New U–Pb zircon ages and a revised integrated age model for the late Miocene northern Taranaki coastal section, New Zealand

ABSTRACT

New LA–ICP–MS U–Pb zircon ages of tuffs, from the northern Taranaki coastal section constrain depositional ages of the Mount Messenger Formation to 9.69 ± 0.12 and 9.43 ± 0.17 Ma at the Mohakatino and Tongaporutu rivers, respectively, and the Urenui Formation near Waiau Stream to 8.45 ± 0.10 Ma. Our preferred new radiometric–biostratigraphic age model comprises lower and upper intervals characterised by relatively low and high un-decompacted sedimentation rates of 102 ± 96 m/Ma (MSWD = 6.0) and 1358 ± 144 m/Ma (MSWD = 1.32), respectively. The poor linear fit to the lower interval could indicate either variable sedimentation rates, or an unconformity (0.7 ± 0.2 Ma) in the upper Mohakatino Formation or at the base of the Mount Messenger Formation. The 1569 m-thick upper interval between Mohakatino River and Waiau Stream was deposited in 1.16 ± 0.01 Ma between 9.67 + 0.07/−0.05 and 8.52 + 0.06/−0.07 Ma. Predicted ages for intra-Tongaporutuan bioevents (e.g. base Tukemokihi Coiling Zone) do not overlap their assigned ages, highlighting the importance of propagating uncertainties into bioevent absolute ages and opportunities for future improvements to the age model, namely using high-precision geochronology to acquire ages for additional tuffs and improve precision of existing ages.

KEYWORDS:

• U–Pb zircon
• tuff
• Taranaki Basin
• Mount Messenger Formation
• Urenui Formation
• Mohakatino Formation
• Miocene
• Tongaporutuan Stage
• Mohakatino Volcanic Centre
• Northern Taranaki

Introduction

Exposures of late Miocene deep-water sedimentary rocks along the northern Taranaki coast, New Zealand, include the primary reference section for the New Zealand Tongaporutuan Stage (Finlay and Marwick 1947; Crundwell et al. 2004) and exhibit stratigraphic relations for deep-water sedimentary systems of international significance, such as the architecture of slope channel and fan systems (e.g. King et al. 2007). Despite several attempts at using radiometric dating methods to determine absolute ages for tuff beds interbedded with the siliciclastic sandstone and mudstone that dominate the coastal exposures (Bergman et al. 1990; Hansen 1996; Maier et al. 2016), the chronology of the coastal section remains problematic for various reasons including, the paucity of horizons containing autocrystic zircons (i.e. primary tuffs) for radiometric dating, sedimentary reworking, and mass slope failure.

In this paper, we report new U–Pb zircon ages for several tuffs from the northern Taranaki coastal section. An improved age model for the coastal section generated by integrating our new ages with existing radiometric and biostratigraphic data, further refines sedimentation rate estimates and the duration of a possible unconformity. The new U–Pb zircon ages also allow the absolute ages of several intra-Tongaporutuan bioevents (e.g. Tukemokihi Coiling Zone) to be independently evaluated.

Background

Stratigraphy

At the broadest stratigraphic level, the rocks of Zealandia (Luyendyk 1995; Mortimer et al. 2017) can be divided into Austral Superprovince ‘basement’ rocks, and the unconformably overlying Zealandia Megasequence ‘cover’ rocks (Mortimer et al. 2014). The basement is subdivided into the Western and Eastern provinces, which are further subdivided into tectonostratigraphic terranes and batholiths (Landis and Coombs 1967; Coombs et al. 1976; Howell 1980; Bradshaw 1989; Mortimer et al. 1999; Mortimer 2004). The cover rocks represent a ∼100 Ma transgressive–regressive first-order cycle and are further subdivided into five supergroups (Momotu, Haerenga, Waka, Māui and Pākihi) that correspond to one or more second-order cycles (Mortimer et al. 2014). The Neogene Māui Supergroup is of interest to this paper. It comprises sedimentary rocks that record marine regression and a regional change to siliciclastic-dominated deposition and subduction-related volcanic rocks, both of which reflect the initiation and development of the modern Australian–Pacific plate boundary through Zealandia (e.g. King 2000).

Northern Taranaki coast

Geology

The sedimentary rocks exposed along the northern Taranaki coast belong to the Miocene Waiiti Group (part of the Māui Supergroup; King and Thrasher 1996), consisting of several formations; the Manganui, Mohakatino, Mount Messenger, and Urenui formations (Figures 1 and 2). The Manganui Formation ranges from early Miocene (Altonian) to Pliocene and is dominated by deep-marine mudstone with thin sandstone and limestone stringers that represents the prevailing middle–lower slope depositional environment into which the interfingering Mohakatino and Mount Messenger formations were deposited (Figure 1A; Figure 2). The Mohakatino Formation is characterised by volcaniclastic rocks derived from the Miocene submarine Mohakatino Volcanic Centre (Bergman et al. 1992), and ranges in age from middle to late Miocene (Lillburnian–Tongaporutuan). The upper Mohakatino Formation interfingers with the late Miocene (Tongaporutuan) Mount Messenger Formation (Figure 1A), which is dominated by siliciclastic mudstone and sandstone totalling ∼875 m (Figure 3). The sandstone beds are interpreted as gravity flow deposits that formed in basin-floor and base-of-slope fans (∼1000–2000 m water depth; Figure 1A). Interbedded mudstone represents both pelagic sedimentation and the fine-grained portion of gravity flows (King et al. 1993; King and Thrasher 1996; King et al. 2007; Rotzien et al. 2014; Masalimova et al. 2016). Significant intervals of the lower Mount Messenger Formation, between the Mohakatino and Tongaporutu rivers, consist of mass transport deposits (MTDs) collectively called the Otukehu MTD (Browne et al. 2006; King et al. 2011; Masalimova et al. 2016). Molluscan and benthic foraminiferal events observed in the late Miocene coastal exposures of the Mount Messenger Formation along the northern Taranaki coastline were originally used to define the Tongaporutuan Stage boundaries (Finlay and Marwick 1947). However, these faunal criteria are now considered to be unsuitable as biostratigraphic markers and planktic foraminiferal events are now used to delineate the Miocene stages (Crundwell et al. 2004). Mudstone makes up most of the overlying Urenui Formation, which is also of Tongaporutuan age and has a thickness of ∼800 m (Figure 3). The Urenui Formation was deposited in a slope setting (∼200–1000 m depth), which included conglomerate- and sandstone-filled canyons and channel lithofacies (King et al. 1993; King et al. 2007). Thin beds of tuff are a minor but conspicuous constituent of both the Mount Messenger and Urenui formations.

