Referee comment on preprint essd-2021-139

“Mineral, thermal and deep groundwater of Hesse, Germany”

General comments:

The authors assembled a large database on groundwater chemistry in the German federal state of Hesse with a focus on high mineralization, high temperature and deeply sourced groundwater, i.e. the subsurface waters with a geothermal and/or balneological application potential. The data was collected, processed and extended from literature including partly old publications with a very limited public availability. The database and the associated article document large efforts invested in its compilation and maintenance. The work is technically and formally sound (with minor exception, see specific and technical comments). To the reviewer´s knowledge, a suchlike database did not exist until now. Considering data sources and compilation efforts, this database is unique and hardy reproducible. I also see a high degree of usefulness here, not only for technical geothermal/balneological questions, but also for regional hydrogeological and hydrochemical approaches in general. Researchers and practitioners with an interest in e.g., a specific trace element, gas, or isotope system will find this database a valuable resource for an overview of existing data. The database is readily available for download in different formats. It is structured and formatted in a meaningful way, data sources are comprehensibly documented. Given the presented specific spatial information for the groundwater samples, export to – and use in –geographical information systems for regionalization exercises etc. are possible. Another big advantage is the (credibly promised) future extension of the database whereby the participation of both state and federal geosciences authorities is a real asset. It would be great to have such a database also for other federal states, or even the entire country.

In the article, the authors offer a comprehensive overview of the geology and hydrogeology of the federal state as a framework to enable the reader a proper evaluation and use of the data. They also give a meaningful example of potential database usage by statistically processing the dataset distinguished by hydrogeological regions of the federal state.

Specific comments:

P1, L26: “almost exclusively from publications” – where is rest of the data from? Consider reformulation.

P1, L27/28: “reduce uncertainties” – unclear in this context, please reformulate.

P2, L3: suggest to rather start like “The physical and chemical properties of groundwater determine its suitability…”

P2, L8: suggest to write “… we developed the database … of the federal state of Hesse, Germany.”

P2, L20: can you give a reference for the mentioned software?

P2, L25: recommend to generally omit “see” from references to figures, tables etc.

P2, L30: recommend to use the term “Southern German Scarplands” rather than “Cuesta Landscape”. It is the more widely used term, at least in my experience.

In the Introduction, the authors offer a very good and detailed regional geological and hydrogeological overview as a framework to the dataset, based on a thorough literature review (most of which is literature written in German of course which, however, makes this overview even more valuable for the international readership).

P6, L29: you talk about 100 m below ground level, I suppose. Please clarify.

The description of data compilation, sources and database structure in the Methods chapter is comprehensible.

P8, L1: not sure if “impoundment” is the right word here. May “sealing” or “packing” be more to the point? Please check.

Eq. 1: recommend to explain equation variables directly below the equation.

In the Results chapter, the authors offer a good description of data distribution and statistics in the different hydrogeological units defined in the federal state, and also mention some noteworthy regional peculiarities. It is a useful overview and example of the application capabilities and limitations of the presented database.

Data availability: the provided link worked without a problem, I was able to download the database and work with it.

P13, L8: “temporally” – do you mean in terms of covered stratigraphy? If so, better use that word.

Fig. 4: any explanation for the 50 °C outlier mentioned earlier?

Database itself: well-arranged, meaningfully formatted. Of course, not all information is available for all the samples due to differences in the (historical) records. But it will be very useful for many regional and supra-regional applications. The offered lithological and geological information (including regional geological/stratigraphical names) is valuable, as is the information on chemical groundwater type and quality of analysis.

In column K, I suggest to concretize “Altitude” – I guess it is the elevation of the drilling/sampling site?   

Technical corrections:

P1, L22: “In the future,”

P1, L28: “to estimate”

Fig. 1: I am missing the city of Kaiserslautern in the map… add, or omit from the caption. Also, add a green cross to the legend, or explain it in the caption.

P7, L6: “is available in Hesse”

P8, L11: omit “they”

P8, L26: “…balances … are marked …”

P12, L31: “Schäffer et al.”

Figs. 3-5: add axis labels!

Table 3: please check the used font, it is obviously different from the rest of the article.

I hope my comments and suggestions can help the authors to further improve this article.

