June 2023 LIP of the Month

Paleogeography and high-precision geochronology of the Neoarchean Fortescue Group, Pilbara, Western Australia

Jennifer Kasbohm* ^a,1, Blair Schoene ^a, Scott A. Maclennan ^a,2, David A.D. Evans ^b, Benjamin P. Weiss ^c

^a Department of Geosciences, Princeton University, Princeton, NJ 08544, USA

^b Department of Earth & Planetary Sciences, Yale University, New Haven, CT 06511, USA

^c Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

¹ Present address: Department of Earth & Planetary Sciences, Yale University, New Haven, CT 06511, USA

² Present address: School of Geosciences, University of the Witwatersrand, Johannesburg, South Africa

* Corresponding author, jennifer.kasbohm@yale.edu

This is a summary and modification of: Kasbohm, J., Schoene, B., Maclennan, S., Evans, D.A.D., Weiss, B.P. (2023). Paleogeography and high-precision geochronology of the Neoarchean Fortescue Group, Pilbara, Western Australia. Precambrian Research 394, 107114. doi.org/10.1016/j.precamres.2023.107114

Introduction

The acceptance of plate tectonics irreversibly changed the way geoscientists understand Earth’s mantle dynamics and lithospheric movements. However, a point of contention in the study of Earth history is whether plate tectonic processes were operating during the Archean (Brown et al., 2020), and if so, whether they were fundamentally different from those of the present. Did a hotter Earth allow for more vigorous convection and rapid plate motion (Davies, 1992), or were tectonics slowed by dehydration and thickening of the mantle lithosphere (Korenaga, 2003)? Furthermore, determining the timing and frequency of magnetic field reversals during the Archean is hampered not only by the paucity of rocks from this eon, but also by the required retention of a primary magnetic signature, the ability to obtain age constraints, and the preservation of somewhat continuous sequences that document multiple polarity intervals, leading to only sparse documentation of these events (Gallet et al., 2012; Hulot et al., 2010, and references therein).

A paleomagnetic study integrated with high-precision geochronological data (e.g., Swanson-Hysell et al., 2019) has the potential to quantify rates of plate motion during the Archean. Collecting both paleomagnetic and geochronological samples within a detailed stratigraphic context may provide a continuous time series for changes in paleogeography, and removes the need for regional correlation. Continental flood basalts are ideal targets for this type of study because basalt is a faithful paleomagnetic recorder that retains a record of paleohorizontal and is erupted in a layered stratigraphic manner.

The ca. 2.8-2.7 Ga Fortescue Group of the Pilbara craton in Western Australia (Figure 1) has been recognized as one of the oldest and best-preserved Archean flood basalt successions hypothesized to be sourced from a continental rift (Blake, 1993), with an estimated basaltic volume of 250,000 km³ (Thorne & Trendall, 2001). The Fortescue succession has the potential to yield insights not only into the rate of cratonic motion occurring in the Archean, but also into how these processes may have affected evolving life on Earth prior to the Great Oxidation Event. The Fortescue Group has been subject to prior lithological (Thorne & Trendall, 2001), geochronological (Blake et al., 2004), and paleomagnetic study (Strik, 2004; Strik et al., 2003) that suggest its suitability for a detailed stratigraphic approach (Figure 1b).

Blake (2001) divided the Fortescue Group into 12 unconformity-bounded packages, which were later dated with SHRIMP U-Pb zircon geochronology (Blake et al., 2004). Guided by Blake’s stratigraphic framework, Strik et al. (2003) published the first detailed Fortescue Group paleomagnetic study, which yielded an apparent polar wander path for the ca. 60 Myr depositional history of the group and the earliest stratigraphically documented reversal in Earth’s geomagnetic field. There are two intervals in the Fortescue Group where the Pilbara craton appears to have been moving rapidly (Strik, 2004; Strik et al., 2003). However, analytical errors on the order of millions of years for the current ages for the Fortescue Group inhibit the calculation of Pilbara drift rates in the short time span of these potentially rapid intervals. Additionally, the difficulty in correlating rocks from distant regions of the Fortescue Group (Figure 1) precludes the possibility of integrating the current paleomagnetic and geochronological datasets.