Figure 1. (A) Early Tongaporutuan (late Miocene, 10 Ma) paleogeographic map (Arnot and Bland 2016) of the Taranaki Basin showing the present-day coastline and the location of the northern Taranaki coastal section; (B) Simplified geological map of the northern Taranaki coastal section (Townsend et al. 2008) showing the locations of both new and existing U–Pb zircon tuff samples. Coordinates are given in the New Zealand Transverse Mercator (NZTM) projection relative to the New Zealand Geodetic Datum 2000 (NZGD2000).

Figure 2. Chronostratigraphic diagram illustrating the relationships between late Cenozoic lithostratigraphic units in the Taranaki Basin.

Figure 3. Simplified lithostratigraphic column for the northern Taranaki coastal section (after Maier et al. 2016) showing the positions and error-weighted mean ²⁰⁶Pb/²³⁸U zircon ages (corrected for common Pb using ²⁰⁷Pb and initial ²³⁰Th disequilibrium; 2σ uncertainties) of existing and new tuff samples (* = sample did not yield any late Miocene zircons). Ages for existing samples have been recalculated as detailed in the text.

The Waiiti Group records a regional change from carbonate sedimentation that dominated the Oligocene–Early Miocene (Waka Supergroup), to siliciclastic-dominated deposition. The drivers of this change were uplift and erosion of the eastern Taranaki Basin margin and central New Zealand related to convergence/shortening across the Australian–Pacific plate boundary (King and Thrasher 1996; Stagpoole and Nicol 2008). Mineralogical, geochemical, and detrital zircon U–Pb geochronological studies indicate that the provenance of the Waiiti Group is dominantly volcanogenic and quartzofeldspathic low-grade metasedimentary rocks of the Permian–Jurassic Torlesse and Waipapa composite terranes, which presently occur east of the Taranaki Basin (King and Thrasher 1996; Rotzien et al. 2018). In contrast, the underlying volcaniclastic Mohakatino Formation and tuffs within the Mount Messenger and Urenui formations were derived from the Mohakatino Volcanic Centre (Nodder et al. 1990; Bergman et al. 1992; Maier et al. 2016).

The Mohakatino Volcanic Centre is a NNE-trending chain of >20 submerged and buried middle–late Miocene calc-alkaline intermediate (basaltic andesite–andesite) stratovolcanoes located in the offshore area of the northern Taranaki Basin (Figure 1A) (Bergman et al. 1992; King and Thrasher 1996). Volcanism in the northern Taranaki Basin migrated progressively in a SSE direction during the Miocene as part of the regional southward migration of Australia–Pacific plate boundary arc volcanism (Ballance 1976; Bergman et al. 1992; Hayward et al. 2001; Giba et al. 2010).

Geochronology

Radiometric dating methods have been applied to the northern Taranaki coastal section to (1) constrain absolute ages for tuff beds (Bergman et al. 1990; Hansen 1996; Maier et al. 2016), and (2) investigate the provenance of the siliciclastic sedimentary rocks (e.g. Rotzien et al. 2018). Below we briefly summarise the results of these studies.

Previously published K–Ar ages have been determined on bulk hornblende separates from tuffs in the Mohakatino and Mount Messenger formations. The K–Ar ages (22.1–14.5 Ma) of the Mohakatino Formation are inconsistent with either or both stratigraphic and biostratigraphic constraints (Bergman et al. 1990; Hansen 1996). Likewise, two of the three K–Ar ages for the Mount Messenger Formation are older than the Tongaporutuan Stage (14.79 ± 0.37 and 12.93 ± 0.32 Ma), while the remaining K–Ar age falls within the Tongaporutuan (11.26 ± 0.28 Ma), neither this age nor any of the others are consistent with the relative stratigraphic positions of the sampled beds. The discrepancies between the K–Ar and stratigraphic ages are attributed to the presence of older, inherited hornblende, excess Ar, and/or Ar loss (Bergman et al. 1990; Hansen 1996).

Maier et al. (2016) reported Sensitive High-Resolution Ion Microprobe-Reverse Geometry (SHRIMP-RG) U–Pb zircon ages for tuffs from the Mohakatino, Mount Messenger and Urenui formations. The reported U–Pb zircon ages range from 10.63 ± 0.65 to 8.66 ± 0.86 Ma, consistent with the Tongaporutuan stratigraphic age range determined from foraminifera-based biostratigraphy and show a general younging-upwards trend (Figure 4 of Maier et al. 2016). However, the SHRIMP-RG U–Pb zircon ages are relatively imprecise, limiting their usefulness in providing absolute constraints for an age model of the northern Taranaki coastal section.

