Technical note: An Assessment of Hg<sup>II</sup> to Preserve Carbonate System Parameters in Organic-Rich Estuarine Waters

Moore, Christopher S.; Byrne, Robert H.; Yates, Kimberly K.

doi:https://doi.org/10.5194/egusphere-2022-1493

Preprints

https://doi.org/10.5194/egusphere-2022-1493

Preprints

13 Jan 2023

| 13 Jan 2023

Technical note: An Assessment of Hg^II to Preserve Carbonate System Parameters in Organic-Rich Estuarine Waters

Christopher S. Moore, Robert H. Byrne, and Kimberly K. Yates

Abstract. This work assesses the effectiveness of sample preservation techniques for measurements of pH_T (total scale), total dissolved inorganic carbon (DIC_T), and total alkalinity (A_T) in organic-rich estuarine waters. Using HgCl₂-treated and untreated water samples, measurements of these carbonate system parameters were examined over a period of three months. Over this duration, continued respiration of dissolved organic matter (DOM) in untreated samples created large discrepancies in DIC_T concentrations, while DIC_T was effectively constant in treated samples. Changes in A_T were observed for both treated and untreated samples, with treated samples showing the greatest variation. In response to changing A_T / DIC_T ratios, pH_T changes were observed in both treated and untreated samples but were relatively small in treated samples. Improved accuracy of results in organic-rich estuarine waters that reflect the in situ carbonate system characteristics of the samples at the time of collection can be achieved when samples obtained for DIC_T and A_T analysis are collected and stored separately. Accurate analyses of DIC_T can be obtained by filtration and preservation with HgCl₂. Accuracy of A_T analyses can be improved by filtration and storage in polypropylene bottles at 4 °C without adding HgCl₂. Quality of pH_T measurements can be improved by prompt analysis in the field and, if this cannot be accomplished, then samples can be preserved with HgCl₂ and measured in the laboratory within one week.

Received: 22 Dec 2022 – Discussion started: 13 Jan 2023

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 preprint. The responsibility to include appropriate place names lies with the authors.

Christopher S. Moore, Robert H. Byrne, and Kimberly K. Yates

Status: closed

RC1:
'Comment on egusphere-2022-1493', Anonymous Referee #1, 07 Feb 2023

1. Line 33-38 Why 3 months experiment? Mos et al., 2021 also did research/observation over six months. I am just wondering, if it is possible, do you also thinking of extend the study up to six months?
2. Line 70 how do you define the ratio of HgCl2 of 6.5%? The ratio of HgCl2 in treated water samples matters the samples storage and impact CO2sys.m parameters? (Normally we preserve water samples for trace metal analysis, e.g. mercury analysis, to add 0.5% of HCl. Perhaps this ratio could be adjusted little lower?)
3. Line 102-103 ideally, it would be better to run three treated samples and three untreated samples to get the average values.
4. Line 141-143 instrument use for Nutrient analyses?
5. Figure 3 and Figure 4 I am not sure if tidal situation may also impact A_T and pH difference between the HgCl2-treated and untreated samples at the beginning of the measurement period? The first wtaer sampling may be impacted more by river source, and the 2^nd sampling may be impacted more by ocean water?
6. Line 207 Table1. The measured inorganic nutrients sampled from each Niskin bottle at the time of sample collection (Mean ± STD)
7. Line 281-282 In the Figure 4, pH values of HgCl2-treated samples seems not showing much difference between the first week and the 3months even with variation during the 3 month. I think acceptable quality measurements of pH_T can be preserved with HgCl2 and measured in the laboratory up to 3 months?

Citation: https://doi.org/10.5194/egusphere-2022-1493-RC1
- AC1: 'Reply on RC1', Christopher Moore, 16 Mar 2023
  
