Petrogenesis of amphibole megacrysts in lamprophyric intraplate magmatism in southern New Zealand

ABSTRACT

Sr-Nd-Hf-Pb isotopes and major/trace element data are presented for amphibole megacrysts from two separate instances of lamprophyric intraplate magmatism in Zealandia (Cretaceous Westland Dike Swarm and Cenozoic Alpine Dike Swarm). Megacrysts from both regions have comparable geochemical and isotopic signatures (⁸⁷Sr/⁸⁶Sr_(i) = 0.702754–0.703371, εNd_(i) = +3.9 to +5.5, εHf_(i) = +5.0 to +8.2, ²⁰⁶Pb/²⁰⁴Pb_(i) = 18.909–20.526, ²⁰⁷Pb/²⁰⁴Pb_(i) = 15.642–15.678, ²⁰⁸Pb/²⁰⁴Pb_(i) = 38.824–40.271), regardless of them having traversed contrasting crustal terranes at different times and tectonic settings. Several megacryst cores may represent early-crystallised phenocrysts, whereas rims crystallised in equilibrium with amphibole microphenocrysts during final emplacement of host melts. Strontium isotopic offsets between megacrysts and host rocks indicate secondary hydrothermal alteration in some lamprophyres. Correlations between host rock – megacryst geochemistry as well as evidence for amphibole fractionation in Zealandia lamprophyres indicate that megacrysts likely crystallised from melts similar to their hosts. Isotopic (Pb, Nd) and Mg# offsets suggest a non-cognate relationship and reflect interaction between isotopically heterogeneous lamprophyric melts and/or cumulate amphibole in the lower crust. Modelling suggests that most alkaline intraplate magmatism throughout Zealandia may be derived from moderately heterogeneous SCLM sources that experienced localised varying degrees of subduction-related carbonatitic metasomatism between the Jurassic to Middle Cretaceous during Gondwana break-up.

KEYWORDS:

• Amphibole
• kaersutite
• lamprophyre
• Zealandia
• megacryst
• SCLM
• Gondwana
• intraplate

Introduction

Lamprophyres are a group of igneous rocks that are volumetrically minor and inherently restricted to continental settings. These rocks were largely considered as enigmatic curiosities up until the works summarised in Rock (1991), which defined lamprophyres as volatile-rich mantle-derived alkaline igneous rocks carrying hydrous minerals such as amphibole and biotite within a feldspar-rich groundmass. Since then, lamprophyres have been used extensively to constrain the nature and origin of the sub-continental lithospheric mantle (SCLM) through their geochemistry (e.g. Foley et al. 2002). Moreover, lamprophyre xenolith and xenocryst assemblages provide direct samples of the SCLM (e.g. Griffin et al. 2004; Pandey et al. 2018).

An ideal location for the study of lamprophyres and their crystal and xenolith assemblages is New Zealand. Since the Late Cretaceous and separation of Zealandia from Australia and Antarctica, repeated intraplate volcanism has erupted sporadically and indiscriminately within different crustal blocks of Zealandia (e.g. Timm et al. 2010; van der Meer et al. 2016). Such prolonged and widespread intraplate alkaline volcanism in Zealandia has provided an invaluable tool for probing the nature of the SCLM and its evolution during the past ~ 100 Ma. Several studies have argued for the tapping of a HIMU-like source component based on Zealandia intraplate alkaline volcanism displaying geochemical compositions and Sr, Nd, Hf ± Pb isotopic signatures similar to ocean island basalts (e.g. Hoernle et al. 2006; Sprung et al. 2007; van der Meer et al. 2017; Scott et al. 2020). However, the nature of this HIMU-like component and the mechanisms behind its repeated tapping remain debated. Nevertheless, the emplacement of comparable intraplate lamprophyres at different times in Zealandia as well as in different crustal blocks may have important implications for the apparent offset in crustal and mantle histories identified through studies of mantle xenoliths in Zealandia.

The presence of megacrystic amphibole is a characteristic feature of some of Zealandia intraplate lamprophyres. The petrogenetic nature of amphibole megacrysts in alkaline intraplate lamprophyres remains a debated subject. Intraplate mafic alkaline volcanism carrying amphibole megacrysts and amphibole-bearing mantle xenoliths has been reported worldwide (e.g. Kesson and Price 1972; Boettcher and O’Neil 1980; Irving and Frey 1984; Dautria et al. 1987; Shaw and Eyzaguirre 2000; Dobosi et al. 2003; Coltorti et al. 2007; Han et al. 2008; Fulmer et al. 2010; van der Meer et al. 2013). Amphibole megacrysts almost exclusively belong to the magnesio-hastingsite-pargasite-kaersutite series and have been interpreted as either: (1) high-pressure cognate phenocrysts (e.g. Binns 1969; Merrill and Wyllie 1975; Irving and Frey 1984; Mayer et al. 2014; Ulrych et al. 2018); (2) xenocrystic fragments of hydrous amphibole-rich pegmatitic veins in peridotites from the SCLM (e.g. Wilshire and Trask 1971; Best 1974; Boettcher and O’Neil 1980; Shaw and Eyzaguirre 2000; Dobosi et al. 2003); or (3) antecrysts that evolved in open system magma-plumbing systems (e.g. Ubide et al. 2014).

This study focuses on megacrystic amphiboles entrained in spatially and temporally decoupled pulses of intraplate alkaline magmatism in southern New Zealand. The aims of this paper are: (1) to constrain variations in major element, trace element and Sr-Nd-Hf-Pb isotope characteristics of megacrystic amphiboles occurring in different intrusives of South Island, New Zealand; (2) to determine the relationship between megacrystic amphiboles and their host rocks; and (3) thereby attempt to constrain the petrogenesis of these megacrystic amphiboles with concomitant implications for the origin of their host magmas.

Geological setting

The islands of New Zealand represent parts of the largely submerged (∼94%) Zealandia continent (>3.5 × 10⁶ km²) in the southwestern Pacific (Figure 1A) (Mortimer et al. 2017). This continent was situated adjacent to the Australian and Antarctica continents prior to the breakup of the Gondwana supercontinent in the Late Cretaceous. Zealandia formed by the accretion of batholith intrusions and fore-arc sedimentary and volcanic terranes onto the eastern convergent margin of Gondwana during the Phanerozoic and Mesozoic (ca. 520–100 Ma) (Mortimer 2004). This enduring accretionary subduction-dominated regime ended around 105–100 Ma and was substituted by crustal extension, which continued until the separation of Zealandia from West Antarctica and Australia and the onset of sea-floor spreading at ca. 85 Ma (Gaina et al. 1998; Waight et al. 1998a, 1998b; Laird and Bradshaw 2004; van der Meer et al. 2016). Sea-floor spreading resulted in the opening of the Tasman Sea and Zealandia consequently drifted northwest to its present-day location (Gaina et al. 1998). Zealandia is divided in two major provinces based on the age of basement rocks. Namely, the older early Palaeozoic (Cambrian to Early Devonian) Western Province (WP) and the younger Permian to mid-Cretaceous Eastern Province (EP) (Mortimer 2004) (Figure 1). The Median Batholith, the plutonic root of a long-lived Devonian to Early Cretaceous cordilleran arc (Mortimer 2004; Scott 2013), separates and stitches both provinces (Figure 1B).

Figure 1. A, Map of present-day Zealandia showing occurrences of intraplate magmatism. B, Map of southern New Zealand with sample locations modified from Cooper et al. (1987). Ages for WDS samples are from (van der Meer et al. 2013), whereas ages for ADS samples are determined based on several studies (e.g. Cooper et al. 1987; Adams and Cooper 1996; Hoernle et al. 2006; Timm et al. 2010; Cooper 2020).

Since the cessation of the subduction-dominated regime in the Middle Cretaceous, intraplate volcanism has been widespread on the Zealandia continent (van Der Meer et al. 2018). Mafic intraplate volcanic rocks in Zealandia predominantly have alkaline compositions and Sr-Nd-Hf-Pb signatures approaching HIMU-like compositions, with ⁸⁷Sr/⁸⁶Sr_(i) ∼ 0.703–704, εNd_(i) = +2 to +8, εHf_(i) = +2 to +10, ²⁰⁶Pb/²⁰⁴Pb_(i) = 18.5–20.7 and ²⁰⁸Pb/²⁰⁴Pb_(i) = 38.5–41 (Price 2003; Hoernle et al. 2006; Panter 2006; Sprung et al. 2007; Timm et al. 2009, 2010; McCoy-West et al. 2010; van der Meer et al. 2017; Hoernle et al. 2020; Scanlan et al. 2020; Scott et al. 2020). The sources behind such signatures in Zealandia intraplate magmas are still under debate. Some authors favour melting of a metasomatically enriched SCLM source (Panter 2006; McCoy-West et al. 2010; Scott et al. 2016a; van der Meer et al. 2017) while others favour melting of an asthenospheric source (Hoernle et al. 2006; Timm et al. 2010; Hoernle et al. 2020) or a combination of both (Finn et al. 2005; Sprung et al. 2007). The trigger of Cretaceous intraplate magmatism has generally been attributed to decompression of lithosphere thinned during extension associated with the opening of the Tasman Sea (Waight et al. 1998a, 1998b; van der Meer et al. 2016, 2017). Initiation of Cenozoic intraplate magmatism has been attributed to either the destabilisation of a hydrous lithospheric mantle (Panter 2006; Sprung et al. 2007; van der Meer et al. 2017) or decompression of the asthenosphere following delamination of the SCLM (e.g. Hoernle et al. 2006; Timm et al. 2010).