Figure 4. Representative cathodoluminescence images of zircons from (A) lower Mount Messenger Formation, Mohakatino River (P88711); (B) lower Mount Messenger Formation, Tongaporutu River (P88712); Urenui Formation, Waiiti Stream (P88716); and (D) Urenui Formation, Waiau Stream (P88717). Circles indicate the size and location of laser spot analyses. Labels dates in millions of years. Those <1500 Ma are ²⁰⁶Pb/²³⁸U dates, corrected for initial ²³⁰Th disequilibrium and/or common Pb using ²⁰⁷Pb, while those >1500 Ma are uncorrected ²⁰⁷Pb/²⁰⁶Pb dates. All uncertainties are 2σ.

Rotzien et al. (2018) undertook a detailed provenance study of the upper Mount Messenger Formation using a combination of mineralogical, geochemical, and LA–ICP–MS U–Pb detrital zircon geochronological data. They also compared their detrital zircon U–Pb data with those of the overlying Urenui Formation (Maier 2012). Although the detrital zircon dates provide important constraints on the provenance of the Mount Messenger and Urenui formations, none of the 949 zircon grains that were analysed have ages overlapping the depositional ages of the sampled rocks (Maier 2012; Rotzien et al. 2018). The ages of the youngest zircons from the Mount Messenger and Urenui formations are 37.2 ± 0.5 Ma (sample MM3) (Rotzien et al. 2018) and 18.0 ± 0.8 Ma (sample U-C-5) (Maier 2012). This indicates either that the Mohakatino Volcanic Centre was not a significant source area for sandstones of the Mount Messenger and Urenui formations, or the Mohakatino contribution was significantly diluted by the volume of siliciclastic sediment input.

Sampling and analytical methods

Sampling

A summary of the key details of each new sample are given in Table 1. Additional sample information can be found online in the GNS Science Petlab database (https://pet.gns.cri.nz/) (Strong et al. 2016). Eight tuff samples were collected from the Mohakatino, Mount Messenger and Urenui formations, between Mokau River in the north to Waiau Stream in the south, of which four were prepared for U–Pb zircon dating (Table 1; Figure 1B). The selected tuffs were collected from the lower Mount Messenger Formation at the Mohakatino River (Awaawanui Stream, adjacent to State Highway 3; P88711) and Tongaporutu River (south side of river; P88712). Those from the Urenui Formation were collected at Waiiti (P88716) and Waiau streams (P88717). Sample P88711 was selected because it is from an interval for which neither biostratigraphic nor U–Pb zircon data have previously been obtained, and together with sample P88712, brackets the interval between the Mohakatino and Tongaporutu rivers that includes the base Tukemokihi Coiling Zone (TCZ), and Otukehu MTD from which sample MM-C-12 of Maier et al. (2016) was collected (Figure 3). Sample P88716 was selected because it comes from the lower Urenui Formation for which Maier et al. (2016) obtained imprecise and dispersed ages (their samples U-C-10 to -27; Figure 3). Sample P88717 was analysed because it was collected from ∼400 m above the nearest sample for which existing age data are available (Figure 3).

Table 1. Details of new tuff samples, northern Taranaki.

Field number	Collection number	Locality	Lat. (°)	Long. (°)	Formation	Strat. height (m)	Lithology
MWS-17-03	P88711	Awaawanui Stream, SH3, tributary of Mohakatino River	−38.7301	174.6154	Lower Mount Messenger	120	Plagioclase–hornblende–lithic tuff
MWS-17-04	P88712	South bank of Tongaporutu River	−38.8168	174.5890	Upper Mount Messenger	284	Plagioclase–hornblende–lithic tuff
MWS-17-08	P88716	Sea cliffs south of Waiiti Stream mouth	−38.9252	174.4715	Urenui	1225	Plagioclase–hornblende–pyroxene–lithic tuff
MWS-17-09	P88717	Base of sea cliffs east of Waiau Stream mouth	−38.9939	174.3448	Urenui	1689	Plagioclase–hornblende–lithic tuff

Note: Stratigraphic heights are relative to base of the stratigraphic column (Figure 3). Coordinates are relative to the WGS84 (G1762) datum.

U–Pb zircon geochronology

Sample preparation and analysis generally followed the methods of Sagar and Palin (2011). Mudstone adhering to the tuffs was removed where possible using a hammer, chisel, and compressed-air scribe. The vast majority of mudstone was removed; however, it was neither practical nor possible to remove all of it. Contamination of the tuffs by zircons from the mudstone that overlap the eruption ages is considered highly unlikely because (1) none of the 949 detrital zircon grains from Mount Messenger and Urenui formation sandstones yielded late Miocene ages (Maier 2012; Rotzien et al. 2018), and (2) the mudstones contain only rare small zircons due to the high specific gravity of zircon (ρ = 4.65 g/cm³), compared with the clay minerals and silt-sized siliciclastic grains, that would be too small to analyse.