  RC1: Anonymous Referee #1, 07 Feb 2023
  “1. Line 33-38 Why 3 months experiment? Mos et al., 2021 also did research/observation over six months. I am just wondering, if it is possible, do you also thinking of extend the study up to six months?”
  Mos et al., 2021 performed their experiments over a segmented period of 0, 1 and 6 months, while our study performed continuous analyses on a daily (initial week) and weekly time scale over a period of 3 months after sample collection. The experimental duration was constrained by the water budget provided by two 30-L Niskin bottles and the desired resolution of results in that timeframe. Typically, when sampling for carbonate system parameters in estuarine conditions, we try to analyze water samples within the first two weeks and do not wait longer than 3 months.
  “2. Line 70 how do you define the ratio of HgCl2 of 6.5%? The ratio of HgCl2 in treated water samples matters the samples storage and impact CO2sys.m parameters? (Normally we preserve water samples for trace metal analysis, e.g. mercury analysis, to add 0.5% of HCl. Perhaps this ratio could be adjusted little lower?)”
  The ratio of 6.5% HgCl₂ was determined by a certified stock solution of Mercuric Chloride saturated solution made by Lab Chem. Cat #: LC166201 Lot #: G102-16. The solution was 6.5% mercuric chloride (CA 7487-94-7) and 93.5% water (CAS 7732-18-5). Based on our findings and the findings of Mos et al., 2021, where water samples of different water types were treated with different ratios of Hg^II and analyzed over different time periods, adjusting the ratio of HgCl₂ added to water samples of certain water types would still be problematic as it would be dependent on the samples’ fraction of organic alkalinity within the DOM concentration, ionic strength, and potentially other unknown factors.
  “3. Line 102-103 ideally, it would be better to run three treated samples and three untreated samples to get the average values.”
  Analyses from three bottles of each sample type would be more statistically significant. This was not performed primarily due to the overall water budget and the experimental design. We are highly confident in the reported level of error. During each day’s laboratory analyses, certified reference material (CRM) was analyzed before and after sample bottle analyses to calculate accuracy and precision. Low sample volume requirements for C_T analyses allowed duplicate C_T sample analysis to be performed as an additional instrument performance check to ensure that differences were within precision standards based on CRM replicates. This strengthened our confidence in the results and allowed us to report standard deviations of the average of duplicate samples obtained from each sample bottle. Next, since the experiment allowed analysis of identical water samples for the two sample types collected at the same time over elapsed time, we can be confident in the resulting best fit linear trend, slope, and significance tests of the C_T and A_T results.
  “4. Line 141-143 instrument use for Nutrient analyses?”
  The instrument used was a Seal Analytical Autoanalyzer 3.
  “5. Figure 3 and Figure 4 I am not sure if tidal situation may also impact AT and pH difference between the HgCl2-treated and untreated samples at the beginning of the measurement period? The first wtaer sampling may be impacted more by river source, and the 2nd sampling may be impacted more by ocean water?”
  Although this location is indeed tidally affected, the two water sample types were collected in adjacent 30 L Niskin bottles within 5 minutes of one another whereby the first Niskin was closed at 12:45 EST on 1/26/2021. At that time, 1.03 miles upstream, a USGS monitored stream gauge station (ID 023060013) displayed a gage height of 0.67’ above sea level on the incoming tide of a mixed semi diurnal tidal cycle. The flow direction did not change until 15:00 EST that day, which marked the beginning of that days next outgoing tide. At the time of collection, the existence of two small natural springs within the park, both upstream and downstream of the study site, was not known. This groundwater input likely influenced the carbonate system parameter results and nutrient results observed between the two Niskin bottles at the time of collection.
  “6. Line 207 Table1. The measured inorganic nutrients sampled from each Niskin bottle at the time of sample collection (Mean ± STD)”
  Thank you, this will be added to the table caption.
  “7. Line 281-282 In the Figure 4, pH values of HgCl2-treated samples seems not showing much difference between the first week and the 3months even with variation during the 3 month. I think acceptable quality measurements of pHT can be preserved with HgCl2 and measured in the laboratory up to 3 months?”
  We agree that there is little difference between initial laboratory pH_T measurements and those made at 91 days after collection, but considering the variability observed within that 91 day time frame compared to that seen in the first week, in future work we would be confident in pH_T results obtained within 1 week.
  
  Citation: https://doi.org/10.5194/egusphere-2022-1493-AC1
CC1:
'Comment on egusphere-2022-1493', Lauren Barrett, 07 Feb 2023

Sorry if this is a double-post, I just don't see my comment from before so I think I accidentally timed out! Hi! Really interesting paper and an important topic to consider. I have a few questions / comments. Line 102: "samples were analyzed in the order they were initially collected." It is my understanding that gas-sensitive samples are always taken from the Niskin first to avoid the effects of increased headspace. Do you think choice this may have affected your time-series analysis? Line 105: You say sample water was "poured" from the original container into the UV/Vis cuvette. Can you be more specific about this? Pouring would create turbulence and possibly affect CO2. It seems this was taken into consideration in the way DIC was sampled. Line 211: You say that the decrease of TA over time in the treated samples is due to complexation of organic bases with Hg(II). Do you have any DOM data to corroborate this hypothesis? Figure 5: If you believe your treated DIC and pH and your untreated TA samples are the most accurate, can you explain the closer agreement between treated TA and TA calculated from treated DIC and pH? It seems that agreement stays relatively robust for up to ~60 days. If there is a significant contribution of organic bases to measured alkalinity, I would expect the measured TA to be higher than calculated TA, but this is not the case for either the treated or untreated samples. Line 243: Without DOM / organic alkalinity data, is there a reason you assume Figure 3 indicates that there is substantial organic alkalinity and thus the algorithms used to produce Figure 5 are inaccurate, instead of vice versa? In other words, could it not be that there isn't a lot of organic alkalinity (as indicated by Figure 5) and the differences in Figure 3 are attributed to something else? Thank you!!

Citation: https://doi.org/10.5194/egusphere-2022-1493-CC1
- AC2: 'Reply on CC1', Christopher Moore, 16 Mar 2023
  
