Electronic Supplement to
A 2300-Year Paleoearthquake Record of the Southern Alpine Fault and Fiordland Subduction Zone, New Zealand, Based on Stacked Turbidites

by Philip M. Barnes, Helen C. Bostock, Helen L. Neil, Lorna J. Strachan, and Matthew Gosling

We provide here supporting details of the laboratory methods and analytical techniques that we applied to sedimentary cores. This includes use of a Geotek multi-sensor core logger, x-ray radiography, particle size analysis, carbonate content, mineralogy and foraminifera composition, AMS radiocarbon dating, and ²¹⁰Pb and ²²⁶Ra activities.

We also provide supporting material in five figures with associated captions. Fig. S1 provides photographic evidence of former glacial age and Holocene seafloor substrates on the Fiordland margin of New Zealand. Locations are provided in Figs 3 and 4 of the main article. Fig. S2 presents the computer codes we ran in OxCal 4.1 Sequence to develop age models of three cores, T49, T18 and T14. Figs. S3 to S5 present the OxCal Sequence multi-plots for each of the three cores.

Finally, Fig. S6 and associated caption, and Table S1 are presented in support of sediment transport analysis under extreme storm waves and currents on the Fiorldand coast.

---

Laboratory and analytical methods applied to sediment cores

The multi-sensor logging was undertaken at the University of Otago in 2008, some 4 years after the collection of the cores. Consequently, the cores were characterised by some loss of moisture and bedding-parallel desiccation cracking. Whilst these aspects effected some absolute physical properties down individual cores (i.e., density, p-wave velocity), they did not unduly affect the relative changes in physical properties. The MSCL was calibrated against aluminium core for gamma ray measurements. The split cores were X-rayed at NIWA using an Ultra EPX-F2800 portable veterinary generator with a Varian PaxScan 4030E digital imaging system, with exposures performed at 100 kV and 0.25 ms duration. The cores were sampled for grainsize analyses at 10 mm intervals down core, decreasing to 5 mm intervals through sand layers. Discrete sediment samples were analysed using a Beckman Coulter LS13320 laser particle sizer. Some of the samples were also analysed for mineralogy at the University of Auckland using the X-Ray Diffraction system. Carbonate content was determined at NIWA using the carbonate bomb method adapted from Jones and Kaiteris (1983) with an accuracy of ± 2%.

The >150 μm fraction from samples throughout the cores were visually examined under a binocular microscope to identify benthic foraminifera, rock fragments, organic matter and other sedimentary particles. Following the methods outlined by Hayward et al. (1999, 2010), key benthic foraminifera were identified by depth range to ascertain the depth of the source sediment within sand and silt layers including: (1) coastal to shallow shelf depths, 0-100 m (Bulimina marginata, Pileolina radiata, Miliolinella subrotundata, Elphidium novozealandica, Rosalina bradyi, Zeaflorilus parri); (2) deeper shelf depths, >120 m (Bulimina striata); and (3) slope to deep waters, >300 m (Bolivinita quadrilatera, Cibicides wuellerstorfi, Karrieriella bradyi, Martinottiella communis).

Samples, typically ~10 mm thick, were taken from immediately below all sand horizons for radiocarbon dating. In five selected sand-silt graded beds we also dated additional samples from lower stratigraphic positions within silt intervals. AMS radiocarbon ages were analysed on mixed planktonic foraminifera that were picked from the >150 mm fraction. The radiocarbon ages were calibrated in CALIB 6.0 (Stuiver and Reimer, 1993; Stuiver et al., 2005) using the Marine09 database (Reimer et al., 2009) and a local reservoir age (ΔR) of -38 ± 30 (Stuiver and Braziunas, 1993) (Table 1). All uncertainties are presented at 2 standard deviations. In addition, shell material and foraminifera were dated from widely distributed core sites across the margin (Fig. 3, Table 1) to evaluate the distribution and rates of post-glacial sedimentation.

²¹⁰Pb and ²²⁶Ra activities were determined by gamma spectroscopy at East Carolina University (J.P Walsh, pers. comm.), on samples from the top 150 mm of each of the three targeted cores to determine if modern sediment (c. <150 yrs) was present in the uppermost deposits (Table 3). Samples were initially dried at 60^oC, homogenized, packed and sealed before counting for at least 24 hours. Gamma counting was conducted on low-background Germanium detectors coupled with a multi-channel analyzer. Detectors were calibrated using natural matrix standards (IAEA-300, 314, 315) at each energy level of interest in the standard counting geometry for the associated detector, and activities were corrected for self-adsorption using a direct transmission method (Cutshall et al., 1983; Cable et al., 2001). Excess ²¹⁰Pb activities were determined by subtracting total ²¹⁰Pb (46.5 kev) from that supported by ²²⁶Ra. ²²⁶Ra activities were determined by allowing samples to equilibrate for greater than 3 weeks prior to counting. ²²⁶Ra were then determined indirectly by counting the gamma emissions of its grand daughters, ²¹⁴Pb (295 and 351 keV) and ²¹⁴Bi (609 keV).

