the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Mar 2024
Support system for heat pump planning in response to drought conditions
Abstract. Access to extensive data enables advanced research on the impacts of climate change, specifically drought, on groundwater resources and their consequential effects on water quality. While scientific studies offer insights into predicting groundwater shortages at a high level, the data often remain inaccessible and incomprehensible to potential users. To bridge this gap, this study proposes a relatively straightforward method for assessing the risk of groundwater availability for heat pump usage. The method involves creating informative color-coded charts illustrating periods of potential excessive groundwater level decline near utilized wells. Additionally, it provides the ability to monitor changes in the risk of groundwater level reduction within a predefined observational period.
During droughts, groundwater levels can significantly drop, impacting groundwater availability and potentially reducing heat pump efficiency. This, in turn, may lead to system overheating, decreasing effectiveness, and causing damage. Exceeding critical groundwater levels may result in well infrastructure damage, affecting water quality and energy extraction efficiency. Excessive well exploitation often leads to chemical and mechanical clogging, further influencing well performance.
In contrast to commonly used hydrogeological drought indicators, this method focuses on a probabilistic model, simplifying calculations as it only requires historical groundwater level data. By applying statistical tests and distribution functions, the study evaluates the risk of extreme groundwater level reduction. The proposed method categorizes risk into very high, high, moderate, and low levels, providing a practical tool for users and groundwater management.
The study area, located in the northwest Eurasian continent, encompasses diverse geological and hydrogeological settings. Utilizing data from 27 groundwater observation points, spanning from 1980 to 2020, the research identifies periods and regions at risk of groundwater depletion. The findings highlight specific points vulnerable to high or very high risks, emphasizing the importance of groundwater management strategies.
By analysing monthly, quarterly, and seasonal risk variations and comparing results between the decades 2001–2010 and 2011–2020, the study unveils critical insights into groundwater dynamics. Points such as 15 exhibit pronounced risk increases, indicating potential overexploitation or insufficient replenishment. Notably, certain points display decreasing risks, showcasing positive trends that align with effective groundwater management practices.
This comprehensive probabilistic approach provides valuable information for stakeholders, empowering them to make informed decisions in selecting a sustainable energy source.
- Preprint
(759 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on hess-2024-41', Anonymous Referee #1, 21 Apr 2024
Review of HESS-preprint hess-2024-41 by Kubiz J and Karczewski M: Support system for heat pump planning in response to drought conditions.
The authors present a statistical analysis of variations of groundwater tables on the basis of multiple groundwater wells distributed over the Oder basin. They motivate their study mainly with the application and operation of heat pumps. The main argumentation line is that very low groundwater tables can lead to an enhanced risk of heat pump failure or at least will substantially reduce the heat pump efficiency. Hence, the authors focus their analysis on minimum groundwater tables over different time periods and analyse and visualise a certain risk of groundwater table minima falling below certain predefined thresholds. In order to communicate the associated drought risk to users the authors defined and applied a simple four-scale risk index visualising the likelihood of a given point in space and time to have a groundwater table minima are below a critical threshold.
While the general topic of drought induced groundwater table lowering and the potential impacts on various aspects of groundwater usage including the installation and operation of heat pumps is certainly of interest to readers of HESS, the applied method is in my opinion not sufficient to give robust results to support any planning. Furthermore, the way the methods are described as well as the discussion of the results have to be improved to match the quality of HESS.
Below I raise my major concerns related to the presented manuscript. I think they are serious enough to recommend, that the manuscript needs a major workover. Hence, I would suggest to reject the current form of the manuscript.
Major comments:
- Method –> given the main motivation of the work being the usage of heat pumps, an analysis of relative changes of groundwater tables only as well as relating groundwater minima to percentile thresholds seems to be not sufficient. In my opinion the method needs to be extended to consider the absolute changes as well. Basically an analysis solely based on relative changes gives the same weighting to wells with small absolute variations which might not be a big problem for heat pump application than wells that are characterized by large absolute groundwater table variations. Hence, it might be the case that wells with only small absolute changes might be characterised with larger risks than ones with large variations.