  CC1: Lauren Barrett, 07 Feb 2023
  “Sorry if this is a double-post, I just don't see my comment from before so I think I accidentally timed out! Hi! Really interesting paper and an important topic to consider. I have a few questions / comments. Line 102: "samples were analyzed in the order they were initially collected." It is my understanding that gas-sensitive samples are always taken from the Niskin first to avoid the effects of increased headspace. Do you think choice this may have affected your time-series analysis?”
  We agree that parameters sensitive to gas exchange (pH_T and C_T) should be sampled first to avoid the influence of gas exchange, which is why field pH_T water samples and sample bottles were sampled immediately after opening each Niskin for collections. All water samples were collected from each Niskin within a period of 45 minutes. If gas exchange did appreciably influence the water samples, we would have seen the effects of gas exchange in HgCl₂-treated C_T samples collected over the 45 minute collection period. Figure 2 shows that treated C_T samples were essentially invariant for the duration of the experiment (3,095.5 ± 3.2 µmol kg^-1).
  “Line 105: You say sample water was "poured" from the original container into the UV/Vis cuvette. Can you be more specific about this? Pouring would create turbulence and possibly affect CO2. It seems this was taken into consideration in the way DIC was sampled.”
  When filling the 10 cm cylindrical optical glass pH cell with sample water, the sample bottle neck was first wiped clean with a Kim Wipe and the pH cell was rinsed three times with a small amount of sample water. After rinsing, the sample was poured into the pH cell and allowed to overflow while carefully observing for any visibly trapped air bubbles. Once complete the cell was immediately capped without air bubbles and placed in a thermal housing for temperature equilibration prior to analyses. Importantly, duplicate pH analyses were performed on C_T and A_T reference material prior to sample analysis to establish an average daily precision of ± 0.012. The resulting average sample precision of our pH_T data was ± 0.015. While this level of precision is somewhat worse than what can be spectrophotometrically achieved in the open ocean (± 0.001), given the spatial and temporal heterogeneity of estuarine/riverine waters it should be satisfactory.
  “Line 211: You say that the decrease of TA over time in the treated samples is due to complexation of organic bases with Hg(II). Do you have any DOM data to corroborate this hypothesis?”
  We do not have DOM data to corroborate this, although the Hillsborough County Environmental Protection Commission performs monthly water quality assessments 0.75 miles upstream (Station 270) and 0.82 miles downstream (Station 1510) of the study site that include total organic carbon (mg/L) measurements (sum of the particulate and dissolved organic carbon). These measurements did not take place during the time of this study, presumably due to COVID-19 restrictions, however monthly records beginning 9/2018 to 1/2023 at these two sites show organics were present at both stations in notable concentrations: a range of 45.9 mg/L with an average of 12.7 ± 7.4 (n=76). Additionally, the sampling location has well known and consistent DOM sources upstream - terrestrial inputs from the Green Swamp, a region of 560,000 acres that drain to create the headwaters of four major rivers including the Hillsborough River, and groundwater input from numerous springs of varying magnitude.
  “Figure 5: If you believe your treated DIC and pH and your untreated TA samples are the most accurate, can you explain the closer agreement between treated TA and TA calculated from treated DIC and pH? It seems that agreement stays relatively robust for up to ~60 days. If there is a significant contribution of organic bases to measured alkalinity, I would expect the measured TA to be higher than calculated TA, but this is not the case for either the treated or untreated samples.”
  The reason for the relatively good agreement between measured HgCl₂-treated A_T and A_T calculated from HgCl₂-treated C_T and pH_T is mainly due to covariance of HgCl₂-treated pH measurements and HgCl₂-treated A_T over time. HgCl₂-treated C_T measurements were essentially invariant throughout the experiment while, in contrast, HgCl₂-treated A_T decreased throughout the experimental time period (likely due to the complexation of DOM by Hg^II). This complexation process released H⁺ ions causing pH_T to decrease through time. As stated in the manuscript, the observation that calculated A_T is greater than measured A_T for both Hg^II-treated and untreated samples is inconsistent with the fact that CO2sys.m calculations do not account for organic alkalinity contributions. This observation, that measured A_T is not larger than calculated A_T, points to potential problems with CO₂-system dissociation constants at low salinities. In contrast to the abundance of CO₂ system dissociation constants at salinity ≥ 20, only one set of dissociation constants has been thoroughly parameterized for calculations made at salinities that are well below 20.
  “Line 243: Without DOM / organic alkalinity data, is there a reason you assume Figure 3 indicates that there is substantial organic alkalinity and thus the algorithms used to produce Figure 5 are inaccurate, instead of vice versa? In other words, could it not be that there isn't a lot of organic alkalinity (as indicated by Figure 5) and the differences in Figure 3 are attributed to something else? Thank you!!”
  This was initially considered yet ruled out as we have good reason to believe that DOM was present at substantial concentrations in the water samples, even without direct measurements. During sample collection, the sample collection locale had well known DOM sources upstream via terrestrial input from the Green Swamp, groundwater input from a large magnitude spring, and several small magnitude springs immediately upstream and downstream of the sample site within Lowry Park. Importantly, filtered water samples exhibited a brownish yellow color. Additionally, as noted in our manuscript, strong affinity between organic ligands, such as are found in the DOM of natural waters, and Hg^II are well documented. Finally, our observation of A_T decreases in organic-rich HgCl₂-treated samples is in accord with Mos et al., 2021 results and interpretation.
  
  Citation: https://doi.org/10.5194/egusphere-2022-1493-AC2
RC2:
'Comment on egusphere-2022-1493', Anonymous Referee #2, 11 Feb 2023
The manuscript describes a technical approach with the aim of preservation of estuarine/marine water samples for analysis of the carbonate system species. The authors used HgCl₂ as a preserving agent and then perform chemical analysis of ainly pH, A and DIC over 3 months.
The use of HgCl₂ is not new and has been using several times in the past with the aim of DOC analysis, however the approach of this technical note nis somehow interesting.
I’ve some concerns that I should need clarification.
The authors should clear point the novelty of this work compared, for instance with the Mos paper.

What was the amount of HgCl₂ added?

The authors are assuming that DIC changes are a consequence of DOC/DOM mineralization, which makes sense. However, there is no clear assumption that there is not evidence that there is no CO₂ dissolution. Probably the work had this issue in consideration but a clear sentence on this issue should be had to the manuscript. Also, and still considering this problem, this work with benefit a lot from determinations of DOC in the analyzed samples.

pH is measured using a spectroscopic technique instead of a normal high-resolution ISE. Is there any specific requirements for using this method? More details about calibration should be provided.

In figure 3, there is a clear tendence of At decrease in HgCl₂ tretaed samples. For untreated, the authors assume that the AVG At is constant. For the last ones it does not look that way. It seems that until day 49 the AT levels are constant but after a decrease is clear observed. Probably the use of a statistic test would clarify this doubt.

It is well known that Hg(II) easily complex with DOM. It is difficult to know how much Hg (II) is complexed (stochiometric all Hg can be complexed because concentration of DOM is higher thar Hg2+) and, consequently we will not know the DOM available for mineralization. Again, DOC measurements would benefit this work.

In summary, this technical note needs this clarification before being considered for publication.
Citation: https://doi.org/10.5194/egusphere-2022-1493-RC2
- AC3: 'Reply on RC2', Christopher Moore, 16 Mar 2023
  
  RC2: Anonymous Referee #2 11 Feb 2023
  “The manuscript describes a technical approach with the aim of preservation of estuarine/marine water samples for analysis of the carbonate system species. The authors used HgCl2 as a preserving agent and then perform chemical analysis of ainly pH, A and DIC over 3 months.
  
