Oceanic Crust Metamorphism on Macquarie Island

From: Metamorphic Evolution of Australia: A Review and Metamorphic Map

Geoscience Australia Report 2015

Ben Goscombe, Richard Blewett, Peter Skirrow, Geoff Fraser, Chris Carson, Karol Carzanota, Ben Wade, John Everard, Terry Brennan, Others at Geoscience Australia

The PT evolution of an example of oceanic crust in Australian territory (Macquarie Island) has been constrained using pre-existing and new petrological and mineral data for the OZMP (Goscombe and Everard, 1997, 1999, 2001). Oceanic crust metamorphism conditions in each crustal rock unit are constrained by PT calculations using THERMOCALC and a range of conventional geothermobarometers, as well as qualitative estimates from the developed mineral assemblages (see below). These PT constraints are applied to a range or different parageneses in different post-magmatic textural settings to document the different stages of ocean crust metamorphism. For example: (1) ductile mylonites and shear bands, (2) sub-ductile shear bands, (3) vein-filled sub-ductile dilational faults, (4) dry gouge faults, (5) dilation cleavages and fractures, (6) veins, (7) vesicle vug infill and (8) penetrative alteration assemblages in the rock mass. In addition, methods of PT estimation are applied to magmatic parageneses to define starting conditions at crust formation and immediately after. These parageneses include: (1) modelled magma temperatures, (2) magmatic crystallization, (3) near-crystallization temperatures and (4) re-equilibrated temperatures. Note that post initial crust formation magmatic events such as: (1) late-D1 gabbro veins/dykes and phlogopite-feldspar-zircon pegmatite veins, and (2) D2 upper-crustal dolerite dykes, accompanied the cooling history and over-print the already formed retrograde features. In addition to the complex evolution and the diversity of post-magmatic mineral growth settings and episodes, there are further complications such as hydrothermal alteration parageneses being formed in both hot upwelling fluid flow and cold down welling fluid flow at the same site.

Parageneses in the different textural settings correlate with different deformation, magmatic and hydrothermal episodes during evolution of Macquarie Island crust (Goscombe and Everard, 1999, 2001).

[1] Starting with crust formation at 8.9-9.4 Ma, based on: microfossil age correlations from inter-pillow sediments of 8.9-9.4 Ma age (Quilty et al., 2008) and minimum estimates at cooling through ~800 ºC, from U-Pb zircon age determinations of 8.6-9.0 Ma in massive gabbros (Schwartz et al., 2010; Dijkstra et al., 2010). Some of the older K-Ar and Ar-Ar determinations from basalts and dykes (Duncan and Varne, 1988) overlap with these crust formation constraints and may also be crystallization ages: 9.2, 9.6 and 9.7 Ma.

[2] High-grade post-crystallization evolution at depth at the inception of oceanic crust metamorphism. For example, late-stage gabbro veins with syn- to post- crystallization extensional mylonite fabrics in their cores are correlated with U-Pb zircon ages from associated late-stage phlogopite-feldspar-zircon pegmatite veins of 8.4-8.8 Ma age (Goscombe, 1999; Armstrong et al., 2004). Consequently, the initiation of oceanic crust metamorphism is bracketed between crust formation (8.9-9.4 Ma) and late-stage gabbro veins (8.4-8.8 Ma), which cut already serpentinized harzburgite.

[3] Oceanic crust metamorphism tracks through a continuum of further episodes of mineral growth during rapid near-isobaric cooling in the ridge axis to off-axis environment. Cooling of the upper crust (basalts and sheeted dykes) continued through ~600 ºC at approximately 6.5-7.2 Ma and ~500 ºC at 5.7-6.7 Ma based on Ar-Ar and K-Ar whole rock data (Duncan and Varne, 1988) and assuming very rapid cooling rates. Cooling of the lower crust (gabbros, troctolite and ultramafics) continued to be rapid, passing through ~310 ºC at 6.0-6.1 Ma and ~120 ºC at 4.7±0.5 Ma based on zircon and apatite FT data (Armstrong et al., 2004). Cooling rates are approximately: 275-550 ºC/Ma from crust formation to late gabbro vein shearing at 850 ºC, 132-208 ºC/Ma to 600 ºC, ~200 ºC/Ma to 500 ºC, ~292 ºC/Ma to 310 ºC, 69-136 ºC/Ma to 120 ºC and 27.5-30.0 ºC/Ma during exhumation to the surface. Oceanic crust metamorphic features were also over-printed by two further episodes.