This paper focuses on intraplate lamprophyric dikes carrying megacrystic amphiboles in Zealandia. Such dikes are both spatially and temporally widespread, occurring in both the Western and Eastern Provinces of the South Island, and intruding in the Late Cretaceous and in the Cenozoic, respectively (e.g. Cooper et al. 1987; Waight et al. 1998a; van der Meer et al. 2013; Cooper 2020). Lamprophyric to phonolitic intraplate magmatism in the WP has been reported in the alkaline Westland Dike Swarm (WDS) centred on the Paparoa and Hohonu Ranges (Figure 1B). These dikes vary in width from 1 cm to 15 m and display sharp intrusive contacts with their host Cretaceous granitoids (e.g. Waight et al. 1998a). Intrusion ages in the WDS have been constrained by ⁴⁰Ar/³⁹Ar ages from groundmass and megacrystic kaersutite in lamprophyric dikes to a first event at 90–84 Ma, with a second pulse around 72–68 Ma (van der Meer et al. 2013, 2016). The generation and emplacement of these dikes has been interpreted to be related to the extensional regime dominating during the opening of the Tasman Sea (Waight et al. 1998a; van der Meer et al. 2016). Similar lamprophyric dikes hosting megacrystic amphibole have been documented in the Alpine Dike Swarm (ADS) located in the EP (Figure 1B). The ADS differs slightly from the WDS as it also includes carbonatite dikes (Cooper and Paterson 2008; Cooper 2020). Lamprophyric dikes of the ADS vary in thickness from centimetres to several tens of metres and show sharp intrusive contacts with the Haast Schist host rock (e.g. Cooper 2020). The ADS is significantly younger than the WDS and was emplaced around 25–23 Ma (Cooper et al. 1987; Adams and Cooper 1996; Hoernle et al. 2006; Timm et al. 2010; Cooper 2020) in a thickened lithosphere, and intruded along shear zones associated with the Alpine Fault plate boundary (Cooper 2020). ADS magmatism was contemporaneous with widespread Cenozoic intraplate activity in the South Island (Price 2003; Hoernle et al. 2006; Sprung et al. 2007; Timm et al. 2010; Scanlan et al. 2020; Scott et al. 2020).

The Zealandia SCLM has been investigated through studies of mantle xenoliths entrained by alkaline magmas (e.g. McCoy-West et al. 2013, 2015, 2016; Scott et al. 2014a, 2014b; Scott et al. 2016a, 2016b; Dalton et al. 2017; Scott et al. 2019; Scott 2020) as well as the exhumed Anita Peridotite (e.g. Czertowicz et al. 2016). While the Anita Peridotite has been interpreted to represent Cretaceous sub-arc lithospheric mantle, spinel-facies peridotite xenoliths revealed signs of widespread but variable mantle metasomatism in the Zealandia SCLM. Geochemical evidence from clinopyroxene suggests that peridotites reacted with carbonatite/CO₂-rich silicate melts, which imprinted HIMU-like Sr, Nd and Pb isotopic signatures onto the SCLM (Scott et al. 2014a, 2014b; McCoy-West et al. 2016; Scott et al. 2016b; Dalton et al. 2017). Present-day Moho depth has been constrained to vary between 35 and 48 km (Salmon et al. 2013) in the ADS and has been assumed to be comparable during the Cenozoic (e.g. ∼ 37 km in Scott et al. 2014b). Late Cretaceous Moho depth in the WDS has not been constrained, but may be comparable to ∼ 30 km as in averaged extensional settings (Christensen and Mooney 1995).

Materials and methods

Methods

Major element compositions of amphibole were determined using a JEOL JXA-8200 Superprobe at the Department of Geosciences and Natural Resource Management, University of Copenhagen. Trace element compositions of amphibole were determined by LA-ICP-MS at GEUS (Geological Survey of Denmark and Greenland). Isotopic compositions of amphibole megacrysts were determined at the University of Copenhagen (Sr, Nd), University of Cape Town (Pb) and University of Cologne (Hf). A detailed description of methodologies and standards as well as data that support the findings of this study are openly available in Figshare at 10.6084/m9.figshare.12057843.

Studied samples

Westland Dike Swarm (WDS)

Four samples originating from a lamprophyre ± minor phonolite dike suite in the WDS were chosen due to the occurrence of megacrystic amphiboles. Samples QH12, KST-2, MOA-16 and HOH-1 were collected in the area around Lake Brunner in the Hohonu Dike Swarm (Figure 1B). Hohonu Dike Swarm host rocks were emplaced in the Late Cretaceous (∼ 88 Ma) (van der Meer et al. 2013). The KST-2 sample was collected from a composite dike composed of both lamprophyre and phonolite (Waight et al. 1998a). Sample QH12 originates from a lamprophyric loose boulder collected near the KST-2 dike. Sample OU79943 was collected in the Bonar Range (Scott et al. 2011), southwest of the Hohonu Dike Swarm and is a lamprophyric dike dated to ∼ 69 Ma (van der Meer et al. 2013) (Figure 1B). WDS samples are characterised as camptonitic and carry amphibole megacrysts, as well as spinel-facies peridotite xenoliths (Tulloch and Nathan 1990; van der Meer et al. 2013; Scott et al. 2016b) (Figure 2).

Figure 2. A, Field photo of light-grey MOA-16 lamprophyre sample from the WDS, first presented in van der Meer et al. (2013). Abundant visible mantle xenoliths (light brown) and amphibole megacrysts (black). B, Field photo from the Fish River locality in the ADS with a 15 cm long chisel for scale. The photo shows a grey-green ultramafic lamprophyre with abundant amphibole megacrysts and mantle xenoliths.

Alpine Dike Swarm (ADS)

Amphibole megacrysts were obtained from three samples belonging to a lamprophyre + phonolite + carbonatite dike suite in the ADS. Samples FIS-4, MAC-1 and LWA-12 originate from lamprophyres from Fish River, MacFarlane River and Lake Wanaka, respectively (Figure 1B). They were chosen because they carry both megacrystic amphiboles and spinel-facies peridotite xenoliths (Scott et al. 2014b; Scott et al. 2016a) (Figure 2B).

Results

General petrography

Investigated WDS camptonites comprise a groundmass consisting of sub-millimetre sized subhedral to euhedral brown amphibole microphenocrysts together with abundant feldspars, minor opaque phases, apatite and secondary epidote ± chlorite (Figure 3). Larger phases consist of orange to brown pleochroic megacrystic amphiboles, as well as occasional <5 mm sized anhedral clinopyroxene and up to 2 mm sized epidote ± carbonate ocelli (Figure 3). Investigated samples contain variable amounts of centimetre-sized amphibole megacrysts, which are represented by heavily fractured subhedral to anhedral grains with irregular edges (Figure 3). When present, rims are <0.5 mm thin and show similar pleochroic colours to groundmass microphenocrysts. Most fractures in megacrystic amphiboles contain groundmass material (Figure 3). Additionally, amphibole megacrysts show indications of chemical re-equilibration along fractures where pleochroic colours match those seen in rims and microphenocrysts (Figure 3). Smaller 1–5 mm-sized anhedral amphibole grains have similar characteristics to the associated megacrysts and are therefore interpreted to represent fragmented megacrysts. Additionally, these amphibole fragments have rims grown uniformly around anhedral cores with no systematic differences in dark/light brown pleochroism between cores and rims.

Figure 3. Representative thin section of WDS samples (specifically from sample KST-2). The photo shows two amphibole megacrysts (Amph), as well as epidote ± carbonate ocelli (Oc) and rounded clinopyroxene grains (Cpx) in an amphibole + plagioclase + epidote + chlorite + opaque phase groundmass. Anhedral >1 mm sized zoned amphibole grains are interpreted to represent megacryst fragments.

The petrography of the ADS camptonites has been provided in detail by Cooper (2020), who describes them as consisting of a feldspar-rich + clinopyroxene + kaersutite + biotite groundmass hosting euhedral phenocrysts of olivine, kaersutite, clinopyroxene and biotite. Kaersutite phenocrysts show both normal and reverse zoning patterns with Mg# (Mg# = 100·Mg/(Mg + Fe)) ranging from 40 to 67 (Cooper 2020).

