High-precision U–Pb zircon age constraints on the duration of rapid biogeochemical events during the Ludlow Epoch (Silurian Period)

Precise determinations of the rates and durations of Palaeozoic biogeochemical events are largely unavailable. Here, we present two new high-precision U–Pb (zircon) dates from volcanic ash deposits from the Ludlow Series (Silurian System) of Podolia, Ukraine, that yielded weighted mean 206Pb/238U dates of 424.08 ± 0.20 (0.29) [0.53] Ma and 422.91 ± 0.07 (0.21) [0.49] Ma (analytical, tracer and total uncertainties). These new dates bracket the largest post-Cambrian global carbon cycle perturbation (Lau Excursion) and constrain the ‘Ludlow Rise’ in 87Sr/86Sr. These chronostratigraphically well-controlled dates improve the calibration of the Silurian time scale and provide the first determinations of the rates of biogeochemical change during the Ludlow Epoch. Supplementary material: U–Pb geochemical methods, data and CL imagery are available at http://www.geolsoc.org.uk/SUP18798.

Changes in the global C and Sr isotopic composition of the ocean, preserved in the stratigraphic record as changes in δ 13 C and 87 Sr/ 86 Sr, can be used as tools for global chronostratigraphic correlation as well as oceanographic and global climate proxies (McArthur et al. 2012;Saltzman & Thomas 2012). Palaeozoic records of δ 13 C and 87 Sr/ 86 Sr contain some of the largest geochemical perturbations in Earth history, but the ultimate cause(s) of these events remain enigmatic (Munnecke et al. 2010). The general lack of precise radio-isotopic age determinations from Palaeozoic strata is a critical factor in our inability to determine the cause-and-effect relationships responsible for these intervals of significant global change. However, recent advances in radio-isotopic methods (Mattinson 2005(Mattinson , 2010Condon et al. 2007), and improved chronostratigraphic correlation of radioisotopic data, make it possible to obtain increasingly precise temporal constraints for Palaeozoic geochemical events.
The Silurian Period was one of the geochemically and biologically most dynamic intervals of Earth history (Munnecke et al. 2010;Cooper et al. 2014), and the Ludlow Epoch contains the largest perturbation of the global carbon cycle (a positive δ 13 C excursion known as the 'Lau Excursion') and one of the most rapid increases in the strontium isotopic composition ( 87 Sr/ 86 Sr) of the ocean during the past 500 myr (Cramer et al. 2011a,b;McArthur et al. 2012;Saltzman & Thomas 2012). At present, five high-precision chemical abrasion-isotope dilutionthermal ionization mass spectrometry (CA-ID-TIMS) 206 Pb/ 238 U dates from zircons are available from Silurian strata, but all of those are limited to the Llandovery and Wenlock series .
Here, we present new, chronostratigraphically well-controlled, high-precision isotope-dilution U-Pb (zircon) dates from two volcanic ash fall (K-bentonite) deposits from the Ludlow Series of Podolia, Ukraine. U-Pb dates were calibrated using gravimetric principles of isotope dilution combined with the chemical abrasion pretreatment method of Mattinson (2005) for the effective elimination of Pb loss (CA-ID-TIMS). These new high-precision dates improve the calibration of the Silurian time scale, effectively bracket the Lau δ 13 C Excursion and, when combined with recently published dates from near the Wenlock-Ludlow boundary , constrain the duration of the Ludlow Rise in 87 Sr/ 86 Sr as well.
The volcanic ash fall deposits, now preserved as K-bentonites in Podolia, are important regional correlation tools , and are the southeasternmost occurrence of lower Palaeozoic bentonites in this part of Europe (Huff et al. 2000). Chronostratigraphic correlation and REE geochemical fingerprinting of bentonites throughout NW Europe indicate that the volcanic activity recorded in the Podolia succession represents a source area distinct from those preserved in the UK, Sweden, Poland and the Baltic States (Huff et al. 2000;Cramer et al. 2012). Whereas most northwestern European explosive volcanism during the Silurian was limited to the Llandovery and Wenlock and related to the closure of the Iapetus Ocean, the Podolia bentonites are largely within the Ludlow and Pridoli and probably originated from a distinct subduction-related tectonomagmatic setting to the SE (present coordinates) within the Rheic Ocean along the Mugodzhar or Kipchak arcs (Şengör et al. 1993;Huff et al. 2000;Cramer et al. 2012).
