the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
A sub-fossil coral Sr/Ca record documents meridional variability of the Intertropical Convergence Zone in the eastern Indian Ocean
Abstract. Sea surface temperature (SST) variability in the south-eastern tropical Indian Ocean is crucial for rainfall variability in Indian Ocean rim countries. A large body of literature has focused on zonal variability associated with the Indian Ocean Dipole (IOD), but it is unclear whether meridional shifts in the position of the Intertropical Convergence Zone (ITCZ), which at present co-vary with the IOD, may also occur independently. We have developed a new, monthly resolved Sr/Ca record from a sub-fossil coral cored at Enggano Island (Indonesia, 5° S, 102° E). Core sections containing diagenetic phases are omitted from the SST reconstruction. U/Th dating shows that the Sr/Ca-based SST record extends from 1917–1868 and from 1861–1823 with a relative age uncertainty of ±2.4 years (2σ). At Enggano Island, coastal upwelling and cooling in austral spring is coupled to the position of the ITCZ, and impacts SST seasonality. The sub-fossil coral indicates an increase in SST seasonality due to enhanced austral spring cooling between 1917 and 1855, which we attribute to stronger SE winds and a northward shift in the position of the ITCZ in austral spring. A nearby sediment core indicates SST cooling and a shallowing of the thermocline prior to ~1930. These results are consistent with an increase in the North-South SST gradient in the eastern Indian Ocean calculated from historical temperature data, that is not seen in the zonal SST gradient. We conclude that the relationship between meridional and zonal variability in the eastern Indian Ocean is non-stationary and influenced by long-term temperature trends.
- Preprint
(4499 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 10 Jun 2024)
-
RC1: 'Comment on cp-2024-25', Anonymous Referee #1, 18 Apr 2024
reply
This study aims to understand how the meridional SST gradient in the eastern equatorial Indian Ocean has changed over the past ~200 years. To do so, the author developed a Sr/Ca record using fossil corals collected from Enggano Island for analysis. The authors found that there was an increase in seasonality strength and an earlier seasonal SST maximum between 1855-1917 compared to the modern period and 1823-1854 CE and attributed the change to an earlier onset of austral spring and strengthened SE winds, which consequently imply a northward shift of ITCZ and a stronger meridional SST gradient. On the other hand, there is no conclusive evidence that changes in zonal SST gradient (e.g., IOD) played a role in changing SST in Enggano Island and meridional gradient despite previous suggestions. Therefore, they concluded that the meridional SST gradient and zonal SST gradient are not always coupled and require more analyses on the mechanisms that drive the meridional SST gradient.
I think this new record is a useful contribution to better understand climate variability in the Indian Ocean and complement existing records in nearby areas. I also think the analysis of meridional SST gradient in the NE Indian Ocean is interesting. The screening for coral quality/diagenesis is also extensive. That said, I have some suggestions/comments, that I hope would help improve the manuscript. There are also several inconsistencies within the text and figures/tables, which needs to be corrected. Otherwise, it is difficult to judge the results and conclusion of this study.
Overall comments:
- I find referencing the years in reversed chronological order (e.g., 2008 to 1930 instead of 1930 to 2008) confusing. There are also several instances where the years are referred in chronological order instead, for instance L317 “Between 1854 and 1923…”. I suggest making this consistent throughout the text, and preferably in chronological order (i.e., older to younger).
- Within the manuscript, there are multiple instances where statistical significance is mentioned (e.g., 99% confidence levels). However, in most cases, it is unclear how this was determined. Even when a Monte Carlo approach was used, it was also unclear how it was carried out. It would be helpful if this can be clarified.
- The results of this study hinge on an accurate chronology and constraints on the annual cycle. While there are multiple instances within the manuscript where dating uncertainty was mentioned, as far as I can tell, there are no testing of how robust the results were against dating uncertainties. It would be nice to see sensitivity tests to check this. Furthermore, it is currently unclear how the annual cycle was derived (see the comment L181-183 below for more details), which makes me unsure of the results. Additionally, the chronology is derived based on the assumption that the internal chronology and U/Th ages have a 1:1 relationship. While in Figure A6, it shows that they correspond to each other fairly well, there are also instances where the U/Th diverges from the 1:1 line (the 2sigma of KNFa(8/11), KNFa(5/11) do not line up with the regression line). So, I wonder how accurate this assumption is in this case? Moreover, it would be great to provide the estimated extension rates of these corals, just so we can make sure it is indeed possible to estimate monthly SST changes.
- Given that the change in SE wind strength, and the shift in ITCZ are supposed to correspond to changes in the South Asian monsoon, I wonder if it will also be helpful if you can show the South Asian monsoon also changed concurrently with these changes.
Specific comments:
L31-32: A ‘reference period’ needed to compare with in order to claim ‘an increase in SST seasonality due to enhanced austral spring cooling’.
