Double dating in the Middle Pleistocene: assessing the consistency and performance of the carbonate U–Th and U–Pb dating methods

Pollard et al. submitted to Geochronology

General comments

Pollard et al present a suite of new U-Th and U-Pb ages for samples from a cave site in Italy  - Corchia Cave - that has seen numerous dating campaigns by the Melbourne group over the past two decades or so.    The researchers provide valuable insight for the geochronology community by addressing: (1) the relative merits of the two uranium-series disequilibrium methods for carbonates that formed in the time window 430 to 630 ka;  (2) the potential causes and consequences of ²³⁸U/²³⁵U variation in speleothems. 

The results and interpretation are worthy of publication in GChron and the text is in good shape.  The findings are robust and based on state-of-the-art technology and protocols that have been reviewed elsewhere.  Figures and tables are well-presented.  

It is fair to say that the results are not strictly novel or definitive for a host of reasons outlined below, but they do present an elegant exploration of the factors determining uncertainty (particularly, relative precision) for such geochronological tools.

Ahead of what might be considerable time and effort, a researcher with a wide range of state-of-the-art dating techniques available to them would ideally select the most efficient and robust options to constrain the age of geological material.   For methods based on the closed-system return to equilibrium in the uranium-series decay chain, high abundance of the parent isotopes and minimal or well-constrained initial abundance of relevant daughter isotopes in the decay chain is required.  For U-Th methods, high U and low Th are a pre-requisite for higher precision age determination towards the older portions of the effective time window.  For U-Pb methods, the precise and accurate determination of the radiogenic Pb component is dependent on a variety of factors that makes a priori assessment of the suitably of material difficult without some form of pre-screening; high [²³⁸U], low [Pb], a range in U/Pb and a state of ²³⁴U/²³⁸U disequilibrium that is distinguishable from unity today.  It is no surprise that the conditions for U-Pb ages are more challenging to assess a priori.

This is a welcome exploration of ‘parameter-space’ for those considering U-Th-Pb methods for dating of Quaternary-aged carbonates and will promote forward-modelling approaches.

Specific comments

On the term  ‘double-dating’.

I accept that this term is becoming more widely used, but please acknowledge that U-Th and U-Pb methods are not independent – in time we will be measuring all isotopes in the decay chain to assess the state of disequilibrium from U to Pb!  The key here is that both U-Th and U-Pb protocols are being applied to sub-samples of calcite formed at the same time.

It would be appropriate to signpost similar strategies of combined or ?double-dating approaches from recent and deeper past (e.g.  U, Th – He vs U/Th; U-Pb vs U-Th-Sm/He). You might even include ESR – U-Th comparison studies (see some refs below).

Age consistency?

Is it justifiable to consider these as independent ages and use a simple  Z-test to assess the statistical difference?  I don’t know what the solution is here, but much of the uncertainty is shared (e.g. measured ²³⁴U/²³⁸U).    

You draw upon additional data (stable isotope variation) to demonstrate the accuracy (or consistency) between U-Pb and U-Th ages – i.e. tie-points and stratigraphic position, Is this data to appear in a future publication?  Is there comparison with the timing of Milankovitch forcing involved?  One must take the success of this strategy at face value because the data are not illustrated.

Age precision?

Section 4.2.  The  nature of  contribution to age uncertainty differs considerably between the two chronometers.  This is the crux of the paper, but at times the text needs reworking.  You describe here a specific case with low initial Th, were this to be more significant, there would be more similarity between the sources of uncertainty.

Use of the term ‘predictable uncertainties’ is awkward.    There are fewer ‘degrees of freedom’ in your U-Th Corchia case, where initial Th is minimal and uncertainties are dictated by solutions of the age equation towards its upper limit.   Your work is exploratory but does not offer definitive rules or solutions.

There is no assessment of the variation in initial ²⁰⁷Pb/²⁰⁶Pb..  or reported data.  It would be useful to make a comment here.  I presume the age calculations on not based on anchored common Pb.

On causes of variation in ²³⁵U/²³⁸U and ²³⁴U/²³⁸U.

