Detrital Zircon U-Pb Ages in the East China Seas: Implications for Provenance Analysis and Sediment Budgeting

Round 1

Reviewer 1 Report

Dear authors,

 

the manuscript "Detrital zircon U-Pb ages in the east China seas: Implications for provenance analysis and sediment budgeting" is a very concise description of a very common problem in provenance analysis. Combined with the impressive data set I highly recommend this manuscript for publication after some minor revisions:

 

• please check for the language, there are several little things, I marked some of them
• sometimes there are +/- rapid changes in the catchments or the discharge that can cause variation in the detrital zircon signal; they can be possible triggers for some of the phenomena you described. An additional reference you could consider is:

Blum, M., et al. (2018): Allogenic and Autogenic Signals in the Stratigraphic Record of the Deep-Sea Bengal Fan. Scientific Reports 8, Article number: 7973.

• could you please add some lines about the limits of your model when using only small numbers of best ages, because your spread concerning these numbers is quite high (60-267); so how reliable is the model when using 60 vs. 267 best ages? are there any differences/no differences?
• please also find some more remarks in the annotated PDF

Sincerely,

the reviewer

Comments for author File: Comments.pdf

Author Response

Dear reviewer,

 

We are grateful for your comments that are very helpful to improve our work. The response to your comments is listed below.

 

Yours sincerely,

 

Xiangtong

 

 

• please check for the language, there are several little things, I marked some of them.

Response :Thanks for your help to improve the manuscript. I have corrected the places as you commented and change some figure colours to make them discernible.

 

• sometimes there are +/- rapid changes in the catchments or the discharge that can cause variation in the detrital zircon signal; they can be possible triggers for some of the phenomena you described. An additional reference you could consider is:

Blum, M., et al. (2018): Allogenic and Autogenic Signals in the Stratigraphic Record of the Deep-Sea Bengal Fan. Scientific Reports 8, Article number: 7973.

 

Response: I fully agree with you that rapid changes in the catchments could lead to variation in the detrital zircon signals. The variations also depend on the spatial and time scales. In our study, as we focused on very large catchment scales, this means the nature might help to smooth out allogenic signals and generate more homogenized autogenic signals. However, we think that the homogenization process in each catchment should be different and that’s why we can observe distinct U-Pb age distributions in different catchments. In terms of time scale, our scale is shorter than that of Blum et al. (2018). This means it is difficult for us to observe very large provenance changes as Blum et al. (2018) did. Nevertheless, we consider this study is significant as it provides a present reference to refer for long-term provenance studies in the East Asia continental margin seas. I am grateful for your recommendation of the very interesting work by Blum et al. (2018) and we added it in our references and hopefully it might help us to come up some new ideas for future study.

 

could you please add some lines about the limits of your model when using only small numbers of best ages, because your spread concerning these numbers is quite high (60-267); so how reliable is the model when using 60 vs. 267 best ages? are there any differences/no differences?

 

Response : We did add some lines to indicate the limits of our model, which include uncertainty of source signals, small number of best ages, etc. To minimize the influences of sample differences, we used the normalized KDE approach, which is believed less impacted by the sample numbers and it has become a practical way to compare samples with different age numbers. This point was also concerned by another reviewer. To make it clear, we added the source signal plots in Figure 3 (new) and the mixing coefficients in Table 2(new). In terms of the differences in the best ages, both in source and sink signals, we think they are influenced by the differences in sample numbers (n), and a possible way to solve the problem ultimately is to conduct so called large-n (n>1000) analysis as suggested some scholars (e.g., Pullen et al., 2014). However, we believe this is still challenging as it is generally limited by machine time to conduct such kind of analyses on a lot of samples and thousands of analyses. Although Matthews et al. (2017) had introduced a fast ablation protocol (15s per sample), it has not been widely used as far as I know. To realize this goal, some problems are needed to be solved, including the manual selection of laser spots, data reduction, common lead correction and best age filtering.

 

Reviewer 2 Report

This paper deals of sediment provenance near Chinese coast through U-Pb zircon ages in sediments. The collected data are very good and the paper was well written. The sections 1 and 2 are very persuasive, while the sections 3.3; 3.4; 4.2 must be better explained.

The crucial node of this work is the use of DZmds of Mathlab to calculate the similarity of age data distributions in source area and in depositional basins. The data distribution of source area derived from literature but no summaries are showed in the manuscript. The input data are unknow.  The pre-conditions for validating conclusions are lacking. The reader must trust the conclusions without any control of the input data. This initial problem conditions the validity of subsequent deductions. The figure 5 shows the results of comparations but the MDS configuration are not explained. If the aim of this paper is “the combination of the U-Pb zircon geochronology with mixing models” the authors must explain the application of this method.

