the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Jan 2025
Regional-scale groundwater analysis with dimensionality reduction
Abstract. Given the importance of groundwater for freshwater provision and groundwater-dependent ecosystems, understanding climate effects on groundwater changes at a regional scale is essential. In this paper, we propose a new way of applying dimensionality reduction for such purpose, not over the collected data, nor over any calibrated models, but over the misfits between the modeled and observed groundwater levels. This methodology highlights local differences in climate-groundwater relations and can be used to identify regions with different vulnerabilities in a data-driven way.
The approach takes gridded groundwater level data and gridded precipitation and evapotranspiration data as input. Linear water balance models are set up for each grid cell in an independent way. The misfits between the water balance model simulations and groundwater levels are used for the dimensionality reduction-based regionalization, with which areas of different groundwater behavior are identified.
We demonstrate the potential of our methodology in the Berlin-Brandenburg region, Germany, where groundwater is a major freshwater source at risk. We show that groundwater level changes are linearly related to climatic variations at a monthly scale, even in areas with strong anthropogenic influences. The dimensionality reduction further reveals an approximate regionalization of groundwater behavior, which can be used as a basis for more detailed investigations.
Competing interests: At least one of the authors is guest editor of the special issue "Current and future water-related risks in the Berlin–Brandenburg region".
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(1753 KB) - Metadata XML
-
Supplement
(218 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2024-4031', Hesam Soltan Mohammadi, 10 Mar 2025
- AC1: 'Reply on RC1', Márk Somogyvári, 27 Jun 2025
-
RC2: 'Comment on egusphere-2024-4031', Anonymous Referee #2, 17 May 2025
The manuscript of Somogyvári et al. presents a methodology for investigating groundwater dynamics at the regional level, applying dimensionality reduction on the differences between a set of groundwater level observations in the Berlin-Brandenburg region (Germany) and the respective levels simulated by a simple water balance model. Though using a subsurface water model based on the physically-based concept of mass balance, the approach is classified as data-driven since it needs several observations, which makes the method hardly applicable in data-scarce areas.
The methodology presented is sound overall, and the results are clear enough. They highlight that linear models can be applied to describe the response of the groundwater dynamics to the driving climate input for most of the analysed region. At the same time, the areas needing further studies are detected, even though their dynamics are not disentangled.
Overall, the manuscript represents an interesting methodological contribution to regional-scale groundwater analysis in regions with good data availability. It can be published after the following minor comments are addressed.
L54: “Assessing the groundwater behavior at a regional/mesoscale level is still relatively understudied”. Even though the authors already added a reference, this statement could be more strongly justified.
Figure 1: I believe that this figure needs to be completely revisited. In the text, especially in the Results and Discussion section, several toponyms are used, which make it very difficult for the reader to follow the thread of reasoning. All these toponyms must be shown clearly in a much more prominent and clearer Figure 1a. In addition, a third panel should be added (let’s say, figure 1c), with appropriate geolithological information related to the hydrogeology of the region analysed.
Section 2.1: It is explained that five separate aquifers at different depths can be identified in the analysed region. The study addresses only the top unconfined one. It is unclear if/how the methodology adopted could also be used for the lower aquifers, presumably less affected by climate drivers. Please explain and discuss.
Section 3.1 (and, consequently, Section 4): It is unclear how much the kriging algorithm (i.e., universal kriging with spherical semivariogram) influences the results achieved, especially for the zones with fewer observations. Please explain and discuss.
Section 4.1: Looking at Figure S1, one can quickly agree that a one-month time lag is the most suitable in general (even though I would have used a discrete color bar rather than continuous since only integer values are in the map). However, some zones diverge from this general rule. It is unclear if the method could consider different time lags for different zones in the same study area.
Fig. 5: The maps are difficult to interpret, and the colors are quite confusing, I think, particularly for color-blind people.
L417: “…the city of Berlin is also in contrast with its surroundings”. Unclear. If I understand, the city of Berlin lies within the borders of the inner polygon, and I can see that the predominant color is green, as in much of the region.
Citation: https://doi.org/10.5194/egusphere-2024-4031-RC2 - AC2: 'Reply on RC2', Márk Somogyvári, 27 Jun 2025
Status: closed
-
RC1: 'Comment on egusphere-2024-4031', Hesam Soltan Mohammadi, 10 Mar 2025
The manuscript by Somogyvári et al. presents an interesting approach to analyzing regional groundwater dynamics using dimensionality reduction on discrepancies between simple water balance models and observed groundwater levels. Applied to Berlin-Brandenburg, the method identifies spatial patterns in model misfits, potentially highlighting anthropogenic influences or unique hydrogeological conditions. By leveraging readily available climate and groundwater data, techniques like principal component analysis and multidimensional scaling offer insights into groundwater trends without requiring extensive prior knowledge. This data-driven approach could be valuable for assessing groundwater systems and their climate responses, helping to guide future studies.
Overall, the manuscript covers a compelling topic and presents a solid methodological framework that could appeal to a wide audience. It is well-structured and written, but I suggest the following improvements to enhance clarity and completeness:
- Please explicitly and precisely state the clear objectives of the study in one or two sentences in the Introduction.
- A clear map showing both Berlin and Brandenburg at the beginning would be helpful. In Figure 1 (and Figure 3), Berlin is labeled, but Brandenburg is not. Additionally, the white font color in Figure 1 makes it difficult to read the city name. It would also be good to ensure that figure labels are properly placed and annotations are aligned.
Minor formatting issues, such as in lines 188, 348, and 374, should also be addressed. - The study region is defined by administrative borders rather than hydrogeological ones. While practical for data availability, this could introduce biases. Could you discuss the implications of this choice?
