Weak Points in the Flood Risk Modelling Chain - an initiative by the IAHR Committee on Flood Risk Management

Introduction      Your contribution      Collected weak points

Table of contents

How to cite

Surname, N. (202X). Title of the contribution. In: Molinari, D. and Haun, S. (eds.): Weak Points in the Flood Risk Modelling Chain - an initiative by the IAHR Committee on Flood Risk Management.

Indirect and intangible impacts in flood risk assessment of complex systems     Back to Top

DOI: 10.3850/IAHR02262022-03380000

Author: Chiara Arrighi (University of Florence), Marcello Arosio (IUSS Pavia)

Description of the weak point:

Direct damages are associated with the physical impacts and generally estimated by stage-damage functions. Vulnerability and damage functions have been developed for many critical assets e.g., residential buildings, allowing for the improvement of risk maps and cost-benefit analysis. All the consequences beyond the direct damage are considered indirect, such as the loss of connectivity within the broader infrastructure or economic network. The understanding of indirect and intangible damages, i.e., those hardly monetizable, is less advanced.

How do we account for intangible damages, such as the loss of spiritual, aesthetic or social values? Do we fully understand how indirect impacts propagate in time and space and how flood characteristics, e.g., water depth, velocity, type of flood (pluvial vs river), lag time, hydrograph dynamics) are affecting them? And how do we merge direct and indirect impact metrics to inform decision makers? These are questions researchers are addressing.

Description of contribution:

We illustrate two examples of research works which addressed some gaps in indirect impacts modelling. One focuses on indirect impact in time and the second on indirect impact in space. The first work highlights the strong dependence among flood severity, indirect losses, post-event recovery time, risk and resilience with a quantitative resilience model for cultural heritage (CH). It provides an application to the art city of Florence where social appreciation of CH has been assessed by means of crowd-sourced data, to also obtain a proxy for intangible exposure values. A stage-idleness vulnerability function is developed to simulate indirect impacts to each cultural building and feeds the resilience model which brings the single indirect loss into a systemic city-perspective. The model yields the recovery time by using visitors to CH as a state variable, which extends indirect flood losses up to one year approximately for severe scenarios.

The second work proposes an analysis of risk of complex urban systems exposed to both pluvial and fluvial floods based on the construction of a graph and the study of its properties. The graph allows for propagating the impacts throughout an affected system. Indeed, the use of a graph allows estimating, besides direct losses (of elements within a flooded area), also indirect losses to elements outside the affected area that rely on services provided by directly hit elements, which may have lost some capacity to provide those services. The propagation and quantification of impacts through a graph allows understanding the system risk mechanisms and weaknesses that can translate into larger indirect consequences. The direct and indirect impacts can then be computed based on three levels of vulnerability: (i) the physical vulnerability of a directly affected element; (ii) the vulnerability associated to the link between an affected element and its receivers; and (iii) the vulnerability of the service-receiving element.


Paper 1: Resilience of art cities to flood risk: A quantitative model based on depth-idleness correlation

Paper 2: Assessment of the Disaster Resilience of Complex Systems: The Case of the Flood Resilience of a Densely Populated City

Challenges at the interface between hydraulic and damage models      Back to Top

DOI: 10.3850/IAHR02262022-03380001

Author: Daniela Molinari and Francesco Ballio (Department of Civil and Environmental Engineering – Politecnico di Milano)

Description of the weak point:

Hydraulic models usually supply results that are only partly usable by damage models. One aspect is related to the types of hazard variables supplied (example: damage models require information on persistence of water and water quality, i.e. presence of sediments or contaminants, that is usually not supplied by hydraulic models) and their scale (damage models work at different scale of aggregation, from the scale of the single exposed elements to the municipality scale, while hydraulic models generally work at the meso scale). Another aspect is related to the need of having seasonal hazard scenarios, as damage strongly depends on the period of the year when the flood occurs, while hydraulic models work with the concept of the yearly return period.

Description of contribution:

Two works are reported, emphasizing the problem(s) at stake. In the first one, a damage model is developed for the estimation of flood damage to crops. The paper highlights (i) how damage strongly depends on hydraulic variables that are usually not supplied by hydraulic models (the duration of the flooding event, at first) and (ii) the variability of damage with the vegetative stage of the crops at the time of the event, that is with the period of the year. In the second work, with the objective of performing the CBA of some implemented structural measures, the effort has been made to adequate the scale of the hydraulic analysis to the requirements of damage modelling (i.e., both models work at the micro scale), highlighting strengths and limitations. Nonetheless, the CBA exercise further stresses the importance of having seasonal hazard scenarios and related probability of occurrence.


