Abstract
This thesis reports the detailed study of 700km² of high grade metamorphic terrain in Doubtful Sound, Fiordland, and a more general synthesis of Fiordland geology as a whole.
Geophysical evidence from the Doubtful Sound area, i.e. gravity (Woodward 1972; and work by J. Coggan and the writer), seismic refraction (Davey and Broadbent, pers. comm.), earthquake data (Smith, 1971; Scholz et al., 1973) and geothermal activity suggests that the Doubtful Sound area is comprised of lower crustal rocks and that mantle rocks are less than 10km below the surface. The rocks have been brought to the surface along an active high angle thrust fault running along the line of the Fiordland west coast.
Detailed geological mapping has been carried out between Doubtful Sound and Breaksea Sound. Six mapping units were recognised which are referred to as the Malaspina, Turn Point and the Waipero Cove Gneisses constituting the basement, and separated from the cover sequence of Kellard Point and Deep Cove Gneisses by the Doubtful Sound Thrust. The Straight River Granite intrudes Deep Cove Gneisses in the Dagg Sound area. The basement units are sub-divided on metamorphic criteria rather than on conventional lithostratigraphic grounds.
The Malaspina Gneisses are feldspathic hornblende granulites with the assemblage antiperthitic plagioclase + olive green or brown hornblende + clinopyroxene + orthopyroxene + ilmenite and magnetite intergrowths + apatite + biotite. Anorthositic veins or thicker pegmatites studded with garnet porphyroblasts and large euhedral hornblende crystals cross-cut the feldspathic hornblende granulites often forming a network. The feldspathic gneiss in contact with these anorthositic veins has a reaction zone parallel to the margins of the veins. The mineral assemblage in the reaction zone is antiperthitic plagioclase + garnet + clinopyroxene + quartz + rutile + apatite. The garnet forms distinctive coronas around clinopyroxene. This assemblage is diagnostic of the high pressure granulite facies. Ultramafic gneisses interbanded with the feldspathic gneisses are interpreted as synmetamorphic intrusive sheets; they are predominantly hornblendites with bands and patches of hornblende eclogite. Some hornblendite sheets are layered with olivine-rich cumulate layers at their base; one such hornblende harzburgite contains hornblende lherzolite nodules. Malaspina Gneisses make up the Granulite Zone.
The Turn Point Gneisses are partially retrogressed equivalents of the Malaspina Gneisses, feldspathic gneisses have two assemblages: (1) plagioclase+ hornblende+ relict clinopyroxene + ilmenite and magnetite intergrowths + biotite + sphene + epidote + quartz + apatite; (2) plagioclase + garnet + relict clinopyroxene + quartz + biotite + hornblende + rutile + sphene +epidote + apatite. These two assemblages have formed by the partial hydration of feldspathic hornblende granulites and garnet granulites respectively. Ultramafic gneisses show similar reaction textures as the feldspathic gneiss:- sodic clinopyroxenes show symplectic breakdown to form less sodic clinopyroxene or hornblende; garnet has kelyphitic rims of hornblende and plagioclase. The Turn Point Gneisses make up the Transition Zone.
The Waipero Cove Gneisses are amphibolites; they have similar mineral assemblages to the Turn Point Gneisses but all clinopyroxene has been replaced by hornblende and biotite. Ultramafic gneisses do not have garnet or clinopyroxene. The Waipero Cove Gneisses make up the Amphibolite Zone.
The Malaspina Gneisses are distinguished by the presence of orthopyroxene in certain lithologies; the Turn Point Gneisses are distinguished where orthopyroxene has been replaced by amphibole and biotite and where relict clinopyroxene is still present. The Waipero Cove Gneisses do not contain pyroxene.
Malaspina, Turn Point and Waipero Cove Gneisses show relict igneous layering and xenoliths of microgabbro, diorite, and in one case a schist-like fragment; thus the basement feldspathic gneisses are interpreted as plutonic rocks.
The sequence Malaspina, Turn Point, and Waipero Cove Gneisses represents increasing hydration or amphibolitisation of the granulite facies basement under amphibolite facies conditions.
