The paper provides a use case for web-based high-level hydraulic analysis tool, designed to conduct numerical simulation from the data provided by user himself in order to access the outcomes of severe flooding. The case study suggests such assessment with the dataset on severe flooding event happened in February 2016 in the Agura River in Portugal.

The paper is a significant leap in shifting towards a new paradigm in hydrology (see e.g. (Rigon et al., 2022)): from Models as an Application (MaaA), e.g. installed on one’s desktop, to Model as a Service (MaaS) – high-level environment designed for numerical experiments and results’ visualization, located on the remote server and using graphical or API interface for user interaction (e.g. data handling and model setup).

The paper is well-organized and provides key features of the work motivation and the basic concept of the RiverCure portal, the data used for the numerical experiments and how the dataset is organized. Furthermore, apart from the input data the authors included the model output into the dataset as an instance of how the initial data could be utilized.

Having said that, I see several drawbacks in the dataset and the describing paper.

The paper structure could be improved. Section 2.2 Input data that describes only the spatial data is followed by Section 2.3 Output data, next Section 3 is called Data records and contains Section 3.1 Input data and Section 3.2 Output data again, which is very confusing. I suggest the authors combine the two sections to describe the data consistently – for input and output, spatial and temporal data separately.

The dataset handling should be improved. It took me a while to understand how I could download the data from the hydroshare.org website. The download process might be better documented for users not familiar with hydroshare.org or Bagit download tool. I suggest the authors prompt the download in Section 5 Data and code availability.

The dataset contents should also be improved. The listed spreadsheets Agueda_hydrometric_PonteRedonda.xlsx and Agueda_hydrometric_Ribeiro.xlsx contain only the streamflow discharge timeseries, contrary to what is stated on lines 159 – 167. The data spans for 16 days of hourly records at two gauges for streamflow discharge for the entire flooding event in February 2016

The river discharge and rain gauges locations could have been provided as a spatial coverage as well (e.g. geojson).

As of May 8^th, 2023, an attempt to load the layers to QGIS desktop 3.28.3 (Windows 10) via the provided links ended up as a failure (Web Map Service https://geoserver.hydroshare.org/geoserver/HS-937927473a3a4e66a07a2e2fdd9d581e/wms?request=GetCapabilities, Web Coverage Service https://geoserver.hydroshare.org/geoserver/HS-937927473a3a4e66a07a2e2fdd9d581e/wcs?request=GetCapabilities). Please check the data availability.

The http://rivercure.inesc-id.pt/ portal is a well-designed but not very useful tool unless you get an instant guest access. Several days after I requested the access to DemoOrganization it is still pending. Without the access visiting the portal narrows to browsing some satellite maps. However, the RiverCure portal functionality may not be the main aim of the dataset and paper.

I suggest the authors address the mentioned issues, revise the dataset and the manuscript accordingly, and both could be accepted for publication after consistently improved.

Line-wise and figure-wise technical comments:

L157:     special – spatial

Fig. 4:    Please round the values in the map legend

References:

Rigon, R., Formetta, G., Bancheri, M., Tubini, N., D’amato, C., David, O., and Massari, C.: HESS Opinions: Participatory Digital eARth Twin Hydrology systems (DARTHs) for everyone - a blueprint for hydrologists, Hydrol. Earth Syst. Sci., 26, 4773–4800, https://doi.org/10.5194/hess-26-4773-2022, 2022.

In their manuscript “Flood simulation with the RiverCure approach: The open dataset of the Águeda 2016 flood event”, Ricardo and colleagues present a web platform designed to integrate hydrodynamic and morphodynamic modelling tools and input data, together with an exemplary application in the context of a hydrodynamic simulation of the Águeda river flood in 2016. For this purpose, the authors provide both the input data (DEM, surface roughness, model domain) and the model output (water level, flow velocity, etc.).

The manuscript is fairly well written and organised. I appreciate the effort of the River Cure Portal (RCP) as a platform to allow for the web-based integration of data and models in the context of hydrodynamic simulations. The data is accessible via https://www.hydroshare.org which I was not aware of before, but which appears to adhere to FAIR principles. The dataset itself, though, is insufficiently documented in some parts, and the presentation of the data in the manuscript appears partly incorrect (see below under technical comments).

The main issue of this contribution is, however, that the presented dataset does not at all meet the review criterion of “significance”, neither with respect to the aspects of “uniqueness” nor “usefulness”. As for “uniqueness”, the model input data for the case of the Águeda river flood largely consists of data which is available to the public via other channels (the global ASTER-DEM, public land use and land cover data from which Manning-Strickler’s friction coefficient is derived, public rain gauge (or discharge?) records). The usage and pre-processing of these data to produce the case study input is undemanding. The output data, in turn, cannot be considered unique either, as it can be re-produced by any hydrodynamic model at moderate computational cost. In fact, the existence of RCP makes it even easier to produce the published model output. The criterion of “usefulness” is of course more subjective, but I do not see that the published dataset allows for any new approaches to address pressing research questions in the field. The authors themselves do not highlight any specific research questions or application cases for the data except that it “[..] can be used as a starting point to design other experiments and tools, and to explore the RiverCure Portal”.

I would like to emphasise, again, that I welcome the concept and implementation of the RCP (although I cannot say how it compares to alternatives frameworks). However, neither the design of the RCP nor the included HiSTAV model are within the scope of ESSD which is about the publication of unique and useful data. I therefore recommend rejecting the manuscript.

