Editorial Estuarine and Coastal Morphodynamics

Revista da Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management 15(1) – March 2015

Table of Contents

Editorial Invited Editorial Board 5 Estuarine and coastal morphodynamics Journal Editorial Board

Articles Xavier Bertin 9 Processes controlling the seasonal cycle of wave-dominated André B. Fortunato inlets Guillaume Dodet

Magnus Larson 21 Model of the evolution of mounds placed in the nearshore Hans Hanson

Paulo Victor Lisboa 35 Anthropogenic influence on the sedimentary dynamics of a Elisa Helena Fernandes sand spit bar, Patos Lagoon Estuary, RS, Brazil

A. Bio 47 Methods for coastal monitoring and erosion risk assessment: L. Bastos two Portuguese case studies H. Granja J.L.S. Pinho J.A. Gonçalves R. Henriques S. Madeira A. Magalhães D. Rodrigues

Martha Guerreiro 65 Evolution of the hydrodynamics of the Tagus estuary (Portu- André Bustorff Fortunato gal) in the 21st century Paula Freire Ana Rilo Rui Taborda Maria Conceição Freitas César Andrade Tiago Silva Marta Rodrigues Xavier Bertin Alberto Azevedo

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1) – March 2015

A. Maia 81 Cost-benefit analysis of coastal defenses on the Vagueira

C. Bernardes and Labrego beaches in North West Portugal M. Alves

José E. Verocai 91 Addressing climate extremes in Coastal Management: The Monica Gómez-Erache case of the Uruguayan coast of the Rio de la Plata System Gustavo J. Nagy Mario Bidegain

Elida Regina de Melo e Silva 109 Análise da estabilidade da Praia do Janga (Paulista, PE, Daniele Laura Bridi Mallmann Brasil) utilizando ferramenta computacional Pedro de Souza Pereira

Yana Friedrich Germani 121 Vulnerabilidade costeira e perda de ambientes devido à ele- Salette Amaral de Figueiredo vação do nível do mar no litoral sul do Rio Grande do

Lauro Júlio Calliari Sul Carlos Roney Armanini Tagliani

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):5-7 (2015)

http://www.aprh.pt/rgci/pdf/rgci-595_Editorial.pdf | DOI:10.5894/rgci595

Editorial

Estuarine and coastal morphodynamics

Coastal waters are permanently in motion in response to issue) and a sand spit (Lisboa & Fernandes, this issue) tides, waves and wind forcing. When the bed is com- are analyzed using satellite images. Empirical equilib- posed of mobile sediments, as is usually the case, the rium models continue to be used due to their simplicity water motion generates sediment fluxes. The spatial and low cost. This type of approach is illustrated in a variability of these fluxes modifies the local bathy- study of beaches in Brazil (Silva et al., this issue). metry, creating areas of erosion and accretion. Finally, However, these simple tools are progressively being these morphological changes affect the waves and cur- replaced by process-based models. A simple approach rents themselves, thereby creating a feedback loop be- consists in combining circulation process-based models tween the morphology and the hydrodynamics. The with empirical models of bed evolution, as is shown in term “morphodynamics” refers to the study of this Guerreiro et al. (this issue). More sophisticated models interaction between the morphology, the sediments solve conservation equations for the hydrodynamics, fluxes and the hydrodynamic processes. the bottom evolution and the sediment transport. These Early studies of coastal morphological evolution were equations can be solved analytically in very simple naturalist, based on observations or on the notion of cases (Larson & Hanson, this issue), but in general so- equilibrium. Equilibrium models provide coarse predic- phisticated numerical methods are required. Numerical tions of equilibrium states of the system, typically es- process-based models are now routinely used in scien- tablished on empirical relationships. Because the focus tific and engineering studies. For instance, these models was usually on equilibrium, the term “morphodynam- can provide insight into physical mechanisms (Bertin et ics” was rarely used. The field of coastal and estuarine al., this issue), simulate the consequences of human morphodynamics emerged around 35 years ago, and interventions (Lisboa & Fernandes, this issue) or be became a very active area of research in the past decade used to design the deposition of dredging spoils (Larson (Figure 1), with many applications in coastal manage- & Hanson, this issue). ment. Morphodynamics is a complex field, and many pro- Morphodynamic studies rely on different tools (in situ cesses remain poorly understood. Scientists must there- data, physical and numerical models), which are often fore continue to shed light on the behavior of coastal used in conjunction, as illustrated in this issue. A grow- systems. Bertin et al. (this issue) provides an example ing number of in situ measurement devices are now of how process-based numerical models can be used to available, as is illustrated in Bio et al. (this issue). Re- explain this behavior for the particular case of wave- mote sensing is also playing an increasingly prominent dominated inlets. Engineers need to predict the impacts role due to its capability to provide data in harsh condi- of human interventions on the environment or to verify tions and to the growing availability of low cost images. if the observed behavior of a coastal system can be For instance, the evolution of a beach (Silva et al., this attributed to such interventions. Lisboa & Fernandes

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):5-7 (2015)

Figure 1 - Evolution of the research in estuarine and coastal morphodynamics: number of papers per year referenced in the Science Citation Index Expanded until 2014 (search string: Topic=((estuar* or coast* or inlet* or beach) and (morphodynamic or morphodynamics)), performed on March 10, 2015). The peaks in 2009 and 2011 are partly explained by the large number of papers published on the subject in the special issues of the Journal of Coastal Research associated with the In- ternational Coastal Symposiums that occurred in those years. These papers originate mostly from the USA, Europe (Neth- erlands, England, France, Spain, Italy, Portugal and Germany), Brazil and Australia

(this issue) show how an intervention in a tidal inlet morphodynamics to a diverse audience of coastal scien- affected a coastal lagoon, while Larson & Hanson (this tists, practitioners and managers. issue) describe a new model that can help design the While the field of morphodynamics has evolved enor- deposition of dredging spoils in the coast. Also, coastal mously over the past decade, much still needs to be managers need projections on how sea level rise and done. The limits of predictability of the models remain extreme events will affect coastal systems. Guerreiro et modest, partly due to large computational costs, and al. (this issue) shows how sea level rise and sedimenta- partly due to model simplifications. These simplification will change tidal propagation and extreme water tions are often related to the need to limit computational levels in a Portuguese estuary. Germani et al. (this is- costs, but also to insufficient knowledge on the physical sue) assesses the vulnerability of a stretch of Brazilian processes. Hence, models need to be improved in terms coast to sea level rise. Verocai et al. (this issue) ad- of efficiency and accurate representation of the relevant dresses the problem of extreme sea levels in the Uru- physical processes. Simultaneously, further research is guayan coast. Both sea level rise and extreme sea levels required to provide qualitative and quantitative under- are putting an enormous stress on the world’s coast- standing on the processes and their interactions. Also, lines, whose protection has prohibitive costs. Cost- the lack of data is a common limitation in most mor- benefit analyses, such as the one described in Maia et phodynamic studies. Hence, new developments in in al. (this issue), can thus help coastal managers make the situ and remote sensors, as well as techniques to extract best decision. better information from the measurements, are needed This thematic issue originated in the meeting “2ª Con- to provide better data at a lower cost. ferência sobre Morfodinâmica Estuarina e Costeira”, held in Aveiro, Portugal, in 2013. This series of bien- References nial meetings brings together a diverse audience of Bertin, X.; Fortunato, A.B.; Dodet, G. (this issue) - Processes scientists, engineers and managers to share their experi- controlling the seasonal cycle of wave-dominated inlets. ence and ideas on coastal morphodynamics. This issue Revista Gestão Costeira Integrada / Journal of Integrated aims at extending the geographical reach of the confer- Coastal Zone Management, 15(1):9-19. DOI: ence to provide an overview of the state-of-the-art in 10.5894/rgci524.

6 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):5-7 (2015)

Bio, A.; Bastos, L.; Granja, H.; Pinho, J.L.S.; Gonçalves, Lisboa, P.V.; Fernandes, E.H. (this issue) - Anthropogenic J.A.; Henriques, R.; Madeira, S.; Magalhães, A.; Rodri- influence on the sedimentary dynamics of a sand spit bar, gues, D. (this issue) - Methods for coastal monitoring and Patos Lagoon Estuary, RS, Brazil. Revista Gestão erosion risk assessment: two Portuguese case studies. Re- Costeira Integrada / Journal of Integrated Coastal Zone vista Gestão Costeira Integrada / Journal of Integrated Management, 15(1):35-46. DOI: 10.5894/rgci541. Coastal Zone Management, 15(1):47-63. DOI: Maia, A.; Bernardes, C.; Alves, M. (this issue) - Analysis of 10.5894/rgci490. cost-benefit of coastal defense in Vagueira and Labrego Germani, Y.F.; Figueiredo, S.A.; Calliari, L.J.; Tagliani, beaches (NW of Portugal). Revista Gestão Costeira Inte- C.R.A. (this issue) - Vulnerabilidade costeira e perda de grada / Journal of Integrated Coastal Zone Management, ambientes devido a elevação do nível do mar no sul do 15(1):81-90. DOI: 10.5894/rgci521. Brasil. Revista Gestão Costeira Integrada / Journal of In- Silva, E.R.M.; Mallmann, D.L.B.; Pereira, P.S. (this issue) - tegrated Coastal Zone Management, 15(1):121-131. DOI: Análise da estabilidade da Praia do Janga (Paulista, PE, 10.5894/rgci540. Brasil) utilizando ferramenta computacional. Revista Guerreiro, M.; Fortunato, A.B.; Freire, P.; Rilo, A.; Taborda, Gestão Costeira Integrada / Journal of Integrated R.; Freitas, M.C.; Andrade, C.; Silva, T.; Rodrigues, M.; Coastal Zone Management, 15(1):109-120. DOI: Bertin, X.; Azevedo, A. (this issue) - Evolution of the 10.5894/rgci492. tidal dynamics of the Tagus estuary (Portugal) in the 21st Verocai, J.E.; Gómez-Erache, M.; Nagy, G.J.; Bidegain, M. century. Revista Gestão Costeira Integrada / Journal of (this issue) - Addressing climate extremes in Coastal Integrated Coastal Zone Management, 15(1):65-80. DOI: Management: The case of the Uruguayan coast of the Rio 10.5894/rgci515. de la Plata System. Revista Gestão Costeira Integrada / Larson, M.; Hanson, H. (this issue) - Model of the evolution Journal of Integrated Coastal Zone Management, of mounds placed in the nearshore. Revista Gestão 15(1):91-107. DOI: 10.5894/rgci555. Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33. DOI: 10.5894/rgci530.

Invited Editorial Board

André B. Fortunato Chair of the Invited Editorial Board Xavier Bertin Giovanni Coco Ana Vila Concejo Invited Editor Invited Editor Invited Editor João M. Dias Elisa H. Fernandes Magnus Larson Invited Editor Invited Editor Invited Editor Ana Matias Anabela Oliveira Paulo Silva Invited Editor Invited Editor Invited Editor

Editorial Board

J. Alveirinho Dias Ulisses Azeiteiro Monica F. Costa Tomasz Boski Editor-in-Chief Associate Editor Associate Editor Advisor Editor

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):9-19 (2015)

http://www.aprh.pt/rgci/pdf/rgci-524_Bertin.pdf | DOI: 10.5894/rgci524

Processes controlling the seasonal cycle of wave-dominated inlets*

@, Xavier Bertin@, a; André B. Fortunatob; Guillaume Dodeta

ABSTRACT This paper reviews the physical processes controlling wave-dominated inlets, based on several studies conducted at two inlets located on the West Coast of Portugal. Once the observed hydrodynamics and morphological changes are reasonably simulated, numerical experiments are performed to explain the development of the inlet during fair weather conditions and its shoaling and closure during winter storms. The former behaviour is explained by a tidal distortion that promotes ebb- dominance while the latter is explained by the combination of several wave-related processes: (1) the “bulldozer effect” due to the shore-normal component of wave forces; (2) the presence of lateral barotropic pressure gradients, accelerating longshore flows towards the inlet; (3) wave blocking during the ebb and (4) a rise in mean sea level, peaking in late autumn. Recent results also suggest that infragravity waves may play a major role. Key words: Tidal inlets, wave-induced processes, wave blocking, infragravity, Óbidos lagoon, Albufeira lagoon.

RESUMO Processos que controlam o ciclo sazonal de embocaduras dominadas pelas ondas. Apresenta-se uma revisão dos processos físicos que controlam a dinâmica de embocaduras dominadas pelas ondas, com base em vários estudos conduzidos em duas embocaduras localizadas na zona costeira centro de Portugal. As análises são con- duzidas com modelos morfodinâmicos, que reproduzem adequadamente a hidrodinâmica e a evolução morfológica observadas. Experiências numéricas são efetuadas para explicar o desenvolvimento das embocaduras durante o verão marítimo e a sua colmatação e fecho durante tempestades de inverno. O primeiro comportamento é explicado pela distorção da maré que promove a dominância da vazante, enquanto o segundo se explica pela combinação de vários processos relacionados com a agitação marítima: (1) o “efeito bulldozer” devido à componente transversal à costa das forças devidas às ondas; (2) a presença de gradientes laterais de pressão barotrópica, que aceleram as correntes longitudinas em direcção à embocadura; (3) o bloqueamento das ondas durante a vazante e (4) o ciclo sazonal do nível médio do mar, cujo pico ocorre no outono. Resultados recentes sugerem ainda que as ondas infragravíticas podem ter um papel importante nesta dinâmca. Palavras-chave: embocaduras de maré, processos induzidos pelas ondas, bloqueamento das ondas, ondas infragravíticas, lagoa de Óbidos, Lagoa de Albufeira.

@ Corresponding author to whom correspondence should be addressed. a UMR 7266 LIENSs, CNRS-La Rochelle University, 2 rue Olympe de Gouges, 17000 La Rochelle, France. e-mails: Bertin ; Dodet b Laboratório Nacional de Engenharia Civil, Avenida do Brasil, 101, Lisboa, Portugal. e-mail:

* Submission: 13 JUN 2014; Peer review: 25 JUL 2014; Revised: 11 DEC 2014; Accepted: 13 DEC 2014; Available on-line: 16 DEC 2014

Bertin et al. (2015)

1. Introduction understanding of the main physical processes that drive morphological changes. The economic and environmental importance of tidal inlets has been growing worldwide, while their sustain- This paper reviews the knowledge gained from previous able management faces many conflicting challenges, modelling-based studies performed at two wave- such as the maintenance of open navigation routes, the dominated inlets located on the west coast of Portugal stability of the adjacent shoreline or the water renewal (Bertin et al., 2009b; Bruneau et al., 2011; Dodet et al., in the back-barrier lagoons. The combined action of 2013; Fortunato et al., 2014) and synthesizes the phys- waves and tides and the presence of shallow channels ical processes controlling the seasonal cycle of wave- often drive a fast and intense sediment dynamics and dominated inlets, including their enlargement during make their behaviour difficult to predict. These prob- fair weather conditions and their shoaling or even clos- lems are particularly relevant at wave-dominated inlets, ure during winter months. where this intense dynamics can drive fast and large morphological changes within a few days/weeks. Even- 2. Study sites tually, wave-dominated inlets can episodically or seas- This study is based on results obtained at two wave- onally close, although the responsible physical pro- dominated inlets located on the West Coast of Portugal: cesses remain only partly understood. the Albufeira Lagoon Inlet and the Óbidos Lagoon Inlet To improve the understanding of tidal inlet dynamics, a (figure 1). Tides are semi-diurnal and range from 0.5 m promising avenue is the development and application of to 3.5 m (meso-tidal). When tides propagate into the morphodynamic modelling systems. These modelling lagoons, the semi-diurnal tidal constituents are severely systems consist of a set of modules to simulate tidal hy- damped, with the amplitude of M2 typically decreasing drodynamics, wave propagation, sediment transport and by 50 to 80 % (figure 1) at both lagoons (Oliveira et al., bottom evolution. The morphodynamic modelling of 2006; Bertin et al., 2009b; Dodet et al., 2013). Tidal amplitude in the lagoon experiences a seasonal cycle, tidal inlets already met several successes over the last with a maximum at the end of the summer and a mini- decade (Cayocca, 2001; Dastgheib et al., 2008; Bertin mum at the end of the winter at Óbidos. In the Al- et al., 2009a; Bruneau et al., 2011). Nevertheless, to bufeira Lagoon Inlet usually closes in autumn (fig- date, the successful simulation of tidal inlet closure is ure 1). In contrast to semi-diurnal constituents, quarter- restricted to simplified/empirical approaches (Ranasin- diurnal and fortnightly non-linear tidal constituents de- ghe et al., 1999) and/or synthetic tidal inlets (Walstra et velop inside the lagoons, resulting in a strongly dis- al., 2009). This problem suggests that the dominant torted tidal signal, with ebb lasting 7 to 8 hours and physical processes are not all captured by most model- floods 4.5 to 5.5 hours. Freshwater discharges are usu- ling systems. However, the numerical modelling of ally negligible compared to the dynamics induced by wave-dominated inlets made significant progresses over waves and tides, particularly in Albufeira where the la- the last decade, and resulted namely in an improved goon is closed from autumn to mid-spring.

Figure 1 - Bathymetric map of the central western coast of Portugal, aerial view of the Albufeira and Óbidos lagoon inlets and time-series of the amplitude of the constituent M2 inside those lagoons based on water level measurements. Figura 1 - Batimetria da zona centro da costa Portuguesa, vista aérea das embocaduras das lagoas de Óbidos e Albufeira e séries temporais da amplitude da constituinte M2 nestas lagunas, baseadas em medições de altura de água. 10 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):9-19 (2015)

The continental shelf in front of both inlets is very nar- tails on these field measurements and data processing row (< 20 km), which causes these inlets to be exposed can be found in Oliveira et al. (2006), Bertin et al. to a very energetic wave climate, particularly in winter. (2009b), Dodet et al. (2013) and Fortunato et al. (2014). Based on a 57-year wave numerical hindcast (Dodet et al., 2010), the mean annual deep water (10.0º W; 3.2. Numerical modelling system 38.0º N; ~3000 m deep) significant wave height (HS), mean direction (MWD) and peak period (TP) are re- The numerical results presented in this study rely on the spectively 1.9 m, 312º, and 10.5 s. During winter (resp. unstructured-grid modelling systems MORSYS2D summer) the corresponding values are: 2.5 m, 305º, and (Fortunato & Oliveira, 2004; Bertin et al., 2009b) and 12.1 s (resp. 1.3 m, 319º and 8.4 s). Also, the presence SELFE (Dodet et al., 2013; Roland et al., 2012). These of coarse sediments (i.e., d50 in the range 0.5-1.0 mm) modelling systems share the same philosophy and fully causes the adjacent beaches to display steep faces, couple a spectral wave model (either SWAN, Boiij et which favour energetic wave breaking and large sedi- al., 1999; or WWMII, Roland et al., 2012), a 2DH cir- ment transport rates. culation model (either ELCIRC, Zhang et al., 2004; or This severe wave climate combined to the meso-tidal SELFE, Zhang et al., 2011) and a sediment trans- range and shallow channels leads to very dynamic in- port/bottom update model (either SAND2D, Fortunato lets, with channel migration which can reach & Oliveira, 2004; Bertin et al., 2009b; or SED2D, Do- 50 m.week-1(Fortunato et al., 2014). Both systems are det, 2013). For both sites, the unstructured grids have a characterized by a seasonal cycle, with an enlargement spatial resolution ranging from about 500 m along the and deepening of the main channel during the summer open boundary to 5 m along the coast in order to ad- period and a strong shoaling during the winter period equately represent the surf zone and wave-induced (Bertin et al., 2009b; Fortunato et al., 2014). flows. The numerical procedure can be seen on figure 2 and includes three main steps: 3. Data and methods 1 - First, the propagation of short waves is simulated 3.1. Field measurements using a spectral wave model forced along its open boundary by time-series of spectra originating from At both sites, several field campaigns were carried out the regional wave model of Dodet et al. (2010). The over the last decade, where pressure sensors, electro- spectral wave model is fed by fields of elevation and magnetic current-meters and ADCPs were deployed currents originating from the circulation model. over both the flood and the ebb deltas, in order to char- 2 - The horizontal circulation is simulated using SELFE acterize wave and tide transformation along their or ELCIRC, which is forced along its open boundary propagation through the inlets. Pressure sensors were by the 16 main tidal constituents whose amplitude also deployed during several months inside the lagoons and phase are computed with the regional tidal in order to characterize properly the seasonal evolution model of Bertin et al. (2012). A full coupling is of tidal amplitude. Finally, repetitive bathymetric and achieved with the spectral wave model through topographic surveys were carried out to quantify the gradients of radiation stresses, horizontal viscosity fast morphological changes observed at both sites. De- and bottom friction.

Figure 2. Flowchart of the modelling system procedure. Figura 2. Esquema do modelo morfodinâmico.

11 Bertin et al. (2015)

3 - Sand fluxes are computed based on time-series of 4.1. Inlet development during fair weather condi- currents, water levels and wave parameters by tions means of total transport empirical formulae (e.g. In order to understand why wave-dominated inlets en- Soulsby and Van Rijn, Soulsby, 1997). The Exner or sediment continuity equation is then solved using a large during fair weather conditions, we performed syn- node-centred finite volume method. thetic simulations at the Óbidos lagoon under tidal forcing only. We considered simplified tides represented 4. Results and discussion by M2 only, whose amplitude was set to 0.75 m (mean neap conditions) and 1.5 m (mean spring conditions). Previous studies conducted by our team have demon- Time series of water depth at the inlet reveal firstly that strated that our modelling system was capable to repro- tides are strongly distorted at the inlet, with a shorter duce waves and currents at the studied sites with a nor- flood than ebb. This distortion is stronger for spring malized root mean square error (hereafter NRMSE) of tides with an ebb duration of 7.5 h and a flood duration the order of 10-15 %. Water levels are predicted more of 5.0 h. According to classical theories on tidal distor- accurately, with a NRMSE of the order of 5 % (Bertin tion for estuaries (e.g., Friedrichs & Aubrey, 1988), et al., 2009b; Bruneau et al., 2011; Dodet et al., 2013). Morphological predictions have larger errors and their longer ebb would result in higher current velocities dur- accuracy deteriorates with time along the simulation, ing flood. Yet, the opposite behaviour is observed at although the main patterns are reproduced qualitatively. both inlets, with slightly larger velocities occurring dur- In particular, the enlargement of the main channel during the ebb and lasting more than maximum flood ve- ing fair weather conditions and its shoaling during win- locities (figure 3-C). This paradoxical behaviour is re- ter months are well captured (Bertin et al., 2009a; lated to the fact that maximum flood occurs for a water Bruneau et al., 2011). In some cases, the model can also depth twice as large as maximum ebb, which causes reproduce the meandering of the channels (Bruneau et ebb currents to be stronger so that mass conservation is al., 2011) and the inlet migration (Bertin et al, 2009c). ensured. Higher velocities in shallower depth during Under these conditions, the dominant physical pro- ebb cause associated sand fluxes to be 1.4 to 2.0 times cesses responsible for these morphological changes are larger than on flood (figure 3-D). As a consequence, assumed to be well captured by our modelling system. under tidal forcing only, the Óbidos and Albufeira La- In this section, we present the results of numerical ex- goon inlets remain strongly ebb-dominated from ve- periments that aim at describing and quantifying these locity and sediment transport viewpoints, with a processes.

Figure 3 - (A) Bathymetry of the Óbidos Lagoon Inlet in July 2001 showing the profile where model outputs were averaged, (B) mean water depth across the inlet, (C) mean depth-averaged velocity across the inlet and (D) sediment fluxes integrated accross the inlet. Figura 3 - (A) Batimetria da embocadura da lagoa de Óbidos em julho de 2001, mostrando o perfil onde os resultados do modelo foram integrados, (B) profundidade média na embocadura, (C) média das velocidades médias na vertical e (D) fluxos sedimentares integrados através da embocadura.