Kind regards

This paper describes the collection process and organization of data about groundwater in Hesse state in Germany. The data includes valuable chemical, thermal, and regional data, with a long legacy, of about 200 yearts. This data is valuable  and may assist future scientific studies, as well as operational perspectives.

The paper is well written and well organised. The geological background of the region is well explained and the processes of data collection and organization are also explained.

The scientific novelty of the paper is limited, but I guess this type of paper is within the scope of ESSD.

In addition, the authors assume that many terms are understood by the readers, but I am not sure this is the case. For example, in Page 2, Line 15 you use the term  'formation water', without explbation. Same is for the term 'mineral water'. I think these terms are not broadly used and their scientific definition may not be clear for many readers. This is the case for many other terms. Please gp over the paper and make sure you use the common terms  / explain terms which are not trivial. 

In page 3, Line 7 you say 'mineralisation increases. What does it mean? increase of EC? TDS? Density? Please clarify.

P8, L1 – it is not clear what the 'Light grey' referring to….?

The authors present a large and detailed dataset of groundwater chemistry in Hesse. The dataset is compiled from many different sources and is, to my knowledge, unique. The manuscript gives a comprehensive and detailed overview of the region and enables readers without a strong background in hydro-geology to make use of the presented dataset.

I have a few minor comments and a few technical comments. These comments are followed by some questions and suggestions concerning the actual dataset.

With best regards,

Mirko Mälicke

Minor Comments:

• p.2 L.20 As a non-geologist I am not aware of the ‚GeoODin‘ software and further reference or a brief description would be helpful (maybe a footnote).
• p.p L.5-15 This clear description of the geological conditions could further be enhanced by a geological cross-section graph, if available. At least my imagination is not enough to picture this from the text, only.
• p.4 L.14-28: Is this detailed description general knowledge of the area or the result of some kind of geological survey(s), that could be referenced here?
• p.6 L.31 For me it is not clear, which ‚nine analyses‘ are referenced here and where the numbers come from.
• p.7 L.17 What exactly is a ‚secondary data source‘ in this context? Are these sources for auxiliary data/metadata/background information or the data itself? If the latter, I would suggest compiling an overview table of used sources for the appendix.
• p.8 L.2-3 If metadata was missing in the original reference, how could this information be added by the authors? Is this compiled from the ‚secondary data sources‘ (see comment above); if yes I would suggest to clearly make this connection here.
• p.8 L.8: Where does the ‚own information‘ about water type originate from?
• p.9 L.14 From my limited geo-hydrological understanding it is hard for me to imagine a use-case for borehole analysis data with missing depth information. How can this data turn helpful? Also, 0.6%, how many datasets are these?
• p.9 L.33: The authors state that no time series of measurements are considered. Is there a reason for this? Secondly, how exactly are the measurements reduced?
• p.9 L.34: Here, likely again due to my limited geo-hydrological understanding, it is not clear why a dataset with an electrical balance error larger than 5% is deleted (and not i.e. flagged). Does this imply that such a measurement has to be wrong, or is it just imprecise?
• p.10 L.1 After conditioning, only roughly half of the datasets were kept in the database. This seems like a lot of cleanup to me. It would be very helpful to understand, which of the three criteria given on p.9 L.32-34 ‚removed‘ most of the datasets. Were most datasets simply reduced (criteria 2.) or erroneous (1, 3)? Finally, if a considerable amount of the datasets have been deleted due to the 1st and 3rd criteria, did this affect mostly older datasets?

Figures:

• Figure 1: I am having some difficulties making use of this Figure. The sampling points are hard to identify and I almost overlooked them on first sight. I would suggest taking some contrast out of the figure, by using less-saturated fill colors for the geological units. Finally, I think it is enough to have Frankfurt marked on the map for orientation purposes. I found the amount of cities and the fact that they are abbreviated rather confusing.
• Figures 3,4,5: I think an axis label can make these figures even clearer
• Figure 3: Just out of curiosity: Do the authors have any possible explanation for the ‚cold‘ outlier?

Technical points:

• p.2 L.18 Please give the full name of ‚TDS‘ on the first occurrence.
• p.7 L.6: I think here the German ‚Hessen‘ is used accidentally, instead of ‚Hesse‘
• p.8 L.1 Please clarify the difference between ‚location‘ and ‚position‘ here.
• p.8 L.1 – 27 The authors present a lot of helpful information about the database here, and how the datasets are distributed. I think it would be helpful to add absolute amounts (of datasets) to the presented percentage numbers.