The cover sequence lies structurally above the basement: the Kellard Point Gneisses are made up of amphibolite facies metasediments and thick marbles with calc-silicate gneisses which have been intruded by pre-or syn-metamorphic acid to ultrabasic intrusives. Where marble is found in contact with the basement then it is always mylonitised; where marble is absent then an epidotised, actinolitised crush zone is found separating massive Waipero Cove Gneisses of the basement from banded quartzofeldspathic and micaceous magmatic metasediments of the Deep Cove Gneisses. This horizon of sheared gneisses and marbles has been mapped throughout the Doubtful Sound area and is interpreted as a decollement zone called the Doubtful Sound Thrust. The Deep Cove Gneisses are only distinguished from Kellard Point Gneisses by their lack of thick marbles although it is realised that this may be controlled by sedimentation or structure. By reconstructing sediment types it is suggested that the cover sequence represents a (off-shore) continental shelf environment.
The Straight River Granite is thought to have been intruded late in the amphibolite facies metamorphism of the cover.
The Doubtful Sound Area has had a complex structural evolution. The basement had undergone at least two deformations during M1 granulite facies metamorphism before the cover sequence was thrust westward over the basement along the Doubtful Sound Thrust. The cover sequence experienced M2 amphibolite facies metamorphism at the time of thrusting and was thrown into a series of recumbent folds while the basement below the Doubtful Sound Thrust was hydrated and amphibolitised. Other thrusts (the Mt. Troup Thrust in the east and the Straight River Thrust in the west) lie above and parallel to the Doubtful Sound Thrust.
Later mylonitisation occurred along deep fault zones in the western half of the area and the various thrusts were reactivated. Conditions allowed the M3 growth of chlorite. The thrusts and retrograds in the basement were warped into open synclines and anticlines during this time.
The M2 amphibolite facies metamorphism and migmatisation of the cover sequence is thought to have been dated by Aronson (1968) who estimated a zircon age of 320-350 m.yr. for a gneissic granite boulder from Deep Cove which this writer considers to be representative of the granitoid portion of nearby outcrops of gneissic metasediment. This date corresponds with the Devonian to Carboniferous Tuhua Orogeny. Since the basement was retrogressively metamorphosed from granulite to amphibolite at this time, the basement granulites must have originally crystallised earlier in the Tuhua or during a pre-Tuhuan orogeny. A Precambrian Malaspina Orogeny is invoked to account for the M1 granulite facies metamorphism. Thus the Fiordland granulites may be old Precambrian similar in age to the Indian, Australian, and Antarctican granulites of Gondwanaland.
The Rangitata Orogeny is probably represented in Doubtful Sound by uplift approximately 95m.y. ago (Aronson 1968, biotite Rb-Sr age from Deep Cove) and mylonitisation along the Mesozoic Alpine Fault. These movements were repeated during the Kaikoura Orogeny which commenced 4 ± 2m.y. ago (Sheppard et al., 1975)
Whole rock geochemical studies show that the feldspathic basement gneisses have the composition of rather sodic gabbroic diorites which have transitional characteristics between calc-alkalic and mildly tholeiitic magmas. Ultramafic gneisses have the composition of basanites and olivine tholeiites. It is not thought that the feldspathic and ultramafic gneisses were once both of the same magma series. The cover rocks analysed show a wide variation composition from granite through to peridotite and from sandstone through to limestone.
It is postulated that a hydrothermal fluid, evolved in the cover through metamorphic dehydration reactions during progressive amphibolite facies metamorphism, infiltrated some 2km into the basement and was absorbed during the transformation of granulite to amphibolite mineralogy. Infiltration metasomatism occurred and the following elements were relatively enriched in the basement: Si, K, C, H, U, Th, Hf, Cs, Nb, Y, Rb; the following elements were relatively depleted in the basement: Al, Fe, Mg, Ca, P: the following elements remained constant: Na, Ti, Mn, Pb, Cu, Ni, Zn, Sr, Zr, R.E.E. Depletion occurred in a zone 0.5km thick below the Doubtful Sound Thrust by back diffusion through the infiltrating fluid.
Metamorphosed igneous and sedimentary rocks have been distinguished using Misra diagrams, i.e. TiO₂ versus (FeO + Fe₂O₃/FeO + Fe₂O₃ + MgO) and TiO₂ versus MnO diagrams.