12 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):9-19 (2015) stronger ebb-dominance for spring tides compared to very weak. In front of the inlet, large wave forces on the neap tides. This ebb-dominance favours the flushing of ebb delta are no longer balanced by a barotropic pres- sediments at the inlet mouth and explains why wave- sure gradient because the wave-induced setup is spread dominated inlets enlarge during fair weather conditions, within the lagoon. As a result, a strong residual forcing when wave-related processes are not dominant. occurs on the ebb delta, which was referred to as “bulldozer effect” by Hageman (1969). This phenomenon is 4.2. Inlet shoaling during winter period well captured by our modelling system, which is able to reproduce the migration of ebb-delta sandbars towards 4.2.1. The “bulldozer effect” the lagoon (Bertin et al., 2009b). Field observations revealed that wave-dominated inlets in Portugal experience shoaling during winter storms, 4.2.2. Lateral barotropic pressure gradients occasionally leading to their closure (Bertin et al., Figure 4-D shows that wave forces induce a setup at 2009b; Fortunato et al., 2014). In order to understand which wave-related processes can induce the shoaling adjacent beaches of the order of 10 % the wave height of wave-dominated inlets during winter storms, we per- at the breaking point. At the inlet, this setup is inter- formed a synthetic simulation at the Óbidos Lagoon rupted, which induces a lateral barotropic pressure gradient. This barotropic pressure gradient is not com- considering energetic (Hs = 3.0 m; Tp = 12 s) shore- normal wave conditions with a 1.1 m tidal amplitude pensated by any wave forces so that a strong residual (mean annual tidal range). Figure 4-B shows that wave forcing directed towards the lagoon occurs on both dissipation over the ebb-delta and at adjacent beaches sides of the inlet. These pressure forces result in an ac- results in large gradients of radiation stress (wave for- celeration of longshore currents towards the lagoon, ces) directed onshore (figure 4-C). These forces induce which tend to push large quantities of sediments into a setup reaching 0.25 m at adjacent beaches, which it- the lagoon. At high tide, this phenomenon is further en- self induces a barotropic pressure gradient. At adjacent hanced by wave refraction over the ebb-delta, which beaches, this barotropic pressure gradient nearly balan- causes wave-induced longshore currents to converge ces wave forces (figure 4-F) and the residual forces are towards the inlet (not shown, Bertin et al., 2009b).

Figure 4 - (A) Bathymetry (m), (B) significant wave height (m), (C) wave forces (Pa), (D) free surface elevation (m), (E) barotropic pressure gradient (Pa) and (F) resultant forces (Pa) at the Óbidos Inlet for shore-normal offshore waves of Hs = 3.0 m. Figura 4 - (A) Batimetria (m), (B) altura significativa (m), (C) forças devidas às ondas (Pa), (D) elevação da superfície livre (m), (E) gradiente de pressão barotrópica (Pa) e (F) resultante das forças (Pa) na embocadura da Lagoa de Óbidos para ondas incidentes perpendiculares à costa com Hs = 3.0 m.

13 Bertin et al. (2015)

4.2.3. Wave blocking during ebb Dodet el al. (2013) investigated the impact of these phenomena on the sediment dynamics of tidal inlets. On A field campaign was carried out at the Albufeira La- flood, the presence of waves superimposed on tidal cur- goon Inlet in September 2010 where pressure transdu- rents in the main channel increases sediment transport cers were deployed on both the ebb and the flood deltas rates. During the ebb, waves are first dissipated by (figure 5-A). A time series of significant wave heights whitecapping, and then blocked, so that the transport on the flood delta (PT2, figure 5-B) revealed firstly that capacity of ebb currents is no longer enhanced. Over a the wave height inside the lagoon is modulated along tidal cycle, these processes decrease the capacity of the the tidal cycle and becomes almost nil at low tide. This inlet to flush sediments out of the lagoon. Nevertheless, behaviour is due to wave energy dissipation by break- further research is needed to determine whether this ing on the ebb delta, which increases as water level de- conclusion can be extended to other wave conditions creases. Furthermore, H are not symmetrical for a s and inlet configurations. Further experiments will have given water level and experience a fast drop when ebb to be carried out during more energetic wave conditions currents start to establish. To better understand this be- and/or at wider inlets, where larger waves can propa- haviour, the modelling system was run with and without gate in the main channel. current feedback on wave propagation. Figure 5-B shows that the fast drop in H is only reproduced if the s 4.2.4. Infragravity waves feedback of currents in the wave model is taken into account. The modelling results show firstly that waves A spectral analysis of pressure transducer and current propagating against currents experience whitecapping data collected at the Albufeira Lagoon Inlet in Septem- dissipation (Dodet et al., 2013). This process is repre- ber 2010 (figure 5-A) showed that a significant part of sented in our modelling system following the approach spectral energy was found in the infragravity band of Westhuysen (2012). Two hours after the beginning (0.004 Hz - 0.04 Hz, figure 6). This phenomenon is par- of the ebb, wave heights inside the lagoon become al- ticularly clear on the third tidal cycle, where more en- most nil. The analysis of modelled tidal currents reveals ergy is found on the infra-gravity band than the gravity that, in the inlet main channel, waves propagate against band (figure 6). This low-frequency energy is expressed tidal currents locally exceeding 2.0 m/s. Such large ve- as low-frequency fluctuations of the free-surface eleva- locities almost correspond to the wave group velocity, tion, wave heights and current velocities. These fluctu- which causes the waves to be blocked locally (Dodet et ations were particularly visible in the measured data at al., 2013). PT2, when comparing the 1-min running averaged with

Figure 5 - (A) Bathymetry of the Albufeira Lagoon Inlet, (B) measured water depth and (C) measured and modelled time series of significant wave heights on the flood delta in September 2010. Figura 5 - (A) Batimetria da embocadura da Lagoa de Albufeira, (B) altura de água medida e (C) séries temporais, medidas e simuladas, da altura significativa no banco de enchente em setembro de 2010.

14 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):9-19 (2015) the 60 min running averaged time-series (figure 7). research is needed. Namely, comparisons at other sites Such oscillations reached up to 10 %, 20 % and 50 % of are required and the implementation of infragravity the 60 min filtered signal, for free-surface elevation, waves in modelling systems (e.g., Roelvink, 2009) wave heights and currents, respectively. While the would be a promising perspective. physical processes that explain these infragravity fluctuations require further investigation, Dodet (2013) es- 4.2.5. Variations in mean-sea level timated their contribution to the sediment dynamics of the inlet – aside from the effect of wave-current interac- Previous studies have shown that at both Óbidos (Bertin tions. In this context, the total transport induced by et al., 2009b) and Albufeira (Dodet et al., 2013), the waves and currents was computed at PT2 (figure 5A), setup induced by wave breaking in front of the inlet using the Soulsby and Van Rijn formula (Soulsby, propagates inside the lagoon and reaches roughly 10 % 1997), fed on the one hand by the 60 min filtered time- of the wave height at breaking. Energetic waves and series of elevation, Hs and velocity (the wave orbital winter storms can thus induce variations of mean sea- velocity was computed using the linear theory) and on level (hereafter MSL) of a few tens of centimetres. the other hand the 1 min filtered time-series that include Since both inlets are very shallow, such variations in the low-frequency oscillations. During the third tidal MSL are expected to impact tidal asymmetry and cycle, where low frequency fluctuations were the larg- thereby sediment dynamics significantly. To better est, the total sediment transport was locally up to eight quantify this phenomenon, morphodynamic simulations times as large when taking into account low-frequency were performed at the Óbidos inlet under tidal forcing fluctuations and twice as large when integrated over a only and varying MSL between -0.2 m and +0.4 m. tidal cycle. Changes in tidal asymmetry were characterized through It can also be noted that these low-frequency fluctu- the ratio between the amplitude of M4 and M2. The im- ations appear mostly during the flood, which suggests pact on sediment transport was characterized through that infragravity waves are also damped or blocked by the ratio between the sediments flushed at the inlet dur- ebb currents. Therefore, in terms of inlet morphody- ing the ebb and entering the lagoon during the flood. namics, this behaviour implies that these low-frequency Results show that higher water levels reduce tidal fluctuations rather contribute to fill the lagoon with asymmetry (figure 8-B), strongly decrease ebb- sediments. This process can play a key role during dominance, and the ratio between ebb and flood sand storms, where it can significantly contribute to inlet transport across the inlet tends to 1 (figure 8-C). closure. However, according to the authors’ knowledge, Because floods occur, on average, at higher water levels it is the first time that the importance of infragravity than ebbs (Fortunato & Oliveira, 2007), the water flows waves at tidal inlets is demonstrated and thus further more freely into, than out of, the lagoon. Floods are

Figure 6 - Energy spectra of the free surface elevation at PT2 (figure 5A), showing that a large part of the energy is located in the infragravity band. Figura 6 - Espectro de energia da superfície livre na estação PT2 (Figura 5A), mostrando que grande parte da energia está localizada na banda infragravítica.

15 Bertin et al. (2015)

Figure 7 - 1-min and 60-min running averaged time-series of water depth, velocity, and total sediment transport (Qt) at PT2 (figure 5). Figura 7 - Médias de banda móvel de 1 e 60 minutos da altura de água, velocidade e fluxo sedimentar total (Qt) em PT2 (Figura 5).

Figure 8 - (A) Bathymetry of the Óbidos lagoon with the cross-section where sand transport was integrated, (B) ratio between the amplitudes of M4 and M2 in the lagoon and (C) residual sand transport across the inlet (ebb/flood). Figura 8 - (A) Batimetria da Lagoa de Óbidos mostrando a secção transversal onde o transporte de areia foi integrado, (B) quociente entre as amplitudes da M2 e da M4 na laguna e (C) transporte de areia residual através da embocadura (vazante/enchente). therefore shorter than ebbs, which would contribute to tates the water outflow, thereby reducing the ebb dur- higher velocities on flood than on ebb. However, mass ation and the flood dominance. For instance, the mean conservation also implies that the higher water depths water level in the lagoon explains 50 % of the differ- on flood reduce the flood velocities relative to the ebb ence between ebb and flood durations in the Albufeira currents. While the former process is usually the domi- lagoon (Fortunato et al., 2014), with longer floods cor- nant one in large estuaries and inlets, we found the lat- responding to higher mean water levels. On the other ter to dominate in our two shallow lagoons. Raising the hand, the relative differences between the water depths mean sea level has therefore two opposite effects. On on ebb and flood decrease, which reduces the distinc- the one hand, the higher water depth in the inlet facili- tion between ebb and flood velocities due to continuity.

16 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):9-19 (2015)

This process is illustrated on Figure 8C. Since tidal Besides the rise in mean sea level associated to seasonal asymmetry is dominated by the mass conservation ef- effects and storms, significant fortnightly variations of fect in shallow lagoons, raising the mean water level water level can also occur inside the lagoon due to non- reduces the predominance of ebb currents in these sys- linear tides. These fortnightly tides are particularly tems. strong in very shallow systems. In the Albufeira lagoon, In addition to the wave-induced setup which develops their amplitude reaches several tens of centimetres, with in the nearshore, MSL experiences a seasonal cycle crests occurring on spring tides and troughs on neap along the west coast of Portugal. A permanent tide tides (Fortunato et al., 2014). The importance of these gauge is located in the Cascais marina between both constituents is shown at the Albufeira lagoon in a simu- studied sites (figure 1). A Demerliac filter was applied lation forced by M2 and S2 tidal constituents. The am- to a time series of water level originating from this sta- plitude of the fortnightly constituent MSf reaches about tion and daily and monthly MSL were computed (fig- 15 cm. The associated velocity at the tidal inlet throat ure 9). This figure reveals that mean sea-level reaches exceeds 30 cm/s. This velocity is in phase with the ele- its maximum in autumn and its minimum in late winter, vations (Figure 10B), indicating that this constituent is a with a 0.2 m difference between both. Although outside standing wave. These strong velocities affect the sedi- the scope of this study, investigations in progress in our ment fluxes significantly. On spring tides they increase team show that this seasonal cycle results from the sediment fluxes on ebb and decrease them on flood superimposition of atmospheric forcing and steric ef- (Figure 10C). The opposite occurs on neap tides (Fig- fects. The Albufeira Lagoon Inlet usually closes in ure 10D). Hence, ebb dominance from a sediment autumn (e.g., Dodet et al., 2013): it is likely that the de- viewpoint is not only higher on spring tides due to the crease in ebb-dominance related to a higher MSL can stronger fourth-diurnal constituents, but also due to the contribute significantly to inlet closure. fortnightly constituents.

Figure 9 - Daily and Monthly mean sea-level in Cascais between 2000 and 2007. Figura 9 - Médias diárias e mensais do nível do mar em Cascais entre 2000 e 2007.

Figure 10 - (A) Batimetry of the Albufeira lagoon inlet, relative to mean sea level. (B) Water levels inside the lagoon obtained in a simulation forced by M2 and S2. MSf velocities at the inlet throat are also shown as a dashed line. Sediment fluxes across the tidal inlet estimated with the Engelund and Hansen (1967) formula on spring (C) and neap tides (D). Figura 10 - (A) Batimetria da embocadura da Lagoa de Albufeira, relativa ao nível médio do mar. (B) Níveis dentro da laguna obtidos numa simulação forçada pelas constituintes M2 e S2. As velocidades associadas à constituinte MSf no centro do canal da embocadura são mostradas a tracejado. Os fluxos sedimentares através da embocadura foram calculados com a fórmula de Engelund e Hansen (1967) em marés vivas (C) e mortas (D).

17 Bertin et al. (2015)

5. Conclusions and future works for Coastal Margin Observation and Prediction, respectively. Topographic and hydrodynamic data were obtained, thanks to This study presented a synthesis of the physical pro- the combined efforts of many individuals from the Faculty of cesses controlling the seasonal cycle of wave- Science of the University of Lisbon, the National Laboratory dominated inlets and revealed firstly that the inlet de- of Civil Engineering, the University of Algarve and the Insti- velopment during fair weather conditions was caused tuto Hidrográfico. The time series of sea surface elevation at by a huge tidal distortion that promotes ebb-dominance. Cascais were provided by the Portuguese Geographic Insti- The shoaling or closure of wave dominated inlets dur- tute. ing winter storm was explained by the superimposition of several wave-related processes: (1) the “bulldozer References effect” due to the shore-normal component of wave for- Behrens, D.K.; Bombardelli, F.A.; Largier, J.L.; Twohy, E. ces acting over the ebb shoals; (2) the presence of lat- (2009) - Characterization of time and spatial scales of a mi- eral barotropic pressure gradients, accelerating long- grating river mouth. Geophysical Research Letters, shore flows towards the inlet; (3) wave blocking during 36(9):L09402 . DOI: 10.1029/2008GL037025 the ebb; (4) the reduction of ebb dominance by a high Behrens, D.K.; Bombardelli, F.A.; Largier, J.L.; Twohy, E. mean sea level in autumn and during storms; and (5) the (2013) - Episodic closure of the tidal inlet at the mouth of presence of infragravity waves. In addition, fortnightly the Russian River – A small bar-built estuary in California, overtides generated by non-linear processes enhance Geomorphology, 189(1):66-80. DOI: 10.1016/j.geomorph. 2013.01.017 ebb-dominance on spring tides and have the opposite effect on neap tides. This effect helps explaining the Bertin, X.; Oliveira, A.; Fortunato, A.B. (2009a) - Simulating morphodynamics with unstructured grids: description and validation stronger tendency for inlet closure on neap tides. While of a modelling system for coastal applications. Ocean model- the impact of processes (1), (2) and (4) on inlet mor- ling, 28(1-3):75-87. DOI: 10.1016/j.ocemod.2008.11.001 phodynamics were already quantified at the Óbidos La- Bertin X.; Fortunato, A.B.; Oliveira, A. (2009b) - A modeling-based goon Inlet, the contribution of wave blocking and infra- analysis of processes driving wave-dominated inlets. Continen- gravity waves remains to be investigated. Also, this tal Shelf Research, 29(5-6):819-834. DOI: 10.1016/ study mostly investigates cross-shore wave induced j.csr.2008.12.019 processes while oblique waves can induce strong inlet Bertin, X.; Fortunato, A.B.; Oliveira, A. (2009c) - Morphodynamic migration. The subsequent lengthening of the channel Modeling of the Ancao Inlet, South Portugal. Journal of Coastal can alter tidal propagation and limit the ability of the Research, SI56:10-14, Coastal Education & Research Founda- inlet to flush sediments. It is expected that the proper tion, Lisboa, Portugal. Available on-line at http://www.jstor.org/ sta- representation of all these processes in morphodynamic ble/25737527 modelling systems will allow simulating the closure of Bertin, X.: Bruneau, N.; Breilh, J.F.: Fortunato, A.B.; Karpytchev, wave-dominated inlets. M. (2012) - Importance of wave age and resonance in storm surges: the case Xynthia, Bay of Biscay. Ocean Modelling, However, hydrodynamic conditions display relatively 42(4):16-30. DOI: 10.1016/j.ocemod.2011.11.001 modest variations along the West Coast of Portugal so Booij, N.; Ris, R.; Holthuijsen, L. (1999) - A third-generation that the physical processes analysed in this study may wave model for coastal regions. 1. Model description and vali- be partly site-specific. Their importance will have to be dation. Journal of Geophysical Research: Oceans, investigated at other wave-dominated inlets, such as in 104(C4):7649–7666. DOI: 10.1029/98JC02622 SW Australia, South Africa, California and Central and Bruneau, N.; Fortunato, A.B.; Dodet, G.; Freire, P.; Oliveira, A.; Southern America. Bertin, X. (2011) - Future evolution of a tidal inlet due to changes in wave climate, Sea level and lagoon morphology (Óbidos Also, other processes not discussed herein can affect the lagoon, Portugal). Continental Shelf Research, 31(18):1915- morphodynamics of tidal inlets. For instance, channel 1930. DOI: 10.1016/j.csr.2011.09.001 meandering increases the risk of inlet closure (Behrens Chaumillon, E.; Ozenne, F.; Bertin, X.; Long, N.; Ganthy, F. et al., 2009, 2013), while fresh water flows promote its (2014) - Wave climate and inlet channel meander bend con- opening (Shuttleworth et al., 2005). Also, the curvature trol spit breaching and migration of a new inlet: La Coubre sand spit, France. Journal of Coastal Research, SI70:109- of the channel affects the inlet migration (Chaumillon et 114. Available on-line at http://www.cerf-jcr.org/images/stories/ al., 2014). Overall, the morphodynamics of wave- 2014_ICS_Proceedings/JCR_SI_70_019_Chaumillon_et_al.pdf dominated inlets is a complex problem, and its control- Dastgheib, A.; Roelvink, J.A.; Wang, Z.B. (2008) - Long term pro- ling mechanisms remain only partly understood. cess-basedmorphological modeling of the Marsdiep Tidal Basin. Marine Geology, 256(1-4):90-100. DOI: 10.1016/j.margeo. Acknowledgements 2008.10.003 Dodet, G.; Bertin, X.; Taborda, R. (2010) - Wave climate variability This work was part of a project funded by the Portuguese in the North-East Atlantic Ocean over the last six decades. Foundation for Science and Technology (FCT): 3D- Ocean Modelling, 31(3-4):120-131. DOI: 10.1016/j.ocemod. MOWADI (PTDC/ECM/103801/2008). Researches in pro- 2009.10.010 gress are part of the research project ANR JC DYNAMO Dodet, G.; Bertin, X.; Bruneau, B.; Fortunato, A.B.; Nahon, A.; Ro- (agreement ANR-12-JS06-00008-01). The wave models land, A. (2013) - Wave-current interactions in a wave- (SWAN) and the circulation models (ELCIRC/SELFE) were dominated tidal inlet. Journal of Geophysical Research: provided by Delft University of Technology and the Center Oceans, 118(3):1587–1605. DOI: 10.1002/jgrc.20146 18 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):9-19 (2015)

Dodet, G., (2013) - Morphodynamic modelling of a wave-dominated Ranasinghe, R.; Pattiaratchi, C.; Masselink, G. (1999) - The seasonal tidal inlet: the Albufeira Lagoon. 181p., PhD thesis, La Ro- closure of tidal inlets: Wilson Inlet, a case study. Coastal Engi- chelle University, La Rochelle, France. Unpublished neering, 37(1):37-56. DOI: 10.1016/S0378-3839(99)00007-1 Engelund, F.; Hansen, E. (1967) - A monograph on sediment trans- Roelvink, J.A.; Reniers, A.; van Dongeren, A.; de Vries, J. van port in aluvial streams. 63p., Teknisk Forlag, Technical Univer- Thiel; McCall, R.; Lescinski, J. (2009) - Modeling storm im- sity of Danmark, Copenhagen, Denmark. Available on-line at pacts on beaches, dunes and barrier islands. Coastal Engineer- http://repository.tudelft.nl/view/hydro/uuid%3A81101b08-04b5-4082- ing, 56(11-12):1133-1152. DOI: 10.1016/j.coastaleng.2009.08. 9121-861949c336c9/ 006 Fortunato, A.B., Oliveira, A. (2004). A modeling system for tidally Roland, A.; Zhang, Y.J.; Wang, H.V.; Meng, Y.; Teng, Y-C.; driven long-term morphodynamics. Journal of Hydraulic Re- Maderich, V.; Brovchenko, I.; Dutour-Sikiric, M.; Zanke, U. search, 42(4):426-434. DOI: 10.1080/00221686.2004.9728408 (2012) - A fully coupled 3D wave-current interaction model on Fortunato, A.B.; Oliveira, A. (2007) - Case study: promoting the unstructured grids. Journal of Geophysical Research: Oceans, stability of the Óbidos lagoon inlet. Journal of Hydraulic Engi- 117(C11):C00J33. DOI: 10.1029/2012JC007952 neering, 133(7):816-824. DOI: 10.1061/(ASCE)0733-9429 Shuttleworth, B.; Woidt, A.; Paparella, T.; Herbig, S.; Walker, D. (2007)133:7(816) (2005) - The dynamic behaviour of a river-dominated tidal inlet, river Murray, Australia, Estuarine, Coastal and Shelf Science, Fortunato, A.B.; Nahon, A.; Dodet, G.; Pires, A.R.; Freitas, M.C.; 64(4):645-657. DOI: 10.1016/j.ecss.2005.04.007 Bruneau, N.; Azevedo, A.; Bertin, X.; Benevides, P.; Andrade, C.; Oliveira, A. (2014) - Morphological evolution of an ephem- Soulsby, R., (1997) - Dynamics of Marine Sands: A Manual for eral tidal inlet from opening to closure: the Albufeira inlet, Practical Applications. 249p., Thomas Telford, Wallingford, Portugal. Continental Shelf Research, 73:49-63. DOI: U.K. ISBN: 978-0727725844. 10.1016/j.csr.2013.11.005 Tung, T.T.,Walstra, D.J., Graaff, J. van de and Stive, M. (2009) - Friedrichs, C.T.; Aubrey, D.G. (1988) - Non-linear tidal distortion in Morphological modeling of tidal inlet migration and closure. shallow well-mixed estuaries: a synthesis. Estuarine, Coastal Journal of Coastal Research (ISSN: 0749-0258), SI56:1080– and Shelf Science, 27(5):521-545. DOI: 10.1016/0272- 1084. Coastal Education & Research Foundation, Lisboa, Portu- 7714(88)90082-0 gal. Available on-line at http://www.cerf-jcr.org/images/stories/1080. 1084_T.T.Tung_ICS2009.pdf Hagemann, B.P. (1969) - Development of the western part of the Westhuysen, A. (2012) - Spectral modeling of wave dissipation on Netherlands during the Holocene. Geologie en Mijnbouw, negative current gradients. Coastal Engineering, 68:17–30. 48:373-388. DOI: 10.1016/j.coastaleng.2012.05.001 Oliveira, A.; Fortunato, A.B.; Rego, J.R.L. (2006) - Effect of mor- Zhang, Y.; Witter, R.W.; Priest, G.P. (2011) - Nonlinear tsunami- phological changes on the hydrodynamics and flushing proper- tide interaction in 1964 Prince William Sound tsunami. Ocean ties of the Óbidos lagoon (Portugal). Continental Shelf Re- Modelling, 40(3–4):246–259. DOI: 10.1016/j.ocemod.2011.09. search, 26(8):917-942. DOI: 10.1016/j.csr.2006.02.011 005

