the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Feb 2026
Technical note: Quantifying System Limits: Flood Oriented Operational Stress Testing under Intensified Precipitation Regimes in Central Europe
Abstract. The rise of short-duration extreme rainfall events across Central Europe has exposed critical weaknesses in traditional flood risk management methods that depend on fixed return periods and design thresholds. In response, this paper introduces FLOOD-ST: a scenario-based operational stress testing framework designed to evaluate the functional limits and vulnerabilities of water infrastructure systems under realistic, high-impact flood scenarios.
FLOOD-ST combines synthetic rainfall archetypes, varying antecedent soil moisture conditions and infrastructure-specific scenarios to systematically evaluate system performance. By integrating elements of functional stress testing, the framework diagnoses weak points and illuminates cascading failure pathways across interconnected assets. Unlike traditional hazard mapping, FLOOD-ST does not aim to replace existing tools but complements them by offering a diagnostic view of system behaviour during extreme conditions. The framework is platform-independent, modular, and scalable, making it suitable for both urban and rural catchments, especially those characterised by high system complexity, short warning times, or critical interdependencies.
This paper describes the conceptual foundation, simulation workflow, and real-world applications of FLOOD-ST, demonstrating its potential to support adaptive planning, emergency preparedness, and climate-resilient water management.
- Preprint
(1424 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-6378', Anonymous Referee #1, 12 May 2026
-
AC1: 'Reply on RC1', Jens Reinert, 13 May 2026
Dear Anonymous Referee #1,
We sincerely thank you for the careful reading of our Technical Note and for the constructive and detailed comments. We are pleased that you recognize the relevance of the proposed FLOOD-ST framework and the value of moving beyond conventional flood hazard mapping towards a more diagnostic assessment of system behavior, weak points, and cascading effects.
Your comments are very helpful for sharpening the scope, structure, and practical interpretation of the manuscript. In particular, we appreciate the suggestion to clarify the main contribution more explicitly, to distinguish more clearly between the state of research and our own methodological contribution, and to better define the practical limits and implementation levels of FLOOD-ST.
Below, we provide a response to the main points raised and indicate how we plan to revise the manuscript accordingly.
General comment 1: Restructure Introduction
We thank the referee for this important comment and agree that the Introduction would benefit from a clearer and more linear structure. In the current version, the background, problem formulation, methodological motivation, and contribution of the manuscript are partly interwoven. We agree that this makes the central argument less direct than it should be.
In the revised manuscript, we will reorganize the Introduction by first presenting the broader background and motivation, namely the limitations of conventional flood hazard mapping under non-stationary and infrastructure-sensitive flood conditions. We will then define the specific methodological gap addressed by the Technical Note: the need for a structured diagnostic approach that evaluates system behavior, weak points, operational limits, and cascading effects under variable stress conditions. Finally, we will present FLOOD-ST as our proposed contribution to this problem.
We will therefore revise the order of reasoning in the Introduction, especially around the passages currently located in lines 45–48 and 61–65. The revised Introduction will more clearly separate background, problem statement, state of research, and the own contribution of the manuscript.
General comment 2: Main contribution and key message
We fully agree with the referee that the central contribution should not be understood primarily as the promotion of FLOOD-ST as a label, but as the systematic extension of flood hazard and risk-oriented modelling towards scenario-based diagnostics of system behavior.
In the revised manuscript, we will therefore emphasize more clearly that the main contribution lies in the structured consideration of multiple stressors and system states that are often insufficiently represented in conventional hazard mapping. These include, among others, different rainfall archetypes, antecedent soil moisture conditions, infrastructure performance and failure, clogging, operational limits, and cascading effects.
We will revise the manuscript accordingly so that this key message becomes more visible throughout the paper. FLOOD-ST will be presented less as a stand-alone concept to be promoted and more as a framework that operationalizes this broader methodological shift from static hazard representation towards diagnostic stress testing of flood-prone systems.