The two bentonites studied here, M12 and C6 (Fig. 1), were collected from the outcrops Malynivtsi 150 and Ataky 117 (Fig. 1c), respectively (legacy samples from Huff et al. 2000). The nearby section, Zhvanets' 39, also contains the M12 bentonite as well as a record of the Lau Excursion described by Kaljo et al. (2007), 2014-094r apid-communicationSpecial10.1144/jgs2014-094High-precision U-Pb zircon age constraints on the duration of rapid biogeochemical events during the Ludlow Epoch Special which further refined the global chronostratigraphic correlation of these strata. The δ 13 C data of Kaljo et al. (2007) place bentonite M12 at a position immediately below the onset of the Lau Excursion within the Ludfordian Stage. Bentonite C6 comes from a position above the end of the Lau Excursion and was placed by Kaljo et al. (2014) within the upper part of the Ludfordian Stage. Whereas chemostratigraphy cannot provide a secondary confirmation of this correlation, sample C6 comes from a level just below the position of the Ludlow-Pridoli boundary at the top of the Pryhorodok Formation Huff et al. 2000;Kaljo et al. 2007Kaljo et al. , 2012Małkowski et al. 2009; see discussion by Kaljo et al. 2014).

Methods and results.
All zircon samples were analysed in the Isotope Geology Laboratory at Boise State University using CA-ID-TIMS methods (Schmitz & Davydov 2012). We utilized a vigorous one-step (195 °C, 12 h) chemical abrasion pretreatment on single CL-imaged zircon crystals prior to dissolution for the effective elimination of Pb loss (Mattinson 2005). Isotope dilution utilized the EARTHTIME U-Pb tracer (ET535) and ages are reported with respect to the 238 U decay constant of Jaffey et al. (1971). Sample ages and uncertainties (2σ) are reported as ± X (Y) [Z] Ma, where X is the internal error, Y is the internal plus tracer calibration error, and Z is the internal plus tracer plus decay constant uncertainty.
CL-imaging of zircon crystals from sample M12 revealed a homogeneous population of nearly equant to highly elongate prismatic grains with planar, oscillatory internal zoning. Occasional CL-bright cores were seen in more equant prismatic grains, although little sign of zoning truncation was observed. Seven grains selected for a range in size and CL-zoning character survived the chemical abrasion process. Although these samples had relatively low U and radiogenic Pb contents, all analyses yielded concordant and equivalent U-Pb dates with a weighted mean 206 Pb/ 238 U age of 424.08 ± 0.20 (0.29) [0.53] Ma (n = 7; MSWD = 0.76; probability of fit = 0.60), which is interpreted to estimate the eruption and depositional age of the volcanic deposit (Table 1; Fig. 2).
CL-imaging of zircon crystals from sample C6 revealed a homogeneous population of elongate prismatic grains with regular, planar, weakly oscillatory zoning. Occasional CL-bright, sector-zoned cores were seen in more equant grains, and these grains were avoided for their potential to contain inherited components. Eight of the elongate prismatic grains of a variety of CL brightnesses were selected for CA-TIMS analysis. All analyses yielded concordant and equivalent U-Pb dates with a weighted mean 206 Pb/ 238 U age of 422.91 ± 0.07 (0.21) [0.49] Ma (n = 8; MSWD = 1.74; probability of fit = 0.095), which is interpreted to estimate the eruption and depositional age of the volcanic deposit. The more precise age of this sample compared with sample M12 is due to the higher U and radiogenic Pb contents of these grains.
Discussion. The dates from bentonites M12 and C6 provide the first precise calibration of the duration of the Lau positive δ 13 C excursion and demonstrate that the entire δ 13 C excursion was of the order of 1 myr (    duration of 4 myr. These new dates are remarkably consistent with the CONOP scaling and calibration presented by Melchin et al. (2012), which was based on limited radioisotopic information from the Ludlow and Pridoli series, and they demonstrate exceptionally rapid rates of change in the global chemical systems of C and Sr during these Silurian biogeochemical events.
