Investigation into Groundwater Resources in Southern Part of the Red River’s Delta Plain, Vietnam by the Use of Isotopic Techniques

Round 1

Reviewer 1 Report

This manuscript summarizes a geochemical investigation of the groundwater system within the southern Red River Delta plain in Vietnam.  The aquifer system consists of a Holocene age unconfined aquifer which is separated from the underlying Pleistocene, Neogene, and Triassic aquifers by a confining unit.  The authors found that the Holocene aquifer is rapidly recharged by rainwater and river water but is impacted by salt-water intrusion near the coast.  The underlying aquifers appear to be hydrologically connected (probably via numerous faults in the area) and contain very old (10,000 to 14,000yr) water that is slowly recharged in the higher elevation areas to the northeast.  Some areas of these aquifers also contain saline water, but this does not appear to be from an oceanic source.  The authors suggest it is trapped saline pore water from an earlier time period.

Overall this is a nice contribution on an important topic.  It is locally and regionally important for water managers in Vietnam, but also may be of interest to a larger audience interested in the plumbing of similar systems.  There are a lot of details and data to sort through and the authors do a good job of putting together a story that fits the pieces together.  I think they are on target with most of their interpretations, but there are a few of which I’m a bit skeptical.  In some cases the existing data collected could also be used to more effectively test several of the interpretations.  Below I outline some suggestions and comments that might be used to strengthen the submission further.

Minor points

1)      Figure 3 – I do not think this figure is needed to understand the rest of the paper. 

2)      Line 144-145: This sentence states that you are quantifying the fractionation of 13C in DIC due to exchange with biogenic CO2.  This is too limiting.  You are measuring the 13C in DIC to determine the source of inorganic carbon, which could be atmospheric CO2, dissolution of carbonates, or biological in origin.

3)      Page 5 (and to some extent pages 8-9):  There is quite a bit written here about how water and carbon isotopes work.  The content is accurate, but I rarely see such lengthy explanations of the foundational principles of isotopic systems.  If you are looking for a place to cut, you could get rid of much of this and simply cite the foundational literature.

More substantial points

1)      It would be good to see a summary of water levels for all the samples so we can calculate differences in vertical gradients. Figure 10 is great, because it shows the connectivity of the two aquifers.  However, these are the only hydraulic head data provided in the paper.

2)      Discussion of figure 5a in results section: The seawater mixing trend could also be explained equally well as an evaporation trend.  This would be the trend for rainwater that underwent evaporation prior to recharge.  This is common for recharge from some lakes, lagoons, swampy regions, possibly some rivers, etc.  Later in the manuscript you provide additional evidence of seawater infiltration, but this graph alone does not establish that as the only unique interpretation.  Another thought here would be to save the interpretation until later altogether.  It is not customary to provide interpretation in the results section – just the discussion section.

3)      Discussion of figure 5b in the results section: What water isotopic composition is expected for paleo-water and why?

4)      Figure 7: In some cases it is hard to corroborate the interpretations of the data as they are shown in the Piper diagrams.  I suggest creating the key cross plots needed to support the interpretations either instead of the Piper diagrams or in addition to them.  For example, if you are looking at waters that mix with seawater versus waters that evolve to a Na-bicarbonate concentration you should be looking at simple cross plot graphs like Na vs Cl or Na vs Ca, or Cl vs HCO3, or more complex but illustrative graphs such as Na/Cl vs 1/Cl, Na/Ca vs 1/Cl or Cl/HCO3, etc.  The simplicity of cross plots would make it easier to demonstrate relationships and help with interpretation.

5)      Figure 8: This is a key diagram, because it can “irrefutably” establish whether you do or do not have mixing with seawater.  Something to check on here is whether you need to have Cl on a log scale to establish a linear mixing model.  When comparing concentration data to isotope data it is common to use log concentration, as mixing is not always a straight line with raw concentration data.  In this particular case I am not sure, but it is worth checking because you don’t want to have the wrong percentages in the mixing model.  You use a log scale for the same parameters in Figure 9, so I’m not sure why you did not use the log scale for Figure 8. Finally, I agree with the interpretation here, but I am not sure where the local precipitation data point came from.  There is no local precipitation point in Figure 5.

6)      One of the interpretations in this paper that is not well-supported is the source of saline water in the Pleistocene and Neogene aquifers.  The authors suggest it is entrapped aquifer/aquatard deposits of saline water, but what evidence do we have to support this?  What is the expected isotopic composition of trapped water of Holocene age?  If it is saline, wouldn’t it be trapped ocean water and therefore have a seawater isotope signature very similar to the present day seawater?  The lack of information here represents an important gap.  Perhaps the concentration of Br relative to Cl may shed some light on the source of the saline waters.

