the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Aug 2026
HP-CatCH: linking hydropower infrastructure to their hydrological catchments in Switzerland
Abstract. Detailed hydropower modeling in mountainous regions such as Switzerland is essential to optimize resource planning and inform the energy transition. To foster this process, we introduce HP-CatCH, an open-access dataset that maps topological configurations of hydropower cascades in Switzerland and links hydropower infrastructure to the associated hydrological catchments. In total, the components and topology of 48 cascades are defined, including 106 reservoirs and covering 95 % and 93 % of Swiss hydropower storage and pumped-storage capacity, respectively. Additionally, 601 run-of-river hydropower plants are modeled. The potential of the dataset is demonstrated by reconstructing historical inflows into reservoirs and run-of-river hydropower plants using simulated runoff from the hydrological model PREVAH. The resulting inflows exhibit large spatiotemporal variability, highlighting the relevance of such detailed and spatially explicit hydropower representation for energy system modeling and hydroclimatic impacts studies.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Earth System Science Data.
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
(11174 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (extended)
- RC1: 'Comment on essd-2026-448', Andrea Galletti, 22 Sep 2026 reply
Data sets
HP-CatCH: linking hydropower infrastructure to their hydrological catchments in Switzerland Y. Y. Haddad et al. https://doi.org/10.3929/ethz-c-000800479
Model code and software
HP-CatCH Y. Y. Haddad https://github.com/yasserhaddad/HP-CatCH
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 149 | 85 | 113 | 347 | 150 | 112 |
- HTML: 149
- PDF: 85
- XML: 113
- Total: 347
- BibTeX: 150
- EndNote: 112
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Haddad Y. Y. et al. present a dataset linking hydropower plants and reservoirs in Switzerland to their contributing hydrological catchments, HP-CatCH. This dataset facilitates the assessment of inflows to specific hydropower systems or cascades, as also demonstrated by the provided supporting analysis.
The manuscript addresses a gap left by previous similar contributions by providing full access to its information, as well as by unifying information related to different plant types and reservoirs, while at the same time avoiding redundancy with existing datasets by coherently referring to them within HP-CatCH.
By design, HP-CatCH is only aimed at constructing topologies useful to determine the natural plus diverted inflow into hydropower systems, while it does not provide technical specifications on the systems listed therein, such as pressure channel capacity or reservoir volume description.
As a topological dataset, I feel that the introduction could lead a bit more directly into its proposed research gap. Moreover, the methodological description of how the topology is derived currently does a poor job of explaining itself to the reader, limiting the transferability of the approach. Finally, while the authors claim that this approach is transferable (and I assume they refer to the topology determination, rather than to HP-CatCH itself, which is limited to the geographic scope of Switzerland), they should clarify the requirements for such transferability (what mainly comes to my mind is the existence of an underlying catchment discretization similar to EZGNR, with a similar hierarchical structure, and a way to correctly link it to at least intake points and reservoirs).
I think that this work is well suited for publication in ESSD, but major revisions are needed to ensure it taps its full potential. More specific comments follow.
Introduction
lines 15-23: a list of generic definitions is provided here. This neither describes the state of the art in Swiss hydropower datasets nor highlights a gap. I would remove it or trim it significantly. Definitions are, however, more suited to a dataset description than to an introduction, so in case the authors want to keep these, I'd consider moving them to the methods section, where deemed appropriate.
lines 46-51: these should be moved earlier, like after line 30. This is not a good closure before stating your research aim, as datasets are not central to any of the cited papers.
lines 42-45: I agree that the cited study provides a valuable blueprint. Therefore, this should be stated at the beginning of the "dataset state of the art", not at the end of it. So my suggested reordering is to move ll. 46-51 after line 30, and ll. 42-45 immediately afterwards.
line 31: "technical representation" does not really represent the scope and actual content of HP-CatCH. I suggest lowering this claim a bit, otherwise it risks mismatching the actual contribution of your dataset.
Data and Methods
line 73: the cascades where hydrological aggregation is performed should be labeled in the dataset, so that the corresponding contributing basins can be interpreted correctly.
lines 74-78: it is hard to understand the nature of this datum. Specifically, "each component is assigned to its corresponding hydrological sub-basin" is not clear: to me, this would mean the entire portion of area draining into it, but this is what is described as "all contributors" later. Therefore, I would assume that each component is assigned to its corresponding topographical unit (the 2-km2 unit where the reservoir or intake physically is). Next, it is not clear what these units are: they are described as sub-basins (as in, e.g., GIS-extracted basins with a threshold area of 2 km2), in which case it would be clearer to specify that they have an average area of 2 km2. This underlying datum is essential to the transferability of your approach, so describing its topological-topographic properties thoroughly is essential.
line 85: is the daily, 500 m runoff used by the model or rather produced by it? In case it is really used, the authors should clarify better what the output of the model is. Moreover, how is the aggregation to 2 km2 performed? On the same note, please clarify whether q_i in Eqs. at ll. 111–113 represents the incremental runoff generated within sub-basin i, obtained by spatially integrating PREVAH grid-cell runoff over that polygon, rather than routed discharge at the sub-basin outlet. This distinction is important, since summing outlet discharge from hierarchically nested sub-basins would itself lead to double counting.
line 95: correct reference to Fig. 1 ("Fig." missing).
