Content uploaded by Pedro Poseiro
Author content
All content in this area was uploaded by Pedro Poseiro on Oct 29, 2014
Content may be subject to copyright.
HIDRALERTA Project – A Flood Forecast and Alert System in
Coastal and Port Areas
C.J.E.M. Fortes1, M.T. Reis1, P. Poseiro1, R. Capitão1, J.A. Santos2, L.V. Pinheiro1, J. Craveiro1,
A. Rodrigues3, A. Sabino3, S. F. Silva3, J.C. Ferreira3, P.D. Raposeiro3, C. Silva3, M.C.
Rodrigues4,6, A. Simões5, E.B. Azevedo6, F. Reis6
1 LNEC – Laboratório Nacional de Engenharia Civil, jfortes@lnec.pt; treis@lnec.pt; pposeiro@lnec.pt;
rcapitao@lnec.pt; lpinheiro@lnec.pt; jcraveiro@lnec.pt
2 ISEL – Instituto Superior de Engenharia de Lisboa, jasantos@dec.isel.pt
3 Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia, a.rodrigues@fct.unl.pt;
andre.sabino@gmail.com; s.silva@campus.fct.unl.pt; jcrf@fct.unl.pt; praposeiro@fct.unl.pt; cpsilva@fcsh.unl.pt
4 Azorina – Sociedade de Gestão Ambiental e Conservação da Natureza, SA, maria.cs.rodrigues2@azores.gov.pt
5 Universidade dos Açores, LAMTEC, anabela@uac.pt
6 Centro do Clima, Meteorologia e Mudanças Globais da Universidade dos Açores, C_CMMG,
mcrodrigues@uac.pt; edubrito@mail.angra.uac.pt; vieirareis@uac.pt
Abstract: This paper describes the recent advances on the research project HIDRALERTA - "Flood
Forecast and Early Warning System in Coastal and Port Areas", whose main objective is developing a
system for forecasting, warning and assessment of risks associated with wave overtopping and flooding
in coastal and port areas, supported by measurements/predictions of waves and water levels in these
areas. The case study here presented is the port and bay of Praia da Vitória, at the Terceira Island,
Azores. Methodologies and accomplishments of the HIDRALERTA project achieved so far, with the
focus on its more relevant results, are herein described.
Keywords: warning system, wave conditions, wave overtopping, flooding, risk assessment
Introduction
The wave overtopping/flooding evaluation in coastal and port areas is a very
important issue in Portugal. In fact, due to the length of the Portuguese coastline, the
severity of the sea conditions and the relevance of the coastal zone for socio-
economic activities, it is important to study wave induced risks and, in particular,
overtopping and flooding due to wave action. In Portugal, emergency situations
caused by adverse sea conditions are frequent, endangering the safety of people and
goods, with negative impacts for the society, the economy and the environment.
For these reasons, it is imperative to empower the national authorities with an
adequate forecast and early warning system that allows the identification of
emergency situations and enables the selection by the authorities of measures to avoid
loss of lives and minimize damages.
Moreover, the forecast and early warning system also act as a long-term
management tool since it can simulate the response to future scenarios related to
climate changes, such as the increase of the mean sea level and/or of the storm
severity, which will increase the probability of coastal flooding. Such a system it will
contribute to complying with the directive 2007/60/CE from the EU of 2007-10-23 in
what concerns the preparation by the member states of flooding risk maps before
December 22th, 2013 (Chap.III-Art.6-8) and of risk management plans, including
forecast, alert and warning systems, before 2015-12-22 (Chap.III-Art.7-3). This
framework does justify the HIDRALERTA project.
The work within this project consists in:
1. Creating a user-friendly tool, which follows Zózimo et al. (2008), that allows:
Construction of risk maps that are decision-support tools for the authorities. These
maps are constructed by considering either long-term series of sea-wave forecasts
or predefined scenarios associated with climate change and/or extreme events;
Real-time evaluation of overtopping/flooding emergencies, issuing warning
messages to the authorities when the safety of people, goods or activities in these
areas is likely to be at stake;
2. Development of a prototype system, HIDRALERTA system, in a WebGIS
environment, for the port and bay of Praia da Vitória, Terceira Island, Azores,
and for São João da Caparica beach, Costa da Caparica, Lisbon.
In the next sections, the HIDRALERTA system modules are described, as well as
recent developments which have been made therein. One application case (port and
bay of Praia da Vitória) is also presented.
HIDRALERTA Methodology
The HIDRALERTA system encompasses four main modules, see Figure 1 and
Poseiro et al. (2013a), namely I – Wave Characterization; II – Wave run-up /
Overtopping; III - Risk Assessment and IV - Warning System. The system is
developed in Python language and it is implemented in a WebGIS platform.