Figure 1: In most cases, the figure is referred in the text when discussing about anomalies during IOD events. So, I wonder if it would be better to change the subplots b-e into OLR and SST anomalies so that they can better serve their purposes?
L111-L112: It is unclear to me how Aug-Oct temperature and symmetry is inferred based on Figs 2-3 – both do not show mean Aug-Oct temperatures.
Figure 3: I suggest checking the consistencies between this figure, the caption, and the main text on which months are mentioned and used for analyses. In the text and the caption, Aug-Oct was mentioned, whereas the figure label suggests Sept-Nov. Additionally, the vertical dashed lines don’t seem to be located at the same months for the plot of each location in subplot c.
L154-155: I think the sentence is missing a verb (e.g., were carried out’).
L181-183: This is actually *inconsistent* with Cahyarini et al. (2021) and Pfeiffer et al. (2021). Both studies tied September to Sr/Ca maxima and May to Sr/Ca minima. But here, it suggests only September was tied to Sr/Ca maxima. Please clarify which way it was done.
L188-195: Fig A6 should probably be referred somewhere here so the readers can go to that figure to get a better sense how this was done.
L233: ‘were’ -> where
L237: I think the Appendix/Supplementary figures referenced here are incorrect.
L311: repeated ‘(Fig. A8)’.
L312-320: Should ‘1854 and 1923’ be 1854-1823 instead? Otherwise, this will be referring to the same overlapping period as the previous sentence (1917-1855). Additionally, I would like more quantification on the comparisons between the distributions instead of simply relying on visualizations. Tests such as a Kolmogorov Smirnov test (or its variant) would be helpful here.
Figure 6: I only see one type of line in subplot (a) with two green and blue lines. I do not see red solid and dashed lines.
L350-354: I wonder if there’s a more objective way to ‘separate’ these time periods? Right now, it seems a bit arbitrary and relies on visualization. One suggestion perhaps would be to analyze the wavelet power of annual cycle and identify periods that are weaker to a ‘reference period’. This can be pulled out from the wavelet spectra.
Figure 7: Why is the seasonal cycle shown here span from July to January 2 years later? There are several mistakes in the caption (e.g., ‘dashed green lines’ for core PB, no explanation of ‘dark grey and dashed lines’ in subplot a). Additionally, which ‘modern’ (coral or observation) is used in (d) and (f)?
Figure 8: in subplot c, there is a discrepancy on which years were used for analysis: it was labelled 1917-1869 in the figure whereas in the caption 1917-1855 was referred.
Figure 9: I don’t quite understand what “not the large 95% confidence levels of the SiZer test” means. It would also be helpful to explain what the horizontal lines associated with the change point indicators mean. Additionally, there needs to be clarification on how the SiZer analysis was carried out. As far as I recall, SiZer applies a range of Gaussian filters with varying bandwidths and calculate the trends based on those bandwidths. But in the caption, it only mentions about a 21 year running average to show the data and not any other information related to SiZer.
L440: I am not sure if age uncertainty is the main issue here. By just looking through the KNFa record, there is no ‘major’ anomalies (<-4C) between 1869-1917. Given that the absolute dating uncertainties are almost <10 years, I don’t think age uncertainty is an issue here. In fact, the remaining paragraph does not discuss about age uncertainty. So, I suggest modifying the first sentence of this paragraph.
L444-446: I don’t think this is an accurate statement. My understanding is that the Enggano record is better in capturing meridional SST gradient compared to previous Mentawai records. So, it is logical that the Enggano record might not detect IOD changes. That said, a comparison between meridional and zonal SST gradients can be achieved by making use of the Enggano record for meridional changes and the recently southern Mentawai record (Abram et al., 2021 Nature) that is supposed to record IOD changes.
L456: hereinafter, ‘foraminifers’ should be ‘foraminifera’.
L461: Would (p>0.05) mean it is not significantly different?
L461-462: I think it would be helpful to test statistically if the relative abundance of thermocline dwelling forams changed significantly.
L464: ‘linkedn’ -> linked
Figure 11: I wonder if the abundance of mixed layer forams timeseries is needed, since it was never discussed in the text?
Table A1: What do those asterisks next to KNFa(3/11) and KNFa(7/11) mean?
Figure A6: I don’t think the equation displayed here is correct. Plugging in any years prior to 1934.1 will result a negative y, which isn’t supposed to happen here.
L601: As far as I can tell, I don’t see any U/Th ages that correspond to 1823 both in the figure and in table A1.
L619: ‘(a)’ should be (b) here.
Citation: https://doi.org/10.5194/cp-2024-25-RC1
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
142 | 31 | 12 | 185 | 7 | 6 |
- HTML: 142
- PDF: 31
- XML: 12
- Total: 185
- BibTeX: 7
- EndNote: 6
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1