Major elements

Amphibole megacrysts and groundmass microphenocryst amphiboles in all samples are classified as predominantly ferri-kaersutite and minor magnesio-hastingsite according to the classification scheme of Hawthorne et al. (2012) using calculations from Locock (2014) (Figure 4). A compilation of major element data of 177 amphibole grains is provided at 10.6084/m9.figshare.12057843. Investigated amphiboles show compositions rich in Al₂O₃ (11–15 wt.%), CaO (10–13 wt.%), TiO₂ (3–7 wt.%) together with particularly variable FeO^tot (9–18 wt.%) and MgO (8–14 wt.%) and minor K₂O (1–2 wt.%) and Na₂O₃ (2–4 wt%) concentrations. High variability in FeO^tot and MgO contents is reflected by amphibole megacryst core Mg# compositions (Figure 4). While amphibole megacrysts in some samples show high Mg# with little to no internal variability [i.e. WDS samples HOH-1 (68–69) and MOA-16 (70); ADS samples FIS-4 (69–70), MAC-1 (71–72) and LWA-12 (68)], amphibole megacryst cores in WDS sample OU79943 show moderately variable Mg# (62–69), and WDS samples QH12 and KST-2 exhibit highly variable amphibole megacryst core Mg# compositions (48–66 and 47–66, respectively) (Figure 4). Investigated amphibole megacryst rim compositions in WDS samples are identical to or approaching host rock microphenocryst compositions (with Mg# = 61–72). Amphibole megacryst rim compositions show higher Mg# compared to core compositions in WDS samples QH12 and KST-2, whereas the opposite is observed in sample MOA-16.

Figure 4. Mg# vs. Ti (apfu) diagram showing studied amphibole megacryst cores, rims and microphenocrysts following the classification after Hawthorne et al. (2012). Ti (apfu) determined using calculations from Locock (2014). Mega-C = megacryst core analysis; Mega-R = megacryst rim analysis; Micro = microphenocryst analysis. Black symbols represent WDS samples and red symbols represent ADS samples.

Trace elements

Trace element data for 83 amphibole grains are available at 10.6084/m9.figshare.12057843. Primitive mantle normalised incompatible trace element patterns of amphibole megacrysts show positive anomalies in Ba, Nb, Ta and Ti, as well as distinct negative anomalies in Th, U, Pb, Zr and Hf (Figure 5A). Megacrystic amphiboles have similar convex-upward chondrite normalised REE patterns with a maximum at Nd [Nd_(n)= 38–111 (avg. = 69)], enriched LREE [Nd_(n)/Lu_(n)= 5–22 (avg. = 9)], depleted HREE [Gd_(n)/Lu_(n)=3–10 (avg. = 5)] and no Eu anomaly (Figure 5B). Host rock microphenocrysts are over twice as enriched in all REE [(^totREE_(n) =600–1050 (avg. = 700))] relative to megacrysts [(^totREE_(n) =180–480 (avg. = 310))] (Figure 5B), while megacryst rims show intermediate compositions approaching microphenocryst compositions (Figure 5B). It is evident that ADS and WDS megacrysts share comparable trace element signatures (Figure 5C). However, WDS megacryst cores are more enriched in MREE and particularly HREE (avg. Nd_n/Lu_n = 7) compared to ADS megacrysts, resulting in steeper LREE-HREE slopes in ADS samples (avg. Nd_n/Lu_n = 16) (Figure 5D).

Figure 5. A, Trace element analyses of amphibole cores, rims and microphenocrysts normalised to primitive mantle (McDonough and Sun 1995). B, REE analyses of amphibole cores, rims and microphenocrysts normalised to chondrite (McDonough and Sun 1995). The solid line in each field shows the averaged composition. C, Trace element analyses of WDS and ADS megacrystic amphibole cores compared to WDS and ADS lamprophyre compositions normalised to primitive mantle (McDonough and Sun 1995). D, REE analyses of WDS and ADS megacrystic amphibole cores compared to WDS and ADS lamprophyre compositions normalised to primitive mantle (McDonough and Sun 1995). Lamprophyre trace element and REE data are from Timm et al. (2010), van der Meer et al. (2016, 2017) and Cooper (2020).

Radiogenic isotopes

To compare source characteristics of amphibole megacrysts occurring in two separate geological domains in southern New Zealand, amphibole megacrysts were separated and measured for Sr-Nd-Hf-Pb isotopes (Table 1). Figure 6 shows age-corrected Sr-Nd-Hf-Pb isotopic data of amphibole megacrysts related to published data from relevant WP and EP lithologies as well as BSE, HIMU and MORB compositions. The effect of age-corrections is relatively minor compared to the variations observed in the data set. Amphibole megacrysts from the WDS show limited variation in Sr-Nd-Hf isotopic signatures with ⁸⁷Sr/⁸⁶Sr_(88-69Ma)= 0.702754–0.703371, ¹⁴³Nd/¹⁴⁴Nd_(88-69Ma)= 0.512748–0.512807 (εNd_(88-69Ma)= +3.9 to +5.5) and ¹⁷⁶Hf/¹⁷⁷Hf_(88-69Ma)= 0.282869–0.282950 (εHf_(88-69Ma)= +5.0 to +8.2). These values overlap with more constrained ADS amphibole megacryst signatures, which have ⁸⁷Sr/⁸⁶Sr_(25-23Ma)= 0.702842–0.702991, ¹⁴³Nd/¹⁴⁴Nd_(25-23Ma)= 0.512808–0.512828 (εNd_(25-23Ma)= +3.9 to +4.3) and ¹⁷⁶Hf/¹⁷⁷Hf_(25-23Ma)= 0.282899–0.282953 (εHf_(25-23Ma)= +5.0 to +6.9) (Figure 6A,B). With regards to Pb isotopes, WDS amphibole megacrysts show less radiogenic signatures with ²⁰⁶Pb/²⁰⁴Pb_(88-69Ma) = 18.909–19.400, ²⁰⁷Pb/²⁰⁴Pb_(88-69Ma)= 15.642–15.657 and ²⁰⁸Pb/²⁰⁴Pb_(88-69Ma) = 38.824–39.290 compared to ADS amphibole megacrysts, which also show a larger spread in Pb signatures with ²⁰⁶Pb/²⁰⁴Pb_(25-23Ma) = 19.494–20.526, ²⁰⁷Pb/²⁰⁴Pb_(25-23Ma) = 15.661–15.678 and ²⁰⁸Pb/²⁰⁴Pb_(25-23Ma) = 39.272–40.271 (Figure 6C,D). Megacrystic amphiboles plot within the Sr-Nd-Hf-Pb general Zealandia intraplate magmatism arrays (Figure 6); however, most WDS amphibole megacrysts show slightly less radiogenic Sr signatures compared to WDS intraplate volcanism (Figure 6A). ADS amphibole megacrysts, as well as intraplate magmatism, generally show similar Sr-Nd-Pb signatures compared to peridotite xenoliths hosted by intraplate magmatism (Figure 6). WDS amphibole megacrysts partly overlap with peridotite xenoliths in Sr-Nd signatures and only overlap with Pb signatures of the Anita Peridotite (Czertowicz et al. 2016) (Figure 6). Hf signatures in amphibole megacrysts are generally less radiogenic than SCLM xenoliths, except for xenoliths collected in the Burke River in the ADS, which show comparable Hf signatures to both amphibole megacrysts and Zealandia intraplate magmatism (Scott et al. 2014b; Scott et al. 2016a) (Figure 6B).

Figure 6. Age-corrected isotopic signatures of amphibole megacryst compared to published data from relevant Zealandia lithologies [i.e. EP Cenozoic intraplate volcanism (Price 2003; Hoernle et al. 2006; Sprung et al. 2007; Timm et al. 2009, 2010; Scanlan et al. 2020; Scott et al. 2020); EP Cretaceous intraplate volcanism (McCoy-West et al. 2010; Hoernle et al. 2020); EP peridotite xenoliths (Scott et al. 2014b; Scott et al. 2016a); WP Cretaceous intraplate volcanism (Timm et al. 2010; van der Meer et al. 2016, 2017; Hoernle et al. 2020); WP exhumed Anita Peridotite (Czertowicz et al. 2016)]. Amphibole megacryst signatures are age-corrected to the ages reported in Table 1. Literature data are age-corrected to ages reported in respective publications, except for Pb peridotite clinopyroxene data, for which only measured isotopic ratios are available. Approximate Tasman, MORB, HIMU and MORB fields are age-corrected to 60 Ma and are from Mortimer et al. (2012), Stracke et al. (2005) and Chauvel et al. (2008), respectively. Present-day NHRL is from Hart (1984) and the εHF vs. εNd mantle array is from Vervoort and Blichert-Toft (1999). Estimated reproducibility is shown on each figure.

Table 1. Radiogenic isotope results of Zealandia amphibole megacrysts.