Quartz, K-feldspar, plagioclase, amphiboles, garnet pyroxene, olivine, biotite, chlorite, muscovite, scapolite, epidote, tourmaline, sphene, rutile, ilmenite, magnetite, limonite, spinel, serpentine, staurolite and carbonate have been analysed on the electron microprobe. Distribution of Fe, Mg, Ti, and Na between various mineral pairs has been studied. Using the Wood-Banno two pyroxene geothermometer combined with the Raheim-Green garnet-clinopyroxene geothermometer/ geobarometer it is concluded that the basement granulites equilibrated during metamorphism at about 775°C and 8kb load pressure (≡ ~30km depth); amphibolite facies gneisses of the cover and basement equilibrated (and the rims of basement granulite minerals re-equilibrated) at about 660°C assuming a pressure of 7kb (≡ ~23km depth).
K(D(gnt)Mg(-cpx))and K(D(gnt)Mg(-hbl)) values from other granulite and amphibolite facies areas compare well with the basement granulite and cover amphibolite in Doubtful Sound. K(D(gnt)Mg(-cpx)), K(D(gnt)Mg(-hbl)) K(D(gnt)Mg(-bio)) and K(D(gnt)Ti(-bio)) values of rims of mineral pairs from the basement compare well with K(D) values for cores and rims of the same mineral pairs from the cover, i.e. the basement was partially retrogressively recrystallised under the same conditions of metamorphism as the cover.
After consideration of field relationships and the major and trace element chemistry (including R.E.E.) it is suggested that hornblende granulite gneisses originated through partial melting of hydrous garnetiferous peridotite at depths greater than 70km; the magma rose through the upper mantle and fractionated mainly olivine and garnet. The differentiated magma intruded the lower crust and crystallised as a gabbroic diorite batholith. Regional metamorphism culminating at 775°C and 8kb formed granulite facies assemblages. Water pressure was less that total pressure but was sufficiently high to allow a small amount of partial melting. The anorthositic melt so formed segregated into a tensional joint system forming pegmatites. The composition of the gabbroic diorite (now feldspathic hornblende granulite gneiss) adjacent to the pegmatites was altered such that oxides, orthopyroxene, and hornblende simultaneously became unstable in the presence of plagioclase and reacted to form feldspathic garnet granulite gneisses. Model reactions are presented which explain the omphacitic nature of the clinopyroxenes found in feldspathic and ultrabasic gneisses.
During the granulite facies metamorphic episode, partial melting of hydrous mantle peridotite produced a water saturated basanitic liquid which rose, fractionating olivine (and clinopyroxene) such that when it intruded the lower crust and cooled, it crystallised largely as hornblendite. Under conditions of granulite facies metamorphism it is likely that some residual melts from the hornblendites crystallised as garnet rich hornblende eclogites, and that other hornblende eclogites were produced by subsolidus reactions between hornblende and plagioclase in the hornblendites to produce omphacite and garnet.
The status of the basement rocks with respect to the metamorphic facies concept is confused. Most workers would assign the feldspathic garnet granulites to a high pressure granulite subfacies and assign the hornblende granulites to a lower pressure or lower temperature granulite subfacies; the eclogite rocks could be assigned to the eclogite facies. Silica undersaturated xenoliths in the granulites have amphibolite facies mineralogy. However this work has shown that all these basement rocks equilibrated together under the same conditions of 775°C and 8kb and between 4 and 8kb H₂O pressure.
The geophysical, geological and geochemical evidence all support the hypothesis that the segment of granulite facies metamorphosed crust presently exposed in Doubtful Sound was formally part of the lower continental crust.
On the basis of the occurrence of certain minerals including kyanite, sillimanite, margarite, spinel, staurolite and the K(D) values of various mineral pairs it is proposed that the cover rocks have been metamorphosed at 675°C and 6.5kb (load pressure= H₂O pressure). Under these conditions it is likely that the quartzofeldspathic metasediments partially melted to give migmatite veining.
The regional geology of Fiordland is discussed with repsect to the basement and cover relationship discovered in Doubtful Sound. It is thought that the Doubtful Sound Thrust can probably be traced from Resolution Island to the Pembroke Valley: a distance of 160km. The mapping of formations in Fiordland by Wood (1960, 1962, 1966) is discussed and rejected and a new regional scheme is proposed. The geology of Fiordland is compared and contrasted with Nelson and Westland. The significance of the Alpine Fault in Fiordland is discussed.