their initiation. Numerical models suggest far-field forcing may cause subduction initiation at weak pre-existing structures, whereas inferences from modern subduction zones suggest initiation through spontaneous lithospheric gravitational collapse. Measuring the time lag between initial lower plate burial and incipient extension in the upper plate should prove diagnostic in characterizing the subduction initiation mode. In modern systems, rocks that directly recorded initial lower plate burial should be found at the subduction interface and are inaccessible. Here we investigate a fossil system, the archetypal Semail ophiolite of Oman, which expose both lower and upper plate relics of incipient subduction stages. We show that burial of the lower plate predated upper plate extension and formation of Semail oceanic crust by at least 8 Myr, using geochronology of lower and upper plate material. Such a time lag requires far-field forced subduction initiation and provides for the first time unequivocal, direct evidence for a subduction initiation mechanism in the geological record.

No abstract

Granulitized eclogite-facies rocks exposed in the Ama Drime Massif, south Tibet, were dated by Lu-Hf garnet geochronology. Garnet from the three samples analyzed yielded Lu-Hf ages of 37.5 ± 0.8 Ma, 36.0 ± 1.9 Ma, and 33.9 ± 0.8 Ma. Eclogitic garnet growth is estimated at ca. 38 Ma, the oldest age for burial of the lower Indian crust beneath Tibet reported from the central-eastern Himalaya. Granulite-facies overprinting followed at ca. 15-13 Ma, as indicated by U-Pb zircon ages. Unlike ultrahigh-pressure eclogites of the northwest Himalaya, the Ama Drime eclogites are not characteristic of rapid burial and exhumation of a cold subducted slab. The rocks instead resulted from crustal thickening during the early stages of continental collision, and resided in the lower-middle crust for >20 m.y. before they were exhumed and reheated. These new data provide solid evidence for the Indian crust having already reached at least ~60 km thickness by the late Eocene.

Determining early orogenic processes within the Pamir-Tibet orogen represents a critical step toward constructing a comprehensive model on the tectonic evolution of the region. Here we investigate the timing and cause of prograde metamorphism of Cenozoic metamorphic rocks from the Pamir plateau through Lu-Hf geochronology, U-Pb rutile thermochronology, and garnet thermometry. Regional prograde metamorphism and heating to 750-830 °C, as constrained by thermometry, occurred between 37 and 27 Ma. Prograde growth of garnet first occurred in the South Pamir and spread to the Central Pamir during the following 10 m.y. The early metamorphism is attributed to high mantle heat flow following the ca. 45 Ma break-off of the Indian slab south of the Pamir. Our investigation confirms a long-lived thermal history of the Pamir deep crust before the Miocene, and provides a causal link between break-off, enhanced mantle heat flow, and prograde heating of the subduction hanging wall. INTRODUCTIONThe Pamir-Tibet orogen is Earth's largest and highest plateau and a prime natural laboratory for investigating how plate dynamics, regional tectonics, and surface uplift and erosion interact. Refining models for this orogen, and collisional orogens in general, requires knowledge about how the crust thickens, and how its thermal and mechanical structure changes during and after collision. Tectonic processes occurring at depth particularly impact regional dynamics. However, investigating these processes in the Pamir-Tibet orogen is hindered by the scarcity of exposed Cenozoic metamorphic rocks and the general difficulty in reconstructing their prograde his-

Metamorphic rutile from granulite facies metapelitic rocks of the Archean Pikwitonei Granulite Domain (PGD; Manitoba, Canada) provides constraints on the systematics of trace elements in rutile during high-temperature conditions and subsequent slow cooling. Compositional profiles and maps of the Zr concentrations in rutile grains (120-600 lm) from three metapelitic gneisses were acquired by electron probe micro-analysis, using a spatial resolution of down to 2 lm. Simultaneously, profiles were analysed for Nb, Cr and V, which have significantly different diffusion characteristics in rutile. The profiles of all elements show relatively homogeneous concentrations within most grains, but significant inter-grain differences even within a single thin section. Some rutile grains display a slight concentration decrease from a neighbouring garnet towards the matrix for all measured elements. The lack of diffusion profiles for all analysed elements shows that these are highly immobile in rutile and that distributions of these elements are primary and preserve prograde information. The Nb and Cr concentrations overlap with ranges that are ascribed to different provenances indicating that source discrimination based on these elements is not possible in all cases. High retentiveness for Zr implies that the Zr-in-rutile geothermometer is highly robust to diffusive re-equilibration, even during very slow cooling (<2°C Ma )1 ) from granulite facies conditions. Most grains have high Zr contents (3000-4600 ppm). Differences between high Zr contents suggest that during growth under vapour-absent conditions there may not be saturation of Zr in rutile, even if zircon is present. Therefore, several rutile grains need to be analysed in a sample to obtain a useful minimum peak temperature. The highest Zr concentrations correspond to 900°C. This is significantly higher than previous peak temperature estimates of 820°C based on two-feldspar thermometry. On a regional scale this implies that part of the PGD was affected by ultra-high temperature (UHT) metamorphism. It also implies that rutile is able to preserve primary compositions even to UHT conditions. This study shows that, if combined with textural information, Zr-in-rutile has the potential to be a very useful tool for estimating rutile crystallization temperatures and peak metamorphic conditions. For granulite facies rocks, Zr-in-rutile yields more reliable peak metamorphic temperatures than most other exchange geothermometers, which tend to partially re-equilibrate by diffusion during cooling.

No abstract

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as

The layered agpaitic nepheline syenites (kakortokites) of the Ilímaussaq complex, South Greenland, host voluminous accumulations of eudialyte-group minerals (EGM). These complex Na-Ca-zirconosilicates contain economically attractive levels of Zr, Nb and rare-earth elements (REE), but have commonly undergone extensive autometasomatic/hydrothermal alteration to a variety of secondary mineral assemblages. Three EGM alteration assemblages are recognized, characterized by the secondary zirconosilicates catapleiite, zircon and gittinsite. Theoretical petrogenetic grid models are constructed to assess mineral stabilities in terms of component activities in the late-stage melts and fluids. Widespread alteration of EGM to catapleiite records an overall increase in water activity, and reflects interaction of EGM with late-magmatic Na-, Cl- and F-rich aqueous fluids at the final stages of kakortokite crystallization. Localized alteration of EGM and catapleiite to the rare Ca-Zr silicate gittinsite, previously unidentified at Ilímaussaq, requires an increase in CaO activity and suggests post-magmatic interaction with Ca-Sr bearing aqueous fluids. The pseudomorphic replacement of EGM in the kakortokites was not found to be associated with significant remobilization of the primary Zr, Nb and REE mineralization, regardless of the high concentrations of potential transporting ligands such as F and Cl. We infer that the immobile behaviour essentially reflects the neutral to basic character of the late-magmatic fluids, in which REE-F compounds are insoluble and remobilization of REE as Cl complexes is inhibited by precipitation of nacareniobsite-(Ce) and various Ca-REE silicates. A subsequent decrease in F– activity would furthermore restrict the mobility of Zr as hydroxyl-fluoride complexes, and promote precipitation of the secondary zirconosilicates within the confines of the replaced EGM domains.