the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Nov 2023
Short communication: Resolving the discrepancy between U–Pb age estimates for the ‘Likhall’ zircon bed, a key level in the Ordovician timescale
Abstract. The ‘Likhall’ zircon bed is a rare case of a single-age zircon population from a carbonate rock, which in this case is contextualised with remarkable biotic and environmental changes as well as meteorite bombardment of Earth after an asteroid breakup in space. Published chemical-abrasion, high-precision isotope-dilution, thermal ionization mass spectrometry (CA-ID-TIMS) U–Pb age estimates disagree at the typical precision of <0.1 % for a 206Pb/238U date, which has led to discrepancies in the interpretation of the timing of events and their possible cause–effect relationships. We evaluate here the relative strengths and weaknesses, and discrepancies in the so far published datasets, propose strategies to overcome them and present a new U-Pb dataset with improved precision and accuracy. Ultimately, we find that domains of residual Pb-loss are a significant source of age-offset between previously published data, amplified by differences in data evaluation strategies. Our new dataset benefits from an improved chemical abrasion protocol resulting in a more complete mitigation of decay-damage induced grain portions, and points to a weighted mean age estimate of 466.37±0.14/0.18/0.53 Ma for the ‘Likhall’ zircon population. This age is intermediate between previous estimates, but outside of analytical uncertainty, and provides a firm tie point for the Ordovician timescale.
- Preprint
(979 KB) - Metadata XML
-
Supplement
(11 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-2597', Anonymous Referee #1, 04 Jan 2024
This study presents a new data set U-Pb zircon dates for the Ordovician carborate-rich ‘Likhall’ zircon bed, as well as a critical assessment of two published U-Pb zircon data sets for the same zircon population. The quality of the new U-Pb data in this study is top-notch, but the manuscript needs improvement in two other aspects: mineralogical support, and attention to detail in interpretation and discussion. The detailed comments linked to the text are below.
Lines 43-44. The paper by Liao et al. (2020) is mentioned in the manuscript 33 times, but is omitted from the reference list. Please add it.
Lines 45-47. “Most importantly, chemical abrasion procedures also differ, being 180°C for 12hrs (Lindskog et al., 2017) and 190°C for 15hrs (Liao et al., 2020). “. Conditions of chemical abrasion include not only leaching time and temperature, but also annealing time and temperature. The latter can be just as important, and should be discussed.
Lines 57-60. “The 206Pb/238U age of the natural reference zircon material Temora reported by Liao et al. (2020) from the mixed TIMS–MC-ICP-MS analysis (417.19±0.15 Ma) is in agreement with the most recent estimates of 417.310 ± 0.074 Ma (von Quadt et al., 2016) and 417.353±0.052 Ma (Schaltegger et al., 2021).”. While this justification is reasonable, it is compromised by the fact that Temora is a natural material, and has been demonstrated to be heterogeneous. The role of using different spikes can be (and must be) evaluated more precisely and reliably by direct comparson of calibrations of the two spikes. Liao et al. (2020) used the spike that was described by Huyskens et al. (2016). That spike was calibrated against two gravimetric U-Pb solutions that were also used by Condon et al. (2015) to calibrate their ET2535 spike. The difference in calibration therefore includes only the calibration procedure itself (about 0.05% or less combined calibration uncertainties from both studies), but not the uncertainty in composition of gravimetric solutions.
Lines 72-74. “Single grains of ‘Likhall’ zircon crystals free of visible inclusions and cracks were hand-picked under a binocular microscope at a magnification of ×20 to ×40 from the same petri dish as the material analyzed by Lindskog et al. (2017). The size of individual fragments was variable, with length ranging from ~50μm to ~300μm.”. CL/BSE images, or at least optical images, of the analysed zircons would be useful.
Line 96. “(206Pb/238U date: 100.173 ± 0.003 Ma; Schaltegger et al., 2021)”. This is unfortunate that the primary gravimetric data for the ET age reference solutions are still not published, and we have to use the mean results from previous studies as a reference value.
Line 97. “n = 32/40”. This is a massive amount of rejection. What mean value and uncertainty do you get if you include all data?
Line 124. “4 Geological Setting”. Why is this section placed after "results"? It would be more appropriate to put it in the more conventional place after Introduction and before Methods.
Lines 152-153. “Subsequently, we only use high Pb*/Pbc (>50) analyses for our age interpretation and will also apply this strategy to the previously published datasets.”. Setting the cut-off of Pb*/Pbc at 50 would be reasonable for evaluation of the data of this study only. But it excludes almost all data from the two previous studies. I suggest to repeat evaluation at two cut-off levels, e.g., 15 and 50 (i.e., <15, 15-50, and >50 bands). The lower level would include sufficient number of data points from the published studies and yield a more objective and informative comparison.
Lines 184-185. “This may be the case even if the optimal calibration of the chemical abrasion procedure (12hrs at 210°C, Widmann et al., 2019) is utilized.”. These conditions were established as optimal by analysis of just one zircon reference material. How can we be sure that they are optimal for any zircon in general?
Lines 191-193. “Such an effect of incomplete removal of metamict domains biased by Pb-loss may possibly be detected through analytical discordance between the two decay schemes, provided that the analytical precision, especially of the 207Pb/235U decay series, was sufficient.”. This reasoning is valid only in the absence of inherited material.