General comment 3: Scope and practical limitation of FLOOD-ST
We thank the referee for this valuable suggestion. We agree that the manuscript currently presents FLOOD-ST in a broad conceptual way, while the examples and the most strongly developed methodological elements are primarily related to flash-flood-oriented and hydrodynamically explicit applications.
In the revised manuscript, we will clarify that FLOOD-ST is conceptually modular and can in principle support different flood-generation mechanisms. However, we will also state more explicitly that the strongest practical basis of the present Technical Note lies in applications where rapidly changing hydrological conditions, infrastructure performance, hydraulic routing, operational controls, and cascading effects are of central importance. This particularly includes flash-flood-oriented applications in small to medium-sized catchments and urban or peri-urban environments with relevant hydraulic infrastructure.
We will add a more explicit practical limitation to avoid overstating the current level of demonstrated applicability. In particular, we will clarify that applications to fully urban pluvial systems require an adequate representation of urban drainage systems, infiltration processes, surface flow paths, and, where relevant, surface–subsurface interactions. Where such components are not represented, FLOOD-ST should be applied only in a more limited diagnostic sense.
General comment 4: Practical applicability and smaller project scales
We agree that this is an important point. The presented examples are indeed embedded in larger research and project contexts, and a full implementation of FLOOD-ST may require substantial data, modelling expertise, and computational resources. We will therefore expand the discussion of practical applicability and scalability.
In the revised manuscript, we plan to introduce different levels of FLOOD-ST implementation. A minimum implementation could consist of extending an existing flood hazard model by one or more explicitly defined stressors, such as antecedent soil moisture or altered rainfall duration, provided that the results are interpreted diagnostically in terms of system behavior and weak points. A more advanced implementation would systematically combine multiple stressors, such as rainfall archetypes, soil moisture states, infrastructure performance, and operational conditions. A full system-oriented FLOOD-ST application would additionally include coupled system components, operational rules, cascading effects, and object-based impact indicators.
This clarification will also address the referee’s question of when a user is applying FLOOD-ST and when they are simply adding one feature to a conventional hazard mapping workflow. We will define FLOOD-ST not by the number of stressors alone, but by the diagnostic logic: the explicit use of scenario variations to identify system behavior, weak points, thresholds, operational limits, and impact pathways.
Specific and technical comments
We thank the referee for the detailed specific and technical comments. These suggestions are very helpful and will be addressed carefully during the revision of the manuscript.
In particular, we will revise the potentially misleading use of the term “drainage”, add supporting references where requested, reconsider the placement of the tipping point definition, and clarify the role of calibration within FLOOD-ST. We will also harmonize the terminology between Table 1 and Figure 1, reduce unnecessary repetition, and provide clearer guidance on the selection and definition of rainfall archetypes.
Furthermore, we will revise the discussion of the examples to make clear that they illustrate selected diagnostic pathways rather than a complete application of all FLOOD-ST dimensions. We will also clarify the treatment of pluvial impacts and specify that such applications require an adequate representation of relevant urban drainage and infiltration processes. The limitation section will be revised to better distinguish general limitations of flood modelling from FLOOD-ST-specific issues, particularly regarding expertise, data availability, model complexity, and scalability.
The technical corrections regarding wording, terminology, capitalization, punctuation, and the explanation of the “2021-like” scenario will also be implemented in the revised manuscript.
Once again, we thank Anonymous Referee #1 for the constructive and detailed assessment. The comments will help us to make the manuscript more focused, more consistent, and more useful for both scientific and applied flood risk communities. We believe that the proposed revisions will substantially improve the clarity of the Technical Note, particularly with regard to the main contribution, practical scope, scalability, and limitations of FLOOD-ST.
Citation: https://doi.org/10.5194/egusphere-2025-6378-AC1
-
AC1: 'Reply on RC1', Jens Reinert, 13 May 2026
-
RC2: 'Comment on egusphere-2025-6378', Anonymous Referee #2, 27 May 2026
This Technical Note presents FLOOD-ST, a scenario-based operational stress testing framework for diagnosing functional limits, weak points, and cascading failure pathways. The topic is relevant, particularly given the increasing concern over short-duration extreme rainfall, compound flood drivers, and the limitations of return-period-based design approaches. The conceptual framing is promising, and the two application contexts suggest that the framework is grounded in ongoing practice. However, several methodological components require clarification. The manuscript would be strengthened by clarifying the scenario-generation logic, providing a more explicit interface with climate change.