[4] Late-stage isolated dolerite dykes accompanied a stress switch to transtension (D2), at some stage during transition from spreading ridge to transpressional plate margin between 5.0-8.4 Ma. These dykes have no discernable metamorphic expression within either the dykes or country rocks.

[5] Late-stage brittle transpressional deformation (D3) was due to dextral shearing with minor reverse thrusting of the Indo-Australian plate over the Pacific plate, starting from ~5.0 Ma on the basis of seafloor anomalies. D3 brittle gouge faults are generally dry and do not develop new mineral growth or veins and so have no obvious metamorphic expression.

[6] Macquarie Island crust was exposed between ~340,000 to ~700,000 years ago based on the oldest quartz thermoluminescence data from palaeo-beach deposits (Adamson et al., 1996). These indicate rapid exhumation of Macquarie Island occurred in the range 0.7-5.0 Ma.

Pressure estimation methods

Pressure (crustal depth) determinations for the different rock associations are important to constrain the thermal gradient at crust formation and at different episodes in the crustal evolution. No reliable and accurate method to calculate pressure is available for Macquarie Island rocks. Nevertheless, a number of conventional geobarometers and THERMOCALC were applied to magmatic assemblages to estimate near-crystallization pressures and retrograde assemblages formed during oceanic crust metamorphism. Calibrations including: total Al and Al6+ in hornblende (Johnson and Rutherford, 1989), Al6+ in hornblende (Spear, 1981), Si in white mica (Massone and Schreyer, 1987), plagioclase-clinopyroxene (Rollinson, 1981; Holland, 1980), plagioclase-hornblende (Plyusnina, 1982) and hornblende-chlorite (Laird, 1989). Surprisingly a significant number of these samples returned results consistent with the expected crustal level and consequently verify the qualitative method used to assign a depth-range to the different crustal rock associations (below).

A qualitative depth gauge, applicable specifically to Macquarie Island crust, was constructed on the basis of estimated thickness of the different rock associations. Average thickness of each of the different rock associations (and thus crustal depths) were derived by synthesis of both mapped thicknesses (Goscombe and Everard, 1999) and global average thicknesses, normalized to a global average oceanic crust thickness of 8 km (Brown and Mussett, 1981). Pressures were calculated from the derived crustal depths of the different rock associations, by assuming 1 kb = 3.5 km, based on average crustal rock density of 2.8 g/cm³. The resultant qualitative depth gauge calibration has the following median depths and pressures for each rock association, at the time of initial crust formation:

Rock association                              Depth: Z (km)                    Pressure: P (kb)

Basalt rock association                      0.850±0.425                      0.243±0.121

Sheeted dykes rock association          2.600±0.450                      0.743±0.129

Microgabbro rock association             3.625±0.125                      1.036±0.036

Massive gabbro rock association        5.125±1.375                      1.464±0.393

Layered gabbro rock association        6.750±0.250                      1.929±0.071

Troctolite rock association                 7.175±0.175                      2.050±0.050

Wehrlite rock association                  7.550±0.200                      2.157±0.057

Dunite rock association                     7.875±0.125                      2.250±0.036

Base of crust                                   8.000                                   2.286

Harzburgite rock association             10.00±2.000                      2.857±0.571

Late-stage gabbro veins and phlogopite pegmatite veins were intruded into the harzburgite association after a significant component of early exhumation, cooling and tilting of harzburgites blocks in the off-axis environment (Goscombe and Everard, 1999, 2001). Emplacement of these late-stage veins is assumed to have been at depths similar to emplacement of the massive and layered gabbro rock associations (~1.6 kb). Mantle harzburgites crystallized at depths greater than the average crustal thickness of 8 km, and an arbitrary crystallization depth of 10 km has been chosen.

The qualitative pressure gauge is broadly consistent with the available geobarometer calculations. Geobarometer results from harzburgite return an average pressure of 3.18 kb (n=3), wehrlite return an average pressure of 2.05 kb (n=4), troctolite returns a pressure of 2.07 kb, massive and layered gabbros return an average pressure of 2.39 kb (n=3), granulite mylonites in late-stage gabbro veins return an average pressure of 1.65 kb (n=15) and basalt returns a pressure of 0.28 kb.