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015)

http://www.aprh.pt/rgci/pdf/rgci-530_Larsen.pdf | DOI:10.5894/rgci530

Model of the evolution of mounds placed in the nearshore*

@, Magnus Larson@, a; Hans Hansona

Abstract A one-dimensional mathematical model is presented that describes the cross-shore evolution of a long linear mound of non- cohesive sediment placed in the nearshore and exposed to non-breaking waves. The equation for the transport rate accounts for wave asymmetry and gravity, and the net rate is expressed by reference to an equilibrium profile shape. Simplifications reduce the governing equation for mound evolution to the diffusion equation, for which analytical solutions are available for various initial shapes of the mound. Temporal and spatial dependencies governing mound evolution are obtained from the analytic solutions; for example, a doubling of the wave height implies a certain mound response in 1/8 of the time compared to the original conditions. The governing equation is also solved numerically in order to avoid schematization of the forcing, initial, and boundary conditions. Both the analytical and numerical models are compared with data on mound evolution from several sites around the world. Model predictions agree with trends in measurements of four mounds at widely different sites. An example is given concerning the application of the analytical model for preliminary mound design. The formulation presented also applies to infilling of dredged trenches with lengths much greater than their widths. Keywords: sediment transport, dredging, profile evolution, beach nourishment, mound design

Resumo § Modelação da evolução de bancos de areia na zona costeira A evolução morfodinâmica de um banco de areia existente na zona costeira do oceano antes da zona de rebentação é descrita recorrendo a um modelo matemático unidimensional. As taxas de transporte sedimentar são calculadas relativamente a um perfil de equilíbrio e têm em consideração a assimetria das ondas e o termo gravitacional. A equação de balanço que rege a evolução da morfologia do fundo, sob certas hipóteses, reduz-se a uma equação de difusão que tem solução analítica conhecida para diferentes configurações do banco. A evolução espaço temporal do banco de areia é determinada em função de diferentes parâmetros a partir da solução analítica: por exemplo, verifica-se que a duplicação da altura da onda conduz a uma evolução mais rápida da morfologia do banco (em1/8 do tempo) relativamente à situação de referência. Para condições mais gerais das condições iniciais, das condições de fronteira e dos forçamentos, a equação é resolvida numericamente. Ambas as soluções, analítica e numérica, são comparadas com dados que reportam a evolução de bancos de areia de diferentes locais no mundo. Os resultados obtidos mostram que os modelos descrevem a tendência das observações efetuadas em quatro bancos de areia distintos. Como exemplo de aplicação, o modelo analítico desenvolvido é considerado no projeto pre- liminar de um banco de areia. A formulação apresentada pode também ser aplicada para estudar a evolução da morfologia de escavações resultantes de dragagens com uma configuração em que o seu comprimento é maior do que a sua largura. Keywords: transporte de sedimentos, dragagem, evolução de perfil, alimentação artificial, projeto de banco de areia

@ Corresponding author, to whom correspondence should be addressed. a Lund University, Department of Water Resources Engineering, Box 118, S-221 00 Lund, Sweden. E-mails: Larson [email protected]; Hanson

* Submission: 25 JUN 2014; Peer review: 19 AUG 2014; Revised: 2 MAR 2015; Accepted: 3 MAR 2015; Available on-line: 5 MAR 2015 § Translation of title, abstract and captions by Paulo Silva on behalf of the authors and the Editorial Board

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015)

formula proposed by Madsen (1991, 1993) is employed 1. Introduction and combined with the sand volume conservation Placement of material in the nearshore is an economical equation to yield a diffusion-type equation for the alternative for bypassing of material dredged as part of mound evolution. After further simplification, analyti- inlet channel maintenance and for maintaining beaches. cal solutions are obtained that provide insight for Material placed in the nearshore need not be exactly preliminary design of mounds. The transport formula is compatible with that on the beach, because sorting by also incorporated in a numerical model to solve the full waves and currents will tend to move coarser sediments equations without approximations in a general model of onshore and finer sediments offshore. Different profile evolution. A numerical approach will allow for a strategies exist for nearshore sand placement, including more detailed evaluation of different mound designs, shoreface nourishment (Van Duin et al., 2004; Grunnet although this is not explored in the present study. et al., 2005) and the placement of sand as an offshore A derivation of the sediment transport equation for mound (bar), where the latter method will be describing the transport rate under non-breaking waves, investigated here. Recognizing the positive attributes of accounting for wave asymmetry and gravity, is linear bars for serving as a reservoir of sand in provided first. The transport equation is combined with promoting beach growth, as well as their wave the sand volume conservation equation to yield a dissipation function, reports on nearshore mound diffusion equation from which analytical solutions can performance have been given by, for example, be obtained. These solutions reveal characteristic Zwamborn et al. (1970), McLellan (1990), Otay (1995), parameters that describe the main features of mound Foster et al. (1996), and Barnard et al. (2006). response to local wave forcing. Four applications Although material placed within the nearshore becomes (Silver Strand, California; Cocoa Beach and Perdido part of the littoral system, benefits to the beach have not Key, Florida; and Maunganui Beach, New Zealand) are been well quantified. If a nearshore mound is intended then given, where the diffusion model is employed to to be stationary, it is referred to as stable whereas if it describe temporal mound response using the diffusion designed to move, it is called active (Hands and coefficient as a fitting parameter. An expression for the Allison, 1991; Beck et al., 2012). Here, a mound is diffusion coefficient is derived based on the analysis for regarded as stable if the cross-shore sediment transport these sites. Subsequently, an example illustrating is sufficiently small to induce only negligible changes preliminary design of an offshore mound, using in the mound shape according to a predefined criterion analytical solutions to the diffusion model, is discussed. (for example, over 10 years). Movement of an active The mound evolution is also simulated for the Silver mound can take place as translation of its center-of- Strand and Cocoa Beach cases with an existing mass (across shore and alongshore) and significant numerical model (see Larson and Kraus, 1989), which dispersion or deflation in relief (e.g., Smith and Gailani, was extended to numerically solve the sediment 2005). In the present study it is assumed that the mound transport equation for the mound. By employing the consists of sand, whether it is stable or active. numerical model, schematizations carried out in the Furthermore, the study specifically considers mounds analytical model regarding forcing, initial, and subjected to transport by non-breaking waves, that is, boundary conditions are avoided and a more realistic mounds placed offshore and seaward of the surf zone evolution is obtained, especially regarding the spatially where the cross-shore morphological development is varying sediment diffusion around the mound. dominated by non-breaking wave conditions. Such mounds are alternatively referred to as offshore 2. Equation for sediment transport in the offshore mounds. Since longshore transport gradients are not It is assumed that the sediment transport in the offshore, considered, the analysis presented is most applicable to seaward of the surf zone, is mainly a function of wave mounds that are constructed as long, linear, shore- asymmetry and gravity (compare Niedoroda et al., parallel bars where changes primarily take place across 1995). The wave asymmetry tends to produce net on- the shore. shore transport, whereas gravity contributes to the off- The objective of this study is to derive and validate a shore transport. If there is a balance between these two simple cross-shore model for estimating the response of mechanisms, an equilibrium slope is obtained where the mounds constructed of dredged material, typically of onshore transport due to asymmetry corresponds to the sand, to non-breaking waves. Such a model would be tendency for increased offshore transport due to gravity. useful for preliminary design of offshore mounds as In order to develop a sediment transport equation for well as for estimation as to whether a specific mound is the mound, the bed load formula by Madsen (1991, active or stable. It is assumed that the main mechanisms 1993) is used as a starting point. Thus, the instanta- controlling the sediment transport are wave asymmetry neous bed-load transport rate per unit width (qB) may be and gravity. A transport equation based on the bed-load expressed as,

22 Larson & Hanson (2015)

(1) In order to employ Eq. 7 for calculating the net transport, the variation in shear stress during a wave cycle must be specified to determine Ion and Ioff (see Eqs. 4 and 5). Also, to include the effects of wave asymmetry where s (= ρs/ρ) is the specific gravity of the sediment, g a higher-order wave theory should be employed. In the acceleration due to gravity, d the sediment diameter, β the simplified approach taken here, the shear stress will be local beach slope (= ; positive if the waves propa- directly related to the bottom orbital velocity (ub) ac- gate upslope), φm the friction angle for a moving grain, t time, and ψ the Shields parameter defined as, cording to , where fw is a friction factor. Wave asymmetry will be described by two sinusoidal (2) with different peak values (up for the positive onshore part of the flow when , and un for the offshore where τ is the instantaneous bed shear stress and ρ and ρ b s negative part of the flow when ). Neglecting are densities for sediment and water, respectively. The subscript cr,β refers to incipient conditions for sediment the critical shear stress for incipient motion in Eqs. 4 movement at the slope b. Although Eq. 1 was specifically and 5, the integrals may be solved by replacing the developed for bed load transport, the overall dependence shear stress with the velocity, using the appropriate on the shear stress to the power 3/2 has often been used to amplitudes, yielding the following expression, calculate the total load, which may give the equation derived in the following more general applicability. (8) In order to determine the net transport during a wave where CB (= ) is a coefficient that cycle, the instantaneous transport rate is integrated sepa- depends on the sediment properties. If the asymmetry is rately for the period of onshore transport (qB,on) and off- not too strong, the sum of the integrals in Eq. 8 may be shore transport (qB,off). Taking offshore transport to be positive (x-axis positive pointing offshore), the net trans- approximated by , where . port during a wave cycle becomes qB,net = qB,off - qB,on. For simplicity, uo is taken to be the bottom orbital ve- Thus, the net transport may be expressed as, locity as given by linear wave theory. Thus, Eq. 7 may be written, (3) (9) where AB (= ) is a coefficient that depends on the sediment properties and the integrals are where Kc (= ABCB) is a coefficient that in practice will defined by, be used for calibration of the model. Thus, the net transport rate is proportional to bottom orbital velocity cubed and the deviation from the equilibrium slope. The (4) transport model may be compared to a formula given by Kobayashi (1982) for computing trench infilling by bedload transport on a gently sloping bottom.

(5) 3. Analytical solution to offshore mound response 3.1 General Solution where the shear stress is onshore directed when The sediment transport relationship for the offshore and offshore directed when . At equilibrium (Eq. 9), in combination with the sand volume conservation equation, may for certain conditions be simplified qB,on = qB,off, leading to the following expression for the so that analytical solutions can be obtained for the pro- local equilibrium beach slope (dh /dx), after rearranging e file evolution around the mound. Analytical solutions, Eq. 3: although describing idealized situations, are useful for determining combinations of parameters that govern the (6) characteristic time and space scales of profile response. These quantities can be employed for first-order esti- Using Eq. 6, Eq. 3 may be expressed as after some ma- mates of profile response or for preliminary design of nipulation: offshore mounds. If the transport equation is combined with the sand volume conservation equation and certain (7) simplifications made, a diffusion equation will result for which analytical solutions are available. 23 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015)

Assuming that the bottom orbital velocity is approxi- explicitly solved for simple mound configurations. For mately constant (uo = uoc) in the area of interest, and example, the evolution of a rectangular mound (placed expressing the response of the mound with respect to on top of an existing EBP) is given by the following the equilibrium beach profile (EBP), with solution (Larson et al., 1987), being the height of the mound over the EBP, yield the sediment transport equation: (15)

(10) where Δzo is the initial mound height over the sea bot-

The selection of a suitable value on uoc for field condi- tom, a half the mound width and erf the error function. tions will be discussed in connection with data com- If the initial height of the mound is given with a nega- parison. In order to compute mound response to wave tive sign, the solution will instead describe the filling by action, Eq. 10 is combined with the sand volume con- cross-shore sediment transport of a long trench dug in servation equation given by: the offshore.

(11) 3.2 Characteristic Quantities for Mound Response Leading quantities governing the response of an off- Substituting Eq. 10 into Eq. 11 yields, shore mound or trench under cross-shore sediment transport can be identified by non-dimensionalizing (12) solutions to Eq. 12, providing insight to the governing time and space scales. Also, these quantities allow where: comparison of the performance of different mound designs. The evolution of an offshore mound having an (13) initial width a, will be governed by the non-dimensional 2 time scale t’=εd t/a . Two mounds having the same con- Equation 12 is formally identical to the diffusion equa- figuration but differing in size will display the same tion, and analytical solutions are available that cover a non-dimensional evolution in time, if appropriately large number of initial and boundary conditions. Larson scaled. The control exerted by geometrical parameters et al. (1987) presented several such solutions for the on the evolution of a mound (or trench) can be assessed one-line model of shoreline change, which reduces to by comparing the non-dimensional quantities. For ex- the diffusion equation under certain assumptions. They ample, the maximum non-dimensional height of two discussed previously published and new solutions re- mounds with the same initial geometric shape will be lated to the shoreline evolution resulting from the the same after time t’. Translating this relationship into placement of a beach fill in the nearshore such that the dimensional time yields, shoreline is initially out of equilibrium with the wave climate. Several of the solutions for shoreline change (16) have direct analogies with mounds (or, alternatively, dredged linear trenches) in the offshore. Under the as- sumption that Eq. 12 is valid for describing the cross- where the indices 1 and 2 refer to two different mounds. shore response of a long offshore mound (or a dredged This equation shows that by doubling the width, a offshore trench) placed uniformly alongshore, the solu- mound can withstand four times as long a period of the tions presented by Larson et al. (1987) for various same wave action before experiencing the same relative beach fill configurations are applicable and will de- decrease of the maximum height. The diffusion co- scribe the mound (trench) evolution (see also, Kobaya- efficient εd enters linearly, but inversely, so that a doub- shi, 1982). Thus, the following general solution of the ling of εd causes the time for the mound to experience a one-dimensional diffusion equation (Carslaw and Jae- certain reduction to be halved. In a preliminary design ger, 1959; Crank, 1973) describes the evolution of a situation, this equation is useful for examining the evo- mound (trench) in the offshore, lution of mounds with different geometric characteristics at a particular site (mounds exposed to the same (14) wave climate). Dean (1991) reviews a similar relation for behavior of rectangular beach fill, in which the where f(x) is the initial shape of the mound (trench) and width of the fill has the same functional dependence as ξ a dummy integration variable. This integral may be mound width in controlling evolution of the feature.

24 Larson & Hanson (2015)

By expressing εd in terms of the local wave climate, the of the material moved onshore (see Fig. 1). During this impact of the wave properties can be assessed. By lin- period the wave climate was mild, and no major storms ear wave theory, were recorded. Thus, these data constitute an excellent

set for testing the analytical model (Eq. 14) developed (17) to predict mound evolution in the offshore under non- breaking waves. All profiles shown here were measured along Survey Line 5 (extending across the central por- where L is the wavelength, H the wave height, and T the tion of the mound; see Larson and Kraus, 1992), where wave period. Again, comparing two cases and equating longshore perturbations were judged to be the smallest. the non-dimensional time t’ gives: The median grain size d50 of the placed material was 0.20 mm. (18)

In the case of shallow water, Eq. 18 may be further simplified to yield:

(19)

Equation 19 illustrates the influence of the wave height and the water depth for comparing the evolution of two mounds of identical initial shape. For example, doubling the characteristic wave height causes a mound to respond in 1/8 of the time as compared to the original conditions. Similarly, placement of an offshore mound in deeper water increases the response time as depth to the 3/2 power, as compared to a base condition. 3.3 Comparison with Field Data Comparisons of predictions were first made to measurements from two field sites where offshore mounds were placed. Because detailed information on the forcing could be obtained for these sites, simulations with Figure 1 - Comparison of measurements and analytical solu- a numerical profile evolution model were also carried tion, Silver Strand, CA out (discussed further below) to examine predicted Figura 1 - Comparação entre as medições e a solução ana- mound response without introducing the simplifications lítica, Silver Strand, CA of the analytical model. In the following, short descrip- An optimum value for the diffusion coefficient, εd=15 tions of the two data sets are given; additional informa- m2/day, was determined through visual fitting against tion is given together with the presentation of the nu- the measured profiles. Fig. 1 illustrates the agreement merical simulation results. between the measurements and analytical solution ob- Measurements of the profile through time made at a tained by superimposing initially trapezoidal line seg- mound off Silver Strand State Park (Andrassy, 1991; ments as discussed by Larson et al. (1987). The surveys Larson and Kraus, 1992) were evaluated with the diffu- were carried out approximately 27, 55, and 119 days sion model. An EBP was determined in accordance after the post-construction survey (used as the initial with Larson et al. (1999) and subtracted from the pro- profile here). As seen in Fig. 1, the analytical solution file surveys to isolate the mound evolution. Wave produces symmetric diffusion of the mound, the result measurements were carried out between January and of specifying a constant diffusion coefficient (i.e., uoc May 1989 during which four surveys were taken constant). In a numerical approach, as shown below, εd (890119, 890215, 890315, and 890518, in YYMMDD can be made a function of water depth, thereby produc- format). These measurements were employed to derive ing more rapid diffusion in shallow water that better a statistically representative value on the forcing pa- describes the skewed shape and onshore migration of rameter included in the diffusion coefficient used in the the mound. However, despite various simplifications analytical model (see next section). The January survey (e.g., schematization of initial, boundary, and forcing was made just after construction of the offshore mound conditions) the analytical solution captures the overall was completed. Subsequent surveys revealed that most response of the mound, and it can be applied to obtain 25 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015) estimates of quantities such as the decrease in the Beach, two other data sets were identified for analysis maximum mound height and reduction in mound vol- of εd. These two mounds were located at Maunganui ume, within the original boundaries of the mound. Beach off the coast of New Zealand (Foster et al., Cocoa Beach near Cape Canaveral served as a benefi- 1996) and at Perdido Key, Florida (Otay, 1995; Work cial-use site for dredged material on three occasions and Otay, 1996). Analysis of data from these two sites 2 between June 1992 and June 1994. The first placement produced εd = 25 and 1 m /day for Maunganui Beach was carried out in June 1992 in the northern half of the and Perdido Key, respectively. Thus, fortunately, a authorized site (approximately between survey lines 0 wide range of εd-values was obtained in the analysis, and +3500), whereas the second and third placements corresponding to a variety of conditions at the different were conducted over longer time periods and broader sites. areas. Data pertaining to the first disposal were considered here for further validation of the diffusion model. One survey was made immediately after construction of the mound followed by two surveys 136 and 291 days after the mound placement. An EBP was determined in accordance with Larson et al. (1999) and subtracted from the surveys to isolate the mound response. Evolu- tion of Survey line 1,500 located in the central portion of the mound was analyzed, and the median grain size of the fill material was 0.14 mm. No wave measurements were carried out in connection with the profile surveying, but a wave hindcast showed that the mound was mainly exposed to non-breaking waves during the measurement period. Fig. 2 indicates the agreement between the analytical model of mound evolution and the measured profile. A diffusion coefficient value of 8 m2/day produced satisfactory description of the mound response. As for the Silver Strand mound, the analytical model predicted Figure 2 - Comparison of measurements and analytical solution, Cocoa Beach, FL some seaward diffusion not observed in the measure- Figura 2 - Comparação entre as medições e a solução ana- ments, attributed to overestimation of εd in this region. lítica, Cocoa Beach, Florida Because the evolution of the mound is described with respect to an equilibrium profile that is monotonically Representative wave quantities at Maunganui increasing (depth) with distance offshore, the predicted Beach were estimated based on various data sources diffusion of the mound at its seaward side will always which reported results of wave measurements off the be over-predicted. The greater the vertical difference New Zealand East Coast (see Table 1 in Foster et al., between the elevation of the equilibrium profile at the 1996; straight-forward averaging was employed to ob- shoreward and seaward side of the mound, the larger tain the representative wave quantities). However, Fos- this effect will be. However, the overall evolution of the ter et al. (1996) pointed out that the data records did not mound is well described by the analytical solution, cre- contain many storms implying that the wave height ating confidence in the simple diffusion model for first might have been somewhat underestimated. At Perdido estimates of how the mounds placed in the offshore Key, wave measurements were available from two respond to the action of non-breaking waves. wave gages, where the longest record encompassed 4

years. Values employed here were computed from 3.4 Dependence of Mound Diffusion on Wave Condi- mean wave quantities reported by Otay (1995). The tions mean significant wave height and mean wave period Ta In order to apply the analytical model for preliminary obtained at the different sites were input to calculate the design of offshore mounds, it is necessary to estimate significant wave height Hs,m and associated bottom or- the diffusion coefficient. Although relative comparisons bital velocity uoc,m at a location corresponding to the can be made based on the characteristic quantities pre- initial maximum mound height (having the associated sented (e.g., Eqs. 18 and 19), it is of practical value to water depth hm). Since the bottom orbital velocity to the quantify the absolute evolution of a mound. However, power 3 is governing the sediment diffusion (see Eq. there are few data sets on mound evolution suitable for 13), it is probably better to work with a representative determining εd. In addition to Silver Strand and Cocoa wave height based on when calculating uoc,m (see

26 Larson & Hanson (2015)

Walstra et al., 2013); however, for applications at sites point for Silver Strand from the overall trend of the with limited data it is more likely that is available points in Fig. 3. compared to . Thus, this statistical measure was employed here. Thus, all wave quantities employed in this study were measured or hindcast for the period when the profile surveys were carried out, except Maunganui Beach which relied on more general estimates of the wave conditions. Table 1 summarizes the environmental conditions and sediment characteristics at the four different sites. Table 1 - Environmental conditions for the four sites (subscript s denotes significant wave height and subscript m denotes quantities taken at a water depth corresponding to the peak of the initial mound). Tabela 1 - Condições hidrodinâmicas e características dos sedimentos nos quatro bancos de areia (o subscrito s refere-se à altura significativa da onda e o subscrito m refere o valor das grandezas na profundidade do cume do banco de areia inicial).