Specific comments
- A central component of FLOOD-ST is the definition of rainfall archetypes such as stationary cloudbursts, translating convective storms, and multi-cell compound events. This is conceptually useful, but the manuscript should explain how these archetypes are selected, parameterized, and justified for a given catchment. For example, how are rainfall depth, duration and spatial pattern determined? Are they based on observed events, expert judgment, or rainfall generators? This issue is important because the current formulation could be perceived as subjective. The authors note that future work may explore machine learning, and stochastic exploration, but the current framework would benefit from a clearer baseline method for selecting plausible but severe scenarios.
- The manuscript motivates FLOOD-ST using the intensification of short-duration extreme rainfall. However, the connection between this motivation and the actual design of rainfall stressors remains underdeveloped. Recent work on nonstationary IDF curves, and sub-daily precipitation changes could provide a stronger quantitative basis for defining rainfall triggers. For example, nonstationary IDF information could be used to specify time-varying or climate-adjusted rainfall depths, or stress increments relative to current design standards. This would help connect FLOOD-ST to existing engineering practice while still allowing the framework to move beyond fixed return periods. The authors should clarify how nonstationary rainfall information can inform the selection of “plausible worst-case” or “future-relevant” scenarios. See the following paper for reference. https://doi.org/10.1002/wat2.1519
- A following comment: the manuscript states that FLOOD-ST is climate-adaptive and can use future analogues, ensemble forecasts, or climate-adjusted storm archetypes. However, it is not clear how future climate change is actually incorporated into the framework.
- The current framework relies on representative rainfall archetypes and synthetic rainfall generators. A potentially valuable addition would be stochastic storm transposition, especially for generating physically plausible high-impact rainfall scenarios while preserving observed storm space–time structure. This could address one of the manuscript’s key limitations: the potential subjectivity in defining representative storm archetypes. Stochastic storm transposition could generate a large library of transposed historical storms. This connection would not necessarily turn FLOOD-ST into a conventional probabilistic risk framework. Rather, stochastic storm transposition could serve as an objective scenario-generation layer for stress testing. The authors may consider discussing this option, where they discuss scenario discovery and stochastic exploration. See 1016/j.jhydrol.2026.135617 and 10.1016/j.jhydrol.2020.124816
- The comparison in Table 1 is useful, but some contrasts appear overly simplified. For example, traditional flood modelling is described as “linear” and “deterministic,” and as neglecting flood peaks and timing. In practice, many hydrological and hydraulic models used in hazard mapping are nonlinear and explicitly simulate hydrographs, routing, and timing, even if their regulatory outputs are often summarized as return-period maps.
- The authors should ensure consistency in the spelling and capitalization of FLOOD-ST/Flood-ST.
Citation: https://doi.org/10.5194/egusphere-2025-6378-RC2 -
AC2: 'Reply on RC2', Jens Reinert, 27 May 2026
Dear Anonymous Referee #2,
We sincerely thank you for carefully reviewing our Technical Note and for your constructive and helpful comments. We are pleased that you recognise the relevance of FLOOD-ST as a scenario-based operational stress testing framework and acknowledge its potential for diagnosing functional limits, weak points, and cascading failure pathways under increasingly complex flood-generating conditions.
Your comments are particularly helpful in identifying where the methodological description of FLOOD-ST should be made more explicit. We agree that the manuscript would benefit from a clearer explanation of the scenario-generation logic, a stronger connection to nonstationary rainfall information and climate change, and a more precise discussion of how rainfall archetypes can be selected and justified in practice.