Apart from that, I have a few technical comments which are, however, not to be considered comprehensive:

• The data in the data repository is insufficiently documented. Not all files contain metadata, and there is no document that provides an overarching overview of the dataset as an entrypoint to users.
• Section 3.1 (3): this subset of data is desribed as precipitation data (the term “udometric” is highly uncommon), but the actual data files appear to contain discharge data - quite confusing. None of these files contain any metadata.
• Appendix: The RSL-based “Context”-definition of the data might be interesting within the RCP, but I think it is irrelevant in the context of the data presentation.
• l. 68: “included in the above input data list” - where is that?
• l. 71: “altimetry” - do you mean DEM?
• l. 136: “periodicity”? Do you mean “interval”?
• l. 208: the reference Ricardo et al. (2022) is missing in the reference list

Review report of the article "Flood simulation with the RiverCure approach: The open dataset of the Águeda 2016 flood event"

The manuscript describes an approach to combine data analysis and simulation as a service. To offer simulations as services is in principle a new way forward and may pave the way for many applications in natural systems analysis.

The manuscript shows a case study of a re-simulated flood event in 2016 in the Agueda river basin. In my opinion, the innovative aspect of the presented case study lays in the combination of data and simulation services, rather than in the dataset itself.

The dataset itself offers an opportunity for studying the flood impacts, e.g. by overlaying flood depths with the damaged buildings to elaborate vulnerability functions or to study the disruption of human mobility during the flood event. Thus, the dataset is worth to describe and beeing published.

However, the manuscript and also the data need to be improved before being useable by other scientists or practicioners. In the following, I am listing the main critical points that should be considered on the revision of the manuscript:

-A more stringent division/distinction between data and modeling is needed. At the moment, the method (the model) is in the foreground. A restructuring of the manuscript in that the data itself becomes a more prominent position might be worth to consider.

-The validation of the output data must be much more imporved. At least, the simulated flow depths must be compared with observed ones at sample sites. If there is a dataset of affected buildings or blocked roads available, the output dataset can be compared with these proxy data that can be used for validation.

- It would be of interest if the data portal can also store sensor data that document the flood event (eg., event documentation data, drone images, flow depth measurements, sediment transport and deposition, changes in cros-section profiles, etc.). This would make the dataset more interesting and valuable for developing and evaluating (benchmarking) newly developed simulation models.

- The main criticism of the simulation model is the poor representation of the river channel geometry. It is not clear which cross-section data was used. The simplification of the overall assumed channel depth of 5 m is not accurate. The presentation of local case study data is valuable only if it is really accurate and at a high spatial resolution. This approach of river geometry simplification is used for global models but not for local-scale models - or for river reaches that are dominated by downstream sea or lake levels. River conveyance capacity is one of the most sensitive factors in flood risk analysis. Moreover, please add more information on the mesh size. This all would make the data more comprehensible and more useful for others.

Minor points:

-The abbreviation HiSTAV must be explained before beeing used the first time and the model must be cited.

-Figure 1: Add a few explanations in the figure captions, explain abbreviations.

-line 86: what is an "in-house" model

-line 138: Please explain why the correction factor +0.5 has been chosen

-line 179: "bam break"?

-line 182: Who used the data from the media to map the limits of the flooded areas? Has this been done for validation?

-figure 3: add scale bar and a better legend. Please add also the site of the gauging station.

-figure 4: add scale bar and a proper legend

-How did you asess the role of insflow 1 and 2 from the measured flow in the gauging station?

-Figure 5: Please add the toponyms "Ponte Redonda" and "Ribeiro" in figure 3

- Please describe the REST API in more detail. This would be an innovative feature for data sharing and especially for interoperability.

Dear Editor,

Flood simulation with the RiverCure approach: The open dataset of the Águeda 2016 flood event

By

Ana M. Ricardo, Rui M.L. Ferreira, Alberto Rodrigues da Silva, Jacinto Estima, Jorge Marques,

Ivo Gamito, Alexandre Serra

The paper presents a WEB GIS platform “RiverCure”, that is designed to manage data and to simulate floods using an in house developed code HiSTAV. The data set that is proposed for the Agueda 2016 flood event, it includes all the input data and the output data.

General Comments

Unfortunately, this paper possesses an in review paper in preprint (https://www.preprints.org/manuscript/202305.1472/v1), which includes a lot of information that is missing in the present manuscript, but it is not cited at least in review. In addition, more than half of the figures are very similar.

Since the data are based on a poor quality DEM, I do not see the interest in sharing the data and the results, since a lot of issues must be linked to this modelling, some of the data is already in free access it seems. Furthermore, not enough information is provided to assess the quality of the data set and results especially when this argument is given: “The Águeda 2016 flood event was reported mainly on the Portuguese media allowing them to know reasonably well the limits of the flooded areas.” The Web-GIS platform seems a very nice contribution, but unfortunately the dataset is not well described and poorly convincing. In addition, since the platform seems efficient, new data based on high-resolution DEM and better described hydraulic parameters, would be of larger interest and with new fast computer it is a nonsense to provide data that results of computation, since models are evolving.

Specific comments

• Line 26: here systemic vulnerability will be a better word.
• Figure 2 does not include enough details since it is linked to the data.
• Line 92: please give the full definition name of CFL I guess: Courant–Friedrichs–Lewy
• Figure 4: improve please.
• Line 197: sampled hydrograph is confusing when used for modelling, what about ground truth?