Variability in the Chemical Composition of Spring Waters in the Postomia River Catchment (Northwest Poland)

Reviewer 1 comments

Responses to the Reviewers

Major Comments

Showing the variability of the chemical composition in the form of an isoline map (depending on the land use) can be a good basis for further discussions.

Thank you for the comment.

Indicated way of presenting the results in the form of isolines is very interesting as it can show spatial variation. However, given the specificity springs location presented in Figure 1, the distribution is somewhat clustered. Therefore, the presentation for the entire catchment area in the form of isolines is subject to high uncertainty. In many cases, extrapolation of the results will be carried out, which is subject to high uncertainty. Such a presentation (isoline map) is more recommended for the results of studies obtained from monitoring networks distributed throughout the whole catchment.

Describing anthropogenic factors that may affect the condition of groundwater, surface runoff from communication routes is not mentioned (Introduction). Both winter road maintenance substances and pollution emitted by traffic can affect the quality of groundwater. What is the distance of springs from communication routes? Can road pollution affect the chemistry of spring waters?

Thank you for the comment.

As suggested, the introduction and discussion sections were completed regarding the impact of traffic roads on groundwater quality.

As suggested, an analysis of the distances of the sources to the roads was carried out. The results showed that these distances ranged from 2 to 160 m, with a mean value of 57 m. Correlation analysis showed only correlations between the distance of the springs to the road and the occurrence of sulphate in the waters (positive correlation) and alkalinity and HCO₃^- (negative correlation).

Introduction

Studies also demonstrate the negative impact of roads and road transport on groundwater quality (Uliasz-Misiak et al. 2022; Wang et al. 2018). Also, the effects of winter road maintenance substances on groundwater quality are highlighted (Jamshidi et al. 2020).

Methods

A correlation analysis was carried out to assess the impact of roads on spring water pollution. This analysis was intended to answer the question of whether the content of MEs, TEs and REEs is related to the distance of the spring from the road. In addition, an analysis of the differences in spring water contamination of MEs, TEs and REEs was carried out for two groups of sources located at distances up to 50 m (Group 1) and above 50 m (Group 2).

Results

The analysis of MEs, TEs and REEs content in springs was performed in the context of distance from roads. The springs are located between 2 and 160 m from roads with an average value of 57 m. Correlation analysis showed relationships between the distance of the spring to the road and the concentration of sulphate in the waters (positive correlation) and alkalinity and HCO₃^- (negative correlation). A final analysis of the significance of differences between the contents of MEs, TEs and REEs in the springs was performed for two groups - 1 (springs up to 50 m from roads – 14 springs) and 2 (springs more than 50 m from roads – 14 springs). The results showed significant differences at the 0.05 significance level for Cl^-, SO₄^2- and K⁺. Higher values of these ions were found in springs located more than 50 m from the road.

Discussion

The study showed that, in general, the influence of roads on the concentrations of MEs, TEs and REEs in spring waters was limited to only three elements (SO₄²-, alkalinity and HCO₃-). There were also no differences between spring water quality up to and above 50 m with the exception of K+, Cl- and SO₄²-. The researchers (Uliasz-Misiak et al. 2022) showed that road transport can affect the quality of groundwater up to 50 m from roads. Whereas Wang et al. (2018) show that Cu, Pb, Zn, Cr, Cd, Na⁺, K⁺ and Cl^- in groundwater were negatively affected by road transportation. The maximum affected distance of these pollutants varied from 15 to 100 m

2.3.Chemical analysis

Please specify the limit of determination, the accuracy of the method and the measurement uncertainty of the elements that have been determined.

Thank you for the comment.

Parameters for the ICP-QQQ method

Limit of determination (ppb) 0.0009-0.21

Accuracy of the method [%] 0.7-1.9

Uncertainty of measurement [%] 0.9-2.5

The LODs for the methods are listed in Table 1.

3.5. Statistical analysis

Some information should be presented more precisely.

Please write more details about "standardized chondrite patterns", explain how they were created and how they relate to the topic of this paper.

Thank you for the comment. Revised as suggested.

