the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Jun 2026
Precision measurement of δ234U in annually banded tropical corals
Abstract. Tropical corals preserve geochemical 238U/234U ratios that provide valuable records of past seawater uranium isotope compositions. Variations in the coral skeletons and thus seawater are likely indicative of freshwater contributions from submerged groundwater discharge and river runoff, or reflecting coral diagenesis. Advances in multi-collector ICP-MS allow precise determinations of 238U/234U ratios, enabling the reconstruction of subtle (typically >1 ‰) environmental changes in these marine records. In this study, we evaluate the reliability of coral-based δ234U records across multiple genera, sampling strategies, and intra-skeletal variability. Analyses of reference material NBS-CRM-112A demonstrate reproducibility within ±0.4 ‰. Replicate sampling across coral structures and colonies indicates measurable intra-band heterogeneity (±0.6 ‰), although local hydrodynamics and submarine groundwater discharge can introduce small inter-colony offsets. No species-dependent isotope fractionation was detected, underscoring the robustness of δ234U as a geochemical proxy. These findings demonstrate that coral skeletons provide reliable archives of up to sub-annual δ234Usw for detecting subtle climatic and hydrological signals.
Competing interests: One of the coauthors is a member of the editorial board of Geochronology.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(807 KB) - Metadata XML
-
Supplement
(579 KB) - BibTeX
- EndNote
Status: open (until 08 Sep 2026)
- RC1: 'Comment on egusphere-2026-3519', Anonymous Referee #1, 07 Aug 2026 reply
-
RC2: 'Comment on egusphere-2026-3519', Anonymous Referee #2, 30 Aug 2026
reply
The manuscript presents a potentially useful methodological assessment of high-resolution δ234U measurements in annually banded tropical corals. The combination of analytical replicates, full procedural replicates, intra-skeletal comparisons, different sampling strategies, multiple coral taxa, and neighboring colonies provides a valuable dataset for evaluating the applicability of coral δ234U as a proxy for seawater uranium isotope variability. However, several important issues need to be resolved before the manuscript can support its main conclusions. In particular, the analytical framework needs to be more carefully reconciled with the recent methodological advances of Hu et al. (2025). This is especially important because the manuscript aims to interpret δ234U variations at the ~0.5–1‰ level, whereas Hu et al. demonstrated that choices in tailing correction and decay constants can introduce systematic offsets of approximately 0.2‰. In addition, the current manuscript does not consistently distinguish analytical precision, external reproducibility, procedural reproducibility, and natural skeletal heterogeneity. Consequently, some statements concerning the detection of sub-permil environmental variability and the absence of species effects appear stronger than warranted by the data. I therefore recommend major revision before acceptance, given that the authors can address my comments below:
Major points:
1) The analytical accuracy and tailing correction need to be quantitatively reconciled with Hu et al. (2025). This is, in my view, the most important issue. The manuscript states that the low-mass tail of 238U is reconstructed using a PCHIP interpolation based on half-mass measurements and explicitly contrasts this approach with Hu et al. (2025). However, Hu et al. demonstrated that the shape of the uranium tail in the vicinity of mass 234 cannot necessarily be inferred accurately from conventional interpolation between adjacent half masses. Their Th-analog approach directly constrained the tail at the position corresponding to 234U and showed that an inappropriate half-mass correction can bias measured 234U/238U ratios by approximately 0.22‰ near secular equilibrium and ~0.20‰ for modern seawater. The PCHIP approach may indeed reproduce the true tail better than a simple exponential interpolation, but this is currently asserted rather than demonstrated at the precision relevant to this manuscript. Because the environmental signals discussed here are commonly only ~0.5–1‰, a potential systematic bias of ~0.2‰ is not negligible. I strongly recommend that the authors quantitatively demonstrate that the PCHIP correction provides an unbiased estimate of the tail specifically at m/z ~ 234. This could be achieved, for example, by directly comparing PCHIP-derived corrections with the Th-analog approach of Hu et al., or by analyzing an independent secular-equilibrium material and/or reference solutions spanning substantially different δ234U values. At minimum, the manuscript should quantify the difference expected between the PCHIP and Th-analog approaches and propagate this uncertainty into the environmental interpretations.
2) The normalization strategy and the claim of absolute accuracy are currently unclear and potentially circular. The manuscript appears to use two reference materials in different roles. The Methods state that measurements were bracketed using HU-1, whereas later the authors state that “all isotope measurements were normalized to NBS-CRM-112A.” This needs to be described much more explicitly. In particular, if CRM-112A is used to define the normalization of the measured isotope ratios, then the observation that 66 analyses of CRM-112A yield a mean of −38.5‰ cannot independently demonstrate analytical accuracy. It would mainly demonstrate precision/reproducibility around the assigned normalization value. Yet the Discussion currently uses these 66 measurements as the principal evidence for accuracy. The HU-1 results may provide an especially useful independent check. The manuscript states that its HU-1 corresponds to approximately −0.9‰ relative to secular equilibrium. Hu et al., however, obtained approximately −1.25 to −1.27‰ for HU-1 using the revised method. The ~0.3–0.4‰ difference is comparable to the entire external reproducibility claimed in this manuscript and should therefore be reported and discussed rather than overlooked.