Laser ablation inductively-coupled plasma mass spectrometry (LA–ICP–MS) U–Th–Pb–trace element (TE) spot analyses of zircon were undertaken at the Otago Community Trust Centre for Trace Element Analysis, University of Otago. TEMORA 2 (Black et al. 2004) and NIST SRM610 (Jochum et al. 2011) were used to calibrate U–Th–Pb isotopic and TE data, respectively. Notable differences in the analytical methodology used here and that detailed by Sagar and Palin (2011), along with details of the data processing using Iolite 2.5 (Paton et al. 2011) and VizualAge 2015.6 (Petrus and Kamber 2012) are provided in Supplementary File 1. All dates <1500 Ma stated in the text are ²⁰⁶Pb/²³⁸U dates that have been corrected for common Pb (*) using the ²⁰⁷Pb method (Williams 1998) and Pb isotope evolution model of Stacey and Kramers (1975), while those >1500 Ma are uncorrected ²⁰⁷Pb/²⁰⁶Pb dates (Spencer et al. 2016). Uranium decay constants (Jaffey et al. 1971) and the ²³⁸U/²³⁵U ratio used are those recommended by Steiger and Jäger (1977). Miocene dates have additionally been corrected for initial ²³⁰Th-disequilibrium (**) following the method of Schärer (1984), using a Th/U_magma ratio of 3 ± 1, which covers the expected compositional range of mafic–intermediate arc volcanic rocks (e.g. Hawkesworth et al. 1997), and a ²³⁰Th decay constant of 9.1705 ± 0.0138 (Cheng et al. 2013). Isoplot 4.15 (Ludwig 2009) was used to calculate error-weighted mean ages and associated 2σ uncertainties that incorporate only analytical errors (i.e. no constant external uncertainties were propagated). Combined Gaussian-summation probability density function–histogram plots were also generated using Isoplot 4.15. Error-weighted mean ages were calculated using only ²⁰⁶Pb**/²³⁸U dates that are concordant within error (i.e. uncorrected ²⁰⁶Pb/²³⁸U and ²⁰⁷Pb/²³⁵U dates overlap within 2σ). Use of the terms zircon autocrysts, antecrysts and xenocrysts, and inherited zircon follows Miller et al. (2007). Detrital zircons are those that may be introduced into the tuffs during eruption, transport, and/or after deposition by bioturbation. Unless otherwise stated, all uncertainties for U–Pb dates and ages are given at the 2σ level.

Integrated age model

We have revised the age model for the northern Taranaki coastal section by integrating our new LA–ICP–MS U–Pb zircon ages with existing U–Pb zircon and biostratigraphic ages from Maier et al. (2016). For consistency with our new data, we have recalculated ²⁰⁷Pb- and ²³⁰Th-corrected ²⁰⁶Pb/²³⁸U zircon dates and error-weighted mean ages from the existing uncorrected data, using the constants and methods described above. These recalculated data are provided in Supplementary File 3 (SF3). Sample Q18/f43 was assigned an age of 8.91 Ma, which is the mean calculated from the range (8.85–8.96 Ma) reported by Maier et al. (2016). The biostratigraphic ages were conservatively assigned 2% uncertainties, because errors were neither estimated nor quantified previously. The stratigraphic positions of all samples are relative to the datum in the upper Mohakatino Formation used by Maier et al. (2016) (Figure 3) and were assigned absolute uncertainties of 5 m. Linear regressions of the available radiometric and biostratigraphic age data were calculated using the method of York (1969) in Isoplot 4.15, assuming that uncertainties are not correlated (i.e. error correlation coefficient, ρ = 0). Age model uncertainties are 2σ if the probability of fit (P) is >5%, or 95% confidence intervals if P < 5%. All quoted sedimentation rates are un-decompacted.

Results

Description of samples

A comprehensive description of all the samples collected for this study can be found in Sagar et al. (2018). Below is a brief description and classification of the samples, which follows the nomenclature of White and Houghton (2006). All the samples are crystal–lithic tuffs dominated by subhedral–euhedral plagioclase grains, which display oscillatory, normal and reverse zoning, and simple and/or multiple twinning. Pale-brown to medium-green pleochroic hornblende is the second-most abundant crystal type. Deep red-brown to yellow-brown pleochroic ‘oxyhornblende’ is present in some samples (e.g. P88708). Pyroxene is present in tuffs from the upper part of the lower Mount Messenger Formation (Waikiekie Stream area; samples P88713 to P88715, inclusive) and lower part of the Urenui Formation (Waiiti Stream; sample P88716). The most common lithic grains are plagioclase-phyric cryptocrystalline volcanic rocks fragments. All original glassy material in the matrix and lithic grains have been altered to orange–brown palagonite, clay minerals and/or carbonate.

U–Pb zircon geochronology

A summary of the ages assigned to the new and previously reported samples is given in Table 2. The full U–Th–Pb–TE zircon dataset is given in Supplementary File 1 (SF1). Below, the new U–Pb spot data are described in detail.

Table 2. Summary of the ages used in the new integrated age model for the northern Taranaki coastal section.

Sample	Strat. height (m)	Ref.	Type	Method	Age	± 2σ	n	MSWD	Mean spot 1σ (%)
R17/f498	15	Maier et al. (2016)	Biostrat.		10.65	0.21
R18/f65A	74	Maier et al. (2016)	Biostrat.		10.39	0.21
M-C-30	80	Maier et al. (2016)	U–Pb zircon	SHRIMP-RG	10.71	0.65	3	1.04	6.3
R18/f66	91	Maier et al. (2016)	Biostrat.		10.39	0.21
P88711	120	This study	U–Pb zircon	LA–ICP–MS	9.69	0.12	31	2.10	4.3
MM-C-12	265	Maier et al. (2016)	U–Pb zircon	SHRIMP-RG	9.76	0.28	25	1.16	9.2
Q18/f142	279	Maier et al. (2016)	Biostrat.		9.45	0.19
P88712	284	This study	U–Pb zircon	LA–ICP–MS	9.43	0.17	17	1.20	4.8
Q18/f168	387	Maier et al. (2016)	Biostrat.		9.44	0.19
Q18/f9583	438	Maier et al. (2016)	Biostrat.		9.42	0.19
Q18/f42	632	Maier et al. (2016)	Biostrat.		9.39	0.19
Q18/f1	1225	Maier et al. (2016)	Biostrat.		8.96	0.18
U-C-10	1225	Maier et al. (2016)	U–Pb zircon	SHRIMP-RG	8.52	0.87	2	0.71	7.6
Q18/f43	1251	Maier et al. (2016)	Biostrat.		8.91	0.18
U-C-17	1280	Maier et al. (2016)	U–Pb zircon	SHRIMP-RG	9.03	0.22	9	1.6	5.4
U-C-27	1290	Maier et al. (2016)	U–Pb zircon	SHRIMP-RG	8.83	0.51	6	1.7	6.5
P88717	1689	This study	U–Pb zircon	LA–ICP–MS	8.45	0.10	28	1.90	3.4