To determine the enrichment of spring waters with REEs, their concentrations were standardized in relation to the average concentrations of REEs in chondrite (Nakamura 1974). Then the results were visualised on graphs to show their pattern. This allowed the identification of anomalies, i.e. higher concentrations in relation to reference values and in relation to the preceding and following REEs. Anomalies in REEs concentration enable the identification of potential sources of pollution and the influence and type of natural processes [41].

Minor comments

Figure 3. should be corrected. The points cannot be distinguished.

Thank you for the comment. Revised as suggested.

Reviewer 2 comments

Responses to the Reviewers

Major Comments

There are numerous English errors. The manuscript needs to be edited by an English-speaking editor.

Thank you for the comment. It was revised as suggested.

The article has been proofread.

More information about the sites is needed. There is not enough geological information, e.g., what is the bedrock? What is the genesis of the springs? Have these springs been sampled before? The water chemistry results suggest to me that the springs are in a carbonate terrain; if so, are there other karst features in the study area? Any information on flowpaths, recharge zones, etc.? Has dye tracing been done for any of the springs? If this indeed a karst region, there is a large and robust literature about water quality in karst springs that the authors should utilize.

Thank you for the comment.

The geological characterisation of the study area was made in chapter 2 of the manuscript. The springs are located in the Polish Plain and are not karst springs. For the first time, water samples were taken for chemical analysis, where several indicators were determined. No dye tracking was performed during the study.

Single samples were collected from each spring, so the authors can’t account for any seasonality in the water quality data. Elevated nitrate levels in some of the springs suggest there could be seem important seasonal effects.

Thank you for the comment.

We collected one-time water chemistry samples from 28 springs in the Postomia region of Poland. Therefore, we did not characterise the variability of parameters over time.

The Results sections are poorly written. They consist primarily of listing the range and mean values of each constituent with little context. This is very tedious for the reader. These Sections would benefit greatly by creation of some data Tables, which would allow the reader to more easily compare the different samples. Then the text could focus on the most interesting/meaningful results.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised. The earlier description of the results has been removed an replaced by a more synthetic description as suggested by the reviewern. In addition, a table presenting the value of descriptive statistics has been included.

The REE data is probably the most exciting and novel part of the study, as these elements are rarely analyzed for. However, they are presented without much context, i.e., why were REEs analyzed for in the first place? What did you hypothesize they would tell you about your study area? Are there any anthropogenic sources in your study area? I’m also a little suspicious about finding statistically significant differences for REE data considering that they are at such low concentrations. Is a difference of 0.03 μg/L truly significant, even if your statistical test results say so?

Thank you for the comment. It was revised as suggested.

The description of REEs in the methods section has been added. Furthermore, it was highlighted that the study of REEs in spring waters has a pioneering character. In general, there are sources of REEs in the analysed catchment through those related to traffic. To determine the enrichment of spring waters with REEs, their concentrations were standardised in relation to the average concentrations of REEs in chondrite [46]. Then the results were visualised on graphs to show their pattern. This allowed the identification of anomalies, i.e. higher concentrations of reference values in relation to the preceding and following REEs. Anomalies in REEs concentration enable the identification of potential sources of pollution and the influence and type of natural processes [41].

It would be useful to calculate saturation indices for calcite and possibly dolomite to help interpret the ME data.

Thank you for the comment.

Saturation indices for calcite were calculated. The LSI

The index ranges from -0.029 to 1.1, indicating that the

spring waters are undersaturated (89%) and

supersaturated (11%) concerning calcium

carbonate (CaCO 3 ).

The Discussion reads like the explanations were taken out of a textbook. More emphasis on site conditions is needed.

Thank you for the comment. It was revised as suggested.

The author’s own results supplemented the discussion section. The results obtained were related to those obtained by other authors.

The Conclusions section reads more like a Summary. In my opinion, Conclusions should highlight the significant and novel aspects of the study results, not just re-hash the Results and Discussion sections.

Thank you for the comment. Revised as suggested.

Minor comments

12: Replace “North West” with “Northwest”.

Thank you for the comment. Revised as suggested.

14: Replace “contents” with “concentrations”. This error is found throughout the manuscript.

Thank you for the comment. Revised as suggested.

16: Delete “structure”. The cation is usually listed first, so change HCO3- - Ca2+ to Ca2+ - HCO3-.

Thank you for the comment. Revised as suggested.