Lines 205-206. “We therefore infer that the data of Liao et al. (2020) and Lindskog et al. (2017) were affected by Pb-loss domains and/or relict inclusions that were not penetrated during chemical abrasion.”. The part of this statement about inclusions equally applies to the current study, because there is no certainty in whether the chemical abrasion procedure of Widnann et al. (2019), or any specific chemical abrasion protocol, guarantees removal of inclusions of other minerals (which can potentially include some minerals that are even more difficult to dissolve than zircon, e.g., cassiterite).
Line 218. “(Schaltegger et al., 2021).”. The reference to Gaynor et al. (2022) Chem. Geol., who discussed this topic in detail, would be more appropriate here.
Lines 229-231. “Evidence for potential inclusions (before chemical abrasion) is provided by imaging of zircon crystals analysed by Lindskog et al. (2017) and Liao et al. (2020), matching our observations during mineral selection.”. More details are needed here. What evidence? What inclusions (composition, size, abundance etc.)? What methods were used to study these inclusions?
Lines 240-242. “These observations are in line with the observations by McKanna et al. (2023a) and the Lindskog et al. (2017) data (Fig 4a), that chemical abrasion at high temperatures is necessary to effectively remove inclusions that are deeply seated within the zircon crystal.”. This does not make much sense. If the inclusions were embedded in zircon during crystallization, and are not confined to high-U growth zones, then they are hosted by crystalline zircon material. They would be released only during complete zircon dissolution. No chemical abrasion would help there.
Lines 250-251. “(v is beyond the scope of this study).”. Why? This is most unfortunate. This could be one of the most interesting parts of this study. And it is not too late to add this comparison during revisions.
Lines 251-252. “… and the presence of Pbc-rich inclusions”. The presence of inclusions is not an "analytical effect". It is a mineralogical / geochemical feature.
Lines 255-256. “Lindskog et al. (2017) preferred a data interpretation based on the statistically most robust weighted mean age, representing the largest number of statistically valid analyses”. Please briefly explain the underlying geochemical assumptions of this, and every other approach to data evaluation. Without that, the comparison of these approaches looks arbitrary.
Line 278. “The youngest concordant zircon U–Pb analysis interpretation can be useful in volcanic samples.”. Any interpretation of an age spectra based on a single data point is highly suspicious, especially since there may be still poorly understood factors that can make some grains particularly prone to Pb* loss (I suspect deformation is one of these factors).
Lines 291-293. “The two datasets produced with the EARTHTIME isotopic tracer ET2535 show better comparability despite divergent chemical abrasion procedures, pointing to a systematic effect of different tracer calibration as well.”. This statement is unfounded. To compare tracer calibrations, you must look at tracer calibrations directly, not at some proxies. And the calibration data for both tracers are published (Condon et al. 2015, and Huyskens et al. 2016, respectively). Both calibrations were done against the same mixed gravimetric reference solutions. So if there is indeed a discrepancy between these spikes, it does not exceed the uncertainty of the calibration process itself (low 0.0x% values).
Line 296. “We can conclude that Pb*/Pbc is of fundamental importance for the ‘Likhall’ zircon datasets.”. This is the main conclusion of this study: to date reliably, we must get the ratio of radiogenic Pb to common Pb as high as possible (although this is not a novel idea). Furthermore, it is likely to be the only thing that really matters here. You have to admit that the evaluation of the roles of all other factors is tentative at best, and say so directly.
Line 322. “(c) our new data show higher Pb*/Pbc”. This point is the most important, and you should emphasize it, maybe by bringing it to the front. Higher precision is largely a result of lower common Pb correction. Whether the difference in chemical abrasion protocols actually influenced the data cannot be reliably deduced from the existing data sets.
Citation: https://doi.org/10.5194/egusphere-2023-2597-RC1 -
AC1: 'Reply on RC1', André Navin Paul, 28 Feb 2024
(RC1) This study presents a new data set U-Pb zircon dates for the Ordovician carborate-rich ‘Likhall’ zircon bed, as well as a critical assessment of two published U-Pb zircon data sets for the same zircon population. The quality of the new U-Pb data in this study is top-notch, but the manuscript needs improvement in two other aspects: mineralogical support, and attention to detail in interpretation and discussion. The detailed comments linked to the text are below.
(AC) Dear reviewer, we firstly would like to thank you for taking the time and effort to review and comment on the manuscript. Below, we are responding to your line specific comments.(RC1) Lines 43-44. The paper by Liao et al. (2020) is mentioned in the manuscript 33 times, but is omitted from the reference list. Please add it.
(AC) We are adding the missing reference details.(RC1) Lines 45-47. “Most importantly, chemical abrasion procedures also differ, being 180°C for 12hrs (Lindskog et al., 2017) and 190°C for 15hrs (Liao et al., 2020). “. Conditions of chemical abrasion include not only leaching time and temperature, but also annealing time and temperature. The latter can be just as important, and should be discussed.
