Late Jurassic detrital zircons from the Haast Schist and their implications for New Zealand terrane assembly and metamorphism

Abstract

The youngest detrital zircon age groups from three samples within the Haast Schist in northwest Otago are Late Jurassic (154, 155, 160 Ma), as determined by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) analysis of U–Pb isotopes in individual zircons. It is inferred that this is the maximum age of sedimentation for these samples, which is within the range of ages for Haast Schist metamorphism (145–180 Ma). This maximum sedimentation age is at least 50 Ma younger than the previously inferred depositional ages for the Caples and Rakaia terranes which are the protoliths of Haast Schist. The zircon age populations within the samples are also different from those found within the Rakaia and Caples terranes, implying different sedimentary sources and possibly a different terrane. The detrital zircon populations are comparable to those found within the Waipapa Terrane in the North Island.

Keywords:

• Haast Schist
• aspiring lithologic association
• detrital zircon
• Caples Terrane
• Jurassic

Introduction

The term ‘terrane’ was introduced to New Zealand geology by Coombs et al. (1976) to refer to large Late Palaeozoic–Mesozoic tectonostratigraphic units with uncertain palaeogeographic relationships. In this 1976 paper, the ‘Torlesse’ Terrane (subsequently subdivided into Rakaia and Pahau terranes by Bishop et al. 1985) is clearly shown as allochthonous with respect to the rest of the terrane assembly, and uncertain displacements are suggested between several other terranes. Following the recognition of allochthonous terranes in western North America (Coney et al. 1980), the term ‘suspect terrane’ was defined as a tectonically bounded sequence of distinctive stratigraphy with uncertain palaeogeographic relationships to the continent. There have been many subsequent reviews and displacement models for New Zealand terranes e.g. Howell (1980); Bishop et al. (1985); Bradshaw (1989); Vaughan et al. (2005).

Two major tectonostratigraphic terranes, the Caples and Rakaia, are overprinted by and form the protolith of the Haast Schist, generally considered to have formed by convergence of the two terranes during the Middle–Late Mesozoic (Coombs et al. 1976; Wood 1978; Mortimer 1993; Little et al. 1999). Metamorphic grade increases towards the middle of the Haast Schist in Otago, from zeolite/prehnite-pumpellyite facies in the southwest (Caples Terrane) and northeast (Rakaia Terrane) to upper greenschist facies at the approximate contact of the two terranes (Fig. 1). Craw (1984) recognised a zone in the central schist in northwest Otago that has a distinctive lithological composition with abundant metabasites, manganiferous cherts and pelitic sediments. He termed this the Aspiring lithologic association (ALA) and considered it a more distal assemblage between the Caples Terrane and more proximal parts of what is now the Rakaia Terrane (Little et al. 1999; Nishimura et al. 2000; Gray & Foster 2004) (Fig. 1). Norris & Craw (1987) reviewed its character and structural relationships with the surrounding terranes, and concluded that it should be termed a suspect terrane in light of the probable allochthonous relationships of the terranes on either side. On the basis of geochemistry, Mortimer & Roser (1992) suggested that it was part of the Rakaia Terrane; on the basis of isotope geochemistry, Adams & Graham (1997) and Adams et al. (2009) however argued that its closest correlative was the Waipapa Terrane of the North Island. Since in northwest Otago the ALA lies between the inboard Caples Terrane, derived mostly from an oceanic island arc, and the outboard Rakaia Terrane, derived largely from the active margin of the Gondwana continent (Adams et al. 2009), its status and history are critical to understanding terrane accretion along the SW Pacific portion of the Gondwana margin.

Figure 1 Sample locations from the Rees and Matukituki Valley and the terrane boundaries in Otago. Dates are maximum time of sedimentation for samples from this note and from Adams et al. (2009). The metamorphic cooling ages are from Little et al. (1999). Approximate Alpine Schist overprint boundary from Mortimer (2000). Terrane boundaries are from Turnbull (2000). Shading shows the ALA and the Caples and Rakaia terranes.