Sample	Westland Dike Swarm				Alpine Dike Swarm
	QH12	KST-2	HOH-1	OU79943	MAC-1	FIS-4	LWA-12
Age (Ma)	∼ 88	87.7 ± 0.3	86.8 ± 1	69 ± 2.4	∼ 25	∼ 25	23.3 ± 0.1
Rb (ppm)	23.8	16.2	27.6	12.2	18.6	38.2	12.3
Sr (ppm)	693	619	918	560	438	508	578
^87Rb/^86Sr	0.09928	0.07578	0.08687	0.06324	0.1227	0.2171	0.06159
^87Sr/^86Sr	0.703015	0.702849	0.702957	0.703433	0.702886	0.703031	0.703013
± 2SE (abs.)	0.000006	0.000008	0.000005	0.000006	0.000006	0.000006	0.000005
^87Sr/^86Sr(i)	0.702891	0.702754	0.702850	0.703371	0.702842	0.702954	0.702991
Sm (ppm)	8.90	5.58	5.71	9.18	6.35	4.64	5.44
Nd (ppm)	38.3	21.7	22.1	34.4	27.9	19.3	21.1
^147Sm/^144Nd	0.1404	0.1553	0.1562	0.1612	0.1374	0.1451	0.1558
^143Nd/^144Nd	0.512888	0.512886	0.512880	0.512821	0.512830	0.512845	0.512853
± 2SE (abs.)	0.000006	0.000006	0.000005	0.000002	0.000007	0.000005	0.000004
εNd	4.9	4.8	4.7	3.6	3.7	4.0	4.2
^143Nd/^144Nd(i)	0.512807	0.512797	0.512791	0.512748	0.512808	0.512821	0.512828
εNd(i)	5.5	5.3	5.2	3.9	3.9	4.2	4.3
Lu (ppm)	0.310	0.225	0.210	0.233	0.0980	0.0772	0.123
Hf (ppm)	4.62	2.46	2.59	3.93	6.17	2.90	2.69
^176Lu/^177Hf	0.009535	0.01300	0.01153	0.008408	0.002254	0.003777	0.006505
^176Hf/^177Hf	0.282966	0.282947	0.282906	0.282880	0.282916	0.282901	0.282956
± 2SE (abs.)	0.000004	0.000004	0.000004	0.000004	0.000008	0.000005	0.000004
εHf	6.9	6.2	4.7	3.8	5.1	4.6	6.5
^176Hf/^177Hf(i)	0.282950	0.282926	0.282887	0.282869	0.282915	0.282899	0.282953
εHf(i)	8.2	7.4	6.0	5.0	5.6	5.0	6.9
^238U/^204Pb*	4.078	2.105	4.027	2.250	6.116	5.437	2.330
^235U/^204Pb*	0.0296	0.0153	0.0292	0.0163	0.0444	0.0394	0.0169
^238Th/^204Pb*	19.66	21.75	8.32	16.27	31.60	28.09	21.67
^206Pb/^204Pb	19.456	19.123	18.963	19.422	19.518	20.092	20.535
± 2SE (abs.)	0.001	0.001	0.001	0.001	0.001	0.002	0.001
^207Pb/^204Pb	15.657	15.643	15.654	15.649	15.661	15.679	15.672
± 2SE (abs.)	0.001	0.001	0.001	0.001	0.001	0.001	0.001
^208Pb/^204Pb	39.175	38.919	38.865	39.346	39.311	39.825	40.298
± 2SE (abs.)	0.003	0.003	0.003	0.003	0.003	0.004	0.003
^206Pb/^204Pb(i)	19.400	19.094	18.909	19.398	19.494	20.071	20.526
^207Pb/^204Pb(i)	15.657	15.642	15.654	15.649	15.661	15.678	15.672
^208Pb/^204Pb(i)	39.090	38.824	38.829	39.290	39.272	39.791	40.271

*Approximated parent/daughter ratios calculated with laser ablation data and natural isotopic abundances. Age-corrected data (i) are calculated with ages reported in the second row. WDS emplacement ages are from van der Meer et al. (2013, 2016); ADS emplacement ages are from published data from Cooper et al. (1987), Adams and Cooper (1996), Hoernle et al. (2006), Timm et al. (2010) and Cooper (2020).

Discussion

The petrogenesis of amphibole megacrysts

Kaersutitic amphibole megacrysts are a common occurrence in lamprophyric magmas worldwide. There is no single consensus on their origin, which has led to three main theories being proposed: (1) a cognate relationship between amphiboles and host rocks, with amphiboles representing high-pressure phenocrysts formed during magma ascent or in deep-seated magma chambers prior to eruption (e.g. Binns 1969; Merrill and Wyllie 1975; Irving and Frey 1984; Mayer et al. 2014; Ulrych et al. 2018); (2) a non-cognate relationship between amphiboles and host rocks, with amphiboles representing xenocrystic fragments from amphibole-rich veins in the upper mantle (e.g. Wilshire and Trask 1971; Best 1974; Boettcher and O’Neil 1980; Ben Othman et al. 1990; Shaw and Eyzaguirre 2000; Dobosi et al. 2003); and (3) amphibole megacrysts representing antecrysts that crystallised in deep and complex magma-plumbing systems (e.g. Ubide et al. 2014).

Mantle amphiboles often occur as major constituents in up to 30 cm thick hydrous hornblendite (± phlogopite) and/or amphibole-clinopyroxenite veins cross-cutting lherzolite xenoliths (e.g. Roden et al. 1984; Dautria et al. 1987; Nielson et al. 1993; Ionov and Hofmann 1995; Vaselli et al. 1995; Witt-Eickschen et al. 1998; Downes 2001; Dobosi et al. 2003; Witt-Eickschen 2003; Coltorti et al. 2004). Such xenoliths are interpreted to represent fragments of hornblendite and amphibole-pyroxenite veins from metasomatically enriched SCLM regions, analogous to hydrous veins reported in situ in the Lherz ultramafic massif in France (e.g. Zanetti et al. 1996; Bodinier 2004). Interestingly, occurrences of amphibole and phlogopite veins have also been reported as rare components to peridotite xenoliths entrained by Zealandia intraplate magmas (Dickey 1968; Tulloch and Nathan 1990; Scott et al. 2014b), as well as occurring in in situ hornblendite and orthopyroxene-amphibole veins in the Anita Peridotite ultramafic massif in Fiordland (Czertowicz et al. 2016). Moreover, the studied ADS and WDS amphibole megacrysts share strikingly similar trace element and REE patterns to upper mantle vein amphiboles from several studies (Figure 7), indicating similar formation processes, which warrants a more detailed look into megacryst – host melt relationships. To that end, we investigate available isotopic data from three WDS host rock – megacrystic amphibole pairs (QH12, KST-2 and OU79943), as well as available major and trace element data in amphibole microphenocrysts and megacrysts from four WDS samples (QH12, KST-2, OU79943 and MOA-16).

Figure 7. A, Trace element analyses of WDS and ADS megacrystic amphibole cores compared to SCLM vein amphiboles from the literature. Data are normalised to primitive mantle (McDonough and Sun 1995). B, REE analyses of WDS and ADS megacrystic amphibole cores compared to SCLM vein amphiboles from the literature. Data are normalised to chondrite (McDonough and Sun 1995). Black lines with solid markers: amphibole in hydrous veins found within worldwide lherzolite xenoliths. Black lines with empty markers: amphibole from in situ amphibole-rich veins in the Lherz ultramafic massif, France. Purple line with empty marker represents amphibole composition from a hornblendite vein found in situ within the Anita Peridotite.

Isotope systematics between WDS amphibole megacrysts and host rocks

Overlapping Sr-Nd-Hf-Pb signatures in both ADS and WDS amphibole megacrysts and intraplate magmatism arrays from Zealandia demonstrate that amphibole megacrysts are closely linked to the Zealandia intraplate magmatic system (Figure 6). In detail, phenocrysts should theoretically record the same isotopic signatures as the host melts in which they crystallised. However, melts rarely remain closed systems during ascent through the mantle and/or crust, where processes such as crustal contamination and element remobilisation are common. Thus, primary isotopic signatures of host melts are often modified during ascent. In this section, age-corrected isotopic signatures of three available amphibole megacryst – host rock pairs from the WDS are investigated in order to determine if amphibole megacrysts are in equilibrium with their respective host melts.