7)      One major improvement would be to include the saturation indices for carbonate and other minerals in the aquifer system in a table.  This would help support several key interpretations regarding carbonate dissolution or precipitation (see below).

8)      Figure 9b shows that a group of very dilute (in terms of chloride) samples change their Na/Cl ratio substantially, which is attributed to the exchange of Ca in the water for two Na ions on clay.  This is the typical pathway suggested for the development of Na-bicarbonate waters.  However, the authors need to rule out the possibility that you have silicate weathering contributing the Na and Ca is simply stagnant or being tied up in carbonate precipitation. Alternately Na may remain constant and Ca is being lost from the system.  Having saturation index calculations and additional cross plots could address these possibilities.

9)      Page 17: the precipitation of siderite is used to explain the lack of correlation of Fe2+ and HCO3.  This is only one of many possible explanations, and perhaps the least likely.  Siderite needs to be substantially over-saturated to precipitate in most groundwater systems.  Other examples of things that could impact this relationship include the adsorption of Fe(II) to other minerals, precipitation of pyrite, precipitation of calcite or dolomite, etc.  One could shed more light on this with additional saturation index info and more cross plots.  One way to link the creation of Fe(II) to biological activity would be to plot F(II) versus d13C DIC.  It would also be illuminating to see borehole depth vs Fe(II) and Fe(II) versus sulfate.

10)   The d13C of DIC is not effectively used as part of the story.  Figure 11 seems rather “tacked on” like the only reason for the figure was to show the d13C DIC data because there seems to be no other need to talk about Mg/Ca ratios.  Frankly, I do not see any relationship between d13C DIC and the Mg/Ca ratios that can be supported from Figure 11.  Why not show the d13C DIC of the Holocene aquifer for comparison?  The effect of precipitation of carbonates is mentioned, but this is not supported by saturation indices.  I would also like to see plots of Mg/Ca ratio and 1/Cl or similar plots to better assess which fluids are interacting.

11)   Line 608 – here it indicates that the hydrochemistry is controlled in part by sulfate reduction but earlier (line 559) it states that sulfate is in low concentration and it was not discussed in favor of a discussion of Fe(II) and ammonia.

Author Response

First of all we would kike to express many our sincere thanks to you for your time spent to have a look at and make very valuable comments in order we could improve our manuscript. Generally, we accepted all your suggestions to supplement more information to link the experimental data and for better understanding geochemical processes occurring in aquifers of the study region. These would very much useful for us in teaching undergraduate and post-graduate students of our University. Followings are our responses in details.

1) Figure 3 - I do not think this figure is needed to understand the rest of the paper.

Yes, we agree with you that figure. 3 is not so important information to attach to this paper, and we deleted it keeping the description of geology of the region to emphasize on the existence of faults which facilitate the inter-aquifer leakage as we observed.  

2) Line 144-145: This sentence states that you are quantifying the fractionation of ¹³C in DIC due to exchange with biogenic CO₂.  This is too limiting.  You are measuring the ¹³C in DIC to determine the source of inorganic carbon, which could be atmospheric CO₂, dissolution of carbonates, or biological in origin.

We agree with your comment and we modified our statement as it was written in the Red colour.

3) Page 5 (and to some extent pages 8-9):  There is quite a bit written here about how water and carbon isotopes work. The content is accurate, but I rarely see such lengthy explanations of the foundational principles of isotopic systems.  If you are looking for a place to cut, you could get rid of much of this and simply cite the foundational literature.

We have shorten the description.

More substantial points

1) It would be good to see a summary of water levels for all the samples so we can calculate differences in vertical gradients. Figure 10 is great, because it shows the connectivity of the two aquifers.  However, these are the only hydraulic head data provided in the paper.

We provided hydraulic water head measured in boreholes during sampling campaigns in Table 1 and gave a short description for the data.

2) Discussion of figure 5a in results section: The seawater mixing trend could also be explained equally well as an evaporation trend. This would be the trend for rainwater that underwent evaporation prior to recharge.  This is common for recharge from some lakes, lagoons, swampy regions, possibly some rivers, etc.  Later in the manuscript you provide additional evidence of seawater infiltration, but this graph alone does not establish that as the only unique interpretation.  Another thought here would be to save the interpretation until later altogether.  It is not customary to provide interpretation in the results section - just the discussion section.

We removed the interpretation for delta deuterium vs. delta oxygen-18 relationship trend from “Results section” to “Discussion section” and combined with other observations and calculations to show groundwater in Holocene aquifer is really being affected by seawater intrusion.

3) Discussion of figure 5b (it now becomes figure 4b) in the results section: What water isotopic composition is expected for paleo-water and why?

Actually, we don’t know how would the water isotopic composition of paleo-water be, because the absolute range of d²H and d¹⁸O in paleo- water (that has long transit time or was connate) was not available in literatures. However, we learned that, due to climate change the heavy isotopic signatures of old age water would be enriched, so that water line for this type of water would lye parallel and under recent meteoric water line. So, a trend of water isotopic compositions connecting points from recent precipitations and points characterizing enriched d¹⁸O (far below the recent meteoric water line) was usually interpreted as a mixing between paleo- and recent precipitation, e.g. Yusever at al. (…).    

4)Figure 7 (Piper diagram): In some cases it is hard to corroborate the interpretations of the data as they are shown in the Piper diagrams.  I suggest creating the key cross plots needed to support the interpretations either instead of the Piper diagrams or in addition to them.  For example, if you are looking at waters that mix with seawater versus waters that evolve to a Na-bicarbonate concentration you should be looking at simple cross plot graphs like Na vs. Cl or Na vs. Ca, or Cl vs. HCO₃, or more complex but illustrative graphs such as Na/Cl vs. 1/Cl, Na/Ca vs. 1/Cl or Cl/HCO3, etc.  The simplicity of cross plots would make it easier to demonstrate relationships and help with interpretation.

We agree with you that the use of Piper diagram some time is very hard to categorize chemistry of water. We removed the diagram and replaced by cross plots of molar [Na⁺] vs. [Cl^-] and molar [HCO₃^-] to [Cl-] ratio vs. [Cl^-]. 

5) Figure 8 This is a key diagram, because it can “irrefutably” establish whether you do or do not have mixing with seawater.  Something to check on here is whether you need to have Cl on a log scale to establish a linear mixing model.  When comparing concentration data to isotope data it is common to use log concentration, as mixing is not always a straight line with raw concentration data.  In this particular case I am not sure, but it is worth checking because you don’t want to have the wrong percentages in the mixing model.  You use a log scale for the same parameters in Figure 9, so I’m not sure why you did not use the log scale for figure 8.

We didn’t use log scale for chloride concentration in figure 8 (now it becomes figure 7) because it better to demonstrate 3 end-members of the mixing: seawater-river’s water- precipitation. If it changes to log scale it would look not so impressive. We keep figure 8 (7) the same as it was.

 Finally, I agree with the interpretation here, but I am not sure where the local precipitation data point came from.  

We inserted a point showing the weighted (by event amount) isotopic compositions of local precipitation in figure 4a, b. In fact its was said in the text before we show the contribution of each end-member to water in Holocene aquifer, lines 392-395.

6) One of the interpretations in this paper that is not well-supported is the source of saline water in the Pleistocene and Neogene aquifers.  The authors suggest it is entrapped aquifer/aquitard deposits of saline water, but what evidence do we have to support this?  What is the expected isotopic composition of trapped water of Holocene age?  If it is saline, wouldn’t it be trapped ocean water and therefore have a seawater isotope signature very similar to the present day seawater?  The lack of information here represents an important gap.  Perhaps the concentration of Br relative to Cl may shed some light on the source of the saline waters.

We supplemented a fact to prove there was no recharge from Holocene aquifer to deep Pleistocene aquifer by statistical processing data using mean d¹⁸O and it standard deviation in water in the two aquifers. Additional, we provided our calculation for molar [Ca] to [Mg] as well as [Br] to [Cl] in water of deep aquifers to show freshening of groundwater in deep aquifers is on-going but not salinization.  

7) One major improvement would be to include the saturation indices for carbonate and other minerals in the aquifer system in a table.  This would help support several key interpretations regarding carbonate dissolution or precipitation (see below).

We supplemented SI of some minerals possibly existed in aquifers deposits by inserting Table 3 and Table 4 and used these data to  interprete related observations.

8) Figure 9b (now it becomes figure 8b) shows that a group of very dilute (in terms of chloride) samples change their Na/Cl ratio substantially, which is attributed to the exchange of Ca in the water for two Na ions on clay.  This is the typical pathway suggested for the development of Na-bicarbonate waters.  However, the authors need to rule out the possibility that you have silicate weathering contributing the Na and Ca is simply stagnant or being tied up in carbonate precipitation. Alternately Na may remain constant and Ca is being lost from the system.  Having saturation index calculations and additional cross plots could address these possibilities.

We used data presented in Table 4 to show Ca in boreholes Q220t, Q227a, Q228a, Q229a is gained, but not lost, so Ca to Na cation exchange could be the fact.

9) Page 17: the precipitation of siderite is used to explain the lack of correlation of Fe²⁺ and HCO3.  This is only one of many possible explanations, and perhaps the least likely.  Siderite needs to be substantially over-saturated to precipitate in most groundwater systems.  Other examples of things that could impact this relationship include the adsorption of Fe(II) to other minerals, precipitation of pyrite, precipitation of calcite or dolomite, etc. One could shed more light on this with additional saturation index info and more cross plots. One way to link the creation of Fe(II) to biological activity would be to plot Fe(II) versus d¹³C DIC.  It would also be illuminating to see borehole depth vs. Fe(II) and Fe(II) versus sulfate.

Yes, we have already interpreted that siderite precipitation is one of reasons why [Fe²⁺} does not close correlate with {HCO₃^-]. In this case calcite, aragonite, dolomite in water from some boreholes made in Holocene aquifer are also precipitating (Table 3). We could not be able to calculate SI for pyrites because we didn’t analyze sulphide in samples, but we believe pyrite also precipitates was water is strongly smelled with sulfur.

Graph of d¹³C vs. {Fe²⁺} was inserted to show reduction of sulphate by organic matters could be one more of processes controlling chemistry of groundwater in the region

10) The d¹³C of DIC is not effectively used as part of the story.  Figure 11 seems rather “tacked on” like the only reason for the figure was to show the d¹³C DIC data because there seems to be no other need to talk about Mg/Ca ratios.  Frankly, I do not see any relationship between d¹³C DIC and the Mg/Ca ratios that can be supported from Figure 11.  

We agree with you that incongruent dissolution of high Mg-biogenic calcite in this case was not important, so we deleted figure 11 and replaced by graph d¹³C vs. [SO₄].

Why not show the d¹³C DIC of the Holocene aquifer for comparison? 

In fact in this study we analyzed d¹³C only for the purpose of correction in estimation of ¹⁴C-age of water in deep aquifers. Water in Holocene aquifer was expected to be modern, so d¹³C in DIC of water from Holocene aquifer was not analyzed for all samples, so we didn’t have many data of d¹³C in DIC in water of shallow aquifer to compare.  

The effect of precipitation of carbonates is mentioned, but this is not supported by saturation indices.  I would also like to see plots of Mg/Ca ratio and 1/Cl or similar plots to better assess which fluids are interacting.

Data of SI for most available minerals in aquifer deposits have been inserted and discussed in the text

11) Line 608 - here it indicates that the hydrochemistry is controlled in part by sulfate reduction but earlier (line 559) it states that sulfate is in low concentration and it was not discussed in favor of a discussion of Fe(II) and ammonia.

We are very sorry for the confusion. In new version of manuscript we showed scatter plot of molar [SO₄^2-] vs. [HCO₃^-] to show reduction of sulphate by organic matters in aquifers that made [SO₄^2-] to be low.

We have asked an original English-speaking teacher working for a secondary school in Hanoi to check and correct for our English, but I am sure that you are still not satisfied with our new version. Please understand our situation!

Dear Prof.,  we learn that you are a top expert of geochemistry, so if it is possible please let’s know your address (E-mail) in order we could communicate in future to get from you comments and instruction in hydro-geological and isotopes hydrology researches. We are currently conducting research into salt intrusion along the sea coat and thermal an minerals water resources in highland.

One again thank you very much and looking forwards to having your advices in future!

Author Response File: Author Response.docx

Reviewer 2 Report

Sugesstions and corrections are included within the text.

I like your paper!

Best regards...

Comments for author File: Comments.pdf

Author Response

Many thanks for you time spent to read our manuscript and make valuable comments. We have corrected for the confusing things in new version per your request.

Thank you again and we look forwards to having your advice and comments to our researches in future.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors did a nice job of addressing the earlier comments and I think improve the manuscript in some key areas.  I do not have any further comments or recommendations.  However, the paper, particularly the new sections, could use some minor editorial polishing.  

Author Response

Thank you so much for your comments and following your suggestion we have made the correction as follows:

Lines 123-124 were corrected to:

“(2) determination of composition of deuterium (d²H) and oxygen-18 (d¹⁸O) in water to delineate…” ;

Line 130 was corrected to:

“During the water cycle, composition of hydrogen and oxygen in water  will be changed…”

(Revised version is attached!)