Temperature estimation methods

Qualitative temperature estimates

Temperature estimates have been made by comparison between the documented petrology and the known T stability range of low-grade mineral species from both empirical, theoretical and experimental data in the literature. Though empirical zeolite data is typically referenced with respect to depth, zeolite stability is strongly T dependent despite the apparent depth (P) correlation. This is because the average thermal gradients in oceanic crust are so high (100-500 ºC/km) that a slight variation in depth equates to large temperature variation. The T stability range of the different minerals have been synthesised into a correlation diagram that is used as the basis for qualitative T estimates from Macquarie Island petrology and a composite petrogenetic grid appropriate for oceanic crust. Sources for low-T mineral stability are: (1) Theoretical petrogenetic grids in NCMASH (Liou et al., 1987) and CASH (Frey, 1987). (2) Experimental reactions in NCMASH (Liou et al., 1987), CASH (Perkins et al., 1980; Chatterjee, 1976) and CMASH (Powell et al., 1993). (3) Experimental and theoretical mineral reactions in a range of calcareous and basaltic systems (numerous sources).

The different textural parageneses considered in the analysis of Macquarie Island rocks, are: disseminated alteration minerals in the rock matrix, fracture and disjunctive cleavage infill, vein infill, vesicle infill, vug infill, shear bands, vein shears, fault mineralogy, vein-fault mineralogy and ductile mylonite fabrics. Most metamorphic mineral parageneses are considered isolated mineral phases that may have crystallized at different times and in different compositional sub-domains (i.e. on the margins of different primary minerals) and are not necessarily true equilibrium assemblages. Consequently, metamorphic parageneses are treated as partial assemblages that can be roughly assigned to a deformation-metamorphism period in the evolution of the crust.

Temperature estimates for different mineral parageneses are made on the basis of inclusive overlap of the experimental stability range of the different mineral phases to narrow the possible temperature range. Successive temperature ranges were assigned sequentially from high-T to low-T minerals to ensure that the overlap process determines the maximum temperature attained by the parageneses. Two exceptions are: (1) mylonite assemblages are treated as equilibrium assemblages and (2) zeolite mineralogy in vesicles and vugs are used to define the maximum possible temperature range, not the lower estimate based on inclusive over-laps. Where temperature ranges of different minerals do not overlap, the parageneses (such as vesicles, vugs or veins) are separated into a high-T and low-T episodes. The low-T episode may reflect breakdown of the earlier high-T parageneses.

Quantitative temperature estimates

In addition to petrology based qualitative T estimates, a range of quantitative T calculations has been applied to both magmatic and metamorphic assemblages. THERMOCALC v2.0b (Powell & Holland, 1988) was applied to igneous assemblages and granulite mylonites with generally reliable results, but large errors. Conventional geothermometers gave results consistent between different geothermometers and different calibrations, as well as the expected phase stability range and are thus considered generally reliable results. The geothermometers applied are: orthopyroxene-clinopyroxene (Finnerty and Boyd, 1987; Yamada and Takahashi, 1984; Brey and Kohler, 1990), olivine-clinopyroxene (Kawaski and Ito, 1994; Powell and Powell, 1974), olivine-spinel (Ballhaus et al., 1991), olivine-orthopyroxene (Podvin, 1988), hornblende-plagioclase (Plyusnina, 1982; Holland and Blundy, 1994), hornblende-orthopyroxene (Perchuk et al., 1985), biotite-clinopyroxene (Perchuk et al., 1985), Ca in orthopyroxene (Brey and Kohler, 1990), Al in orthopyroxene (Witt-Eickchen and Seck, 1991) and orthopyroxene-olivine-spinel (Witt-Eickchen and Seck, 1991). Magmatic crystallization temperatures are either sourced from modelled (dry solidus) phase stability at high T or pooled average crystallization temperatures calculated using geothermometers and THERMOCALC. Re-equilibrated temperatures are recognised by lower-T groupings of geothermometer and THERMOCALC result, indicating equilibration after cooling from the more typical crystallization conditions.

[TOP]