(AC) We agree with the reviewer and expand this section to include relevant information. Works by e.g. Nasdala et al. (2001; 2002), Ende et al. (2021) and others indeed have demonstrated a time-temperature dependence of the annealing efficiency. Moreover, too high annealing temperature at ambient pressure is reported to reset U-Pb systematics (Mattinson et al., 2001; 2011) (>1100°C). However, between the Liao et al. (2021) and Lindskog et al. (2017) data, the same temperature (900°C) was applied. The only difference is the duration (48 vs 72h), which may have an (at this point) unquantifiable effect on the efficiency of thermal annealing prior to partial dissolution.(RC1) Lines 57-60. “The 206Pb/238U age of the natural reference zircon material Temora reported by Liao et al. (2020) from the mixed TIMS–MC-ICP-MS analysis (417.19±0.15 Ma) is in agreement with the most recent estimates of 417.310 ± 0.074 Ma (von Quadt et al., 2016) and 417.353±0.052 Ma (Schaltegger et al., 2021).”. While this justification is reasonable, it is compromised by the fact that Temora is a natural material, and has been demonstrated to be heterogeneous. The role of using different spikes can be (and must be) evaluated more precisely and reliably by direct comparson of calibrations of the two spikes. Liao et al. (2020) used the spike that was described by Huyskens et al. (2016). That spike was calibrated against two gravimetric U-Pb solutions that were also used by Condon et al. (2015) to calibrate their ET2535 spike. The difference in calibration therefore includes only the calibration procedure itself (about 0.05% or less combined calibration uncertainties from both studies), but not the uncertainty in composition of gravimetric solutions.
(AC) We are adding more detail on the inter-comparability of the two labs’ performance. Mainly through the reported ET100 results, which are in line with the consensus value.(RC1) Lines 72-74. “Single grains of ‘Likhall’ zircon crystals free of visible inclusions and cracks were hand-picked under a binocular microscope at a magnification of ×20 to ×40 from the same petri dish as the material analyzed by Lindskog et al. (2017). The size of individual fragments was variable, with length ranging from ~50μm to ~300μm.”. CL/BSE images, or at least optical images, of the analysed zircons would be useful.
(AC) We agree that CL/BSE and optical images can be useful, but considering that two previous studies have provided such without any benefit to the U-Pb age interpretation we see no additional scientific value of these in this study. Essentially, we made no observations concerning crystal characteristics that may add to the already published details.(RC1) Line 96. “(206Pb/238U date: 100.173 ± 0.003 Ma; Schaltegger et al., 2021)”. This is unfortunate that the primary gravimetric data for the ET age reference solutions are still not published, and we have to use the mean results from previous studies as a reference value.
(AC) We agree that the community will benefit from publication of the primary gravimetric data for the ET reference solutions.(RC1) Line 97. “n = 32/40”. This is a massive amount of rejection. What mean value and uncertainty do you get if you include all data?
(AC) The mean value and uncertainty without any data rejection is 100.1639±0.0043 Ma, compared to the preferred reported age of 100.1678±0.0046 Ma. We note that in lines 97-98 we mention that one batch of the ET 100 is rejected, for an anomalously young average age. The argument of rejection of this batch from overall consideration stems from the process and interlab comparison observations in Schaltegger et al. (2021). The ET 100 solutions are routinely prepared in batches, which are split into aliquots after processing (1 batch yielding 5-8 aliquots), as reported and recommended in Schaltegger et al. (2021) (process described in detail in that study). Schaltegger et al. (2021) reported data and discussion of similar observations from different labs, an issue that remains to be evaluated and solved by the EARTHTIME community. We will ensure that the full ET 100 data is available in the supplementary repository or similar.(RC1) Line 124. “4 Geological Setting”. Why is this section placed after "results"? It would be more appropriate to put it in the more conventional place after Introduction and before Methods.
(AC) We are adjusting the sequence of the two items.(RC1) Lines 152-153. “Subsequently, we only use high Pb*/Pbc (>50) analyses for our age interpretation and will also apply this strategy to the previously published datasets.”. Setting the cut-off of Pb*/Pbc at 50 would be reasonable for evaluation of the data of this study only. But it excludes almost all data from the two previous studies. I suggest to repeat evaluation at two cut-off levels, e.g., 15 and 50 (i.e., <15, 15-50, and >50 bands). The lower level would include sufficient number of data points from the published studies and yield a more objective and informative comparison.
(AC) We thank the reviewer for this suggestion and will consider this calculation and comparison of the expanded datasets in a revised manuscript version at incremental Pb*/Pbc higher cut-offs.(RC1) Lines 184-185. “This may be the case even if the optimal calibration of the chemical abrasion procedure (12hrs at 210°C, Widmann et al., 2019) is utilized.”. These conditions were established as optimal by analysis of just one zircon reference material. How can we be sure that they are optimal for any zircon in general?
(AC) This is of course a very fair comment, as the data in Widmann et al. (2019) are strictly valid for Plesovice only. The repeated application, however, demonstrates that our repeatability has improved since we have adopted these. We are rephrasing to include the possibility that compositional and other effects can influence chemical abrasion results.(RC1) Lines 191-193. “Such an effect of incomplete removal of metamict domains biased by Pb-loss may possibly be detected through analytical discordance between the two decay schemes, provided that the analytical precision, especially of the 207Pb/235U decay series, was sufficient.”. This reasoning is valid only in the absence of inherited material.
(AC) We are rephrasing to state that this assumes absence of inherited material.(RC1) Lines 205-206. “We therefore infer that the data of Liao et al. (2020) and Lindskog et al. (2017) were affected by Pb-loss domains and/or relict inclusions that were not penetrated during chemical abrasion.”. The part of this statement about inclusions equally applies to the current study, because there is no certainty in whether the chemical abrasion procedure of Widmann et al. (2019), or any specific chemical abrasion protocol, guarantees removal of inclusions of other minerals (which can potentially include some minerals that are even more difficult to dissolve than zircon, e.g., cassiterite).