Presented here are the results of an integrated stratigraphic, paleomagnetic, and geochronological study aiming to quantify the minimum velocity of the Pilbara during these two intervals of potentially rapid motion. We present six new or updated high-quality paleomagnetic poles for the Fortescue Group. With four new high-precision U-Pb CA-ID-TIMS ages, we provide improved velocity constraints that exceed both modern and Mesoarchean drift rates. We show that the Tumbiana stromatolite colony developed 2721.23±0.88/0.88/6.9 million years ago as the Pilbara craton drifted from 51.5±7.0° to 32.1±5.7° paleolatitude. Finally, we revisit the classification of the Fortescue Group as a large igneous province, and comment on its history of magnetic field reversals. Whether or not the Fortescue Group meets the criteria for classification as a large igneous province, it can provide numerous insights into the tectonics, magmatism, and state of the geodynamo of the Archean Earth system.

Methods

Two field seasons in summers 2013 and 2014 were undertaken in the Fortescue Group, in which stratigraphic sections were measured and sampled at four different localities. Two sections were measured at dm-scale across the Package 0–1 boundary at Glen Herring Gorge (GHG; 900 m) and Coongan River (CR; 900 m) (Figure 2a). For Packages 7–8, two sections were measured on opposite sides of the Meentheena Centrocline (Figure 2b); the Meentheena Centrocline North (MCN; 900 m) and the Meentheena Centrocline South (MCS; 600 m) sections. Unit thicknesses and lithologies, as well as the stratigraphic positions of paleomagnetism sample sites and geochronology samples, were recorded in each section (Figure 3).

Paleomagnetism

Eight hundred forty-six paleomagnetic cores were drilled from 75 sites, and oriented by magnetic and solar compasses. Samples underwent thermal demagnetization and were analyzed at the MIT Paleomagnetism Laboratory using a cryogenic DC-SQuID magnetometer (sensitivity with sample holder ∼10⁻⁹ Am²) with automated sample changer (Kirschvink et al., 2008). Each sample’s magnetic components were resolved with principal component analysis (Kirschvink, 1980), using software created by Jones (2002) — least-squares fits were applied to successive demagnetization steps that represented the characteristic remanent magnetization (ChRM) of each sample. Mean declination (D), inclination (I), precision parameters (k), and 95% confidence limits (α95) were calculated to find the average ChRM of each sampling site using Fisher (1953) statistics.

Geochronology

Twenty-one interflow units were logged and collected as potential geochronology samples in all Package 0-1 and 7-8 stratigraphic sections, and from the Package 2 Spinaway Porphyry. Four samples yielded zircons of Neoarchean age (~2.7 Ga). The stratigraphic positions of each geochronology sample are shown with squares on Figure 3. Colored squares represent samples that yielded zircon ages and are discussed below; greyed-out squares represent samples that either did not yield zircon or yielded only detrital or inherited grains. Geochronology methods are as described in Kasbohm & Schoene (2018). Concordia plots with some zircon images from each sample are provided in Figure 4.

Results & Discussion

New high-quality Archean paleomagnetic poles for the Pilbara Craton

By yielding new paleomagnetic data and high-precision ages, our study updates the position and ages of five poles and provides one new pole that we propose for inclusion in databases of reliable paleomagnetic poles for Precambrian tectonic reconstructions (Evans et al., 2021). In Table 1, poles are graded based on the quality criteria of Van der Voo (1990) and the reliability criteria of Meert et al. (2020). We follow Evans et al. (2021) in suggesting that poles receive an A grade when they possess an age constrained within 40 Ma (Q1, not necessarily R1), a sufficient number and statistical quality of poles (Q2, not necessarily R2), demagnetization and principal component analysis (Q3), field stability tests (Q4, R4), and structural control (Q5, R5). Poles with B grade lack one or more of these criteria, and we highlight further work that would be needed for promotion to A grade.