  The use of HgCl2 is not new and has been using several times in the past with the aim of DOC analysis, however the approach of this technical note nis somehow interesting.
  
  I’ve some concerns that I should need clarification.”
  
  “1. The authors should clear point the novelty of this work compared, for instance with the Mos paper.”
  Using state-of-the-art spectrophotometric techniques (pH_T and A_T) over a highly-resolved 3 month timescale, our work compares carbonate system parameters (pH_T, A_T and C_T) measured in low salinity estuarine water with and without additions of HgCl₂. Mos et al., 2021 analyzed two of the three parameters where pH was measured using NIST scale (potentiometric) and A_T was measured using Gran potentiometric titrations on three occasions (at 0,1 and 6 months). Our observations of C_Tconstancy in HgCl₂-treated samples along with declining A_Tin HgCl₂-treated samples provides a substantially insightful view of CO₂ sample preservation processes.
  This information will be added to the revised manuscript.
  “2. What was the amount of HgCl2 added?”
  For HgCl₂-treated samples, 100 microliters of 6.5% HgCl2 was added to 300 mL sample bottles. This information will be added to the revised manuscript.
  “3. The authors are assuming that DIC changes are a consequence of DOC/DOM mineralization, which makes sense. However, there is no clear assumption that there is not evidence that there is no CO2 dissolution. Probably the work had this issue in consideration but a clear sentence on this issue should be had to the manuscript. Also, and still considering this problem, this work with benefit a lot from determinations of DOC in the analyzed samples.”
  Two outcomes lead us to believe that CO₂ dissolution was inconsequential in this work. Of the measured parameters sensitive to CO₂ dissolution that were unaffected by respiration of organic matter and complexation by Hg^II, HgCl₂-treated C_T measurements results remained essentially invariant (3,095.5 ± 3.2 µmol kg^-1) (Samples were analyzed in the order in which they were filled). A statement will be added to the manuscript addressing this point. Though this manuscript lacks determination of DOM concentrations, its presence was undeniable given the environmental setting, upstream and downstream measurements of TOC made by the EPCHC (See CC1), and carbonate system results for untreated samples that exhibited changes due to respiration of organic matter that are consistent with those of the Redfield equation.
  Although direct DOM measurements would somewhat alleviate uncertainties surrounding the significance of interactions between organics and Hg^II, future observations of A_T in organic-rich systems will be best served with the inclusion of titrimetically-derived organic acid/base concentrations and pK values.
  “4. pH is measured using a spectroscopic technique instead of a normal high-resolution ISE. Is there any specific requirements for using this method? More details about calibration should be provided.”
  Spectrophotometric pH_T measurements have been directly compared to potentiometric methods by repeated analyses at low ionic strengths and are shown to be more accurate (French et al., 2002). Spectrophotometric pH_T is also used to calibrate pH electrodes, including ISFETs, in seawater (Easley and Byrne, 2012) and estuarine waters (Martell-Bonet and Byrne, 2020). Spectrophotometric pH indicators are molecularly calibrated using calibration buffers and eliminate bias caused by variations in the liquid junction potentials of potentiometric pH systems. This information will be added to the revised manuscript.
  “5. In figure 3, there is a clear tendence of At decrease in HgCl2 treated samples. For untreated, the authors assume that the AVG At is constant. For the last ones it does not look that way. It seems that until day 49 the AT levels are constant but after a decrease is clear observed. Probably the use of a statistic test would clarify this doubt.”
  Line 168 – 169 states the significance test results of regression at 95% confidence level of p < 0.01 to statistically show that the A_T of untreated samples was constant. Also, the results show that the A_T for untreated samples is constant for a sufficiently long time period that untreated samples can be returned to the lab for analysis. A_T analysis of untreated samples in a suitably short time frame is clearly preferable to A_T analysis of HgCl₂-treated samples.
  “6. It is well known that Hg(II) easily complex with DOM. It is difficult to know how much Hg (II) is complexed (stochiometric all Hg can be complexed because concentration of DOM is higher thar Hg2+) and, consequently we will not know the DOM available for mineralization. Again, DOC measurements would benefit this work.”
  We agree that we do not know the extent of complexation by Hg^II. However this work shows for the first time, at high resolution, that Hg^II addition, which is the standard method of sample preservation for carbonate system water samples, rapidly affects the pH and A_T of estuarine water samples after collection. Also, this work suggests that organic alkalinity titrations (i.e., the fraction of DOM that actively exchanges hydrogen ions) performed after removal of all inorganic carbon, in conjunction with A_T measurements, is likely preferable to measurements of DOM in conjunction with A_T. Text will be added to the manuscript to highlight these conclusions and should benefit future work in this realm.
  “In summary, this technical note needs this clarification before being considered for publication.”
  New text and clarifications will be added to the manuscript to address the reviewers’ comments.
  
  Citation: https://doi.org/10.5194/egusphere-2022-1493-AC3

Status: closed

RC1:
'Comment on egusphere-2022-1493', Anonymous Referee #1, 07 Feb 2023

1. Line 33-38 Why 3 months experiment? Mos et al., 2021 also did research/observation over six months. I am just wondering, if it is possible, do you also thinking of extend the study up to six months?
2. Line 70 how do you define the ratio of HgCl2 of 6.5%? The ratio of HgCl2 in treated water samples matters the samples storage and impact CO2sys.m parameters? (Normally we preserve water samples for trace metal analysis, e.g. mercury analysis, to add 0.5% of HCl. Perhaps this ratio could be adjusted little lower?)
3. Line 102-103 ideally, it would be better to run three treated samples and three untreated samples to get the average values.
4. Line 141-143 instrument use for Nutrient analyses?
5. Figure 3 and Figure 4 I am not sure if tidal situation may also impact A_T and pH difference between the HgCl2-treated and untreated samples at the beginning of the measurement period? The first wtaer sampling may be impacted more by river source, and the 2^nd sampling may be impacted more by ocean water?
6. Line 207 Table1. The measured inorganic nutrients sampled from each Niskin bottle at the time of sample collection (Mean ± STD)
7. Line 281-282 In the Figure 4, pH values of HgCl2-treated samples seems not showing much difference between the first week and the 3months even with variation during the 3 month. I think acceptable quality measurements of pH_T can be preserved with HgCl2 and measured in the laboratory up to 3 months?