(AC) The reviewer is correct in pointing out that any zircon may exhibit relict Pb-loss after chemical abrasion. We already present evidence of Pb-loss in the ‘Likhall” zircons in Fig. 1 of the original submission. Mentioned in line 204-205 is the observation that at least 2 analyses of our new dataset exhibit Pb-loss, which demonstrates that the Widmann procedure did not remove all Pb-loss in this study. We will slightly rephrase to clarify.(RC1) Line 218. “(Schaltegger et al., 2021).”. The reference to Gaynor et al. (2022) Chem. Geol., who discussed this topic in detail, would be more appropriate here.
(AC) The reviewer is incorrect here. The study by Gaynor et al., 2022 discuss Pb*/Pbc and its effect on the total uncertainty. The study by Gaynor et al. (2022) does not however model the effect of changing blank isotopic correction (IC) on the absolute dates. Since we refer to the effect of changing the blank IC effect on the absolute date, we continue to refer to Schaltegger et al. (2021). We will expand this section by stating that higher Pb*/Pbc reduces the dependency of both the absolute date and its uncertainty to clarify.(RC1) Lines 229-231. “Evidence for potential inclusions (before chemical abrasion) is provided by imaging of zircon crystals analysed by Lindskog et al. (2017) and Liao et al. (2020), matching our observations during mineral selection.”. More details are needed here. What evidence? What inclusions (composition, size, abundance etc.)? What methods were used to study these inclusions?
(AC) We are explaining in more detail the nature of our observations, which are primarily observations made and noted during multiple picking sessions for the annealing, chemical abrasion, washing and dissolution steps. During the picking process, zircons with visible cavities and inclusions (these may be melt or mineral inclusions) are identified. While we take great care in examining zircons selected for analysis under the microscope, we cannot exclude with absolute certainty that sub-visible inclusions are present.(RC1) Lines 240-242. “These observations are in line with the observations by McKanna et al. (2023a) and the Lindskog et al. (2017) data (Fig 4a), that chemical abrasion at high temperatures is necessary to effectively remove inclusions that are deeply seated within the zircon crystal.”. This does not make much sense. If the inclusions were embedded in zircon during crystallization, and are not confined to high-U growth zones, then they are hosted by crystalline zircon material. They would be released only during complete zircon dissolution. No chemical abrasion would help there.
(AC) We are rephrasing to clarify that the dependency of inclusion accessibility depends on structural features and not primarily on time and temperature. Observations and experience accrued over time show that 1) inclusions are very often (not always) surrounded be altered/recrystallized zircon, an 2) linked to the surface by cracks.(RC1) Lines 250-251. “(v is beyond the scope of this study).”. Why? This is most unfortunate. This could be one of the most interesting parts of this study. And it is not too late to add this comparison during revisions.
(AC) We agree that it would be exciting to produce a more refined time scale for the overall time interval that encompasses the ‘Likhall’ bed, but there is simply not enough reliable contextual data. More specifically, considering the data discussed in our study: Computing a model age based on data that may be flawed in some form is not advisable. In the Liao et al. dataset we’d assume presence of Pb-loss, in Lindskog et al. higher dependence on initial/blank Pb IC and in our new dataset Pb-loss and inheritance. These variables make it too difficult/unrealistic to make any meaningful progress at the moment.(RC1) Lines 251-252. “… and the presence of Pbc-rich inclusions”. The presence of inclusions is not an "analytical effect". It is a mineralogical / geochemical feature.
(AC) We are clarifying that “analytical and mineralogical effects…” are considered.(RC1) Lines 255-256. “Lindskog et al. (2017) preferred a data interpretation based on the statistically most robust weighted mean age, representing the largest number of statistically valid analyses”. Please briefly explain the underlying geochemical assumptions of this, and every other approach to data evaluation. Without that, the comparison of these approaches looks arbitrary.
(AC) We are not sure what the reviewer means with “geochemical” assumptions, however we are happy to expand the sections by briefly mentioning that:i) The underlying assumption is that some un-resolved Pb-loss may affect the youngest analyses and that inheritance may explain the oldest range of concordant zircons analysed (Samperton et al., 2015).
ii) The underlying assumption is that all Pb-loss is effectively removed and that inheritance or protracted growth in the magma chamber is responsible for the older analyses.
iii) The underlying assumption is that all concordant analyses reflect growth from the same magmatic system (Samperton et al., 2015).
iv) The underlying assumption is that all zircon analyses except the youngest is affected by inheritance or growth during protracted magma chamber activity (Schaltegger et al., 2009).
(RC1) Line 278. “The youngest concordant zircon U–Pb analysis interpretation can be useful in volcanic samples.”. Any interpretation of an age spectra based on a single data point is highly suspicious, especially since there may be still poorly understood factors that can make some grains particularly prone to Pb* loss (I suspect deformation is one of these factors).
(AC) As with any age interpretation, there is either statistical, geological or philosophical factors that are weighed. We here simply test previously published and executed interpretation strategies and explore their relative differences, without making any recommendation, thus don’t see any need for changes.(RC1) Lines 291-293. “The two datasets produced with the EARTHTIME isotopic tracer ET2535 show better comparability despite divergent chemical abrasion procedures, pointing to a systematic effect of different tracer calibration as well.”. This statement is unfounded. To compare tracer calibrations, you must look at tracer calibrations directly, not at some proxies. And the calibration data for both tracers are published (Condon et al. 2015, and Huyskens et al. 2016, respectively). Both calibrations were done against the same mixed gravimetric reference solutions. So if there is indeed a discrepancy between these spikes, it does not exceed the uncertainty of the calibration process itself (low 0.0x% values).
