The Hard Rock Group, Department of Earth Sciences, University of Oxford
Metamorphic modelling projects
These projects have the broad aim of understanding how the Earth's crust
evolves and transforms itself, by studying the chemistry and microstructure
of small samples taken from large zones of tectonic importance, such as
collisional mountain belts, subduction zones, or volcanic arcs. The examples
may be modern or ancient; in some respects the processes have changed over
time. Our group's main focus is on the younger mountain belts of the
The advent of self-consistent thermodynamic data for common minerals and
of improved software for calculating equilibria
(e.g. THERMOCALC, Theriak/Domino) has made the calculation of
phase relationships fairly straightforward. However, the rocks that make up
young and ancient mountain belts are complex systems capable of recording a
long history of deformation and recrystallisation. Extracting this history
tests the techniques of equilibrium calculation to their limits, and in most
cases all one can do is infer an approach to equilibrium, or a local or
partial equilibrium, supported by a careful study of microstructures and of
the spatial variation of mineral compositions.
To be truly useful, our sequence of events needs the absolute timescale
supplied by radiometric dating, and our preferred method is U-Pb (or U-Th-Pb)
geochronology on suitable accessory minerals, such as monazite. Monazite has
a number of advantages: it is of widespread occurrence, has very little common
Pb, holds on to its radiogenic Pb at high temperature, and can be dated with
high precision, even by in-situ methods in thin section. Its textural
associations and chemistry can commonly be used to link its growth to that
of the major minerals in a specimen.
On this page we describe some of the recent studies where we put these
techniques to work.
Karakoram Metamorphic Complex
Our modelling activity increased markedly when Richard Palin's project was reconfigured to tackle a number of geological-tectonic questions across the Himalayan system using the same set of petrological and geochronological tools. One of these studies (Palin et al. 2012) built on work Richard undertook during his Masters research, aimed at characterising and comparing the P-T evolution of different units within the Karakoram Metamorphic Complex.
One rock in particular, a garnet-staurolite-kyanite schist from the Baltoro region, was studied in detail to extract a P-T path from its growth-zoned garnets. In the published paper, fractionation of components into the growing garnet had not been taken into account, as the rock contained only 4% garnet by volume. Nevertheless Richard later recalculated the path with fractionation included. Although the calculated paths for the two cases were very similar, he showed that the fit to the modal amounts of minerals and the relative timing of appearance of new minerals was significantly better when fractionation had been taken into account.
Work on the Karakoram metamorphic rocks continues. We are finding significant differences in age of peak metamorphism and P-T trajectory between distinct units. The Baltoro region, for example, shows garnet growth over an interval of markedy increasing pressure over the few Myr prior to the intrusion of the Baltoro granites. We have seen something similar in the Zanskar region of the Himalaya (Walker et al. 2001, Tectonics), where garnet growth accompanied the stacking of fold nappes during the prograde history.
E Tibet: the Danba Culmination
The Danba Culmination was the first of three main areas in eastern Tibet to be studied by Owen Weller during his DPhil work (Weller et al., 2013). It is an oval area some 50km in diameter in the Songpan-Garze terrane (see Tibet map below), forming a structural culmination that shows a concentric set of Barrovian metamorphic zones, cored by migmatites. There had been debate about the age of metamorphism and possibility of more than one metamorphic cycle.
In the event, this area proved to be a learning experience about the subtleties of even a simple metamorphic setting. The structural and microstructural history was straightforward and reflected shortening and homogeneous thickening of a fairly uniform sedimentary sequence during prograde metamorphism. There was no evidence for multiple tectonic or metamorphic cycles in either microstructure or age determination. It was possible to determine a metamorphic field gradient and determine some segments of P-T path. Did these fit the text-book model of a simple Barrovian metamorphism of this type, involving nested clockwise P-T paths and a field gradient at an angle to the near-peak P-T trajectories? Well, no. The P-T paths were sub-parallel to the field gradient, not strongly curved and not obviously nested.
It emerged, with the help of monazite chronology, that the probable reason was the slow burial and long time-scale for this metamorphism, which allowed thermal relaxation approximately to keep pace with thickening. High temperatures were maintained for at least 12 Myr.
SE Tibet: the Belts of Basong Tso
The Lhasa Block of southern Tibet (see map below) is now known to be divided in two by an E-W suture zone containing high-pressure rocks, including the Sumdo eclogites (currently under study by Owen Weller). Owen also studied a traverse at the eastern end of the northern Lhasa Block near the Basong Tso, where two metamorphic belts of similar age but contrasting history are juxtaposed, their actual junction concealed by a 4km-wide unit of low-grade turbidites in tectonic contact.
The southern belt, the Basong Tso complex, consists largely of protomylonitic rocks with peak metamorphic conditions around 690°C and 9 kbar, in the field of kyanite stability. Monazite ages cluster around 202 Ma, giving no indication of a multistage history. However, garnets in the dated schist show two clear growth stages, with compositionally distinct euhedral cores that contain sillimanite inclusions. Phase diagram modelling indicated that the core zones grew at <7 kbar, 630°C, and that the overgrowth records prograde growth from ~620°C at 9 kbar, implying an isothermal pressure increase that left no record in garnet growth.