Site Hs,m Ta hm uoc,m d50 (m) (sec) (m) (m/sec) (mm) Figure 3 - Diffusion coefficient for different mounds Silver 0.78 8.7 3.8 0.59 0.20 Strand Figura 3 - Coeficiente de difusão, εd, para diferentes bancos

Cocoa 1.26 8.8 4.7 0.84 0.14 Beach 4. Preliminary design of nearshore mounds Preliminary design of mounds can be carried out based Maunganui 1.33 9.0 3.5 1.05 0.29 Beach upon the analytic solutions, by which key parameters controlling mound response can be estimated. The ana- Perdido 0.55 6.3 3.8 0.39 0.30 Key lytical solutions were obtained through simplifications, and the limitations of these solutions should be realized. However, reasonable results were achieved for the field Equation 13 gives a theoretical relationship for the sites investigated; and the solutions should provide dependence of εd on uoc, and Fig. 3 shows this relation- acceptable first estimates of the mound response if ship plotted for the analyzed data sets. Although the cross-shore sand transport under non-breaking waves is scatter is considerable, the clear trend indicates that Eq. the dominant transporting mechanism. Also, the pre- 13 based on the mean local wave conditions may yield vailing wave and sediment conditions should not devi- predictions on εd that exhibit its main behavior, al- ate too much from the field cases summarized in Table though the equation should be used with care and pre- 1. For detailed analysis and design of offshore mounds ferably a range of values on εd employed to establish a numerical approach should be taken. the variability in mound response. The bottom orbital The solution presented by Larson et al. (1987) for a velocity employed was calculated from the mean sig- collection of line segments can be applied for any initial nificant wave height at the peak of the initial mound mound shape. Here, only the example of an initially during the measurement period. A least-square fit of triangular mound will be discussed (the solution for a Eq. 13 to the data points yielded Kc=0.0024. rectangular mound is given by Eq. 15). Fig. 4 illustrates Efforts were made to estimate Kc individually for the time evolution of the non-dimensional maximum each case and to relate these Kc-values to various non- mound height and non-dimensional mound volume for dimensional parameters including grain size, but no a triangular mound. Height and volume were normal- clear relationship could be established. Presently avail- ized with their values at time t=0, and the volume ex- able data are limited and do not support adoption of presses the amount of material within the original expressions for εd that are more comprehensive than Eq. boundaries of the mound, between x= -a and a. With 8. The mound at Silver Strand was placed on top of a knowledge of the typical wave climate (mean signifi- natural bar, whereas the other mounds were placed fur- cant wave height and period) and the dimensions of the ther offshore where the profile depth was monotonically mound (height and width), Fig. 4 can be entered to es- increasing with distance offshore. This circumstance timate the height and volume after a certain time. Simi- may have contributed to the apparent deviation of data larly, the water depth of placement can be optimized to 27 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015) achieve a specified spread of material (i.e., volume As an example, consider a typical U.S. East Coast wave reduction) by reference to this figure. To simplify esti- climate with an average deep-water significant wave mation of the diffusion coefficient when applying Fig. height of 1 m and average wave period of 8 sec. Placing 4, Fig. 5 was constructed by converting to deep-water an initially triangular mound with the peak in h=4 m 2 wave height Ho and wavelength Lo, and neglecting re- water depth gives h/Lo=0.04 and εdT/Ho Lo/Ho=0.087 fraction. Thus, from the wave conditions in deep water, from Fig. 5. Thus, εd is calculated to be approximately -4 2 2 εd may be estimated at any water depth through Fig. 5. 1.1⋅10 m /sec (=9.5 m /day). Assuming an initial mound width of 100 m (a=50 m), mound response after 1 month can be determined from Fig. 4. The non- dimensional time is given by t’=9.5 30/502=0.11. From Fig. 4 it is seen that the remaining volume (ΔV/ΔVo, where ΔV is the mound volume and subscript o denotes the initial conditions) is about 90% of the volume placed originally, and the maximum height is about 60% of the initial height (Δz/Δzo). Fig. 4 is valid for an initially triangular mound. Other mound shapes (e.g., rectangular) would display somewhat different evolution, especially regarding the decrease in Δz. However, the evolution of the remaining volume, being an integrated quantity, is less sensitive to the initial mound shape. The analytical solutions describing the time response of the mound can be applied to design both stable and active mounds, where the stability (or activity) of the mound should be defined in terms of changes in geometric mound properties over certain time scales. Finally, it is again noted that the solutions presented are equally applicable for trenches (“nega-

tive” mounds) if the basic mechanisms controlling the Figure 4 - Evolution of relative height and volume of initially triangular mound sand transport and bathymetric response are the same. Figura 4 - Evolução da altura relativa e do volume de um 5. Numerical simulation of mound response banco de areia triangular. 5.1 Modified SBEACH Model To investigate the sediment transport model for mound evolution in the offshore without the approximations necessary to arrive at analytical solutions, the model was incorporated in an existing numerical model of profile evolution and dune erosion (SBEACH; see Lar- son and Kraus, 1989). The original SBEACH model does not include the effects of wave asymmetry and gravity in the offshore, but the sediment transport in this region is primarily a function of seaward diffusion of sediment from the surf zone that is mainly characteristic for offshore transport and profile erosion. The main advantage of using a numerical approach is that the initial, boundary, and forcing conditions can be made arbitrary allowing for more realistic simulations (i.e., the diffusion coefficient varies in time and space). Also, the transport of material in the surf and swash zone could be incorporated, although the former was of minor importance for the transport on the mound in the cases investigated. Thus, the overall objective of the Figure 5 - Diffusion coefficient as function of normalized numerical simulations was to assess the applicability of water depth the sediment transport model under more general condi- Figura 5 - Valores do coeficiente de difusão em função da tions without limitations in characterizing the forcing or profundidade normalizada profile configuration. 28 Larson & Hanson (2015)

Two field sites (Silver Strand, CA, and Cocoa Beach, brated to 0.03 by visually minimizing the difference FL) where the response of offshore mounds to waves between the final measured and calculated profile. was monitored in detail were available for calibration Overall the model prediction was satisfactory, although and validation of the sediment transport model the accumulation above mean sea level was not well developed for the offshore. However, the equation was described (not the focus of this study, but the entire first generalized to random waves through a wave-by- profile shown since it is included in the simulation). wave calculation algorithm (Dally, 1992), resulting in This is likely a result of the model’s limited ability to expressions for the mean transport rate analogous to simulate accumulation in the swash zone. However, the what Larson (1996) obtained for the surf and swash actual profile probably also experienced some changes zone. The sediment transport routine was then included associated with longshore transport, not included in the in SBEACH to calculate the necessary hydrodynamic present simulations. The diffusion of the mound, with quantities and to predict cross-shore sediment transport most of the material moving onshore, was correctly rates in the surf zone (if necessary). Background reproduced similarly to the analytical model, but the information for the simulations are given in the possibility to vary the diffusion coefficient in time and following as well as summaries of the simulation space improved the simulated profile evolution with the results. numerical model. Employing a transport equation in the form given by Eq. 9 requires predictive expressions for how he varies with wave and sediment properties. For the surf zone, Dean (1977, 1987) proposed relationships based on median grain size or sediment fall speed (see also Bruun, 1954). Larson et al. (1999) developed a predictive equation for the profile shape under non- breaking waves, where the depth at breaking hb constituted the main parameter in calculating the shape (hb is the shoreward boundary for the portion of the profile exposed to non-breaking waves). This equation was adopted in the present simulations to derive the local EBP slope at every time step (used in Eq. 9).

5.2 Silver Strand, California The numerical model was calibrated and validated with measurements for the mound placed off Silver Strand State Park (Andrassy, 1991; Larson and Kraus, 1992). As previously described, four surveys were taken between January and May 1989, together with wave Figure 6 - Calculated and measured profiles at Silver Strand measurements. After placement, the mound diffused after 4 months and most material moved onshore. Measured significant wave height, mean wave period, and mean incident Figura 6 - Comparação entre o perfil simulado e o observado em Silver Strand após 4 meses wave angle were available about every three hours (in 10.9 m water depth) for 114 days. The water level was To validate the model, comparisons were made with the not recorded, but an hourly time series of tidal eleva- intermediately measured profiles, and Figs. 7 and 8 tions was generated with a numerical model (DRP, illustrate the results for surveys made at 890215 and 1994). The simulations started with the measured pro- 890315, respectively. Agreement is judged to be good, file at 890119, and comparisons were made between especially regarding the overall mound shape. For the calculated and measured profiles for the other three survey made 890215, the trough seaward of the mound surveys. The only calibration coefficient was Kc, all was more pronounced in the measurements, implying other coefficients in the model were held constant in the that the model fills up the trough somewhat too quickly. simulations at their default values as recommended in the SBEACH manual (Rosati et al., 1993). The time 5.3 Cocoa Beach, Florida step was 20 min, and grid cell size was 10 m. The me- The second case investigated with the numerical model dian grain size was set at 0.20 mm, in accordance with was the mound placement at Cocoa Beach, previously field samples. simulated for the analytical model validation. The pro- Fig. 6 displays the calculated profile after the entire file surveys taken off Cocoa Beach only included an simulation period (890518) together with the measured area that started in about 4 m water depth and extended initial and final profile. The coefficient Kc was cali- to a little more than 8 m. Trial simulations were carried 29 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015)

selected initial shape surely oversimplified the real profile configuration; however, after some simulation time the profile took on a more realistic shape. Also, the inshore shape did not markedly affect the response of the mound, although there was clear shoreward trans-

port of material from the mound to shallower water. A time period of about 4.5 months (920728 to 921211) was selected for the simulations where the initial profile corresponded to the post-disposal profile after the first mound placement. No wave or water level measurements were available, so numerical hindcasts were utilized. WIS wave information available every 3 hr (10-m water depth) was combined with calculated hourly tidal elevations (DRP, 1994). It should be noted that the use of WIS data might have overestimated the wave conditions since the effect of bottom friction in the propagation from offshore to the end of the grid was not included (Dally and Osiecki, 2006). Wave angle was not included in the simulations, and the time and

Figure 7 - Calculated and measured profiles at Silver Strand length steps were 20 min and 10 m. A representative after 1 month median grain size of 0.15 mm was employed in all simulations based on sediment sampling. The calculated Figura 7 - Comparação entre o perfil simulado e o observado em Silver Strand após 1 mês waves across the profile demonstrated that only few waves broke in the vicinity of the mound and that the

transport in this area due to the breaking waves was small.

First, the Kc-value from the Silver Strand simulations was input. However, the predictions produced onshore transport rates that seemed to yield excessively large accumulation in the inshore portion of the profile as well as a diffusion of the mound that exceeded the measurements. Thus, Kc was re-calibrated, and a value of 0.004 was obtained as optimal for reproducing the measurements. As expected, the difference in Kc-value between Cocoa Beach and Silver Strand value agreed with the results from fitting the analytical model to the field data, which also produced a larger value on the diffusion coefficient for Silver Strand that deviated from the other studied sites. The Kc-value obtained for the Cocoa Beach simulations with the numerical model was in agreement with the overall estimate from the analytical model, indicating that this value is most likely more representative than the Kc-value from the Silver Strand simulations.

Fig. 9 illustrates the initial profile together with the Figure 8 - Calculated and measured profiles at Silver Strand after 2 months final calculated and measured profiles, and Fig. 10 shows a blow-up for the portion of the profile were the Figura 8 - Comparação entre o perfil simulado e o measurements were carried out. The calculated inshore observado em Silver Strand após 2 meses portion of the profile displays a shape typical of a com- out for the surveyed area only (neglecting the inshore posite EBP subject to random waves. A concave shape portion of the profile), but it proved difficult to formu- appears in the region where breaking waves dominate, late the shoreward boundary condition for such a situa- followed by a change in the curvature going seaward as tion and significant pile-up of material occurred at this the bar region is approached and where non-breaking boundary. Thus, a hypothetical profile was constructed waves start to control the profile shape. Further seaward shoreward of the survey area by linearly extrapolating the profile again attains a concave shape, although in the most inshore portion of the measured profile. The the present case the mound perturbs the profile here. 30 Larson & Hanson (2015)

values had been selected on a coefficient (Kc) characterizing the mound diffusion. The mound diffusion was

related to and Kc appeared as a multiplier in the expression for the diffusion coefficient. For three of the four field cases studied, Kc attained quite similar values; however, for the Silver Strand case Kc exhibited a significantly larger value. One possible reason for this deviation was that the mound at Silver Strand was placed in an area (on top of an existing bar) with more active sand transport, with breaking waves occasionally affecting the transport for the period of study. Overall, the relationship derived for the diffusion coefficient may be used for scoping-mode predictions of the mound response, if the mound is placed outside the surf zone. In such predictions it may be useful to investigate the mound response for a range of Kc-values to establish the sensitivity of the response to this coefficient. The numerical simulations reproduced the mound evo- Figure 9 - Calculated and measured profiles at Cocoa Beach lution satisfactorily, although there was some discre- after 4,5 months pancy in the trough area directly shoreward of the Figura 9 - Comparação entre o perfil simulado e o mound. The optimum Kc-value for Cocoa Beach was observado em Cocoa Beach após 4.5 meses significantly lower than the corresponding value found for Silver Strand. This value is in agreement with ob-

servations made in fitting the analytical model to the field data, indicating that a value of Kc=0.004 should be representative for the transport in the offshore. Again, one reason for obtaining a larger value for the Silver Strand data might be that the material was placed on top of an existing bar, where the transport activity is expected to be higher than at locations further offshore. Other reasons for the discrepancy in Kc-values between the two sites investigated might be: (1) differences in wave climate and sediment characteristics between the two sites, and (2) waves were measured at Silver Strand, but hindcasted at Cocoa Beach. More comparisons with laboratory or field data are needed before reliable values for Kc can be established and the dependence of Kc on different environmental factors re- solved. The difference between the analytical and numerical model in predicting the mound response for Silver Strand and Cocoa Beach was not markedly large. The constant diffusion coefficient employed in the analyti- Figure 10 - Calculated and measured profiles at Cocoa Beach cal model produced a symmetric evolution of the after 4.5 months (blow-up of Fig. 9) mound that was not in agreement with the observations. Figura 10 - Comparação entre o perfil simulado e o A numerical model is able to describe a varying diffu- observado em Cocoa Beach após 4,5 meses (pormenor da sion coefficient since the boundary and forcing condi- Figura 9) tions can be arbitrary. Also, the numerical model will include the entire profile that is typically of interest in 6. Discussion of simulation results more detailed studies, not only the mound region. The analytical model was based on marked simplifica- 7. Conclusions tions of the governing processes and how the initial, boundary, and forcing conditions were employed in the A one-dimensional mathematical model was developed model. In spite of this, the evolution of the mounds for calculating the time-averaged net cross-shore trans- investigated could be well reproduced after appropriate port rate and evolution of mounds placed where non-

31 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):21-33 (2015) breaking waves prevail. At the mound, the governing References equation for the cross-shore transport rate balances Andrassy, C.J. (1991) - Monitoring of a nearshore disposal mound contributions by wave asymmetry and gravity, and the at Silver Strand State Park. Proceedings Coastal Sediments ‘91, rate of change of bottom elevation is controlled or ASCE, 1970-1984. bounded by reference to an equilibrium profile. The Barnard, P.L.; Hanes, D.M.; Lescinski, J.; Elias, E. (2007) - Moni- resultant equation can be reduced to the diffusion equa- toring and modeling nearshore dredge disposal for indirect beach nourishment, Ocean Beach, San Francisco. Proceedings tion for which many analytical solutions exist, if the 30th Coastal Engineering Conference, ASCE, 4192-4204. initial, boundary, and forcing conditions can be sche- Beck, T.M.; Rosati, J.D.; Rosati, J. (2012) - An update on nearshore matized. Such solutions yield characteristic parameters berms in the Corps of Engineers: Recent projects and future controlling the evolution of a mound (or trench) sub- needs. CHETN-XIV-10, Coastal and Hydraulics Laboratory, U.S. jected to wave action. The analytical model can be used Army Engineer Research and Development Center, Vicksburg, MS, U.S.A. Available on-line at http://www.dtic.mil/cgi- to investigate the influence of various design and envi- bin/GetTRDoc?AD=ADA559368 ronmental parameters. For example, it can easily be Bruun, P. (1954) - Coast erosion and the development of beach shown that the time scale of change is directly propor- profiles. Technical Memorandum No. 44, Beach Erosion Board, tional to the square of mound width, inversely propor- U.S. Army Corps of Engineers. U.S.A. tional to wave height cubed, and directly proportional to Carslaw, H.; Jaeger, J. (1959) - Conduction of heat in solids. 510p., the water depth to the three-half power. The analytical Clarendon Press, Oxford, U.K. model was validated with field data on mound evolu- Crank, J. (1973) - The Mathematics of Diffusion. 2nd ed., Clarendon Press, Oxford. Available on-line at http://www.ees.nmt.edu/outside/ tion from different sites. courses/hyd510/PDFs/SupplementaryReadings/Crank_1975_Diffusion.p The model can also be solved numerically to account df for local changes in forcing and bathymetry. This was Dally, W.R. (1992) - Random breaking waves: Field verification of done for two of the field sites using a modified version a wave-by-wave algorithm for engineering application. Coastal Engineering, 16(4):369-397. DOI: 10.1016/0378-3839(92) of SBEACH where the transport in the offshore was 90060-8 included. The simulations results with SBEACH im- Dally, W.R.; Osiecki, D.A. (2006) - Development & validation of proved compared to the analytical model, especially on hindcast-driven nearshore wave information. 9th International the seaward side of the mound. Also, in the numerical Workshop on Wave Hindcasting and Forecasting, Victoria, approach the entire beach profile is included, which Canada. yields more complete information on the profile re- Dean, R.G. (1977) - Equilibrium beach profiles: U.S. Atlantic and sponse required in more detailed studies. Gulf coasts. Department of Civil Engineering, Ocean Engineer- ing Report No. 12, Univ. of Delaware, Newark, DE. A key parameter of the analytical model, the diffusion Dean, R.G. (1987) - Coastal sediment processes: Toward engineering coefficient, was determined from four data sets that solutions. Proceedings of Coastal Sediments '87, ASCE, 1-24. span a wide range of wave conditions. Based on these Dean, R.G. (1991) - Principles of beach nourishment. In: Komar, limited data, an empirical formula was developed for P.D. (editor), Handbook of Coastal Processes and Erosion, the diffusion coefficient that should be useful in scop- pp.217-231, CRC Press, Boca Raton, FL, U.S.A. ing-mode studies of mound analysis and design. The DRP (1994) - Tidal constituents database – East Coast, Gulf of transport coefficient in the SBEACH model, corres- Mexico, and Carribean Sea. Dredging Research Technical Notes, DRP-1-13. Coastal Engineering Research Center, U.S. ponding to the diffusion coefficient in the analytical Army Engineer Waterways Experiment Station, Vicksburg, MS. model, gave values consistent with the latter coefficient. Foster, G.A.; Healy, T.R.; De Lange, W.P. (1996) - Presaging beach It is concluded that the analytical model has applic- nourishment from a nearshore dredge dump mound, Mt. Maun- ability for preliminary design in determining the time ganui Beach, New Zealand. Journal of Coastal Research, scale and movement of material placed in the form of 12(2):395-405. offshore linear mounds and that the numerical model Grunnet, N.M.; Ruessink, B.G.; Walstra, D.J.R. (2005) - The influ- can be used for more detailed investigations of mound ence of tides, wind and waves on the redistribution of nourished sediment at Terschelling, The Netherlands. Coastal Engineer- response. ing, 52(7):617-631. DOI: 10.1016/j.coastaleng.2005.04.001 Hands, E.B.; Allison, M.C. (1991) - Mound migration in deeper Acknowledgments water and methods of characterizing active and stable berms. Proceedings Coastal Sediments ’91, pp.195-1999, ASCE, This work was partially funded under the Regional Sediment U.S.A. Management Program of the U.S. Army Engineer Research Kobayashi, N. (1982) - Sediment transport on a gentle slope due to and Development Center, and partially by the Swedish research waves. Journal of Waterway, Port, Coastal and Ocean Engi- council Formas (Project No. 214-2009-802: The impact of neering, 108(WW3):254-271. climate change on coastal flooding and erosion: processes, Larson, M. (1996) - Model of beach profile change under random modeling, and measures). Mr. Randall A. Wise provided the waves. Journal of Waterway, Port, Coastal, and Ocean Engi- beach profile data for Cocoa Beach, FL. The review comments neering, 122(4):172-181. by Dirk-Jan Walstra and Mohamed Dabees helped to greatly Larson, M.; Kraus, N.C. (1989) - SBEACH: Numerical Model for improve the paper and are much appreciated. Simulating Storm-Induced Beach Change. Report 1: Empirical

32 Larson & Hanson (2015)

Foundation and Model Development. Technical Report CERC- Otay, E.N. (1995) - Monitoring results of a nearshore disposal berm. 89-9, Coastal Engineering Research Center, U.S. Army Engi- Proceedings Coastal Dynamics ’95, pp.546-558, ASCE, U.S.A. neer Waterways Experiment Station, Vicksburg, MS, U.S.A. Rosati, J.D.; Wise, R.A.; Kraus, N.C.; Larson, M. (1993) - Larson, M.; Kraus, N.C. (1992) - Analysis of cross-shore movement SBEACH: Numerical model for simulating storm-induced beach of natural longshore bars and material placed to create long- change. Report 3: User's manual. Instruction Report CERC-93- shore bars. Technical Report DRP-92-5, U.S. Army Engineer 2, Coastal Engineering Research Center, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, U.S.A. Waterways Experiment Station, Vicksburg, MS, U.S.A. Larson, M.; Kraus, N.C.; Hanson, H. (1987) - Analytical solutions of Smith, E.R.; Gailani, J.Z. (2005) - Nearshore placed mound physi- the one-line model of shoreline change. Technical Report CERC- cal model experiment. DOER Technical Notes Collection 87-15, U.S. Army Engineer Waterways Experiment Station, (ERDC TN-DOER-D3), U.S. Army Engineer Research and De- Vicksburg, MS, U.S.A. velopment Center, Vicksburg, MS, U.S.A. Larson, M.; Kraus, N.C.; Wise, R.A. (1999) - Equilibrium beach pro- Van Duin, M.J.P.; Wiersma, N.R.; Walstra, D.J.R: Van Rijn, L.C.; files under breaking and non-breaking waves. Coastal Engineer- Stive, M.J.F. (2004) - Nourishing the shoreface: observations ing, 36(1):59-85. DOI: 10.1016/S0378-3839(98)00049-0 and hindcasting of the Egmond case, The Netherlands. Coastal Madsen, O.S., (1991) - Mechanics of cohesionless sediment transport Engineering, 51(8-9):813-837. DOI: in coastal waters. Proceedings of Coastal Sediments '91, pp.15-27, 10.1016/j.coastaleng.2004.07.011 ASCE, U.S.A. Walstra, D.J.R.; Hoekstra, R.; Tonnon, P.K.; Ruessink, B.G. (2013) Madsen, O.S. (1993) - Sediment transport on the shelf. Sediment - Input reduction for long-term morphodynamic simulations in Transport Workshop DRP TA1, Coastal Engineering Research wave-dominated coastal settings. Coastal Engineering, 77, 57- Center, Vicksburg, MS, U.S.A. 70. DOI: 10.1016/j.coastaleng.2013.02.001 Work, P.A.; Otay, E.N. (1996) - Influence of nearshore berm on McLellan, T.N. (1990) - Nearshore mound construction using th dredged material. Journal of Coastal Research, SI7:99-107, Fort beach nourishment. Proceedings 25 Coastal Engineering Con- Lauderdale, FL, U.S.A. Article Stable URL: http://www.jstor.org/ ference, pp.3722-3735, ASCE, U.S.A. stable/25735392 Zwamborn, J.A.; Fromme, G.A.W.; FitzPatrick, J.B. (1970) - Under- Niedoroda, A.W.; Reed, C.W.; Swift, D.J.P; Arato, H.; Hoyanagi, K. water mound for the protection of Durban’s beaches. Proceedings th (1995) - Modeling shore-normal large-scale coastal evolution. Ma- 12 Coastal Engineering Conference, pp.975-994, ASCE, U.S.A. rine Geology, 126(1-4):181-199. DOI: 10.1016/0025- 3227(95)98961-7

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

http://www.aprh.pt/rgci/pdf/rgci-541_Lisboa.pdf | DOI:10.5894/rgci541

Anthropogenic influence on the sedimentary dynamics of a sand spit bar, * Patos Lagoon Estuary, RS, Brazil

@, Paulo Victor Lisboa@, a; Elisa Helena Fernandesa

Abstract Known to be transition zones between the sea and the mainland, estuaries are recognized for their socio-economic and ecological importance as well as their complexity and vulnerability to anthropogenic influences, harbor activities included. It is reasonable to consider that these influences produce changes in the dynamics of the region, resulting in modifications related to the biota, sediment transport, and flows between the estuary and the adjacent coastal zone. The possible changes in the sediment budget of the Patos Lagoon estuary can affect fragile features present in the estuary such as sand spits. The objective of this work is to evaluate the effect of the modernization work carried out at Rio Grande Harbor access channel on a specific sand spit called Pontal Sul as an indicator of the sediment dynamics alteration in the Patos Lagoon estuary. The study is based on numerical modeling experiments with the TELEMAC-3D model and remote sensing techniques (Quickbird images). Re- sults indicate that the Rio Grande Harbor modernization work changed the Pontal Sul erosion and deposition rates and longitudinal growth. The results suggest that the modification carried out at Rio Grande Harbor access channel changed the Patos Lagoon sediment dynamics. Key Words: Anthropic influence, erosion, deposition, harbors, numerical modeling.