21:       Not sure what “marked contents” is supposed to mean. Elevated concentrations?

Thank you for the comment. Revised as suggested.

22:       HMs not defined. Heavy metals?

Thank you for the comment. Revised as suggested.

47:       They can also be important ecologically.

Thank you for the comment. Completed as suggested.

59:       Geochemical and also biological processes.

Thank you for the comment. Completed as suggested.

68:       It’s not clear what region you are referring to. The region studied by Gao et al.?

Thank you for the comment.  The study by Gao et. al [16] presents temporal variations of spring water in karst areas: A case study of Jinan Spring Area, Northern China.

78:       I am not familiar with the term “alimentation area”. Are you referring to watershed?

Thank you for the comment. Revised as suggested.

80:       MEs and TEs need to be defined.

Thank you for the comment. Revised as suggested.

82:       Replace “abiotic” with “physical”.

Thank you for the comment. Revised as suggested.

86:       What does “left” tributary mean? Are you referring to a direction?

Thank you for the comment. The left” tributary means the left-bank tributary of the Warta River

100:     More geological detail is needed. How thick are the Pleistocene deposits? Are the aquifers confined? What is the bedrock? Is the bedrock used as an aquifer? Where do the springs form, i.e., do they form where the Pleistocene deposits meet the bedrock surface? It might be useful to show a geologic cross-section.

Thank you for the comment.

The geological characterization of the study area was made in section 2 of the manuscript. The springs are located in the Polish Plain and are not karst springs. For the first time, water samples were taken for chemical analysis, where such a number of indicators was determined. No dye tracking was performed during the study.

110:     Not sure what “no isolation from the surface” means. Does that mean the sandy deposits are at or near land surface?

Thank you for the comment.

Yes “no isolation from the surface” means that the sandy deposits are near the land surface.

115:     How were the springs selected for sampling? Do they all represent the same origin, i.e., the same geological factors?

Thank you for the comment. Revised as suggested.

All springs identified during the desktop study and field studies were selected for analysis.

137:     Section 2.4. I’m not sure this needs to be part of the main text, could probably move to supplemental material.

Thank you for the comment.

Section 2.4 Reagents and certified reference material is an important part of the methodology and should be part of the main text.

154:     Can you explain what is meant by “descending” and “ascending” springs?

Thank you for the comment.

Ascending spring - a spring to which water rises under the influence of pressure in rock cavities (pores or fractures) in the last section and emerges where the surface crosses the static water table or the aquifer below the water table.

Descending spring - a spring into which water flows under the action of gravity from the supply area down through the aquifer to the point of discharge.

156:     I’m curious how land use was determined for each spring. Was it where the spring discharged? Or did you also consider land use in the springshed? That to me would make more sense, and it might be likely that you’d have mixed land uses. Have there been any studies to delineate springsheds for any of the sampled springs?

Thank you for the comment.

The determination of springsheds was difficult, without a detailed geological survey. Therefore, the land use structure was defined as the dominant land use form in a 50m diameter circle around the source.

160: Eliminating outliers without a compelling reason seems like a bad idea. There should be a scientific reason to remove them, not just a statistical reason.

Thank you for the comment. The authors' previous studies show that outliers strongly influence the result of CA and PCA analysis.  Therefore, the outliers were not finally removed but, as stated in the methods, their values were replaced. According to Tabachnick and Fidell [47], outliers are easiest to eliminate by removing them from further analysis. However, especially in multidimensional statistical methods, their conversion should be considered. Elliott and Stettler [48] proposed to replace outliers with possible values. The outliers were replaced as follows: in the case of the highest val-ues, the second highest value was taken, and it was increased by 1%, while in the case of the lowest values, the second lowest value was taken, and it was reduced by 1%.

209:     That’s a huge range in pH, suggesting some major differences in hydrogeological terrain.

Thank you for the comment. The geological characterization of the study area was made in section 2 of the manuscript. The springs are located in the Polish Plain and are not karst springs. For the first time, water samples were taken for chemical analysis, where such a number of indicators was determined.

