the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Evaluation of a socio-hydrological water resource model for drought management in groundwater-rich areas
Abstract. Groundwater is a drought resilient source of water supply for many water users globally. Managing these highly-used groundwater stores is complicated by the episodic nature of droughts and by our limited understanding of water systems’ response to extreme events. Models are useful tools to simulate a range of prepared drought interventions, however, we need to ensure robust representation of surface water and groundwater storage, their users, and management interventions for drought resilience. A robust modelling approach is therefore essential for decision-making in groundwater management.
In this study, we present a Socio-Hydrological Water Resource (SHOWER) model for drought management in groundwater-rich regions. We evaluate SHOWER using a response-based and a data-based model evaluation in Great Britain which considers the modelling uncertainty, dynamic impact of management and modelling setups available. In the response-based evaluation, we first examined the model consistency with our understanding of the system functioning and the influence of modelled management scenarios on model simulations. Secondly, we tested the accuracy of heavily influenced discharge and groundwater level simulations in three catchments representative of typical hydrogeological conditions and water management practices in Great Britain (data-based evaluation). In the response-based method, we have found consistent simulations for all model setups and identified which parameters were influential to model output at what times. The data-based analysis shows that calibration can be focused on either source-specific or combined model outputs using a ‘best overall’ calibration approach that captures groundwater levels and low flows. The source-specific calibrations result in the highest and narrowest KGE ranges for discharge and groundwater (KGE: 0.75–0.84 and 0.62–0.95 respectively) with larger ranges using a `best overall' approach (KGE: 0.55–0.79 and 0.27–0.91). Integrated water management interventions have significant impact on flows and groundwater beyond parameter uncertainty and show leverage to reduce droughts by minimising shortages in water demand. With the modular and open-access structure of SHOWER we aim to provide a useful new tool for groundwater managers to explore their management interventions further, increasing drought resilience strategies using a robust modelling approach.
- Preprint
(8315 KB) - Metadata XML
-
Supplement
(5015 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1645', Dan Myers, 26 Apr 2025
-
AC1: 'Reply on RC1', Doris Wendt, 28 Oct 2025
This study did a thorough evaluation of the SHOWER model for modeling groundwater responses to water management scenarios in real catchments in groundwater-rich areas. I found the modeling approach to be robust, well-documented and technically sound. I believe it is exemplary work of interest and value for HESS readers. I have two main comments.
We thank the reviewer for their positive response to our work and welcomed the two main comments, see below.
My first comment: Clarify the knowledge gap. In the introduction, particularly around page 2, lines 40-75, and in Table S1, a variety of existing modeling approaches and limitations are introduced. At this point, it sounds like the paper simply combines and evaluates a groundwater model, rainfall-runoff model, and water management practices model, with a calibration including of management interventions in real catchments. If other models are already doing these, either individually or in combination, then the specific novel advancement of this work should be clearer. I suggest adding a little more background on specific models or cases that do similar things, potentially including those mentioned in Table S1 or models like SWAT-MODFLOW and ParFlow used in other areas. Then, describe more clearly how the SHOWER model and/or the analyses in this study go beyond the previous work to fill a specific knowledge gap (perhaps something related to water management and droughts in real catchments).
We appreciate the concerns raised by the reviewer and believe that we can address this by clarifying what can be modelled well using the current modelling tools available and what we think is missing given the highly managed groundwater systems in the UK. For example, there is limited water management influence included in most of the listed rainfall-runoff models in lines 41-42 and most of the recent modelled water management influence focuses on surface water abstractions, for example in Rameshwaran et al. (2022) or on reservoirs by Salwey et al. (2024). Groundwater is assumed to be largely uninfluenced by abstractions, responding as it would do in a natural system, which is not the case for many regions in the UK particularly in drought conditions (Wendt et al. 2020).
We intended to introduce SHOWER to fill this research gap as it has a simple rainfall-runoff structure that includes (drought) management strategies for both surface water and groundwater stores and fluxes. Another key advancement is that we incorporated different groundwater model options to better simulate groundwater flow in large aquifers in the UK that are used for drinking water supply. With the three different options, we could adapt baseflow generation to account for karstic, fracture or largely porous flows instead of using a linear baseflow release that is currently implemented in other rainfall-runoff models. These adaptations have shown improvements in groundwater-rich areas in the UK and abroad (Wendt et al. 2021; Stoelzle et al. 2015). We will rephrase the second paragraph in the introduction to clarify the research gap and bring in more examples from Table S1.