Resumo Influência antropogênica sobre a dinâmica sedimentar de um pontal, estuário da Lagoa dos Patos, RS, Brasil Conhecidos por serem zonas de transição entre o mar e o continente, os estuários são ambientes referidos pela sua impor- tância sócio–econômica e ecológica, além de sua complexidade e vulnerabilidade frente a influência antrópica, onde se destacam as atividades portuárias. É razoável considerar que estas influências produzam modificações nos padrões ambientais da região, podendo resultar em modificações que interferem desde a biota até o transporte de sedimentos, e nos fluxos entre a região estuarina e a zona costeira adjacente. Essa possível modificação no balanço sedimentar do Estuário da Laguna dos Patos, pode se refletir em frágeis feições presentes no estuário, como os pontais. O objetivo desse trabalho é avaliar o efeito das obras de modernização do Canal de Acesso ao Porto do Rio Grande sobre a dinâmica sedimentar do Pontal Sul, com base em técnicas de modelagem numérica e geoprocessamento. Também foi analisada a variação das áreas (erodidas e acrescidas) e da linha de costa do referido pontal. O estudo foi realizado utilizando imagens orbitais e o modelo numérico tridimensional TELEMAC-3D. Os resultados indicam que as obras de modernização do Porto do Rio Grande, alteraram as taxas de erosão e deposição no Pontal Sul, além de evidenciar uma tendência de crescimento longitudinal do pontal. Estes resultados são um indicativo do impacto da referida obra de modernização na dinâmica sedimentar do estuário da Lagoa dos Patos. Palavras chave: Influência antrópica, erosão, deposição, portos, modelagem numérica

@ Corresponding author, to whom correspondence should be addressed: a Universidade Federal do Rio Grande (FURG) – Instituto Oceanográfico – IO – Laboratório de Oceanografia Física Costeira e Estuarina, Av. Itália s/ n°, CEP 96200-000

* Submission: 29 JUL 2014; Peer review: 27 AUG 2014; Revised: 14 FEB 2015; Accepted: 5 MAR 2015; Available on-line: 6 MAR 2015

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

1. Introduction Harbor, financial resources were provided by the Bra- zilian Government to improve the harbor draft and Known as transition zones between the sea and the main- safety navigation conditions. Both jetties located at the land, estuaries are recognized for their socio-economic Patos Lagoon mouth (Figure 1) were extended (370 and and ecological importance as they receive nutrient 700 m) and made convergent; also, the access channel contributions from the continent and the ocean, stimulat- was made progressively deeper with maximum depth of ing primary production and profitable fisheries (Mi- 18 m outside and minimum depth of 16 m inside the randa et al, 2002). The strategic location of estuarine jetties. This modernization work finished in 2010 and is areas promotes the growth of major urban centers in described in detail by Fernandes et al. (2012). their surroundings, where the development of harbor activities can be highlighted. In this sense, estuaries These investments were of great importance because are susceptive to natural and anthropogenic processes, Rio Grande Harbor is the only seaport with favorable being also subject to a combination of both impacts. geographical location in the South Atlantic, and is con- Brazil has more than 35 harbors along 8500 kilometers nected to all southern Brazil and several Latin Ameri- of coastline. Although they have different features (types can countries. It is expected, however, that alterations in of load, infrastructure and environmental conditions), the estuarine geomorphology will change the existing they share common problems related to siltation, dredg- circulation patterns in the area, affecting the biota, the ing operations and their consequences to the environment sediment transport patterns and the interaction between and water quality. Brazilian harbors are a logistical bot- the estuary and the coastal zone. tleneck for the national economy, limiting the Brazilian Sand Spits are geomorphological features observed in ability to compete internationally. Hence, in order to micro-tidal areas under the influence of bidirectional reduce the Brazilian harbor operational costs, investing wind along the main axis of the lagoon. Inside the Patos in infrastructure is essential. Lagoon, the predominant wind and wave direction and Recently, environmental and commercial demands, the initial shape of the lagoon are considered the most combined with the need of maintaining and expanding important factors for the genesis and evolution of these the current capacity of several Brazilian harbors, have features (Zenkovich, 1967). Near the mouth of the Pa- led to strong investments. Particularly at Rio Grande tos Lagoon estuary, where micro-tidal conditions are

Figura 1 - Área de estudo. Figure 1 - Study area.

36 Lisboa & Fernandes (2015) observed, the Pontal Sul sand spit constitutes a deposi- goon. In opposition, this pattern does not occur during tional feature resulting from the interaction between high river discharge episodes (> 4000 m3 s-1) because continental and oceanic processes (Antiqueira & Cal- the SW wind effect is overwhelmed by the river dis- liari, 2004). Understanding the geomorphological evo- charge, which produces a high pressure gradient to- lution of these features in coastal regions subject to wards the ocean. In this situation, the estuarine zone significant anthropogenic impacts is of great impor- occurs at the adjacent coastal region (Moller & Cas- tance because their rapidly response to changes in the taing, 1999). environment (Antiqueira & Calliari, 2004). Thus, a The tidal signal in the region is mixed (mainly diurnal), study of the local hydrodynamics is crucial for under- having a mean range of 0.47 m (Moller et al., 2001); standing the sediment budget in the region. therefore, characterizing a microtidal system. The tidal Given the context, this paper aims to evaluate the effect effect is of secondary importance for the dynamics of of the modernization work carried out at Rio Grande the system, being restricted to the coastal region and the Harbor access channel on the sedimentary dynamics of estuarine portion of the Patos Lagoon. The access chan- the Pontal Sul sand spit as an indicator of the anthropo- nel acts like a filter damping the tidal signal as it moves genic contribution to the processes shaping the system. up the estuary (Fernandes et al., 2004). 2. Study area The morphological characteristics of the Patos Lagoon show a dominance of coarse sediment in the shallow The Patos Lagoon (Figure 1), located on the southern- areas and fine grains in the deeper ones, especially in most part of Brazil (between 30 °S and 32 °S), is the the estuarine zone. The west margin of the Patos La- largest choked coastal lagoon in the world (Kjerfve, goon is covered by medium to coarse sands (Martins et 1986), being 250 km long and 40 km wide, covering an al., 1987), carried out mainly by the Camaquã River. 2 area of 10.360 km . The lagoon is considered a shallow At the east and north margins, the sediment is com- system with a mean depth of 5 m and higher depths posed of fine sand from the coastal-marine depositional occurring in the estuarine channels. The system is con- system. nected to the South Atlantic Ocean through a narrow The estuarine region, where the Rio Grande Harbor is channel, which is less than 1 km wide and about 18 m located (Figure 1), occupies about 10% of the Patos deep. Lagoon area; characterized by an increase in cross sec- The Lagoon drains a basin of approximately tion with the distance from the mouth (Möller et al, 2 201,626 km , being the Guaíba and Camaquã Rivers the 2001). The Lagoon and the estuary are separated by a main tributaries in the north and central lagoon, respec- morphological step formed by sandbanks situated tively (Figure 1). The hydrographic basin rivers have a around the region called Ponta da Feitoria (Figure 1). midlatitude pattern: high discharge in late winter and Apart from the contribution of the main Lagoon, the early spring followed by low to moderate discharge São Gonçalo Channel is an important source of sedi- through summer and autumn (Moller et al., 2001). The ment to the Patos Lagoon estuary (Hartmann et al., mean annual freshwater contribution in the north of 1990). The sedimentology of the estuarine area is more Patos Lagoon is 2000 m3 s-1; seasonal variations can be diverse than in the lagoon, presenting sand in the shal- observed from 700 m3 s-1 during summer (late Decem- low banks and a combination of silt and clay in the ber–March) up to 3000 m3 s-1 during spring (Septem- deeper areas (Calliari et al., 2009). ber–early December). Moller et al. (1996) observed 3 -1 river discharge peaks of 12,000 and 25,000 m s dur- 3. Materials and methods ing El Niño years. The Patos Lagoon circulation is mainly controlled by a The study of the dynamics of geomorphological fea- combination of setup and setdown driven by local tures, sand spits included, can be carried out based on winds and non-local wind action in the coastal region direct observations in the field, remote sensing and and by the river discharge in the north of lagoon (Fer- application of numerical models. Each of these tech- nandes et al., 2002; 2005; Moller et al., 1996). The niques has advantages and disadvantages and is com- wind action has an important role during low and mod- plementary to each other. This study is based on remote erate river discharge (< 3000 m3 s-1). Moller et al. sensing and numerical modeling techniques. (2001) demonstrated that the NE wind causes a 3.1. Digital processing of images barotropic pressure gradient towards the ocean, resulting in a freshwater discharge onto the continental shelf. Orbital images from the Quickbird satellite sensor In contrast, SW wind causes an increase in the water (Digital Globe), acquired from the IFRS (InstitutoFed- level (Ekman transport) on the coast, causing the estab- eral do Rio Grande do Sul) database, with a panchro- lishment of a barotropic pressure gradient toward the matic spatial resolution of 0.61 m, in three bands of continent, resulting in a flood regime at the Patos La- colors (green, red and blue), for the years of 2004,

37 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

2007, 2009 and 2012 were used to construct the recent torized, one can extract the eroded and/or increased area migratory behavior of the sand spit area. by the generation and subtraction of polygons. Thus, The images were georeferenced using ArcGis®9.3. An from the intersection of two coastlines, a series of poly- orbital image already registered with ground control gons is created so that they represent the regions where points (GCP) was used as base map in the correction accretion and/or erosion process are occurring. By ana- process; afterwards, the images were projected to the lysing the changes in the polygon, it is possible to quan- WGS84 – UTM Zone 22S projection. The choice of tify the variation between two shorelines. Hence areas ground control points corresponds to an important stage where both positive and negative changes occur can be of the work. The accuracy applied to extract the shore- calculated from the decomposition of this complex line is dependent upon the accuracy applied in the iden- polygon into a series of simpler sub-polygons. tification the GCPs. At least 15 control points were Vectorised files (shapefiles) were generated using distributed in the regions close to the sand spit for each ArcMAP 9.3. The vectorised shoreline in the images image. was used to generate and calculate the polygons. To calculate the spit area variation rates, in every interval 3.1.1. Error inherent to the photogrammetric process under analysis, the polygon that resulted from the oldest In order to measure inaccuracies resulting from photo- date was subtracted from the one that resulted from the grammetric process, the standard image geospatial posi- most recent one. As a result, the final area value, in tioning proposed by the US Geographic Data Commit- square meters, represents the eroded/accreted area in tee (FGDC – STD, 1998) was adopted. This standard that period. uses the residual error (Root Mean Square - RMS) to quantify the average error of the georeferencing process 3.2. Computational modeling (Table1). The hydrodynamic model used in this study was the TELEMAC-3D, developed by EDF - Laboratoire Na- Table 1 – RMS values calculated for each image composi- tional d'Environnement et Hydraulique Company Elec- tion. tricité de France (EDF), combined with the sediment Tabela 1 – Valor de RMS calculado para cada composição da transport model SediMorph, developed by Federal Wa- imagem terways Engineering and Research Institute and the Years Mean Squared Error (cm) Institute für Wasserwesen from the University of the German Armed Forces, Munich. 2004 1.1509 The TELEMAC-3D solves the Navier-Stokes equations 2007 1.0208 considering local variations on the free surface of the 2009 0.9824 fluid using the advection-diffusion equations for trans- 2012 1.0345 porting tracer concentrations (temperature, salinity or suspended sediment). The hydrostatic pressure and 3.1.2. Extraction of shoreline and area calculation for Boussinesq approximations were used to solve the mo- each image mentum equations (Hervouet, 2007). The SediMorph model is a tri-dimensional morphody- The shoreline determination and its dynamical behavior namic and fractionated sediment transport model (Mal- over time are essential for numerous activities related to cherek et al., 2005). This model calculates the bed research, engineering and planning. As the shoreline is roughness as a function of grain size and shape, the vulnerable to accretional and erosional processes, the bottom shear stress as function the roughness and inten- Change Polygon Method (Smith & Cromley, 2012) was sity of the flow, the rates of bottom transport, the rates used to evaluate the shoreline variation in this study. of erosion and deposition and the bed evolution. The erosion process is responsible for the loss in sedimentary areas when compared to other years, whist the The rate of erosion is calculated according the formula- accretion process is related to the increase of the beach tion of Partheniades (1965): strip caused by sediment accumulation. Due to the difficulty to detect the shoreline, its extrac- 1 tion was done using a constant indicator in each image called the high-water line (maximum run-up of the where, is the rateof erosion and is a shear stress. The water). Among all other possible indicators of shoreline formulation assumes that the sediment is eroded when reported by Boak & Turner (2005) the wet/dry interface the bottom shear stress is greater than the critical shear on a sand beach proved to be the most consistent one. strain. Based on the method proposed by Smith & Cromley The rate of deposition is based on the Partheniades (2012), from two distinct coastlines previously vec- (1965) formulation:

38 Lisboa & Fernandes (2015)

configuration before (Figure 2A) and another after the modernization work (Figure 2B). The simulations did 2 not include the wave effect. where C is the concentration of sediments along the A brief description of the initial and boundary water column, is the bottom shear stress, and is the conditions used in the simulations is provided in this critical shear strain for the deposition of sediments. section, while more details are presented in Marques et The SediMorph model was coupled to the hydrody- al. (2009, 2010). namic model TELEMAC-3D by Marques et al., 2010. Initial salinity and temperature fields were obtained Within the coupling, the hydrodynamic model calcu- from the OCCAM Project (Ocean Circulation and lates the flow velocity, flux of material in suspension Climate Advanced Modeling Project), and prescribed and deposition flux whereas the morphodynamic in three dimensions over the entire domain (Figure 3). model calculates sediment transport near the bottom, In the coastal area, salinity and temperature fields erosion flux, updating the friction with the new bed (OCCAM project) represent the influence of the La characteristics and bathymetry. The main parameters Plata river discharge over the SBS (Southern Brazilian considered are presented in Table 2. Shelf) and the Brazil Current (BC) along the continen-

Table 2 – Parameters used in the models set-up. tal shelf at the northern oceanic boundary. Water lev- Tabela 2 – Parâmetros usados para a calibração do modelo els were set to 0.4 m, which is about the mean value of the tide in the study region; finally, null velocity fields Time step 90 seconds were prescribed in the entire domain. At the continen- Coriolis coefficient -7.70 x 10-5 N m-1 s-1 tal boundaries, river discharge from the National Wa- Horizontal turbulence model Smagorinsky ter Agency’s page (www.ana.gov.br, ANA) was pre- Vertical turbulence model Mixing length (Jet) scribed for the main tributaries in the studied area (Camaquã and Guaíba rivers). At the oceanic bound- Scale mixing length 10 m ary, amplitude and phase data for the major five tidal Law of bottom friction Nikuradse components (K , M , N , O and S ), obtained from -6 -1 -1 1 2 2 1 2 Coefficient of Wind influence 6 x 10 N m s the Grenoble Model FES95.2 (Finite Element Solution Settling deposition law Van Leuseen - v.2004) were prescribed. Bottom transport model Hunziker The model surface boundary layer was forced with time Erosion law Quadratic and space varying winds from the NOAA Reanalysis Sediment transported in Fine Silt database (www.cdc.noaa.gov/cdc/reanaly-sis) with 1.75 suspension degrees of spatial resolution at 4 hours intervals. Data covering from 25 °S 42 °W to 38 °S 55 °W was col- The two coupled hydrodynamic and morphodynamic lected and interpolated using a cubic in-terpolation to a simulations, both 225 days length, were performed con- 1 km resolution mesh. Afterwards, for each nodal point sidering exactly the same initial and boundary condi- of the mesh, the meridional and zonal wind compo- tions. In order to analyze the results in a comparable nents of the wind were extracted and the resulting vec- way, each simulation was carried out based on a differ- tor was computed by the model using a constant coeffi- 6 ent numerical mesh (Figure 2), one representing the cient of wind influence of 6x10 .

Figure 2 - Mesh with the old (A) and actual (B) scenario. Figura 2 - Malha com os cenários Antiga (A) e Atual(B)

39 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

The initial conditions for SediMorph were prescribed quantify the behavior of the erosion and accretion proc- considering initial zero concentration of suspended ess before and after the modernization work carried out solids in the entire domain. Boundary conditions were in Rio Grande Harbor access channel, the time interval considered for the concentrations of suspended solids was fractionated over 3 years periods. Also, an overall along the continental open boundaries, where the river analysis of the 9 years period was presented. discharge is prescribed. Constant values of concentra- Figure 4 presents the Pontal Sul sand spit eroded and -3 tion of suspended solids were considered: 0.5 kg m increased areas and the balance between them for the -3 for the discharge of the Guaíba River and 0.3 kg m period between 2004 and 2012. Results presented in for the Camaquã River (Niencheski & Windom, Figure 4A indicate that for the period between 2004 and 1994). Initial bottom sediment distribution was pre- 2007 (before Rio Grande Harbor modernization work) scribed using measurements taken between 1975 and the Pontal Sul sand spit exhibited an erosive behavior 1979 in the Patos Lagoon and Estuary (Toldo, 1994). with a reduction of 26550 m² in the total area. This The classes of sediments prescribed were distributed result corresponds to an averaged annual erosion rate of as fine silt, fine sand and coarse sand (Figure 3B) in approximately 8850 m²/year. Figure 5A shows the the lagoon and estuarine region. At the coastal area, eroded (red) and the increased (yellow) areas for the the bottom was defined as fine sand (Figure 3B). same period. This analysis indicates that the Pontal Sul sand spit becomes slender and longer with time. Knowing that the perimeter of the sand spit in the earliest period (2004) was 2643.46 m, the mean reduction of the shoreline observed in this period was 5.46 m, resulting in an annual averaged erosion rate of 1.82 m/year. Results presented in Figure 4B indicate that for the period between 2007 and 2009 (the period over which the modernization work occurred), the Pontal Sul sand spit trend was still erosive (eroded area of 29370 m²), with an averaged annual erosion rate of 14685 m²/year. Figure 5B shows the eroded (red) and the increased (yellow) areas for the period. Results indicate that the Pontal Sul sand spit continues to grow in the north- south direction. The perimeter of the shoreline in 2007 was 3753.29 m, and the mean reduction of the shoreline observed in this period was 5.69 m, resulting in an annual average erosion rate of 2.48 m/year. Results presented in Figure 4C indicate that for the period between 2009 and 2012 (the modernization work was finished in 2010), a reduction in the erosion rates, Figure 3 - Boundary Conditions for TELEMAC-3D model but not enough to change the erosive tendency of Figura 3 - Condições de Contorno do modelo TELEMAC-3D period. The total area eroded was 19110 m², with an average annual rate of 9555 m²/year. Figure 5C shows The calibration and validation of hydrodynamic model the eroded (red) and the increased (yellow) areas for the (TELEMAC3D) and morphodynamic model period, which indicates a decrease in the erosion (SediMorph) coupling was presented in detail by tendency when compared to the previous periods. The Marques et al. (2010), where the authors compared the perimeter of the shoreline in 2009 was 3004.93 m, and model results with hydrodynamic field data and the average reduction of the shoreline observed in this suspended sediment concentrations extracted from period was 2.70 m, resulting in an annual mean of satellite images. Results from the calibration and 0.9 m/year. validation tests indicated that the coupled models well The overall balance over time (2004 – 2012) indicated reproduce the behavior of the system. that Pontal Sul sand spit eroded area was higher than the increased area (Figure 6). The total eroded area was 4. Results 64750 m², with an average annual erosion rate of 4.1. Sand spit morphological variations 8093 m²/year. Figure 7 strengthens this result showing the eroded (red) and the accreted (yellow) areas for the Variations on the shoreline position and on the sand spit whole period. The initial perimeter of the shoreline in area were analyzed from 2004 to 2012. In order to 2004 was 2643.46 m, and the average reduction of the

40 Lisboa & Fernandes (2015)

Figure 4 - Pontal Sul eroded and increased area and the balance between them for A) 2004 to 2007, B) 2007 to 2009, C) 2009 to 2012 Figura 4 - Área erodida e acrescida do Pontal Sul, e o balanço entre elas entre A) 2004 e 2007, B) 2007 e 2009, C) 2009 e 2012.

Figure 5 - Pontal Sul eroded and increased area and the balance between them between 2004 and 2012. Figura 5 - Área erodida e acrescida do Pontal Sul, e o balanço entre elas entre 2004 e 2012. shoreline observed in this period was 16.88 m, resulting Harbor access channel, numerical simulations were in an annual erosion rate of 2.11 m/year. Thus, the carried out using both domains (see Figure 2) with overall analysis of the studied period indicated that the exactly the same initial and boundary conditions. predominant process is the erosion of the sand spit west margin; although, accretion also occurs in specific ar- 4.2.1 Residual Ebb Current Velocity eas, promoting the longitudinal growth of the sand spit. The dominance of NE winds over the region induces the dominance of ebb currents in the estuarine region. 4.2. Hydrodynamic Modeling The mean ebb current velocity calculated by the nu- In order to compare the hydrodynamics and sediment merical model for both scenarios, for the simulated dynamics of the Patos Lagoon estuary before and after period, is presented in Figure 8. The highlighted area the modernization work carried out at Rio Grande along the sand spit shows a reduction in the mean ebb

41 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

Figure 6 -Evolution of the sand spit area between 2004 and 2012 superimposed on a QuickBird image from 2004,and its spectral composition (bands). Figura 6 - Evolução da área do Pontal Sul entre os anos de 2004 e 2012, sobreposta a uma imagem QuickBird do ano de 2004, com suas composições espectrais (Bandas). current velocity when the present configuration of Rio Grande Harbor access channel was taken in account. Therefore, the model results confirm the tendency observed in section 4.1, suggesting that before the modernization work the ebb currents that tended to erode the west margin of the sand spit were stronger. Modeling results suggest that after the work finished, the erosion potential was reduced by the reduction of the mean ebb current velocity along the west margin of the spit. Furthermore, modeling results confirm the Figure 7 - Evolution of the sand spit area between A) 2004 growing trend observed in the northern part of the sand and 2007 (superimposed on a QuickBird image from spit because the recirculation created in this region after 2004); B) 2007 and 2009 (superimposed on a QuickBird the modernization work (highlighted in Figure 8B) image from 2007); C) 2009 and 2012 (superimposed on a appears to keep the sediment trapped in the area. QuickBird image from 2009), with its spectral composition (bands). 4.3. Morphodynamic modeling. Figura 7 - Evolução da área do Pontal Sul entre A) 2004 e 4.3.1. Bottom Shear Stress 2007 (sobreposta a uma imagem QuickBird de 2004); B) 2007 e 2009 (sobreposta a uma imagem QuickBird de Figure 9 presents the mean bottom shear stress 2007); C) 2009 e 2012, sobreposta a uma imagem calculated by the model using the old (Figure 9A) and QuickBird de 2009), com suas composições espectrais new (Figure 9B) configuration of Rio Grande Harbor (Bandas). access channel. Warm (cold) colors indicate higher (lower) values. 4.3.2. Mean Erosion and Deposition Fluxes Results indicate that there was an overall reduction in Figure 10 presents the mean deposition flux calculated the bottom shear stress in the main channel when by the morphodynamic model considering the old considering the new configuration of Rio Grande (Figure 10A) and new (Figure 10B) configuration of Harbor access channel. A similar behavior was also Rio Grande Harbor access channel. Warm (cold) colors observed in the area close to the Pontal Sul west margin indicate higher (lower) values. central part (Figure 9B). These results corroborate the Modelled results indicate that there was an overall in- erosion tendency decrease observed in the satellite crease in the mean deposition flux along the main images. 42 Lisboa & Fernandes (2015)

Figure 8 - Mean ebb current velocity calculated by the model for the old (A) and actual (B) scenario, superimposed on a QuickBird image from 2012. A highlight is given for the recirculation observed at the north of Pontal Sul. Figura 8 -Velocidade média de vazante calculada pelo modelo para o cenário antigo (A) e atual (B), sobreposta a uma imagem QuickBird do ano de 2012. Em destaque a recirculação observada ao norte do Pontal Sul.