213:     In the previous paragraph you state the average pH was 7.67, here you say it was 7.70. Which is it? Would probably be better to report median values unless the data are normally distributed, which it doesn’t seem like they are.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised. The earlier description of the results has been removed and replaced by a more synthetic description as suggested by the reviewer. In addition, a table presenting the value of descriptive statistics has been included.

Fig. 2:  Do the whiskers represent the 10^th and 90^th percentiles, or the 5^th and 95^th? You should indicate the number of samples (N) in each location grouping in the caption. Is “agriculture” the same as “meadow”? In line 157 you state that there were only 4 springs in this category. I don’t think you can construct a box-and-whisker diagram with so few data points, can you?

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised. The earlier description of the results has been removed and replaced by a more synthetic description as suggested by the reviewer. In addition, a table presenting the value of descriptive statistics has been included.

222:     Section 3.2. In this section the data are described in an uninteresting way, basically going through each ME, reporting it’s mean and range with little context or interpretation. The reader gets bogged down in all the numbers in the text. A table showing all the data would be more useful for the reader than having you list every mean and range value. Looking at Figure 4, I think you should focus your discussion on those MEs where there is something interesting to show, e.g., nitrate and maybe chloride and sodium. For most of the other MEs, there doesn’t appear to be significant differences in concentration among the categories.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised. The earlier description of the results has been removed and, as suggested by the reviewer, replaced by a more synthetic description. In addition, a table presenting the value of descriptive statistics has been included.

223:     The cation is usually listed first, so change HCO₃^- - Ca²⁺ to Ca²⁺ - HCO₃^-.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised. The earlier description of the results has been removed and replaced by a more synthetic description as suggested by the reviewer. In addition, a table presenting the value of descriptive statistics has been included.

Fig. 3:  The piper diagram suggests a limestone terrain, is that correct? Are there karst features (other than springs) in the study area?

Thank you for the comment.

The geological characterisation of the study area was made in chapter 2 of the manuscript. The springs are located in the Polish Plain and are not karst springs. For the first time, water samples were taken for chemical analysis, where several indicators were determined. No dye tracking was performed during the study.

227:     Are you reporting nitrate as NO₃^- or as NO₃-N? Make it clear.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised.

231:     Report median instead of mean. This should be done for all parameters unless you can show that the data are normally distributed.

Thank you for the comment. It was revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised.

277:     Section 3.3. Same comments as for Section 3.2.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 have been revised.

321:     What do you mean by “extreme”? 12.6 μg/L for arsenic is not particularly elevated.

Thank you for the comment. It was revised as suggested.

332:     Section 3.4. Same comments as for Sections 3.2 and 3.3.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 has been revised.

367:     Section 3.5. Same comments as previous.

Thank you for the comment. Revised as suggested.

371:     I’m confused. In the previous sentence you say there were 12 outliers for MEs but here you say there were no outliers for MEs. All in all, I’m a little uncertain why so much emphasis is being given to outliers.

Thank you for the comment. Revised as suggested.

The analysis show that in physical parameters were 3 outliers and in the group of MEs - 12 outliers, the TEs - 16 outliers and the REEs - 15 outliers, respectively.

378:     It is still unclear to me why the outliers need to be removed. Is there something about the specific tests that are unable to satisfactorily deal with outliers?

Thank you for the comment. The authors' previous studies show that outliers strongly influence the result of CA and PCA analysis.  Therefore, the outliers were not finally removed but, as stated in the methods, their values were replaced. According to Tabachnick and Fidell [47], outliers are easiest to eliminate by removing them from further analysis. However, especially in multidimensional statistical methods, their conversion should be considered. Elliott and Stettler [48] proposed to replace outliers with possible values. The outliers were replaced as follows: in the case of the highest val-ues, the second highest value was taken, and it was increased by 1%, while in the case of the lowest values, the second lowest value was taken, and it was reduced by 1%.

381:     In line 156 you state there was only one slope spring. It seems to me you can’t include such a small group as a distinct population in your statistical tests.

Thank you for the comment. Revised as suggested. Statistical analysis was performed for valley springs, springs formed by river beds and scarp-foot springs. The analysis for all analysed elements in the distinguished spring types showed statistically significant differences between valley springs and scarp-foot springs.

383:     Again, it would be useful to have a table with the results presented instead of just in the text, much harder to visualize.

Thank you for the comment. Revised as suggested.