Figure 8 shows age-corrected Sr-Nd-Hf-Pb isotopic signatures for three WDS amphibole megacryst – host rock pairs relative to ADS megacryst amphiboles, as well as fields depicting both age-corrected Cretaceous and Cenozoic lamprophyric intraplate magmatism arrays. Amphibole megacrysts plot within the respective intraplate arrays, consistent with megacrysts being part of the Zealandia intraplate magmatic system. Host rocks show variable but consistently more radiogenic ⁸⁶Sr/⁸⁷Sr_(i) compared to associated amphibole megacrysts (especially so in sample OU79943) (Figure 8A). Enriched whole rock Sr signatures are interpreted as being a consequence of varying degrees of secondary hydrothermal alteration in WDS lamprophyres, as proposed by van der Meer et al. (2016, 2017) and consistent with petrographic evidence for secondary calcite. The OU79943 amphibole megacryst shows small offsets in Hf and Pb, as well as comparable Nd signatures relative to its host rock. As the OU79943 host rock is rather evolved and may have been subject to hydrothermal alteration ± crustal contamination (van der Meer et al. 2016, 2017), such slight discrepancies between host rock and amphibole megacrysts isotopic signatures may be interpreted to be secondary. Thus, the OU79943 amphibole megacryst may either have crystallised as a phenocryst in a host melt, which subsequently experienced secondary alteration, or the OU79943 megacryst could be xenocrystic but derived from a source with similar isotopic signatures. The QH12 megacrystic amphibole has comparable Hf and Nd isotopic signatures while showing substantially more radiogenic Pb signatures relative to its host rock (Figure 8C). Such Pb discrepancy between host rock and megacryst suggests that the analysed QH12 amphibole megacryst is not cognate to its host melt. In a similar manner, the KST-2 amphibole megacryst shows more evolved Nd signatures and slightly different Pb and Hf signatures compared to its host rock, also indicating that it did not crystallise from its host melt. Thus, based on host melt – amphibole megacryst isotopic signatures, at least two WDS amphibole megacrysts (KST-2 and QH12) cannot have crystallised from their host melts as reflected in the whole rock dike composition, and must have been collected by their host melts during ascent and/or storage in the lithosphere.

Figure 8. Age-corrected amphibole megacryst – host rock isotopic relationships using the ages indicated in Table 1. A, ¹⁴³Nd/¹⁴⁴Nd_(i) vs. ⁸⁶Sr/⁸⁷Sr_(i) diagram showing ¹⁴³Nd/¹⁴⁴Nd_(i) discrepancy between KST-2 amphibole and host rock isotopic signatures as well as variable ⁸⁶Sr/⁸⁷Sr_(i) discrepancies between WDS amphiboles and host rock isotopic signatures. B, εHf_(i) vs. εNd_(i) diagram showing εHf_(i) discrepancies between host rock and amphibole signatures in samples KST-2 and OU79943, as well as a large εNd_(i) discrepancy between host rock and amphibole signatures in sample KST-2. C, ²⁰⁸Pb/²⁰⁴Pb_(i) vs. ²⁰⁶Pb/²⁰⁴Pb_(i) diagram showing a major Pb isotopic discrepancy between QH12 amphibole megacryst and host rock signatures, as well as discrete ²⁰⁶Pb/²⁰⁴Pb_(i) isotopic discrepancies between host rock and amphibole signatures in samples KST-2 and OU79943. WP Cretaceous intraplate volcanism data from van der Meer et al. (2016, 2017) and Hoernle et al. (2020). EP Cenozoic intraplate volcanism data from Price (2003), Hoernle et al. (2006), Sprung et al. (2007), Timm et al. (2009, 2010), Scanlan et al. (2020) and Scott et al. (2020).

Geochemical systematics between WDS amphibole megacrysts and host rocks

Mg# variability in amphibole megacrysts. If amphibole megacrysts crystallised as phenocrysts in lamprophyric host melts, one would expect early-formed phenocrysts to have high Mg# and that subsequently formed rims and microphenocrysts to show lower Mg#, following a normal melt evolution. Amphibole megacrysts from sample MOA-16 show high Mg# (70) cores and low Mg# (61–68) rims and microphenocrysts (Figure 4), indicating that they may likely have crystallised as earlier phenocrysts within the host melt. Contrastingly, sample OU79943 exhibits a large population of megacryst cores with slightly lower Mg# (62–64) compared to rim and microphenocryst Mg# (65–69) (Figure 4). Moreover, amphibole megacryst cores show anomalously large variation towards low Mg# within WDS samples QH12 and KST-2 (i.e. Mg# 47–66) (Figure 4), as well as showing substantially lower Mg# in amphibole megacryst cores relative to both rims and microphenocrysts (Figure 4). Host lamprophyres typically evolve to lower MgO with increasing SiO₂. While subtle but consistently higher Mg# in OU79943 rims and microphenocryst can be interpreted to be caused by a single injection of a more primitive melt after megacryst crystallisation, neither the chemical evolution of QH12 and KST-2 host melts nor the injection of a primitive melt alone can account for the large Mg# variability in these amphiboles. Thus, consistent with the previously described isotopic discrepancies in megacryst – host rock signatures, low-Mg# QH12 and KST-2 amphibole megacryst cores did not crystallise in their respective host melts.

Trace element systematics between megacryst core, rims and microphenocrysts. Similar major and trace element characteristics of megacryst amphibole rims and microphenocrysts (Figures 4 and 5A,B) indicate that megacryst rims most likely grew concurrently with amphibole microphenocrysts in associated host rocks. This is supported by fractured megacryst cores showing signs of chemical reequilibration along fractures and showing matching pleochroism with megacryst rims and microphenocrysts (Figure 3). This suggests that cores were fractured before concurrent crystallisation of both microphenocrysts and megacryst rims. The absence of Eu anomalies in all amphiboles suggests that groundmass feldspar was the last crystallising phase and that no feldspars crystallised in tandem with either amphibole megacrysts or microphenocrysts (Figure 5A). Investigated megacryst cores show similar but lower trace element and REE patterns compared to rims and microphenocrysts (Figure 5A,B). Given the incompatibility of these elements in both amphibole and host melts, early-formed phenocrysts should have lower trace elements and REE abundances compared to microphenocrysts and megacryst rims, which represent later crystallising phases. Thus, the general trace element systematics between amphibole megacryst cores, rims and microphenocrysts seem to depict a cognate relationship between amphibole megacrysts and host melts (Figure 5). However, low-Mg# and host rock – megacrystic amphibole isotopic discrepancies in QH12 and KST-2 are inconsistent with a cognate model.

Based on the observations above, host rocks and amphibole trace element patterns are used to further constrain mineral-melt relationships. For this purpose, we calculate theoretical trace element compositions of amphiboles that are in equilibrium with host melt compositions at different degrees of crystallisation using the equation for fractional crystallisation [C_Amph = C_Host·D_Amph·F^DAmph−1 (where C_Amph is the concentration of an element in an amphibole; C_Host is the concentration of that element in the host melt for a remaining melt fraction F; and D_Amph is the partition coefficient of a given element in amphibole)]. Any potential pyroxene or olivine is omitted from the models as these are assumed to have minimal influence due to low partition coefficients of the trace elements of interest. Additionally, petrographic observations revealed that olivine is absent and pyroxenes are scarce in these samples. Feldspars are assumed to form after amphibole crystallisation based on the absence of Eu anomalies in all investigated amphiboles. Relevant published amphibole partition coefficients were collated from Adam and Green (2006) (for Rb, Ba, Th, U, Nb, Ta, La, Ce, Pb, Sr, Nd, Sm, Ho, Y, Yb and Lu) and Fulmer et al. (2010) (for Pr, Eu Gd, Dy and Er).

Modelled amphibole trace element patterns approaching groundmass microphenocryst compositions constrain the timing of crystallisation to the last ∼20% of remaining melt of host magmas (Figure 9A,B). Calculated amphibole compositions in equilibrium with host rocks at different amounts of remaining melt (25%–80%) generally cover the range of megacrystic amphibole core compositions in all three available host rock – amphibole pairs (Figure 9). Thus, these models show that even though low-Mg# QH12, KST-2 and potentially OU79943 amphiboles did not form in their respective host rocks, they must have crystallised in melts with similar trace element compositions to investigated host lamprophyres.

Figure 9. Modelling of theoretical fractionated amphibole trace element compositions compared to amphibole compositional ranges within each sample. Data are normalised to primitive mantle (McDonough and Sun 1995). A, Modelling of fractionated amphibole compositions calculated using QH12 host rock chemistry. B, Modelling of fractionated amphibole compositions calculated using OU79943 host rock chemistry. C, Modelling of fractionated amphibole compositions calculated using KST-2 host rock chemistry. Bulk rock data from van der Meer et al. (2016). Ti and K are omitted here due to large variations in published partition coefficients.