Figure 9 - Bottom shear stress calculated by the morphodynamic model considering the old (A) and new (B) scenarios, superimposed on a QuickBird image from 2012. Figura 9 - Tensão de cisalhamento de fundo calculada pelo modelo morfodinâmico para o cenário antigo (A) e para o cenário atual (B), sobreposta a uma imagem QuickBird do ano de 2012.

Figure 10 - Mean deposition flux calculated by the morphodynamic model for the old (A) and new (B) scenarios. Results superimposed on QuickBird image from 2012. Figura 10 - Fluxo médio de deposição calculado pelo modelo morfodinâmico para o cenário antigo (A) e novo (B). Resultados sobrepostos imagem QuickBird do ano de 2012.

43 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

Figure 11 - Mean erosion flux calculated by the morphodynamic model for the old (A) and new (B) scenarios. Results superimposed on QuickBird image from 2012. Figura 11 - Fluxo médio de erosão calculado pelo modelo morfodinâmico para o cenário antigo (A) e novo (B). Resultados sobrepostos imagem QuickBird do ano de 2012. channel when considering the new configuration informed as initial condition for the model. The (Figure 10B) of Rio Grande Harbor access channel. comparison between the bed evolution results Moreover, an increase in the mean deposition flux was calculated by the morphodynamic model for the old observed at the north of the Pontal Sul sand spit. As a (Figure 12A) and new (Figure 12B) scenarios indicate result, these results are in accordance with the increased significant differences in the north of the Pontal Sul deposition observed in this area in the satellite images region. It is evident that after the modernization work (Figure 5C). (Figure 12B) the material deposition in the sand spit The Figure 11 presents a similar analysis for the area is favored. modelled mean erosion flux. Although results do not Results presented in this study were consistent and of indicate expressive differences for the main access great potential for future applications and research. channel, a decrease in the mean erosion flux is evident However, some considerations should be made in along the Pontal Sul west margin when considering the relation to the advantages and disadvantages of the new scenario (Figure 11B). These results are in methodologies employed and, also, regarding sugges- accordance with the decrease in the erosion tendency tions for the continuity of the present study. observed in satellite images after the modernization The sediment transport study in coastal regions can be work (Figure 5C). done by means of field data collection, remote sensing and numerical modeling. Each of these methods has 4.3.3. Bed Evolution advantages and disadvantages in relation to operational The bed evolution calculation takes in account the bal- costs, logistics involved and complexity of application. ance between the deposition flux and the erosion of the In this study, both methods considered (remote sensing suspended sediment and the transport of material and numerical modelling) were of low cost and fairly

Figure 12 - Bed Evolution calculated by the model for the old (A) and new (B) scenarios superimposed on QuickBird image from 2012. Figura 12 - Evolução do fundo calculada pelo modelo morfodinâmico para o cenário antigo (A) e atual (B), sobreposta numa imagem QuickBird do ano de 2012.

44 Lisboa & Fernandes (2015) simple logistics and at the same time providing good 6. Conclusions spatial and temporal resolution results to subsidize The idea of studying geomorphological changes of spe- analysis. The consistency of both methodologies was cific coastal features to infer changes in the dynamics of validated by the convergence between the results on coastal systems resulting from anthropogenic contribu- the observed and modelled variations in the sand spit tions in coastal environments proved to be a valuable shoreline resulting from the erosion and deposition strategy. The combination between remote sensing in- balance. formation and numerical modeling for specific scenar- 5. Discussion ios proved to be a low cost method with high spatial resolution. Thus, the combined methodology was effi- Sedimentary transport studies are usualy accomplished cient in evaluating the effect of the modernization work on field-based approaches, remote sensing techniques, carried out at Rio Grande Harbor access channel on the and/or numerical modeling. The present study achieved sedimentary dynamics of Pontal Sul sand spit as an relevant results linking both remote sensing and indicator of the anthropogenic contribution to the dy- numerical modeling techniques. Antiqueira & Calliari namics of the system. (2004) studied the dynamics of the Pontal Sul sand spit The modernization work carried out in the access chan- between 1947 – 2003; during this period, the authors nel to Rio Grande Harbor did change the overall sedi- observed alternating periods of accretion and erosion ment dynamics of the system. Specifically for the Pon- promoting variations in the feature, resultig in a net tal Sul spit, it is possible to infer that this feature tends growth of 154 % in spit area anf increase of 304.4 m in to become more slender and longer within time due to its longitudinal axis (N-S). The present study the reduction in the erosion tendency observed along corroborates to the dynamic behavior of the Pontal Sul the west coast of the spit and the less intense recircula- spit between 2004 to 2012, which presented tendency of tion pattern established at the northern end. erosion in the southern and central parts and of accretion in the northern part. However, based on the These preliminary results also suggest that siltation in satellite images and in the modelled results, it was the main channel increased due to the modernization evident that the Rio Grande Harbor modernization work. Therefore, it is essential to recognize the impor- seemed to change the expected patterns by reducing the tance of carrying out the adequate investigation of the erosion tendency along the west margin of the spit and environmental impacts of such kind of coastal work increasing the deposition tendecy in the northern area. before its execution. Moreover, it was evident that changes in mean ebb flow, which is predominant in the area (Moller et al., Acknowledgments 2001), was the mechanism responsible for the observed The authors would like to express their gratitude to FINEP for alterations in the erosion and deposition fluxes, as ex- sponsoring the TRANSAQUA Project plained by Partheniades (1965). (www.transaqua.furg.br), contract 01.11.0141.01, and to CNPq trough contracts 551436/2011-5 (EHF) and 308274/2011-3 Lumborg & Pejrup (2005) described a similar mecha- (EHF). nism in the estuary of Lister Dyb, between Denmark and Germany. The authors used numerical modeling References techniques to simulate the transport of sediment in the Antiqueira, J.A.F.; Calliari, L.J. (2004) - Geomorphologic Evolution region, observing that the current velocity and shear of a Sand Spit Located in the Mouth of a Chocked Costal stress were the main responsible for the remobilization Lagoon. Lagoa dos Patos: Southern Brazil. Journal of Coastal of sediments, which would eventually reach the re- Research (ISSN: 0749-0208), SI39:255-258 [ICS 2004 Pro- gions of low energy and deposit. ceedings], Itajaí, Santa Catarina, Brazil. The effect of anthropogenic contributions in the Boak, E.H.; Turner, I.L. (2005) - Shoreline Definition and Detection: dynamics of costasl systems was also mentioned by A Review, Journal of Coastal Research, 21(4):688-703. DOI: Lorenzo et al. (2007). The authors studied a sand spit 10.2112/03-0071.1. system in Spain and observed that hydrodinamic Calliari, L.J.; Winterwerp, J.C.; Fernandes, E.; Cuchiara, D.; Vinzon, parameters such as current velocity and tidal range, as S.B.; Sperle, M.; Holland, K.T. (2009) - Fine grain sediment transport and deposition in the Patos Lagoon Cassino beach well as morphodynamic parameters, were affected by sedimentary system. Continental Shelf Research, 29(3):515-529. an anthropogenic contribution, resulting in changes in DOI: 10.1016/j.csr.2008.09.019 the accretion trend during the studied period. More Fernandes, E.H.L.; Dyer, K.R., Möller, O. O. (2005) - Spatial Gradi- recently, Zaromskis & Gulbinskas (2010) observed ent in the flow of southern Patos Lagoon. Journal of Coastal Re- that the progressive dredging carried out at Klaipida search, 21(4):759-769, DOI: 10.2112/006-NIS.1 harbor changed the local sediment dynamics, affecting Fernandes, E.H.L.; Dyer, K.R.; Möller, O.O.; Niencheski, L.F.H. the maintenance of the Curonian Spit due to the deficit (2002) - The Patos Lagoon hydrodynamics during El Niño event (1998). Continental Shelf Research, 22(11-13):1699-1713. DOI: of sediment supplement. 10.1016/S0278-4343(02)00033-X

45 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):35-46 (2015)

Fernandes, E.H.L.; Mariño-Tapia, I.; Dyer, K.R.; Moller, O.O. Martins, L.R.; Villwock, J.A.; Calliari, L.J. (1987) - Ocorrência de (2004) - The attenuation of tidal and subtidal oscillations in the lama na praia do Cassino, (RS). Anais Hidrográficos (ISSN Patos Lagoon estuary. Ocean Dynamics, 54(3-4):348-359. DOI: 0373-9260), 33:1-22, Ministério da Marinha, Rio de Janeiro, RJ, 10.1007/s10236-004-0090-y Brazil. Fernandes, E.H.L.; Moller, O.O.; Cuchiara, D.; Marques, W.C. (2012) - Avaliação das Condições de Navegabilidade do Canal Miranda, L.B.; Belmiro, M.C.; Kjerfve, B. (2002) - Princípios de de Acesso ao Porto de Rio Grande Após as Obras de Moder- Oceanografia Física de Estuários. 424p., Ed. USP, São Paulo, nização. Technical Report, Rio Grande, RS, Brazil. Unpublished. SP, Brazil. ISBN: 978-8531406751 Hartmann, C., Harkot, P.R.G. (1990) – Influência do canal São Möller, O.O., Castaing, P. (1999) - Hydrographical characteristics of Gonçalo no aporte de sedimentos para o estuário da Laguna dos the estuarine area of Patos Lagoon (30°S, Brazil). In: G. M. E. Patos – RS. Revista Brasileira de Geociências (ISSN: 0375- Perillo, M. C. Piccolo & M. Pino-Quivira (eds.), Estuaries of 7536), 20(1-4):329-332, Sociedade Brasileira de Geologia São South America, pp.83-100, Springer Berlin Heidelberg, Heidel- Paulo, SP, Brazil. Available on-line at berg, Germany. ISBN: 978-3642601316. DOI: 10.1007/978-3- http://sbgeo.org.br/pub_sbg/rbg/vol20_down/2001-2-3-4/2001329.pdf 642-60131-6_5 Hervouet, J.M. (2007) - Hydrodynamics of Free Surface Flows: Möller, O.O.; Castaing, P.; Salomon, J.C.; Lazure, P. (2001) - The Modelling With the Finite Element Methods. 360p., Wiley, New influence of local and non- local forcing effects on the subtidal York, NY, U.S.A. ISBN: 978-0470035580. circulation of Patos Lagoon. Estuaries, 24(2):297-311, DOI: 10.2307/1352953. Kjerfve, B. (1986) – Comparative Oceanography of Coastal La- goons. In: Douglas A. Wolfe, Estuarine Variability, pp. 63-81, Niencheski, L.F.; Windom, H.L. (1994) - Nutrient flux and budget in Ed Estuarine Variability, Academic Press, Inc, New York, NY, Patos Lagoon Estuary. The Sci. Tot. Environ., 149:53-60. DOI: USA. ISBN: 978-0127618906. Available on-line at http://dx.doi.org/10.1016/0048-9697(94)90004-3(1965) - Erosion http://geotest.tamu.edu/userfiles/167/61.pdf and deposition of cohesive soils. Journal of the Hydraulics Divi- Lorenzo, A.A.; Pagés, J.L. (2007) - Erosion and Accretion of Beach sion (ISSN: 0044-796X), 91(1):105-139, American Society Civil and Spit Systems in Northwest Spain: A Response to Human Ac- Engineers, New York, NY, U.S.A. tivity. Journal of Coastal Research, 23(4):834-845. Smith, M.J.; Cromley, R.G. (2012) - Measuring historical coastal DOI: 10.2112/04-0236.1 change using GIS and the change polygon approach. Transac- Lumborg, U.; Pejrup, M. (2005) - Modelling of cohesive sediment tions in GIS, 16(1):3-15. DOI: 10.1111/j.1467- transport in a tidal lagoon – an annual budget, Marine Geology, 9671.2011.01292.x 218:1-6. DOI: 10.1016/j.margeo.2005.03.015 Toldo Jr., E.E. (1994) - Sedimentação, Predição do Padrão de Malcherek, A., Piechotta, F., Knoch, D. (2005) - Mathematical Ondas, e Dinâmica Sedimentar da Antepraia e Zona de Surfe do module Sedimorph – Validation Document – version 1.1. Techni- Sistema Lagunar, da Lagoa dos Patos, RS. Porto Alegre. 189p., cal Report. The Federal Waterways Engineering and Research Tese de Doutorado em Geociências, Universidade Federal do Institute. Karlshure, Hamburg, Ilmenau, Germany. Unpublished. Rio Grande do Sul, Porto Alegre, RS, Brazil. Unpublished. Marques, W.C.; Monteiro, I.O.; Fernandes, E.H. (2009) - Numerical Žaromskis, R., Gulbinskas, S. (2010) - Main patterns of coastal zone modeling of Patos Lagoon plume, Brazil. Continental Shelf Re- development of the Curonian Spit, Lithuania. Baltica (ISSN: search. 29(3):556-571. DOI: 10.1016/j.csr.2008.09.022 0067-3064), 23(2):149-156. Vilnius, Klaipédia, Lithuania. Avail- able on-line at http://www.geo.lt/geo/uploads/media/149-156.pdf Marques, W.C.; Monteiro, I.O.; Fernandes, E.H.; Möller, O.O. Zenkovitch, V.P. (1967). Processes of coastal development. 738p. (2010) - Straining and advection contributions to the mixing Translated by D. G. Fry and edited by J. A. Steers. Oliver and processo of the Patos Lagoon costal plume, Brazil. Journal of Boyd, Edinburgh, Scotland, U.K. Geofhysical Research, 115(C06019):1-23. DOI: 10.1029/2009JC005653.

46 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015)

http://www.aprh.pt/rgci/pdf/rgci-490_Bio.pdf | DOI:10.5894/rgci490

Methods for coastal monitoring and erosion risk assessment: * two Portuguese case studies

A. Bio @, a; L. Bastos a, b; H. Granja a, c; J.L.S. Pinho d; J.A. Gonçalves a, b; R. Henriques c, e; S. Madeira a, f; A. Magalhães a, g; D. Rodrigues d

Abstract Coastal zones are naturally dynamic and mobile systems exposed to natural factors (river flows, waves and storms) as well as human interventions that continuously reshape their morphology. Erosion phenomena related to extreme weather events and sediment scarcity are common, threatening buildings and infrastructures, as well as beaches, ecosystems and valuable wetland; conditions that pose challenges to coastal security and defence. Regular monitoring of coastal areas, assessment of their morphodynamics and identification of the processes influencing sediment transport are thus increasingly important for a better understanding of changes and evolutionary trends in coastal systems. This demands a multi-disciplinary approach involving researchers with expertise in coastal processes and state- of-the-art observation technologies. In this paper state-of-the-art surveying methods for an efficient quantification of changes in coastal environments are described and evaluated, and two NW-Portuguese case studies are presented. Survey methods included: topographic surveys based on terrestrial videogrammetric mobile mapping and aerial photogrammetry; sub-tidal bathymetry with sonar imagery using an Autonomous Surface Vehicle (ASV); as well as field observations, with sediment sampling and beach characterisation. In the first case study, erosion/accretion patterns in the Douro estuary sand spit were analysed, considering its breakwater, river flow, wave and wind effects. Prior to the construction of a detached breakwater, the spit’s morphodynamics was related to extreme river flow events, wave and wind conditions; afterwards the spit stabilized its shape and increased its area and volume. In the second case study the coast of Vila Nova de Gaia was broadly analysed, including the shoreface, foreshore and dunes, the characterization of major features and a short-period analysis of installed dynamics. Results obtained from field data, topographical surveys and numerical wave models were combined for an erosion risk assessment, using a methodology specifically developed for the study area. Both monitoring programs achieved their proposed objectives and provided valuable information to the local authorities, as gathered and processed information constitutes a valuable database for coastal planning and for ICZM purposes. They demonstrate the potential of several approaches, supported by advanced technologies, for the study of complex coastal morphodynamic processes. Keywords: monitoring techniques; Digital Elevation Models; morphodynamics; coastal erosion; risk assessment; Douro Estuary; Vila Nova de Gaia; Portugal.

@ Corresponding author: a Universidade do Porto, CIMAR/CIIMAR – Centro Interdisciplinar de Investigação Marinha e Ambiental, Rua dos Bragas, 289, 4050-123 Porto, Portugal b Universidade do Porto, Faculdade de Ciências, Departamento de Geociências, Ambiente e Ordenamento do Território, Rua Campo Alegre 687, 4169 - 007 Porto, Portugal c University of Minho, Department of Earth Sciences, Campus de Gualtar, 4710-057 Braga, Portugal d University of Minho, Department of Civil Engineering, Campus de Gualtar, 4710-057 Braga, Portugal e Universidade do Porto, Centro de Geologia, Rua Campo Alegre 687, 4169 - 007 Porto, Portugal f Universidade de Trás-os-Montes e Alto Douro, Department of Mathematics, Apartado 1013, 5001-801 Vila Real, Portugal g Universidade do Porto, Faculdade de Ciências, Observatório Astronómico, Monte da Virgem, 4430−146 Vila Nova de Gaia, Portugal

* Submission: 14 JAN 2014; Peer review: 6 FEB 2014; Revised: 26 MAR 2014; Accepted: 3 APR 2014; Available on-line: 11 APR 2014

Bio et al. (2015)

Resumo Métodos de monitorização e análise de risco de erosão costeira: dois casos de estudo portugueses Zonas costeiras são sistemas naturalmente dinâmicos e móveis, expostos a fatores naturais (fluxos de rios, ondas e tempestades) e intervenções humanas, que alteram continuamente a sua morfologia. Fenómenos de erosão relacionados com temporais e escassez de sedimentos são comuns. Eles ameaçam construções e infraestruturas, praias, ecossistemas e zonas húmidas, o que constitui um desafio para a segurança e defesa costeiras. Uma monitorização regular de áreas costeiras, com avaliação da sua morfodinâmica e identificação dos processos que influenciam o transporte de sedimentos, visando uma melhor compreensão das alterações e tendências evolutivas nos sistemas costeiros, torna-se portanto cada vez mais importante. Para tal necessita-se de uma abordagem multidisciplinar e investigadores especializados em processos costeiros e tecnologias de observação de ponta. O presente trabalho descreve e avalia métodos de monitorização de última geração para uma quantificação eficiente de alterações em ambientes costeiros e apresenta dois estudos de caso Portugueses. Os métodos de observação incluem: levantamentos topográficos terrestres em modo cinemático baseados em videogrametria; levantamentos topográficos aéreos baseados em fotogrametria; batimetria sub-tidal e imagens de fundo obtidas com sonar a partir de um veículo autónomo de superfície; e observações de campo com análise de sedimentos e caracterização de praias. O primeiro caso de estudo refere- se à análise de padrões de erosão e acreção na restinga do estuário do Douro, tendo em conta efeitos causados pelo molhe, o caudal do rio, ondas e vento. Antes da construção de um molhe destacado, a morfodinâmica da restinga estava relacionada com eventos extremos de caudal, agitação marítima e vento. Após a construção, a forma da restinga estabilizou e observou-se um aumento da sua área e do seu volume. No segundo caso de estudo, a costa de Vila Nova de Gaia foi amplamente estudada, incluindo shoreface, foreshore e dunas, com a caracterização dos principais atributos e uma análise de curto-período da dinâmica instalada. Os resultados obtidos a partir de dados de campo, dos levantamentos topográficos e de modelos numéricos de ondas foram combinados numa análise de risco com métodos especificamente desenvolvidos para a área de estudo. Ambos os programas de monitorização atingiram os seus objetivos e geraram informação relevante para as autoridades locais. A informação recolhida e processada constitui uma base de dados valiosa para o planeamento costeiro e a Gestão Integrada de Zonas Costeiras. Os estudos demonstram o potencial das diversas abordagens, apoiadas por tecnologias avançadas, para o estudo dos processos complexos de morfodinâmica costeira. Palavras chave: técnicas de monitorização; Modelos Digitais de Elevação; morfodinâmica; erosão costeira; análise de risco; estuário do Douro; Vila Nova de Gaia; Portugal.

1. Introduction Coastal erosion is complex and depends on the off and on-shore environment, including wave energy and Being land-ocean interfaces, and river-ocean interfaces direction, weather and climate, the materials that make at estuaries, coasts constitute transitional zones of high up the coast and the supply or extraction of sediments, ecological, as well as economical importance. Exposed as well as the influence of manmade structures to different phenomena that continuously reshape their (Archetti & Zanuttigh, 2010; Granja & Pinho, 2012). morphology, coasts are naturally dynamic and mobile Monitoring should consider these interactions and be as systems. They constantly adapt in response to natural comprehensive as possible, including the submerged forcing factors (currents, waves, winds and storms), but part of the shore to obtain important information on also to human interventions, like sand mining, and slope, bathymetry and the bottom features and materials construction of edifications and defence structures. (Holland et al., 2009). Many coasts suffer from erosion processes, causing coastal retreat that threatens beaches, wetlands, marshes For a responsible and sustainable coastal management, and coastal ecosystems, as well as buildings and infra- it is necessary to understand the coastal system, its structures, hence stressing local economies. For physical processes (e.g., sediment transport patterns, instance, almost all coastal EU Member States have sediment sources and sinks), forcing factors (e.g., wave problems with coastal erosion, with more than 20% of action, storm surges, currents) and their interrelations the evaluated European coastline affected (Niesing, and effects. Coastal morphology and its dynamics have 2005). In an effort to mitigate erosion, soft (e.g. beach to be followed and quantified along time, and erosion nourishment) and hard defences (e.g., groins or break- risks have to be evaluated considering present and waters) are built against sea impacts. Hard defence potential future (e.g., climate change) scenarios structures are however static and interfere with natural (Nicholls et al., 2007; Dodet et al., 2010; Pereira & coastal dynamics and the accommodation space for Coelho, 2013). adaptation, and their medium to long-term effects are Several methods have been proposed for coastal frequently unexpected as there is often not enough morphodynamic monitoring. Coasts can be analysed in quantified knowledge about the local coastal dynamics terms of shoreline dynamics only (Boak & Turner, and the forcing factors and processes driving it. 2005) or in terms of beach morphology changes.