- Method –> my second point related to the method is the selection of the two hkr thresholds (70 and 90 percentiles, respectively). In my opinion they are purely subjective and it has to be confirmed by proper data analysis that these thresholds are good proxies for severe ground water drought as was mentioned in lines 157ff. Also the usage of the quantiles for risk classification needs to be motivated and discussed in much more detail.
- Data -> the description of the data is rather poor. Basically it does not become clear, what data exactly are available at the specific locations, e.g. temporal resolution, time period covered by the data, number of missing values, etc. I guess this would have been more interesting information to be included in Table 2 instead of other parameters listed, but never discussed (e.g. the subsurface properties or the location in the reservoir). Moreover, it is mentioned that the data is affected by gaps, and hence should not be seen as a time series (lines 155 ff). However, the extreme value analysis is based on time series since - at least to my understanding – the annual minima for e.g. the months or seasons have been selected and included. Hence, it would be of major importance to clarify, how many data points went into the analysis for each of the wells.
- Stationarity of data -> Given the fact that only stationary data was included into the analysis; I was surprised that the analysis of two subsequent decades revealed substantial differences in the associated risk of groundwater levels falling below certain thresholds. How can this be explained? Furthermore, it does not become clear what actually has been analysed in the case of the decadal minimum. Given the description in the manuscript, it would be the single minimum value only, however, I assume that it is probably the annual minimum values over the decade. But as already mentioned previously, a more detailed description of the data and data analyses is needed.
- Discussion of results –> In the current form, the manuscript basically contains no discussion of the results. Given the fact that the method itself inherits some rather subjective decisions/thresholds, the discussion should at least be used to tackle some of the uncertainties alongside the analysis chain, e.g. the variation in thresholds, variation of extreme value distribution (by the way, it does not become clear how the best distribution has been selected), absolute vs relative changes in the ground water table, etc. In its current from the chapter is mainly a description of the two tables.
- Language –> Although I am not a native speaker myself, I have the feeling that a major rewriting and language crosscheck would be needed for the manuscript. There are also quite a few typos in the manuscript and also some references are not cited in a correct way in the text.
Citation: https://doi.org/10.5194/hess-2024-41-RC1 -
AC1: 'Reply on RC1', Justyna Kubicz, 08 Aug 2024
- Method –> given the main motivation of the work being the usage of heat pumps, an analysis of relative changes of groundwater tables only as well as relating groundwater minima to percentile thresholds seems to be not sufficient. In my opinion the method needs to be extended to consider the absolute changes as well. Basically an analysis solely based on relative changes gives the same weighting to wells with small absolute variations which might not be a big problem for heat pump application than wells that are characterized by large absolute groundwater table variations. Hence, it might be the case that wells with only small absolute changes might be characterised with larger risks than ones with large variations.
We agree that the inclusion of just relative changes has its limitations, and some nuance might get lost. The goal of the paper was to present a general method that can give some insight into risk estimation but failed to present limitations of this method. We’ve included a paragraph in conclusions addressing this issue.
The inclusion of absolute values would give more detailed answer, but in our opinion the selection of weights is not necessarily straightforward, and would require more knowledge about specificity of certain well. Similar result might be achieved by selecting more rigorous quantile level. For example 99.9% for very stable wells. The selection of both of those values would have to be done on case by case basis.
- Method –> my second point related to the method is the selection of the two hkr thresholds (70 and 90 percentiles, respectively). In my opinion they are purely subjective and it has to be confirmed by proper data analysis that these thresholds are good proxies for severe ground water drought as was mentioned. Also the usage of the quantiles for risk classification needs to be motivated and discussed in much more detail.