Evidence from host lamprophyre geochemical evolution. Bulk rock compositions of intraplate lamprophyres from both WDS and ADS show evidence of amphibole fractional crystallisation with decreasing Dy/Yb vs. increasing SiO₂ (Figure 10). This is further supported by Cooper (2020) who argued for fractionation of MREE-rich kaersutite (+ olivine, clinopyroxene, and titano-magnetite) from ADS melts in order to produce the compositional evolutionary trends between lamprophyres and phonolites of the ADS. Moreover, fractionation of such phases in lamprophyres would also account for the occurrence of anhedral clinopyroxene grains found within WDS host lamprophyres (Figure 3), which may represent remnants of fractionated phases. Considering that phonolites also occur in the WDS (Waight et al. 1998a) and that both ADS and WDS lamprophyre-phonolite arrays show decreasing Dy/Yb vs SiO₂ (Figure 10), it is tempting to envision the involvement of kaersutite fractionation in magma reservoirs to account for the lamprophyre-phonolite compositional range in both regions, in which investigated amphibole megacrysts represent remnants of such fractionated cumulate phases. WDS lamprophyres show a wide range of Mg# (37–68; van der Meer et al. 2017), which can account for the crystallisation of WDS megacrystic amphiboles with variable Mg# (47–70). Additionally, WDS Sr-Nd-Hf-Pb isotopic arrays overlap isotopic discrepancies between QH12 and KST-2 host rock and megacrystic amphiboles (Figure 8). Thus, indicating that amphibole megacrysts likely crystallised from lamprophyric melts related to the WDS.

Figure 10. Lamprophyre to phonolite raw Dy/Yb vs. SiO₂ arrays in the WDS and ADS. Grey arrow shows the negative correlation between Dy/Yb and SiO₂, indicative of amphibole fractionation. WDS bulk rock data is from van der Meer et al. (2016, 2017) and ADS bulk rock data from Timm et al. (2010) and Cooper (2020).

Amphibole thermobarometry

Tentative geothermobarometric calculations based on the partitioning of Al in amphibole were performed on megacryst cores using equations from Putirka (2016) coupled with the amphibole hygrometer from Ridolfi and Renzulli (2012), with the caveat that it is unclear how applicable these thermobarometers are to lamprophyric systems. Pressure estimates on individual amphiboles can yield uncertainties of ± 4 kbar (Putirka 2016), which are not trivial. Thus, we attempt to mitigate potential uncertainties by averaging pressure estimates of several amphiboles from each sample as recommended by Putirka (2016). Results suggest that ADS amphibole megacrysts crystallised from magmas having 4–6.1 wt.% H₂O at pressure conditions of 8.3–9.6 kbar and temperatures of 1031–1053°C, whereas WDS amphibole megacrysts crystallised in pressure conditions of 7.5–8.6 kbar and temperatures of 1024–1062°C in magmas containing 4.7–5.6 wt.% H₂O. These results are consistent with the experimental study of Pilet et al. (2010), which demonstrated that kaersutite in basanitic melts with 5–6 wt.% H₂O is stable at conditions under 1130°C and 15 kbar. By following a conventional conversion factor of 1 kbar ∼ 3.7 km of crust, ADS and WDS amphibole megacryst are estimated to have crystallised around 31–36 km and 28–32 km depth, respectively. Based on present-day estimations of 35–48 km Moho depth (Salmon et al. 2013) and assumed Moho depths of ∼ 37 km during the Cenozoic in the ADS region (e.g. Scott et al. 2014b), it can be inferred that ADS amphibole megacrysts may have crystallised at depths equivalent to the lowermost crust. The crystallisation setting of WDS amphibole megacrysts is harder to constrain as Moho depths during the Late Cretaceous have not been explicitly studied. However, assuming an averaged extensional setting with a ∼ 30 km thick crust (Christensen and Mooney 1995), the crystallisation of WDS megacrystic amphiboles may have occurred close to the crust/mantle boundary.

Implications for amphibole megacryst petrogenesis in Zealandia

Investigating amphibole megacryst geochemical and isotopic characteristics alone is not adequate to fully determine their petrogenesis. Evidently, detailed investigations of isotopic and geochemical relationships in megacryst – host rock pairs are required to properly distinguish phenocrysts from xenocrysts. Even though we currently do not possess a full geochemical dataset of investigated amphibole megacrysts and their respective host rocks, the strong similarities between both amphibole megacrysts and host rocks from the WDS and ADS infer that similar processes may act in both regions. Even though the presented geochemical evidence suggests that amphibole megacrysts crystallised from lamprophyric melts, we cannot ignore that megacrystic amphiboles share remarkable trace element and REE compositional similarities to SCLM vein amphibole compositions (Figure 7). As such, two models for the petrogenesis of amphibole megacrysts are considered.

The lamprophyre model. Considering tentative geothermobarometric estimations, megacrystic amphibole may have crystallised at lower crustal depths. Thus, low-Mg# amphibole megacrysts likely fractionated from low-Mg# lamprophyric melts in lower crustal magma-plumbing systems. Repeated melting of an isotopically moderately heterogeneous source region initiated the injection of more primitive lamprophyric host melts and promoted repeated magma mingling in lower crustal settings, resulting in both the variable Mg# and isotopic spread in megacryst cores as well as the high Mg# in megacryst rims and microphenocrysts. Alternatively, QH12 and KST-2 amphibole megacrysts may have fractionated in magma reservoirs from precursor low-Mg# lamprophyric melts that remained emplaced in deeper settings than the investigated WDS lamprophyric host rocks. Subsequent ascent of WDS host melts may have reactivated such reservoirs and collected xenocrystic cumulate amphibole megacrysts. Models involving different generations of lamprophyric melts in the WDS are further supported by the fact that the KST-2 lamprophyre host rock is part of a composite dike showing diffuse contacts with an associated amphibole megacryst-bearing phonolitic unit (Waight et al. 1998a; van der Meer et al. 2013), indicating the simultaneous presence of magmas of contrasting compositions and indicating that some degree of magma mixing and/or magma reservoir reactivation could have occurred in the WDS. This is also consistent with the presence of oscillatory zoning and sieve textures in plagioclase phenocrysts in doleritic members of the Hohonu Dike Swarm (Waight 1995)

The vein model. Low-Mg# megacrystic amphiboles could represent xenocrysts sampled from a veined SCLM. Such amphiboles would have crystallised in precursor hydrous melts, which were emplaced as metasomatic veins in the SCLM. Subsequent infiltration of host lamprophyric melts used the rheologically weaker veins as conduits and collected both fractured megacrystic amphiboles as well as peridotite xenoliths during their ascent through the SCLM. Recent studies have proposed that amphibole-rich veins in the SCLM may represent important components in the sources for lamprophyric magmatism in Zealandia (Panter 2006; Sprung et al. 2007; Scott et al. 2016a; van der Meer et al. 2017) and it is tempting to link the investigated amphibole megacrysts to such SCLM lithologies. In such a model, lamprophyre compositions would inherit similar trace element and REE patterns to amphibole compositions, as amphibole would be the main contributor of such elements. This is partly the case (Figure 5), but an amphibole-rich SCLM source also requires an additional Th, and U-rich component to account for high concentrations in Zealandia lamprophyres. In a vein model, isotopic discrepancies between host rock – megacryst pairs would be expected if lamprophyres were derived from a heterogeneously metasomatised and veined SCLM source. Namely, WDS and ADS lamprophyres would record mixed isotopic signatures of heterogeneous melting sources, in which amphibole xenocrysts record primary isotopic signatures of individual source components. However, reconciling a vein model with the presented data raises some issues. Firstly, despite showing comparable trace element and REE patterns to megacrystic amphiboles, published vein amphiboles show exclusively high Mg# (70–90) (e.g. Ionov and Hofmann 1995; Zanetti et al. 1996; Coltorti et al. 2004). A single hornblendite vein in the Anita Peridotite hosts low-Mg# (52) amphibole (Czertowicz et al. 2016). However, the amphibole also displays heavily contrasting trace element and REE patterns compared to the studied megacrysts, which argues for disparate formation mechanisms (Figure 7). One could argue for Fe-rich basanitic melts to have formed amphibole-rich metasomes with low-Mg# in order to accommodate the vein model (Frost 2006). However, such arguments remain speculative. Additionally, albeit being fairly uncertain, geobarometric estimations may place the formation depth of amphibole megacrysts in the lowermost crust instead of the SCLM.

In both models, amphibole megacryst cores that display higher than or comparable Mg# compared to microphenocryst compositions may represent cognate phenocrysts within host rocks. Additionally, amphibole megacrysts rims and host rock microphenocrysts likely crystallised during final emplacement of investigated WDS lamprophyres. The presented data cannot decisively constrain one single petrogenetic model for amphibole megacryst formation, but we currently favour the lamprophyre model as it more elegantly explains the data.

Comparison of WDS and ADS amphibole megacrysts

Both WDS and ADS megacrystic amphiboles fall into the classification of the magnesio-hastingsite – kaersutite series. Moreover, amphibole megacrysts from both regions show remarkably similar trace element and REE patterns (Figure 5C,D) as well as overlapping restricted ranges of Sr-Nd-Hf isotopic signatures (i.e. ⁸⁷Sr/⁸⁶Sr_(i) = 0.702754–0.703371, εNd_(i) = +3.9 to +5.5 and εHf_(i)= +5.0 to +8.2) (Figure 6), which suggests that they formed under comparable conditions and from similar sources – despite an age difference of over 50 Ma. Such compositional similarities may simply be related to the megacrysts being hosted by compositionally and isotopically similar host lamprophyres in both regions (Figures 5C,D and 6). However, given the different settings in which host lamprophyre dike swarms were emplaced as well as being both spatially and temporally decoupled, such amphibole megacrysts are a peculiarity worth investigating.