---

Supporting figures and captions: Sediments photographs and OxCal sequences

Figure S1. Camera photographs of various seafloor substrates (a, b, c, e and f), and core from the surface of glacial outwash fan (d). (a) Boulder gravel substrate exposed on the crest of Bligh Ridge (camera transect T70, see Fig. 4f in the article), interpreted as relict (former glaciation) ice-carved debris. (b) Thin sediment veneer partially covering boulder gravel on the landward flank of Bligh Ridge (camera transect T70). (c) Relict gravel substrate, dated ~16 cal. ka, on the Caswell fan glacial outwash surface (Tan0405#31, Fig. 3) (Barnes, 2009). (d) Sediment core of relict surface gravel in (c) (Tan0405#41, Fig. 3). (e) Photograph of sandy sediment body on the upper slope, representative of the source region from which basin turbidites were triggered. White feature on carbonate mound is a filter feeding hydrocoral Calyptopore veticulata. (f) Burrowed muddy sediment in channel axis below Looking Glass Basin (Tan0405#26, Fig. 3).

Figure S2. Codes developed and run in OxCal 4.1 Sequence (Bronk Ramsey, 2008) for the construction of age models for cores T49, T18 and T14, collected in 2004 from offshore Fiordland, New Zealand. The models incorporate radiocarbon dates (R-dates) from foraminifera and excess ²¹⁰Pb data. The turbidite event numbers, referred to as S1 to S10, L1 to L15, and G1 to G8, correspond to the turbidite beds interpreted in Figures 5 to 7.

Figure S3. Ages of turbidite emplacement events in sedimentary core T14 from George Basin, offshore Fiordland, modelled using OxCal 4.1 Sequence (Bronk Ramsey, 2008). See Figure S2 for OxCal Sequence input code, Table 1 for raw radiocarbon dates, and Table 3 for OxCal output ages. The turbidite event numbers, referred to as G1 to G8, correspond to the turbidite beds interpreted in Figure 7. The lightly shaded probability distributions visible are derived from the input radiocarbon ages, whereas dark colour distributions are model outputs. The crosses are the median ages, whilst the horizontal bar is the error range at 95% confidence.

Figure S4. Ages of turbidite emplacement events in sedimentary core T18 from Looking Glass Basin, offshore Fiordland, modelled using OxCal 4.1 Sequence (Bronk Ramsey, 2008). See Figure S2 for OxCal Sequence input code, Table 1 for raw radiocarbon dates, and Table 3 for OxCal output ages. The turbidite event numbers, referred to as L1 to L15, correspond to the turbidite beds interpreted in Figure 6. The lightly shaded probability distributions visible are derived from the input radiocarbon ages, whereas dark colour distributions are model outputs. The crosses are the median ages, whilst the horizontal bar is the error range at 95% confidence.

Figure S5. Ages of turbidite emplacement events in sedimentary core T49 from Secretary Basin, offshore Fiordland, modelled using Oxcal 4.1 Sequence (Bronk Ramsey, 2008). See Figure S2 for OxCal Sequence input code, Table 1 for raw radiocarbon dates, and Table 3 for OxCal output ages. The turbidite event numbers, referred to as S1 to S10, correspond to the turbidite beds interpreted in Figure 5. The lightly shaded probability distributions visible are derived from the input radiocarbon ages, whereas dark colour distributions are model outputs. The crosses are the median ages, whilst the horizontal bar is the error range at 95% confidence.

---

Supporting figure, caption, and Table: Sediment transport analysis under storm waves and currents

Figure S6. Annual frequencies of ocean-wave bed orbital velocities (u) at 50 m depth intervals for the offshore Fiordland region, derived from a 45-year wave hindcast model (Richard Gorman, NIWA, personal communication). Thin lines define the upper envelope of the velocity occurrences at each depth, and are extrapolated to longer RIs (lower annual frequency). The horizontal black line is the mean RI of turbidites in cores T49, T18 and T14, whilst the thickness of this line represents 95% confidence. The band represents the minimum and maximum range of turbidite RI from any of the three cores (see Table 3). The stars reflect fine and medium sand sediment transport conditions modelled in this study from analysis of bed shear velocities u_* and bed shear stresses t_crb/crs (Table S1). Note that sediment transport analysis at 10 m depth intervals indicates that at the mean recurrence interval of 160 ± 340 years, combined extreme storm-waves and currents shift fine and medium sand in water depths to a maximum of 160 m. The vertical dashed lines are approximate current velocity thresholds for sand transport based on HjulstrÖms relationships. In comparison with the results in Table S1, we tested more extreme storm waves, in the event that the extrapolation of 45-year hindcast data to 160 year RIs underestimates the largest storms. We find that hypothetical waves with H_sig of 15 m and period T of 18 s can shift sediment to about 270 m water depths, whilst larger storm waves of H_sig 20 m and T 20 s, could shift sediment to about 350 m.

Table S1. Sediment transport analysis for different water depths*.

* SEDtrans05 program outputs (Nueumeier et al., 2008), for fine sand (0.125 mm) particle size at 50 m water depth intervals off the Fiordland coast, and incorporating peak tidal currents (0.05 m s^-1 at 343º) and extreme storm waves (significant wave height H_sig = 10m, period T = 14s, direction 240º) expected at a 160 ± 40 year recurrence interval (the mean turbidite recurrence at 95%, from Table 3). See text for explanation of model input parameters and discussion. Letters in square brackets [ ] are Sedtrans05 output codes. The crital shear stresses (τ_crb/crs) for fine sand (0.125 mm) bedload and suspended sediment transport relate to the respective critical shear velocities (u_*) (codes USTCRB and USTCRS) via τ_crb/crs = u²_*ρ where ρ is the fluid density in kg m^-3 (1026) (e.g., Li and Amos, 2001). The sediment transport model used here is from Engelund and Hansen, 1967), and the resulting sediment transport direction is relative to North.

---

[ Back ]