In the revised manuscript, we will therefore expand the description of rainfall archetype selection and parameterisation. We will clarify that archetypes should not be selected arbitrarily, but should be derived from a transparent combination of observed historical events, regional hydroclimatic characteristics, design rainfall statistics, expert knowledge, and, where available, rainfall generators or climate-adjusted precipitation information. We will also specify more clearly how rainfall depth, duration, temporal distribution, and spatial pattern can be defined for a given catchment and diagnostic question.
We agree that the link between the motivation of intensified short-duration rainfall and the actual design of rainfall stressors should be strengthened. In response, we will revise the manuscript to explain how nonstationary IDF curves, sub-daily precipitation scaling, and climate-adjusted design rainfall information can be used to define plausible worst-case or future-relevant rainfall stressors. This will help connect FLOOD-ST more explicitly to existing engineering practice, while preserving its diagnostic purpose beyond fixed return-period-based hazard mapping. We thank the referee for the suggested reference, which we will consider in this context.
We also agree that the incorporation of future climate change should be described more concretely. We will therefore clarify that climate change can enter FLOOD-ST at different levels, for example through adjusted rainfall depths or intensities, altered temporal rainfall patterns, modified antecedent moisture conditions, future analogue events, ensemble-based rainfall scenarios, or climate-adjusted storm archetypes. In addition, we will clarify that future-relevant stressors are not limited to the precipitation side. Changes in the catchment area itself, such as urban expansion, increasing surface sealing, changes in land use, modified drainage capacity, or shifts in the composition and vulnerability of the urban fabric, can also be incorporated as stressors or scenario dimensions. In this sense, FLOOD-ST can combine climate projections with catchment development scenarios to assess how future hydro-meteorological forcing and changing exposure or system conditions jointly affect flood system performance. We will make it clear that FLOOD-ST does not require a single fixed climate scenario but provides a framework for translating climate and catchment-change information into stressors relevant to system diagnostics.
The suggestion to discuss stochastic storm transposition is very valuable. We agree that this approach could provide an objective and physically plausible scenario-generation layer by preserving the space–time structure of observed storms while exploring their potential impact under different spatial placements. We will add this option to the discussion of scenario discovery and stochastic exploration and clarify that such approaches can complement FLOOD-ST without turning it into a conventional probabilistic risk framework. Instead, they can support the generation of plausible high-impact stress scenarios for diagnostic stress testing. We will consider the suggested references in this context.
We further thank the referee for pointing out that some contrasts in Table 1 may be overly simplified. We agree that traditional hydrological and hydraulic models are not inherently linear or purely deterministic, and that many models used in hazard mapping explicitly simulate hydrographs, routing, and timing. The intended contrast was not between model physics, but between the dominant use of such models in regulatory or planning contexts, where outputs are often condensed into return-period-based hazard maps. We will revise Table 1 accordingly to avoid overgeneralization and to distinguish more carefully between modelling capability and typical hazard-mapping practice.
Finally, we will ensure consistent spelling and capitalization of FLOOD-ST throughout the manuscript.
Once again, we thank Anonymous Referee #2 for the constructive assessment. We believe that the suggested revisions will substantially improve the methodological clarity of the Technical Note, especially with regard to scenario generation, climate-change integration, and the practical use of rainfall archetypes in FLOOD-ST.
Citation: https://doi.org/10.5194/egusphere-2025-6378-AC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 752 | 593 | 79 | 1,424 | 83 | 104 |
- HTML: 752
- PDF: 593
- XML: 79
- Total: 1,424
- BibTeX: 83
- EndNote: 104
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors introduce the FLOOD-ST framework that moves away from conventional flood hazard mapping and towards evaluating system behavior, identifying weak points and their cascading effects under variable scenarios and conditions. The paper is well written and drives on a comprehensive methodology section. It addresses important issues in flood modeling and management and provides valuable insight into the topic. However, some aspects should be improved before publication.
General comments:
Furthermore, line 61-64 presents a small part of the applied methodology, but line 65 then presents the main focus of the submission. After that, more background and recent developments and papers are presented again. I understand that it is often impossible to have the introduction in strictly linear order, but this submission would heavily benefit from disentangling the state of research and the own contribution more precisely. The raised issues in lines 45 and following and line 61 and following can be transferred to most parts of the introduction.