The text previously presented in subsections 3.1., 3.2, 3.3 and 3.4 and 3.5 has been revised. Moreover, Table 1 has been completed.

421:     Not sure elevated levels of TEs and REEs suggest anthropogenic contamination. It could reflect difference in geological strata. High levels of TEs are commonly found in naturally brackish waters.

Thank you for the comment. This study shows that anthropogenic sources have a decisive influence on elevated TEs and REEs. To show the influence of other factors further studies need to be carried out, in particular the geological structure. These studies have a preliminary character.

Fig. 8:  Not sure what I’m looking at here. What does chondrite normalized mean?

Thank you for the comment. The way REEs are normalised is that each measured value of REEs is divided by a reference value in this case in chondrite. The values of REEs in chondrite can be found in Taylor and McLennan (1985) are: La 0.367; Ce 0.957; Pr 0.137; Nd 0.711; Sm 0.231; Eu 0.087; Gd 0.306; Tb 0.058; Dy 0.381; Ho 0.085; Er 0.249; Tm 0.036; Yb 0.248; Lu 0.038. Then, diagrams were prepared for the normalised values to identify anomalies.

446:     Since HCO₃^- is calculated from alkalinity, a strong positive correlation is always to be expected.

Thank you for the comment. Yes such a correlation is obvious.

481:     The cation is usually listed first, so change HCO₃^- - Ca²⁺ to Ca²⁺ - HCO₃^-.

Thank you for the comment. Revised as suggested.

482:     In your study site section you don’t mention any rock or mineral types.

Thank you for the comment.

The geological characterization of the study area was made in chapter 2 of the manuscript. The springs are located in the Polish Plain and are not karst springs.

485:     Quaternary should be upper case.

Thank you for the comment. Revised as suggested.

488:     Unclear what is meant by “weakly acidic reaction are directly determined by underground waters, with no addition of mineral deposits in the headwater area”. Please clarify.

Thank you for the comment. This paragraph has been deleted.

505:     In my experience, elevated levels of NH₄⁺ are generally naturally occurring (from paloesols).

Thank you for the comment. Our studies shows that concentrations of NH4+ ions were high in waters of almost all springs. Elevated ammonium ion levels are obviously influenced by natural factors related to the presence of humic substances and decomposition of organic matter.

512:     Nitrate may indicate pollution, but not ammonium necessarily. Not sure what is meant by “permanent.” If the source(s) of contamination is reduced, then pollution should decrease.

Thank you for the comment. “Permanent” means having an impact over a long period of time with approximately the same intensity.

517:     Not sure what is meant by “irrational water and sewage management.” Improper?

Thank you for the comment. Revised as suggested.

537:     I would not call As concentrations of 12.6 μg/L particularly elevated. Especially considering how heterogeneous it can be.

Thank you for the comment. Revised as suggested.

A concentration of 12.6 μg/L is more extreme than elevated.

549:     While a source is needed, metals pollution usually requires appropriate geochemical conditions or the input of chelating agents.

Thank you for the comment. The objective of the study was to determine the sources of TEs REEs in the springs waters. Preliminary studies were also aimed at investigating the influence of land use on spring water contamination. The authors plan further studies which are intended to find out the detailed geological structure and to carry out a systematic study of water chemistry.

577:     The Conclusions are really more of a Summary of what we’ve just read. I would prefer to see Conclusions.

Thank you for the comment. Revised as suggested.

Reviewer 1 comments

Responses to the Reviewers

Minor comment

The triangles in the figure 2 shown overlap. The position of the points is poorly legible. It would be better if the points were marked with contour triangles rather than filling in the inside.

Thank you for the comment. Figure 2 has been revised according to the suggestion. The points were marked with contour triangles.

Reviewer 2 comments

Responses to the Reviewers

Minor comment

The number of agricultural sites (4) is probably too small to make meaningful comparisons with the other 2 land use types.

Thank you for the comment. However, as written in the methods section, the assessment of differences in the concentrations of MEs, TEs, and REEs in the designated groups of springs employed a non-parametric Kruskal-Wallis test coupled with a post-hoc Dunn test at a significance level of P ≤ 0.05. The statistical analysis takes into account the size of the samples to be compared.