Rapid Neoarchean Pilbara plate motion?

There is a significant change in the paleogeography of the Pilbara craton during emplacement of the Package 0-1 Mount Roe Basalt flows. Between the lower and upper Package 0 directions, there is 25.5±15.3° (2835±1697 km) of displacement between the VGPs, and between Upper Package 0 and Package 1, there is 54.8±15.4° (6092±1716 km) of displacement between the VGPs (Figure 5). Since all zircons dated from GHG and CR were inherited, a plate velocity across these two transitions cannot be calculated at this time.

The Pilbara craton moved from a paleolatitude of 50.8±6.7° during Lower Package 0 emplacement, to 68.2±15° for the BRDS (Black Range Dyke Suite) and 76.3±13.7° during Upper Package 0, and to 48.9±7.1° during Package 1. We can attempt to constrain a minimum rate of the second portion of this motion, between the BRDS and Grand Mean Package 1, using the Wingate (1999) age of 2772±2 Ma for BRDS and our new high-precision Package 2 age of the Spinaway Porphyry of 2762.43±0.58 Ma, the latter used as a minimum conservative age estimate for Grand Mean Package 1. Following the approach of Swanson-Hysell et al. (2014), we use a Monte Carlo simulation to calculate a minimum rate of plate motion 23±20 cm/a (Figure 6a and b). This interpretation, which does not depend upon the craton crossing the polar circle, allows for a combination of local vertical axis rotation as well as translation. However, our model reveals that explaining the offset between poles through rotation alone (requiring negligible translation) is implausible, as that model would fall within the extreme lower tail of our confidence interval. We also note that 23±20 cm/yr (and other rates described below) are minimum velocity estimates, as paleolongitude is not constrained in these calculations. Therefore, the rate of Pilbara plate motion we calculate between the BRDS and Grand Mean Package 1 poles is at least comparable if not more rapid than the fastest plate motion observed on Earth today, of ~20 cm/a (Zahirovic et al., 2015).

In contrast to the hypothesis of vertical axis rotation described above, we can also calculate a rate for Pilbara plate motion that may have been dominated by translation (Figure 6c and d). Evans et al. (2017) interpreted the sequence of VGPs from Package 0, Upper Package 0/BRDS, and Package 1 as translational plate motion of the Pilbara craton across the polar circle, rather than the block rotation advocated by Strik (2004). To assess this interpretation, we therefore update the Monte Carlo estimation described above to force cratonic motion across the pole. When we calculate the rate of motion between our Upper Package 0 (with the Wingate (1999) age of the overlapping BRDS pole) and Grand Mean Package 1 poles (with the same Package 2 age constraint as above) this interpretation yields a minimum translational rate of 64±23 cm/a (Figure 5). We also calculate this rate using individual BRDS dyke ages obtained by Wingate (1999) and estimate a rate 51.5±25.8 cm/a when the oldest dyke age of 2774.7±4.6 Ma is applied to Upper Package 0, and a rate of ~32 meters/a when the youngest dyke age of 2768.0±5.6 Ma is applied. The latter rate is extraordinarily fast because the lower precision SHRIMP age nearly overlaps with our TIMS age for Package 2. Whether the oldest individual BRDS age or the BRDS weighted mean age of 2772±2 is applied to our Upper Package 1 pole, the rate of Pilbara plate motion we calculate over this interval far exceeds any plate motion observed on Earth today.

Between VGPs for Package 7 basalt and Package 8, we calculate a displacement of 19.4±9.0° (2157±1004 km; Figure 5), which falls between the initial estimate of Strik et al. (2003) of 27.2° of movement across this interval and a revised estimate by Strik (2004) of a 14.4° shift. Our only geochronology sample from Package 7, K295, overlies the Package 7 basalt paleomagnetic site and overlaps in age with our results from Package 8, and thus does not provide a meaningful constraint on the plate velocity for this interval. Even though none of the lapilli tuffs we sampled at MCN or MCS in the lowermost volcaniclastic unit of Package 7 yielded zircons, it may be fruitful to sample this lithology in other regions with the goal of obtaining an improved age constraint for Package 7 and an updated plate velocity estimate for this interval.