You have recognised the potential complication of time elapsed since inheritance of the ground water signal in calcite when considering the relationship between ẟ²³⁴U (measured) and ẟ²³⁸U.   The controls on the source of U and range in U concentration in speleothems are poorly constrained and worthy of further exploration, and valuable comments are made here about the potential factors influencing the ẟ²³⁸U.

Technical comments

Line 26:  You refer to radiometric tools.  Add information here on other tools (astro,  geomagnetic and others).

Line 27. Utility determined by the material..  vague – expand.

Line 34: Middle Pleistocene (ca. 400-650 ka).  Please consider the official term and dates associated with GSSP etc.  see ref Head (2021) and elsewhere.  Maybe you could refer to Chibanian (Middle Pleistocene).

 Line 44:  Is it better to declare that  ²³⁰Th activity approaches secular equilibrium with parent nuclides?

Line 45:  The U-Th technique does not impose a limit of 650 ka, this is determined by the measurable extent of disequilibrium in the uranium-series decay chain.

Line 49: delete ‘currently’

Line 53: Consider using the following…  ‘suited to dating older material, for which sufficient time has passed for significant accumulation of radiogenic Pb, it is also suitable for middle Pleistocene material that has (²³⁴U/²³⁸U) activity ratios that are ……’    

Line 59:  delete ‘precise’

Line 110: delete ‘vertically’

Line 120 : suggests

Line 126:  consider ‘ to aid in the identification of’

Line 136:  ‘sub-sample taken from the centre of the isochron’ – rephrase.

Line 141:  Consider use of ‘U and Th isotopic analysis was performed’

Line 143:  You declare a range of sample masses, but constant U abundance.  Had you pre-screened the material for [U]?

Line 153:  Where were U and Th isotopic analyses conducted?  MN, Melbourne,

Line 160:  Is it the lower dynamic range or the operational upper limit of applied current that precludes use of standard gain calibration methods. 

Line 164.  Substitute ‘were’ for ‘was’.

Line 176. Please clarify this statement ‘corrected U ratios were normalised to CRM-112A’.

Line 179. Please clarify (or expand upon) the following    ‘For the purposes of comparison ..  a correction for initial ²³⁰Th was not applied’.

Line 194:  On the use of ²³⁵U/²³⁸U as internal mass bias.  What value of ²³⁵U/²³⁸U are you using?   Does this come from your U-Th analysis?

Line 202:  What do you mean by ‘the data itself’?  The rationale is probably embedded in Powell et al (2020) and Pollard et al (2023). Can you come up with a more useful phase here?

Line 362:  Declare the value of ²³⁴U/²³⁸U that you are using to standardise data in the text.  It is mentioned in Fig 6.  You might also refer to the UID associated with Li and Tissot.

Some useful references?

Makhubela TV, Kramers, JD (2022) Testing a new combined (U,Th)–He and U/Th dating approach on Plio-Pleistocene calcite speleothems. Quaternary Geochronology, 67,  101234. https://doi.org/10.1016/j.quageo.2021.101234

Bender M, Taylor F, Matthews RK (1973) Helium-uranium dating of corals from middle Pleistocene Barbados reef tracts. Quaternary Research, 3,  142-146, https://doi.org/10.1016//0033-5894(73)90060-4

Bender MJ, Fairbanks RG, Taylor  FW, Matthews  RK, Goddard  JG, Broecker WS (1979 )Uranium-series dating of the Pleistocene reef tracts of Barbados. Geological Society of America Bulletin, 90, 577-594

Radtke U, Grün R, Schwarcz HP (1988).  Electron Spin Resonance Dating of the Pleistocene Coral Reef Tracts of Barbados. Quaternary Research. 29, 197-215. doi:10.1016/0033-5894(88)90030-0

Pozzi JP, Rousseau L, Falguères C. et al. (2019) U-Th dated speleothem recorded geomagnetic excursions in the Lower Brunhes. Sci Rep 9, 1114 (2019). https://doi.org/10.1038/s41598-018-38350-4

Head MJ  (2021).  Review of the Early–Middle Pleistocene boundary and Marine Isotope Stage 19. Prog Earth Planet Sci 8, 50 https://doi.org/10.1186/s40645-021-00439-2