Specific Comments:
Line 42: The term drainage is misleading here. In an urban context (which FLOOD-ST also aims to cover), drainage does not increase flood peaks. I assume the authors want to mention the drainage through small ditches which are typical at agricultural lands, but I think the term in itself can be misleading here. I suggest the authors to either omit the term drainage here, or further explain which type of drainage they are pointing at.
Line 83-90: While I agree with the statements here, the authors should provide and cite sources to support the statements.
Line 94-97: The definition of tipping points could better be moved to the methodology section.
Line 168-172: This is an important new distinction in FLOOD-ST which might even be worth mentioning in the Motivation chapter. The authors should comment on whether model calibration might not actually support the framework. Even if calibration can only be fulfilled for a specific set of variables, which might differ at different scenarios (e.g. antecedent soil moisture), it can still be used to validate the model representation of other, more fixed variables, e.g. topography, roughness etc.
Table 1 and Figure 1: Figure 1 repeats a lot of information from Table 1 (Output, Outcome). Furthermore, Basis in Table 1 seems to be Input in Figure 1. The authors should try to limit repetition, but keep consistency in their wording when describing FLOOD-ST.
Line 198-207: If I understand correctly, the authors suggest to keep duration and rainfall amount constant within one archetype. However, the authors should also give some suggestions on how many different archetypes for one model domain are advised and provide further guidelines on how to define one archetype.
Line 280 and Line 281: FLOOD-ST seems to promote testing different stressors, and explicitly mentions that investigating average antecedent moisture conditions are also of high importance (line 235-238) and that this even presents “the full diagnostic capability” of FLOOD-ST. However, both given examples seem to mainly focus on the extreme spectrum, e.g. only investigating AMC III. The authors should therefore comment, why FLOOD-ST was not fully applied in the given examples and whether fully applying FLOOD-ST is even manageable. Furthermore, I don’t understand why these examples are given here, when there is the explicit section 2.4 “Current Application”.
Line 287: The way that FLOOD-ST is presented does not necessarily cover pluvial impacts. Pluvial flooding is mainly dependent on rainfall amount and the capacity of the drainage system combined with soil infiltration. Since the drainage system is only very implicitly covered by FLOOD-ST (maybe within infrastructure), I think the authors would do better to exclude pluvial flooding. Otherwise, the impact of underground drainage systems should be discussed more in detail within the framework.
Lines 332-338: This is mainly a repetition of information that has been mentioned multiple times before. The authors could skip this. In general, the benefits of FLOOD-ST have been covered before in the submission, so the authors might rephrase also the title of this overall section (possibly “Co-benefits”).
Lines 361-363: This mainly starts to touch on my concern in the general comments. I do not think data quality is a specific limitation of FLOOD-ST, but to flood modeling in general. The authors should more consider whether FLOOD-ST is appropriate for simplified methods on a coarse scale and should add more on the required expertise for applying FLOOD-ST and whether this might present another aspect of limitation.
Line 394 states that the stress tests are carried out on the full hydrologic system, which conflicts to different statements in the manuscript that seem to hint that different individual of FLOOD-ST can also be applied. Again, if it is always on the full hydrologic system, then this would put high limits on expertise and resources.
Technical Comments:
Line 52: Avoid the repetition of the word short here. E.g. you could change to “quick hydrological responses with often only short-term discharge records” or “short hydrological response times with often only recent discharge records”.
Line 62: I think the authors should replace banking sector with financial sector, as mentioned later.
The authors should recheck their spellings when it comes to capitalization, e.g., line 83 “(Flash) Floods”, line 123 “Scenarios”, line 269 “Indicators”.
Line 280: While a “2021-like” scenario might be understood in Luxembourg, Belgium, the Netherlands and Germany, the authors should better explain the term and explicitly refer to the flood event from 2021.
Line 287: delete the comma between on and river, and add a comma before “or cascading”.