In the world literature of recent years, we find results of groundwater drought studies based on various methods, including the threshold level method. Initially, this method was used to determine river low flows, but over time its application has been extended to groundwater levels. According to this method, a drought is a period in which the groundwater level is at or below an assumed threshold level [Hisdal et al. 2004, Fleig et al. 2006, van Loon 2013, 2015]. In the assumptions of the TLM method, the threshold for determining drought is related to the environmental element or economic sector affected by the drought, e.g. the amount of water needed for the proper functioning of water-dependent ecosystems, the amount of water needed for irrigation of agricultural areas, for industrial and energy purposes, or the amount of drinking water needed [Lloyd-Hughes 2014, Hisdal et al. 2004]. Determining the level of drought for each of these elements would be extremely difficult, mainly due to the lack of relevant information. Therefore, for practical reasons, thresholds are set for the analysed area or study point. This assumes a high correlation between surface water and groundwater. If the studies show that the selected quantiles are the best for indicating surface water scarcity, they will represent groundwater well. For Poland these are Q70 and Q90. [Strzebońska-Ratomska, 1994, Hisdal i in. 2004, Fleig i in. 2006, Tallaksen i in. 2009, Tokarczyk 2010].
- Data -> the description of the data is rather poor. Basically it does not become clear, what data exactly are available at the specific locations, e.g. temporal resolution, time period covered by the data, number of missing values, etc. I guess this would have been more interesting information to be included in Table 2 instead of other parameters listed, but never discussed (e.g. the subsurface properties or the location in the reservoir). Moreover, it is mentioned that the data is affected by gaps, and hence should not be seen as a time series. However, the extreme value analysis is based on time series since - at least to my understanding – the annual minima for e.g. the months or seasons have been selected and included. Hence, it would be of major importance to clarify, how many data points went into the analysis for each of the wells.
Each site had daily measurements of groundwater level. Given missing data we opted for calculating weekly averages. Table below shows the number of datapoints and year span for each site. On this dataset we estimated the distribution of monthly groundwater level minimas for each month, season separately for years 2000s and 2010s.
Calculation of quantile does not require time series data. Quantile is a common value derived from probability distribution, which can show values that can be considered high in specific scenarios. We do agree that having data in time series form can bring additional insight, but the purpose of this paper is to show what could be done with data that have certain limitations.
No of point
Location
Weeks available (year span)
No of point
Localisation
Weeks available (year span)
1
Sarbicko - 1
1159 (2001-2020)
15
Borówiec
1142 (2001-2020)
2
Sarbicko - 2
1159 (2001-2020)
16
Głazów
1037 (2001-2020)
3
Czachurki
1154 (2001-2020)
17
Jaskrów
1022 (2001-2020)
4
Szubin
1043 (2001-2020)
18
Radolin
1042 (2001-2020)
5
Chachalnia
892 (2001-2018)
19
Kamieńsk
998 (2001-2020)
6
Bogdaszowice
1034 (2001-2020)
20
Kochcice
1044 (2001-2020)
7
Złoty Potok
1044 (2001-2020)
21
Murzynowo
891 (2001-2018)
8
Żółwia Błoć
897 (2001-2018)
22
Słońsk
900 (2001-2018)
9
Gądno
930 (2001-2018)
23
Ujście
1034 (2001-2020)
10
Międzyzdroje
667 (2008-2020)
24
Stęszew
1027 (2001-2020)
11
Koszewko
622 (2005-2018)
25
Międzychód
1043 (2001-2020)
12
Dobrzyń
695 (2004-2018)
26
Turowo
1034 (2001-2020)
13
Lasów
703 (2004-2018)
27
Dźwirzyno
1044 (2001-2020)
14
Spore
1158 (2001-2020)
- Stationarity of data -> Given the fact that only stationary data was included into the analysis; I was surprised that the analysis of two subsequent decades revealed substantial differences in the associated risk of groundwater levels falling below certain thresholds. How can this be explained? Furthermore, it does not become clear what actually has been analysed in the case of the decadal minimum. Given the description in the manuscript, it would be the single minimum value only, however, I assume that it is probably the annual minimum values over the decade. But as already mentioned previously, a more detailed description of the data and data analyses is needed.