Paper 1: AGRIDE-c, a conceptual model for the estimation of flood damage to crops: development and implementation

Paper 2: Cost–benefit analysis of flood mitigation measures: a case study employing high‑performance hydraulic and damage modelling 

Prediction of flow variables truely representative for damage estimation, including inside the buildings Back to Top

DOI: 10.3850/IAHR02262022-03380002

Arthor: Benjamin J. Dewals (University of Liege, Belgium)

Description of the weak point:

A gap in flood modelling is the characterization of danger inside the buildings. Buildings are usually schematized as impermeable obstacles to the flow. In reality, the flow inside the buildings may be quite different from the flow around the building; and this affects the damage mechanisms and the likelihood of fatality. Initiatives exist on this particular aspect of “flood modelling”; but it remains far from common practice.

Description of your contribution:

This point relates to various on-going research, such as within the French ANR project DEUFI, the coordinator and partners of which are acknowledged for fruitful discussions.

Fluvial processes as inventible parts in assessing riverine floods      Back to Top

DOI: 10.3850/IAHR02262022-03380003

Author: Stefan Haun (University of Stuttgart)

Description of the weak point:

Flood risk assessment and flood hazard mapping are inventible steps for obtaining a sustainable integrated flood management concept. However, so far the occurring sediment transport in rivers and associated hydromorphological bed changes (fluvial processes) during floods are not taken into account in a proper way when obtaining flood inundation maps. The reason is twofold: First, sediment transport and hydromorphology are not taken into account in available regulations and directives (e.g., within the Flood Directive 2007/60/EC, launched by the European Union), and second, it is not trivial to simulate hydromorphological changes during flood events properly.

Description of contribution:

Sediment transport in rivers is strongly depending on the occurring flow conditions and the availability of sediments. Usually, during high flow periods the highest sediment transport rates occur in natural river courses. In addition, high discharges (with a return period between 2 and 5 years) account for the bed formation of the fluvial system. These processes may effect floods and the associated flood risk by:
(i) A reduction of the capacity of the river, as a result of sediment depositions and a possible local damming in combination with associated backwater effects.
(ii) Occurring fluvial processes (either in the form of bed forms or an entirely deformation of the river channel), as a result of the discharge and the sediment supply from upstream.
These alterations in the geometry of the river channel may happen during the flood event, but may also change the river bed in a sustainable manner, with consequences for mid or long term periods. As a direct result the flood frequency, but also the severity of floods may change.

Although meanwhile several researchers proved that it is necessary to take sediment transport processes and morphological bed changes into account for an accurate estimate of the impact of floods, the still open questions are:
• When and how do we need to consider sediment transport and morphological changes in simulating flood events?
• How to deal with mid or long term bed changes and when do we have to perform new simulations of flood inundation maps?
Especially with an outlook on more frequent and more extreme flood events in the future, as a result of global change, these questions need to be addressed.

Improving flood crest forecasting: a data-driven approach      Back to Top

DOI: 10.3850/IAHR02262022-03380004

Author: Marian Muste (IIHR-hydroscience & Engineering)

Description of the weak point:

The need for improved and timely forecasts of the flood hazard is critical for communities located in flood-prone areas at a time when flood hazards are increasingly widespread and more frequent worldwide. The new forecasting approach presented herein is rooted in the innovative interpretation of the flow physics during flood wave propagation. In particular, the approach takes advantage of the separation of directly measured index velocity and stage hydrographs occurring at a site prone to hysteresis (see Figure 1) (Muste & Kim, 2020). Using data mining and machine learning applied to existing hydrographs, relationships can be developed to determine the magnitude and timing of the flood crest using the index-velocity peak (that always precedes the stage peak) as a flagging point. Application of the algorithm to directly measured data at an existing index-velocity site enabled forecasting of flood crest characteristics in good agreement with actual measurements (Muste & Kim, 2020).

Description of the contribution:

We illustrate the proof-of-concept of the new data-driven forecasting approach with data acquired at the USGS station #05558300 on the Illinois River. We selected this site because the stream at this location is prone to hysteresis and because the data acquisition and processing is made with rigorous and uniform protocols over extended periods of time, therefore providing a reliable benchmark dataset.