Figure 1. Schematic representation of the HIDRALERTA system
Wave characterization module
The objective of this module is to determine 180-hours-in-advance sea wave
characteristics, to be used by the warning system, and also to provide adequate
handling of long term time series of wave characteristics for the definition of risk
maps.
For the warning part, the HIDRALERTA system is able to determine sea wave
characteristics inside a given port or at a certain location on the coast by using
numerical models for sea wave propagation. The use of one or more numerical
models for the propagation depends on the study region characteristics and on the
phenomena involved in the propagation. The numerical models used by the system
are the nonlinear spectral model of generation and propagation of waves in coastal
areas, SWAN (Booij et al., 1999) and the linear mild-slope model of wave
propagation in port areas, DREAMS (Fortes, 2002) which will be replaced later by
the nonlinear Boussinesq model BOUSS-WMH, Pinheiro et al. (2011).
The implemented procedure is as follows: SWAN is forced by the 180-hours-in-
advance sea wave characteristics estimated by WAVEWATCH III (Tolman, 1999), a
numerical model for sea wave prediction at a regional level, and by the wind fields,
also at a regional level, as well as tide levels, obtained respectively from NAVGEM
(Whitcomb, 2012), and XTide (Flater, 1998) models. XTide provides astronomical
tide levels, so it was considered a constant storm surge to obtain te final tide level.
WAVEWATCH III estimated offshore wave conditions and the wind field, obtained
from NAVGEM, are both provided through The Fleet Numerical Meteorology and
Oceanography Center (FNMOC). FNMOC delivers forecast wave data from
WAVEWATCH III up to 180 hours and historic data since September 2003 until the
present, with a 1º resolution. It also delivers wind data from NAVGEM up to 180
hours and historic data since January 2004 until the present, with a 0.5º resolution.
Then, these values are transferred to the coast using the SWAN model and finally into
the harbour basins (where the reflection effects are important) with the DREAMS
model.
For the risk assessment, the system uses the long time series estimated either by the
WAVEWATCH III model or, preferably, by in-situ measurements, although the latter
is not common since only a few locations exist with wave buoys and even when they
exist, there usually are important gaps in the data records that precludes its use in
practice. In both cases, the wave characteristics are to be transferred by SWAN and
DREAMS (or BOUSS WHF) models, following the above procedure.
Recent developments in this module consisted of: a) Establishment of an automatic
download procedure (Poseiro, 2013) for the 180-hours-in-advance sea wave
predictions of WWIII (FNMOC), with tides being provided by the XTide model and
winds being provided by NAVGEM model (FNMOC); b) Coupling of WWIII,
SWAN and DREAMS models and validation of each module, as well as its coupling
through comparison with in-situ measurements. To this effect, campaigns for
measuring waves in Costa da Caparica, like the one that took place on October 29th,
2012, and campaigns for construction of beach profiles, like the one that took place on
April 29th and 30th, 2014, are to be used to feed the system; c) Automation of
processes for creating layouts generated by the numerical models as well as creation
of files with values obtained for each parameter at each point; d) Development of an
ARTMAP neural network with fuzzy logic (Santos, 2013) to be tested for determining
the wave propagation conditions inside ports which can substitute the use of wave
propagation numerical models in some cases.
Wave run-up/overtopping module
The objective of this module is the determination of wave run-up/overtopping on
coastal and port areas. In the HIDRALERTA system, wave run-up/overtopping
determination follows two different approaches, in case of port or coastal areas.
For port areas, the tool NN_OVERTOPPING2 (Coeveld et al., 2005) is employed
and it is based on neural network modelling. This tool was developed as part of the
CLASH European project (Crest Level Assessment of coastal Structures by full scale
monitoring, neural network prediction and Hazard analysis on permissible wave
overtopping) to predict Froude-scaled mean wave overtopping discharges, q, and the
associated confidence intervals for a wide range of coastal structure types (such as
dikes, rubble-mound breakwaters and caisson structures). The input needed to run
NN_OVERTOPPING2 includes the wave/water level conditions in front of each
structure and its geometrical characterization. For coastal areas (simple beaches or
beaches with coastal defence structures), empirical formulas are applied to evaluate
wave run-up/overtopping. The formulas applied for wave run-up evaluation are
presented at Neves et al. (2013). The flood levels are obtained by adding the wave
run-up estimation to the astronomical tide level and the storm surge.