Note: Biostratigraphic and SHRIMP-RG U–Pb zircon ages are from Maier et al. (2016). All U–Pb zircon ages have been corrected for common Pb using ²⁰⁷Pb and initial ²³⁰Th-disequilibrium.

Mount Messenger Formation, Mohakatino River (Awaawanui Stream; P88711)

Zircon grains are subhedral to euhedral equant grains to elongate prisms. Length to width ratios (L:W) and lengths range from 1:1 to 5:1 and 100–500 µm, respectively, although L:W of 2:1–3:1 and lengths of 200–300 µm are more typical. Cathodoluminescence (CL) imaging reveals that the majority of grains consist of a variably resorbed interior (‘core’) with broad zoning mantled by an oscillatory-zoned exterior (‘rim’; Figure 4A). Both the cores and rims appear relatively bright in CL images.

Forty dates from the cores (n = 11) and rims (n = 29) of 29 crystals range from 8.44 ± 1.32–91.8 ± 3.9 Ma. The oldest date is from a crystal that appears relatively dark under CL and is clearly an inherited or detrital grain because it is approximately an order of magnitude greater than the Mount Messenger Formation stratigraphic age and, therefore, not further considered. The remaining dates are all late Miocene and form a broad peak centred at ∼9.7 Ma on a Gaussian-summation probability density function (PDF) plot (Figure 5A). Dates from the rim and core of the same crystal are indistinguishable within uncertainty (Figure 4A) and, hence, both were used to calculate the error-weighted mean age. Thirty-one concordant dates yield an error-weighted mean age of 9.69 ± 0.12 Ma (mean squared weighted deviation, MSWD (Wendt and Carl 1991) = 2.10; Figure 5A).

Figure 5. Frequency histogram–Gaussian-summation probability density function (PDF) plots of ²⁰⁶Pb**/²³⁸U zircon dates and error-weighted mean ages of tuffs analysed as part of this study. (A) Urenui Formation, Waiau Stream (P88717); (B) Mount Messenger Formation, Tongaporutu River (P88712); (C) Mount Messenger Formation, Mohakatino River (Awaawanui Stream; P88711). All ²⁰⁶Pb/²³⁸U dates have been corrected for both common Pb using ²⁰⁷Pb and initial ²³⁰Th disequilibrium (**); uncertainties are 2σ.

Mount Messenger Formation, Tongaporutu River (P88712)

Zircon crystals range from anhedral to euhedral and equant (L:W = 1:1) to elongate (L:W = 4:1). The most abundant crystals are equant and stubby, ∼100–300 µm-long prisms comprising relatively bright broad- or sector-zoned cores with or without thin oscillatory-zoned rims (Figure 4B). Relatively small (≤100 µm-long) dark oscillatory-zoned equant and stubby prisms are subordinate (Figure 4B).

The cores and rims of 26 crystals yielded 30 dates that range widely from 8.54 ± 0.87 to 687.6 ± 25.5 Ma. Six of the dates are Mesozoic or older and are all from the small, CL dark crystals (e.g. Figure 4B) and, therefore, interpreted to be from inherited or detrital crystals and were not considered further. The majority of dates (n/N = 24/30) are late Miocene, were obtained from CL bright crystals, and form a narrow unimodal peak on a PDF plot (Figure 5B). The dates from the rim and core of the same crystal are indistinguishable within uncertainty (Figure 4B). An error-weighted mean age of 9.43 ± 0.17 Ma (MSWD = 1.20) was calculated using the 17 concordant Miocene dates (Figure 5B).

Urenui Formation, Waiiti Stream (P88716)

Zircon crystals show a diverse range of sizes, morphologies and textures (Figure 4C). However, whole and fragmental crystals comprising cores with a range of textures and rims displaying oscillatory zoning are the most common. The CL of many crystals response is subdued compared with the other three samples (Figure 4A–D).

Twenty analyses on rims of 20 crystals yielded a large spread of dates from 109.5 ± 3.8–1984 ± 17 Ma, 16 of which are <10% discordant and can be divided into five broad peak groupings (Figure 6): (1) Early Cretaceous (∼112 Ma, n = 1), (2) Late Triassic–Middle Jurassic (∼231–166 Ma, n = 3); (3) Late Devonian–Early Permian (∼378–291 Ma, n = 6), (4) early Cambrian–Late Ordovician (∼517–445 Ma, n = 3), and (5) Pre-Cambrian (∼1977–688 Ma; n = 3; Figure 6). Given the late Miocene stratigraphic age of the Urenui Formation, all the dates are interpreted to be from inherited and/or detrital zircon.

Figure 6. Frequency histogram–PDF plot of zircon ages acquired for sample P88716, Urenui Formation, Waiiti Stream. Ages <1500 Ma are ²⁰⁶Pb/²³⁸U dates corrected for common Pb using ²⁰⁷Pb, while those >1500 Ma are uncorrected ²⁰⁷Pb/²⁰⁶Pb dates. All uncertainties are 2σ.

Urenui Formation, Waiau Stream (P88717)

The morphology, size and L:W of zircons are similar to those of sample P88711 (Figure 4A, D). Likewise, the majority of crystals show similar textures to those of P88711, with bright broad- or sector-zoned cores and oscillatory-zoned rims (Figure 4A, D).

Thirty-nine dates from 30 crystals (29 rims and 10 cores) are all late Miocene, ranging from 7.80 ± 0.41–9.28 ± 0.50 Ma. An error-weighted mean age of 8.45 ± 0.10 Ma (MSWD = 1.90) was calculated from 28 concordant dates (Figure 5C).

Integrated age model

The results of the linear regressions are given in Table 3. Visual assessment of a plot of stratigraphic position versus age suggests a significant change in gradient in the lower Mount Messenger Formation at ∼120 m (sample P88711; Figure 7). Regressing a straight line through the data for the lower (15–120 m) and upper (120–1689 m) parts of the section gives sedimentation rates of 102 ± 96 m/Ma (P = 0.04%, MSWD = 6.0) and 1358 ± 144 m/Ma (P = 20.6%, MSWD = 1.3). The high MSWD and low P for the lower part of the section reflects scatter in the data outside of that which can be accounted for by the assigned uncertainties or linear regression. However, the acceptable MSWD and higher P for the upper part of the section suggests that most of the variation in the data can be explained by the linear regression and uncertainties. Linear regression of our three new U–Pb zircon ages, which are for samples within the upper interval only (120–1689 m), yields a sedimentation rate of 1305 ± 154 m/Ma (P = 19.5%, MSWD = 1.68), indistinguishable within error of that calculated for the full radiometric–biostratigraphic dataset.

Figure 7. Stratigraphic height versus age of the northern Taranaki coastal section Mohakatino, Mount Messenger and Urenui formations. Linear un-decompacted sedimentation rates and uncertainties (2σ) were calculated using new and existing U–Pb zircon tuff ages and existing biostratigraphic ages (Maier et al. 2016) as described in the text. Existing U–Pb zircon ages recalculated as outlined in the methods section (data in Supp. File 3). All U–Pb zircon ages have been corrected for common Pb using ²⁰⁷Pb and initial ²³⁰Th-disequilibrium.

Table 3. Age data linear regression results, northern Taranaki coastal section.

Strat. Interval (m)	Lowest sample	Highest sample	n	Error type	Gradient (m/Ma)	± 2σ	y-intercept (m)	± 2σ	P (%)	MSWD
0−1689	R17/f498	P88717	17	95% CI	−1085	267	10854	2501	6.66E-19	8.8
0−120	R17/f498	P88711	5	95% CI	−102	96	1122	967	0.04	6.0
120−1689	P88711	P88717	13	2σ	−1358	144	13258	1316	20.6	1.32
*120−1689	P88711	P88717	3	2σ	−1305	154	12716	1396	19.5	1.68

Note: See text for explanation of methods and results. *only LA–ICP–MS U–Pb zircon ages used. P = probability of fit.

Discussion

U–Pb zircon geochronology: Eruption ages and detrital zircons

The U–Pb zircon ages decrease from the 9.69 ± 0.12 Ma at Mohakatino River, through 9.43 ± 0.17 Ma at Tongaporutu River, to 8.45 ± 0.10 Ma at Waiau Stream, consistent with the relative stratigraphic positions of the samples (Figure 3; Figure 5A–C). Our new ages are also consistent with the existing independently-determined biostratigraphic and SHRIMP-RG U–Pb zircon ages for the northern Taranaki coastal section (Table 2; Figure 7).

The crystal-rich nature of the dated tuffs suggests that they are dominated by juvenile pyroclastic material, although, differences in the degree of alteration, rounding and sorting between and within samples suggests the presence of some reworked material (Sagar et al. 2018). Therefore, most zircons in these tuffs are likely to be autocrystic. Accordingly, we consider our new U–Pb zircon ages, within their uncertainties (0.10–0.17 Ma), to record the timing of eruption and deposition of the tuffs and, therefore, constrain the absolute ages of the enclosing strata. This interpretation is supported by the good agreement between the new U–Pb zircon ages and biostratigraphic ages (Figure 7).

Sample P88716 did not contain any autocrystic zircons, indicating that the erupted magma was undersaturated with respect to and did not crystallise zircon and, therefore, would have dissolved any inherited grains (e.g. Miller et al. 2003). The zircon present in sample P88716 must, therefore, be detrital in origin, incorporated either during eruption or transport, or introduced by subsequent bioturbation. The latter is considered unlikely due to the paucity of zircon in the enclosing mudstone (see background and methods sections). Incorporation of detrital zircon from pre-existing Urenui Formation sediments and/or contemporaneous terrestrial siliciclastic sediment input during transport is considered most probable, because the P88716 zircon age spectrum includes Permian–Jurassic peaks that are common in the Eastern Province terranes (Ireland 1992; Wysoczanski et al. 1997; Cawood et al. 1999; Pickard et al. 2000; Adams et al. 2002a, 2002b; Wandres et al. 2004a, 2004b; Wandres et al. 2005; Adams et al. 2007; Adams et al. 2009a, 2009b), particularly the Torlesse and Waipapa composite terranes that were the major sediment sources for the Urenui and Mount Messenger formations (Maier 2012; Rotzien et al. 2018). For example, the age peaks at ∼222 and ∼231 Ma are common in the central North Island Waipapa Composite Terrane (Adams et al. 2009b). The presence of detrital zircons suggests deposition of this tuff via gravity flow processes, rather than direct settling through the water column to the slope floor. The 112 Ma zircon could have been derived from either the Pahau Terrane (Cawood et al. 1999; Pickard et al. 2000; Wandres et al. 2004a, 2009b), or Separation Point or Rahu suite plutons in the northern South Island (e.g. Muir et al. 1994).

New integrated age model for the northern Taranaki coastal section: Sedimentation rates and a possible unconformity

Our preferred age model consists of two intervals; a lower interval (0–120 m) characterised by a relatively low sedimentation rate, and an upper interval (120–1689 m) of relatively high linear sedimentation rate (Figure 7). The poor linear fit for the lower interval (MSWD = 6.0; Table 3; Figure 7), suggests that further subdivision into two sub-intervals – a lower sub-interval of relatively high sedimentation rate and an upper sub-interval characterised by a relatively low sedimentation rate – may be warranted. Alternatively, there may be an unconformity at the base of the Mount Messenger Formation or within the uppermost Mohakatino Formation (Figure 7; Maier et al. 2016). Age data for samples R18/f66 and P88711 indicates that an unconformity could represent a period of ∼0.7 ± 0.2 Ma (Figure 7). The sampling resolution is, unfortunately, too low to conclusively distinguish these possibilities.

In contrast, there is a good linear fit to the age data for the upper interval (MSWD = 1.32; Table 3; Figure 7), indicating that the 1569 m-thick interval of Mount Messenger and Urenui formations cropping out between Mohakatino River and Waiau Stream was deposited in 1.16 ± 0.01 Ma between 9.67 + 0.07/−0.05 and 8.52 + 0.06/−0.07 Ma, which equates to a mean linear (un-decompacted) sedimentation rate of 1358 ± 144 m/Ma.

Implications for the New Zealand Geological Timescale

The good linear fit (MSWD = 1.68) to our new U–Pb zircon ages (Table 3; Figure 7) affords the opportunity to independently assess the timing of three intra-Tongaporutuan bioevents in the northern Taranaki coastal section. These are the (1) base and (2) top of the Tukemokihi Coiling Zone (TCZ), and (3) the highest occurrence (HO) of planktic foraminifera Globoquadrina dihiscens. The TCZ is one of two intra-Tongaporutuan intervals defined by Globoconella miotumida with ≥20% dextral shells (Crundwell and Nelson 2007) and has lower and upper boundaries of 9.45 and 9.39 Ma, respectively, calibrated against the Geomagnetic Polarity Timescale (GPTS) of Ogg (2012) (Maier et al. 2016). The HO G. dihiscens marks the base of the informal late Tongaporutuan Stage and has a GPTS-calibrated (Ogg 2012) age of 8.96 Ma based on the age of Chron C4An.43 at Ocean Drilling Program (ODP) Site 1123 (Crundwell and Nelson 2007; Raine et al. 2015a, 2015b).

The base TCZ, top TCZ, and HO G. dihiscens occur at 279, 632, and 1225 m, respectively, in the northern Taranaki coastal section (Figure 7; Maier et al. 2016), corresponding to predicted ages of 9.53 + 0.07/−0.05, 9.26 + 0.03/−0.02, and 8.80 + 0.02/−0.03 Ma. None of the predicted ages overlap the assigned ages. The predicted ages of the top TCZ and HO G. dihiscens are significantly younger than their assigned ages by 0.10–016 and 0.13–0.19 Ma, respectively. In contrast, the base TCZ predicted age is greater than the assigned age by 0.02–0.14 Ma. However, the age range of Chron C4An.43 at ODP Site 1123 (9.105–8.771 Ma) overlaps the predicted age for the HO G. dihiscens, thus, providing independent evidence that the absolute age of this bioevent is accurate. This highlights the importance of propagating uncertainties into the absolute ages assigned to bioevents and suggests if this were done for the TCZ there may in fact be agreement between the predicted and assigned ages of this bioevent, too.

The discrepancy between the predicted and assigned base TCZ ages is smaller compared with those of the top TCZ and HO G. dihiscens. This difference is probably due to close bracketing of the base TCZ by samples P88711 (159 m below) and P88712 (4 m above). In contrast, the top TCZ and HO G. dihiscens are between ∼400 and ∼1000 m above or below the nearest control points (Table 2). Age predictions for all the intra-Tongaporutuan bioevents, particularly the top TCZ and HO G. dihiscens, could be improved through further development of the age model described in the next section, because despite the excellent linear fit to our U–Pb ages, it is unlikely that sedimentation proceeded at precisely the rate predicted by the U–Pb zircon ages (1305 ± 154 m/Ma) during deposition of the entire 1569 m-thick interval.

Future improvements to the new age model

There are several ways in which the new age model presented here could be improved. Firstly, uncertainties in the bioevent ages have been neither quantified nor estimated, and so for the purposes of this paper we conservatively assigned them 2% (2σ) errors. The bioevents were assigned ages based on the age model of ODP Site 1123, which is located on Chatham Rise, Pacific Ocean, presently >1000 km southeast of Taranaki Basin, recalibrated against the GPTS (Ogg 2012; Maier et al. 2016). One source of uncertainty in the biostratigraphic ages is that in the magnetostratigraphy of ODP Site 1123 from the GPTS. Uncertainties such as these could be quantitatively propagated from the ODP Site 1123 age model to the biostratigraphic ages. Secondly, the sedimentation rates and possible unconformity in the lower part of the section could be constrained through higher resolution sampling within the uppermost Mohakatino Formation and basal Mount Messenger Formation. Next, the ages of tuffs dated by the SHRIMP-RG method (Maier et al. 2016) are imprecise (Table 2; Figure 7) and could be improved by re-dating the same horizons whilst ensuring higher analytical precision and analysing greater numbers of zircon grains per sample. Dating of tuffs using the ⁴⁰Ar/³⁹Ar method could resolve whether excess Ar and/or Ar loss is responsible for the discrepant K–Ar ages. ⁴⁰Ar/³⁹Ar analyses of several single hornblende crystals from each tuff using a modern high-precision multi-collector noble gas mass spectrometer (e.g. Thermo Scientific ARGUS VI) would resolve inherited and/or detrital hornblende age components, thus having the potential to yield precise and accurate eruption ages. However, ⁴⁰Ar/³⁹Ar analyses are costly and time consuming (due to the need for sample irradiation) compared with U–Pb zircon geochronology using the SHRIMP and, particularly, the LA–ICP–MS methods. Finally, further refinement of the sedimentation rate and bioevent age predictions in the upper part of the section could be achieved by dating of additional tuffs from the upper Mount Messenger Formation from Waikiekie Stream to Pariokariwa Point (∼500–1000 m), lower Urenui Formation between Pariokariwa Point and Waiiti Stream (∼1000–1200 m), and upper Urenui Formation from Mimi River to Onaero River (∼1300–1500 m; Figures 3 and 7). The latter two coastal sections are, however, difficult to access other than from offshore. Mass transports deposits (e.g. sample MM-C-12) should be avoided when collecting any new samples for dating, because the difference between depositional ages and MTD emplacement ages is unconstrained.

Conclusions

Our new LA–ICP–MS U–Pb zircon ages constrain the depositional ages of the Mount Messenger Formation to 9.69 ± 0.12 Ma at Mohakatino River and 9.43 ± 0.17 Ma at Tongaportu River, and Urenui Formation near Waiau Stream to 8.45 ± 0.10 Ma. Our preferred age model for the northern Taranaki coastal section, developed by integrating our new ages with existing biostratigraphic and SHRIMP U–Pb zircon ages, comprises lower and upper intervals characterised by relatively low and high un-decompacted sedimentation rates of 102 ± 96 m/Ma (MSWD = 6.0) and 1358 ± 144 m/Ma (MSWD = 1.32), respectively. The poor linear fit to the lower interval could indicate either variable sedimentation rates, or the presence of an unconformity in the upper Mohakatino Formation or at the base of the overlying Mount Messenger Formation. The duration of this unconformity is constrained to 0.7 ± 0.2 Ma by a biostratigraphic age (10.39 Ma; R18/f66) and one of our new U–Pb zircon ages (9.69 ± 0.12 Ma; P88711). The good linear fit to the upper interval indicates that the Mount Messenger and Urenui formations between Mohakatino River and Waiau Stream were deposited in 1.16 ± 0.01 Ma between 9.67 + 0.07/−0.05 and 8.52 + 0.06/−0.07 Ma. The absence of autocrystic zircons and presence of Eastern Province detrital zircons in sample P88716 (Urenui Formation, Waiiti Stream) suggests that this tuff was deposited by gravity flow processes.

An age model for the upper interval based exclusively on our three U–Pb zircon ages allows independent evaluation of the timing of three intra-Tongaporutuan bioevents. The base TCZ (279 m), top TCZ (632 m), and HO G. dihiscens (1225 m) have predicted ages of 9.53 + 0.07/−0.05, 9.26 + 0.03/0.02, and 8.80 + 0.02/−0.03 Ma, respectively, none of which overlap their assigned ages. However, the age range of Chron C4An.43 at ODP Site 1123 (9.105–8.771 Ma) overlaps the predicted age for the HO G. dihiscens, providing independent corroboration of the absolute age of this bioevent. This highlights the importance of propagating uncertainties into the absolute ages assigned to bioevents and the opportunities for future improvements to the age model by using high-precision geochronological methods to date additional tuffs and previously-dated tuffs whose existing SHRIMP U–Pb zircon ages have large uncertainties.

Acknowledgements

Discussions with Chris Adams (GNS Science) and J. Michael Palin (University of Otago) helped in the writing of this manuscript. David Flynn (Victoria University of Wellington) assisted with CL imaging. U–Th–Pb–TE zircon data acquisition would not have been possible without the expert help of Malcolm Reid (University of Otago). Rose Turnbull (GNS Science) is thanked for assistance with U–Th–Pb–TE zircon data acquisition. This research is part of the GNS Science Sedimentary Basin Research programme, funded from the Strategic Science Investment Fund (SSIF) administered by the New Zealand Government’s Ministry of Business, Innovation and Employment (MBIE). Kyle Bland and Sarah Milicich (GNS Science) are thanked for providing timely and insightful internal reviews of earlier versions of the manuscript. Comments from Associate Editor Richard Wysoczanski, Andy Tulloch and an anonymous reviewer greatly helped us to further refine the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Matthew W. Sagar http://orcid.org/0000-0001-8178-1619

Greg H. Browne http://orcid.org/0000-0003-0865-6146

Malcolm J. Arnot http://orcid.org/0000-0003-1152-7417

Diane Seward http://orcid.org/0000-0002-6465-4084

Dominic P. Strogen http://orcid.org/0000-0002-1744-1289

Additional information

Funding

This work was supported by Ministry of Business, Innovation and Employment: [Grant Number Strategic Science Investment Fund (SSIF)].