The reason for this peculiar outcome was uneven year span of those two stations. The change in groundwater levels in those years were enough to affect the significance of minima but not enough to affect stationarity. Since those are two different tests, the result while unusual is not impossible.
- Discussion of results –> In the current form, the manuscript basically contains no discussion of the results. Given the fact that the method itself inherits some rather subjective decisions/thresholds, the discussion should at least be used to tackle some of the uncertainties alongside the analysis chain, e.g. the variation in thresholds, variation of extreme value distribution (by the way, it does not become clear how the best distribution has been selected), absolute vs relative changes in the ground water table, etc.
The authors agree that the selection of thresholds in groundwater analysis often involves subjective decisions that can significantly influence the outcomes. These thresholds might include percentile values used to define extreme conditions or baseline levels for relative change calculations. The variability in choosing these thresholds can lead to different interpretations of groundwater risk. It is essential to conduct sensitivity analyses to understand how varying these thresholds affects the results and to ensure that the chosen thresholds are scientifically justified. Use in research the chosen thresholds based on literature or expert judgment helped in establishing their relevance. The use of percentile methods in predicting groundwater levels has advantages and disadvantages. Brenner et al. (2016) used percentile levels and the VarKarst model operated on a daily time step. He found that that percentile approach is able to reliably predict groundwater level exceedances across all considered time scales up to their 75th percentile. It fails when the 95th percentile of groundwater exceedance levels is considered.
Studies by Stebońska-Ratomska (1994) in Poland demonstrated the validity of using the 90th percentile in groundwater studies as an indicator of drought and scarcity. She also demonstrated the similarity between groundwater and surface water. She showed that similar thresholds can be used in groundwater studies as for surface water.This method has been in use with good results in other studies in Poland (Zdralewicz, Lejcuś 2008). The authors are aware that there is a need for further research into the reliability of the results. The percentiles chosen are only examples. It is recommended to adapt them to the conditions of the research site.- Language –> Although I am not a native speaker myself, I have the feeling that a major rewriting and language crosscheck would be needed for the manuscript. There are also quite a few typos in the manuscript and also some references are not cited in a correct way in the text.
The text was submitted for proofreading
Citation: https://doi.org/10.5194/hess-2024-41-AC1
-
AC1: 'Reply on RC1', Justyna Kubicz, 08 Aug 2024
-
RC2: 'Comment on hess-2024-41', Marina Mautner, 12 Jul 2024
This article develops a visual aid to represent groundwater decline potential at a given groundwater well with heat pump usage. The tool is meant to be useful in times of local drought and to help users avoid the negative effects on heat pumps of low-quality groundwater often encountered during excessive groundwater usage. The method uses historical groundwater level data and a probabilistic model to categorize current conditions into very high to low level risk. The article is accurately titled and presents a tool that is relevant for groundwater management practitioners in regions that use heat pumps as an energy source.
While the tool is an interesting approach, there does not seem to be any comparison done with between the proposed tool and the previous indicators that it is supposedly replacing. The study does not evaluate the performance of the tool in terms of the decision-making that occurs as a result of using the tool, which makes it difficult to evaluate whether or not the tool serves its purpose. Either the authors need to include this testing and analysis with the current article, or the article should be presented as a case study that evaluates the conditions of the study site only. The former option would make the article fit for review within this journal while the latter would be appropriate for a journal focused on case studies.
In terms of the quality of the manuscript, the flow of the introduction does not provide a clear picture of the previous research and placing of the current study within the literature. Specifically, there are quite a few statements about the motivation to conduct such research, however, it is not clear what other research on the development of this type of tool has been carried out and what current practices are in the field. Similarly, the article mentions the other indices that are currently used and that they require other inputs but does not list what these inputs are nor what the benefits of the previous indicators are. The introduction should include more information on what kinds of tools are best suited for practitioners, specifically, what is novel about providing levels of risk instead of some other prediction tool for groundwater quality? Additionally, why is it better to exclude more variables and focus on a single variable (groundwater levels)? Finally, what was the motivation for the probabilistic approach taken, what other studies have taken this approach, and what is their performance? There should be more references to research in the field of groundwater quality evaluation for heat pump usage for decision-making, not only references to describe the need for the tool generally. Overall, the introduction as is should be consolidated to spend less time on motivation while also expanded to improve the technical background and literature review regarding the methods proposed. I also suggest moving lines 73-78 starting with “Due to that,” to the end of the introduction and including a short description of the actual statistical approach used to develop the tool.