Recent advances consisted in developing Fortran programs to implement the
empirical formulas in order to couple both these formulas and the neuronal network
tool with numerical models as well as to automatically display the results. The new
neural network tool, of the ARTMAP type with fuzzy logic, to calculate the
overtopping of maritime structures (Santos et al., 2013) is being further developed to
become an alternative to the NN_OVERTOPPING2 tool, especially for small mean
overtopping discharges. In addition, new formulations have been developed for
estimating run-up and overtopping at seawalls built on land and in very shallow water
(Mase et al., 2013), which will also be incorporated in the system.
Risk assessment module
The objective of this module is the evaluation of the overtopping/flooding risk. Risk
assessment is based upon the concept of risk level given by Risk level = Probability
level x Consequences level, this is, risk level is the product of the probability level
associated with the exceedance of a pre-set threshold for the mean overtopping
discharge per unit length of the structure crest (or for the flood level) by the
consequences level associated to such exceedance (Raposeiro, et al., 2010).
For assessing the risk of overtopping or flooding in a given port or coastal area, the
procedure followed is: a) to evaluate the nature of the activities developed in the area
sheltered by the structure (or beach) and the impact of the overtopping/flooding on the
safety of people and infrastructure; b) to establish the thresholds for the allowable
mean overtopping discharge (or for the flood level) for each type of structure (or
beach) and activity; c) to assess the probability level of occurrence of these discharges
(or flood levels); d) to assess the consequences level; e) to assess the risk level.
The thresholds for the allowable mean overtopping discharge are set based on
existing recommendations (Pullen et al., 2007), which take into account the nature of
the activities in the area sheltered by the structure (or beach), the characteristics of the
overtopped structure and the need to ensure the safety of people and infrastructure
located in this area. The thresholds for the flood levels are related to the topography of
the area (crest levels of the beach/dune). The probability levels associated with the
exceedance of a pre-set threshold for the mean overtopping discharge/flood levels are
defined according Raposeiro et al., (2010). The consequences of overtopping/flooding
have been estimated using a methodology that allows a simple qualitative evaluation
of the consequences level associated with hazardous events in the area under study
(Raposeiro et al., 2010).
However, in that methodology there is no prioritization or allocation of weights to
the different environmental, economic and social aspects that can be important for a
given study area in the occurrence of hazardous events that exceed pre-set thresholds.
In order to complement the qualitative method to get the consequences level, Poseiro
et al. (2013b) applied a methodology developed by Antunes (2012) and Craveiro et
al. (2012) based on a multi-criteria analysis that enables a spatial analysis, a
classification and an assignment of weights to each of the aspects that characterize a
given study area. This methodology for the establishment of the consequences map
consists on the construction of a spatial index of human pressure on the port/coastal
area through the application of the Analytic Hierarchy Process (AHP).
Recent developments include the application of AHP method in the port and bay of
Praia da Vitória (Azores), and the application of Coastal Vulnerability Index and
Hazard Assessment to obtain Coastal Risk in sandy beaches with and without coastal
defence structures in Costa da Caparica (Lisbon) by using a geo-referenced database
and the multi-criteria analysis.
Warning system module
The objective of this module is to assess and disseminate warnings of sea-wave
overtopping to the authorities. The system integrates all the information that is
required from, and generated by, the methodology components. It is made of two
components: the data evaluation that integrates and processes the data from the
remaining modules and the user interface. Whenever wave run-up/overtopping
exceeds a pre-set threshold in a specific area, it sends warning messages to the
authorities.
The warning system deals with the following tasks: a) Data acquisition from the
data sources; b) Trigger the wave overtopping determination component; c) Store risk
assessment results; d) Disseminate current warning conditions through the following
channels: Website, Twitter account, and Email; e) Maintain the zone characterization,
using a map with the overtopping consequence layer and threat areas;
The data evaluation component assumes an emergency occurrence whenever the
overtopping / flooding threshold is exceeded in one or more sections of the analysed
structures. This identification of an emergency situation thus leads to the generation
of graphs, charts and reports, which are then transmitted to the interaction component
and presented to the user who will evaluate the situation. The user interaction
component is materialized in a web application, in which the whole warning system is
parameterized. The application is designed for use in traditional and mobile web
browsers adapting the information in accordance with the characteristics of the
client’s device.
Recent developments include the integration of both components of the warning
system. A first version is presented in Sabino et al. (2014). The data evaluation
component consists of a series of Python scripts that engage all models and automate
all the procedures for the operation of the system. Additionally, the web interface is
being designed using public domain tools: the web development platform Django
(also based on Python), for the development of interactive component; and the
database management system PostgreSQL (with the PostGIS spatial extension).