Detrital zircons are unique in their ability to date and fingerprint the protoliths of the Haast Schist due to the destruction of the stratigraphy by metamorphism and the rarity of fossils within these terranes. Adams et al. (2007, 2009) found that the youngest detrital zircon population in a sample is close to the inferred depositional age in the Mesozoic terranes of New Zealand, due to their proximity to active margins.

In Otago, the Haast Schist is divided into the Otago Schist and Alpine Schist. The Otago Schist was metamorphosed in the Jurassic and exposed in the Cretaceous, whereas the Alpine Schist may have been metamorphosed as late as the Cretaceous and has been exposed in the last five million years from uplift along the Southern Alps (Mortimer 2000). The ALA has been overprinted by both the Otago and Alpine schist metamorphism (Fig. 1). In this note, we report detrital zircon age spectra from samples from the Haast Schist and compare them with published data from the Caples and Rakaia terranes and other terranes in New Zealand.

Method

Three samples of schist were collected from Muddy Creek in the Rees Valley (OU:42218, OU:81372, OU:81373) and one sample from the Matukituki Valley north of Aspiring Hut (OU:81365) (Fig. 1). Each sample was processed separately. Samples were crushed and sieved to a fraction between 50 and 250 µm. The laser ablation inductively coupled plasma mass spectrometer (LA-ICP-MS) at the Research School of Earth Sciences, Australian National University was used, following procedures described most recently by Sagar & Palin (2011). All zircons within the microscope field of view were hand-picked and mounted, whether they were definite or probable zircons. Crystals were set and polished in an epoxy mount and analyses were only taken from the core of each mounted crystal (Fig. 2). The standard zircon TEMORA2 and NIST610 glass were used for correction of depth-related inter-element and isotopic fractionation of Pb, Th and U and concentration determination, respectively.

Figure 2 Cathodoluminescence of four representative zircons with a Jurassic age from the Rees Valley samples OU:81372 and OU:81373 (University of Otago JEOL JXA-8600 Superprobe). The U–Pb dates and Th/U ratio were obtained from spot analysis in the core of the zircons.

Spot analyses with internal mean square weighted deviate (MSWD) values in excess of 10 were rejected from further consideration. Concordance was calculated on the basis of agreement between ²⁰⁷Pb/²³⁵U and ²⁰⁶Pb/²³⁸U dates within their reported 1σ internal errors. Analyses that were more than 10% discordant were not included in the age calculation.

Results

Detrital zircons have different age patterns between the Rees Valley and Matukituki Valley. The three Rees Valley (OU:42218, OU:81372, OU:81373) samples have principal age peaks of 154, 155 and 160 Ma and smaller peaks around 250 Ma (Table 1, Fig. 3). The ratio of c. 150 Ma and c. 250 Ma ages varies between samples. One of the Rees Valley samples (OU:81372) has no zircon ages older than 300 Ma (Fig. 3). The Matukituki Valley sample (OU:81365) has prominent age peaks at c. 250 and c. 350 Ma and two zircons from the Precambrian. Zircons less than 200 Ma are completely absent from this sample.

Figure 3 Histograms and probability density diagrams of detrital zircon ages from the Haast Schist. The size of a zircon age population which has a 95% chance of being identified is given as X_L from Dodson et al. (1988) and Andersen (2005). Dates in the top right of each box are older than the range of the graphs. The ranges used for peak ages are in Table 1.

Table 1 Detrital zircon samples from the Rees and Matukituki valleys. Age of peaks include all analyses within the specified age range with less than 10% discordances. For example, the age given for OU:42218 includes all viable analyses with a date younger than 180 Ma. New Zealand Transverse Mercator grid references.