Pollard et al. conduct a systematic comparison of U–Th and U–Pb dating methods, focusing on their overlapping age range (∼430–630 ka) to evaluate their consistency and effectiveness for dating Pleistocene carbonates. The authors use 3 stalagmites from Corchia Cave, Italy, and adopt state-of-the-art analytical protocols, including i) U–Th measurement protocol that collects the low abundance 234U+ and 230Th+ ion beams in a Faraday cup fitted with a 10¹³ Ω resistor, and ii) U–Pb protocol using ‘stacked resin’ technique described in Engel et al. (2020) along with computing full U–Pb Tera-Wasserburg. 

U–Th and U–Pb dating methods are extensively used for radiometric dating of Pleistocene carbonates, offering some of the most accurate and precise ages for this period. These techniques will remain essential for establishing reliable chronologies of carbonates. The manuscript is well-written, with a detailed methodology, clear figures, and robust results. Therefore, I recommend its publication as it represents a valuable contribution to the geochronology community by providing the following key insights: 

• U–Th age uncertainties tend to increase in an approximately exponential manner over the time interval of ∼430–630 ka. On the other hand, U–Pb age uncertainties are influenced by uncertainty in the isochron slope, which is in turn largely controlled by the U/Pb spread of data points along the isochron, and the availability of highly radiogenic material.
• The U–Pb method (via ID MC-ICP-MS) achieves precision surpassing that of U–Th method prior to the latter reaching its upper age limit. The authors acknowledge that the exact overlap point will vary depending on sample types and study sites.
• The study highlights the importance of carefully considering 238U/235U values during data processing and age calculation. By using a high precision all-Faraday-cup protocol for U–Th measurements which permits measurement of accurate and precise 238U/235U values appropriate for assessing natural 238U/235U variability, the authors address the potential consequences of 238U/235U variation in speleothems. They find  that speleothem 238U/235U values are often significantly lower than the commonly used value of 137.88.

Specific comments 

Line 6: thousand of years (ka)

Line 28: applications 

Line 48: worth mentioning briefly why U–Pb chronometer is currently less widely applied than U-Th for this time interval. What does UoM stands for, I assume University of Melbourne? 

Line 245: The U-Pb age of sample CCB-6-1 seems to be reliable based on good spread in U/Pb ratios along the isochron and data point scatter. However, the discussion on the synchronicity of the stable isotope profiles among the 3 stalagmites lacks sufficient context. The authors should elaborate on the basis of this agreement, and provide some more context about the good agreement between isotopic variations amongst coeval speleothems from Galleria delle Stalattiti. 

Line 260-263: the MSWD and p values listed here for the best-fit regression line and for goodness-of-fit of the paired ages to a 1:1 line are different than the values listed in Figure 3. Please revisit either the text of the figure to reflect the correct values. 

Line 274: I think the text here refers to figure 4, not 5. Please correct. 

Line 313: specify for the reader that this Figure S4 in the Supplemental Material 

Line 381: include here the references of Woodhead et al (2006) and Stirling et al (2007) for the speleothem MO-1/3 and KOZ. 

Line 387: see Supplemental figure S5 

Line 388: worth mentioning what is the value for this correlation in Stirling et al (2007).

Figures and tables

Table 1:  note in the table legend that CI stands for confidence interval 

Figure 1: it is difficult to see the grey shading indicating the 95% confidence interval of the black line; consider using a darker shade.

Figure 2:  in the legend, specify what MSWD stands for and that n is for number of samples.

Table 2: this table is not referenced in the manuscript. In the table legend, what does “3 Assuming Guassian distributed analytical uncertainties” refers to? Number 3 is not listed in the table.

Figure 6: For consistency in the figure legend and enhance readability, I suggest including the following information:

• ”consensus value in geochronology”( Steiger & Jager (1977)
• average terrestrial zircon (Hiess et al (2012)
• average crustal rock (Tissot and Dauphas) 2015
• average oven water (Tissot and Dauphas) 2015

In the figure caption replace “Li et al” with “Li and Tissot”