Hydralerta system: Application to Praia da Vitória
The port of Praia da Vitória is located on the east coast of the Terceira island in The
Azores archipelago, Figure 2a. Both breakwaters that protect the harbour define a
roughly rectangular port basin with about 1 km x 2 km. For this port, the system was
applied in the two possible components: warning (modules I, II and IV) and risk
assessment (modules I, II and III), both for the overtopping phenomenon.
The warning system is running permanently for Praia da Vitória. The sea-wave
characterization module runs every day to predict 180 hours of wave characteristics at
the port entrance and into the port, together with wind field and tide level predictions.
Every 3 hours, the system creates a layout with significant wave height and angle of
wave attack, Figure 2b. Once the wave characteristics in the port are available, every
3 hours, the second module, which predicts the run-up/overtopping associated to those
wave characteristics, is applied. For each set of wave/water level characteristics,
NN_OVERTOPPING2 provides information on mean wave overtopping discharge, q,
Figure 2c, for each of the studied cross-sections of the structures. If the mean
overtopping discharge exceeds the pre-set threshold, a warning is issued, Figure 2d.
The WebGIS is presented at Figure 2e.
a) b) c)
d) e)
Figure 2. a) Praia da Vitória aerial view; b) Module I: DREAMS; c) Module II:
NN_OVERTOPPING2; d) Module IV: Warning map generated; e) WebGIS
The risk assessment of Praia da Vitória permits the construction of risks maps in this
bay. In the qualitative risk evaluation, it was considered a five-year period with sea-
wave data from 2008 to 2012, as well as the effects of local wind and the
astronomical tide level. The methodology described in Poseiro et al. (2013a) was
applied to eight structures along the port and bay of Praia da Vitória, from D1 to D8
(Figure 3a). Example of a risk map is presented at Figure 3b, which shows the regions
where mitigation measures are to be implemented. The AHP methodology (Poseiro et
al. 2013b) was employed to generate the consequences map (Figure 3c).
Figure 3. a) Main structures of Praia da Vitória; b) Risk map; c) Consequences map (AHP)
Conclusions
This communication describes the recent developments of the HIDRALERTA
system, a novel system currently being implemented as an early warning application
and also to assess the risk of flooding in coastal and port regions. The system,
implemented in a WebGIS environment, follows the basic idea of using wave
forecasts (up to 180 hours) to calculate the effects of waves on the coast, particularly
in terms of wave overtopping and flooding. Once wave overtopping and flooding are
evaluated, they are compared with pre-defined thresholds, to build warning maps and
risk maps, and, if necessary, to issue warning messages.
Here we have described the application of the system to the Praia da Vitoria bay, in
Terceira Island, Azores. It shows that HIDRALERTA system has the potential to
become a useful tool for the management of coastal and port areas, due to its fast and
efficient capacity to effectively issue warning and to evaluate risks. In the framework
of the HIDRALERTA project, the system has also been applied to low-line areas,
namely sandy beaches and dunes systems under pressure and higher vulnerable to
climate changes impacts such as Costa da Caparica, either as a warning system or as a
risk evaluation tool, but it can be easily extended to other locations. In fact, it has
been applied to other Portuguese locations, such as the ports of Ponta Delgada
(Azores) and Sines, and the Praia da Galé coastal area.
At this point, the project is developing: : a) the replacement of the DREAMS linear
wave model by the BOUSS-WMH nonlinear wave model; b) carry out overtopping
tests on physical models for other types of structures, being the data produced within
these tests used in evaluating the performance of empirical, neuronal network or
numerical tools; c) improve the methodology for constructing maps of consequences;
d) create maps to enable illustration of the spatial distribution of successive volume
thresholds, which will be complemented maps of consequences, and consequently
maps of risk of overtopping/flooding; and finally, e) set suitable levels (thresholds) to
issue a warning.
Acknowledgements
The work described in this publication was supported by the “Fundação para a
Ciência e a Tecnologia”, Portugal, through project HIDRALERTA, ref. PTDC/AAC-
AMB/120702/2010. The participation of all members of the HIDRALERTA project is
acknowledged. The help of the Port of Lisbon Authority (APL), Port of Azores, City
Council of Praia da Vitória and Faculty of Sciences of the University of Lisbon
(FCUL) on providing local wave and tide data and topographic surveys, respectively,
is also acknowledged.
References
Antunes, Ó. (2012). Análise Multicritério em SIG para Determinação de um Índice Espacializado de Pressão
Antrópica Litoral. Casos de Espinho, Caparica e Faro. MsC Thesis, FCSH, UNL.
Booij, N., Ris, R.C. e Holthuijsen, L.H. (1999). A third-generation wave model for coastal regions, Part I,
Model description and validation. J. Geographical Research, C4, 104, 7649-7666.