Citation: https://doi.org/10.5194/egusphere-2022-1493-RC1
- AC1: 'Reply on RC1', Christopher Moore, 16 Mar 2023
  
  RC1: Anonymous Referee #1, 07 Feb 2023
  “1. Line 33-38 Why 3 months experiment? Mos et al., 2021 also did research/observation over six months. I am just wondering, if it is possible, do you also thinking of extend the study up to six months?”
  Mos et al., 2021 performed their experiments over a segmented period of 0, 1 and 6 months, while our study performed continuous analyses on a daily (initial week) and weekly time scale over a period of 3 months after sample collection. The experimental duration was constrained by the water budget provided by two 30-L Niskin bottles and the desired resolution of results in that timeframe. Typically, when sampling for carbonate system parameters in estuarine conditions, we try to analyze water samples within the first two weeks and do not wait longer than 3 months.
  “2. Line 70 how do you define the ratio of HgCl2 of 6.5%? The ratio of HgCl2 in treated water samples matters the samples storage and impact CO2sys.m parameters? (Normally we preserve water samples for trace metal analysis, e.g. mercury analysis, to add 0.5% of HCl. Perhaps this ratio could be adjusted little lower?)”
  The ratio of 6.5% HgCl₂ was determined by a certified stock solution of Mercuric Chloride saturated solution made by Lab Chem. Cat #: LC166201 Lot #: G102-16. The solution was 6.5% mercuric chloride (CA 7487-94-7) and 93.5% water (CAS 7732-18-5). Based on our findings and the findings of Mos et al., 2021, where water samples of different water types were treated with different ratios of Hg^II and analyzed over different time periods, adjusting the ratio of HgCl₂ added to water samples of certain water types would still be problematic as it would be dependent on the samples’ fraction of organic alkalinity within the DOM concentration, ionic strength, and potentially other unknown factors.
  “3. Line 102-103 ideally, it would be better to run three treated samples and three untreated samples to get the average values.”
  Analyses from three bottles of each sample type would be more statistically significant. This was not performed primarily due to the overall water budget and the experimental design. We are highly confident in the reported level of error. During each day’s laboratory analyses, certified reference material (CRM) was analyzed before and after sample bottle analyses to calculate accuracy and precision. Low sample volume requirements for C_T analyses allowed duplicate C_T sample analysis to be performed as an additional instrument performance check to ensure that differences were within precision standards based on CRM replicates. This strengthened our confidence in the results and allowed us to report standard deviations of the average of duplicate samples obtained from each sample bottle. Next, since the experiment allowed analysis of identical water samples for the two sample types collected at the same time over elapsed time, we can be confident in the resulting best fit linear trend, slope, and significance tests of the C_T and A_T results.
  “4. Line 141-143 instrument use for Nutrient analyses?”
  The instrument used was a Seal Analytical Autoanalyzer 3.
  “5. Figure 3 and Figure 4 I am not sure if tidal situation may also impact AT and pH difference between the HgCl2-treated and untreated samples at the beginning of the measurement period? The first wtaer sampling may be impacted more by river source, and the 2nd sampling may be impacted more by ocean water?”
  Although this location is indeed tidally affected, the two water sample types were collected in adjacent 30 L Niskin bottles within 5 minutes of one another whereby the first Niskin was closed at 12:45 EST on 1/26/2021. At that time, 1.03 miles upstream, a USGS monitored stream gauge station (ID 023060013) displayed a gage height of 0.67’ above sea level on the incoming tide of a mixed semi diurnal tidal cycle. The flow direction did not change until 15:00 EST that day, which marked the beginning of that days next outgoing tide. At the time of collection, the existence of two small natural springs within the park, both upstream and downstream of the study site, was not known. This groundwater input likely influenced the carbonate system parameter results and nutrient results observed between the two Niskin bottles at the time of collection.
  “6. Line 207 Table1. The measured inorganic nutrients sampled from each Niskin bottle at the time of sample collection (Mean ± STD)”
  Thank you, this will be added to the table caption.
  “7. Line 281-282 In the Figure 4, pH values of HgCl2-treated samples seems not showing much difference between the first week and the 3months even with variation during the 3 month. I think acceptable quality measurements of pHT can be preserved with HgCl2 and measured in the laboratory up to 3 months?”
  We agree that there is little difference between initial laboratory pH_T measurements and those made at 91 days after collection, but considering the variability observed within that 91 day time frame compared to that seen in the first week, in future work we would be confident in pH_T results obtained within 1 week.
  