The Lau Excursion appears to be the largest post-Cambrian positive δ 13 C excursion in terms of absolute values as well as total change from baseline. Peak values >+8.0‰ are routinely observed in sections that contain a thick record of the excursion (Samtleben et al. 1996;Wigforss-Lange 1999;Kaljo et al. 2007;Jeppsson et al. 2007). The duration of 1.17 ± 0.21 myr (Table 2) between K-bentonites M12 and C6 provides a maximum duration for the Lau Excursion (Fig. 3). The excursion does not extend exactly to K-bentonite C6, however (Fig. 1), and a conservative estimate places the duration of the excursion at c. 1 myr. With a total duration of 1 myr, the ascending limb of the excursion is likely to have lasted no more than 500 kyr, which indicates a rate of increase of the order of +2.0‰ per 125 kyr. This rate is very fast when compared with steady-state equations of the marine carbon cycle alone and potentially rules out hypotheses such as the carbonate weathering hypothesis (e.g. Kump & Arthur 1999) owing to the prohibitively large amount of carbonate that would need to be weathered in such a short interval of time to produce changes of the magnitude recorded here (e.g. Cramer & Saltzman 2007). However, the additional forcing of a large volcanic event that can flux thousands of Gt of CO 2 into the atmosphere has been modelled to produce rapid and large positive shifts in δ 13 C by promoting oceanographically induced anoxia that in turn promotes efficient remobilization and recycling of PO 4 (e.g. Payne & Kump 2007). Such a scenario represents both oceanographically induced anoxia owing to higher atmospheric CO 2 as well as increased primary productivity, which both lead to enhanced organic carbon burial. Whereas this scenario as a model can produce positive changes in δ 13 C of similar rates and magnitudes to those seen in the Ludlow, the present lack of stratigraphic evidence for such a volcanic event cannot be overlooked and this hypothesis requires substantial further investigation.
Previous age models for the Silurian included poor constraints on the duration of the Ludlow Epoch and therefore uncertain esti-mates of the duration of the Ludlow Rise. Wren's Nest Hill Bentonite 15 ) and K-bentonite M12 (this study) bracket this interval and provide precise temporal control for the duration of the Ludlow Rise. Conservatively, the total change in 87 Sr/ 86 Sr of 0.00025 (0.70845-0.7087, Fig. 1) took place over 4 myr (3.78 ± 0.38 myr, Table 2), which indicates a rate of change in 87 Sr/ 86 Sr of 0.0000625 myr −1 , making this is one of the most rapid positive changes in 87 Sr/ 86 Sr confidently constrained by high-precision radio-isotopic data (McArthur et al. 2012).
There are a few similarly exceptional rates of positive change in 87 Sr/ 86 Sr known, such as the Cambrian Rise, the Early Triassic Rise and the Cenozoic Rise (see McArthur et al. 2012). A variety of potential causative mechanisms for the Ludlow Rise were presented by Cramer et al. (2011b), including the possibility of errors in the calibration of the Ludlow Epoch. The new dates presented here, combined with the Wren's Nest Hill Bentonite 15 date , limit the duration of the Ludlow Rise and effectively discount this possibility. The rapid increase in 87 Sr/ 86 Sr during the Ludlow is near the upper limit for positive rates of change in the global 87 Sr/ 86 Sr value of the oceans (McArthur et al. 2012) and was probably the result of a combination of factors acting in unison. An increase in the global Sr flux owing to increased weatherability along the Caledonian-Appalachian collisional front, an increase in the 87 Sr/ 86 Sr value of the global riverine Sr flux owing to the predominantly quartzofeldspathic material weathering from that collisional front during the early stages of Old Red Sandstone deposition, and a small but concomitant decrease in sea-floor spreading could have all contributed to the Ludlow Rise (see discussion by Cramer et al. 2011b). Each of these factors would have had the effect of driving marine 87 Sr/ 86 Sr values higher and probably acted in unison to produce the Ludlow Rise.

Conclusions
The two new high-precision U-Pb (zircon) dates presented here provide the first precise calibrations of the duration of the Ludlow Epoch (4.95 ± 0.33 myr), the duration of the largest post-Cambrian positive δ 13 C excursion (Lau Excursion, c. 1 myr) and the duration of one of the fastest rates of positive change in 87 Sr/ 86 Sr  Ruppel et al. (1998), Qing et al. (1998), Azmy et al. (1999) and Cramer et al. (2011b) shown with the four high-precision U-Pb dates of Cramer et al. (2012) and the new dates provided here. It should be noted that the time scale at left has been modified from Cramer et al. (2011a) to match the newly available radiometric data. Biostratigraphy after Cramer et al. (2011aCramer et al. ( , 2012 and Melchin et al. (2012).
(Ludlow Rise, 3.78 ± 0.38 myr). The rate of positive change in δ 13 C during the Lau Excursion is calculated at c. +2.0‰ per 125 kyr, and the rate of positive change in 87 Sr/ 86 Sr during the Ludlow Rise is calculated at +0.0000625 myr −1 . A combination of events during the Ludlow Epoch could produce the changes observed in the 87 Sr/ 86 Sr record but the δ 13 C record remains more enigmatic and requires further investigation. These conservatively calculated rates of change provide the first calibrations of the durations of these Silurian biogeochemical events and provide important end-members for future modelling of Palaeozoic global change.