3) The use of the Cheng et al. (2013) λ234 value should be reconsidered after Hu et al. (2025). The manuscript calculates δ234U using λ234 = 2.82206 × 10-6 yr-1 from Cheng et al. (2013). Hu et al. (2025), however, revised the 234U half-life and corresponding decay constant after identifying a systematic bias in previous 234U tail corrections. Given that Hu et al. is already cited throughout this manuscript, it is difficult to justify retaining the previous λ234 without discussion. The difference between the old and revised λ234 corresponds to approximately 0.2‰ on the activity-ratio scale, which is small for traditional U-series applications but not small relative to the <1‰ variations emphasized here.
4) Analytical precision, external reproducibility, procedural reproducibility, and natural heterogeneity are not consistently distinguished in the current manuscript. For example, internal uncertainties are reported as 2σ standard errors, whereas CRM-112A reproducibility is described using a standard deviation of ±0.4‰; Figure 2 additionally reports a 2σM of ±0.06‰. These quantities represent fundamentally different aspects of uncertainty and should not be compared interchangeably. The manuscript concludes that δ234U variations exceeding 1‰ can safely be interpreted as seawater variability, and elsewhere states that variations <1‰ can be detected. These statements are difficult to reconcile quantitatively with the reported replicate variability.
Minor points
- Abstract, “238U/234U ratios” should be “234U/238U ratios”
- Two sections are labeled 2.4, “ICP-MS concentration and calcium assessment” and “MC-ICP-MS U isotope measurements”. The latter should be Section 2.5.
- Figure 1, the text reports a maximum measured concentration of 174.4 ppb, whereas the figure caption states 128.40 ppb. These numbers should be reconciled.
- The statement that analytical error “decreases exponentially” with uranium concentration should either be supported by an explicit fitted exponential relationship or replaced with a more descriptive formulation such as “decreases strongly/nonlinearly.”
- Please define consistently whether reported uncertainties represent 2SD or 2SEM. Every figure caption should explicitly specify the uncertainty represented by error bars.
- Figure 2 should clearly distinguish the “standard deviation of individual analyses (external reproducibility)” from the uncertainty of the mean (2σM = ±0.06‰). The very small uncertainty on the mean should not be described as the analytical reproducibility of an individual coral measurement.
- Table 3 should report not only Pearson correlation coefficients, but also the number of paired observations and exact p values for each comparison. Because the time series are autocorrelated, the authors should also consider whether nominal Pearson p values overestimate significance.
- The statement that “the temporal resolution is limited solely by sample availability” is too strong. Temporal resolution is also limited by growth rate/band thickness, sampling geometry, age-model uncertainty, U concentration, required U mass, and analytical precision.
- The statement that concordant δ234U among species “confirms the lack of fractionation observed using the 238U/235U ratios” should be softened. Agreement in 234U/238U does not directly constitute confirmation of observations made using a different uranium isotope ratio.
- Please clarify the exact role of the 233U–236U TriSpike. The Methods describe the spike composition but later state that instrumental mass bias was corrected using natural 235U/238U. It would be useful to state explicitly which isotope ratios are used for concentration determination, spike correction, and mass-fractionation correction.
- The manuscript would benefit from explicitly reporting the “234U signal intensity”, rather than only total U concentration. Hu et al. show that uncertainty in the 10to13 Ω amplifier baseline and tail correction directly controls the achievable precision of the small 234U beam. This would make the analytical comparison much more meaningful than solution concentration alone.
- There are several typographical and grammatical issues that should be corrected, such as duplicated words (“the theca”), “corals species,” “reproducability,”, etc.
Citation: https://doi.org/10.5194/egusphere-2026-3519-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 68 | 30 | 14 | 112 | 15 | 9 | 11 |
- HTML: 68
- PDF: 30
- XML: 14
- Total: 112
- Supplement: 15
- BibTeX: 9
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Greve et al. review
General comments
This paper reports sampling and analytical protocols for conducting high precision δ234U analyses (c. ±0.4 ‰) on tropical, shallow water corals. The authors then investigate variations in δ234U in modern and near-modern corals to determine their utility as environmental and hydrological archives.
Coeval samples of twelve species of coexisting corals are found to faithfully record analytically indistinguishable values of seawater δ234U. Small but resolvable variations in δ234U (up to c. 1‰) are observed in growth bands of corals sampled at sub-annual scales. Larger variations in δ234U (up to several ‰) were observed in decadal-scale time series of coral samples from an area of known submarine groundwater discharge (SGD), with distinct values of δ234U recorded in neighboring coral colonies. The authors conclude that corals provide reliable archives of δ234U for detecting subtle climatic and hydrological signals. Much of this is unsurprising but nonetheless it will be useful to those working with coral archives to incorporate these carefully measured results into their thinking.
Line-by-line comments follow, with specific comments followed by technical corrections.
Specific comments
Line 24. On glacial-interglacial timescales, variations in seawater 238U/234U are expected (as the authors are aware) and such variation should be mentioned here.