  Citation: https://doi.org/10.5194/egusphere-2022-1493-AC1
CC1:
'Comment on egusphere-2022-1493', Lauren Barrett, 07 Feb 2023

Sorry if this is a double-post, I just don't see my comment from before so I think I accidentally timed out! Hi! Really interesting paper and an important topic to consider. I have a few questions / comments. Line 102: "samples were analyzed in the order they were initially collected." It is my understanding that gas-sensitive samples are always taken from the Niskin first to avoid the effects of increased headspace. Do you think choice this may have affected your time-series analysis? Line 105: You say sample water was "poured" from the original container into the UV/Vis cuvette. Can you be more specific about this? Pouring would create turbulence and possibly affect CO2. It seems this was taken into consideration in the way DIC was sampled. Line 211: You say that the decrease of TA over time in the treated samples is due to complexation of organic bases with Hg(II). Do you have any DOM data to corroborate this hypothesis? Figure 5: If you believe your treated DIC and pH and your untreated TA samples are the most accurate, can you explain the closer agreement between treated TA and TA calculated from treated DIC and pH? It seems that agreement stays relatively robust for up to ~60 days. If there is a significant contribution of organic bases to measured alkalinity, I would expect the measured TA to be higher than calculated TA, but this is not the case for either the treated or untreated samples. Line 243: Without DOM / organic alkalinity data, is there a reason you assume Figure 3 indicates that there is substantial organic alkalinity and thus the algorithms used to produce Figure 5 are inaccurate, instead of vice versa? In other words, could it not be that there isn't a lot of organic alkalinity (as indicated by Figure 5) and the differences in Figure 3 are attributed to something else? Thank you!!

Citation: https://doi.org/10.5194/egusphere-2022-1493-CC1
- AC2: 'Reply on CC1', Christopher Moore, 16 Mar 2023
  
  CC1: Lauren Barrett, 07 Feb 2023
  “Sorry if this is a double-post, I just don't see my comment from before so I think I accidentally timed out! Hi! Really interesting paper and an important topic to consider. I have a few questions / comments. Line 102: "samples were analyzed in the order they were initially collected." It is my understanding that gas-sensitive samples are always taken from the Niskin first to avoid the effects of increased headspace. Do you think choice this may have affected your time-series analysis?”
  We agree that parameters sensitive to gas exchange (pH_T and C_T) should be sampled first to avoid the influence of gas exchange, which is why field pH_T water samples and sample bottles were sampled immediately after opening each Niskin for collections. All water samples were collected from each Niskin within a period of 45 minutes. If gas exchange did appreciably influence the water samples, we would have seen the effects of gas exchange in HgCl₂-treated C_T samples collected over the 45 minute collection period. Figure 2 shows that treated C_T samples were essentially invariant for the duration of the experiment (3,095.5 ± 3.2 µmol kg^-1).
  “Line 105: You say sample water was "poured" from the original container into the UV/Vis cuvette. Can you be more specific about this? Pouring would create turbulence and possibly affect CO2. It seems this was taken into consideration in the way DIC was sampled.”
  When filling the 10 cm cylindrical optical glass pH cell with sample water, the sample bottle neck was first wiped clean with a Kim Wipe and the pH cell was rinsed three times with a small amount of sample water. After rinsing, the sample was poured into the pH cell and allowed to overflow while carefully observing for any visibly trapped air bubbles. Once complete the cell was immediately capped without air bubbles and placed in a thermal housing for temperature equilibration prior to analyses. Importantly, duplicate pH analyses were performed on C_T and A_T reference material prior to sample analysis to establish an average daily precision of ± 0.012. The resulting average sample precision of our pH_T data was ± 0.015. While this level of precision is somewhat worse than what can be spectrophotometrically achieved in the open ocean (± 0.001), given the spatial and temporal heterogeneity of estuarine/riverine waters it should be satisfactory.
  “Line 211: You say that the decrease of TA over time in the treated samples is due to complexation of organic bases with Hg(II). Do you have any DOM data to corroborate this hypothesis?”
  We do not have DOM data to corroborate this, although the Hillsborough County Environmental Protection Commission performs monthly water quality assessments 0.75 miles upstream (Station 270) and 0.82 miles downstream (Station 1510) of the study site that include total organic carbon (mg/L) measurements (sum of the particulate and dissolved organic carbon). These measurements did not take place during the time of this study, presumably due to COVID-19 restrictions, however monthly records beginning 9/2018 to 1/2023 at these two sites show organics were present at both stations in notable concentrations: a range of 45.9 mg/L with an average of 12.7 ± 7.4 (n=76). Additionally, the sampling location has well known and consistent DOM sources upstream - terrestrial inputs from the Green Swamp, a region of 560,000 acres that drain to create the headwaters of four major rivers including the Hillsborough River, and groundwater input from numerous springs of varying magnitude.
  “Figure 5: If you believe your treated DIC and pH and your untreated TA samples are the most accurate, can you explain the closer agreement between treated TA and TA calculated from treated DIC and pH? It seems that agreement stays relatively robust for up to ~60 days. If there is a significant contribution of organic bases to measured alkalinity, I would expect the measured TA to be higher than calculated TA, but this is not the case for either the treated or untreated samples.”
  The reason for the relatively good agreement between measured HgCl₂-treated A_T and A_T calculated from HgCl₂-treated C_T and pH_T is mainly due to covariance of HgCl₂-treated pH measurements and HgCl₂-treated A_T over time. HgCl₂-treated C_T measurements were essentially invariant throughout the experiment while, in contrast, HgCl₂-treated A_T decreased throughout the experimental time period (likely due to the complexation of DOM by Hg^II). This complexation process released H⁺ ions causing pH_T to decrease through time. As stated in the manuscript, the observation that calculated A_T is greater than measured A_T for both Hg^II-treated and untreated samples is inconsistent with the fact that CO2sys.m calculations do not account for organic alkalinity contributions. This observation, that measured A_T is not larger than calculated A_T, points to potential problems with CO₂-system dissociation constants at low salinities. In contrast to the abundance of CO₂ system dissociation constants at salinity ≥ 20, only one set of dissociation constants has been thoroughly parameterized for calculations made at salinities that are well below 20.
  “Line 243: Without DOM / organic alkalinity data, is there a reason you assume Figure 3 indicates that there is substantial organic alkalinity and thus the algorithms used to produce Figure 5 are inaccurate, instead of vice versa? In other words, could it not be that there isn't a lot of organic alkalinity (as indicated by Figure 5) and the differences in Figure 3 are attributed to something else? Thank you!!”
  This was initially considered yet ruled out as we have good reason to believe that DOM was present at substantial concentrations in the water samples, even without direct measurements. During sample collection, the sample collection locale had well known DOM sources upstream via terrestrial input from the Green Swamp, groundwater input from a large magnitude spring, and several small magnitude springs immediately upstream and downstream of the sample site within Lowry Park. Importantly, filtered water samples exhibited a brownish yellow color. Additionally, as noted in our manuscript, strong affinity between organic ligands, such as are found in the DOM of natural waters, and Hg^II are well documented. Finally, our observation of A_T decreases in organic-rich HgCl₂-treated samples is in accord with Mos et al., 2021 results and interpretation.
  