Please review all references and ensure that you are citing the original literature and that the statements made are an accurate representation of the findings of the source literature. For example, in line 40, the citation of Calero, et al., 2018 is not the source for the statistic provided. Additionally, the statement in Calero, et al., 2018 is regarding the potential for building rehabilitation to provide the most energy efficiency as compared to new construction, however, your statement in the introduction talks about where the most energy is consumed in a building. This is very concerning for two reasons. Firstly, the statement made is too general in scope, which is was what prompted me to check the reference material. And, secondly, it is the job of the authors to ensure that each statement is both substantiated by the material which is referenced and that the reference material is the original source of such information.
In terms of the method proposed, there are some choices by the author that are not clear and may be deficient in their rigor. It is not clear what exactly the authors mean when they refer to “selected sets of values of the minimum levels of groundwater” in the context of the periods chosen. For example, in the case of the monthly sets for January (“1”): does this mean that the minimum value in January is selected for each year and a distribution is made of those values? Or what would the selected set be for a given month? Similarly, how is the minimum set selected for a quarter, season, or decade?
In lines 157-159, are the authors stating that the critical level was calculated as the 70th and 90th percentile values using the historical distributions calculated as the critical levels? Why was this chosen as representative as a drought condition? Isn’t it possible that historically some wells are consistently in a drought condition while others never are? This would mean that for some wells using a value of the percentile would under or overestimate the drought condition as the entire distribution or none of the distribution would indicate a drought condition if this were true. Please explain the rationale for these values and how these cases would be handled in such a case. Additionally, an example of a specific well and a specific month, quarter, season, or decade when describing this method would help to improve the understanding of how the method is being applied.
Finally, it seems that the usefulness of the method would be to calibrate the method with actual impacts to heat pumps or well water quality and subsequent validation. If the method proposed is meant to indicate the risk, it is necessary to compare to the actual risk, not just an indicator related to the risk. However, since this data may be difficult to systematize, instead you may simply discuss how the drop in groundwater table is related to the quality of the water or the functioning of heat pumps generally in the region selected.
Minor comments:
Ensure that proper punctuation and spelling are used throughout, perhaps by using an outside editor. For example, in lines 51-52, the authors should use commas to separate each different type of energy source. In line 86 the comma following “is obtained” should be a period, while in line 159, following the parentheses, the period should be a comma. As well as in line 53, “carries” should be “carrier”. I have not included all examples of incorrect punctuation and spelling, please review the entire manuscript for these errors.
Table 2 should be included as a separate page in landscape format as the text is not readable in certain parts of the table.
It is not immediately clear what the colors of the rows in Table 4 and 5 refer to. Please comment in the text or caption that they refer to which season the month is in or perhaps instead just label the rows in a final column with the text rotated to indicate the season.
Citation: https://doi.org/10.5194/hess-2024-41-RC2 -
AC2: 'Reply on RC2', Justyna Kubicz, 08 Aug 2024
- The study does not evaluate the performance of the tool in terms of the decision-making that occurs as a result of using the tool, which makes it difficult to evaluate whether or not the tool serves its purpose. Either the authors need to include this testing and analysis with the current article.
We provide analysis of mean squared error and mean absolute error attached in the text. To calculate them with use of bootstrap we opted for calculating it for each month separately and then averaging them.