Despite similarities, two key differences between ADS and WDS amphibole megacrysts are notable. The first difference is that amphibole megacrysts from the ADS show lower HREE concentrations relative to WDS megacrysts, which is a trait shared by ultramafic lamprophyres of the ADS (Cooper 2020). The resulting steeper LREE/HREE pattern in ADS megacrysts could be associated to HREE retention by garnet during partial melting in ADS source regions, which lowered the HREE budget of ADS host lamprophyres (Cooper 2020). It is noteworthy that peridotite xenoliths entrained by intraplate alkaline magmatism in Zealandia exclusively represent spinel-facies peridotites (Scott et al. 2020, and references therein). Thus, indicating either that ADS lamprophyres were derived in part from melting of garnet-facies peridotite and only incorporated xenoliths from shallower regions, or that steeper LREE/HREE patterns are caused by other processes. For example, such patterns can also be derived if host melts were (at least partially) sourced by partial melting of a depleted SCLM, which subsequently experienced heterogeneous enrichment of LREE through carbonatitic metasomatism (e.g. Scott et al. 2014a, 2014b; McCoy-West et al. 2015; Scott et al. 2016a, 2016b). The second notable difference is that compared to WDS amphibole megacrysts, ADS amphiboles show more radiogenic ²⁰⁶Pb/²⁰⁴Pb_(i) and ²⁰⁸Pb/²⁰⁴Pb_(i) signatures as well as a larger spread in these isotopic ratios (Figure 6C,D). Radiogenic and variable HIMU-like ²⁰⁶Pb/²⁰⁴Pb_(i) and ²⁰⁸Pb/²⁰⁴Pb_(i) signatures have been reported in both Zealandia intraplate magmatic rocks (Hoernle et al. 2006; Panter 2006; Sprung et al. 2007; Timm et al. 2009, 2010; van der Meer et al. 2016, 2017; Scott et al. 2020) and in SCLM peridotite xenoliths entrained by intraplate magmatism (Scott et al. 2014a, 2014b; McCoy-West et al. 2016; Scott et al. 2016a) (Figure 6C,D). Thus, pronounced HIMU-like Pb signatures in ADS megacrystic amphiboles may be inherited by a local enrichment of ²⁰⁶Pb/²⁰⁴Pb_(i) and ²⁰⁸Pb/²⁰⁴Pb_(i) in the source regions. As the main differences between WDS and ADS megacrystic amphiboles are derived by processes driven by regional differences, the next section will investigate isotopic signatures of amphibole megacrysts in relation to intraplate magmatism in Zealandia in order to place constrains on source characteristics and evolution.

Amphibole megacryst isotopic signatures in a Zealandia setting

Even though amphibole megacrysts share comparable compositions and isotopic signatures, they each belong to two separate instances of Zealandia intraplate magmatism emplaced in different regional settings and at different times. Both WDS and ADS magmatism belong to widespread alkaline intraplate magmatism that has been emplaced throughout Zealandia since the Late Cretaceous, with HIMU-like Sr-Nd-Hf-Pb signatures with ⁸⁷Sr/⁸⁶Sr_(i) ∼ 0.703–0.704, εNd_(i) = +2 to +8, εHf_(i) = +2 to +10, ²⁰⁶Pb/²⁰⁴Pb_(i) = 18.5–20.7 and ²⁰⁸Pb/²⁰⁴Pb_(i) = 38.5–41 (Price 2003; Hoernle et al. 2006; Panter 2006; Sprung et al. 2007; Timm et al. 2009, 2010; McCoy-West et al. 2010; van der Meer et al. 2017; Hoernle et al. 2020; Scanlan et al. 2020; Scott et al. 2020).

WDS and ADS amphibole megacrysts plot within or close to both Cretaceous and Cenozoic intraplate alkaline magmatism arrays in Sr-Nd-Hf-Pb isotopic systems (Figure 6A–D), forming a general array between MORB and HIMU compositions. Additionally, intraplate magmatism and megacrysts share similar Pb-Nd-Sr-Hf signatures to spinel-facies peridotite xenoliths originating from the Zealandia SCLM (e.g. Scott et al. 2014b; McCoy-West et al. 2016; Scott et al. 2016a; van der Meer et al. 2017) (Figure 6). HIMU-like signatures in SCLM xenoliths in Zealandia have generally been attributed to subduction-related carbonatitic mantle metasomatism, broadly constrained to have occurred sometime in the Cretaceous (Scott et al. 2014a, 2014b; McCoy-West et al. 2016; Scott et al. 2016a; Dalton et al. 2017; van der Meer et al. 2017). Even though the origin of the HIMU-like Sr-Nd-Hf-Pb signatures in alkaline intraplate magmatism in Zealandia still remains under debate, we favour derivation from melting of a metasomatically enriched SCLM. As such, variable carbonatitic metasomatism in lamprophyre source regions is believed to have caused heterogeneous enrichments in Th and U sometime during the Cretaceous, which by ingrowth produced the observed spread in radiogenic ²⁰⁶Pb/²⁰⁴Pb_(i) and ²⁰⁸Pb/²⁰⁴Pb_(i) signatures in ADS and WDS host rocks and megacrystic amphiboles while retaining similar ²⁰⁷Pb/²⁰⁴Pb_(i) and Sr-Nd-Hf signatures (Figure 6A,C,D). Interestingly, investigated megacrystic amphiboles show HIMU-like ²⁰⁶Pb/²⁰⁴Pb_(i) and ²⁰⁸Pb/²⁰⁴Pb_(i) signatures that span the entire Zealandia lamprophyre array (Figure 6). Additionally, we suggest that is not coincidental that Cretaceous WDS amphiboles record both the relatively most constrained and least radiogenic Pb signatures, while Cenozoic ADS megacrysts record the most radiogenic and diverse Pb signatures (Figure 6). Due to the very low compatibility of Th and U in amphibole, cognate amphibole megacrysts [with low U-Th/Pb ratios (²³⁸U/²⁰⁴Pb ∼ 2–6 and ²³²Th/²⁰⁴Pb ∼ 8–32)] have negligible ingrowth and therefore preserve the Pb isotopic composition of the parental melts and sources at the time of crystallisation. Assuming that variable enrichment due to metasomatism occurred only a relatively short period prior to WDS magmatism, Late Cretaceous WDS megacrystic amphiboles should record relatively unradiogenic and constrained Pb signatures closer to those of pre-metasomatised sources as Th and U enrichment had limited time to substantially modify Pb signatures. By Cenozoic (∼ 25 Ma) times, megacrystic amphiboles will record more radiogenic and varied Pb signatures since longer time has spanned since heterogeneous U and Th enrichment, resulting in more variable degrees of ingrowth in the mantle sources. Such a relationship is exactly what is observed (Figure 6C,D). Thus, WDS and ADS amphibole megacryst record time-constrained isotopic snapshots of in situ ²⁰⁶Pb and ²⁰⁸Pb ingrowth in mantle sources.

Assuming that megacrystic amphiboles represent early-crystallised phases in Zealandia intraplate lamprophyre magmatic systems, WDS and ADS megacrystic amphiboles should record primary isotopic compositions of sources from which associated host melts originate. Thus, it can be inferred that amphibole megacrysts record moderate heterogeneities in mantle sources that mix to form the variation in Sr-Nd-Hf signatures of intraplate alkaline magmatic arrays. Considering the comparable geochemical and isotopic characteristics of WDS and ADS host rocks and amphibole megacrysts, one can speculate that alkaline intraplate magmatism throughout Zealandia may be derived from similar processes acting in a moderately heterogeneous SCLM reservoir that experienced variable carbonatitic metasomatism.