(AC) We are abandoning this statement.(RC1) Line 296. “We can conclude that Pb*/Pbc is of fundamental importance for the ‘Likhall’ zircon datasets.”. This is the main conclusion of this study: to date reliably, we must get the ratio of radiogenic Pb to common Pb as high as possible (although this is not a novel idea). Furthermore, it is likely to be the only thing that really matters here. You have to admit that the evaluation of the roles of all other factors is tentative at best, and say so directly.
(AC) Neglecting the potential impact of interpretation strategy is an erroneous assessment by the reviewer in our opinion. Reducing the focus on Pb*/Pbc is a rather crude oversimplification. We clearly demonstrate that discrepancy between the interpretation strategies has amplified the difference in the published U-Pb dates when comparing the Liao et al. (2021) and Lindskog et al. (2017) studies.(RC1) Line 322. “(c) our new data show higher Pb*/Pbc”. This point is the most important, and you should emphasize it, maybe by bringing it to the front. Higher precision is largely a result of lower common Pb correction. Whether the difference in chemical abrasion protocols actually influenced the data cannot be reliably deduced from the existing data sets.
(AC) We have no objection in moving point c) to a) and will do so in the revised manuscript.Citation: https://doi.org/10.5194/egusphere-2023-2597-AC1
-
AC1: 'Reply on RC1', André Navin Paul, 28 Feb 2024
-
RC2: 'Comment on egusphere-2023-2597', Anonymous Referee #2, 02 Feb 2024
See general comments and annotations on the manuscript in attached PDF.
-
AC2: 'Reply on RC2', André Navin Paul, 28 Feb 2024
(RC2) The paper presents a third U-Pb zircon data set for zircons recovered from the Ordovician Likhall carbonate horizon in Sweden. The data appears to be superior to previous efforts in that they have successfully increased the ratio of radiogenic to common Pb in a subset of zircons that they use to determine their preferred age. The data appear to be of high quality and the new age an improvement that will have impact on the Ordovician timescale and a debate concerning a possible link between the Great Ordovician Biodiversification Event (GOBE) and the break up and arrival of an L-chondrite asteroid. On these merits alone, the paper should be considered worth publishing.
However, the paper gets seriously sidetracked in its effort to assess the two previously published U-Pb zircon studies of this same horizon (same samples in at least one other study as the author Lindskog provided the zircons to both). In this endevour, they provide a confusing (and in places factually incorrect) discussion of why they believe the different studies have delivered different ages. They ultimately suggest that the other ages are inaccurate but then go ahead anyways and try to assess the different methods of treating a range of ages from a volcanic horizon that only makes sense if all the individual zircon ages are assumed to be accurate.
While the paper expends much effort in the comparison of different data sets and recommendations for the chemical abrasion method, there is only 8 lines of discussion (Section 5.6) of the implication of their urevised age on timescale issues and no mention that I found of the implication for the debate on whether the Ordovician L-chondrite break up had anything to do with the GOBE.
I would recommend that the authors be encouraged to make major revisions to the paper, toning down the detailed assessment of the two previous studies. They can demonstrate by way of the Pbr/Pbc that their data (and uncertainties) are better than previous efforts and then focus on the implications.
I have annotated the manuscript extensively with comments and provide the following more general points keyed to different sections of the discussion that I feel are important.
(AC) Dear reviewer, thank you for taking the time and effort to assess this manuscript submission. Below we respond to your general comments, while your annotations will be considered directly when producing a revised manuscript. We have opted to focus primarily on the multiple factors that can sway a U-Pb age interpretation, to make the most confident interpretation of available data. We are happy to expand the section 5.6 a bit more, but the primary focus of our study lies on showcasing the sources of discrepancy between analyses of practically identical sample materials.(RC2) 5.1 Radiogenic Pb/common Pb ratio (Pb*/Pbc) as a selection criterion
In this section, the authors suggest that the zircons analysed in Lindskog et al had inclusions that were not removed in the partial dissolution step of the chemical abrasion method. While this may be true, they have missed or decided to ignore the point that this study assumed that all non-radiogenic Pb was blank (or modern Pb that was not removed from the zircon before digestion). If it is initial common Pb it would have an Ordovician Pb composition that would have had lower 7/4 and 6/4 ratios resulting in older, not younger, ages. They never recognize this problem and seem to consider common Pb, blank Pb and initial Pb as one and the same.
(AC) We agree that changing the IC of the blank correction will drive the Lindskog et al. (2017) data to either older or younger dates, subject to choosing a Pb blank IC. Given the radiogenic assignment of the Pb blank IC in Lindskog et al. (2017) (6/4 = 18.5), choosing a 500 Ma S&K value will make the Lindskog et al. (2017) data older. We can however, not assess if the original Pb blank IC is over- or underestimated in the Lindskog et al. (2017) study. Recognizing the correlation between U-Pb date and Pb*/Pbc, we interpret it as an artefact of inaccurate Pb blank correction. As the reviewer mentions in the annotated manuscript, we will explain more the effect on the size of the ellipse and the absolute age. Hence, we pursue the track of suggesting a cut-off value for Pb*/Pbc in this section, to reduce bias introduced from choice of Pb blank IC and its random variability carried over into the final age interpretation.(RC2) 5.2 Residual Pb-loss in chemically abraded natural zircon
After concluding in section 5.1 that the Lindskog et al data was troubled by low Pb*/Pbc, they explore here a different tack in that perhaps there were residual domains that had lost Pb that were not removed by the lower temperature etching step employed in this study. But this would have the opposite effect of producing younger ages rather than the older ages they are trying to explain away. L205 states that they conclude that both previous studies “were affected by Pb-loss domains and/or relict inclusions that were not penetrated during chemical abrasion.” This may be true for the Liao et al study that recommended a younger age but not the Lindskog et al data as suggested.