Group	Location	Grid ref.	Peak age	MSWD	N#	Group (Ma)
OU:42218	Rees Valley	E1241300 N5040900	154.1 ± 2.0	2.7	17/31	<180
OU:81372	Rees Valley	E1241850 N5041180	154.5 ± 2.9	6.2	46/67	<180
OU:81373	Rees Valley	E1241700 N5041000	159.7 ± 5.6	4.9	11/37	<180
OU:81373			251.6 ± 7.9	5.7	13/37	220–300
OU:81365	Matukituki V.	E1254600 N5068900	242.5 ± 4.2	3.2	42/65	<300
OU:81365			358.8 ± 6.1	1.4	18/65	300–450

Discussion

Do Jurassic zircon ages represent igneous or metamorphic events?

The Jurassic zircons from the Rees Valley samples are of similar age to the metamorphism of the Haast Schist in which they are found (Little et al. 1999); it is therefore important to know if these zircons are igneous or metamorphic in origin. Zircons from granulite facies rocks can develop metamorphic rims (Rubatto et al. 2001), but the Haast Schist in the locations sampled attain only upper greenschist facies (Mortimer 2000) and no metamorphic rims were observed. Discordance does not vary between the Jurassic and Triassic dates within the Rees Valley samples so Pb loss, while possibly present, does not explain the Jurassic dates. These zircons show typical igneous structure when viewed with cathodoluminescence (Corfu et al. 2003) (Fig. 2) and have typical igneous Th/U ranges of greater than 0.1 in all but two analyses (Ahrens et al. 1967; Hoskin & Black 2000; Rubatto et al. 2001). For the above reasons it is inferred that they eroded from an igneous source and were transported with other sediment. The maximum age of deposition of the Rees Valley samples is therefore Late Jurassic. This is c. 90 Ma younger than the Triassic maximum age of deposition of the Matukituki Valley sample.

Comparison of Rees Valley detrital zircons with other terranes

The Rees Valley sample location was identified as the Caples Terrane from whole-rock geochemistry performed on the same meta-psammite lens by Mortimer & Roser (1992). However, the samples from the Rees Valley reported here have different detrital zircon age distributions to the Caples and Rakaia terranes (Adams et al. 2007, 2009). The most important difference is the group of zircons of Jurassic age that limit the depositional age of Rees Valley samples to younger than c. 154 Ma. In comparison, the youngest zircon populations from the Caples and Rakaia terranes are 210–220 Ma (Adams et al. 2007, 2009; Jugum 2009). The 500–600 Ma detrital zircon peak which is usually observed in the Rakaia Terrane is also missing in the samples from the Rees and Matukituki valleys, as noted by Adams et al. (2009) for their ALA sample. The difference in detrital zircon ages implies that these samples have a different sedimentary source to the known Caples and Rakaia terranes.

The ALA has been correlated with the Waipapa Terrane by Adams & Graham (1997) and Adams et al. (2009) as they share the same tectonostratigraphic position across the Alpine Fault and a similar Sr isotopic signature. Geochemically, the Waipapa Terrane overlaps with both the Caples and Rakaia terranes with respect to Si, K, Na and other major-element discriminants (Coombs et al. 2000). Lithologically, the Rees Valley samples are from an area where pelites are common and have been mapped as part of the ALA previously (Craw 1984). The Waipapa Terrane also shares the abundance of chert and basic volcanic rock associated with the ALA (Black 1994; Adams & Maas 2004).

The complete detrital zircon age pattern from the Jurassic Rees Valley sample (OU:42218) is also very similar to the representative Jurassic Waipapa Terrane sample in Adams et al. (2007) from north Auckland (Fig. 4). Zircons older than Jurassic age in the Rees Valley samples have a similar age distribution to those in the Matukituki Valley sample (Fig. 3). The Skippers River sample (ALA) and Waipapa and Caples terranes samples of Adams et al. (2009) all show these peaks at 250 Ma and a lack of Cambrian zircons, common in Rakaia Terrane samples.

Figure 4 Comparison of detrital zircon ages from sample OU:42218 from the Rees Valley and the Adams et al. (2007) representative Jurassic sample from the Waipapa Terrane in Auckland.