  Citation: https://doi.org/10.5194/egusphere-2022-1493-AC2
RC2:
'Comment on egusphere-2022-1493', Anonymous Referee #2, 11 Feb 2023
The manuscript describes a technical approach with the aim of preservation of estuarine/marine water samples for analysis of the carbonate system species. The authors used HgCl₂ as a preserving agent and then perform chemical analysis of ainly pH, A and DIC over 3 months.
The use of HgCl₂ is not new and has been using several times in the past with the aim of DOC analysis, however the approach of this technical note nis somehow interesting.
I’ve some concerns that I should need clarification.
The authors should clear point the novelty of this work compared, for instance with the Mos paper.

What was the amount of HgCl₂ added?

The authors are assuming that DIC changes are a consequence of DOC/DOM mineralization, which makes sense. However, there is no clear assumption that there is not evidence that there is no CO₂ dissolution. Probably the work had this issue in consideration but a clear sentence on this issue should be had to the manuscript. Also, and still considering this problem, this work with benefit a lot from determinations of DOC in the analyzed samples.

pH is measured using a spectroscopic technique instead of a normal high-resolution ISE. Is there any specific requirements for using this method? More details about calibration should be provided.

In figure 3, there is a clear tendence of At decrease in HgCl₂ tretaed samples. For untreated, the authors assume that the AVG At is constant. For the last ones it does not look that way. It seems that until day 49 the AT levels are constant but after a decrease is clear observed. Probably the use of a statistic test would clarify this doubt.

It is well known that Hg(II) easily complex with DOM. It is difficult to know how much Hg (II) is complexed (stochiometric all Hg can be complexed because concentration of DOM is higher thar Hg2+) and, consequently we will not know the DOM available for mineralization. Again, DOC measurements would benefit this work.

In summary, this technical note needs this clarification before being considered for publication.
Citation: https://doi.org/10.5194/egusphere-2022-1493-RC2
- AC3: 'Reply on RC2', Christopher Moore, 16 Mar 2023
  
  RC2: Anonymous Referee #2 11 Feb 2023
  “The manuscript describes a technical approach with the aim of preservation of estuarine/marine water samples for analysis of the carbonate system species. The authors used HgCl2 as a preserving agent and then perform chemical analysis of ainly pH, A and DIC over 3 months.
  
  The use of HgCl2 is not new and has been using several times in the past with the aim of DOC analysis, however the approach of this technical note nis somehow interesting.
  
  I’ve some concerns that I should need clarification.”
  
  “1. The authors should clear point the novelty of this work compared, for instance with the Mos paper.”
  Using state-of-the-art spectrophotometric techniques (pH_T and A_T) over a highly-resolved 3 month timescale, our work compares carbonate system parameters (pH_T, A_T and C_T) measured in low salinity estuarine water with and without additions of HgCl₂. Mos et al., 2021 analyzed two of the three parameters where pH was measured using NIST scale (potentiometric) and A_T was measured using Gran potentiometric titrations on three occasions (at 0,1 and 6 months). Our observations of C_Tconstancy in HgCl₂-treated samples along with declining A_Tin HgCl₂-treated samples provides a substantially insightful view of CO₂ sample preservation processes.
  This information will be added to the revised manuscript.
  “2. What was the amount of HgCl2 added?”
  For HgCl₂-treated samples, 100 microliters of 6.5% HgCl2 was added to 300 mL sample bottles. This information will be added to the revised manuscript.
  “3. The authors are assuming that DIC changes are a consequence of DOC/DOM mineralization, which makes sense. However, there is no clear assumption that there is not evidence that there is no CO2 dissolution. Probably the work had this issue in consideration but a clear sentence on this issue should be had to the manuscript. Also, and still considering this problem, this work with benefit a lot from determinations of DOC in the analyzed samples.”
  Two outcomes lead us to believe that CO₂ dissolution was inconsequential in this work. Of the measured parameters sensitive to CO₂ dissolution that were unaffected by respiration of organic matter and complexation by Hg^II, HgCl₂-treated C_T measurements results remained essentially invariant (3,095.5 ± 3.2 µmol kg^-1) (Samples were analyzed in the order in which they were filled). A statement will be added to the manuscript addressing this point. Though this manuscript lacks determination of DOM concentrations, its presence was undeniable given the environmental setting, upstream and downstream measurements of TOC made by the EPCHC (See CC1), and carbonate system results for untreated samples that exhibited changes due to respiration of organic matter that are consistent with those of the Redfield equation.
  Although direct DOM measurements would somewhat alleviate uncertainties surrounding the significance of interactions between organics and Hg^II, future observations of A_T in organic-rich systems will be best served with the inclusion of titrimetically-derived organic acid/base concentrations and pK values.
  “4. pH is measured using a spectroscopic technique instead of a normal high-resolution ISE. Is there any specific requirements for using this method? More details about calibration should be provided.”
  Spectrophotometric pH_T measurements have been directly compared to potentiometric methods by repeated analyses at low ionic strengths and are shown to be more accurate (French et al., 2002). Spectrophotometric pH_T is also used to calibrate pH electrodes, including ISFETs, in seawater (Easley and Byrne, 2012) and estuarine waters (Martell-Bonet and Byrne, 2020). Spectrophotometric pH indicators are molecularly calibrated using calibration buffers and eliminate bias caused by variations in the liquid junction potentials of potentiometric pH systems. This information will be added to the revised manuscript.
  “5. In figure 3, there is a clear tendence of At decrease in HgCl2 treated samples. For untreated, the authors assume that the AVG At is constant. For the last ones it does not look that way. It seems that until day 49 the AT levels are constant but after a decrease is clear observed. Probably the use of a statistic test would clarify this doubt.”
  Line 168 – 169 states the significance test results of regression at 95% confidence level of p < 0.01 to statistically show that the A_T of untreated samples was constant. Also, the results show that the A_T for untreated samples is constant for a sufficiently long time period that untreated samples can be returned to the lab for analysis. A_T analysis of untreated samples in a suitably short time frame is clearly preferable to A_T analysis of HgCl₂-treated samples.
  “6. It is well known that Hg(II) easily complex with DOM. It is difficult to know how much Hg (II) is complexed (stochiometric all Hg can be complexed because concentration of DOM is higher thar Hg2+) and, consequently we will not know the DOM available for mineralization. Again, DOC measurements would benefit this work.”
  We agree that we do not know the extent of complexation by Hg^II. However this work shows for the first time, at high resolution, that Hg^II addition, which is the standard method of sample preservation for carbonate system water samples, rapidly affects the pH and A_T of estuarine water samples after collection. Also, this work suggests that organic alkalinity titrations (i.e., the fraction of DOM that actively exchanges hydrogen ions) performed after removal of all inorganic carbon, in conjunction with A_T measurements, is likely preferable to measurements of DOM in conjunction with A_T. Text will be added to the manuscript to highlight these conclusions and should benefit future work in this realm.
  “In summary, this technical note needs this clarification before being considered for publication.”
  New text and clarifications will be added to the manuscript to address the reviewers’ comments.
  