Line 52. Secular variation in the past 234U/238U ratio of seawater on the multi-ka timescale is now well documented (e.g., Chutcharavan et al., 2018) so not all pristine corals are expected to yield initial 234U/238U activity ratios of ~1.45. The authors are clearly aware of this given later sentences, but the statement starting on line 52 needs to be corrected.
Line 55. The phrase “U-series open system behavior…in well-preserved corals…” is self-contradictory; please rewrite.
Also, since young corals have d234U ≈ 450 versus the d234U of CRM-122A ≈ -38.5, it’s important to test that the normalization based on CRM-122A is appropriate for the coral analyses. The authors have done so using their seawater standard, although their results are not mentioned until line 228. It would be helpful to expand line 194-195 to include words to the affect that “The reproducibility was also assessed with repeated measurement of an in-house seawater standard, which yields a mean δ234U value of 145.7‰ with a standard deviation of ±0.3‰, thus comparable in precision to NBS-CRM-112A”, with appropriate repositioning of call outs for Figs. 2 and 3.
Line 189. The authors state that “all isotope measurements were normalized to the NBS-CRM-112A standard”. The authors should briefly describe how this was done and how large the normalization factors were.
Line 206. Please specify the signal levels (in amps or volts with respect to a 10E11 ohm amplifier) typically obtained for 238U in Fig. 1 or state your typical sensitivity in volts/ppb. Such signal levels are more useful than concentrations for assessing mass spectrometer performance and analytical uncertainties. Also please give the approximate amount of 238U consumed during typical analyses.
Line 217. (a) The certified value of CRM-112A should be added to Fig. 2 to illustrate this statement.
Lines 217-219. (b) The authors state that “The methodology applied here allows theoretically for ε-level precision for δ234U measurements …”However, this statement is not supported by the data presented since the scatter observed for replicates of CRM-112A is too high to support ε-level precision. It would be appropriate to omit this statement unless it can be supported by measurements.
The authors go on to state that although the instruments used in this study and that of Hu et al. (2025) are similar, runs of impractical length (c. 3 hours) would be required to achieve ε-level precision with the author’s protocol. But there are some salient differences in the measurement and data reduction protocols in the two studies. For example, Hu et al apparently use much higher intensity 238U beams — ~400 volts versus mostly ~80 volts or less in the current study (judging by Fig. 1). Other things equal, the large beams used by Hu et al provide a 5X improvement in signal/noise ratios on Faradays. This may help explain why excessively long runs would be necessary to attain ε-level precision with the author’s protocol. Another notable difference in the two studies is how corrections for peak-tailing on 234U are made, with measurements at distinct e/m positions and different curve-fitting approaches being used. It’s possible that larger samples are not the sole reason, or even the main one, for the higher precision reported by Hu et al. for CRM-112A and other materials.
In any case, as the authors point out, an external precision of ~ ±0.3–0.4‰, which they demonstrate for CRM-112A, their in-house sea water standard, and replicate coral analyses is adequate to assess variable 234U/238U in corals at the c. 1‰ level, which is the topic of this paper.
Line 226. It’s hard to reconcile a “a total measurement duration of 4 minutes” with an “integration time of 4 s and 60 scans”, which to me implies an on-peak measurement time of 240 s. Please clarify.
Lines 235-238. How do the authors interpret the larger deviations observed for three samples out of 20?
Lines 240-242. Do the authors think they have detected true intra-band variability in the TM2020 coral?
Lines 356-375. This section could be reduced by 50% or more. The main point — that SGD, already a known feature of the region, provides a plausible explanation for the differences in the d234U time series of the two corals — can be stated more succinctly.
Technical Corrections
Line 22. Can delete “geochemical”
Line 27. Add “characterize”; i.e., characterize intra-skeletal variability.
Line 28. Specify confidence level of quoted reproducibility.
Line 34. “Elemental and isotopic compositions…” instead of “Geochemical element and isotope composition…”.
Line 35 “Coral-based”, not “Coral based”. Delete “over”.
Line 42. “freshwater discharge from rivers and submarine groundwater may locally modify marine 234U/238U ratios.”
Comment: chose either 234/238 or 238/234 and then be consistent. Since d234U is defined in terms of the 234/238 ratio, that is probably the better choice.
Line 44. Delete “today”.
Line 47. Sentence beginning with “Lastly…” is vague. Can you be more specific or omit?
Line 49. Should read “absence of, or accurate correction for, initial 230Th.
Line 61. Suggest starting a new paragraph here, where you switch to discussing methods.
Line 86. Are the “skeletal structure replicates” sampled from the same growth band at different positions on the colony? If so, please so state.
Line 89. Delete “and” in “and past seawater…”
Line 169. “interspersed” rather than “interspaced”
Line 170, 171. Ohm is not capitalized when used a unit of measurement. Also, it seems a word is missing after electronic; e.g., electronic source…
Line 262. That these samples are coeval should be stated right away; i.e., Eleven coeval samples….
Line 304. Minor or subtle, not both.
Line 320. “high resolution uranium isotope…”
Line 380. No capital needed for uranium.