The protolith lithology, geochemistry and detrital zircon age distribution of the eastern Rees Valley schist overlap with the Waipapa Terrane. It is therefore suggested here that the Waipapa Terrane does not stop at the Alpine Fault in Nelson, but extends into Otago. It remains possible, however, that the Rees Valley samples are from a Jurassic component of the Caples Terrane or an as yet unknown terrane within the Haast Schist. These possibilities remain open as detrital zircons have only provided a maximum age on the known Caples Terrane (Adams et al. 2009) and terrane boundary faults are obscured by the Haast Schist.

Consequences of detrital Jurassic zircons in the Haast Schist

The estimated age of peak metamorphism of the Haast Schist in Otago varies within the range 145–160 Ma (Gray & Foster 2004) and 170–180 Ma (Little et al. 1999). The maximum sedimentation age for the Rees Valley samples presented here precludes it being overprinted by the Jurassic metamorphism or limits the age of that metamorphism to after c. 154 Ma. The cooling ages given by Little et al. (1999) for samples near the Rees Valley samples are 140 and 130 Ma (Fig. 1). This allows c. 14 Ma for the zircon crystallisation, erosion, deposition, metamorphism and cooling of the Rees Valley samples. This continues the trend of the youngest detrital zircon ages being close to sedimentation ages for Mesozoic New Zealand terranes (Akdams et al. 2007, 2009).

If the ALA is the extension of the Waipapa Terrane southeast of the Alpine Fault (Adams & Graham 1997), then it can be used to shed further light on the terrane relationships within the Haast Schist. In Northland, the Waipapa Terrane is a continuous belt east of the Caples Terrane and generally does not exhibit a penetrative foliation characteristic of the Haast Schist in the South Island (Black 1994). In Otago, however, the ALA is structurally beneath the Caples Terrane and within the structural centre of the exposed Haast Schist (Norris & Craw 1987; Little et al. 1999; Mortimer 2000, 2003; Fig. 5). The new zircon data reinforce the distinctive nature of the ALA within the Haast Schist and point to a correlation with part of the Waipapa Terrane on the northwest side of the Alpine Fault. The ages also indicate a final accretion of these terranes sometime after the Late Jurassic, placing constraints on the age of the Haast Schist metamorphic overprint. The presence of older metamorphic ages in parts of the Rakaia Terrane (Little et al. 1999; Adam et al. 2009) suggests that some metamorphism occurred during assembly of the terrane before its final accretion along the Gondwana margin. Alternatively, during the accretion that led to the metamorphic and structural overprint of the Haast Schist, the ALA may have been underplated beneath the already partially metamorphosed Rakaia and Caples terranes. The underplating of the ALA may have been part of the exhumation process that led to the cooling of the overlying Caples and Rakaia terranes in the Early Cretaceous.

Figure 5 Schematic cross-section of the tectonostratigraphic terranes of New Zealand's South Island (Little et al. 1999; Adams et al. 2002; Mortimer et al. 2002). Above the cross-section is the depositional age range of the terranes. The older terranes are Early Palaeozoic and intruded by plutons ranging in age from Early Palaeozoic to Early Cretaceous.

Supplementary files

Supplementary file 1: Tables S1. Tables 1–12 contain U–Th–Pb isotope data, Tables 1–2 contain zircon analyses for OU:42218, Tables 2–6 for OU:81365, Tables 7–10 for OU:81372 and Tables 11–12 for OU:81373.

Acknowledgements

Dave Craw contributed useful suggestions on the Haast Schist and a sample from his PhD was analysed for this paper. We are grateful to Nick Mortimer for conversations on the terranes in Otago and comments on how to improve this manuscript. James Scott provided assistance with zircon separation and analysis. We would also like to thank the many members of staff in the Department of Geology of the University of Otago and GNS Science, Dunedin. The manuscript was also improved by suggestions from two anonymous reviewers.