- The descriptive discussion on aquifers AQ-1 to AQ-5 is useful, but a visual and precise representation on a map would improve clarity.
- Could you provide references for the hydrogeological details of AQ-1 mentioned in lines 150-152?
- The study focuses on 504 filtered wells, removing the outliers and anthropogenically influenced wells, but there is no sensitivity analysis of how this affects the results. Would fewer or more wells affect the regional patterns identified? Please also provide the initial number of wells before filtering.
- The unit for water outflow in Figure 4 is given in (m)—is this correct?
- The study assumes a constant flux approximation for the groundwater recovery process in Lausitz (1990–2022), despite its non-linear nature. Could you provide further justification or discuss whether a more detailed approach would improve accuracy?
- The manuscript refers to supplementary figure S1, but unfortunately I could not find it in the preprint.
- The manuscript states that a one-month time lag was chosen based on the “best overall fit” but does not define the criteria used for this selection. Could you please elaborate on “best overall fit”?
- Given the variability in aquifer types (confined/unconfined) and geological settings, a constant one-month lag may not be the best choice across the entire study area. Could you discuss this from a process-based perspective rather than relying solely on statistical analysis? Since AQ-1 is the primary focus, do you think this assumption holds for other aquifers as well, or just the best results are shown here?
- The study applies universal kriging with a spherical variogram model, but details on kriging parameters and model selection are missing. Since this interpolation is critical for the entire analysis, could you elaborate on how the Kriging method and best model were chosen and discuss any uncertainties related to this approach? Did you consider the directionality in variography?
- A simple linear model for the water balance has been used throughout the paper. Do you think such an oversimplification can capture the dynamic behavior of groundwater systems, especially in cases with complex hydrogeological processes, in urban areas with anthropogenic influences and in the era of climate change?
Citation: https://doi.org/10.5194/egusphere-2024-4031-RC1 - AC1: 'Reply on RC1', Márk Somogyvári, 27 Jun 2025
-
RC2: 'Comment on egusphere-2024-4031', Anonymous Referee #2, 17 May 2025
The manuscript of Somogyvári et al. presents a methodology for investigating groundwater dynamics at the regional level, applying dimensionality reduction on the differences between a set of groundwater level observations in the Berlin-Brandenburg region (Germany) and the respective levels simulated by a simple water balance model. Though using a subsurface water model based on the physically-based concept of mass balance, the approach is classified as data-driven since it needs several observations, which makes the method hardly applicable in data-scarce areas.
The methodology presented is sound overall, and the results are clear enough. They highlight that linear models can be applied to describe the response of the groundwater dynamics to the driving climate input for most of the analysed region. At the same time, the areas needing further studies are detected, even though their dynamics are not disentangled.
Overall, the manuscript represents an interesting methodological contribution to regional-scale groundwater analysis in regions with good data availability. It can be published after the following minor comments are addressed.
L54: “Assessing the groundwater behavior at a regional/mesoscale level is still relatively understudied”. Even though the authors already added a reference, this statement could be more strongly justified.
Figure 1: I believe that this figure needs to be completely revisited. In the text, especially in the Results and Discussion section, several toponyms are used, which make it very difficult for the reader to follow the thread of reasoning. All these toponyms must be shown clearly in a much more prominent and clearer Figure 1a. In addition, a third panel should be added (let’s say, figure 1c), with appropriate geolithological information related to the hydrogeology of the region analysed.
Section 2.1: It is explained that five separate aquifers at different depths can be identified in the analysed region. The study addresses only the top unconfined one. It is unclear if/how the methodology adopted could also be used for the lower aquifers, presumably less affected by climate drivers. Please explain and discuss.
Section 3.1 (and, consequently, Section 4): It is unclear how much the kriging algorithm (i.e., universal kriging with spherical semivariogram) influences the results achieved, especially for the zones with fewer observations. Please explain and discuss.
Section 4.1: Looking at Figure S1, one can quickly agree that a one-month time lag is the most suitable in general (even though I would have used a discrete color bar rather than continuous since only integer values are in the map). However, some zones diverge from this general rule. It is unclear if the method could consider different time lags for different zones in the same study area.
Fig. 5: The maps are difficult to interpret, and the colors are quite confusing, I think, particularly for color-blind people.
L417: “…the city of Berlin is also in contrast with its surroundings”. Unclear. If I understand, the city of Berlin lies within the borders of the inner polygon, and I can see that the predominant color is green, as in much of the region.
Citation: https://doi.org/10.5194/egusphere-2024-4031-RC2 - AC2: 'Reply on RC2', Márk Somogyvári, 27 Jun 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 263 | 137 | 19 | 419 | 39 | 16 | 36 |
- HTML: 263
- PDF: 137
- XML: 19
- Total: 419
- Supplement: 39
- BibTeX: 16
- EndNote: 36
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript by Somogyvári et al. presents an interesting approach to analyzing regional groundwater dynamics using dimensionality reduction on discrepancies between simple water balance models and observed groundwater levels. Applied to Berlin-Brandenburg, the method identifies spatial patterns in model misfits, potentially highlighting anthropogenic influences or unique hydrogeological conditions. By leveraging readily available climate and groundwater data, techniques like principal component analysis and multidimensional scaling offer insights into groundwater trends without requiring extensive prior knowledge. This data-driven approach could be valuable for assessing groundwater systems and their climate responses, helping to guide future studies.
Overall, the manuscript covers a compelling topic and presents a solid methodological framework that could appeal to a wide audience. It is well-structured and written, but I suggest the following improvements to enhance clarity and completeness:
Minor formatting issues, such as in lines 188, 348, and 374, should also be addressed.