48 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015)

Surveys using conventional techniques, such as classic comprehensive evaluation of morphodynamic trends surveying equipments (Holland et al., 2009; Cowart et and erosion risks in the coastal system. Analyses were al., 2010; Souza & Luna, 2010; Muehe, 2011), are both based on results from the simultaneous and integrated costly and labour-intensive, and generally lack the use of terrestrial mobile mapping surveys, as well as spatial and temporal resolution needed to identify and airborne digital photography surveys, bathymetric quantify short-term changes as well as long-term trends surveys, in-situ field observations, beach profiling and (Short & Trembanis, 2004). Alternatives using fixed sediment sampling. video imagery (Davidson et al., 2007; Archetti & Zanuttigh, 2010) provide lower-cost continuous data, 2.1. Terrestrial mobile mapping surveys i.e. an optimal temporal resolution, but are generally Traditional land surveying methods are time consum- limited to few beach stretches. Remotely sensed data ing, and therefore also expensive. A much more have also been successfully used for coastal monitoring. efficient alternative is the use of a Mobile Mapping Satellite radar, particularly Synthetic Aperture Radar (SAR), have been used to extract shoreline information System (MMS), using a direct georeferencing system and video cameras, which is easily and quickly put into (Niedermeier et al., 2000) and aerial digital photogra- operation under almost any weather conditions. The phy to obtain 3D information through photogrammetric observing methodologies applied in our case studies, analysis (Fletcher et al., 2003; Cowart et al., 2010). evolved with time. In the sand spit study, surveys Laser surveys may deliver similar (or slightly better, started with an on-foot GPS equipment (used until 10-15 cm) accuracy in the vertical component and may 2002), followed by a two antenna system installed on a constitute an alternative to aerial digital photography moto-quad (from 2003 to 2009) (Baptista et al., 2008) and photogrammetry; but they are still extremely and, more recently (since 2009), by a videogrammetric expensive. The generalisation of GPS (Global Position- (Fig. 1), which can be used with different direct geore- ing System) methods, particularly those using RTK- GPS (Real-Time Kinematic GPS), has facilitated analy- ferencing systems, developed at the University of Porto (Madeira, 2007; Madeira et al., 2010). In the second sis of spatial variability allowing precise three- case study, only the two-antenna system was applied. dimensional survey data to be collected both rapidly and at high spatial resolution (Rocha et al., 2009; Automated spatial data collection using terrestrial Barnard & Warrick, 2010). A novel approach to videogrammetric MMS in coastal zones is still at an topographic surveys is videogrammetric mobile map- early stage of development (Zhu & Brilakis, 2009; ping, where 3D coordinates are extracted from stereo Brilakis et al., 2011). Software developed at the video images (Madeira, 2007; Zhu & Brilakis, 2009; University of Porto (UP), Portugal, was applied to Brilakis et al., 2011). automatically extract 3D coordinates (Madeira et al., The main objectives of this paper are: (1) to present an 2010). A process was developed that relies on an overview of state-of-the-art coastal survey techniques, automatic selection and coordination of ground points including novel methods for coastal topography defini- in each pair of photogrammetric images obtained by the tion, such as terrestrial videogrammetric mobile map- system’s video cameras. The DEM was computed ping and aerial digital photogrammetry, and (2) to show creating a TIN (triangulated irregular network) and their potential in morphodynamic monitoring and converting it to a regular grid (with 1 m spacing) by erosion risk assessment and, hence, integrated coastal linear interpolation. management. Two case studies are used to illustrate the With an adequate direct georeferencing system, data applicability of the methods. The first refers to the acquired with the MMS implemented at UP allow 3D Douro Estuary sand spit, where the effects of a de- topographic mapping with accuracies similar to those of tached breakwater on spit morphodynamics were RTK-GPS based methods (Dail et al., 2000). The analysed based on more than a decade of regular method to obtain the DEM from the image-derived surveys and considering river flow, wave and wind coordinates does not depend on the relative changes patterns. The second refers to a comprehensive between the GPS antenna and the ground but only on monitoring program developed to assess erosion risk for the calibration of the cameras (which can be done very the coast of Vila Nova de Gaia. precisely, at the sub-pixel level; Madeira et al. 2009) and the determination of the relative position of the 2. Coastal monitoring methods GPS antenna and the video cameras. In this section, the several survey techniques used in the The relative position of the cameras and the GPS/INS case studies are presented. The first case study, direct georeferencing system can also be accurately concerning the evolution of the Douro spit, used Digital established before mounting the system, and will Elevation Models (DTM) derived from terrestrial remain constant for each installation. DTM height mobile mapping surveys. The second consisted of a accuracies, tested with control points, were at the few- multi-disciplinary monitoring approach, aiming at a centimetre level (RMS = 0.06 m).

49 Bio et al. (2015)

Figure 1 - Terrestrial mobile mapping survey. A: Four-wheel vehicle with two kinematic antennas and videogrammetric equipment; B: Screenshot of the software used to extract topographic coordinates and terrain features. Figura 1 - Levantamento terrestre em modo cinemático. A: Carro equipado com duas antenas cinemáticas e equipamento de videogrametria; B: Imagem de ecrã do programa usado para extrair as coordenadas topográficas e atributos de terreno.

A further advantage is the simultaneous (video) Measurements with the MMS were collected continu- recording of ground features, such as dune vegetation or ously, in kinematic mode, using GPS post-processing storm scars, with a resolution of a few centimetres, relative positioning. Positioning and orientation of the superior to that achieved by traditional aerial photogra- cameras was initially done using a GPS-only dual phy. These records can be of upmost relevance for frequency multi-antenna setting to derive roll and later ecological and geomorphological interpretation of the a GPS/INS georeferencing system based on a dual environment, and can be compiled in GIS databases for frequency GNSS receiver and a navigation-grade IMU further analysis. (Inertial Measurement Unit), which allows determina- 50 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015) tion of the vehicle orientation in all three angular direc- obtain the centre of camera projection and the tions (roll, pitch and heading). Here, we present results attitudinal angles for each image. A boresight alignment obtained using the latter setting. was done with ground control points in order to correct slight systematic effects in the attitude angles of about 2.2. Airborne photogrammetry 0.02 degrees (Gonçalves et al., 2011). Calibration was done in built up areas, which contain features that can Aerial photos were taken using a digital camera (ZI be used as accurate ground control points. Imaging DMC) installed in an airplane flying at about DEMs were computed from the pairs of stereoscopic 1030 m height (Fig. 2). This resulted in high-resolution aerial images after extracting correlated points through images with 10 cm ground sampling distance (GSD). stereomatching, using the BLUH software (Jacobsen, Two strips of images were collected, with a longitudinal 2000; Rieke-Zapp & Nearing, 2005). Notice that stereo- overlap of 60% and a lateral overlap of 30%. Images matching is difficult with images of sandy, low-contrast were collected during spring low tides to maximize the beaches due to light over-exposure. Whenever possible, observable emerged area. surveys were therefore conducted under low-sun condi- Images were directly georeferenced by a GPS/INS tions, when some shadow patterns appear in the sand, system. The GPS post-processing relative positioning which facilitate image correlation. The digital photo- mode was used, and the information was processed to graphs were furthermore orthorectified and combined

Figure 2 - Airborne photography survey. A: Survey airplane; B: digital ZI-DMC camera (sensor with 7680×13824 pixels); C: Flight plan with two rows of overlapping photos (here: ~100 photos covering ~14 km). Figura 2 - Levantamento por fotografia aérea. A: Avião utilizado; B: câmara digital ZI DMC (sensor com 7680×13824 pixéis); C: plano de voo com as duas fiadas de fotografias sobrepostas (neste caso: ~100 fotos que cobrem ~14 km).

51 Bio et al. (2015) into mosaics with 0.1 m resolution, for visual analysis side-scan were also analysed to identify rocky outcrops of ground occupation and patterns. DTM height accura- and sandy patches. cies, tested with control points, were at the few The accuracy of the bathymetric surveys was not decimetre level (RMS = 0.18 m). evaluated by the authors. In controlled environments (harbours), the catamaran showed accuracies at the few 2.3. Bathymetric surveys centimetre level. At the coast and under ideal An Autonomous Surface Vehicle (ASV), a remotely conditions, depth could be measured with an accuracy computer controlled, 4.2 m long catamaran (Fig. 3), of about 15 cm. However, due to the variability of was used for bathymetric surveys (Ferreira et al., 2009). observation conditions in the ocean, in practise, all Vehicle navigation and positioning was done with a existing remote techniques (including echo- sounder, PolaRx RTK GPS, with a horizontal accuracy of 2 cm laser, etc.) usually deliver a worse final accuracy, with and a vertical accuracy of 4 cm. The platform was precisions not better than 20 to 30 cm. Notice that the furthermore equipped with a Tritech digital single beam catamaran-based system is not introducing additional side-scan altimeter sonar; a 500 KHz echo-sounder with errors or limitations, when compared to traditional 6o conical beam width and 1 cm of range resolution, bathymetric hydrographical surveys, and has advanta- measuring bathymetry and capturing acoustic images of ges concerning the depth of the zone that it can survey. the ocean bottom. The ASV was electrically propelled Compared to manned surveys, autonomous computer and the sonar sensor heads mounted at the aft of the controlled robotic surveys allow higher precision in boat without interference from the hull motion and course control. They are furthermore cheaper and safer propeller induced noise. to operate under rough conditions. The system allowed for continuous recording of all data, including navigation data and the observation 2.4. Field observations and sedimentary analyses time. The derived horizontal coordinates were referred For erosion risk assessment coasts need to be observed to the ETRS89 datum and depths referred to CD (Chart and characterized, giving particular attention to geo and Datum). Bathymetry data were used for the bio- indicators of coastal short term change and to the implementation of a wave propagation model. segments most sensitive to erosion. In our case study, Reflectance and shadows of the sonar images from the location, type and state of the beach-dune system pro-

Figure 3 - Bathymetric and seabed survey. A: Autonomous catamaran, platform with GPS RTK and altimeter; B: Side-scan sonar. Figura 3 - Levantamento batimétrico e do fundo. A: Catamarã autónomo, plataforma com GPS RTK e altímetro; B: sonar de varrimento lateral.

52 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015) file, grain size of sediments, regularity of the dune the spit tip in order to: stabilize the estuary margins, ridge, dune vegetation, as well as micro-forms like especially those of the sand spit and the estuary outlet; blowouts, cliffs, washovers, ridge and runnels, and improve navigability and navigation safety under any cusps were recorded. tide condition; reduce propagation of storm waves into Sediment size data were obtained through sieving and the estuary; preserve environmental values (most of the processed using the SEDMAC software (Henriques, sandy spit and the adjacent São Paio Bay constitute a 2004) to obtain commonly used size parameters for Nature Reserve); and reduce the dredging needed to particle populations. Data obtained included relative maintain the navigation channel. and cumulative frequencies in 1⁄2 Ø (phi) intervals. South of the river mouth lies the 15 km long coastal Mean, standard deviation, skewness and kurtosis were stretch of Vila Nova de Gaia (Fig. 4C). Many parts of determined, using the method of moments. The distri- that coast have shown progressive erosion in recent bution of size fractions according to the modified Wentworth scale was also determined. Sample distribution by dimensional 1⁄2 Ø class was done in correspondence to the metric scale. Sediment size and statistics were processed and integrated in a GIS, using the sediment samples’ spatial reference.

3. Case Studies 3.1. Study area The case studies refer to the Douro river barrier sand spit and the coast line south of it (Fig. 4). The Douro River has a funnel-shaped estuary that lies embedded in a steep valley, extending 21.6 km from the Crestuma- Lever dam, which sets an artificial limit to the estuary, to the river mouth (Portela, 2002). The river has a torrential regime, regulated by 15 dams (seven in the Spanish and eight in the Portuguese territory). Water flows to the estuary depend on water levels and on hydropower generation needs. Periods of null-discharge flows are frequent, mostly during summer, as well as floods (with river flood discharges reaching more than 13 000 m3 s-1) in winter or spring, when the dams’ limited water storage capacity is not enough to prevent them. The estuary inlet is protected by a sand spit (Fig. 4C). This spit is rooted on the left bank of the Douro river estuary and points north due to the S-N drift caused by wave diffraction around the ebb-tidal delta and some outcrops, and the breakwater to the North of the outlet. The spit is presently about 300 m wide (east-west) and 700 m long (north-south). Its Eastern flank is stabilized by plants like Elymus farctus and Ammophilla arenaria and connects with the São Paio Bay, a mostly submerged (1−2 m below CD), hydrodynamically calm wetland of ecological importance. Composed of fine to Figure 4 - Case study areas. A: Portugal (box) in SW-Europe; medium-coarse sand, the spit is influenced by subaerial- B: Location of Porto and Vila Nova de Gaia on the NW- fluvio-marine conditions, including; sedimentary Atlantic Portuguese coast; C: GoogleEarth image of the transport by the river, swell and tide effects, as well as two study areas (boxes), the Douro spit in the North and by winds (Silva et al., 2005; Consulmar, 1996). the coast of Vila Nova de Gaia municipality to the South of it. There are numerous records of flood events that caused Figura 4 - Áreas de estudo. A: Portugal (retângulo)no SO da rupture of the spit and of periods when accretion took Europa; B:Localização do Porto e de Vila Nova de Gaia place at the spit tip, causing its progression towards the na costa NO-Atlântica de Portugal; C: imagem North and even obstruction of the navigation channel. GoogleEarth das duas áreas de estudo (retângulos), a Between 2004 and 2008 the Northern breakwater was restinga do Douro no norte e a costa do município de extended and a new detached breakwater was built at Vila Nova de Gaia a sul. 53 Bio et al. (2015) decades (Granja & Carvalho, 2000; Soares de Carvalho position, shape and volume, and assess its evolution. et al., 2006), with coastal retreat reaching, in some These data were later analysed with the objective to cases, six meters per year, challenging coastal security identify relationships between sand spit dynamics and and protection. Several causes for erosion have been river flow regimes, ocean wave and wind patterns, and pointed out: increasing lack of sediment supply from to evaluate the impact of the breakwaters on spit rivers (due to dam construction since the 50’s, dredging morphology and on its vulnerability to river flow, wave and extraction of aggregates); sediment starvation in the and wind conditions. inner continental shelf; influence of coastal defence Using GPS in relative kinematic mode (see Section structures; and an increase in storm surge frequency and 2.1), quarterly terrestrial mobile mapping measurements intensity (though this is controversial). were taken along the instanta-neous water limit (i.e., of At these latitudes (40º–42ºN), the North-Western the sand spit edge) and on a grid of profiles, allowing Portuguese coast is highly energetic (Cruz, 2008), with determination of the spit’s contour and topography, and offshore mean significant wave heights of 2–3 m, and subsequent production of DEM representing its 3D mean wave periods of 8–12 s. In winter, storms surface (Fig. 5). DEM were used to analyse spit generated in the North Atlantic are frequent and can evolution in time, distinguishing between the periods persist for up to five days, with significant wave heights before and after breakwater construction. reaching 8 m (Costa et al., 2001). The tides are semi- Changes in sand spit shape, area, topography and diurnal, with heights ranging from 2 m to 4 m during volume were related to extreme river flows discharges, neap and spring tides, respectively. Waves are usually ocean wave and wind conditions, considering: (i) coming from the W-NW, causing a dominant drift maximum daily mean river flow discharges; (ii) the current from North to South (S-N drift is occasionally number of days with discharge flows higher than observed during W-SW storm periods). This current is 800 m3 s-1; (iii) maximum wave power events (with in some areas inverted due to the presence of obstacles wave power in kW m-1 estimated by ≈ 0.5 kW m3 s-1 × that promote wave diffraction, as happens at the Douro wave height2 × period); (iv) number of 3-hour periods spit. At Vila Nova de Gaia, the main drift direction is with wave power values above 200 kW m-1; (v) N-S, causing accretion northwards of obstacles maximum wind power (which was estimated as the (drainage pipes, breakwaters) and erosion to the South cube of the maximum wind speed; Rasmussen et al., (Granja et al., 2011). 2011); and (vi) the number of half-hour periods with

wind power values above 10 000 (m s-1)3. 3.2. Douro estuary sand spit study Daily average discharge records at the Crestuma-Lever 3.2.1. Monitoring and data analysis dam (supplied by INAG – Portuguese Water Authority, In June 2001 a monitoring program was set up to under permission of the EDP – Portuguese Electrical monitor the Douro Estuary sand spit, measuring its Company) were taken to represent total freshwater flow

Figure 5 - Sand spit survey. A: Distribution of samples collected during in June 2009 (each dot is an observation) and delimitation of the four spit sectors analysed (coordinates are in Datum 73, Hayford-Gauss, IPCC); B: resulting DTM. Figura 5 - Levantamento da restinga. A: Distribuição das amostras recolhidas em Junho de 2009 (cada ponto é uma observação) e a divisão nos quatro sectores analisados (com coordenadas em Datum 73, Hayford-Gauss, IPCC); B: MDT resultante. 54 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015)

into the estuary. Wave data were obtained at 3-hour during breakwater construction, stabilizing from 2008 intervals, from a datawell directional wave buoy located onwards for the eastern spit tip and displaying an about 32 km NW of the sand spit (41o19’N, 8o59’W), increasing trend for the western sectors (Fig. 7). at 83 m water depth (supplied by IH – Portuguese Hy- Analyses revealed that the observed erosion and drographical Institute). Wind velocities were obtained accretion patterns were related to extreme river flow, from NCEP FNL Operational Model Global Tropo- wave and wind conditions, with effects depending on spheric Analyses (National Center for Atmospheric spit section and height, and therefore exposure, and Research, http://rda.ucar.edu/datasets/ds083.2/), at 41º N with distinctly different patterns for the periods before and 9º W, considering the level of 1000 mbar. and after breakwater construction. Before breakwater To distinguish dynamics of different areas of the sand construction, extreme river flow was significantly (p < spit, the spit was divided into four sectors (Fig. 5): the 0.05) correlated with spit erosion. The number of days seaward spit base (SW sector), the seaward tip (NW), with extreme wave power was correlated with an the river facing spit base (SE) connected to the São Paio increase in area and in the volume observed at lower Bay wetland, and the river facing spit tip section (NE). heights (up to 4 m CD) of the NW sector. High wind For each sector and survey, sand spit area and volumes power was related to losses in volume and to erosion in above different heights were determined. Changes be- the SW sector, and to accretion in the NE. After tween surveys were analysed, discriminating between breakwater construction river flow, wave and wind the period before (Dec. 2001 to Jan. 2005) and after effects became overall less significant (see Bastos et al., breakwater construction (June 2007, when most of the 2012 for more detail). The breakwater seems hence to sand-spit breakwater structure was in place, to April protect most of the spit against rough conditions, 2010). leading to a relative stabilization of the spit’s shape and Correlations between these parameters and sand spit topography. The T-shape established after breakwater area and volume dynamics were assessed using construction (due to the creation of a tombolo towards Spearman correlation tests, as most pairs of data failed the detached breakwater and an inland arm), with little the Shapiro-Wilk normality test. All statistical analyses extension to the North, does not obstruct the navigation were carried out using R software (R Dev. Core Team, channel, and favours the navigability between the sea 2009). and the small estuarine harbours. There is, however, increasing silting up of the wetland area SE of the spit,

due to changes in circulation patterns caused by the new 3.2.2. Results and discussion spit shape, threatening the local Nature Reserve. In Considering sand spit shape and volume obtained from recent years, the São Paio Bay, has progressively filled seasonal surveys between 2001 and 2009 (Bastos et al., up with sediment and is losing its wetland character- 2012), three periods could be distinguished. The first, istics (pers. observation). between 2001 and the end of 2003, was very dynamic, Furthermore, the strengthening of the spit bears risks with marked changes of the sand spit tip shape and for the riparian communities. In recent years, the spit thickening and thinning of its body (Fig. 6; see Bastos has further stabilized and increased its area and volume et al., 2012 for more detail). The spit stayed stable (Fig. 6). In the period after the breakwater was built, the during the second period, 2004 to early 2005, which spit’s area increased ~56 000 m2 (i.e. ~23%) and its coincided with a severe drought in continental Portugal volume ~340 000 m3 (~33%). Its W-face has moved with relatively low river flows. Decreasing volumes in 30−45 m westward, while the lower E-face remained the seaward sections and slightly increasing volumes in stable. Prior to the sand spit breakwater construction, the NE spit tip sector were observed (Fig. 7). The third the sand spit used to rupture under severe floods, period was related to the construction of the breakwater allowing discharge of the river flow through a wider attached to the sand spit, which started in March 2005. cross section. A strengthened spit constitutes a more During breakwater construction the spit tip developed stable barrier and may cause floods with higher water an inland arm (along an older breakwater, which runs in levels within the estuary for a given flood discharge. E-W direction along the top of this arm), with a branch The monitoring of the Douro barrier sand spit is an (Fig. 6). This inland arm was eroded by the time of the ongoing project, integrated in a larger monitoring December 2006 survey, after a flood that occurred in framework promoted by the APDL, which includes the previous November, and was breached before the morphodynamical surveys in the Douro estuary, as well April 2010 survey after extreme river flows in the as monitoring of navigation conditions, damages and previous winter. Sand spit areas and volumes oscillated protections of harbours and other manmade structures.

55 Bio et al. (2015)

Dec. 02 Dec. 03 Jan. 05

Dec. 05 Dec. 06

Dec. 07 Dec. 08 Dec. 09

Jan. 11 Fev. 12 Mar. 13

Figure 6 - DTM obtained from surveys carried out before (A), during (B) and after (C) the construction of the breakwater (a selection of yearly winter surveys is presented; more data can be found in Bastos et al., 2012). Figura 6 - MDT obtidos em levantamentos antes (A), durante (B) e após (C) a construção do molhe (apresenta-se uma seleção de levantamentos anuais de inverno; dados adicionais encontram-se em Bastos et al., 2012).

56 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015)

Figure. 7 - Variation in sand spit volume and area for the four sand spit sectors considered. The stacked area plots show volumes for different height layers and the line represents the total area of each sector. Figura 7 - Variação do volume e da área da restinga para os quatro sectores considerados. As áreas empilhadas representam os volumes para camadas de diferentes altitudes e a linha a área total de cada sector.