Table Fitting performance of the considered distribution using mean absolute error (MAE) and mean squared error (MSE)
No of point
Location
Mean absolute error (MAE)
Mean squared error (MSE)
Normal
Weibull
Gamma
Lognormal
Weibull3
Normal
Weibull
Gamma
Lognormal
Weibull3
1
Sarbicko
0.054
0.06
0.053
0.053
0.044
0.0045
0.0054
0.0044
0.0044
0.0032
2
Sarbicko
0.056
0.066
0.058
0.058
0.06
0.0053
0.0064
0.0058
0.0057
0.0063
3
Czachurki
0.049
0.03
0.067
0.079
0.025
0.0042
0.0017
0.0079
0.0115
0.0011
4
Szubin
0.042
0.028
0.065
0.085
0.033
0.0028
0.0012
0.007
0.0119
0.0021
5
Chachalnia
0.113
0.071
0.164
0.198
0.084
0.0231
0.0075
0.049
0.0726
0.0127
6
Bogdaszowice
0.055
0.051
0.06
0.06
0.046
0.005
0.0039
0.006
0.0062
0.0034
7
Złoty Potok
0.212
0.105
0.263
0.291
0.094
0.0596
0.0146
0.0908
0.1113
0.0122
8
Żółwia Błoć
0.083
0.054
0.109
0.127
0.047
0.0113
0.0044
0.0204
0.0285
0.0035
9
Gądno
0.076
0.024
0.084
0.081
-
0.0076
0.0008
0.0095
0.0086
-
10
Międzyzdroje
0.078
0.04
0.086
0.08
0.431
0.0085
0.0026
0.0101
0.0089
0.9626
11
Koszewko
0.071
0.055
0.074
0.076
0.132
0.0074
0.0049
0.0083
0.0088
0.0627
12
Dobrzyń
0.099
0.034
0.124
0.14
0.043
0.0215
0.0019
0.0363
0.0476
0.0037
13
Lasów
0.124
0.04
0.153
0.171
0.042
0.0247
0.0026
0.0395
0.0501
0.0033
14
Spore
0.023
0.019
0.027
0.031
0.023
0.0008
0.0006
0.0013
0.0017
0.0009
15
Borówiec
0.362
0.403
0.358
0.358
0.273
0.1981
0.203
0.1973
0.2013
0.1124
16
Głazów
0.079
0.037
0.1
0.114
0.039
0.0097
0.002
0.0165
0.0218
0.0031
17
Jaskrów
0.075
0.042
0.083
0.08
0.043
0.0092
0.003
0.0113
0.0105
0.0032
18
Radolin
0.022
0.02
0.024
0.022
0.019
0.0007
0.0006
0.0009
0.0007
0.0006
19
Kamieńsk
0.11
0.132
0.113
0.111
0.274
0.0242
0.0301
0.0244
0.0241
0.2803
20
Kochcice
0.1
0.068
0.114
0.114
1.333
0.0173
0.007
0.022
0.022
2.4393
21
Murzynowo
0.061
0.026
0.068
0.07
0.758
0.0056
0.0011
0.0068
0.0072
0.8528
22
Słońsk
0.06
0.025
0.068
0.068
0.024
0.0063
0.0009
0.008
0.0083
0.0009
23
Ujście
0.061
0.06
0.056
0.06
0.065
0.0058
0.0052
0.0048
0.0057
0.0066
24
Stęszew
0.043
0.02
0.049
0.051
0.233
0.003
0.0006
0.0041
0.0044
0.2369
25
Międzychód
0.128
0.044
0.147
0.156
1.317
0.0286
0.0032
0.0383
0.045
8.3143
26
Turowo
0.069
0.07
0.077
0.081
0.062
0.008
0.0076
0.0098
0.0111
0.0065
27
Dźwirzyno
0.031
0.021
0.038
0.04
0.02
0.0024
0.0007
0.0038
0.0048
0.0007
- The article mentions the other indices that are currently used and that they require other inputs but does not list what these inputs are nor what the benefits of the previous indicators are. The introduction should include more information on what kinds of tools are best suited for practitioners, specifically, what is novel about providing levels of risk instead of some other prediction tool for groundwater quality?