In summary, the construction of the forecasting protocol entails the tracing, recording, and determining the following parameters for each individual pulse occurred during the storm: a) magnitude of the stage and index-velocity values at the onset of the index-velocity pulse; b) rates of change for the index velocity and stage during the pulse; c) duration of the rising of the pulse to its peak; d) unsteadiness coefficient; and e) the time interval between the index-velocity peak and the associated stage peak. The construction and usage of the data-driven forecasting algorithm (i.e., does not involve any hydraulic modeling) is shown in Figure 2.

A surrogate validation of the developed protocol was made by comparing hindcasted values using the developed algorithms with the actual values of the storms’ pulses above the Action Stage recorded at the station in 2019. The implementation of the forecast is illustrated with red dotted straight lines in Figures 2b and 2c corresponding to a storm recorded at this station in 2019. The differences between the predicted and recorded values of the stage crest are less than 10%, with more than half of the predicted values within a ~ 5% range. Overall, the differences between the prediction and actual values for the arrival time of the crest are larger than those in predicting the flood crest magnitude. Notably, the lag between hydrograph peaks for the largest storm of 2019 are between 32 and 65 hours, equivalent to 1.3 to 2.7 days, time interval that enable managers and public to evacuate and make last-minute interventions.

Benchmark dataset link:



Figure 1. Hysteresis effects on flow variables:

a) stage vs. free-surface water slope (adapted from Muste et al. 2020); 

b) stage vs. index velocity (adapted from Muste & Kim, 2020); 

c) stage vs discharge (adapted from Muste & Kim, 2020); 

d) hydrograph phase sequencing (adapted from Graf & Qu, 2004). 

e), f), g) relationships for the variable pairs in Figures 1b, 1c, and 1d, respectively, as measured at an index-velocity station. Rising and falling specify stage variation phases (from steady flow to Hmax).

Fig 2.png

Figure 2. Construction and usage of the flood wave crest amplitude and arrival time forecasting protocol

Robustness and transferability of existing damage models      Back to Top

DOI: 10.3850/IAHR02262022-03380005

Author: Daniela Molinari, Francesco Ballio (Department of Civil and Environmental Engineering – Politecnico di Milano)

Description of the weak point:

Several tools are now available for flood damage estimation which are characterized by different levels of robustness and reliability and no model can be, at present, considered as a standard. The choice of the more suitable model(s) to be implemented can be challenging, above all for non-expert users, and may imply significant errors in damage estimates if done without a critical knowledge of models’ limits and usability. Unfortunately, information on how damage models have been derived, calibrated, validate, and how they must be applied is often lacking.

Description of contribution: 

An operative tool to collect and organize information about available flood damage models is presented: the Flood Damage Model Repository (FDM). The FDM is aimed at supporting flood damage modelers, especially non-expert ones, in the choice of the best available model(s) for a specific context under investigation and a specific problem at stake. For each model included in the repository, key information is supplied that, if ignored, can lead to improper use of models and then significant errors in flood damage assessment. Information includes: (i) a description of the context in which the model was derived (i.e. the country of development, the scale of the analysis, the flood type for which the model was developed, the model type and the exposed items considered), (ii) mathematical formulation of the model, (iii) the required hazard, exposure and vulnerability input parameters, (iv) the output of the model, (v) details on calibration and validation (i.e. the context in which the model was calibrated/validated, the dimension of the dataset used in the process and the quality of such data, the estimation error) and (vi) an evaluation of model transferability and applicability by non-expert users. FDM presently focuses only on direct damage.


Integrating human systems in the current flood modelling practices      Back to Top

DOI: 10.3850/IAHR02262022-03380006 

Author: Yared Abayneh Abebe

Description of the Weak Point: In flood risk management (FRM), the likelihoods of adopting and implementing measures that reduce flood hazard, vulnerability and exposure depend on changes in individual and institutional behaviour in response to the potential of flooding and the accompanying impact. Therefore, on the one hand, FRM is dependent on the rules, regulations, policies and implementations that aim to reduce flood risk. On the other hand, it relies on how individuals react towards those aspects and adapt their behaviour. Hence, comprehensive approaches developed to reduce flood impacts should include human adjustment to floods, and focus on human elements such as exposure, vulnerability, capacity and resilience, which are shaped by socio-economic factors. Developing a framework that integrates the human and flood subsystems, and supports modelling multiple stakeholders’ decision makings in FRM has been a major challenge.