(AC) In the available CA studies (e.g. Mattinson et al., 2005; Crowley et al., 2014; Huyskens et al., 2016; McKanna et al. 2024), lower T chemical abrasion (e.g. 180°C or 190°C) results in incomplete Pb-loss removal. Based on this observation we have to make the assumption that the Lindskog et al. (2017) data, which are chemically abraded at 180°C, are affected by this incomplete Pb-loss. It is unreasonable to assume that the higher T chemical abrasion procedure is less efficient than the lower T chemical abrasion. The apparent “old” Lindskog data is potentially an artefact of the blank correction or presence of inheritance. Therefore, we prefer to base our assessment of Pb-loss on the reported duration and temperature of the chemical abrasion, to indirectly assess probable residual Pb-loss. Ultimately, we would state that a concluding assessment of Pb-loss removal in the Lindskog et al. (2017) data is precluded by the lower precision of the data.(RC2) 5.3 Lead blank isotopic composition correction effects on the spread of zircon U–Pb dates
Here correlations between Pbr and Pbc are considered in terms of their absolute abundances without considering the difference between Pbc as Pb blank or Pb initial and that assuming one or the other would have different implications for the discussion. Again, they fail to discuss the Lindskog et al paper having assigned all non-radiogenic Pb as modern terrestrial Pb and that their interpretation of Pb-bearing inclusions of presumed Ordovician age would have lower 6/4 and 7/4 ratios, producing older calculated ages in both decay schemes. This makes the different in ages between the two studies worse rather than explaining it away.
(AC) Initial Pb and blank Pb cannot be distinguished by the ID-TIMS method with high confidence (ignoring indirect attempts, as applied to Pbc-bearing phases, as these are to imprecise for high precision zircon U-Pb dating). Arguably, here, the reviewer is too fixated on changing the Pb blank IC in a particular direction. In lines 217-218 we clearly state that it may have an effect that drives either towards younger or older dates.(RC2) 5.4 The impact of the interpretation strategy on U–Pb zircon ages
This section runs through the different established strategies employed to make a sensible interpretation of a spread of U-Pb zircons commonly found in volcanic ash layers. A combination of issues may be in play here to cause the spread including residence time of zircons (related to magma chamber timescales) and/or detritial grains from an earlier eruptions but issues that have geological explanations. These strategies assume that all the individual ages of zircons in the range are accurate for their respective crystallization and closure. But previous sections have concluded that the 2 previous studies and especially that of Lindskog et al are not reliable ages due to their high Pbc. If they have concluded that at least the Lindskog et al ages are inaccurate then they should not be considered further in this section. Only the zircons with their cut off of >50 for Pbr/Pbc should be considered here.
(AC) We will outline more clearly the underlying hypotheses for each of the interpretation strategies to help clarify for the reader. This comparison is necessary to better understand all individual sources contributing to the discrepancy of the Lindskog et al. (2017) and Liao et al. (2020) ages, before conducting a new, revised age calculation. In section 5.5 we do perform the interpretation of data with Pb*/Pbc >50.(RC2) 5.6 Implications for the Ordovician timescale and the absolute timing of events
This section is eight lines long without discussing the implications for the debate mentioned in the introduction regarding the connection between the Ordovician break up and arrival of the L-chondrite and the GOBE or any implications for the new age on the Ordovician timescale save that it needs to be modified. This should be the most important part of the paper else why bother to get a refined age for this horizon in the first place. As presented, it seems the main justification for this study was to show that that you need higher temperatures and longer times for the chemical abrasion method.
(AC) As indicated above, we will slightly expand this section and will include additional relevant references, but we maintain that the primary focus of our study has been to explore the significant differences in results/interpretations between previously published studies and to add state-of-the-art U-Pb data to resolve the situation at hand. Thus, we mainly discuss the fundamental implications of our new results, and additional details pertaining to the GOBE and L-chondrite breakup may be found in the cited literature. The ‘Likhall’ bed is indeed a key level in Ordovician event stratigraphy and time scale, so our results by themselves comprise a significant contribution to move the field ahead.Citation: https://doi.org/10.5194/egusphere-2023-2597-AC2
-
AC2: 'Reply on RC2', André Navin Paul, 28 Feb 2024
-
RC3: 'Comment on egusphere-2023-2597', Anonymous Referee #3, 29 Feb 2024
-
AC3: 'Reply on RC3', André Navin Paul, 28 Mar 2024
(RC) Thanks to the editor and authors for the opportunity to review this interesting paper. This study revisits a previously dated zircon-bearing carbonate bed, the Likhall bed, which has important implications for the Ordovician time scale but has yielded inconsistent dates between studies. The authors attribute these differences to both
- methodological issues (mainly incomplete chemical abrasion by previous studies, as well systematic differences in tracer calibration between studies using an ET spike and a study using an in-house spike) and
- interpretive differences relating to how one determines a single age for a bed that contains multiple dated zircon grains.
The authors grapple with these problems and put forward a new weighted mean age estimate
for the bed. The study is well-written, concise, and clear. I support publication of the study and make the following (hopefully constructive) suggestions.