Coeveld, E.M., Van Gent, M.R.A. e Pozueta, B. (2005). Neural Network: Manual NN_OVERTOPPING2.
CLASH WP8 – Report, june.
Craveiro, J.L., Antunes, O., Freire, P., Oliveira, F., Almeida, I.D. e Sancho, F. (2012). Comunidades urbanas
na orla costeira: a metodologia multicritério AHP (Analytic Hierarchy Process) para a construção de
um índice de vulnerabilidade social face à ação marítima. Atas do 2º Cong. Ibero Americano de
Responsabilidade Social, ISEG, Lisbon, October, 25-27.
Fortes, C.J.E.M. (2002). Transformações Não Lineares de Ondas em Zonas Portuárias. Análise pelo Método
dos Elementos Finitos. Tese de Doutoramento, Eng. Mecânica, IST.
Mase, H., Tamada, T., Yasuda, T., Hedges, T.S. & Reis, M.T. (2013). Wave runup and overtoping at
seawalls built on land in very shallow water. Journal of Waterway, Port, Coastal, and Ocean
Engineering,139(5), pp. 346-357.
Neves, P, Poseiro, P., Fortes, C.J.E.M., Reis, M.T., Capitão, R., Antunes do Carmo, J.S., Raposeiro, P.,
Ferreira, J.C. (2013). Aplicação da metodologia de avaliação do risco de inundação/galgamento na
praia de São João da Caparica. 8ªs JPECP, LNEC, Lisbon, 10-11 October.
Pinheiro L., Fortes C.J.E.M., Santos J.A., Fernandes L., Walkley M. (2011) Boussinesq-type numerical
model for wave propagation near shore and wave penetration in harbors. MARTECH 2011, 1st Int.
Conf. on Maritime Tech. and Engineering, Lisbon, May.
Poseiro, P., Fortes, C.J.E.M., Reis, M.T., Santos, J.A., Simões, A., Rodrigues, C. e Azevedo, E. (2013a). A
methodology for overtopping risk assessment in port areas: Application to the Port of Praia da Vitória
(Azores, Portugal). Proc. SCACR 2013, LNEC, 4-7 june.
Poseiro, P., Fortes, C.J.E.M., Santos, J.A., Reis, M.T. e Craveiro, J. (2013b). Aplicação do processo de
análise hierárquica (AHP) à análise das consequências de ocorrência de galgamentos. O caso da baía
da Praia da Vitória. 8ªs JPECP, LNEC, 10-11 October.
Poseiro, P. (2013). Procedimentos para a execução do sistema HIDRALERTA: Desenvolvimento de
aplicações em Python. NPE- DHA-LNEC. May.
Pullen, T., Allsop, N.W.H., Bruce, T., Kortenhaus, A., Schuttrumpf, H. e Van der Meer, J.W. (2007).
EurOtop: Wave Overtopping of Sea Defences and Related Structures: Assessment Manual.
Environment Agency, UK, Exper. Netwerk Waterkeren, NL, Kuratorium fur Forschung im
Kusteningenieurwesen, DE, August.
Raposeiro P.D, Fortes, C.J.E.M., Reis, M.T. e Ferreira, J.C. (2010). Development of a methodology to
evaluate the flood risk at the coastal zone. In Geographic Technologies Applied to Marine Spatial
Planning and Integrated Coastal Zone Management, Calado, H. e Gil, A. (Eds.), University of The
Azores (Ponta Delgada), August, 129-137. ISBN: 978-972-8612-64-1.
Sabino, A., Rodrigues, A., Poseiro, P., Fortes, J. Reis, T. and Araújo, J., Wave Overtopping Analysis and
Early Warning Forecast System, Proceedings of the 14th International Conference on Comp. Science
and Its Applications (ICCSA 2014), June 30-Jul 3, Guimarães.
Santos, F.L., Reis, M.T., Fortes, C.J., Lotufo, A.D.P. e Maciel, G.F. (2013b). Estudo de Galgamentos
Utilizando Redes Neurais Artificiais do Tipo Artmap com Lógica Fuzzy. 8as JPECP, LNEC, 10-11
October.
Tolman, H. (1999). User Manual and System Documentation of WAVEWATCH-III. version 1.18,
NOA/NWS/NCEP/OMB Technical Note 166, 110p., Washington, E.U.A.
Zózimo, A.C., Fortes, C.J.E.M. e Neves, D.R.C.B. (2008). GUIOMAR: Geographical User Interface for
cOastal and MARine modeling. Recent developments and assessment of potential geographical errors.
Proc. Mediterranean Days of Coastal, Port Eng., Palermo, 7-9 October.