Description of contribution: 

A modelling framework called Coupled fLood-Agent-Institution Modelling framework (CLAIM) and a methodology to build holistic human-flood interaction models that provide new insights into FRM policy analysis and decision-making are developed.
- The modelling framework helps to decompose the elements that make up human-flood systems. The framework defines the coupled system as a complex adaptive system and conceptualizes the drivers of flood hazard, vulnerability and exposure as factors that shape the complex interaction between and within the component subsystems.
- In the methodology that accompanies the framework, the human subsystem is modelled using the agent-based modelling approach (ABM). ABM incorporates heterogeneous actors and their actions and interactions with the environment and flooding. It also provides the possibility to analyse the underlying institutions that govern the actions and interactions in managing flood risk. The flood subsystem is modelled using a physically-based, numerical model. The ABM is dynamically coupled to the flood model to investigate and understand the interaction between the subsystems.

The coupled ABM-flood modelling approach contributes to FRM in multiple ways.
(i) it presents a simulation that shows how flood risk evolves over time in response to actors’ behavioural change, environmental change including urbanization and climate change, and measures implemented.
(ii) it provides a holistic view of flood risk in which one could study how the social, economic, governance and hydrological makeup of an area contributes to the risk. It could capture all aspects of flood risk – flood hazard and communities’ vulnerability and exposure.
(iii) it puts emphasis on social institutions and provides policy analysis for decision makers. Coupled ABM-flood model provides a platform to test existing and draft flood risk reduction policies. The new insights gained from simulation outputs could contribute to better FRM policy design.

Paper link:



Figure 1. The CLAIM framework showing interactions among humans (agents and institutions), their urban environment, the physical processes that generate flood and external factors.

Field observations of individual stability in inundated urban environment during a major flood     Back to Top

DOI: 10.3850/IAHR02262022-03380007


Hubert Chanson (The University of Queensland, Australia)  

Richard Brown (Queensland University of Technology, Australia)

Description of the weak point: 

The flooding of urbanised areas constitutes a major hazard to populations and infrastructure. Flood  flows during urban inundations have been studied only recently, and few studies considered the  impact of water flows on populations, including residents, emergency, and swift rescue personnel.  The risk to the populations is expected to increase with urban developments in flood‐prone areas,  while squatting areas in some countries is highly vulnerable to flooding. It has been argued that  current design guidelines are obsolete and inadequate, because of the absence of real‐world data  sets collected during major flood disasters. Thus, there is a need for more data on human instability  in floods during real events, to update present guidelines and make them safer. 

Description of contribution: 

In January 2011, the city of Brisbane (Australia) experienced a major flood. The event affected many  people (12,000 homes) causing in excess of $500 million of flood damage. A series of detailed  velocity measurements were conducted in an inundated section of the city about the peak of the  flood (Paper 1). As part of the fieldwork, the authors experienced the impact of the flood flow during  the preparation and installation of the instrumentation. For each instance, the individuals used  secured safety ropes and safety handrails to work safely in the inundated urban environment. The  data analyses (Papers 2 and 3) showed that the hydrodynamic conditions were unsafe for  individuals, and that most current guidelines were over‐optimistic and unsafe in real floodwaters  and natural disasters like in January 2011 in Brisbane. Simply, the experience gained by the authors  and the re‐analysis of real flood disaster data shows that, during a natural disaster, the flood flow  conditions can be treacherous and dangerous. A series of more conservative and safer guidelines is  proposed, particularity relevant to flood events in inundated urban environments (Paper 3).  


[1] BROWN, R., and CHANSON, H. (2012). "Suspended Sediment Properties and Suspended Sediment Flux Estimates in an Urban Environment during a Major Flood Event." Water Resources Research,  AGU, Vol. 48, Paper W11523, 15 pages (DOI: 10.1029/2012WR012381)  

[2] CHANSON, H., BROWN, R., and McINTOSH, D. (2014). "Human body stability in floodwaters: the  2011 flood in Brisbane CBD." in "Hydraulic Structures and Society – Engineering Challenges and  Extremes", The University of Queensland, Brisbane, Australia, Proceedings of the 5th IAHR International Symposium on Hydraulic Structures (ISHS2014), 25‐27 June 2014, Brisbane, Australia,  H. CHANSON and L. TOOMBES Editors, 9 pages (DOI: 10.14264/uql.2014.48).