(AC) Dear Reviewer, thank you for taking the time to read and comment on the submission. We appreciate your opinion and constructive comments. Below we respond to your comments.
(RC)
This bed was previously dated by two studies, Lindskog et al. 2017 and Liao et al. 2020, and ages from all three studies are shown in Figure 2.
- This figure shows all the ages with their X uncertainty, but I believe they should all be shown with their Y uncertainty (including tracer calibration uncertainty), as the Liao data were determined using an in-house spike and the Lindskog/Paul data were generated using an ET spike. If the authors don’t wish to or can’t do this, then I suggest caveating this in the caption.
(AC) The expansion to the Y uncertainty will inflate the bars by ca. 3 kyrs (ET datasets), which is too little to be visible/significant. Similarly for the Liao et al. data. We can point this out in the caption, however.
(RC)
Line 95: Can the authors provide a citation supporting the assumption of a magma initial Th/U ratio of 3.5 +/- 1?
Figure 1:
- I suggest recoloring this figure so that the cut-off value of 50 is at an inflection point in the colorscale and/or indicate on the colorscale where 50 is. It is hard to look at this and immediately understand that some of the green grains that overlap with the concordia curve are actually below 50 and should be left out.
- It’s also somewhat confusing that the grains “selected” for analysis here are not the same as the grains selected in Table S1, column “interpretation strategy i.”
o I suggest annotating this column in Table S1 to indicate the three grains that are selected for analysis using the more restrictive Pb*/Pbc screen.
o I suggest adjusting the figure so that all grains that are “selected” using
“interpretation strategy i” have a black dashed ellipse and those that are “selected” using the more strict Pb*/Pbc screen have a black ellipse (as they do now).
- Isn’t a part of the selection criteria also that the ages are within the youngest cluster (line 305)? This should be stated in the caption.
(AC) We will look to make some adjustments to figure. Since a gradient will be challenging to adjust appealingly, we may try a simplified two- or three-color scheme then. Table S1 may be amended to include the Pbc/Pb* >50 cut-off, albeit it only leaves few data points. As for adding strategy i), this may overload the figure. We would however prepare the graphs for all interpretation strategies and make them available in the appendix/online version.
(RC) Line 125: I think the geological setting should be given greater priority within the manuscript and encourage the authors to include a figure showing the field context of the bed sampled. I understand that this has been described elsewhere, but such a figure would increase the value of this manuscript to the reader. Similarly, images of the zircon grains have been previously published elsewhere, but including some images in this manuscript or in the supplement would be helpful.
(AC) The placement of the geological setting will be adjusted. As in inflating the volume of the manuscript, we would much prefer to have this available in the appendix. We have found no meaningful correlation with zircon petrography and U-Pb systematics.
(RC) The authors first consider differences in the chemical abrasion techniques used between studies, and argue that incomplete chemical abrasion of inclusions and metamict zones within zircons resulted in inaccurate ages for the previous studies. Previous studies used lower temperatures for their chemical abrasion procedures than the current study. The main data the authors use to support this idea is shown in Figure 4, which plots Pbc vs Pb* for each of the three studies.
- This is more of an aesthetic suggestion, but I suggest making all the plots the same sizeand aligning their left y-axis.
(AC) We will clean this up, sure.
(RC) I also think that plotting all the data with the same axes or on the same plot would help the authors make their point that their approach has resulted in much smaller Pbc measurements and presumably more precise and accurate dates.
(AC) We have initially opted against this since the scale difference is almost a factor of 10 when all data are displayed. If we find a more appealing and useful graphical solution, we are happy to present that of course.
(RC)
Because chemical abrasion is meant to remove zones of zircon affected by Pb loss, it’s easy to follow this argument for the Liao data, which the authors argue is too young because of incomplete chemical abrasion, but harder to follow it for the Lindskog data, which is too old (and the authors discuss this). They suggest that the “too old” ages come frominaccurate and imprecise blank corrections.
o Line 231: “If we assume that larger zircons contain more Pb* and a larger volume of Pbc-bearing inclusions…” I’m happy to agree that larger zircons would have more Pb* and Pbc, but I’m missing a step in the logic of the sentence here. I think it’s that larger zircons with more Pbc would be more impacted by an erroneousblank correction, but perhaps it can be spelled out a bit more clearly.
(AC) We will attempt to clarify this more.
(RC) The authors consider several different interpretation strategies. They identify five ways that one might determine the age of this bed:
- Weighted mean of a subset of data
- Youngest cluster of overlapping ages at 2-sigma
- Use entire range of concordant zircon analyses as autocrystic growth in magma chamber
- Use the youngest concordant grain as the best proxy for the timing of eruption
- A Bayesian approach as suggested by Keller et al. 2018
- I would rephrase 5 to something like “A Bayesian approach such as those suggested by Keller et al. 2018 or Traylor et al. 2021.”
(AC) We will consider the suggestion to include relevant citations.
(RC) o The authors immediately state that #5 is outside the scope of this study, but I don’t agree. The approach used by Keller et al. 2018 is relatively easy to implement and
a publicly-available Jupiter Notebook provides help at https://github.com/brenhinkeller/Chron.jl?tab=readme-ov-file. I suggest that it be included and discussed; if the authors have a major disagreement with this approach, they should show why.