Our calculated minimum velocities calculated for the Pilbara, as it drifted and/or rotated across the polar circle from the Upper Package 0 to Package 1 position, are noteworthy not only because they exceed modern plate tectonic rates, but also because they are far greater than the average plate velocities calculated for the Pilbara craton during the Mesoarchean (Brenner et al., 2020). For that older interval, minimum velocities are calculated using only three poles with ages ca. 3350, 3180, and 2800 Ma. We suggest that such long-term averages could be overlooking shorter intervals of more rapid motion. While any combination of a hotter Archean Earth with more rapid mantle convection, or more rapid true polar wander, have been invoked previously as possible explanations for rapid drift rates, our results show that Archean plate velocities were not uniformly fast or decreasing monotonically as time progressed.

Is the Fortescue Group a Large Igneous Province?

With an estimated basaltic volume of 250,000 km³ (Thorne & Trendall, 2001), the Fortescue Group invites comparison to other continental flood basalt provinces, such as the Columbia River Basalt Group or the Deccan Traps. Indeed, the basaltic volume of the Fortescue Group is 20% greater than that of the Columbia River Basalt Group (Reidel, 2015). However, Thorne & Trendall (2001) object to the classification of the Fortescue Group as a continental flood basalt due to its paleoenvironment and duration of its emplacement. While the northern exposures of the Fortescue Group suggest predominantly subaerial, ‘continental’ emplacement, the basalts of the southern Fortescue Group are predominantly subaqueous, as if they erupted in a passive continental margin. Also, the Fortescue Group’s emplacement duration of ~60 Ma is far longer than that of any other flood basalt described (Thorne & Trendall, 2001).

In recent decades, large igneous provinces (LIPs) have been defined more broadly to include a wider array of geologic expressions of voluminous magmatism, such as oceanic LIPs and silicic LIPs. Therefore, the subaqueous emplacement and more felsic geochemistry of the Fortescue basalts would not preclude the classification of the Fortescue as a large igneous province. However, the 60 Ma duration of Fortescue emplacement would be its most notable disqualifying factor, as LIPs should be emplaced within a short duration (<5 Ma), or with multiple short pulses over a maximum of a few 10s of Ma (Ernst & Youbi, 2017). A recent review of high-precision geochronology of large igneous provinces finds that thus far, all well-dated LIPs were emplaced in <1 Ma (Kasbohm et al., 2021).

Since the Fortescue Group contains three distinct basaltic formations (Mount Roe, Kylena, and Maddina Basalts), perhaps each of these could be considered an individual LIP (e.g., Ernst et al. 2021). Volume estimates for these units are 72,000 km³, 68,000 km³, and 110,000 km³ respectively (Thorne & Trendall, 2001), though these may be hampered by limited geologic preservation. While the Mount Roe and Kylena Basalts do not satisfy the >100,000 km³ volume cutoff for LIP classification, the Maddina Basalt (Packages 8-10) could qualify as a continental flood basalt with its larger volume. Our geochronology shows a minimum duration of 0.98±1.03 Ma for the emplacement of Package 8 at MCN, suggesting that the rest of the Maddina Basalt could have been emplaced in a few Ma or less.

While the Fortescue Group may or may not meet specific criteria for LIP classification, assessing its emplacement dynamics may yield insights for an early prototype of LIP magmatism. One aspect of more recent LIP emplacement that is not always well-known is the relative timing of intrusive and extrusive magmatism (Kasbohm et al., 2021). By contrast, our paleomagnetic correlation showing the temporal coincidence of the Mount Roe Basalt flows of Upper Package 0 and the intrusive activity of the BRDS conclusively places the widespread BRDS within the temporal and stratigraphic context of the Fortescue Group, and allows for a temporal correlation between dykes and lava flows.