The main pro of this framework is its robustness. There are couple of open ended moments that could be adjusted to fit specific use cases. First, the selection of distributions. The user can pick any of available parametric and non-parametric distributions, and select those that fit the specific case of the analysed region. The second moment is the quantile selection which can be changed to better fit specificity of region. Low risk wells with low variability should have much higher quantile. This is something that should be considered for each use case.
- What is novel about providing levels of risk instead of some other prediction tool for groundwater quality? Additionally, why is it better to exclude more variables and focus on a single variable (groundwater levels)?
The risk-based approach is in line with modern risk management practices that are increasingly used in environmental and resource management.
Providing risk levels rather than using traditional groundwater level prediction tools introduces a novel approach that emphasises uncertainty management and decision support. Traditional prediction tools often provide deterministic results, which can be misleading if uncertainties are not made explicit.
By presenting levels of risk, we provide clearer information about the range of possible outcomes and their associated probabilities. This is particularly important in the context of groundwater levels, where variability and uncertainty can have a significant impact on resource management and planning, such as the operation of heat pumps that rely on stable groundwater levels.
Comprehensive multi-variable models often require extensive data input, which can be difficult to obtain and maintain. Focusing on a single variable reduces the burden of data collection and allows for more streamlined and efficient monitoring and forecasting processes. This is particularly beneficial in regions where data availability is limited.- What was the motivation for the probabilistic approach taken, what other studies have taken this approach, and what is their performance?
The motivation for adopting a probabilistic approach to groundwater research stems from the need to better deal with the uncertainties and variability inherent in hydrological systems.
By providing a range of possible outcomes and their associated probabilities, stakeholders can make more informed decisions, prepare for different scenarios and implement adaptive management strategies.- Please review all references and ensure that you are citing the original literature and that the statements made are an accurate representation of the findings of the source literature. For example, in line 40, the citation of Calero, et al., 2018 is not the source for the statistic provided. Additionally, the statement in Calero, et al., 2018 is regarding the potential for building rehabilitation to provide the most energy efficiency as compared to new construction, however, your statement in the introduction talks about where the most energy is consumed in a building.
Part deemed not required at this time. Citation removed.
- In terms of the method proposed, there are some choices by the author that are not clear and may be deficient in their rigor. It is not clear what exactly the authors mean when they refer to “selected sets of values of the minimum levels of groundwater” in the context of the periods chosen. For example, in the case of the monthly sets for January (“1”): does this mean that the minimum value in January is selected for each year and a distribution is made of those values? Or what would the selected set be for a given month? Similarly, how is the minimum set selected for a quarter, season, or decade?
For table 4 and 5 the monthly minima were calculated, and for each month probability distribution was estimated using all years. For table 6 we combined all observations throughout the decade.