Why is the garnet inner boundary so sharp? Why did no garnet grow as pressure increased? Is this really consistent with a single tectono-metamorphic cycle as indicated by the age determinations? The probable answer emerged from considering the total combined water content of the rocks concerned as a function of P and T. The smooth 'prograde' equilibration of the rocks during the ~2.5 kbar pressure increase would have required an external supply of water. In the absence of fluid, nothing could happen. The resumption of garnet growth at 9 kbar most likely reflected the first access of external water-rich fluid to the rock body.
In contrast, the northern belt (the Zhala complex) showed a high-temperature overprint at relatively low pressure on staurolite-grade assemblages. The early assemblages in both complexes formed at similar conditions, and it's possible that they shared the same early P-T evolution.
Amphiboles in Subduction Zones
Amphiboles are not uncommonly found in eclogites, but they occur in a wide range of textural associations, and show broad compositional variation, often within single crystals. Laura Airaghi and Dave Waters spent a good deal of 2013 looking at the common, relatively coarse-grained, sometimes spongy-textured sodic-calcic and calcic amphiboles found interstitially in the matrix regions of eclogites from the Western Gneiss Complex of Norway. The suite of specimens from our Museum collections covered a range of P-T environments, from low T, 15-20 kbar in the south to the UHP coesite-bearing occurrences in the north.
We used both Theriak/Domino and THERMOCALC to map out the variation of amphibole composition as a function of pressure and temperature, and were able to reproduce almost all of the observed zoning trends.
Some of the lowest-temperature eclogites contained a well-equilibrated Na-Ca-amphibole-bearing assemblage formed at about 650°C, 17 kbar, but the calculated phase diagrams indicated that amphibole would be absent from the peak assemblages of most W Norway eclogites, no matter how high the water activity.
The higher-T eclogites with post-peak amphibole clearly record regionally signifiant hydration processes that occurred at around 16 kbar, outside the stability field of feldspar, and which predate the amphibolite-facies re-equilibration that affected all the country-rocks and the outer parts of most eclogite bodies. Thus there is evidence for recycling of fluids during exhumation of the HP/UHP rocks from the subduction zone, before they stalled within the crust.
Coarse interstitial matrix amphibole formed by influx of aqueous fluid at about 16 kbar during exhumation of eclogite. Note also the atoll garnets, filled with phengite and amphibole, formed at this same stage.
Meanwhile Richard Palin was studying metabasic eclogite from a locality in the UHP Tso Morari complex in the NW Himalaya, and after extracting a P-T path for the prograde evolution (in St-Onge et al. 2012) went on to determine a similar story of aqueous fluid infiltration that generated post-peak Na-Ca-amphibole and clinozoisite (Palin et al. 2014).
Granulites beneath the Semail Ophiolite
The metamorphic sole beneath the Semail ophiolite in Oman/UAE shows an inverted metamorphic gradient with thin high-temperature garnet- and clinopyroxene-bearing amphibolites preserved at the top of the sequence in a number of places. The common association between garnet and its numerous inclusions of Cpx has been known and described for many years, but in reviewing our large collection of specimens in 2011 Dave noticed that this association locally formed discrete porphyroclastic enclaves containing a well-annealed granulite-facies assemblage, which predated the high-temperature hydration and deformation that produced the dominant and foliated hornblende-bearing assemblage (see Cowan et al. 2014). Multi-equilibrium thermobarometry on the amphibole-free association confirmed the high pressures (~12 kbar) previously suggested for the garnet-bearing sole rocks, demonstrating that these were buried far deeper than can be accounted for by the thickness of the ophiolite. These assemblages must reflect the early injection of the distal margin of Oman into hot mantle beneath the newly-formed oceanic lithosphere of the ophiolite.
Granulites in the Kohistan Arc
Over the year 2013/14 Freya George and Dave Waters have been studying the mafic granulites of the Jijal Complex at the deepest levels of the Kohistan arc in northern Pakistan, using the thesis collection of Simon Gough (2002) and building on his work. Simon described two contrasting textural styles of garnet growth in these otherwise rather uniform metagabbroic Grt-Cpx-Pl-Qz-Rt assemblages: isolated equidimensional garnets in the deeper parts (possibly of igneous origin), and corona-style textures of garnet (with quartz) surrounding Cpx in the shallower levels of the Complex. We were able not only to confirm the observations but also to quantify the extent of development of the reaction-related texture (Cpx + Pl => Grt + Qz) with structural height in the Complex using grain-transition probabilities.
Our aim was also to determine P-T paths from the marked zoning in the garnets. This was not so straightforward. The garnet-forming reaction is promoted by an increase of pressure, a decrease of temperature, or any combination of the two. It became apparent that most of the garnet growth, even of the coronas, had taken place at such high temperatures (~900°C?) that any record in the mineral compositions had been lost. The observed zoning was the result of frozen-in diffusional closure, with or without a certain amount of further garnet growth. Modelling these later stages could not be undertaken without serious consideration of the fractionation effects and the extent of exchange between the outer zones of the garnet and matrix phases. Modelled P-T paths proved to be rather sensitive to the amount of garnet rim involved in re-equilibration, so that other criteria (e.g. volumes and compositions of coexisting minerals) had to be used to constrain the most likely paths. The most plausible paths involved approximately isobaric cooling at 12-13 kbar, accompanied by growth of 5-20% more garnet.
The earlier history is consistent with formation of the Jijal gabbro precursors >110 Ma and their burial at high T between 99 and ~91 Ma as a result of magmatic thickening of the arc ('over-accretion'), followed by cooling at depth after burial.