lines 96-102: this explanation is not particularly clear: first, it is not clear why, in the case of reservoir hydropower plants (and pumped-storage, I assume? This should be mentioned explicitly), contributing areas are computed with reference to the reservoir (i.e., all intakes, all the way down to the reservoir), while for RoR plants the end point is the plant itself. If there is a reason for this inconsistency, it should be better clarified. Second, it seems that there are three topological types: reservoir-direct, reservoir-all, RoR-all: I think it could be beneficial to sketch them all in Fig. 1. Third, reservoir-direct accounts for sub-basins draining into the specified reservoir plus all water intakes connected to it: this includes water intakes located outside the hydrological catchment, but is not represented in Fig. 1 (another reason to improve it). Also, what if a water intake for R2 is geographically located in the area that hydrologically contributes to R1? All in all, while the "all contributors" definition is fairly intelligible (basically everything natural or anthropogenic that brings water is considered), the "direct contribution" deserves a more detailed description, and especially coverage of edge cases (transboundary intakes, pumping, cascading intakes, etc.). This would substantially improve the transferability of the approach to other areas where more complex topologies might exist. On a final note, the formal definitions provided are correct but, in my view, do not facilitate the understanding of differences between topologies and elements included/excluded therein.
line 105: use "on the other hand" or a synonym, instead of "moreover".
lines 111-113: as per my previous comment, again the "inflow into hydropower reservoir or hydropower plant" is mentioned here. There is a bit of inconsistency here: reservoirs also have a connected hydropower plant downstream, and I assume the inflow for the plant would be the same as that for the reservoir. However, this is not specified.
Dataset description
The section shows three tables, but not how they interact with each other. As far as I understand, the data contained in the table of which Table 1 is an excerpt constitute the main reference. Here, nodes are numbered incrementally and labelled by cascade, type, etc., while also being cross-referenced to other databases where applicable. Then, Tables 2 and 3 seem to represent the topologies of one reservoir-hydropower system (Table 2) and several RoR plants (Table 3). I think the topological scheme approach used in Table 2 is much easier to follow than that in Table 3, and the fact that these two are different is confusing (if I understood correctly, the EZGNR column is similar to a FlowsFrom, while WASTA is similar to FlowsInto). Would it be possible to align the RoR part to the topology adopted for reservoirs? Out of curiosity, why does Table 2 not include any of the hydropower plants presented in Fig. 2?
I suggest a restructuring of the section as follows: first, describe the two tables qualitatively (one stores all nodes, with IDs and cross-references (Table 1), while Table 2 stores topological details about the cascade). Then describe the attributes of Table 1 and Table 2 as is currently done.
Fig. 3 would benefit from an additional panel where a DEM and the main rivers are depicted, so that even someone not familiar with Swiss orography would be able to make sense of the spatial distribution of the characteristics depicted in the other 5 panels.
Results
line 161: "Panel" missing.
line 162: I would swap the order (larger areas tend to have larger inflows).
Panel (a) of Figs. 4 and 5 is not very informative as is. By using absolute values and summing everything together, larger systems dominate the plot. One way to convey that there is some annual variability could be to use the mean monthly standardized flow (i.e., for each reservoir take the monthly average inflow and divide it by the annual average): this would preserve the patterns and give the same weight to large and small systems. Indeed, perhaps large and small systems would exhibit different annual cycles, but this can't be seen when analyzing the sum of absolute values. The same goes for Fig. 5. More importantly, the depicted numbers are very likely duplicating the same flows over and over, if inflows to multiple systems located on the same river are accounted for (actually, this duplication is the very definition of the difference between direct contributors and all contributors). Again, using standardized flows would improve this. "The magnitude of mean monthly inflow into RoR hydropower significantly exceeds that of reservoirs" cannot really be supported by this analysis, as neither category's numbers represent independent totals.
Panel B of both figures would be better as a scatterplot (mean inflow vs volume and power, respectively), as no added value comes from the spatial visualization in this case.
Discussion and conclusion
lines 186-190: this is presented as a limitation. To me, avoiding redundancy is an extra effort worth acknowledging. If the related data are correctly linked to the relevant datasets, this is an added value and should be framed consistently. At most, the main limitation could be that the dataset currently relies on a static 2023 snapshot of WASTA, if I understood correctly.
lines 181 onwards: topological simplifications should be addressed more systematically here. As also highlighted in my previous comments, several steps in the construction of HP-CatCH involve simplifications or interpretation of the available information, including topological aggregation, the treatment of pumped-storage connections, and the assignment of intakes bypassing reservoirs. In this respect, what does "closest upstream reservoir" mean? Is proximity defined purely in terms of distance, or does it account for cascade membership and internal hierarchy? More generally, since the topology itself is the main novel contribution of HP-CatCH, I think its quality and completeness should be assessed more explicitly. In particular, it would be useful to report, where possible:
- how many nodes/connections/intakes could be directly identified from the underlying datasets;
- how many required manual correction, topological aggregation, or other assumptions;
- how many systems were cross-checked against technical documentation from hydropower operators or other independent sources;
- whether any nodes or connections remain uncertain, and how these cases are identified in the dataset;
- some overall indication of the expected completeness and reliability of the reconstructed topologies.
While this does not require a quantitative validation metric, even just a quality-control summary would substantially improve the transparency of the dataset, allow users to distinguish directly supported information from reconstructed or simplified connections, and strengthen the claimed transferability of the approach.
lines 191-195: the resolution of the underlying stream network clearly affects the inflow to smaller intakes. However, I would expect this effect to substantially decrease when larger catchments are considered.
Code and data: it seems to me that the link provided in lines 208-209 does not work (the one in the references does).