To test the aforementioned hypothesis, we present Nd-Hf evolution models of a theoretical mantle source, using WDS amphibole megacryst initial isotopic signatures coupled with relevant published mantle reservoir parent/daughter ratios [i.e. Tasman-MORB ¹⁴⁷Sm/¹⁴⁴Nd from Mortimer et al. (2012); average MORB ¹⁷⁶Lu/¹⁷⁷Hf and ¹⁴⁷Sm/¹⁴⁴Nd from Chauvel et al. (2008) and DMM ¹⁴⁷Sm/¹⁴⁴Nd and ¹⁷⁶Lu/¹⁷⁷Hf from Salters and Stracke (2004)] (Figure 11A,B). Calculated models are compared to age-corrected compositions for Zealandia intraplate magmas since the Late Cretaceous (Price 2003; Hoernle et al. 2006; Sprung et al. 2007; Timm et al. 2009, 2010; McCoy-West et al. 2010; van der Meer et al. 2017; Hoernle et al. 2020; Scanlan et al. 2020; Scott et al. 2020), present day isotopic signatures of SCLM peridotite xenoliths (Scott et al. 2014b; Scott et al. 2016a) and relevant mantle reservoirs (Figure 11). Strikingly, amphibole megacrysts from the ADS and most Zealandia intraplate magmas since the Late Cretaceous plot along calculated mantle source evolution fields in Nd-Hf space (Figure 11A,B), further indicating that intraplate magmatism in Zealandia is derived from similar sources and processes since the break-up of Gondwana. Late Cretaceous EP Marlborough province magmatism display lower ¹⁴³Nd/¹⁴⁴Nd_(i) and ¹⁷⁶Hf/¹⁷⁷Hf_(i) than predicted source signatures (Figure 11A,B), and may instead be derived from a distinct source originating from the subducting Hikurangi Plateau as proposed by van der Meer et al. (2017). Interestingly, calculated sources evolve to present-day ¹⁴³Nd/¹⁴⁴Nd₍₀₎ and ¹⁷⁶Hf/¹⁷⁷Hf₍₀₎ that match signatures of SCLM peridotite xenoliths sampled in Burke River in the ADS by Scott et al. (2016a) (Figure 11A,B), who proposed that intraplate magmatism in Zealandia may be derived from such a metasomatically enriched and amphibole-bearing SCLM.

Figure 11. A and B, Theoretical Hf and Nd isotope evolution fields for a modelled reservoir using WDS megacryst amphibole initial isotopic ratios coupled with parent/daughter ratios from relevant reservoirs. C and D, Modelled effect of variable U and Th enrichment on Pb isotopic evolutions. Modelled evolution lines are plotted against age-corrected amphibole megacrysts (this study), dredged Tasman MORB evolution from Mortimer et al. (2012); average MORB evolution from Chauvel et al. (2008); approx. DMM evolution from Salters and Stracke (2004) and Stracke et al. (2005) with ²³⁸U/²⁰⁴Pb = 13 from White (1993); age-corrected EP alkaline intraplate volcanism data from Price (2003); Hoernle et al. (2006); Sprung et al. (2007); Timm et al. (2009, 2010); McCoy-West et al. (2010); Hoernle et al. (2020); Scanlan et al. (2020) and Scott et al. (2020); age-corrected WP alkaline intraplate volcanism data from Timm et al. (2010), van der Meer et al. (2016, 2017) and Hoernle et al. (2020); and present-day isotopic signatures of EP SCLM peridotite xenoliths from Scott et al. (2014b) and Scott et al. (2016a).

Intraplate magmatism in the WP and especially EP shows varying degrees of radiogenic Pb-enrichment towards HIMU-like signatures (Figure 11C,D) compared to both Tasman MORB and DMM reservoir evolutions. As discussed, modified Pb signatures have previously been attributed to variable carbonatitic metasomatism in source regions (McCoy-West et al. 2016; Scott et al. 2016a; van der Meer et al. 2017), consistent with reported heterogeneous carbonatitic metasomatism in SCLM xenoliths (e.g. Scott et al. 2014a, 2014b). Additionally, present day ²⁰⁶Pb/²⁰⁴Pb₍₀₎ and ²⁰⁸Pb/²⁰⁴Pb₍₀₎ in SCLM xenoliths exhibit a wide range of signatures, similar to amphibole megacrysts and intraplate magmatism in Zealandia (Figure 11C,D). It is also noteworthy that the least radiogenic Pb signatures in both intraplate magmatism and SCLM xenoliths intersect with the Tasman-MORB Pb evolution fields (Figure 11C,D). As we have established that amphibole megacrysts likely represent phases which crystallised within lamprophyric melts of the ADS and WDS, investigating such spatially and temporally separate instances of low U-Th/Pb megacrystic amphiboles provide unique insight into the Pb isotopic history of sources from which Zealandia lamprophyres are derived. Studied megacrystic amphiboles record the full spectrum of lamprophyre Pb signatures (Figure 6) and are therefore inferred to also record the full extent of heterogeneous carbonatitic U and Th enrichment in lamprophyre source regions. Late Cretaceous WDS megacrystic amphiboles likely record an early snapshot into the isotopic evolution of Pb in lamprophyre sources before much ingrowth had occurred, whereas Cenozoic ADS megacrystic amphiboles record both the effects of prolonged Pb ingrowth as well as the heterogeneous nature of carbonatitic metasomatism in source regions. Additionally, the close isotopic relationship between amphibole megacrysts, intraplate magmatism and SCLM peridotites indicates that intraplate magmatism may be derived from a metasomatically enriched SCLM source.

Thus, using present-day isotopic signatures of SCLM peridotites coupled with parent-daughter ratios from published Zealandia intraplate intrusives (i.e. ²³⁸U/²⁰⁴Pb ∼ 10–120 and ²³²Th/²⁰⁴Pb ∼ 30–300), we can extrapolate appropriate Pb isotopic evolutions to intersect with both ADS and WDS megacrystic amphiboles as well as EP and WP lamprophyre signatures to t ∼ 90 Ma (Figure 11C,D). Modelled Pb ingrowth regressions cover the range of HIMU-like intraplate magmatism Pb signatures throughout most of Zealandia since the Late Cretaceous (Figure 11C,D), except 90–100 Ma EP Marlborough Province magmatism, which has previously been proposed to tap a distinct source with diluted HIMU signatures originating from the subducting Hikurangi Plateau (van der Meer et al. 2017). Pb isotopic regressions also indicate that intraplate magmatism in Zealandia is derived from sources which originally had similar isotopic compositions to Tasman MORB prior to Th and U enrichment (Figure 11C,D). By using Pb regressions to fully encompass Pb ingrowth in intraplate magmatism, we can broadly constrain the timing of metasomatic Th and U enrichment to have occurred in the Jurassic to Middle Cretaceous in both ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb space (Figure 11C,D), consistent with published estimates (Scott et al. 2014a, 2014b; McCoy-West et al. 2016; van der Meer et al. 2017).

In summary, amphibole megacryst and lamprophyre isotopic signatures depict a moderately heterogeneous SCLM source with Tasman MORB-like Nd-Hf parent/daughter isotope ratios. This SCLM source likely originally had similar Pb isotopic compositions to Tasman MORB, which was modified by localised varying degrees of subduction-related heterogeneous carbonatitic U and Th enrichment between the Jurassic and Middle Cretaceous. Our results support studies advocating melting of subduction-modified lithospheric mantle to account for alkaline intraplate magmatism in Zealandia (e.g. Scott et al. 2016a; van der Meer et al. 2017; van der Meer et al. 2019; Cooper 2020).

Conclusions

This study presents a comprehensive Sr-Nd-Hf-Pb isotopic and major/trace element dataset of amphibole megacrysts hosted by two spatially and temporally decoupled instances of alkaline intraplate magmatism in Zealandia. Results provide new insights into the genesis of both amphibole megacrysts and alkaline intraplate magmatism in Zealandia.

1. Geochemical and isotopic evidence suggest that amphibole megacrysts likely represent phenocrysts and/or remnants of fractionated phases belonging to lamprophyric melts.

2. Discrepancies in Mg# and isotopic signatures in host rock – amphibole megacryst pairs are interpreted to be a consequence of interaction between lamprophyric melts and/or cumulates in the lower crust derived from repeated tapping of a moderately heterogeneous mantle source.

3. Amphibole megacrysts from both the Cenozoic ADS and Cretaceous WDS show remarkably similar trace element and REE patterns as well as overlapping narrow ranges of Sr-Nd-Hf isotopic signatures (i.e. ⁸⁷Sr/⁸⁶Sr_(i) = 0.702754–0.703371, εNd_(i) = +3.9 to +5.5 and εHf_(i)= +5.0 to +8.2), suggesting similar formation processes.

4. Diverse HIMU-like Pb isotopic signatures coupled with low Th-U/Pb ratios in amphibole megacrysts record the evolution of ²⁰⁶Pb and ²⁰⁸Pb ingrowth in intraplate sources since U and Th enrichment by carbonatitic metasomatism in the Jurassic to Middle Cretaceous.

5. Source modelling suggest that most alkaline intraplate magmatism throughout Zealandia is derived from moderately heterogeneous SCLM sources that suffered localised and varying degrees of subduction-related carbonatitic metasomatism between the Jurassic to Middle Cretaceous during Gondwana break-up, in agreement with results from recent studies.

Acknowledgements

We would like to thank Antonio M. Álvarez-Valero and an anonymous reviewer for very helpful comments that helped improve the final version of the paper. We thank Toni Larsen, Toby Leeper, Mojagan Alaei and Olga Nielsen for assistance in the laboratory work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

A detailed description of methodologies and standards as well as data that support the findings of this study are openly available in Figshare at 10.6084/m9.figshare.12057843.

Additional information

Funding

This work was supported by Natur og Univers, Det Frie Forskningsråd to T.E. Waight [Grant Number: 10-085245].