  Citation: https://doi.org/10.5194/egusphere-2022-1493-AC3

Christopher S. Moore, Robert H. Byrne, and Kimberly K. Yates

Data sets

CO2 system measurements in Hillsborough River Moore, C. S., Byrne, R. H., and Yates, K. K. https://doi.org/10.5066/P9J9IYFD

Christopher S. Moore, Robert H. Byrne, and Kimberly K. Yates

Viewed

Total article views: 692 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
461	190	41	692	25	27

HTML: 461
PDF: 190
XML: 41
Total: 692
BibTeX: 25
EndNote: 27

Views and downloads (calculated since 13 Jan 2023)

Month	HTML	PDF	XML	Total
Jan 2023	106	28	5	139
Feb 2023	125	30	6	161
Mar 2023	58	25	7	90
Apr 2023	18	8	0	26
May 2023	11	9	0	20
Jun 2023	10	3	1	14
Jul 2023	19	10	1	30
Aug 2023	9	10	0	19
Sep 2023	9	14	0	23
Oct 2023	7	3	0	10
Nov 2023	8	5	1	14
Dec 2023	4	2	0	6
Jan 2024	7	3	0	10
Feb 2024	6	3	0	9
Mar 2024	8	7	2	17
Apr 2024	10	6	4	20
May 2024	9	4	2	15
Jun 2024	7	2	5	14
Jul 2024	9	4	4	17
Aug 2024	9	2	2	13
Sep 2024	3	2	0	5
Oct 2024	4	4	1	9
Nov 2024	1	4	0	5
Dec 2024	4	2	0	6

Cumulative views and downloads (calculated since 13 Jan 2023)

Month	HTML	PDF	XML	Total
Jan 2023	106	28	5	139
Feb 2023	125	30	6	161
Mar 2023	58	25	7	90
Apr 2023	18	8	0	26
May 2023	11	9	0	20
Jun 2023	10	3	1	14
Jul 2023	19	10	1	30
Aug 2023	9	10	0	19
Sep 2023	9	14	0	23
Oct 2023	7	3	0	10
Nov 2023	8	5	1	14
Dec 2023	4	2	0	6
Jan 2024	7	3	0	10
Feb 2024	6	3	0	9
Mar 2024	8	7	2	17
Apr 2024	10	6	4	20
May 2024	9	4	2	15
Jun 2024	7	2	5	14
Jul 2024	9	4	4	17
Aug 2024	9	2	2	13
Sep 2024	3	2	0	5
Oct 2024	4	4	1	9
Nov 2024	1	4	0	5
Dec 2024	4	2	0	6

Viewed (geographical distribution)

Total article views: 690 (including HTML, PDF, and XML) Thereof 690 with geography defined and 0 with unknown origin.

Country	#	Views	%

Cited

Latest update: 30 Dec 2024

Short summary

This paper presents time series measurements of pH, dissolved inorganic carbon, and total alkalinity for organic rich estuarine water. The efficacy of sample preservation is examined by comparing results obtained with and without additions of the HgCl₂. Our results show that addition of HgCl₂ affects titration alkalinity and subsequently pH. Accurate analyses can be achieved using improvements in sample preservation techniques for carbonate system measurements in organic-rich estuarine waters.


Total:	0
HTML:	0
PDF:	0
XML:	0

Technical note: An Assessment of HgII to Preserve Carbonate System Parameters in Organic-Rich Estuarine Waters

Data sets

Viewed

Viewed (geographical distribution)

Cited

1 citations as recorded by crossref.

Technical note: An Assessment of Hg^II to Preserve Carbonate System Parameters in Organic-Rich Estuarine Waters