Dataset:

The dataset itself is presented in good shape and accompanied by a detailed description, that is in line with the presented manuscript. All relevant metadata can be found. The authors use a common format (xlsx) and added a CSV version of the dataset.

However, there are a few minor suggestions for improvement:

• Some of the cells in the Excel file have light grey or dark grey background. Although this is explained at the end of the dataset description, it would be convenient to have the explanation in the Excel file as well. Maybe as a comment, legend or extra page.
• The first row is empty and can be removed
• From l. 1041, the format of the table changes considerably. This makes it really complicated to read the file automatically (i.e. with R or Python).
• L.1046 – 1064: Here I have some difficulties understanding the presented data. It seems like this section should be a sheet of its own. The IDs start over with one again, which leads to duplicated IDs, which is not compatible with a relational data model. The same applies to the section from L.1066 to 1102.
• The column ‘Water containment system’ (column L) defines the value space as ‘(well, spring)’, but I can also find ‘Borehole’, which should be added. The values are capitalized, but the definition of the value space in the header is lowercase. It would be nice to keep this consistent.
• The column ‘Static water level’ has ‘meter’ as a unit, but I can find strings in the values (i.e. ‘Artesian’ or ‘Free overrun’). I would suggest keeping this column atomic and of single data type. If the static water level cannot be expressed in the specified unit, a possible path is to split it up into two columns.
• The column ‘Rock type’ (column R) contains question marks. Most other columns are kept empty if the information is not available. I would suggest either to keep this consistent or add an explanation of the difference if there is any.
• Columns U,V,W,X: these columns are just labeled ‘T.M,B,S’. For me, it is not clear what these columns specify. Maybe short information or a header can be added.
• Columns ‘Cations’ and ‘Anions’ (columns Y and Z) contain some broken formulas. I would suggest checking this formula, and remove the formula from cells that are intentionally empty
• The format of the date-times in column AB (Date of analysis) is not consistent.
• The column ‘Temperature’ (AG and AL) contains the string ‘n.d.’ and empty cells. I assume ‘n.d.’ is the abbreviation of ‘not defined’. How can the temperature be not defined? It can be not observed, like in the empty cells, but what exactly does ‘n.d.’ mean in this context?
• The columns with chemical parameters (AV – DT) are all defined to be of numeric data types, but occasionally contain strings (i.e. ‘Traces’ or ‘<0.5’). What is the difference between ‘Traces’ and ‘< xx’? I personally would interpret i.e. ‘ < 0.5’ (cell BK 350) as ‘below the detection limit of Nitrate’. But in BK 344 (same column) I also find ‘Traces’, which I personally would interpret in the same way. In BK 352 I find the number 0.22, which is also ‘<0.5’. I understand, that it is challenging to harmonize this information from different data sources, nevertheless, I think it is important to make this great dataset more readily usable. A user will have to harmonize this anyway if he wants to load the database into a GIS or using a programming language.
• From column EL to GE, there are a lot of broken formulas.
• First, I want to acknowledge that the authors added an ER diagram of the data model to the dataset, that’s really helpful. However, I am struggling with some details:
  • I think the relations are not really clear. What do they connect and what do they describe? Following the relations, a record of ‘Physical-chemical parameters’ relates a ‘Metadata’ by the fields ‘Temperature’ to ‘Thermal water’. First, I can’t find ‘Thermal water’ in the database and second, I really have some difficulties imagining this relationship. Is this relation a foreign key, like usually noted for database ER diagrams? How can the actual temperature be used for referencing entities? I feel like I don’t really get the idea here.
  • In the Metadata entity, the ‘Coordinates GK H’ self-references the UTM Northing. This is obviously not a self-reference in a technical sense, as the Gauß-Krüger Hochwert will never match the UTM Northing numerically. Additionally, whatever this relation is describing, does not apply to the GK Rechtswert and UTM Easting, as there is no relation indicated. The Easting is rather self-referencing the Altitude. Without further explanation, I cannot really make sense of this information.
  • I think I am completely on the wrong track interpreting the ER diagram, here. As a database engineer, I use ER diagrams a lot for database design and I might have a very narrow view on them. I guess there is only missing a short textbox explaining the meaning of an entity and a relation here and all confusion is cleared up. Maybe a legend is already enough.