 

Captions of Figures 4, 5,6 and 7 lack details.

A table with calculated mixing coefficients could be inserted.

The procedures of calculations and the input data are essential.

 

3.3. Multi-dimensional scaling (MDS)

The MDS results must be explained; what is the output of application? What are the assumptions? How the program works?

3.4. Mixing model of detrital zircon U-Pb age distribution

This section must be explained with major details. The calculations are obscure and the only parameter is R² >0.75. The assumptions to calculate the model are not showed.

 

5.4. Sediment budgeting

The budget of sediments deposed in the different seas has been calculated assuming the mixing coefficients but the reader does not control the previous data.

Some suggestions in the attached file.

 

Comments for author File: Comments.pdf

Author Response

Dear reviewer,

 

We are grateful for your comments that are very helpful to improve our work. We did a lot of modifications on our manuscript as you suggested by: 1) adding a new figure of source signals (Figure 3) ; 2) adding a new table of mixing coefficients; 3) adding more details to Figures 5,6,7 and 8, which were Figures 4, 5, 6 and 7 in the previous manuscript.

 

The response to your comments is listed below.

 

Yours sincerely,

 

Xiangtong

 

.

 

3.3. Multi-dimensional scaling (MDS)

The MDS results must be explained; what is the output of application? What are the assumptions? How the program works?

Response :

Thanks for your comments on this very important aspect and we made some improvement in Section 3.3. To avoid repetition, we discussed the MDS results in Section 5.1. This MDS is a standard statistical method but was introduced to provenance analysis by Vermeesch (2013). The general assumption is to compare the configuration that determines both the distances and the fit to the disparity transformation between all possible sample pairs. There are a lot of complicate mathematical equations to get the output, which is far beyond the scope of this paper. However, our concern is the output of the method, which is a plot or map that is used to visualize the comparison among different samples. As viewed from the MDS map one can easily judge which samples are more similar by the first and second distant neighbors.

To run this program, DZmds, we input the ages from our samples and the river samples in the literature and chose the kernel density estimates (KDEs) at a bandwidth of 25 Myrs to perform a metric cross-correlation comparison to minimize the differences of the best ages among samples and river sources. The goodness of fit is suggested by a Shepard stress of 0.16, which is a fair result according to rules of thumb. When we run this program, we had tried the non-metric comparison or 3-dimension MDS, which could give a better Shepard stress (0.08); however, the outputs of the MDS were similar to the map generated by the metric comparison. For this reason, we did not use the non-metric MDS output or the 3d output. Another reason is that we think the results are good enough to distinguish the dissimilarity among these distributions when the wide range of our samples and the potential sources is considered.

For more details of the MDS results and discussion please refer to Section 5.1.

3.4. Mixing model of detrital zircon U-Pb age distribution

This section must be explained with major details. The calculations are obscure and the only parameter is R² >0.75. The assumptions to calculate the model are not showed.

Response: 

To make this section clear, we added a new figure containing the river source signals and a table with all outputs, including R2 and mixing coefficients. we performed the mixing modeling by using the KDEs shown in Figure 3 as the defined endmembers and the KDEs of our samples as the mixed daughter distributions (Eqn. 1). The model was solved by parametric EMA (Endmember analysis) as it fits the true mixing regime better than nonparametric EMA (Paterson and Heslop, 2015). In our study, the parametric EMAs are defined according to the river KDEs, which might be helpful to minimize the uncertainty on the model’s ability to distinguish source distributions. I fully agree with you that the source inputs are very important to the modeling, and when better source signals are available in the future, some of our results might be updated.

5.4. Sediment budgeting

The budget of sediments deposed in the different seas has been calculated assuming the mixing coefficients but the reader does not control the previous data.

Response:

Thanks for your suggestion on this key point, we will present the mixing coefficients in a new Table (Table 2).

Reviewer 3 Report

The paper is good. I noticed only some grammar imperfections that should be checked by a native English-speaker. Colours in some figures are not easy discernible. All my notes are marked in PDF. If there are some highlighted expressions without comments, they mean that I was not sure if they are correct or not.

Roman Aubrecht

Slovakia

Comments for author File: Comments.pdf

Author Response

Dear Roman Aubrecht,

 

We are grateful for your comments that are very helpful to improve our work. We changed some colours in the figures to make it more discernible and checked the highlighted to make it clear.

 

 

Yours sincerely,

 

Xiangtong

Round 2

Reviewer 2 Report

The revised manuscript is much better than the previous one. The modifications (especially new Table2 and figure 3) in the test and captions of figures, improve the impact of paper and the deductions are supported by data. Now the work is ready for an international audience.