Each site had daily measurements of groundwater level. Given missing data we opted for calculating weekly averages. Table below shows the number of datapoints and year span for each site. On this dataset we estimated the distribution of monthly groundwater level minimas for each month, season separately for years 2000s and 2010s.No of point
Location
Weeks available (year span)
No of point
Localisation
Weeks available (year span)
1
Sarbicko - 1
1159 (2001-2020)
15
Borówiec
1142 (2001-2020)
2
Sarbicko - 2
1159 (2001-2020)
16
Głazów
1037 (2001-2020)
3
Czachurki
1154 (2001-2020)
17
Jaskrów
1022 (2001-2020)
4
Szubin
1043 (2001-2020)
18
Radolin
1042 (2001-2020)
5
Chachalnia
892 (2001-2018)
19
Kamieńsk
998 (2001-2020)
6
Bogdaszowice
1034 (2001-2020)
20
Kochcice
1044 (2001-2020)
7
Złoty Potok
1044 (2001-2020)
21
Murzynowo
891 (2001-2018)
8
Żółwia Błoć
897 (2001-2018)
22
Słońsk
900 (2001-2018)
9
Gądno
930 (2001-2018)
23
Ujście
1034 (2001-2020)
10
Międzyzdroje
667 (2008-2020)
24
Stęszew
1027 (2001-2020)
11
Koszewko
622 (2005-2018)
25
Międzychód
1043 (2001-2020)
12
Dobrzyń
695 (2004-2018)
26
Turowo
1034 (2001-2020)
13
Lasów
703 (2004-2018)
27
Dźwirzyno
1044 (2001-2020)
14
Spore
1158 (2001-2020)
- Are the authors stating that the critical level was calculated as the 70th and 90th percentile values using the historical distributions calculated as the critical levels? Why was this chosen as representative as a drought condition? Isn’t it possible that historically some wells are consistently in a drought condition while others never are? This would mean that for some wells using a value of the percentile would under or overestimate the drought condition as the entire distribution or none of the distribution would indicate a drought condition if this were true. Please explain the rationale for these values and how these cases would be handled in such a case. Additionally, an example of a specific well and a specific month, quarter, season, or decade when describing this method would help to improve the understanding of how the method is being applied.
The quantile levels we included in our paper are propositions, and the robustness of the method makes it easy to use different values as needed. We agree that a single set of quantile values cannot capture the nuances of different well characteristics. Therefore, our goal was to develop a more robust method that allows for adjusting the risk level, for example, 99.9% for wells that are very stable. We included an additional paragraph in the conclusions addressing this issue. The selection of the value itself might be up for discussion and it will very much depend on wells characteristics.
The inclusion of just relative changes has its limitations, and some nuance might get lost. The goal of the paper was to present a general method that can give some insight into risk estimation but failed to present limitations of this method. We’ve included a paragraph in conclusions addressing this issue.
The inclusion of absolute values would give more detailed answer, but in our opinion the selection of weights is not necessarily straightforward, and would require more knowledge about specificity of certain well. Similar result might be achieved by selecting more rigorous quantile level. For example 99.9% for very stable wells. The selection of both of those values would have to be done on case by case basis.
- Finally, it seems that the usefulness of the method would be to calibrate the method with actual impacts to heat pumps or well water quality and subsequent validation. If the method proposed is meant to indicate the risk, it is necessary to compare to the actual risk, not just an indicator related to the risk. However, since this data may be difficult to systematize, instead you may simply discuss how the drop in groundwater table is related to the quality of the water or the functioning of heat pumps generally in the region selected.
A decrease in groundwater table can severely affect water quality, which in turn impacts the operation of heat pumps in a region. Lower groundwater levels can cause the intrusion of saline or mineral-rich waters into freshwater aquifers, leading to increased salinity. The sources of contamination, such as waste dumps and agricultural runoff, further exacerbate the problem by contributing pollutants that degrade the quality of the water (Brindha and Schneider 2019; Chełmicki, 2001, Parisi et al., 2023; van Gend et al., 2021).
Salinisation and hydrogen ions, chlorides, sulphates, and dissolved gases like CO2 and H2Salso which can be of both natural and anthropogenic origin, make groundwater corrosive, accelerating the deterioration of heat pump materials accelerates the degradation of metal materials in heat pumps. As a result, the efficiency of heat pumps decreases and maintenance costs rise (Nogara and Zarrouk 2018). The relationship between these factors underscores the importance of monitoring groundwater levels and quality. (Added in text).
Citation: https://doi.org/10.5194/hess-2024-41-AC2 - The study does not evaluate the performance of the tool in terms of the decision-making that occurs as a result of using the tool, which makes it difficult to evaluate whether or not the tool serves its purpose. Either the authors need to include this testing and analysis with the current article.
-
AC2: 'Reply on RC2', Justyna Kubicz, 08 Aug 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 263 | 58 | 36 | 357 | 20 | 25 |
- HTML: 263
- PDF: 58
- XML: 36
- Total: 357
- BibTeX: 20
- EndNote: 25
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1