(AC) The Keller model is available indeed and calculations may be performed quite easily. However, comparing the three data sets is strictly not feasible. To achieve comparable results, all data sets must have eliminated Pb-loss efficiently, be directly comparable in instrumental setup and of sufficient high precision. This is not possible here. The Keller model will generate an model age that approximates the youngest zircon U-Pb age of the data set. Thus, when rejecting only analyses that a clearly discordant in Concordia space, the Keller model results will be:
ModelAge
2sigma
Liao
464.9969
0.360647
Lindskog
466.7484
0.803836
this study
466.4687
0.261105
We thus need to develop a strategy to unify the data before we can expect comparable results. This, in turn, requires a much larger dataset first, which goes beyond the scope of what we want to discuss here.
(RC) o Traylor et al. refers to modifiedBChron, which would consider all analyses as part of a summed probability density distribution function and then use stratigraphic superposition in a Bayesian model to determine an age that is supposed to be a better representative of geological uncertainty. Using only the new data produced by this study, of course this can’t be applied as we don’t have the benefit of stratigraphic superposition, but if, as the authors state, there are other dated bentonites in the section with clear stratigraphic relationships, why not give it a try, and see how it compares? modifiedBChron is also fairly easy to implement. I suspect the outcome will mostly highlight the importance of getting good ageconstraints elsewhere in the section, but this would be worth highlighting too. The Bayesian approaches introduced by Traylor and Keller are exciting new developments that help us understand what the meaning of a dated ash bed means in stratigraphic context.
(AC) We thank the reviewer for the very constructive thought and consideration. We will be happy to explore the Traylor et al. model in a second research item, to not extend beyond the short communication format chosen here. We will like to explore this when presenting some new bentonite dates from Argentina, that temporally overlap with this work, where we have supporting fossil records and stratigraphy.
(RC) This paper’s ambition is to make the point that the Ordovician timescale must be reconsidered, and it certainly provides evidence in that direction, but at present, this paper misses an opportunity to grapple with these emerging approaches and their implications for the timescale.
(AC) We thank the reviewer and hope to present a different, second, manuscript to the community that may incorporate more of the above mentioned points. Here, we have opted to focus primarily on more technical and interpretational aspects to clarify the U-Pb zircon date of this layer.
(RC) The authors spend a lot of time discussing using Pb*/Pbc as a screening metric and note that it significantly improves accuracy while coming at a significant cost of greatly reducing the number of grains that are viable. Can they give non-zircon geochronologists some sense of how a strict requirement of Pb*/Pbc > 50, for example, would affect the universe of published CA-ID-TIMS zircon data? Would it knock out >50% of published zircon grains, as it does here? Given the expense and time-consuming nature of CA-ID-TIMS analysis, how do they recommend other workers grapple with this?
(AC) Continued progress in laboratory procedures and technical improvements has been a fundamental pillar of the CA-ID-TIMS U-Pb community for the past decades now, and we expect that future data sets will continue that trend. As we gather more and more data on how to avoid common Pb (inclusions and/or consumables), newly generated data are expected to fall above that threshold. For past data, certainly a large number will become less relevant. Seeing more than 75% of data becoming too “imprecise” for future consideration would not be surprising.
(RC) Line 299: “leaves only 8 of 22”—yes and worth stating explicitly that these 8 are not all within the youngest age cluster.
(AC) We will emphasize as suggested.
(RC) The authors describe some of the changes to the Ordovician timescale that their new age requires.
It would be helpful for the authors to visualize these changes in a “before” and “after” figure.
(AC) That is something that we hope to describe more extensively in a follow up submission, when more data is generated and combined from relevant sections.
(RC) The authors also note that the timing of L-chondrite breakup should be revised to c. 467.1 Ma, but the reader isn’t given enough context about why this date should be assigned to
the breakup, (which isn’t the age they give the Likhall bed).
(AC) We may eliminate that statement, as it requires a return to discussion of cosmic ray exposure dating conducted previously. Extending the discussion to the relevant level of depth may be beyond the frame of this short submission, although we did want to mention the approximate estimate by us to the community.
(RC) As the authors motivated this study by mentioning the controversial hypothesis linking meteorite bombardment with Ordovician biodiversification in the introduction, they should return to this hypothesis and discuss the alignment or un-alignment of these two events in light of their new results.
(AC) We will extend this section slightly, hopefully providing a more satisfying answer.
(RC) Finally, the references:
- Liao et al. 2020 is a key study for this paper and the source of much of the legacy data
discussed, but it does not appear in the reference list. This must be addressed before final
publication.
(AC) Adding the missing reference.
(RC) I suggest reformatting the Schmitz et al. citation in the references to superscript 205, 235, 238, etc. Same thing for the von Quadt et al. citation and 10^13.
(AC) Modifying the reference / checking correct super/subscripts throughout.
(RC) On the topic of CA-ID-TIMS dating of cryptotephra from carbonates: The authors might be interested to know about a similar study focused on retrieving volcanic zircon from
carbonate rock: Finzel and Rosenblume 2021, Geology. I think referencing this study would help make the authors’ point that determining best interpretive practices for
carbonate-derived crypotephra ashes is a question with applications beyond the single bed considered here.
(AC) We thank the reviewer for pointing out this reference, we will check it and incorporate where suitable.
Citation: https://doi.org/10.5194/egusphere-2023-2597-AC3
-
AC3: 'Reply on RC3', André Navin Paul, 28 Mar 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 247 | 84 | 25 | 356 | 19 | 12 | 11 |
- HTML: 247
- PDF: 84
- XML: 25
- Total: 356
- Supplement: 19
- BibTeX: 12
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1