[3] CHANSON, H., and BROWN, R. (2018). "Stability of Individuals during Urban Inundations: What  Should We Learn from Field Observations?" Geosciences, Vol. 8, No. 9, Paper 341, 9 pages (DOI:  10.3390/geosciences8090341) 

Field observations of hydrodynamics in inundated urban environments     Back to Top

DOI: 10.3850/IAHR02262022-03380008


Hubert Chanson (The University of Queensland, Australia)  

Richard Brown (Queensland University of Technology, Australia)  

Description of the weak point: 

The vulnerability of urban environments with respect to flooding has been a long‐standing concern of society. The flooding of urbanised areas constitutes a major hazard to populations and  infrastructure. Flood flows during urban inundations have been studied only recently. Few studies considered the interactions between water flows, sediments, buildings and topographic features, and the associated instabilities. In this context, there is some paucity of field data to understand large flood phenomena, including mega‐floods, in particular in terms of velocity and sediment data  in an inundated urban environment.  

Description of contribution: 

During the January 2011 flood of the Brisbane River in eastern Australia (Paper 1), detailed field measurements were performed in the inundated flood plain of the Brisbane River in the Central  Business District (CBD) of the city of Brisbane. Instantaneous velocity and suspended sediment  concentration (SSC) estimate were collected at high frequency for several hours. The results (Papers  2 and 3) showed some unusual long‐period oscillations linked to hydrodynamic resonance induced  by the flooded urban settings, creating very challenging conditions highlighting a number of practical  issues. The data further highlighted the large suspended sediment loads and fluctuations in the  inundated urban setting associated possibly with a non‐Newtonian behaviour.  


[1] CHANSON, H. (2011). "The 2010‐2011 Floods in Queensland (Australia): Observations, First  Comments and Personal Experience." Journal La Houille Blanche, No. 1, pp. 5‐11 (ISSN 0018‐6368).  

[2] BROWN, R., and CHANSON, H. (2012). "Suspended Sediment Properties and Suspended Sediment  Flux Estimates in an Urban Environment during a Major Flood Event." Water Resources Research,  AGU, Vol. 48, Paper W11523, 15 pages (DOI: 10.1029/2012WR012381).

[3] BROWN, R, and CHANSON, H. (2013). "Turbulence and Suspended Sediment Measurements in an  Urban Environment during the Brisbane River Flood of January 2011." Journal of Hydraulic  Engineering, ASCE, Vol. 139, No. 2, pp. 244‐252 (DOI: 10.1061/(ASCE)HY.1943‐7900.0000666). 


Suspended sediment properties and suspended sediment flux estimates in an inundated urban environment during a major flood event

Technical report:

Stability of Individuals during Urban Inundations: What Should We Learn from Field Observations? 




Detailed field measurements of spillway operation during major floods     Back to Top

DOI: 10.3850/IAHR02262022-03380009


Hubert Chanson (The University of Queensland, Australia) 

Description of the weak point: 

During major flood events, the floodwaters must be passed safely above, beneath or beside a dam: this is achieved with a spillway system, designed to discharge safely the extreme reservoir

outflows. Two main functions of a spillway are the safe conveyance of floodwaters and safe dissipation of the kinetic energy of the flow before rejoining the natural river system. Currently, a major knowledge gap is the lack of field measurements in dam spillway in operation during large floods.

Description of contribution: 

Since conveyance and energy dissipation are closely interrelated, detailed measurements must be undertaken in both the steep chute and in the stilling structure. The first part of the present contribution describes a new approach combining laboratory experiments and field observations, called hybrid modelling (Mark II). Optical techniques may be applied to gain some quantitative information on the hydrodynamic properties, although the validation is not trivial. A recent application based upon an optical flow (OF) technique was successful and briefly discussed. Further developments may encompass the usage of a combination of optical techniques, validated with detailed laboratory experiments in large-size physical models.


Hydraulics and Energy Dissipation on Stepped Spillways - Prototype and Laboratory Experience

The Philosophy of Flood Estimation     Back to Top

DOI: 10.3850/IAHR02262022-03380010


David Stephenson (University of Botswana)

Description of the weak point: 

The risk of incorrect flood estimation is as severe as the consequences of a big flood. The method selected for flood calculation can depend on the background of the analyst, but also on the available data. Hence, the model used has a big effect on the estimate.

Description of contribution: 

The method used to calculate floods and their risk has a big effect on the answer, as well as the impact of the flood, or damage. The more important the project the more sophisticated the method should be. Extrapolation for unmeasured extreme floods of a selected probability of non-occurrence cannot   be perfect. The probability and risk tolerated will affect the answer sought. Deterministic, empirical and mathematical solutions can differ considerably and alternative methods may be compared to provide confidence in the answer. The paper reviews a history of flood estimation methods applicable to risk and suggests how to decide the answer for extremes in the face of a number of uncertainties.


Flood Estimation Philosophy


Who are we?                         IAHR privacy policy

Copyright © 2023 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions