Visualizing hydrodynamic fluid simulations within an immersive experience as a scientific dissemination strategy

Abstract

Visualizing results is more important than ever in scientific dissemination. Natural hazards are complex phenomena; their examination and illustration call for a holistic approach when studying them and improving their communication in order to save lives and cost. In this paper, we present an overview of different methodologies ruling the visualization of flash floods and a proposal for integrated workflow for a more immersive experience and better risk perception. Our proposal will contribute to the current state-of-the-art on visualization of flash floods with an integrated visualization platform where precise and realistic flood simulations will be implemented. The contributions will affect small, ungauged steep rivers by the development of a model with a more precise and robust predictive capability at a local scale. The target of the hydraulic modelling executed in this research project is to achieve optimized simulations that could be carried out in the prototype of a serious gaming engine. The incentive for such optimized simulations is that complex hydrodynamic and morphodynamic flood simulations are data and computationally greedy. This contradicts the necessity for low complexity solutions needed in the real-time based scenarios that an immersive experience and on-site decision-making requires. Concluding remarks are that advanced fluid solvers integrated within computer graphics suites would provide the best results in terms of realistic visualization of reliable hydrodynamic simulations. Virtual reality engines can provide an experience arena or visualization, where different disciplines can meet and combine resources and knowledge for a common goal, such as the study and communication of natural hazards.

Full text

264 265
Xuyan Liu, Bjørn H. Skallerud, Victorien E. Prot and Gerhard A. Holzapfel
[14] A. Menzel. Modelling of anisotropic growth in biological tissues. A new approach and
computational aspects. Biomech. Model. Mechanobiol., 3:147–171, 2005.
[15] A. J. Schriefl, G. Zeindlinger, D. M. Pierce, P. Regitnig, and G. A. Holzapfel. Determi-
nation of the layer-specific distributed collagen fiber orientations in human thoracic and
abdominal aortas and common iliac arteries. J. R. Soc. Interface, 9:1275–1286, 2012.
[16] E. K. Rodriguez, A. Hoger, and A. D. McCulloch. Stress-dependent finite growth in soft
elastic tissues. J. Biomech., 27:455–467, 1994.
[17] G. A. Holzapfel. Nonlinear Solid Mechanics. A Continuum Approach for Engineering.
John Wiley & Sons, Chichester, 2000.
[18] G. A. Holzapfel and R. W. Ogden. Constitutive modelling of arteries. Proc. R. Soc. Lond.
A, 466:1551–1597, 2010.
[19] G. A. Holzapfel, J. A. Niestrawska, R. W. Ogden, A. J. Reinisch, and A. J. Schriefl.
Modelling non-symmetric collagen fibre dispersion in arterial walls. J. R. Soc. Interface,
12:20150188, 2015.
[20] O. G¨
ultekin, H. Dal, and G. A. Holzapfel. On the quasi-incompressible finite element
analysis of anisotropic hyperelastic materials. Comput. Mech., 63:443–453, 2019.
[21] T. C. Gasser, R. W. Ogden, and G. A. Holzapfel. Hyperelastic modelling of arterial layers
with distributed collagen fibre orientations. J. R. Soc. Interface, 3:15–35, 2006.
[22] T. C. Gasser and G. A. Holzapfel. Finite element modeling of balloon angioplasty by
considering overstretch of remnant non-diseased tissues in lesions. Comput. Mech., 40:47–
60, 2007.
[23] Abaqus: Dassault Syst `
emes Simulia Corp., Providence, RI, USA, 2019.
[24] A. Javili, B. Dortdivanlioglu, E. Kuhl, and C. Linder. Computational aspects of growth-
induced instabilities through eigenvalue analysis. Comput. Mech., 56:405–420, 2015.
16
10th National Conference on Computational Mechanics
MekIT’19
B. Skallerud and H. I. Andersson (Eds)
1
VISUALIZING HYDRODYNAMIC FLUID SIMULATIONS WITHIN
AN IMMERSIVE EXPERIENCE AS A SCIENTIFIC DISSEMINATION
STRATEGY
ADINA MORARU1, ODDBJØRN BRULAND1, ANDREW PERKIS2 AND NILS
RÜTHER1
1Department of Civil and Environmental Engineering
Norwegian University of Science and Technology (NTNU)
7491 Trondheim - Norway
e-mail: adina.moraru@ntnu.no; oddbjorn.bruland@ntnu.no; nils.ruther@ntnu.no
2Department of Electronic Systems
Norwegian University of Science and Technology (NTNU)
7491 Trondheim - Norway
e-mail: andrew.perkis@iet.ntnu.no
Key words: Visualization of Natural Hazards, Flash Floods in Steep Rivers, Hydrodynamic
Fluid Simulations, Immersive Experience, Quality of Experience, Risk Perception.
Abstract. Visualizing results is more important than ever in scientific dissemination. Natural
hazards are complex phenomena; their examination and illustration call for a holistic approach
when studying them and improving their communication in order to save lives and cost. In this
paper, we present an overview of different methodologies ruling the visualization of flash
floods and a proposal for integrated workflow for a more immersive experience and better risk
perception. Our proposal will contribute to the current state-of-the-art on visualization of flash
floods with an integrated visualization platform where precise and realistic flood simulations
will be implemented. The contributions will affect small, ungauged steep rivers by the
development of a model with a more precise and robust predictive capability at a local scale.
The target of the hydraulic modelling executed in this research project is to achieve optimized
simulations that could be carried out in the prototype of a serious gaming engine. The incentive
for such optimized simulations is that complex hydrodynamic and morphodynamic flood
simulations are data and computationally greedy. This contradicts the necessity for low
complexity solutions needed in the real-time based scenarios that an immersive experience and
on-site decision-making requires. Concluding remarks are that advanced fluid solvers
integrated within computer graphics suites would provide the best results in terms of realistic
visualization of reliable hydrodynamic simulations. Virtual reality engines can provide an
experience arena or visualization, where different disciplines can meet and combine resources
and knowledge for a common goal, such as the study and communication of natural hazards.
1. INTRODUCTION
Natural disasters are responsible for fatalities and economic losses worldwide and, among
them, floods are the most widespread and have caused the highest damages in recent years [1].
Human actions have become a dominant influence on fluvial systems and, together with
potential effects of climate change on flood regime (e.g. spatially restricted extreme rainfalls

266
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
2
affecting especially steep and ungauged rivers), predicting where major geomorphic changes
may occur is very challenging. Associated geomorphic processes cause a significant amount
of the damages related to floods and river response to floods can vary significantly [2].
Research efforts have fruitfully focused on major rivers, while thorough flood risk assessment
(i.e. including full hydro- and morphodynamic approaches) in small and steep ungauged rivers
has been constrained by the inadequacy of fluid solvers to tackle the complexity of the
dynamics affecting such rivers in more affordable applications [3]. New environmental
changes introduce new and more complex scenarios that translate into expertise demands (such
as detecting these changes promptly and increase social awareness of the problem at hand) that
researchers need to address.
Risk perception not only has an important role in shaping natural hazards policies. In order
to increase public knowledge and influence its opinion, it is crucial to understand why people
have diverging attitudes and perceptions related to natural hazards and their possible
consequences. The perception to an extreme event is explained to a certain degree by the direct
personal experience of the damage caused by climate-related events, such as flooding or
landslides, as shown by Lujala et al. [4]. This has the largest impact on the subjects’ belief that
there will occur more natural hazards locally than nationally or globally. It is noteworthy that
merely living in a more exposed area, but not having a personal experience of the phenomenon,
does not affect the population’s concern towards natural hazards.
Available models for flood hazard assessment in steep and ungauged rivers converge in
outcomes that do not convey the estimated risk based on a user-friendly and relatable three-
dimensional real world nor is the analysis often based on the most recent and highly accurate
data at hand [5]. Therefore, the risk perception achieved is hardly ever in accordance with the
hydraulic model presented. In order to fill this gap, the World of Wild Waters (WoWW) project
aims at being the future tool for analyzing and communicating the potential causes and effects
of natural hazards, where the end user is not a scientist, but a land manager or any other user
with no background in hydrology and hydraulics. Its orientation towards the gamification of
natural disasters and its aim at bringing together knowledge on the physical properties of these
with knowledge on digital storytelling and human behaviour emerges into an immersive
experience based on real data, realistic scenarios and simulations. WoWW project acts as a
framework for the setting of the research strategy here presented. As part of WoWW’s work
packages, the investigation scheme described in this document targets the Visualization of
Flash Floods as a strategy to improve the scientific dissemination of one of the most recurring
and of highest impact natural disasters affecting small, ungauged steep rivers in Norway and
worldwide. Particularly, it is aiming to dynamically visualize water flow in small, ungauged
steep catchments and the effect of water forces on their riverbanks and structures in and along
these watercourses based on knowledge about hydraulics and hydraulic modelling and enable
more efficient hydrodynamic fluid simulations for better flood risk assessment and
communication. The latest trends in realistic data presentation and scientific visualization go
hand in hand with the implementation of a flooding scenario in Extended Reality (XR, i.e. the
combination of Augmented/Mixed/Virtual Reality; AR/MR/VR, respectively). Immersive
technologies require very advanced visualization modalities. The outcome of this project will
be implemented into the prototype for a serious gaming (e.g. VR/MR) flood engine.
The aim of the present document is to highlight the need of a holistic approach when
studying natural hazards and improving the communication of these in order to save life and
cost. Hereinto is presented an overview of different methodologies and a proposal of integrated
workflow that will contribute to the current state-of-the-art on visualization of flash floods
affecting small, ungauged steep rivers by developing a model with an improved predictive
capability at a local scale. An efficient and robust flood simulation model should conceive
high-quality and most updated data retrieval from the Internet of Things (IoT), precise flood
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
3
area estimation and representation, reduced computational time scale (i.e. currently of one to
several hours in the case of 2D hydraulic simulations), on-site application and analysis
accuracy [6]. The model will be calibrated with monitoring data collected through the
environmental IoT and further tested within WoWW. Regarding the visualization of the results,
this research will include the development of a system with a geo-referenced 3D environment
on which the model can be implemented and coupled into a virtual reality gaming engine.
In the following sections, we will address key concepts regarding the visualization of natural
hazards. We will also describe the characteristics of numerical models and visualization trends
affecting flash floods and propose an integrated workflow scheme for the visualization of
natural hazards. The implementation of this workflow scheme will be addressed in a step-wise
fashion and we will discuss the suitability of the methodology for WoWW’s purpose.
2. KEY CONCEPTS ON THE VISUALIZATION OF NATURAL HAZARDS
The multidisciplinary aspect of WoWW give rise for a need to unify common terminology.
The most controversial concepts identified were, for instance, simulation, modelling, real-time,
visualization and gamification. Other useful concepts regarding the currently presented topic,
yet not often used in the field of hydraulic engineering (e.g. immersive experience, quality of
experience), are also defined in this section.
A simulation is carried out when a particular set of conditions are created artificially in order
to study or experience something that could exist in reality. In this study, this concept will be
handled in the context of fluid (i.e. water, specifically) simulation and the analysis of its flow
patterns, directions, forces and physical characteristics and properties. We will address,
therefore, hydrodynamic simulations throughout the following sections. For instance,
hydrodynamic modelling has been defined as the mathematical application of momentum,
continuity and transport conservation equations to represent evolving fluid flow in terms of
velocity, density and scalar fields [7] [8].
Real-time refers to an immediate computational response to input data. In the context of
hydrodynamic fluid simulations, real-time refers to the simulation speed being equal to the
speed of the simulated event when occurring in real life. Real-time visualization refers, on the
other hand, to the display of the simulated scenario at real life speed. In this sense, it is
noteworthy that the temporal perception is human-relative, and not computer-relative. As
suggested by Henonin et al. [6], real-time flood assessment may be classified into three
categories depending on the role of hydrodynamic fluid modelling in the hazard evaluation: i)
empirical scenarios-based (i.e. no hydraulic model is used), ii) pre-simulated scenarios-based
(i.e. hydraulic models are used as pre-study tools) and iii) real-time simulations-based (i.e. real-
time forecast with online and real-time simulation models). Although previous hydrodynamic
simulations are often used to assess flood hazard, we will aim at developing an integrated real-
time simulation-based model that can be scaled and reproducible in future prototypes.
Scientific visualization is often mentioned regarding the representation of three-dimensional
phenomena, where the efforts are committed to realistic renderings (i.e. the processing of an
image using colour and shading to translate it into a solid and three-dimensional look) of
volumes, surfaces, illumination sources, and so forth, often with a dynamic (temporal)
component and oriented to subsequent analysis [9]. This interdisciplinary branch is frequently
also considered a subset of computer graphics and its goal is to depict scientific data graphically
so that scientists understand, illustrate and gain insight from their data.
The visualization of flash floods facilitates disclosing information regarding their analysis
and management in a universal fashion. In our case, the display is predominantly of
geographical data, which may overlay a map, terrain model, or even an orthophoto.
Traditionally, experts working on risk assessment have based its representation on colour-
coded maps. Some numerical models use maps or orthophotos to texture the Digital Elevation

266 267
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
2
affecting especially steep and ungauged rivers), predicting where major geomorphic changes
may occur is very challenging. Associated geomorphic processes cause a significant amount
of the damages related to floods and river response to floods can vary significantly [2].
Research efforts have fruitfully focused on major rivers, while thorough flood risk assessment
(i.e. including full hydro- and morphodynamic approaches) in small and steep ungauged rivers
has been constrained by the inadequacy of fluid solvers to tackle the complexity of the
dynamics affecting such rivers in more affordable applications [3]. New environmental
changes introduce new and more complex scenarios that translate into expertise demands (such
as detecting these changes promptly and increase social awareness of the problem at hand) that
researchers need to address.
Risk perception not only has an important role in shaping natural hazards policies. In order
to increase public knowledge and influence its opinion, it is crucial to understand why people
have diverging attitudes and perceptions related to natural hazards and their possible
consequences. The perception to an extreme event is explained to a certain degree by the direct
personal experience of the damage caused by climate-related events, such as flooding or
landslides, as shown by Lujala et al. [4]. This has the largest impact on the subjects’ belief that
there will occur more natural hazards locally than nationally or globally. It is noteworthy that
merely living in a more exposed area, but not having a personal experience of the phenomenon,
does not affect the population’s concern towards natural hazards.
Available models for flood hazard assessment in steep and ungauged rivers converge in
outcomes that do not convey the estimated risk based on a user-friendly and relatable three-
dimensional real world nor is the analysis often based on the most recent and highly accurate
data at hand [5]. Therefore, the risk perception achieved is hardly ever in accordance with the
hydraulic model presented. In order to fill this gap, the World of Wild Waters (WoWW) project
aims at being the future tool for analyzing and communicating the potential causes and effects
of natural hazards, where the end user is not a scientist, but a land manager or any other user
with no background in hydrology and hydraulics. Its orientation towards the gamification of
natural disasters and its aim at bringing together knowledge on the physical properties of these
with knowledge on digital storytelling and human behaviour emerges into an immersive
experience based on real data, realistic scenarios and simulations. WoWW project acts as a
framework for the setting of the research strategy here presented. As part of WoWW’s work
packages, the investigation scheme described in this document targets the Visualization of
Flash Floods as a strategy to improve the scientific dissemination of one of the most recurring
and of highest impact natural disasters affecting small, ungauged steep rivers in Norway and
worldwide. Particularly, it is aiming to dynamically visualize water flow in small, ungauged
steep catchments and the effect of water forces on their riverbanks and structures in and along
these watercourses based on knowledge about hydraulics and hydraulic modelling and enable
more efficient hydrodynamic fluid simulations for better flood risk assessment and
communication. The latest trends in realistic data presentation and scientific visualization go
hand in hand with the implementation of a flooding scenario in Extended Reality (XR, i.e. the
combination of Augmented/Mixed/Virtual Reality; AR/MR/VR, respectively). Immersive
technologies require very advanced visualization modalities. The outcome of this project will
be implemented into the prototype for a serious gaming (e.g. VR/MR) flood engine.
The aim of the present document is to highlight the need of a holistic approach when
studying natural hazards and improving the communication of these in order to save life and
cost. Hereinto is presented an overview of different methodologies and a proposal of integrated
workflow that will contribute to the current state-of-the-art on visualization of flash floods
affecting small, ungauged steep rivers by developing a model with an improved predictive
capability at a local scale. An efficient and robust flood simulation model should conceive
high-quality and most updated data retrieval from the Internet of Things (IoT), precise flood
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
3
area estimation and representation, reduced computational time scale (i.e. currently of one to
several hours in the case of 2D hydraulic simulations), on-site application and analysis
accuracy [6]. The model will be calibrated with monitoring data collected through the
environmental IoT and further tested within WoWW. Regarding the visualization of the results,
this research will include the development of a system with a geo-referenced 3D environment
on which the model can be implemented and coupled into a virtual reality gaming engine.
In the following sections, we will address key concepts regarding the visualization of natural
hazards. We will also describe the characteristics of numerical models and visualization trends
affecting flash floods and propose an integrated workflow scheme for the visualization of
natural hazards. The implementation of this workflow scheme will be addressed in a step-wise
fashion and we will discuss the suitability of the methodology for WoWW’s purpose.
2. KEY CONCEPTS ON THE VISUALIZATION OF NATURAL HAZARDS
The multidisciplinary aspect of WoWW give rise for a need to unify common terminology.
The most controversial concepts identified were, for instance, simulation, modelling, real-time,
visualization and gamification. Other useful concepts regarding the currently presented topic,
yet not often used in the field of hydraulic engineering (e.g. immersive experience, quality of
experience), are also defined in this section.
A simulation is carried out when a particular set of conditions are created artificially in order
to study or experience something that could exist in reality. In this study, this concept will be
handled in the context of fluid (i.e. water, specifically) simulation and the analysis of its flow
patterns, directions, forces and physical characteristics and properties. We will address,
therefore, hydrodynamic simulations throughout the following sections. For instance,
hydrodynamic modelling has been defined as the mathematical application of momentum,
continuity and transport conservation equations to represent evolving fluid flow in terms of
velocity, density and scalar fields [7] [8].
Real-time refers to an immediate computational response to input data. In the context of
hydrodynamic fluid simulations, real-time refers to the simulation speed being equal to the
speed of the simulated event when occurring in real life. Real-time visualization refers, on the
other hand, to the display of the simulated scenario at real life speed. In this sense, it is
noteworthy that the temporal perception is human-relative, and not computer-relative. As
suggested by Henonin et al. [6], real-time flood assessment may be classified into three
categories depending on the role of hydrodynamic fluid modelling in the hazard evaluation: i)
empirical scenarios-based (i.e. no hydraulic model is used), ii) pre-simulated scenarios-based
(i.e. hydraulic models are used as pre-study tools) and iii) real-time simulations-based (i.e. real-
time forecast with online and real-time simulation models). Although previous hydrodynamic
simulations are often used to assess flood hazard, we will aim at developing an integrated real-
time simulation-based model that can be scaled and reproducible in future prototypes.
Scientific visualization is often mentioned regarding the representation of three-dimensional
phenomena, where the efforts are committed to realistic renderings (i.e. the processing of an
image using colour and shading to translate it into a solid and three-dimensional look) of
volumes, surfaces, illumination sources, and so forth, often with a dynamic (temporal)
component and oriented to subsequent analysis [9]. This interdisciplinary branch is frequently
also considered a subset of computer graphics and its goal is to depict scientific data graphically
so that scientists understand, illustrate and gain insight from their data.
The visualization of flash floods facilitates disclosing information regarding their analysis
and management in a universal fashion. In our case, the display is predominantly of
geographical data, which may overlay a map, terrain model, or even an orthophoto.
Traditionally, experts working on risk assessment have based its representation on colour-
coded maps. Some numerical models use maps or orthophotos to texture the Digital Elevation

268
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
4
Model (DEM; i.e. represent the land use on the DEM). However, even though the DEM is
three-dimensional, the calculated parameters are visualized as two-dimensional colour-codes
if the model used solved equations only in two dimensions. The suitability of the models used
in hydrodynamic fluid simulations will be discussed further sections.
Gamification is the process of arranging non-game content by means of game mechanics
with the main objective of positively influencing behavior and enhancing the user’s motivation
[10]. This is usually done by incorporating elements that the user must interact with in order to
achieve certain goals provided by the game instructions. In serious gaming, the purpose is to
create a better understanding of a certain concept or topic; hence, it is often aligned to
educational or business goals. Although the gamification of natural hazards will not be
discussed in the present document, the gamification of floods is contemplated as one-step
beyond the visualization of floods. Therefore, the visualization scheme here presented must
ensure the potential of gamifying its results in order to facilitate an improved immersive
experience.
Immersive Experience is a representation of the (virtually created) reality that allows the
audience to be engaged with the visualized content to the extent of perceiving themselves as
being present in the displayed surrounding environment. An immersive experience ought to be,
among others, accurate, realistic if relevant, emotional, context adaptive, engaging, useful,
interactive, intuitive, etc. [11] [12] [13]. This definition goes very much on the lines of the
updated concept of Quality of Experience (QoE), defined by the Qualinet group of experts as
“the degree of delight or annoyance of the end user of an application or service. It results from
the fulfillment his or her expectations with respect to the utility and/or enjoyment of the
application or service in the light of the user’s personality and current state”. In fact, the concept
of immersive experiences has evolved into the QoE concept, and the trends have shifted in
favor of the latter. QoE better differentiates related terminology such as performance, Quality
of Service and application acceptance, and it focuses its efforts on evaluating the user
experience based on a rigorously designed methodology that contemplates both objective and
subjective metrics. As the Quality of Experience is inherently dependent on system-, human-
and context-influencing factors [14], the content design and its display shall be carefully
conducted taking into account all these factors and their interrelations.
Immersive Media Technology Experiences (IMTE) are game changers when it comes to
transferring knowledge and influencing lifestyles as they tackle these tasks with a human-
centred designed approach [15]. IMTE have been used so far in fields such as entertainment,
medical and biosciences, art, oil and gas, aerospace and naval, automotive, power and traffic,
gaming industries, etc. Nevertheless, this science has not been overly applied in natural
hazards, risk perception and risk assessment-related studies. A recent experiment [16] has
proven that, if the user related emotionally to the experience evaluated, the level of immersion
was higher than that of when being spatially immersed. A thorough content design process,
where risk perception and system factors are interlaced, and with a visualization of natural
hazards that presents the characteristics needed to achieve an immersive experience (e.g.
accurate and context-adaptive in terms of physics; relevant and realistic in terms of scenario
selection; engaging, interactive and able of creating an affinity between the user and the
displayed content) will result in a more effective communication of flood risk.
3. CHARACTERISTICS OF NUMERICAL MODELS AND VISUALIZATION
TRENDS APPLIED TO FLASH FLOODS
The vast advances technology has experienced in the last decades provide a limitless array
of tools when it comes to studying flash floods and river’s response to them. Flash floods are
studied through the definition and characterization of spatial and temporal parameters. In this
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
5
sense, Geographic Information Systems (GIS) and related technology help representing and
visualizing data in real-time and in an understandable way (see for example [17] [18] [19]
[20]). A quick display of information allows a more effective and competent implementation
of alleviation measures and flood management practices. Experts are looking for ways to
integrate the existing tools and use them in interdisciplinary projects in order to make the best
of the immense possibilities these tools provide.
3.1. Characteristics of numerical models for flash flood simulation
Numerical simulations have become a common tool to approach engineering problems for
which there is no available similar analytical solution. One possible criterion when selecting a
suitable model is to see if our approach considers an analysis dependent on field observations
and the statistical analysis of the relationship between these and other characteristics of the
phenomenon (i.e. empirical), regards the total phenomenon at a given point in space (i.e. is
purely analytical) or if it is based on Computational Fluid Dynamics (CFD) and solves
numerically once the parametres are discretized in time and space (i.e. numerical models). In
this research, numerical models will be used and contrasted afterwards with either empirical or
different numerical models. Moreover, we will rather refer to the different models, their
characteristics and suitability based on the continuum mechanics approach and spatial
dimensions the model is contemplating (Table 1). This approach allows taking into
consideration the target and level of simplification required by the analysis in question.
If discretized numerical models into 1-, 2- and 3-dimensional, the most extended form of
CFD equations concerning flood simulations are the Reynolds-Averaged Navier-Stokes
equations (also referred to as RANS) and the Shallow-Water Equations (SWE), which are
derived from the Navier-Stokes equations considering simplifications on the third dimension
[21] [22]. They allow the prediction of the fluid dynamics, e.g. water velocity, water elevation,
water forces, shear stress, stream power, travelled distance by the fluid, hence, if there is
overflow of the riverbanks and where would the overflowing areas be. A recent comparative
study of 1D, 2D and combined 1D/2D models applied to the same study case in HEC-RAS
[23] showed that all three models could successfully reproduce a historic flooding event.
Moreover, the 2D and 1D/2D model could also provide relatively detailed information
regarding flood propagation and velocities on the floodplain.
A detailed comparison of 2D hydraulic modelling packages is provided in [21]. On the other
hand, two-dimensional modelling is not very comprehensive in incorporating secondary
circulation at bends and three-dimensional modelling is favoured to study that behavior (Table
1). For instance, HEC-RAS is unable to work with falls and steps and changing flow regimes,
which is reflected in 2D hydrodynamic simulations of steep slopes [22]. Furthermore, the slope
limitation in one-dimensional numerical simulations carried out in this software is, basically,
because the 1D St. Venant equation is derived with the assumption that the bed slope is very
small and higher slopes come in contradiction with this assumption.
Regarding 3D modelling, often applied to short river stretches in order to solve complex
specific issues, such as vertical turbulence, vortices, secondary circulation, bed mobilization
and bank erosion, the reference frame utilized also differentiates models. For instance, an
Eulerian reference frame is grid-based and fixed in space while the Lagrangian reference frame
is particle-based and takes into account the movement with the local velocity. The advantage
of Lagrangian models is that they do not require spatial discretization and are able to represent
smaller features than the grid size. Moreover, particle-based models provide a higher accuracy
and non-diffusive prediction of convection. Different models use different fluid assumptions,
nevertheless.

268 269
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
4
Model (DEM; i.e. represent the land use on the DEM). However, even though the DEM is
three-dimensional, the calculated parameters are visualized as two-dimensional colour-codes
if the model used solved equations only in two dimensions. The suitability of the models used
in hydrodynamic fluid simulations will be discussed further sections.
Gamification is the process of arranging non-game content by means of game mechanics
with the main objective of positively influencing behavior and enhancing the user’s motivation
[10]. This is usually done by incorporating elements that the user must interact with in order to
achieve certain goals provided by the game instructions. In serious gaming, the purpose is to
create a better understanding of a certain concept or topic; hence, it is often aligned to
educational or business goals. Although the gamification of natural hazards will not be
discussed in the present document, the gamification of floods is contemplated as one-step
beyond the visualization of floods. Therefore, the visualization scheme here presented must
ensure the potential of gamifying its results in order to facilitate an improved immersive
experience.
Immersive Experience is a representation of the (virtually created) reality that allows the
audience to be engaged with the visualized content to the extent of perceiving themselves as
being present in the displayed surrounding environment. An immersive experience ought to be,
among others, accurate, realistic if relevant, emotional, context adaptive, engaging, useful,
interactive, intuitive, etc. [11] [12] [13]. This definition goes very much on the lines of the
updated concept of Quality of Experience (QoE), defined by the Qualinet group of experts as
“the degree of delight or annoyance of the end user of an application or service. It results from
the fulfillment his or her expectations with respect to the utility and/or enjoyment of the
application or service in the light of the user’s personality and current state”. In fact, the concept
of immersive experiences has evolved into the QoE concept, and the trends have shifted in
favor of the latter. QoE better differentiates related terminology such as performance, Quality
of Service and application acceptance, and it focuses its efforts on evaluating the user
experience based on a rigorously designed methodology that contemplates both objective and
subjective metrics. As the Quality of Experience is inherently dependent on system-, human-
and context-influencing factors [14], the content design and its display shall be carefully
conducted taking into account all these factors and their interrelations.
Immersive Media Technology Experiences (IMTE) are game changers when it comes to
transferring knowledge and influencing lifestyles as they tackle these tasks with a human-
centred designed approach [15]. IMTE have been used so far in fields such as entertainment,
medical and biosciences, art, oil and gas, aerospace and naval, automotive, power and traffic,
gaming industries, etc. Nevertheless, this science has not been overly applied in natural
hazards, risk perception and risk assessment-related studies. A recent experiment [16] has
proven that, if the user related emotionally to the experience evaluated, the level of immersion
was higher than that of when being spatially immersed. A thorough content design process,
where risk perception and system factors are interlaced, and with a visualization of natural
hazards that presents the characteristics needed to achieve an immersive experience (e.g.
accurate and context-adaptive in terms of physics; relevant and realistic in terms of scenario
selection; engaging, interactive and able of creating an affinity between the user and the
displayed content) will result in a more effective communication of flood risk.
3. CHARACTERISTICS OF NUMERICAL MODELS AND VISUALIZATION
TRENDS APPLIED TO FLASH FLOODS
The vast advances technology has experienced in the last decades provide a limitless array
of tools when it comes to studying flash floods and river’s response to them. Flash floods are
studied through the definition and characterization of spatial and temporal parameters. In this
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
5
sense, Geographic Information Systems (GIS) and related technology help representing and
visualizing data in real-time and in an understandable way (see for example [17] [18] [19]
[20]). A quick display of information allows a more effective and competent implementation
of alleviation measures and flood management practices. Experts are looking for ways to
integrate the existing tools and use them in interdisciplinary projects in order to make the best
of the immense possibilities these tools provide.
3.1. Characteristics of numerical models for flash flood simulation
Numerical simulations have become a common tool to approach engineering problems for
which there is no available similar analytical solution. One possible criterion when selecting a
suitable model is to see if our approach considers an analysis dependent on field observations
and the statistical analysis of the relationship between these and other characteristics of the
phenomenon (i.e. empirical), regards the total phenomenon at a given point in space (i.e. is
purely analytical) or if it is based on Computational Fluid Dynamics (CFD) and solves
numerically once the parametres are discretized in time and space (i.e. numerical models). In
this research, numerical models will be used and contrasted afterwards with either empirical or
different numerical models. Moreover, we will rather refer to the different models, their
characteristics and suitability based on the continuum mechanics approach and spatial
dimensions the model is contemplating (Table 1). This approach allows taking into
consideration the target and level of simplification required by the analysis in question.
If discretized numerical models into 1-, 2- and 3-dimensional, the most extended form of
CFD equations concerning flood simulations are the Reynolds-Averaged Navier-Stokes
equations (also referred to as RANS) and the Shallow-Water Equations (SWE), which are
derived from the Navier-Stokes equations considering simplifications on the third dimension
[21] [22]. They allow the prediction of the fluid dynamics, e.g. water velocity, water elevation,
water forces, shear stress, stream power, travelled distance by the fluid, hence, if there is
overflow of the riverbanks and where would the overflowing areas be. A recent comparative
study of 1D, 2D and combined 1D/2D models applied to the same study case in HEC-RAS
[23] showed that all three models could successfully reproduce a historic flooding event.
Moreover, the 2D and 1D/2D model could also provide relatively detailed information
regarding flood propagation and velocities on the floodplain.
A detailed comparison of 2D hydraulic modelling packages is provided in [21]. On the other
hand, two-dimensional modelling is not very comprehensive in incorporating secondary
circulation at bends and three-dimensional modelling is favoured to study that behavior (Table
1). For instance, HEC-RAS is unable to work with falls and steps and changing flow regimes,
which is reflected in 2D hydrodynamic simulations of steep slopes [22]. Furthermore, the slope
limitation in one-dimensional numerical simulations carried out in this software is, basically,
because the 1D St. Venant equation is derived with the assumption that the bed slope is very
small and higher slopes come in contradiction with this assumption.
Regarding 3D modelling, often applied to short river stretches in order to solve complex
specific issues, such as vertical turbulence, vortices, secondary circulation, bed mobilization
and bank erosion, the reference frame utilized also differentiates models. For instance, an
Eulerian reference frame is grid-based and fixed in space while the Lagrangian reference frame
is particle-based and takes into account the movement with the local velocity. The advantage
of Lagrangian models is that they do not require spatial discretization and are able to represent
smaller features than the grid size. Moreover, particle-based models provide a higher accuracy
and non-diffusive prediction of convection. Different models use different fluid assumptions,
nevertheless.

270
10th National Conference on Computational Mechanics
MekIT’19
B. Skallerud and H. I. Andersson (Eds)
6
Table 1. Comparison of hydrodynamic models (based on [8] [21] [22] [23] [24] [26]). A model with more dimensions includes also the features of that with less dimensions.
Criterion 1D models 2D models 3D models
Governing equations Bernouilli (steady flow);
1D Saint Venant (unsteady flow)
2D Saint Venant (SWE);
2D Reynolds-Averaged Navier Stokes (RANS)
3D Reynolds-Averaged Navier-Stokes (RANS)
Assumptions/
simplifications
Cross-sectional averaged velocity, topography
is discrete, constant lateral variables
Hydrostatic pressure distribution, Reynolds averaged
over shallow depth, uniform vertical distribution of
velocity, incompressible fluid with uniform density
Conservation of momentum and continuity
(incompressible fluid), Turbulence is modelled instead
of assumed
Features
Needs a predefined flow path, subjective to
cross-sectional spacing, orientation and
location. Free surface calculated with pressure
Topography is continuous over the cross-section. Takes
into account viscosity, shear stress, wall friction, inflow
volume and momentum, turbulence is incorporated
through Manning’s equation. Free surface is calculated
with level set
Grid-based or Particle-based, includes a “multi-physics”
environment. Free surface is calculated with volume of
fluid
Applications
Little spatial variation and uniform flow
simulations (e.g. deep flooding in narrow
streets, long reaches (>100km), etc.)
Large spatial variability and non-uniform flow (e.g. steep
rivers, urba n and coastal areas, wi de flood plai ns). Also
used in morphodynamic studies
In short river stretches and localized complex dynamics
in 3D (e.g. vertical turbulence, vortices, secondary
circulation). Suitable for simulation of weirs. Also in
morphodynamic studies
Accuracy Less accurate if the processes prime in more
than 1D
More accurate than 1D in terms of lateral diffusion of
flood wave and flow recirculation
More detailed information on flow behavior, provides
detailed analysis on streamlines
Advantages
Computationally more efficient, possibility to
couple with 2D models, extensive literature
available. Ideal as a preliminary approach to
more complex problems
Makes better use of topographic information than 1D
models, ability to represent more detailed changes in
velocity and flow depth and direction, flexible meshes,
time discretization, shock-capture schemes
Solves full RANS equations directly, mass and
momentum conservation equations contemplate
additional terms, incorporates tuning parametres of
energy dissipation. Possibility to propose and check
changes in structures (design-oriented)
Limitations
Does not simulate lateral flood wave diffusion,
averages velocity and topography is discrete,
cannot describe different water levels at the
same section, does not describe local hydraulic
behavior, turbulence is disregarded
Does not represent channel bend-induced secondary
circulation, not suitable for large-scale systems (>100km)
or topographic discontinuities, as it neglects vertical
velocity by averaging it. Requires calibration on
discharge coefficient (in case of weirs) which is laborious
and complex
Requires the calibration of the turbulence model and
roughness value (which are key in energy dissipation)
Speed and cost Quick and inexpensive Computational speed and cost dependent on problem
complexity
Computational greedy and expensive
Available Software
e.g. HEC-RAS, Flood Modeller-TUFLOW,
Mike 11, IDSS
e.g. Iber, HEC-RAS, Telemac-Mascaret, Flood
Modeller-TUFLOW, Flow2D, Mike 21, ANUGA, Dual
SPHysics, REEF3D, Autodesk InfraWorks, Gerris,
e.g. REEF3D, Gerris, OpenFOAM, Blender, Flow3D,
Maya-GLU3D, Bifröst
10
th
National Conference on Computational Mechanics
MekIT’19
B. Skallerud and H. I. Andersson (Eds)
7
Depending on the degree of complexity of the pondered scenario, new and more
sophisticated demands have called for new and more integrated models. These models differ
depending on the required scope of the analysis, the scale, available input data and the extent
of the output aimed for. In Figure 1, different numerical models, such as Direct Numerical
Simulation (DNS), Large Eddy Simulation (LES), RANS, Double-Averaged Navier-Stokes
equations (DANS), SWE and Diffusive Wave Equation (DWE), are plotted based on their
applicability to new and more complex fluid modelling demands (i.e. their role in real-time
flood risk assessment and the level of simplification that they include). Often enough, these
involve combining hydrodynamic fluid simulations with algorithms that increase their
efficiency or improve their precision (see for example [27] [28]). The computational cost of
3D models is still a noteworthy disadvantage.
Figure 1. Numerical models according to the analysed spatial scale and their usability in flood risk analysis
(modified after [6] and [29]). The level of complexity of the model is linearly correlated with that of the
simulated scenario and the scale at which this is studied. See text for abbreviations.
3.2. Trends in the visualization of flash floods
Several studies tackle the visualization of flood events (see for example [30] [31] [32] [33]
[34] [35] [36] [37] [38]). However, the tools presented until now lack common principles and
approach to visualization, as well as the unification of data formats due to the wide-ranging
amount of software offered in the market. Suhr [39] defined sound decision-making as a reality-
based, congruent and effective decision-making, as it is founded on “the correct use of correct
data”. Widely used nowadays, sound decision-making consists not only on the accurate
selection of data, but also on the comparison of these in order to detect the best alternative (i.e.
the most advantageous alternative). In flood risk management, sound decision-making requires
the use of reliable decision-support tools; nonetheless, worldwide there is currently no
integrated model [19] [30] [35] for both excellent risk assessment and effective communication
of the potential impact of flood risk in small steep rivers to the stakeholders.
Visualizing large and realistic flood scenarios is, in fact, complex and requires the use of
excellent state-of-the-art graphic tools that allow rendering these as quasi real-time scenarios.
The display of grid computation results obtained in flood modelling arises the need of a unified
visualization scheme with unified standards, e.g. integration of input data formats [31]. If grid
computation is required, it is fundamental to have a pre-designed input workflow, as this

270 271
10
th
National Conference on Computational Mechanics
MekIT’19
B. Skallerud and H. I. Andersson (Eds)
7
Depending on the degree of complexity of the pondered scenario, new and more
sophisticated demands have called for new and more integrated models. These models differ
depending on the required scope of the analysis, the scale, available input data and the extent
of the output aimed for. In Figure 1, different numerical models, such as Direct Numerical
Simulation (DNS), Large Eddy Simulation (LES), RANS, Double-Averaged Navier-Stokes
equations (DANS), SWE and Diffusive Wave Equation (DWE), are plotted based on their
applicability to new and more complex fluid modelling demands (i.e. their role in real-time
flood risk assessment and the level of simplification that they include). Often enough, these
involve combining hydrodynamic fluid simulations with algorithms that increase their
efficiency or improve their precision (see for example [27] [28]). The computational cost of
3D models is still a noteworthy disadvantage.
Figure 1. Numerical models according to the analysed spatial scale and their usability in flood risk analysis
(modified after [6] and [29]). The level of complexity of the model is linearly correlated with that of the
simulated scenario and the scale at which this is studied. See text for abbreviations.
3.2. Trends in the visualization of flash floods
Several studies tackle the visualization of flood events (see for example [30] [31] [32] [33]
[34] [35] [36] [37] [38]). However, the tools presented until now lack common principles and
approach to visualization, as well as the unification of data formats due to the wide-ranging
amount of software offered in the market. Suhr [39] defined sound decision-making as a reality-
based, congruent and effective decision-making, as it is founded on “the correct use of correct
data”. Widely used nowadays, sound decision-making consists not only on the accurate
selection of data, but also on the comparison of these in order to detect the best alternative (i.e.
the most advantageous alternative). In flood risk management, sound decision-making requires
the use of reliable decision-support tools; nonetheless, worldwide there is currently no
integrated model [19] [30] [35] for both excellent risk assessment and effective communication
of the potential impact of flood risk in small steep rivers to the stakeholders.
Visualizing large and realistic flood scenarios is, in fact, complex and requires the use of
excellent state-of-the-art graphic tools that allow rendering these as quasi real-time scenarios.
The display of grid computation results obtained in flood modelling arises the need of a unified
visualization scheme with unified standards, e.g. integration of input data formats [31]. If grid
computation is required, it is fundamental to have a pre-designed input workflow, as this

272
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
8
controls the efficient execution of the visualization tool. The visualization scheme should
integrate visualization requests of any type of application oriented on the computation of
natural hazards.
Recent developments in the visualization of floods include numerical computations and
simulations in 3D. For instance, Ghazali and Kamsin [17] used the combination of SPH
(through the GLU3D plug-in) and 3D computer graphics (in this case, Maya) to simulate and
model flash floods on a three-dimensional geo-referenced environment (i.e. LiDAR DEM).
Moreover, they used most of the software’s potential by testing the realistic visualization of
natural hazards in real-time through an Application Programming Interface (API) in Maya.
Later on, Li et al. [32] used OpenGL for flood simulations in 3D by first creating the 3D terrain
and sky background, and later simulating the water flow and its depth overlaid on the terrain.
Parallel, Ye et al. [40] used SPH to model the water flow during a dam break and their
computational output was embedded in a 3D spatio-temporal GIS application where this and
other flood scenarios where dynamically visualized. Additionally, specific layers could be
added to show public infrastructures within the system. Demir and Krajewski [34] juxtaposed
in their research flood analysis, adaptive real-time communication of flooding conditions and
the interactive visualization of results. Nevertheless, the data provided was very complex,
including river conditions, flood maps, forecasting and related information. Cartoon-style
displays of results in an online platform made the data easily accessible and understandable to
the end user.
Most recently, Macchione et al. [37] represented 2D hydraulic simulations within a so-
called 3D virtual reality environment while aiming at presenting a product potentially useful
for hydraulic engineers for risk communication purposes. Their workflow contemplated a
sensible compromise between the inherent complexity of virtual reality and the need to
represent flooding events in 3D environments to improve the interaction with decision-makers
and to engage people with natural hazards. Zhang et al. [36] [41] went a step further and aimed
to improve data visualization, increase simulation speed and allow real-time interaction during
the simulation process. This lead in the development of a 3D flood simulation platform with
VR technology. They analyzed flood processes, simulated the flow field as well as the breach
flood process and contemplated the emergency plan making. Moreover, their digital platform
not only represented the real-time changing process through the “instruction-response” method
and data interpolation, but also combined the virtual visualization of the data with numerical
modelling in a 3D visual form based on modular software design. Although their methods
involved the use of coding techniques, numerical modelling and virtual reality tools (e.g. C++,
FORTRAN, OSG, VPB, osgGIS, intranet, middleware, etc.) altogether, the simulation speed
and the interaction between numerical models had room for improvement regarding the
achieved level of immersion and complexity of interactive functions.
Wang et al. [5] tackled the need for an improvement of flood risk communication and the
real-time flood risk assessment through a combined simulation-visualization approach
somewhat similar to that of Macchione et al. [37] in terms of presenting 2D hydraulic
simulations in an improved 3D environment. They enhanced decision support by including the
analysis of the uncertainties of such model as well as by increasing the computational
efficiency in data assimilation and calculation with the help of algorithms. Their study did not
limit to the presentation of aesthetically attractive scenarios, where the 3D model was rendered
by means of computer graphic-assisted improvement. A gamification process adds value to
their study and is noteworthy. They added interactive elements, such as user-interactive
features e.g. 3D “drag-and drop” icons, design analysis, 3D stereo panorama, storyboards or
online shared view. These elements allow the user of their application to position elements
wherever they want, measure the inundated area, share the experience as a web link or QR code
that displays a 360º rendered panoramic view, follow a tour of static and dynamic specified
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
9
views or even capture their comments of design plans. All these tools permit a greater
immersive experience and a much more effective communication of the presented hazardous
scenarios. The interaction of user-model through virtual reality increases the stakeholders’
implication in planning and decision-making and provides instant feedback on how to increase
the model’s potential and effectiveness.
The most suitable model should be selected based not only on the most advanced
technologies and the attempts to tackle very complex problems. The model choice should be
ruled by what works best in the given context, considering the level of investment needed (i.e.
data and financially speaking), the hydraulic context and the precision needed for decision-
making. Characteristics of numerical models and model-selection criteria are described in
section 3.1. WoWW is aiming at potential real-time simulation-based solutions, where the
output is both physically and visually realistic and enables an accurate flood risk assessment
and its communication to the stakeholders, whom are oftentimes not part of the scientific
community. For such purpose, the most suitable approach is a combination of that adopted by
Wang et al. [5], implemented in suites that integrate advanced hydrodynamic fluid solvers and
computer graphic rendering that will upgrade the fluid simulations into visually relatable
scenarios (e.g. see [42]).
4. INTEGRATED WORKFLOW SCHEME FOR VISUALIZING NATURAL
HAZARDS
In the present section, we propose a working scheme (Figure 2) that will allow an integrated
study and visualization of Flash Floods in small and ungauged steep rivers and a more
immersive experience for a better risk perception. This will result in a simplified, reproducible
and scalable prototype that can be embedded in a serious gaming engine and help a non-expert
user make decisions based on intuitive and more precise data than that provided by current
gaming engines, often based on computer graphics rather than real hydraulics. In summary, the
objectives of this working scheme are to:
 Optimize existing fluid simulation models.
 Progress towards real-time simulations of flash floods in steep rivers.
 Incorporate optimized simulations in a serious gaming engine.
 Design realistic and dynamically evolving flood scenarios in steep rivers.
 Implement realistic, hydraulics-based scenarios in a prototype that will be
furtherly executed in a gamifying engine.
The first step is to review the imminent trends on serious gaming (VR/MR) applied to the
study of natural hazards. Also, to exhaustively overview the state-of-the-art on hydrodynamic
and morphological simulation models (i.e. CFD) solving the Reynolds Averaged Navier-
Stokes (RANS) equations, the numerical flood models available today, and the visualization
models (2D and 3D) of flash floods that are currently in use within the scientific community.
In doing so we will be able to i) understand what possibilities and limitations the current
simulation models have, ii) chose and/or combine available models in order to achieve speedier
results, iii) gain knowledge on how to couple available CFD models within a serious gaming
engine. The hereby-presented state-of-the-art (see section 3.2) on visualization of floods results
from a first attempt of completing the first branch of the workflow scheme.
Once familiar with the characteristics, potential and limitations of available hydrodynamic
and morphological models, the most suitable numerical model will be chosen based on testing
and comparison of currently used ones among the scientific community (see section 3.1), such

272 273
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
8
controls the efficient execution of the visualization tool. The visualization scheme should
integrate visualization requests of any type of application oriented on the computation of
natural hazards.
Recent developments in the visualization of floods include numerical computations and
simulations in 3D. For instance, Ghazali and Kamsin [17] used the combination of SPH
(through the GLU3D plug-in) and 3D computer graphics (in this case, Maya) to simulate and
model flash floods on a three-dimensional geo-referenced environment (i.e. LiDAR DEM).
Moreover, they used most of the software’s potential by testing the realistic visualization of
natural hazards in real-time through an Application Programming Interface (API) in Maya.
Later on, Li et al. [32] used OpenGL for flood simulations in 3D by first creating the 3D terrain
and sky background, and later simulating the water flow and its depth overlaid on the terrain.
Parallel, Ye et al. [40] used SPH to model the water flow during a dam break and their
computational output was embedded in a 3D spatio-temporal GIS application where this and
other flood scenarios where dynamically visualized. Additionally, specific layers could be
added to show public infrastructures within the system. Demir and Krajewski [34] juxtaposed
in their research flood analysis, adaptive real-time communication of flooding conditions and
the interactive visualization of results. Nevertheless, the data provided was very complex,
including river conditions, flood maps, forecasting and related information. Cartoon-style
displays of results in an online platform made the data easily accessible and understandable to
the end user.
Most recently, Macchione et al. [37] represented 2D hydraulic simulations within a so-
called 3D virtual reality environment while aiming at presenting a product potentially useful
for hydraulic engineers for risk communication purposes. Their workflow contemplated a
sensible compromise between the inherent complexity of virtual reality and the need to
represent flooding events in 3D environments to improve the interaction with decision-makers
and to engage people with natural hazards. Zhang et al. [36] [41] went a step further and aimed
to improve data visualization, increase simulation speed and allow real-time interaction during
the simulation process. This lead in the development of a 3D flood simulation platform with
VR technology. They analyzed flood processes, simulated the flow field as well as the breach
flood process and contemplated the emergency plan making. Moreover, their digital platform
not only represented the real-time changing process through the “instruction-response” method
and data interpolation, but also combined the virtual visualization of the data with numerical
modelling in a 3D visual form based on modular software design. Although their methods
involved the use of coding techniques, numerical modelling and virtual reality tools (e.g. C++,
FORTRAN, OSG, VPB, osgGIS, intranet, middleware, etc.) altogether, the simulation speed
and the interaction between numerical models had room for improvement regarding the
achieved level of immersion and complexity of interactive functions.
Wang et al. [5] tackled the need for an improvement of flood risk communication and the
real-time flood risk assessment through a combined simulation-visualization approach
somewhat similar to that of Macchione et al. [37] in terms of presenting 2D hydraulic
simulations in an improved 3D environment. They enhanced decision support by including the
analysis of the uncertainties of such model as well as by increasing the computational
efficiency in data assimilation and calculation with the help of algorithms. Their study did not
limit to the presentation of aesthetically attractive scenarios, where the 3D model was rendered
by means of computer graphic-assisted improvement. A gamification process adds value to
their study and is noteworthy. They added interactive elements, such as user-interactive
features e.g. 3D “drag-and drop” icons, design analysis, 3D stereo panorama, storyboards or
online shared view. These elements allow the user of their application to position elements
wherever they want, measure the inundated area, share the experience as a web link or QR code
that displays a 360º rendered panoramic view, follow a tour of static and dynamic specified
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
9
views or even capture their comments of design plans. All these tools permit a greater
immersive experience and a much more effective communication of the presented hazardous
scenarios. The interaction of user-model through virtual reality increases the stakeholders’
implication in planning and decision-making and provides instant feedback on how to increase
the model’s potential and effectiveness.
The most suitable model should be selected based not only on the most advanced
technologies and the attempts to tackle very complex problems. The model choice should be
ruled by what works best in the given context, considering the level of investment needed (i.e.
data and financially speaking), the hydraulic context and the precision needed for decision-
making. Characteristics of numerical models and model-selection criteria are described in
section 3.1. WoWW is aiming at potential real-time simulation-based solutions, where the
output is both physically and visually realistic and enables an accurate flood risk assessment
and its communication to the stakeholders, whom are oftentimes not part of the scientific
community. For such purpose, the most suitable approach is a combination of that adopted by
Wang et al. [5], implemented in suites that integrate advanced hydrodynamic fluid solvers and
computer graphic rendering that will upgrade the fluid simulations into visually relatable
scenarios (e.g. see [42]).
4. INTEGRATED WORKFLOW SCHEME FOR VISUALIZING NATURAL
HAZARDS
In the present section, we propose a working scheme (Figure 2) that will allow an integrated
study and visualization of Flash Floods in small and ungauged steep rivers and a more
immersive experience for a better risk perception. This will result in a simplified, reproducible
and scalable prototype that can be embedded in a serious gaming engine and help a non-expert
user make decisions based on intuitive and more precise data than that provided by current
gaming engines, often based on computer graphics rather than real hydraulics. In summary, the
objectives of this working scheme are to:
 Optimize existing fluid simulation models.
 Progress towards real-time simulations of flash floods in steep rivers.
 Incorporate optimized simulations in a serious gaming engine.
 Design realistic and dynamically evolving flood scenarios in steep rivers.
 Implement realistic, hydraulics-based scenarios in a prototype that will be
furtherly executed in a gamifying engine.
The first step is to review the imminent trends on serious gaming (VR/MR) applied to the
study of natural hazards. Also, to exhaustively overview the state-of-the-art on hydrodynamic
and morphological simulation models (i.e. CFD) solving the Reynolds Averaged Navier-
Stokes (RANS) equations, the numerical flood models available today, and the visualization
models (2D and 3D) of flash floods that are currently in use within the scientific community.
In doing so we will be able to i) understand what possibilities and limitations the current
simulation models have, ii) chose and/or combine available models in order to achieve speedier
results, iii) gain knowledge on how to couple available CFD models within a serious gaming
engine. The hereby-presented state-of-the-art (see section 3.2) on visualization of floods results
from a first attempt of completing the first branch of the workflow scheme.
Once familiar with the characteristics, potential and limitations of available hydrodynamic
and morphological models, the most suitable numerical model will be chosen based on testing
and comparison of currently used ones among the scientific community (see section 3.1), such

274
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
10
as grid-based and particle-based (e.g. Smoothed-Particle Hydrodynamics, SPH) models.
Furthermore, an integrated model based on state-of-the-art will be developed given the new
demands on higher simulation speed without compromising the achieved level of complexity.
Developing sharpened hydraulic modelling permits to focus, simplify and standardize the
simulation methods before mentioned, thus, optimize risk assessment studies, as it will display
the main factors controlling river response to flood events and allow to direct the efforts
towards the parameters that are most relevant to study. Another reason to optimize the fluid
simulations is to reduce the simulation time needed in order to achieve increasingly precise
although not necessarily as accurate scenarios. By doing so, we will be one step closer to real-
time simulations in future prototypes. The need of this simplification will be verified through
statistical comparison (R and RStudio) of simplified to full hydrodynamic fluid simulations.
Figure 2. Approximate workflow diagram pr oposed for the research project. The pink and gray sections
correspond to the link between the output of this investigation and the Human Behaviour and Risk Perception
and Immersive and Interactive Experience of Natural Hazards work packages, respectively. The red dashed
rectangle contains the implementation stage of the resulting model in WoWW.
The model will first be developed and further tested in existing study cases affecting
Norwegian rivers (e.g. Tokke in 2009, Flåm in 2014, Utvik and Innvik in 2017, etc.) within
WoWW’s umbrella. Identifying critical locations along steep watercourses by integrating
Geographical Information Systems (GIS) and existing 2D and 3D hydrodynamic fluid
simulation models (e.g. see Table 1 for examples of trending software) will result in simplified
flood simulations that contribute to a more efficient risk assessment. The reconstructed flood
event will be based on steep mountainous areas, where hydrological data will be input and the
fluid simulation will be run in a 3D geo-referenced environment (i.e. terrain, in most cases a
DEM). Afterwards, the elements will be visualized by texturing and adding materials, as well
as incorporating visualization effects such as rain or sky (i.e. world).
The information needed
in order to perform numerical simulations will be collected and implemented through field and
experimental work, or from environmental sources available in the IoT. The data collected and
modelled will need testing and validation with field observations.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
11
Additionally, the simplified and validated numerical hydrodynamic model will be integrated
in a serious gaming engine (by means of e.g. Unity), as represented in the red dashed rectangle
on the right-hand side of Figure 2. As mentioned above, a prototype for a serious gaming engine
will be created in 3D computer rendering software such as Blender or Maya. Available open
source software will be employed during the research, as far as these have enough functionality
(i.e. efficiency and compatibility with needed add-ons). However, certain adjustments might
be required, as existing fluid solvers might need to be complemented with the development of
additional material descriptions and customized computer algorithms in order to achieve the
optimal fluid simulation.
Furthermore, two to four scenarios, if not more, will be tested and presented in a Virtual
Flood Game (i.e. Serious Gaming Engine), which will allow testing the prototype in selected
watercourses. Certainly, the research could be extended to investigating worldwide study cases,
validating the resulting optimized model in any river that might be interesting and meets the
eligibility criteria (i.e. small steep catchment that has flood risk potential and with available
data of the required resolution for further study and implementation in the Serious Gaming
Engine).
In order to achieve the World of Wild Waters’ goal of communicating the implications of
natural hazards as a phenomenon and helping decision-making by means of reaching an end
user without scientific background in a very relatable approach, the research carried out in this
work package is very much interrelated with creating an immersive experience and will use
risk perception as a basis during the content design, hence, contributing to the creation of a
good Quality of Experience (QoE). For this purpose, it is expected to have a close collaboration
with other work packages of the World of Wild Waters (i.e. WP4: Immersive and Interactive
Experience of natural hazards, and WP5: Human Behaviour and Risk Perception of natural
hazards). The results of this investigation will be used in a serious gaming engine and tested in
selected subjects for the purpose of iteration and improving the communication of natural
hazards and the pursue of a proactive and preventive response in the end user.
The application of both 2D and 3D hydrodynamic models to flood risk assessment in small,
ungauged steep rivers, as well as the visualization of results will be exemplified in the
following section. Their suitability for the scope of this research will also be discussed.
5. RESULTING VISUALIZATION OF HYDRODYNAMIC FLUID SIMULATIONS
USING AN INTEGRATED WORKFLOW SCHEME
For clarification purposes of the workflow scheme proposed in this document (Figure 2), in
this section, we will describe an example applied on a step-wise fashion.
A flooding event is characterized in terms of hydrologic and geographic parametres that are
used as an input in hydrodynamic fluid simulations (Figure 3). These are the initial conditions
and define the geometry and boundary conditions defined in the plan set up in any two- or
three-dimensional numerical model. The computational grid needs to respect singularities
present in the terrain and adapt in function of the level of interest of each section (i.e. denser
mesh, break lines and refinement regions shall be used for a higher precision in the fluid
simulation). To go one step closer to real-time flood simulations, hydrological data resulting
from WoWW’s statistics of extremes and hydrologic work packages will be connected to the
IoT and retrieved by means of algorithms included in the developed model. This previous stage
is not shown in the workflow scheme, as it is resulting from simultaneous research within the
World of Wild Waters project.

274 275
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
10
as grid-based and particle-based (e.g. Smoothed-Particle Hydrodynamics, SPH) models.
Furthermore, an integrated model based on state-of-the-art will be developed given the new
demands on higher simulation speed without compromising the achieved level of complexity.
Developing sharpened hydraulic modelling permits to focus, simplify and standardize the
simulation methods before mentioned, thus, optimize risk assessment studies, as it will display
the main factors controlling river response to flood events and allow to direct the efforts
towards the parameters that are most relevant to study. Another reason to optimize the fluid
simulations is to reduce the simulation time needed in order to achieve increasingly precise
although not necessarily as accurate scenarios. By doing so, we will be one step closer to real-
time simulations in future prototypes. The need of this simplification will be verified through
statistical comparison (R and RStudio) of simplified to full hydrodynamic fluid simulations.
Figure 2. Approximate workflow diagram pr oposed for the research project. The pink and gray sections
correspond to the link between the output of this investigation and the Human Behaviour and Risk Perception
and Immersive and Interactive Experience of Natural Hazards work packages, respectively. The red dashed
rectangle contains the implementation stage of the resulting model in WoWW.
The model will first be developed and further tested in existing study cases affecting
Norwegian rivers (e.g. Tokke in 2009, Flåm in 2014, Utvik and Innvik in 2017, etc.) within
WoWW’s umbrella. Identifying critical locations along steep watercourses by integrating
Geographical Information Systems (GIS) and existing 2D and 3D hydrodynamic fluid
simulation models (e.g. see Table 1 for examples of trending software) will result in simplified
flood simulations that contribute to a more efficient risk assessment. The reconstructed flood
event will be based on steep mountainous areas, where hydrological data will be input and the
fluid simulation will be run in a 3D geo-referenced environment (i.e. terrain, in most cases a
DEM). Afterwards, the elements will be visualized by texturing and adding materials, as well
as incorporating visualization effects such as rain or sky (i.e. world).
The information needed
in order to perform numerical simulations will be collected and implemented through field and
experimental work, or from environmental sources available in the IoT. The data collected and
modelled will need testing and validation with field observations.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
11
Additionally, the simplified and validated numerical hydrodynamic model will be integrated
in a serious gaming engine (by means of e.g. Unity), as represented in the red dashed rectangle
on the right-hand side of Figure 2. As mentioned above, a prototype for a serious gaming engine
will be created in 3D computer rendering software such as Blender or Maya. Available open
source software will be employed during the research, as far as these have enough functionality
(i.e. efficiency and compatibility with needed add-ons). However, certain adjustments might
be required, as existing fluid solvers might need to be complemented with the development of
additional material descriptions and customized computer algorithms in order to achieve the
optimal fluid simulation.
Furthermore, two to four scenarios, if not more, will be tested and presented in a Virtual
Flood Game (i.e. Serious Gaming Engine), which will allow testing the prototype in selected
watercourses. Certainly, the research could be extended to investigating worldwide study cases,
validating the resulting optimized model in any river that might be interesting and meets the
eligibility criteria (i.e. small steep catchment that has flood risk potential and with available
data of the required resolution for further study and implementation in the Serious Gaming
Engine).
In order to achieve the World of Wild Waters’ goal of communicating the implications of
natural hazards as a phenomenon and helping decision-making by means of reaching an end
user without scientific background in a very relatable approach, the research carried out in this
work package is very much interrelated with creating an immersive experience and will use
risk perception as a basis during the content design, hence, contributing to the creation of a
good Quality of Experience (QoE). For this purpose, it is expected to have a close collaboration
with other work packages of the World of Wild Waters (i.e. WP4: Immersive and Interactive
Experience of natural hazards, and WP5: Human Behaviour and Risk Perception of natural
hazards). The results of this investigation will be used in a serious gaming engine and tested in
selected subjects for the purpose of iteration and improving the communication of natural
hazards and the pursue of a proactive and preventive response in the end user.
The application of both 2D and 3D hydrodynamic models to flood risk assessment in small,
ungauged steep rivers, as well as the visualization of results will be exemplified in the
following section. Their suitability for the scope of this research will also be discussed.
5. RESULTING VISUALIZATION OF HYDRODYNAMIC FLUID SIMULATIONS
USING AN INTEGRATED WORKFLOW SCHEME
For clarification purposes of the workflow scheme proposed in this document (Figure 2), in
this section, we will describe an example applied on a step-wise fashion.
A flooding event is characterized in terms of hydrologic and geographic parametres that are
used as an input in hydrodynamic fluid simulations (Figure 3). These are the initial conditions
and define the geometry and boundary conditions defined in the plan set up in any two- or
three-dimensional numerical model. The computational grid needs to respect singularities
present in the terrain and adapt in function of the level of interest of each section (i.e. denser
mesh, break lines and refinement regions shall be used for a higher precision in the fluid
simulation). To go one step closer to real-time flood simulations, hydrological data resulting
from WoWW’s statistics of extremes and hydrologic work packages will be connected to the
IoT and retrieved by means of algorithms included in the developed model. This previous stage
is not shown in the workflow scheme, as it is resulting from simultaneous research within the
World of Wild Waters project.

276
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
12
Figure 3. Documentation of the flood event that affected the municipalities of Utvik and Innvik in 2017, a study
case for this project. Water flow direction is from South to North. See [3]
for further reference on the case.
The most suitable numerical model will be selected based on the state-of-the-art and model
comparison presented in section 3, and implemented on open-source fluid simulation software,
when feasible. Figure 4 shows a 2D hydrodynamic simulation in HEC-RAS, whereas figure 5
presents a 3D overly simplified hydrodynamic simulation in Blender. The two-dimensional
hydrodynamic simulation is solving the Shallow Water Equations (SWE) through a
combination of Finite Difference (for orthogonal mesh sections) and Finite Volume (for local
non-orthogonal sections) methods for unsteady flow (see computational mesh in Figure 4,
right), while the three-dimensional fluid simulation is using the Lattice-Boltzmann free surface
Method (LBM) on a quadrangular mesh (Figure 5, right). For further reference, see the manuals
of HEC-RAS and Blender, respectively.
Figure 4. Left: 2D simulation of Utvik flood event (2017) in HEC-RAS (v.5.0.6), where the water depth is
represented in blue tones (darker tones correspond to larger water depths), and cross-section (red line in a)) of
the river channel and the flooded riverbanks. Scale is 50m. Right: close-up of the computational mesh, where
different cell sizes and shapes (orthogonal and non-orthogonal) are observed due to the use of mesh-refinement
and break lines.
The resulting hydrodynamic simulations will consider real-time and an increased efficiency
will be sought out. The knowledge gathered will result in the development of a model that shall
be replicable; hence, the comparison of several study cases will provide an insight on how to
optimize the fluid simulation speed without compromising on precision. Hydrodynamic fluid
simulations can be connected to the environmental data available in the IoT and optimized with
the help of algorithms embedded in the numerical simulation model [5] [20] [28]. The level of
optimization achieved relative to currently existing full hydrodynamic simulations will be
estimated through statistical analysis.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
13
Figure 5. a) Grid-based three-dimensional fluid simulation in Blender; b) detail of the computational (non-
orthogonal) mesh in the computational domain shown in a).
As noted by Macchione et al. [37] and Wang et al. [5], the visualization of floods must
integrate fluid simulations and 3D computer graphic modelling in a common platform that
provides flexible design features and the monitoring of the scenario at hand. The rendering of
the hydrodynamic simulations, regardless of the spatial dimensions considered, will result in a
visually improved and more realistic scenario that will communicate the results in a very
intuitive manner. Three-dimensional creation suites such as Blender or Maya and its Bifröst
Fluids often include fluid simulation physics/fluid solvers that take into account rather
restricted fluid physics (e.g. fluid type, density, buoyancy, velocity, turbulence, vorticity) as
well as their interaction with other elements (be it terrain or other fluids) and emulate the
physical behavior of the fluid visually. The specifications of such software include the
possibility of customizing textures and materials of the modelled elements by means of node-
based schemes (Figures 6 and 7) with a significant degree of automation that do not increase
the complexity of the modelling procedure and improve notably their visualization.
Figure 6. Design of a node-structured water shader scheme in Unity 3D. The procedure is very alike to that used
in Blender.

276 277
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
12
Figure 3. Documentation of the flood event that affected the municipalities of Utvik and Innvik in 2017, a study
case for this project. Water flow direction is from South to North. See [3]
for further reference on the case.
The most suitable numerical model will be selected based on the state-of-the-art and model
comparison presented in section 3, and implemented on open-source fluid simulation software,
when feasible. Figure 4 shows a 2D hydrodynamic simulation in HEC-RAS, whereas figure 5
presents a 3D overly simplified hydrodynamic simulation in Blender. The two-dimensional
hydrodynamic simulation is solving the Shallow Water Equations (SWE) through a
combination of Finite Difference (for orthogonal mesh sections) and Finite Volume (for local
non-orthogonal sections) methods for unsteady flow (see computational mesh in Figure 4,
right), while the three-dimensional fluid simulation is using the Lattice-Boltzmann free surface
Method (LBM) on a quadrangular mesh (Figure 5, right). For further reference, see the manuals
of HEC-RAS and Blender, respectively.
Figure 4. Left: 2D simulation of Utvik flood event (2017) in HEC-RAS (v.5.0.6), where the water depth is
represented in blue tones (darker tones correspond to larger water depths), and cross-section (red line in a)) of
the river channel and the flooded riverbanks. Scale is 50m. Right: close-up of the computational mesh, where
different cell sizes and shapes (orthogonal and non-orthogonal) are observed due to the use of mesh-refinement
and break lines.
The resulting hydrodynamic simulations will consider real-time and an increased efficiency
will be sought out. The knowledge gathered will result in the development of a model that shall
be replicable; hence, the comparison of several study cases will provide an insight on how to
optimize the fluid simulation speed without compromising on precision. Hydrodynamic fluid
simulations can be connected to the environmental data available in the IoT and optimized with
the help of algorithms embedded in the numerical simulation model [5] [20] [28]. The level of
optimization achieved relative to currently existing full hydrodynamic simulations will be
estimated through statistical analysis.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
13
Figure 5. a) Grid-based three-dimensional fluid simulation in Blender; b) detail of the computational (non-
orthogonal) mesh in the computational domain shown in a).
As noted by Macchione et al. [37] and Wang et al. [5], the visualization of floods must
integrate fluid simulations and 3D computer graphic modelling in a common platform that
provides flexible design features and the monitoring of the scenario at hand. The rendering of
the hydrodynamic simulations, regardless of the spatial dimensions considered, will result in a
visually improved and more realistic scenario that will communicate the results in a very
intuitive manner. Three-dimensional creation suites such as Blender or Maya and its Bifröst
Fluids often include fluid simulation physics/fluid solvers that take into account rather
restricted fluid physics (e.g. fluid type, density, buoyancy, velocity, turbulence, vorticity) as
well as their interaction with other elements (be it terrain or other fluids) and emulate the
physical behavior of the fluid visually. The specifications of such software include the
possibility of customizing textures and materials of the modelled elements by means of node-
based schemes (Figures 6 and 7) with a significant degree of automation that do not increase
the complexity of the modelling procedure and improve notably their visualization.
Figure 6. Design of a node-structured water shader scheme in Unity 3D. The procedure is very alike to that used
in Blender.

278
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
14
Efforts are being focused on integrating more complex fluid solvers into these creation
suites. For instance, the 3D solver for the two-phase incompressible Navier-Stokes equations
NaSt3DGPF was successfully coupled with Maya in a toolkit that enables the user to control
the full fluid simulation within Maya’s interface [42] [43]. The solver uses high-order Finite
Difference discretization methods and the rendering techniques result in realistic CFD
visualizations. Moreover, the limitations of the discretization methods seem to be overcome by
introducing an approximation method in stochastic space, with a high convergence order and
a very small pre-asymptotic error, outperforming methods such as Monte Carlo. The use of
coupled toolkits as such is desirable and is contemplated in the workflow scheme here
proposed.
Figure 7. Hydrodynamic fluid simulation in Blender, with textures and materials on terrain and water.
Once the textures and materials have been incorporated to the hydrodynamic fluid
simulation, visualization effects such as rain, a panoramic 360 degrees sky or buildings will
make the resulting scenario more relatable to the non-expert user, hence, ready to use in further
experiments (such as risk perception and QoE) or for direct decision-making.
6. DISCUSSION
Simulating flood scenarios allows iterating research objectives and potential answers to
these. A quick display of information allows a more effective and competent implementation
of alleviation measures and flood management practices. The comparison of frequently used
fluid simulation models (Table 1, Figure 1), as well as the overview on recent related work on
the visualization of flash floods gives an insight into the selection of an optimal model to
simulate flash floods in small, ungauged steep rivers. The suitability of the model depends on
the problem definition and the scale of analysis. For instance, for visualization purposes and
further research on achieving an outstanding immersive experience, the main goal is to obtain
optimized hydrodynamic fluid simulations and of increasing precision.
Two-dimensional hydrodynamic modelling is the best alternative in terms of compromise
between precision and computational expense. Even though 2D models are not very
comprehensive in incorporating complex hydrodynamics, such as secondary circulation at
bends, this feature is not required for a realistic visualization. Therefore, three-dimensional
modelling is not necessarily favoured to represent flood scenarios in a precise and more
understandable fashion. On the other hand, HEC-RAS solves 2D-SWE and is unable to work
with falls and steps and changing flow regimes, which leads to shocks and instabilities when
modelling flash floods in steep slopes [22]. Shock-capturing algorithms (e.g. Total Variation
Diminishing) are not included in this and generally in most of the open-source modelling
packages, with exception of, e.g. Telemac-Mascaret. Nevertheless, three-dimensional solutions
seem to give the most realistic representation, although the computational cost of 3D models
is still a noteworthy disadvantage.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
15
Advanced solvers are often present in modelling packages without a graphical user interface
(e.g. Telemac-Mascaret for 2D or REEF3D for both 2D and 3D). These tools have the
advantage of solving complex physics at optimal speeds and with very satisfactory reliability.
On the other hand, they need to be coupled to post-processors (e.g. Blue Kenue, Paraview) in
order to visualize results. This means additional work for the researcher, often significantly
time-consuming, in order to analyse, interpret and build up based on the simulated outcome.
The lack of an integrated interface obviously hinders decision-making, as the user needs to be
familiar with multiple platform, generally not user-friendly. Therefore, experts are looking for
ways to integrate the existing tools and even use them in interdisciplinary projects in order to
save cost and make the best of the immense possibilities these tools provide. Research efforts
are currently oriented towards combining advanced two- and three-dimensional fluid solvers
with algorithms that increase their efficiency or improve their precision for visualization
purposes (see for example [27]).
Realistic portrayal of flash floods in small steep rivers requires advanced hydrodynamic
simulations and outstanding rendering of these, which is expected to be computationally
expensive, anyhow. Hadimlioglu and King [28], for example, used a flexible mapping engine
to visualize 3D-simulated water depth, which allowed adaptive resolution and the possibility
to select the representation type dynamically. This resulted in an increased efficiency. The
simulations were particle-based (Lagrangian), which seemed to provide more stable outcomes
than grid-based fluid simulations. Their flexible mapping engine also provided increasing
precision of the water level description by means of quadtrees (which is also used in signal
processing in High Efficiency Video Coding, HEVC) and allowing the system to select the
most suitable representation based on the demanded level of detail. Their model is efficient,
provides good precision and has spatial scalability. However, it does not permit a dynamic
change of parameters over time (e.g. changes in discharge, and no adaptive real-time response),
although this could be potentially implemented. Their study alludes to the benefits of using 2D
visualization over more complex and realistic 3D visualization, as it is easier to use. The
visualization of such model is very promising, despite the fact that the sense of being there (i.e.
good immersive experience) will not be achieved unless using three-dimensional visualization
tools.
It is noteworthy that computer graphics specifications are rarely included in fluid modelling
packages, and the latter generally prioritize computational sources destined to the fluid solver,
leaving the display of results in a secondary position. Although accuracy in terms of
mathematics and physics is crucial when representing a realistic scenario, the data
representation should not require the extensive post-processing that rendering often implies.
When it comes to communication of flood risk to the stakeholders, not only precision is needed,
but also an intuitive display of results. The step-wise workflow scheme hereinto presented
(Figure 2) highlights the need to master diverse software, sometimes open-source and user-
friendly, but most of the times insufficient alone for a reliable and integrated presentation of
the risk assessed. This is inefficient and leads to compatibility issues at times, as well as
unsuited for real-time decision-making. Ideally, a unified visualization platform shall be used
in such case. This means to combine, in a unique platform, advanced hydrodynamic fluid
solvers with algorithms that increase their computational speed, and computer graphic
rendering, achieving solid and optimized hydrodynamic fluid simulations that are immediately
depicted into visually relatable scenarios (e.g. see [5] [42] [43]).
Some visual scenarios are too complex or transitory, and an honest representation of them
faces challenges that might be overcome by their implementation through a holistic approach.
The use of risk perception-based knowledge together with a correctly represented content, in a
well-studied context, will provide the unexperienced user the best Quality of Experience.
Natural hazards concern many different target groups, thus, designing a valuable

278 279
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
14
Efforts are being focused on integrating more complex fluid solvers into these creation
suites. For instance, the 3D solver for the two-phase incompressible Navier-Stokes equations
NaSt3DGPF was successfully coupled with Maya in a toolkit that enables the user to control
the full fluid simulation within Maya’s interface [42] [43]. The solver uses high-order Finite
Difference discretization methods and the rendering techniques result in realistic CFD
visualizations. Moreover, the limitations of the discretization methods seem to be overcome by
introducing an approximation method in stochastic space, with a high convergence order and
a very small pre-asymptotic error, outperforming methods such as Monte Carlo. The use of
coupled toolkits as such is desirable and is contemplated in the workflow scheme here
proposed.
Figure 7. Hydrodynamic fluid simulation in Blender, with textures and materials on terrain and water.
Once the textures and materials have been incorporated to the hydrodynamic fluid
simulation, visualization effects such as rain, a panoramic 360 degrees sky or buildings will
make the resulting scenario more relatable to the non-expert user, hence, ready to use in further
experiments (such as risk perception and QoE) or for direct decision-making.
6. DISCUSSION
Simulating flood scenarios allows iterating research objectives and potential answers to
these. A quick display of information allows a more effective and competent implementation
of alleviation measures and flood management practices. The comparison of frequently used
fluid simulation models (Table 1, Figure 1), as well as the overview on recent related work on
the visualization of flash floods gives an insight into the selection of an optimal model to
simulate flash floods in small, ungauged steep rivers. The suitability of the model depends on
the problem definition and the scale of analysis. For instance, for visualization purposes and
further research on achieving an outstanding immersive experience, the main goal is to obtain
optimized hydrodynamic fluid simulations and of increasing precision.
Two-dimensional hydrodynamic modelling is the best alternative in terms of compromise
between precision and computational expense. Even though 2D models are not very
comprehensive in incorporating complex hydrodynamics, such as secondary circulation at
bends, this feature is not required for a realistic visualization. Therefore, three-dimensional
modelling is not necessarily favoured to represent flood scenarios in a precise and more
understandable fashion. On the other hand, HEC-RAS solves 2D-SWE and is unable to work
with falls and steps and changing flow regimes, which leads to shocks and instabilities when
modelling flash floods in steep slopes [22]. Shock-capturing algorithms (e.g. Total Variation
Diminishing) are not included in this and generally in most of the open-source modelling
packages, with exception of, e.g. Telemac-Mascaret. Nevertheless, three-dimensional solutions
seem to give the most realistic representation, although the computational cost of 3D models
is still a noteworthy disadvantage.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
15
Advanced solvers are often present in modelling packages without a graphical user interface
(e.g. Telemac-Mascaret for 2D or REEF3D for both 2D and 3D). These tools have the
advantage of solving complex physics at optimal speeds and with very satisfactory reliability.
On the other hand, they need to be coupled to post-processors (e.g. Blue Kenue, Paraview) in
order to visualize results. This means additional work for the researcher, often significantly
time-consuming, in order to analyse, interpret and build up based on the simulated outcome.
The lack of an integrated interface obviously hinders decision-making, as the user needs to be
familiar with multiple platform, generally not user-friendly. Therefore, experts are looking for
ways to integrate the existing tools and even use them in interdisciplinary projects in order to
save cost and make the best of the immense possibilities these tools provide. Research efforts
are currently oriented towards combining advanced two- and three-dimensional fluid solvers
with algorithms that increase their efficiency or improve their precision for visualization
purposes (see for example [27]).
Realistic portrayal of flash floods in small steep rivers requires advanced hydrodynamic
simulations and outstanding rendering of these, which is expected to be computationally
expensive, anyhow. Hadimlioglu and King [28], for example, used a flexible mapping engine
to visualize 3D-simulated water depth, which allowed adaptive resolution and the possibility
to select the representation type dynamically. This resulted in an increased efficiency. The
simulations were particle-based (Lagrangian), which seemed to provide more stable outcomes
than grid-based fluid simulations. Their flexible mapping engine also provided increasing
precision of the water level description by means of quadtrees (which is also used in signal
processing in High Efficiency Video Coding, HEVC) and allowing the system to select the
most suitable representation based on the demanded level of detail. Their model is efficient,
provides good precision and has spatial scalability. However, it does not permit a dynamic
change of parameters over time (e.g. changes in discharge, and no adaptive real-time response),
although this could be potentially implemented. Their study alludes to the benefits of using 2D
visualization over more complex and realistic 3D visualization, as it is easier to use. The
visualization of such model is very promising, despite the fact that the sense of being there (i.e.
good immersive experience) will not be achieved unless using three-dimensional visualization
tools.
It is noteworthy that computer graphics specifications are rarely included in fluid modelling
packages, and the latter generally prioritize computational sources destined to the fluid solver,
leaving the display of results in a secondary position. Although accuracy in terms of
mathematics and physics is crucial when representing a realistic scenario, the data
representation should not require the extensive post-processing that rendering often implies.
When it comes to communication of flood risk to the stakeholders, not only precision is needed,
but also an intuitive display of results. The step-wise workflow scheme hereinto presented
(Figure 2) highlights the need to master diverse software, sometimes open-source and user-
friendly, but most of the times insufficient alone for a reliable and integrated presentation of
the risk assessed. This is inefficient and leads to compatibility issues at times, as well as
unsuited for real-time decision-making. Ideally, a unified visualization platform shall be used
in such case. This means to combine, in a unique platform, advanced hydrodynamic fluid
solvers with algorithms that increase their computational speed, and computer graphic
rendering, achieving solid and optimized hydrodynamic fluid simulations that are immediately
depicted into visually relatable scenarios (e.g. see [5] [42] [43]).
Some visual scenarios are too complex or transitory, and an honest representation of them
faces challenges that might be overcome by their implementation through a holistic approach.
The use of risk perception-based knowledge together with a correctly represented content, in a
well-studied context, will provide the unexperienced user the best Quality of Experience.
Natural hazards concern many different target groups, thus, designing a valuable

280
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
16
communication of the assessed risk, as well as successfully guiding and influencing their
choices, requires a level of interaction, usability, immersion, and compromise in the
methodology used that is only comprehended by the concept of QoE. Quality of Experience,
in the context of flash flood risk assessment and communication, evaluates the user experience
on the serious gaming engine within WoWW by contemplating both objective and subjective
metrics. QoE does not only evaluate the content presented to the end user, but is rather
inherently dependent on system-, human- and context-influencing factors. It is, nonetheless,
essential to provide precise and usable content, which has been designed taking into account
the rest of the influencing factors (i.e. that fulfills the user’s needs and expectations and is
provided in an intuitive manner), in order to achieve the best communication of flash flood risk
to the stakeholders and support sound decision-making.
7. CONCLUSIONS
 The target of the hydraulic modelling executed in this research project is to achieve
optimized simulations of flash floods affecting small, ungauged steep rivers that could
be carried out in a realistic, scalable and reproducible prototype in a serious gaming
engine.
 Developing an integrated mathematical model with modern 3D graphics combined in a
user-friendly model engine for satisfactory impact assessment will emphasize the
potential use of visualization technology to enhance understanding of engineering
problems for non-experts on hydraulic engineering.
 Visualization results from the study can be self-sustaining and used in the analysis and
estimation of the consequences of flash floods by a vast array of decision-maker
profiles, such as emergency agencies, government, risk managers, risk consultants, and
for educational purposes, risk communication and awareness, gaming industry, and
users with no-scientific background among others.
 The implementation of optimized fluid simulations in other applications will save time
and cost to those seeking for a compromise between precise and accurate simulation
results, such as researchers, companies and the administration. The data obtained from
simplified simulations could be potentially used as a preliminary orientation in
decision-making, permitting to narrow down the research focus when tackling a very
complex problem, hence, increase efficiency by saving time and cost.
 Soil erosion and sediment transport are important parameters to take into account, but
not enclosed in hydrodynamic simulations. Most of the visualized models available do
not include morphodynamics of small and steep ungauged rivers due to their
complexity to be reproduced successfully even in two-dimensional numerical
simulations. Sediment transport is studied parallel in WoWW and the feasibility to
include morphodynamics in the visualization engine will be discussed in future work.
 Developing virtual scenarios gives the opportunity to replicate and analyze complex
real-life models and the human experiences that accompany them. Virtual reality
engines can provide a laboratory of visual experience where different disciplines can
meet and combine resources and knowledge for a common goal, such as the study and
communication of natural hazards. Using Quality of Experience and Immersive Media
Technology Experiences in the context of natural hazards will improve risk assessment
and its delivery for better decision-making.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
17
ACKNOWLEDGEMENTS
The authors thank two anonymous reviewers for their improving suggestions. This
publication is part of the World of Wild Waters (WoWW) project, which falls under the
umbrella of Norwegian University of Science and Technology (NTNU)’s Digital
Transformation initiative.
REFERENCES
[1] UNISDR - United Nations Office of Disaster Risk Reduction. United Nations Office
for Disaster Risk Reduction: 2018 annual report, (2019) 107pp. + annex I (24pp.).
[2] Moraru, A., Ruiz-Villanueva, V. and Furdada, G. Factors controlling streambank
erosion and channel widening: implications for flood hazard. In: Garcia, C.; Gómez-Pujol, L.;
Morán-Tejeda, E.; Batalla, R. J. (Eds.), Geomorfología del Antropoceno: efectos del cambio
global sobre los procesos geomorfológicos. Palma de Mallorca, Balearic Islands: Universitat
de les Illes Balears, Sociedad Española de Geomorfología. (2018) 247–252, ISBN: 978-84-09-
04850-2
[3] Pavlíček, M. and Bruland, O. Mapping areas exposed to erosion and water forces
during extreme floods in steep terrain. In: European Geosciences Union (EGU), EGU2019-
19167. Available at: https://presentations.copernicus.org/EGU2019-19167_presentation.pdf;
last accessed: 24th May 2019. (2019)
[4] Lujala, P., Lein, H. and Rød, J.K. Climate change, natural hazards, and risk perception:
the role of proximity and personal experience. Local Environment, (2015) 20(4): 489–509.
DOI: 10.1080/13549839.2014.887666
[5] Wang, C., Hou, J., Miller, D., Brown, I. and Jiang, Y. Flood risk management in sponge
cities: The role of integrated simulation and 3D visualization. International Journal of Disaster
Risk Reduction, In Press, (2019) 101139. DOI: 10.1016/j.ijdrr.2019.101139
[6] Henonin, J., Russo, B., Mark, O. and Gourbesville, P. Real-time urban flood
forecasting and modelling – a state of the art. Journal of Hydroinformatics, (2013), 15(3):717–
36. DOI: 10.2166/hydro.2013.132
[7] Hodges, B.R. Hydrodynamical Modeling. In: Elias, S.A. (Ed.). Reference Module in
Earth Systems and Environmental Sciences. Elsevier, (2014), 22pp.
[8] Teng, J., Jakeman, A.J., Vaze, J., Croke, B.F. W., Dutta, D. and Kim, S. Flood
inundation modelling: A review of methods, recent advances and uncertainty analysis.
Environmental Modelling & Software, (2017) 90: 201‒216. DOI:
10.1016/j.envsoft.2017.01.006
[9] Friendly, M., and Denis, D.J. Milestones in the history of thematic cartography,
statistical graphics and data visualization. Statistical Consulting Service and Institute for Social
Research of the University of York (Ontario, Canada), (2006) 71pp.
[10] Muntean, C.I. Raising engagement in e-learning through gamification. In: 6th
International Conference on Virtual Learning (ICVL), Cluj-Napoca, Romania, (2011) 323–
329.
[11] Ebrahimi, T., Foessel, S., Pereira, F. and Schelkens, P. JPEG Pleno: Toward an
Efficient Representation of Visual Reality. IEEE Multimedia, (2016) 23(4):14–20. DOI:
10.1109/MMUL.2016.64
[12] Timmerer, C., Ebrahim, T. and Pereira, F. Toward a New Assessment of Quality.
Computer Journal, (2015) 48(3): 108–110. ISSN: 0010-4620
[13] Perkis, A., Hameed, A. Immersive media experiences - what do we need to move
forward? In: Society of Motion Picture and Television Engineers (SMPTE). Los Angeles,
USA, (2018). 1–12.

280 281
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
16
communication of the assessed risk, as well as successfully guiding and influencing their
choices, requires a level of interaction, usability, immersion, and compromise in the
methodology used that is only comprehended by the concept of QoE. Quality of Experience,
in the context of flash flood risk assessment and communication, evaluates the user experience
on the serious gaming engine within WoWW by contemplating both objective and subjective
metrics. QoE does not only evaluate the content presented to the end user, but is rather
inherently dependent on system-, human- and context-influencing factors. It is, nonetheless,
essential to provide precise and usable content, which has been designed taking into account
the rest of the influencing factors (i.e. that fulfills the user’s needs and expectations and is
provided in an intuitive manner), in order to achieve the best communication of flash flood risk
to the stakeholders and support sound decision-making.
7. CONCLUSIONS
 The target of the hydraulic modelling executed in this research project is to achieve
optimized simulations of flash floods affecting small, ungauged steep rivers that could
be carried out in a realistic, scalable and reproducible prototype in a serious gaming
engine.
 Developing an integrated mathematical model with modern 3D graphics combined in a
user-friendly model engine for satisfactory impact assessment will emphasize the
potential use of visualization technology to enhance understanding of engineering
problems for non-experts on hydraulic engineering.
 Visualization results from the study can be self-sustaining and used in the analysis and
estimation of the consequences of flash floods by a vast array of decision-maker
profiles, such as emergency agencies, government, risk managers, risk consultants, and
for educational purposes, risk communication and awareness, gaming industry, and
users with no-scientific background among others.
 The implementation of optimized fluid simulations in other applications will save time
and cost to those seeking for a compromise between precise and accurate simulation
results, such as researchers, companies and the administration. The data obtained from
simplified simulations could be potentially used as a preliminary orientation in
decision-making, permitting to narrow down the research focus when tackling a very
complex problem, hence, increase efficiency by saving time and cost.
 Soil erosion and sediment transport are important parameters to take into account, but
not enclosed in hydrodynamic simulations. Most of the visualized models available do
not include morphodynamics of small and steep ungauged rivers due to their
complexity to be reproduced successfully even in two-dimensional numerical
simulations. Sediment transport is studied parallel in WoWW and the feasibility to
include morphodynamics in the visualization engine will be discussed in future work.
 Developing virtual scenarios gives the opportunity to replicate and analyze complex
real-life models and the human experiences that accompany them. Virtual reality
engines can provide a laboratory of visual experience where different disciplines can
meet and combine resources and knowledge for a common goal, such as the study and
communication of natural hazards. Using Quality of Experience and Immersive Media
Technology Experiences in the context of natural hazards will improve risk assessment
and its delivery for better decision-making.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
17
ACKNOWLEDGEMENTS
The authors thank two anonymous reviewers for their improving suggestions. This
publication is part of the World of Wild Waters (WoWW) project, which falls under the
umbrella of Norwegian University of Science and Technology (NTNU)’s Digital
Transformation initiative.
REFERENCES
[1] UNISDR - United Nations Office of Disaster Risk Reduction. United Nations Office
for Disaster Risk Reduction: 2018 annual report, (2019) 107pp. + annex I (24pp.).
[2] Moraru, A., Ruiz-Villanueva, V. and Furdada, G. Factors controlling streambank
erosion and channel widening: implications for flood hazard. In: Garcia, C.; Gómez-Pujol, L.;
Morán-Tejeda, E.; Batalla, R. J. (Eds.), Geomorfología del Antropoceno: efectos del cambio
global sobre los procesos geomorfológicos. Palma de Mallorca, Balearic Islands: Universitat
de les Illes Balears, Sociedad Española de Geomorfología. (2018) 247–252, ISBN: 978-84-09-
04850-2
[3] Pavlíček, M. and Bruland, O. Mapping areas exposed to erosion and water forces
during extreme floods in steep terrain. In: European Geosciences Union (EGU), EGU2019-
19167. Available at: https://presentations.copernicus.org/EGU2019-19167_presentation.pdf;
last accessed: 24th May 2019. (2019)
[4] Lujala, P., Lein, H. and Rød, J.K. Climate change, natural hazards, and risk perception:
the role of proximity and personal experience. Local Environment, (2015) 20(4): 489–509.
DOI: 10.1080/13549839.2014.887666
[5] Wang, C., Hou, J., Miller, D., Brown, I. and Jiang, Y. Flood risk management in sponge
cities: The role of integrated simulation and 3D visualization. International Journal of Disaster
Risk Reduction, In Press, (2019) 101139. DOI: 10.1016/j.ijdrr.2019.101139
[6] Henonin, J., Russo, B., Mark, O. and Gourbesville, P. Real-time urban flood
forecasting and modelling – a state of the art. Journal of Hydroinformatics, (2013), 15(3):717–
36. DOI: 10.2166/hydro.2013.132
[7] Hodges, B.R. Hydrodynamical Modeling. In: Elias, S.A. (Ed.). Reference Module in
Earth Systems and Environmental Sciences. Elsevier, (2014), 22pp.
[8] Teng, J., Jakeman, A.J., Vaze, J., Croke, B.F. W., Dutta, D. and Kim, S. Flood
inundation modelling: A review of methods, recent advances and uncertainty analysis.
Environmental Modelling & Software, (2017) 90: 201‒216. DOI:
10.1016/j.envsoft.2017.01.006
[9] Friendly, M., and Denis, D.J. Milestones in the history of thematic cartography,
statistical graphics and data visualization. Statistical Consulting Service and Institute for Social
Research of the University of York (Ontario, Canada), (2006) 71pp.
[10] Muntean, C.I. Raising engagement in e-learning through gamification. In: 6th
International Conference on Virtual Learning (ICVL), Cluj-Napoca, Romania, (2011) 323–
329.
[11] Ebrahimi, T., Foessel, S., Pereira, F. and Schelkens, P. JPEG Pleno: Toward an
Efficient Representation of Visual Reality. IEEE Multimedia, (2016) 23(4):14–20. DOI:
10.1109/MMUL.2016.64
[12] Timmerer, C., Ebrahim, T. and Pereira, F. Toward a New Assessment of Quality.
Computer Journal, (2015) 48(3): 108–110. ISSN: 0010-4620
[13] Perkis, A., Hameed, A. Immersive media experiences - what do we need to move
forward? In: Society of Motion Picture and Television Engineers (SMPTE). Los Angeles,
USA, (2018). 1–12.

282
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
18
[14] Qualinet (European Network on Quality of Experience in Multimedia Systems and
Services). Qualinet white paper on definitions of quality of experience, Dagstuhl seminar
12181, (2012) 24pp.
[15] Perkis, A. A QoE cross layer approach to model media experiences. IEEE COMSOC
MMTC E-Letter, (March 2013) 8(2): 6‒9.
[16] Zhang C, Perkis A, Arndt S. Spatial immersion versus emotional immersion, which
is more immersive? In: 9th International Conference on Quality of Multimedia Experience
(QoMEX). Erfurt, Germany: IEEE, (2017) 1–6. DOI: 10.1109/QoMEX.2017.7965655
[17] Ghazali, J. and Kamsin, A. A real time simulation and modeling of flood hazard. In:
12th WSEAS International Conference on Systems, Heraklion, Greece, (2008) 438‒443. ISBN:
978-960-6766-83-1
[18] Wang, L. and Cheng, Q. Web-based hydrological modeling system for flood
forecasting and risk mapping. Geoinformatics 2008 and Joint Conference on GIS and Built
Environment: Monitoring and Assessment of Natural Resources and Environments, (2008)
7145(November 2008): 71450A. DOI: 10.1117/12.812986
[19] Al-Sabhan, W. Designing a Human-Centred, Mobile Interface to Support Real-time
Flood Forecasting and Warning System. PhD Thesis, University of Brunel (UK), (2016) 419pp.
[20] Krajewski, W.F., Ceynar, D., Demir, I., Goska, R., Kruger, A., Langel, C., Mantilla,
R., Niemeier, J., Quintero, F., Seo, B.C., Small, S.J., Weber, L.J. and Young, N.C. Real-time
flood forecasting and information system for the state of Iowa. Bulletin of the American
Meteorological Society, (2017) 98(3): 539–554. DOI: 10.1175/BAMS-D-15-00243.1
[21] Néelz, S. and Pender, G. Benchmarking the Latest Generation of 2D Hydraulic
Modelling Packages. Environment Agency (Bristol, UK), (2013) 182pp.
[22] Brunner, G.W. HEC-RAS 5.0 Hydraulic Reference Manual. (2016) 538pp. Last
access: 10th May 2019; retrieved from: www.hec.usace.army.mil/software/hec-
ras/documentation.aspx
[23] Betsholtz, A. and Nordlöf, B. Potentials and limitations of 1D-2D-coupled 1D-2D
flood modelling in HEC-RAS - A case study on Höje river. MSc Thesis, University of Lund,
(2017) 128pp.
[24] Gómez, M. and Martínez, E. 1D, 2D, and 3D Modeling of a PAC-UPC Laboratory
Canal Bend. In: SimHydro: Modelling of rapid transitory flows, Sophia Antipolis, (2014) 12pp.
DOI: 10.1007/978-981-287-615-7_29
[25] Gharbi, M., Soualmia, A., Dartus, D. and Masbernat, L. Comparison of 1D and 2D
hydraulic models for floods simulation on the medjerda riverin tunisia. Journal of Materials
and Environmental Science, (2016) 7(8): 3017‒3026. ISSN: 2028-2508
[26] Glock, K., Tritthart, M., Habersack, H. and Hauer, C. Comparison of hydrodynamics
simulated by 1D, 2D and 3D models focusing on bed shear stresses. Water (Switzerland),
(2019) 11(226): 19pp. DOI: 10.3390/w11020226
[27] Annis, A., Gonzalez-Ramirez, N., Nardi, F., Castelli, F. Integrating a 2D Hydraulic
Model and GIS Algorithms into a Data Assimilation Framework for Real Time Flood
Forecasting and Mapping. In: La Loggia, G., Freni, G., Puleo, V., De Marchis, M. (Eds.), 13th
International Conference on Hydroinformatics (HIC). Palermo, Italy: EPiC Series in
Engineering, (2018) 36–44. DOI: 10.29007/29nd
[28] Hadimlioglu, I.A. and King, S.A. Visualization of Flooding Using Adaptive Spatial
Resolution. ISPRS International Journal of Geo-Information, (2019) 8(5): 204. DOI:
10.3390/ijgi8050204
[29] Zinke, P. Modelling of flow and levee depositions in a freshwater delta with natural
vegetation. PhD Thesis, Norwegian University of Science and Technology (NTNU), (2011)
309pp.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
19
[30] Lee, J.H.W., Cheung, V., Kuang, C.P., Choi, D.K. W., Wang, W.P., Tu, C.H., Chan,
B. and Choi, Y.K. VISJET & VISFLOOD: Software for Environmental Hydraulic Modeling
and Visualization. In: S. H. Winoto (Ed.), 7th Asian symposium on visualization, Singapore:
Springer, (2003) conf24a155, 1–6.
[31] Pajorová, E., Hluchý, L., Halada, L. and Slížik, P. 3D visualization tool for Virtual
models of natural disasters. In: Aravossis, K., Brebbia, C.A., Gomez, N. (Eds.), Environmental
Economics and Investment Assessment II WIT Press, (2008) 105–114. DOI:
10.2495/EEIA080111
[32] Li, X., Wan, W., Li, L., Zhang, X., Gan, C. and Yu, X. Realization of flood simulation
visualization based on OpenGL. In: 2012 International Conference on Audio, Language and
Image Processing (IEEE), (2012a) 1151‒1154. DOI: 10.1109/ICALIP.2012.6376790
[33] Li, Y., Pan, L., Liu, T. and Wei, C. Three-dimensional GIS based dynamic
visualization simulation system for flood routing. In: Proceedings of 2nd International
Conference on Computer Science and Network Technology (ICCSNT), (2012b) 1409–1412.
DOI: 10.1109/ICCSNT.2012.6526184
[34] Demir, I. and Krajewski, W.F. Towards an integrated Flood Information System:
Centralized data access, analysis, and visualization. Environmental Modelling and Software,
(2013) 50: 77–84. DOI: 10.1016/j.envsoft.2013.08.009
[35] Haynes, P.S. and Lange, E. In-situ flood visualisation using mobile AR. Proceedings
of the IEEE Symposium on 3D User Interfaces (3DUI), (2016) 243–244. DOI:
10.1109/3DUI.2016.7460061
[36] Zhang, S.H., Yuan, R. and Zhang, T.X. Development and Application of a Three-
Dimensional Flood Simulation Platform. In: 11th International Symposium on Ecohydraulics,
Melbourne, Australia, (2016) 8pp.
[37] Macchione, F., Costabile, P., Costanzo, C. and De Santis, R. Moving to 3-D flood
hazard maps for enhancing risk communication. Environmental Modelling and Software,
(2019) 111(January 2019): 510–522. DOI: 10.1016/j.envsoft.2018.11.005
[38] Gissler, C., Peer, A., Band, S., Bender, J. and Teschner, M. Interlinked SPH Pressure
Solvers for Strong Fluid-Rigid Coupling. ACM Transitions on Graphics, (2019) 38(1):1–13.
DOI: 10.1145/3284980
[39] Suhr J. Basic Principles of Sound Decisionmaking [Internet]. Utah, USA; 1981.
Available from: http://leanconstruction.org/media/docs/deliveryGuide/Appendix13b.pdf
[40] Ye, F., Wang, H., Ouyang, S., Tang, X., Li, Z and Prakash, M. Spatio-temporal
analysis and visualization using SPH for dam-break and flood disasters in a GIS environment.
In: 2012 International Symposium on Geomatics for Integrated Water Resource Management
(IEEE), (2012) 1–6. DOI: 10.1109/GIWRM.2012.6349636
[41] Zhang, S.H., Xia, Z.X. and Wang, T.W. A real-time interactive simulation framework
for watershed decision making using numerical models and virtual environment. Journal of
Hydrology, (2013) 493: 95–104. DOI: 10.1016/j.jhydrol.2013.04.030
[42] Zaspel, P. and Griebel, M. Photorealistic visualization and fluid animation: Coupling
of Maya with a two-phase Navier-Stokes fluid solver. Computing and Visualization in Science
(2011), 14(8): 371–83. DOI: 10.1007/s00791-013-0188-1
[43] Griebel, M., Rieger, C. and Zaspel, P. Kernel-based stochastic collocation for the
random two-phase Navier-Stokes equations. In Press, (2019) 1–21. ArX iv ID: 1810.11270, last
access: 8th August 2019.

282 283
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
18
[14] Qualinet (European Network on Quality of Experience in Multimedia Systems and
Services). Qualinet white paper on definitions of quality of experience, Dagstuhl seminar
12181, (2012) 24pp.
[15] Perkis, A. A QoE cross layer approach to model media experiences. IEEE COMSOC
MMTC E-Letter, (March 2013) 8(2): 6‒9.
[16] Zhang C, Perkis A, Arndt S. Spatial immersion versus emotional immersion, which
is more immersive? In: 9th International Conference on Quality of Multimedia Experience
(QoMEX). Erfurt, Germany: IEEE, (2017) 1–6. DOI: 10.1109/QoMEX.2017.7965655
[17] Ghazali, J. and Kamsin, A. A real time simulation and modeling of flood hazard. In:
12th WSEAS International Conference on Systems, Heraklion, Greece, (2008) 438‒443. ISBN:
978-960-6766-83-1
[18] Wang, L. and Cheng, Q. Web-based hydrological modeling system for flood
forecasting and risk mapping. Geoinformatics 2008 and Joint Conference on GIS and Built
Environment: Monitoring and Assessment of Natural Resources and Environments, (2008)
7145(November 2008): 71450A. DOI: 10.1117/12.812986
[19] Al-Sabhan, W. Designing a Human-Centred, Mobile Interface to Support Real-time
Flood Forecasting and Warning System. PhD Thesis, University of Brunel (UK), (2016) 419pp.
[20] Krajewski, W.F., Ceynar, D., Demir, I., Goska, R., Kruger, A., Langel, C., Mantilla,
R., Niemeier, J., Quintero, F., Seo, B.C., Small, S.J., Weber, L.J. and Young, N.C. Real-time
flood forecasting and information system for the state of Iowa. Bulletin of the American
Meteorological Society, (2017) 98(3): 539–554. DOI: 10.1175/BAMS-D-15-00243.1
[21] Néelz, S. and Pender, G. Benchmarking the Latest Generation of 2D Hydraulic
Modelling Packages. Environment Agency (Bristol, UK), (2013) 182pp.
[22] Brunner, G.W. HEC-RAS 5.0 Hydraulic Reference Manual. (2016) 538pp. Last
access: 10th May 2019; retrieved from: www.hec.usace.army.mil/software/hec-
ras/documentation.aspx
[23] Betsholtz, A. and Nordlöf, B. Potentials and limitations of 1D-2D-coupled 1D-2D
flood modelling in HEC-RAS - A case study on Höje river. MSc Thesis, University of Lund,
(2017) 128pp.
[24] Gómez, M. and Martínez, E. 1D, 2D, and 3D Modeling of a PAC-UPC Laboratory
Canal Bend. In: SimHydro: Modelling of rapid transitory flows, Sophia Antipolis, (2014) 12pp.
DOI: 10.1007/978-981-287-615-7_29
[25] Gharbi, M., Soualmia, A., Dartus, D. and Masbernat, L. Comparison of 1D and 2D
hydraulic models for floods simulation on the medjerda riverin tunisia. Journal of Materials
and Environmental Science, (2016) 7(8): 3017‒3026. ISSN: 2028-2508
[26] Glock, K., Tritthart, M., Habersack, H. and Hauer, C. Comparison of hydrodynamics
simulated by 1D, 2D and 3D models focusing on bed shear stresses. Water (Switzerland),
(2019) 11(226): 19pp. DOI: 10.3390/w11020226
[27] Annis, A., Gonzalez-Ramirez, N., Nardi, F., Castelli, F. Integrating a 2D Hydraulic
Model and GIS Algorithms into a Data Assimilation Framework for Real Time Flood
Forecasting and Mapping. In: La Loggia, G., Freni, G., Puleo, V., De Marchis, M. (Eds.), 13th
International Conference on Hydroinformatics (HIC). Palermo, Italy: EPiC Series in
Engineering, (2018) 36–44. DOI: 10.29007/29nd
[28] Hadimlioglu, I.A. and King, S.A. Visualization of Flooding Using Adaptive Spatial
Resolution. ISPRS International Journal of Geo-Information, (2019) 8(5): 204. DOI:
10.3390/ijgi8050204
[29] Zinke, P. Modelling of flow and levee depositions in a freshwater delta with natural
vegetation. PhD Thesis, Norwegian University of Science and Technology (NTNU), (2011)
309pp.
Adina Moraru, Oddbjørn Bruland, Andrew Perkis and Nils Rüther
19
[30] Lee, J.H.W., Cheung, V., Kuang, C.P., Choi, D.K. W., Wang, W.P., Tu, C.H., Chan,
B. and Choi, Y.K. VISJET & VISFLOOD: Software for Environmental Hydraulic Modeling
and Visualization. In: S. H. Winoto (Ed.), 7th Asian symposium on visualization, Singapore:
Springer, (2003) conf24a155, 1–6.
[31] Pajorová, E., Hluchý, L., Halada, L. and Slížik, P. 3D visualization tool for Virtual
models of natural disasters. In: Aravossis, K., Brebbia, C.A., Gomez, N. (Eds.), Environmental
Economics and Investment Assessment II WIT Press, (2008) 105–114. DOI:
10.2495/EEIA080111
[32] Li, X., Wan, W., Li, L., Zhang, X., Gan, C. and Yu, X. Realization of flood simulation
visualization based on OpenGL. In: 2012 International Conference on Audio, Language and
Image Processing (IEEE), (2012a) 1151‒1154. DOI: 10.1109/ICALIP.2012.6376790
[33] Li, Y., Pan, L., Liu, T. and Wei, C. Three-dimensional GIS based dynamic
visualization simulation system for flood routing. In: Proceedings of 2nd International
Conference on Computer Science and Network Technology (ICCSNT), (2012b) 1409–1412.
DOI: 10.1109/ICCSNT.2012.6526184
[34] Demir, I. and Krajewski, W.F. Towards an integrated Flood Information System:
Centralized data access, analysis, and visualization. Environmental Modelling and Software,
(2013) 50: 77–84. DOI: 10.1016/j.envsoft.2013.08.009
[35] Haynes, P.S. and Lange, E. In-situ flood visualisation using mobile AR. Proceedings
of the IEEE Symposium on 3D User Interfaces (3DUI), (2016) 243–244. DOI:
10.1109/3DUI.2016.7460061
[36] Zhang, S.H., Yuan, R. and Zhang, T.X. Development and Application of a Three-
Dimensional Flood Simulation Platform. In: 11th International Symposium on Ecohydraulics,
Melbourne, Australia, (2016) 8pp.
[37] Macchione, F., Costabile, P., Costanzo, C. and De Santis, R. Moving to 3-D flood
hazard maps for enhancing risk communication. Environmental Modelling and Software,
(2019) 111(January 2019): 510–522. DOI: 10.1016/j.envsoft.2018.11.005
[38] Gissler, C., Peer, A., Band, S., Bender, J. and Teschner, M. Interlinked SPH Pressure
Solvers for Strong Fluid-Rigid Coupling. ACM Transitions on Graphics, (2019) 38(1):1–13.
DOI: 10.1145/3284980
[39] Suhr J. Basic Principles of Sound Decisionmaking [Internet]. Utah, USA; 1981.
Available from: http://leanconstruction.org/media/docs/deliveryGuide/Appendix13b.pdf
[40] Ye, F., Wang, H., Ouyang, S., Tang, X., Li, Z and Prakash, M. Spatio-temporal
analysis and visualization using SPH for dam-break and flood disasters in a GIS environment.
In: 2012 International Symposium on Geomatics for Integrated Water Resource Management
(IEEE), (2012) 1–6. DOI: 10.1109/GIWRM.2012.6349636
[41] Zhang, S.H., Xia, Z.X. and Wang, T.W. A real-time interactive simulation framework
for watershed decision making using numerical models and virtual environment. Journal of
Hydrology, (2013) 493: 95–104. DOI: 10.1016/j.jhydrol.2013.04.030
[42] Zaspel, P. and Griebel, M. Photorealistic visualization and fluid animation: Coupling
of Maya with a two-phase Navier-Stokes fluid solver. Computing and Visualization in Science
(2011), 14(8): 371–83. DOI: 10.1007/s00791-013-0188-1
[43] Griebel, M., Rieger, C. and Zaspel, P. Kernel-based stochastic collocation for the
random two-phase Navier-Stokes equations. In Press, (2019) 1–21. ArX iv ID: 1810.11270, last
access: 8th August 2019.