Another crucial aspect of LIP emplacement to consider is the duration of hiatuses. The Package 1-10 subdivisions of Blake (2001) were initially envisioned to be bounded by significant time breaks in the Fortescue stratigraphy. Our geochronology sampling at the boundary between Packages 7 and 8 yield the first high-precision temporal constraints on one of these hiatuses, since K295 is in the uppermost Package 7, and K302 overlies the Maddina Basalts in Package 8. While Blake (2001) noted the Package 7–8 boundary was likely relatively short (1-5 Ma), and subsequent geochronology produced a duration of ~3 Ma (Blake et al., 2004), we show that the time elapsed between these two samples (whose dates overlap within uncertainty) is 0.98±1.03 Ma, suggesting that the actual amount of time elapsed before Package 8 volcanism started is much less. Therefore, it seems that time breaks between these packages is less significant than previously thought, and we cannot preclude the possibility that the contact between Packages 7 and 8 may in fact be conformable.

Neoarchean magnetic field reversals

Our results from the Fortescue Group yields insights into the timing of magnetic field reversals and duration of polarity chrons during the Archean. Strik (2003) documented Earth’s oldest known magnetic field reversal between Packages 1 and 2, which we now constrain with a minimum age of 2762.11±0.66 Ma on the Spinaway Porphyry, and the previously existing maximum age of 2772±2 Ma from the BRDS (Evans et al., 2017; Wingate, 1999).

While polarity reverts to normal after Package 2, the reversely magnetized basalt flow in the Boodalyeri Creek Area sampled by Strik (2004) suggest that an additional reversed polarity interval occurred during Package 7, in the Tumbiana Formation, at ~2722 Ma. The reversely magnetized Package 7 volcaniclastic sediments we sample at the base of MCN and MCS may also may document this interval, if their directions are interpreted to be primary. Normal polarity returns at or before 2721.57±0.64, based on our zircon age from the Tumbiana stromatolites overlying the normally magnetized basalts. These data suggest that Earth’s polarity reversed at least four times between 2772 and 2721 Ma. The durations of each polarity chron defined would potentially be ~10 Ma during the Package 2 reversed interval, ~40 Ma during the normal chron when Packages 3-6 were deposited, and ~1 Ma for the reversed interval in Package 7. The pattern of the long normal chron (during the deposition of Packages 3-6) followed by the brief reversed interval in Package 7 reflects patterns in other thick and densely sampled Precambrian successions, where long intervals (30-40 Ma) of uniform polarity are juxtaposed with intervals of rapid reversals (Elston et al., 2002; Gallet et al., 2012; Pavlov & Gallet, 2010). This alternation between reversing and non-reversing regimes may reflect the sensitivity of the geodynamo to changing heat flux patterns at the core-mantle boundary, in a time prior to the crystallization of the inner core (Gallet et al., 2012).

Conclusions

By using a stratigraphic approach to integrate paleomagnetic and high-precision geochronological data from the Fortescue Group, we provide crucial new constraints on plate velocities, large igneous province magmatism, and magnetic field reversals in the Neoarchean. We present six new high-quality paleomagnetic poles and four high-precision U-Pb ages that show that the Pilbara craton drifted at a minimum rate of 23±20 to 64±23 cm/a (depending on the extent of cratonic rotation versus translation) over an interval of ~10 million years in the Neoarchean. Both rates far exceed both background drift rates for the craton, and modern rates of plate motion. Our new age of 2721.23±0.88/0.88/6.9 Ma for the Tumbiana Formation provides the first high-precision U-Pb zircon age constraint interbedded within a colony of Archean stromatolites. The new constraints described here on plate tectonic rates, timing of magmatism, and magnetic field reversals show how an early prototype of a large igneous province may enlighten future investigations of its successors.

References