My second comment: Highlight results beyond evaluations. The study is very heavy on the technical aspects, and the results are essentially model evaluations, without highlighting further scientific theories or comparisons tested. This is evident in the title (“Evaluation of…”), as well as the methods and results section headers which only go to calibration and evaluation. I suggest bringing the results beyond model evaluation, uncertainties, and parameter sensitivities more prominence. As an idea for this, you could state and test a scientific hypothesis about the groundwater and water management interactions, like a case study in the real catchments to show a scientific application of the model. The paper is already set up with different management scenarios that are tested to see their hydrologic impacts on droughts compared to baseline (e.g., Figures S9 to S11, lines 17-18 in the abstract, and lines 455-458 in the conclusion), so it is possible that no new analyses need to be done, just reframing. Perhaps you could create a hypothesis in the last introduction paragraph (Third,…) about different management scenarios affecting things like drought duration and deficit in different geologies, and then put all these findings into a 3rd results section with a header clearly beyond evaluations (like “3.3. Management scenarios and drought impacts”). This could help show the theoretical advancements in real catchments to potential users of the model.
We greatly appreciate the suggestions provided by the reviewer, in previous versions of the manuscript we had contemplated a similar setup although we moved away from this idea given that the influence of drought management scenarios is already investigated using the Global Sensitivity Analysis (GSA). The aim of the GSA is to investigate 1) changes in parameter sensitivity of discharge and groundwater storage over time, 2) the overall sensitivity of discharge and groundwater to management scenarios and 3) implications for droughts when applying management scenarios (Lines 169-171).
A generic application of the drought management strategies to the case studies seemed therefore double and not the right fit, as water managers in the case study areas are likely to apply various combinations of the modelled drought management scenarios at the same time, whereas (for testing purposes) we applied these scenarios one at the time. Plus, the existing significant pieces of work, the GSA and case study application, make a lengthy paper and therefore we aimed to reduce any duplications.
However, we do appreciate the suggestion and acknowledge that the current result section of the GSA covers the first and second analysis in much detail but does not elaborate much on the effect of drought management scenarios on surface water and groundwater much in the short paragraph (Lines 286-293). Instead of creating the suggested additional header at the end of the result section, we will integrate these suggestions within section 3.1 adding to the third component of the GSA. We will elaborate in more detail on the consequences of the identified leverage in Figure 5, because two quite different aspects of leverage of the integrated scenarios are evident (particularly the conjunctive use scenario), which results in quite different consequences for groundwater drought characteristics in the three groundwater modules (now in S8-S10 in Supplementary material). In the revised paragraph we will elaborate more on this difference in leverage and the consequences for drought characteristics either the karstic and porous module compared to the fractured groundwater module (see also the first point of Reviewer 2).
Minor Comments
We appreciate the provided minor comments and will address them in the full rebuttal.References:
Rameshwaran, P., Bell, V. A., Brown, M. J., Davies, H. N., Kay, A. L., Rudd, A. C., & Sefton, C. (2022) Use of abstraction and discharge data to improve the performance of a national-scale hydrological model. Water Resources Research, 58, e2021WR029787 https://doi.org/10.1029/2021WR029787
Salwey, S., Coxon, G., Pianosi, F., Lane, R., Hutton, C., Bliss Singer, M., McMillan, H., and Freer, J. (2024) Developing water supply reservoir operating rules for large-scale hydrological modelling, Hydrology and Earth Systems Science, 28, 4203–4218 https://doi.org/10.5194/hess-28-4203-2024
Stoelzle, M., Weiler, M., Stahl, K., Morhard, A., and Schuetz, T. (2015) Is there a superior conceptual groundwater model structure for baseflow simulation? Hydrological Processes, 29, 1301–1313 https://doi.org/10.1002/hyp.10251
Wendt, D. E., Van Loon, A. F., Bloomfield, J. P., and Hannah, D. M. (2020) Asymmetric impact of groundwater use on groundwater droughts, Hydrology and Earth System Sciences, 24, 4853–4868 https://doi.org/10.5194/hess-24-4853-2020
Wendt, D. E., Bloomfield, J. P., Van Loon, A. F., Garcia, M., Heudorfer, B., Larsen, J., and Hannah, D. M. (2021) Evaluating integrated water management strategies to inform hydrological drought mitigation, Natural Hazards and Earth System Sciences, 21, 3113–3139 https://doi.org/10.5194/nhess-21-3113-2021
Citation: https://doi.org/10.5194/egusphere-2025-1645-AC1
-
AC1: 'Reply on RC1', Doris Wendt, 28 Oct 2025
-
RC2: 'Comment on egusphere-2025-1645', Anonymous Referee #2, 16 Jul 2025
The authors present the evaluation of a modelling tool for drought management that incorporates groundwater as well as abstractions and water management. The model evaluation is quite thorough and this is described in much detail. I think the manuscript is suitable for HESS and will provide a valuable contribution after some minor revisions.
My main issue is with the fact that the main aims and contribution are not very clear. The manuscript reads more like a report on a modelling tool and the main contribution to the existing literature on socio-hydrology/groundwater modelling, is not highlighted very well. What is the main contribution, is it the integration of groundwater or the integration of management scenarios? The results and discussion focus on general model performance, and there is not much focus on the management scenarios and how they influence water availability and drought characteristics. Especially given the title of the manuscript and the emphasis on the model as a tool for groundwater managers and to “evaluate the impact of groundwater abstraction strategies on hydrological droughts”, I would expect more discussion on how well the management actions and socio-hydrological feedbacks are captured and some results on how the different management strategies influence droughts.
Linked to this, I would like to see a bit more detail on how the management strategies are implemented in the model. Also, it would be good to see the socio-hydrological processes represented in more detail in figure 1. In its current form. it looks like the model is an extensive hydrological model but not a socio-hydrological model, with management just included as scenarios, rather than feedbacks between the hydrology and decision making.
I think this is just a matter of changing the focus of the results and rewriting the introduction and discussion so that it better highlights what the main contribution of this manuscript is.
The manuscript could do with a grammar check and some clearer language. Some examples:
- Line 71-74, it is not entirely clear to me what will be compared with what.
- Line 129: “These scenarios are modelled the following.”
- Line 131, what is meant by “integrates water demand between the stores”?
- Line 391: “ Simulated discharge and groundwater results show a significant influence on simulated discharge and groundwater”.
- Line 435: ≤ should be ≥?
- Line 447: “We also identified (un)influential parameters in the GSA that aided calibration when applying SHOWER to three case study areas in the UK.” Do you mean the parameters aided calibration or that the fact that you identified these parameters has aided the calibration? How has this helped the calibration? This wasn’t clear to me from the methods or results section.
These are just a few examples, the whole manuscript would benefit from a thorough grammar check, because there were many sentences that are difficult to understand.
In Lines 229-230 the authors talk about a step-change around 50-60 simulations, however the figures in the supplement that are referred to seem to indicate this is around 40-50 simulations. In addition, it is not clear to me what is meant by 40-50 simulations. Are these just the 40 best performing parameter combinations? In that case, what does this step change mean? Is it just a strange artefact, or is it related to some change in the parameter values? The relevance of this step change is not clear to me.
The numbers in the legend for the relative sensitivity in Figure 3 are not readable.
The data availability statement mentions that the model code is available from DEWendt/SHOWER, but I am unable to find this.
Citation: https://doi.org/10.5194/egusphere-2025-1645-RC2 -
AC2: 'Reply on RC2', Doris Wendt, 28 Oct 2025
The authors present the evaluation of a modelling tool for drought management that incorporates groundwater as well as abstractions and water management. The model evaluation is quite thorough and this is described in much detail. I think the manuscript is suitable for HESS and will provide a valuable contribution after some minor revisions.
We thank the reviewers for their positive views on the work and the constructive suggestions. In this author response, we have indicated how we plan to revise the manuscript following these suggestions.
My main issue is with the fact that the main aims and contribution are not very clear. The manuscript reads more like a report on a modelling tool and the main contribution to the existing literature on socio-hydrology/groundwater modelling, is not highlighted very well. What is the main contribution, is it the integration of groundwater or the integration of management scenarios? The results and discussion focus on general model performance, and there is not much focus on the management scenarios and how they influence water availability and drought characteristics. Especially given the title of the manuscript and the emphasis on the model as a tool for groundwater managers and to “evaluate the impact of groundwater abstraction strategies on hydrological droughts”, I would expect more discussion on how well the management actions and socio-hydrological feedbacks are captured and some results on how the different management strategies influence droughts.
We appreciate the comment that the manuscript is quite model-heavy at the moment. This is likely a consequence of us attempting to not lengthen the manuscript unnecessarily. In the revised document, we will rephrase section 2.1 that introduces the management strategies (lines 122-136).
Additionally, we will link the initial impact of the drought management strategies in Figure 5 (lines 272-277 for the karstic and porous module and lines 278-285 for the fractured & reservoir module) to the consequence for drought characteristics (Figure 6) and detailed findings per scenario for each groundwater module (Figure S9-S11). We will integrate these comments in the last section of the GSA analysis (Section 3.1) and discuss in more detail how the management actions translate to drought characteristics for the three groundwater modules. See also Reviewer 1, second comment.
We will also review section 4.2 (Discussion) to further elaborate on the influence of management scenarios.
Linked to this, I would like to see a bit more detail on how the management strategies are implemented in the model. Also, it would be good to see the socio-hydrological processes represented in more detail in figure 1. In its current form. It looks like the model is an extensive hydrological model but not a socio-hydrological model, with management just included as scenarios, rather than feedbacks between the hydrology and decision making.
I think this is just a matter of changing the focus of the results and rewriting the introduction and discussion so that it better highlights what the main contribution of this manuscript is.We agree that more detail is needed in section 2.1 (methods) to introduce the drought management strategies and later in section 3.1 how these scenarios are reflected in the results. We believe that the model is an extensive hydrological model that can be used to understand management feedbacks in the hydrological cycle. For example, when maintaining low flows, groundwater use is restricted only when flows fall below a certain threshold. This conditional use of groundwater storage results in higher flows and less groundwater use, with a larger dependency on stored surface water or imported surface water. These management strategies are indeed defined based on common drought management strategies in Wendt et al. (2021). We will make sure that this is clearer from the introduction and discussion as suggested.
The manuscript could do with a grammar check and some clearer language. Some examples:
- Line 71-74, it is not entirely clear to me what will be compared with what.
- Line 129: “These scenarios are modelled the following.”
- Line 131, what is meant by “integrates water demand between the stores”?
- Line 391: “ Simulated discharge and groundwater results show a significant influence on simulated discharge and groundwater”.
- Line 435: ≤ should be ≥?
- Line 447: “We also identified (un)influential parameters in the GSA that aided calibration when applying SHOWER to three case study areas in the UK.” Do you mean the parameters aided calibration or that the fact that you identified these parameters has aided the calibration? How has this helped the calibration? This wasn’t clear to me from the methods or results section.
These are just a few examples, the whole manuscript would benefit from a thorough grammar check, because there were many sentences that are difficult to understand.
We thank the reviewer for their careful reading and will make sure to double-check the spelling and grammar in the next version of the manuscript.
In Lines 229-230 the authors talk about a step-change around 50-60 simulations, however the figures in the supplement that are referred to seem to indicate this is around 40-50 simulations. In addition, it is not clear to me what is meant by 40-50 simulations. Are these just the 40 best performing parameter combinations? In that case, what does this step change mean? Is it just a strange artefact, or is it related to some change in the parameter values? The relevance of this step change is not clear to me.
This is a good point raised by the reviewer that we have investigated whilst doing the analyses -but should have been phrased clearer in the manuscript. When doing 10k simulations, finding the ‘best’ simulations can be achieved in a multitude of ways, which partly depends on the used evaluation criteria. However, defining the exact number of ‘best’ simulations is arbitrary, as this could be the top 20, top 40, top 50 or top 100. When identifying a step-change in the performance metric, meaning we identified a shift in performance between 40-50 and 50-60 simulations of around 0.05 NSElog for the Pang and Trent catchments rather than a gradual decline in performance. Hence we decided that it would be most sensible to use the top 50 to show the best simulation results. This should be clearer in Lines 229-230 where the typo adds confusion (it should state 40-50 simulations). Elsewhere in the text, we do refer to the top 50 simulations. We will double-check the supplementary figures and investigate if the step-change is indeed caused by parameter values and will change the text and supplementary material if we can identify a particularity in parameterisation.
The numbers In the legend for the relative sensitivity In Figure 3 are not readable.
We will make sure the legend is readable in the next version of the manuscript. Thanks for letting us know.
The data availability statement mentions that the model code is available from DEWendt/SHOWER, but I am unable to find this.
My apologies, as this repository is set to be public and therefore should be accessible. I will add the full link to the data availability section to avoid confusion. Please find the link to the repository here: https:/github.com/DEWendt/SHOWER : Operational Drought modelling tool for based on a socio-hydrological water resource modelling
Citation: https://doi.org/10.5194/egusphere-2025-1645-AC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 821 | 100 | 20 | 941 | 43 | 33 | 36 |
- HTML: 821
- PDF: 100
- XML: 20
- Total: 941
- Supplement: 43
- BibTeX: 33
- EndNote: 36
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study did a thorough evaluation of the SHOWER model for modeling groundwater responses to water management scenarios in real catchments in groundwater-rich areas. I found the modeling approach to be robust, well-documented and technically sound. I believe it is exemplary work of interest and value for HESS readers. I have two main comments.
My first comment: Clarify the knowledge gap. In the introduction, particularly around page 2, lines 40-75, and in Table S1, a variety of existing modeling approaches and limitations are introduced. At this point, it sounds like the paper simply combines and evaluates a groundwater model, rainfall-runoff model, and water management practices model, with a calibration including of management interventions in real catchments. If other models are already doing these, either individually or in combination, then the specific novel advancement of this work should be clearer. I suggest adding a little more background on specific models or cases that do similar things, potentially including those mentioned in Table S1 or models like SWAT-MODFLOW and ParFlow used in other areas. Then, describe more clearly how the SHOWER model and/or the analyses in this study go beyond the previous work to fill a specific knowledge gap (perhaps something related to water management and droughts in real catchments).
My second comment: Highlight results beyond evaluations. The study is very heavy on the technical aspects, and the results are essentially model evaluations, without highlighting further scientific theories or comparisons tested. This is evident in the title (“Evaluation of…”), as well as the methods and results section headers which only go to calibration and evaluation. I suggest bringing the results beyond model evaluation, uncertainties, and parameter sensitivities more prominence. As an idea for this, you could state and test a scientific hypothesis about the groundwater and water management interactions, like a case study in the real catchments to show a scientific application of the model. The paper is already set up with different management scenarios that are tested to see their hydrologic impacts on droughts compared to baseline (e.g., Figures S9 to S11, lines 17-18 in the abstract, and lines 455-458 in the conclusion), so it is possible that no new analyses need to be done, just reframing. Perhaps you could create a hypothesis in the last introduction paragraph (Third,…) about different management scenarios affecting things like drought duration and deficit in different geologies, and then put all these findings into a 3rd results section with a header clearly beyond evaluations (like “3.3. Management scenarios and drought impacts”). This could help show the theoretical advancements in real catchments to potential users of the model.
Minor comments:
Title: Consider updating the title to clarify the more novel scientific findings. Evaluation of a model in itself does not suggest the scientific advancement of the work to me, and I think it has the potential to be more generally impactful. Perhaps something like “Socio-hydrological model reveals how water resource management affects drought duration and deficit in real groundwater-rich catchments”
Page 2, lines 29-34. Consider condensing this to basically say highly managed groundwater systems are present around the globe, rather than elaborating on the individual regions. Though it’s fine with me if you prefer to keep it as-is.
Page 20, lines 360-370. I’m not sure what the scientific value of showing that the SHOWER model had similar or better discharge performance to other models is, particularly if those models were simulated using different approaches for factors like input data, time periods, observations, or calibrations. I suggest condensing this.
Page 22, lines 424-425. I suggest rewording as something more direct, like “However, this simplification creates the opportunity for modelers with insufficient time and computing resources for more expensive models to be able to calibrate the model and explore results in detail.”
Page 23, lines 455-458. I really like this conclusions statement. It shows the scientific implications and exciting capabilities of the work well.
I wish the authors the best with this manuscript and their future endeavors.