3.3. Vila Nova de Gaia erosion risk study limitations. The team therefore decided in favour of 3.3.1. Monitoring and risk analysis distributing the obtainable profiles over the whole area (N.B. there are hardly any bathymetric data available In 2007, a multi-disciplinary monitoring program was for that shoreface area). set up to follow and quantify short-term changes of the Sediment samples were collected along transects per- densely occupied and locally erosion-prone Vila Nova pendicular to the coastline. Whenever possible, samples de Gaia coast. The objective was to evaluate morpho- were taken on the dune crest, at the dune base and at the dynamic trends of the coastal system, including the shoreface, the foreshore and the dunes. Four campaigns took place, after fair weather conditions in the autumns high, mean and low tide water levels. Sediment samples of 2008 and 2009, and just after storm winter conditions were processed and analysed to determine grain size in the springs of 2009 and 2010. An innovative set of fractions (as described in Section 2.4). monitoring methodologies was applied, including: Collected data were integrated in a Geographical mobile mapping surveys, airborne digital photography Information System (GIS) for further spatial analysis, surveys for photogrammetric topography, bathymetric using the PT-TM06/ETRS89 (EPSG: 3763) coordinate surveys and in-situ field observations, beach profiling system. To assess morphodynamics during the observa- and sediment sampling (see Section 2). tion period, a different DEM was obtained for each Bathymetric surveys were done along (E-W) profiles, campaign from terrestrial mobile mapping and from more or less perpendicular to the coastline and up to 8 airborne photogrammetric surveys (Fig. 8). Contours, m CD depth (i.e., 10 m MSL), during high tide, to allow partial and total volumes and sediment budgets were driving the ASV as close as possible to the foreshore. determined for different topographic height classes, and Depending on the bottom slope, the profile lengths for each of the coastal segments. varied from 300 m to 2000 m. Profiles were collected at Survey results were used in combination with wave 1 km intervals. A denser observation scheme would be climate data (from the same close-by offshore direc- recommended but was not feasible due to budget tional wave buoy as in the previous study), and a local

57 Bio et al. (2015)

wave propagation model was implemented to assess local wave energy spatial distributions and produce erosion risk maps. Field data and model outputs were integrated, processed and analyzed to assess the coastal vulnerability to erosion and for the elaboration of associated risk maps using a multi-criteria assessment approach. A specific erosion risk assessment methodology was developed for this work, at the University of Minho, based on morphological, hydrodynamic and other indicators that were quantified using the data and results obtained from the monitoring program. Vulnera- bility and exposure indicators were aggregated in order to classify the coast according to three erosion risk classes – low, intermediate and high risk. The applied methodology was based on previous works (Thieler & Hammar-Klose, 1999; Thieler, 2000; Thieler & Pendleton et al., 2005; Rajawat et al., 2006; Hegde & Reju, 2007; Loureiro, 2006; Kumar et al., 2010), but specifically adapted to the local characteristics of the coastal stretch under study. Comparing our methodology with the above cited, there are three main distinguishing aspects: (i) the segmentation for the quantify- cation of indexes (susceptibility, exposure and risk) is also based on specific coastal types, obtained by a judi- cious delimitation based on the beach-dunes-urban front main characteristics, as well as human occupation and administrative boundaries; (ii) a new set of susceptibility variables is proposed, and others used in previous works (eg. tidal range; offshore significant wave height; mean sea level rise) are neglected since they showed almost uniform values along the whole studied coastal stretch, not adding information to the risk assessment; and (iii) the adopted exposure variables resulted from a detailed quantification of the existing building and beach areas extracted from the aerial digital photos. In a first phase a base line was defined, delimiting areas that were strongly occupied from those that have some capability to adapt to the effects of natural forcing factors (mainly winds and waves). The area was divided into 47 elementary segments considering the coastal typology, administrative boundaries and human occupation. For each segment a set of susceptibility indicators was quantified: coastal segment average elevation; width and slope; volume rate computed according to two DEMs obtained in different times; a wave-energy related indicator computed using wave propagation model results; and the percentage of vegetated area. The area of existing buildings in a predefined buffer and the area occupied by sandy beaches in each segment were Figure 8 - DTM of the Vila Nova de Gaia coast for considered as exposure indicators. The erosion risk in- November 2008, obtained through photogrammetry, and dex was finally computed based on two partial indexes: the division of the coast into 18 segments. the vulnerability index and the exposure index (Fig. 9). Figura 8 - MDT da costa de Vila Nova de Gaia para Computation of each vulnerability and exposure indi- Novembro de 2008, obtido por fotogrametria, e a divisão cators within each segment was supported by GIS tools da costa em 18 segmentos. using the relevant field data and DEM. Moreover,

58 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):47-63 (2015)

Susceptibility indicators Exposure indicators Erosion risk

Erosion risk index high intermediate low

Figure. 9 - Vulnerability and impact indicators used for erosion risk analysis; a part of the coast by Gulpilhares is shown. Figura 9 - Indicadores de vulnerabilidade e impacto utilizados na análise de risco de erosão; mostra-se uma parte da costa junto a Gulpilhares. models were applied to analyse local wave propagation the wave energy reaching each segment for a constant dynamics and estimate the susceptibility of the coast to energy source at the offshore model boundary. This a predefined wave event. This way, the model allows new approach allows assessing which coastal segments estimation of the wave energy near the coast (within a are most exposed to wave actions, mainly due to predefined buffer) resulting from the propagation of an shoaling, diffraction and refraction phenomena. offshore wave event. Model simulation results were used to discriminate 3.3.2. Results and discussion coastal segments in terms of susceptibility to wave The Vila Nova de Gaia monitoring program allowed, action. Wave propagation was modelled at specific for the first time, both a quantitative analysis of locations to identify local patterns useful for the under- morphological changes and an erosion risk assessment standing of the observed sedimentary dynamics (Granja for the whole coastal stretch of this municipality, which & Pinho, in press). The numerical model used is based is characterized by small pocket beaches between rocky on a Boussinesq equations formulation (Nawogu & De- outcrops, generating tombolos and creating specific mirbilek, 2001) and is able to simulate most hydrody- local and still poorly understood dynamics. This coastal namic phenomena that are of interest for coastal zones, area presents very diverse segments, with morphody- such as: silting-up; refraction; diffraction; com-plete namic behaviour influenced by local hydrodynamics, and partial reflection/transmission; bottom friction; sometimes conditioned by recent man-made structures non-linear wave-wave interactions; wave breaking; (Granja & Pinho, in press; Granja et al., 2011). Most wave-induced currents; and wave-current interaction. segments have been suffering erosion, some show The model was integrated in a hydroinformatic envi- accretion, partially supported by artificial sand nour- ronment appropriate for coastal processes modelling ishment, e.g. at Granja Beach, where this sand however (Pinho, 2001). The bathymetry used in the model is rapidly lost. The observed erosion trends constitute a construction was derived from the combination of a serious problem for this densely populated and build-up coarse bathymetric database (from Instituto Hidrográ- area, where the natural migration of the coastal system fico, 2010) available for this coastal zone and the is severely restricted. Besides susceptibility to erosion, higher-resolution bathymetric data obtained with the local ecological and economical values have to be ASV. Model results were processed in order to estimate considered and also evaluated.

59 Bio et al. (2014)

Considering the 18 administrative divisions of the The used complementary observation methods allowed: studied coastal stretch (Fig. 8), evolutionary trends, (i) identification of particular morphodynamic pro- based on the analysis of orthophoto-maps from 2005 cesses (using field observations), (i.e. previous to the monitoring program) to 2009, (ii) characterization of the nature of beach sediments showed erosion trends in segments 2, 3, 4 (northern (georeferenced sediment samples), part), 5, 9, 13, 15, 16, 17 and 18, and accretion trends in segments 4 (south), 11 (north) and 14. During this (iii) acquisition of high resolution beach mor-phology monitoring program (2008/2010), in November 2008, (MMS), recent erosion phenomena were observed for the (iv) extension of the DEM to vege-tated and other segments 2 to 9, whereas segments 10 to 13 and the sensitive areas (using the non-invasive aerial northern part of segment 14 were temporarily stable. photogrammetric method), and (v) characteriza-tion of the nearshore bottom (sub-tidal Two segments showed erosion due to man-made bathymetric surveys). structures: segment 5, where a reinforced sewage pipe acts as a groin, affecting sediment circulation by the Through this holistic approach a set of relevant data north-south drift and causing erosion south of the pipe; was obtained, concerning the fundamental morphody- and segment 15 and part of sector 14, which have lost namic processes that occur in the studied coastal sediments due to the breakwater at Aguda Beach. The segments, resulting from wave and wind action as well Aguda Beach breakwater had originally been planned as human interventions (coastal defence structures and as a detached breakwater to protect a small fishing dune fences), allowing erosion risk classification of the harbour. But after some months, still during the works coastal segments according to the aggregated score of of the breakwater construction, a tombolo wedge beach their vulnerability and exposure indexes (Fig. 10). formed against the temporary groin, promoting Erosion risk is higher for southern areas, with northern enlargement of the beach located up-drift and erosion at areas presenting a mixed behaviour. the down-drift coastal segment (Granja Beach). The temporary groin was not removed (Rosa-Santos et al., 2009) and the space between the breakwater and the harbour filled up, rendering the harbour useless (Rosa- Santos et al., 2009; Granja et al., 2011). Between November 2008 and November 2009, segments 3, 5, 10 and 12 showed negative sedimentary budgets, and segments 5, 14 (south) and 15 (affected by the above mentioned structures) suffered marked erosion. Artificial beach nourishment with sand had positive, though very short-lived, effects on the Granja beach. Sedimentary budgets became more favourable by May 2010, with accretions observed particularly in the upper, dry and vegetated part of the beaches (affected by wind conditions), whereas the wet part of the beach (affected by waves and currents) showed erosion in the most vulnerable sectors. Sedimentary analyses showed the presence of quite coarse sands, which are overall homogeneous in size class distribution, both alongshore and cross-shore, possibly related to artificial beach nourishment, with sediments coming from waste landfill sites. In later surveys, coarser sediments were hidden by finer ones, reflecting less ocean wave impact and fresh sediment supply. Sub-tidal surveys showed that beaches with low slope – both above and below the water line – were preferential accretion sites. Interpretation of the sonar data revealed accretion areas on the inner continental shelf (nearshore) and relationships between the Figure 10 - Erosion risk map for the Vila Nova de Gaia coast. morphodynamical behaviours of the nearshore and the Figura 10 - Mapa de risco de erosão para a costa de Vila contiguous foreshore. Nova de Gaia.

60 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):715-731 (2015)

4. Conclusion blue and near-infrared image bands. The innovative (terrestrial) videogrammetric approach used in the pre- Coastal monitoring programs should be set up based on sented case studies also provides a visual record of the the specific objectives and on the characteristics of the area, where ground features such as dune vegetation are area to be monitored. Selection of monitoring methods recorded. These records can be kept in GIS databases will obviously depend on type and accessibility of the and be further analysed. Sub-tidal bathymetric and area, on its size, and on the expected rates of change sonar surveys constitute an underwater extension of the (fast changes may require more frequent, i.e. easier to surveyed area, which is important because the sediment deploy, monitoring). Availability of equipment and types, structure and slope of the submerged area have a know-how will also influence the choice of methods. A direct impact on the degree of erosion and stability of cost-benefit analysis should help deciding whether it is, the contiguous foreshore, mainly due to cross-shore for instance, worth using the more expensive but faster transport. and non-invasive aerial methods, or the more limited but more precise and cheaper terrestrial survey The presented case studies demonstrate the potential of methods. several approaches, supported by advanced technolo- Topographic surveys should have a better than deci- gies, for the study of complex coastal morphodynamic metre precision to be useful. Traditional point or cross- processes. Both monitoring programs achieved their shore profile surveys (with total station technique) have proposed objectives and provided valuable information high precision but are very time and labour-intensive. to the local authorities, as gathered and processed The proposed terrestrial MMS methodology provides information constitutes a valuable database for coastal very precise 3D surveys (our precision of 5 cm is planning and for ICZM purposes. The monitoring comparable to that of Harley et al., 2007) and is easier program for the Vila Nova de Gaia coast is particularly to apply, weather resistant and low cost. Deployment of innovative as it combines several observation tech- the equipment is easy and fast, so that sudden changes, niques in a systematic way, in sequential campaigns, e.g. after severe storms, can easily be assessed (which is providing information for a detailed analysis with a not the case when using traditional techniques). The high spatial resolution. This combination allows a main limitation of this method is the production of detailed and continuous observation in a coastal buffer discrete spatial information, which is not very sensitive delimited by the urban front and the outer shoreface. to small forms when compared with continuous data. The relatively short length of the coastal stretch implies that other erosion risk assessment methodologies based Topography based on aerial stereo high-resolution on relatively rough data (e.g., wave climate variables, digital photography and photogrammetry is less precise tidal characteristics, medium to long term coastline but can provide spatially continuous data, important for dynamics; Thieler, 2000; Kumar et al., 2010), would an effective morphological assessment. Laser surveys fail to distinguish between different segments. constitute an alternative, but although they are more precise, they are currently extremely expensive. In comparison to terrestrial systems, aerial surveys are Acknowledgments more expensive and the equipment is not as easily We thank: the Administração dos Portos do Douro e Leixões deployed. But they have other advantages: they are (APDL) for the financial support of the data acquisition for better suited for regional monitoring, as vast areas (tens Douro Estuary sand spit study; INAG and EDP for the supply of kilometres) can easily be observed in a few minutes; of river data; the Parque Biológico de Gaia, SA. for the and, they are non-invasive. This is particularly import- financial funding of the Vila Nova de Gaia coast monitoring between 2008 e 2010; and Paulo Baptista for his help in the ant when monitoring ecologically sensitive areas, like collection and processing of the GPS data. This research was vegetated dunes, or estuarine marshes. Ground vehicles partially supported by the European Regional Development have to access the site to operate. Aerial photography Fund (ERDF) through the COMPETE – Operational also captures fragmented terrains and observes the Competitiveness Programme and national funds through FCT whole beach, including the intertidal region and fea- – Foundation for Science and Technology, under the project tures in the water, e.g. rocky outcrops. “Pest-C/ MAR/LA0015/2013”, and partially funded by the Project ECORISK (reference NORTE-07-0124-FEDER- Furthermore, the high-resolution images constitute a 000054), co-financed by the North Portugal Regional visual record of the study area, where vegetation, Operational Programme (ON.2 – O Novo Norte), under the buildings and infrastructure can be located and quanti- National Strategic Reference Framework (NSRF), through fied (through image analysis). Classification of, for the European Regional Development Fund (ERDF). Wave instance, vegetation can be done taking advantage of climate analyses were supported by the FCT-funded RAP the additional (though lower- resolution) red, green, project (PTDC/MAR/111223/2009).

61 Bio et al. (2014)

References Gonçalves, J.; Bastos, L.; Pinho, J.; Granja, H. (2011) – Digital aerial photography to monitor changes in coastal areas based Archetti, R.; Zanuttigh, B. (2010) – Integrated monitoring of the on direct georeferencing. 5th EARSeL Workshop on Remote hydro-morphodynamics of a beach protected by low crested Sensing of the Coastal Zone, Prague, Czech Republic. detached breakwaters. Coastal Engineering, 57(10):879–891. http://www.conferences.earsel.org/abstract/show/2689 DOI: 10.1016/j.coastaleng.2010.05.002 Granja, H.; Bastos, L.; Pinho, J.; Gonçalves, J.; Henriques, R.; Bio, Baptista, P.; Bastos, L.; Bernardes, C.; Dias, J. (2008) – Monitoring A.; Magalhães, A. (2011) – Small harbours risks: lowering for sandy shores morphologies by DGPS – a practical tool to fishery and increasing erosion. The case of Portinho da Aguda generate digital elevation models. Journal of Coastal Research, (NW Portugal). Littoral 2010, 09003, DOI: 10.1051/litt/ 24(6):1516−1528. DOI: 10.2112/07-0861.1 201109003 Barnard, P.L.; Warrick, J.A. (2010) – Dramatic beach and nearshore Granja, H.M.; Carvalho, G.S. (2000) – Inland beach migration morphological changes due to extreme flooding at a wave- (“beach erosion”) and the coastal zone management (the dominated river mouth. Marine Geology, 271(1-2):131–148. experience of the northwest coastal zone of Portugal). DOI: 10.1016/j.margeo.2010.01.018 Periodicum Biologorum, 102(suppl. 1):413–424. Bastos, L; Bio, A; Silva, J.; Pinho, J.L.S.; Granja, H. (2012) – Granja, H.; Pinho, J. L. (2012) – Coastal defense in NW Portugal: Dynamics of the Douro estuary sand spit before and after the improbable victory. In: Cooper, J.; Pilkey, H. (eds.), The breakwater construction. Estuarine, Coastal and Shelf Science, Pitfalls of Shoreline Stabilization, pp. 251−266, Coastal 109:53 69. DOI: 10.1016/j.ecss.2012.05.017 − Research Library, vol 3, Springer, USA. ISBN: 978- Boak, E.H.; Turner, I.L. (2005) – Shoreline definition and detection: 9400741232. DOI: 10.1007/978-94-007-4123-2_15 A review. Journal of Coastal Research, 21(4):688 703. DOI: − Granja H.; Pinho J. L. (in press) – A multi-criteria approach for 10.2112/03-0071.1 erosion risk assessment using a new concept of spatial unit Brilakis, I.; Fathi, H.; Rashidi, A. (2011) – Progressive 3D analysis, wave model and high resolution DEMs. In: Charles, J.; reconstruction of infrastructure with videogrammetry. Finkl, W., Makowski, C. (eds.), Advances in Coastal and Automation in Construction, 20(7):884−895. DOI: Marine Resources: Remote Sensing and Modeling. Coastal 10.1016/j.autcon.2011.03.005 Research Library, Springer, USA. ISBN: 978-3319063256. Consulmar (1996) – Estudos das obras necessárias à melhoria da Harley, M.D.; Turner, I.L.; Short, A.D.; Ranasinghe, R. (2007) – acessibilidade e das condições de segurança na barra do Monitoring beach processes using conventional, RTK-GPS and Douro. Consulmar Report, Administração dos Portos do Douro image-derived survey methods: Narrabeen Beach, Australia. In: e Leixões (in Portuguese), Lisboa, Portugal. Unpublished. Woodroffe, C.D.; Bruce, E.M.; Puotinen, M.; Furness, R.A. Costa, M.; Silva, R.; Vitorino, J. (2001) – Contribuição para o (eds.), GIS for the Coastal Zone: A Selection of Papers from Estudo do Clima de Agitação Marítima na Costa Portuguesa. CoastGIS 2006, pp. 151–164. University of Wollongong, Actas das 2as Jornadas Portuguesas de Engenharia Costeira e Australia. ISBN 9781741281378 Portuária. Associação Internacional de Navegação, Sines, Hegde, A.V.; Reju, V.R. (2007) – Development of coastal Portugal (in Portuguese). Available on-line at vulnerability index for Mangalore coast, India. Journal of ftp://ftp.mohid.com/Fortaleza_CD/Bibliografia/Waves/clima_agmar.pdf Coastal Research, 23(5):1106–1111. DOI: 10.2112/04-0259.1 Cowart, L.; Walsh, J.P.; Corbett, D.R. (2010) – Analyzing Estuarine Henriques, R. (2004) – SEDMAC/SEDPC: An application to Shoreline Change: A Case Study of Cedar Island, North support particle size analysis of unconsolidated sediments. 32nd Carolina. Journal of Coastal Research, 26(5):817–830. DOI: International Geological Congress, Florence-Italy, abs. 154-6, 10.2112/JCOASTRES-D-09-00117.1 p. 726. Cruz, J. (2008) – Ocean Wave Energy. Current Status and Future Holland, K.T.; Vinzon, S.B.; Calliari, L.J. (2009) – A field study of Perspectives. Green Energy and Technology. 431p., Springer, coastal dynamics on a muddy coast offshore of Cassinobeach, Berlin. ISBN 978-3-540-74895-3 Brazil. Continental Shelf Research, 29(3):503–514. DOI: Dail, H.J.; Merrifield, M.A.; Bevis, M. (2000) – Steep beach 10.1016/j.csr.2008.09.023 morphology changes due to energetic wave forcing. Marine Geology, 162(2-4):443–458. DOI: 10.1016/S0025-3227(99) Jacobsen, K. (2000) – User Manual Program System BLUH, 444p., 00072-9 Institute for Photogrammetry and Engineering Surveys, University of Hannover. Available via http://www.ipi.uni- Davidson, M.; Van Koningsveld, M.; de Kruif, A.; Rawson, J.; hannover.de/kj.html?&L=1 Holman, R.; Lamberti, A.; Medina, R.; Kroon, A.; Aarninkhof, S. (2007) – The CoastView project: Developing video-derived Kumar, T.S.; Mahendra, R.S.; Nayak, S.; Radhakrishnan, K.; Sahu, Coastal State Indicators in support of coastal zone management. K.C. (2010) – Coastal Vulnerability Assessment for Orissa Coastal Engineering, 54(6-7):463–475. DOI: 10.1016/ State, East Coast of India. Journal of Coastal Research, j.coastaleng.2007.01.007 26(3):523−534. DOI: 10.2112/09-1186.1 Dodet, G.; Bertin, X.; Taborda, R. (2010) – Wave climate variability Loureiro, E. (2006) – Indicadores geomorfológicos e sedimentoló- in the North-East Atlantic Ocean over the last six decades. gicos na avaliação da tendência evolutiva da zona costeira: Ocean Modelling, 31(3-4):120–131. DOI: 10.1016/j.ocemod. aplicação ao concelho de Esposende. 329 p., PhD Thesis, 2009.10.010 Universidade do Minho (in Portuguese). Available on-line at http://hdl.handle.net/1822/7096 Ferreira, H.; Almeida, C.; Martins, A.; Almeida, J.; Dias, N.; Dias, A.; Silva, E. (2009) – Autonomous bathymetry for risk Madeira, S. (2007) – Sistema Móvel Terrestre de Levantamento com assessment with ROAZ robotic surface vehicle. OCEANS 2009 Integração em SIG. PhD Thesis, Porto University (in IEEE Bremen, 1−6. Available on-line at http://wrem2012.isr. Portuguese). Unpublished uc.pt/Papers/Ferreira_WREM2012.pdf Madeira, S.; Gonçalves, J.; Bastos, L. (2009) – Fast Camera Fletcher, C.; Rooney, J.; Barbee, M.; Lim, S.C.; Richmond, B. Calibration for Low Cost Mobile Mapping. Proceedings of the (2003) – Mapping shoreline change using digital ortho- 6th International Symposium on Mobile Mapping Technology photogrammetry on Maui, Hawaii. Journal of Coastal Research, (MMT09), 21−24 July 2009, Presidente Prudente, São Paulo, SI38:106−124. Available on-line at ftp://ftp.soest.hawaii.edu Brazil. Available on-line at http://docs.fct.unesp.br/departamentos/ /coastal/webftp/JCR_MappingShoreline.pdf cartografia/eventos/2009_MMT/_papers/303_Madeira_MMT09.pdf 62 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):715-731 (2015)

Madeira, S.; Gonçalves, J.A.; Bastos, L. (2010) – Photogrammetric Rajawat, A.S.; Bhattacharya, S.; Jain, S.; Gupta, M.; Jayaprasad, P.; mapping and measuring application using MATLAB. Tamilarasan, V.; Ajai; Nayak, S. (2006) – Coastal Vulnerability Computers and Geosciences, 36(6):699−706. DOI: Mapping for the Indian Coast. Second International Symposium 10.1016/j.cageo.2009.06.015 on Geoinformation for Disaster Management. Muehe, D. (2011) Erosão Costeira - Tendência ou Eventos Rasmussen, D.J.; Holloway, T.; Nemet, G.F. (2011) – Opportunities Extremos? O Litoral entre Rio de Janeiro e Cabo Frio, and challenges in assessing climate change impacts on wind Brasil/Coastal Erosion. Revista da Gestão Costeira Integrada, energy – a critical comparison of wind speed projections in 11(3):315−325. DOI: 10.5894/rgci282 California. Environmental Research Letters, 6, 9pp. DOI: 10.1088/1748-9326/6/2/024008. Nawogu, G.; Demirbilek, Z. (2001) – BOUSS-2D: A Boussinesq Wave Model for Coastal Regions and Harbors. Report 1. Rieke-Zapp, D.H.; Nearing M.A. (2005) – Digital Close Range Theoretical Background and User’s Manual. US Army Corps of Photogrammetry for Measurement of Soil Erosion. The Engineers. Engineer Research and Development Center. USA. Photogrammetric Record, 20(109):69−87. DOI: 10.1111/j.1477- Available on-line at http://www.dtic.mil/cgi-bin/Get TRDoc?AD= 9730.2005.00305.x ADA400294. Rocha, C. P.; Araújo, T.C.M.; Mendonça, F.J.B. (2009) – Aplicação Nicholls, R.; Wong, P.; Burkett, V.; Codignotto, J.; Hay, J.; de metodologia alternativa para localizar e monitorar linhas de McLean, R.; Woodroffe, S.R.C. (2007) – Coastal systems and costa usando técnicas de posicionamento pelo GNSS: Um low-lying areas. In: Parry, M.; Canziani, O.; Palutikof, J.; estudo de caso na praia de Sauaçui, Nordeste do Brasil. Revista Linden, P.V.D.; Hanson, C. (eds.) Climate Change 2007: da Gestão Costeira Integrada, 9(1):93-108. DOI: 10.5894/ Impacts, Adaptation and Vulnerability. Contribution of Working rgci151 Group II to the Fourth Assessment Report of the Rosa-Santos P., Veloso-Gomes F., Taveira-Pinto F., Silva R., Intergovernmental Panel on Climate Change. pp. 315-357, Pais-Barbosa J., 2009. Evolution of Coastal Works in Cambridge University Press, Cambridge, U.K. ISBN 978 0521 Portugal and their Interference with Local Morphodynamics, 88010-7, Available on-line at https://www.ipcc.ch/publications Journal of Coastal Research, Special Issue 56:757-761. _and_data/publications_ipcc_fourth_assessment_report_wg2_report_im Available on-line at http://www.cerf-jcr.org/images/stories pacts_adaptation_and_vulnerability.htm /757.761_P.Rosa-Santos_ICS2009.pdf Niedermeier, A.; Romaneessen, E.; Lehner, S. (2000) – Detection of Short, A.D.; Trembanis, A.C. (2004) – Decadal scale patterns in coastlines in SAR images using wavelet methods. IEEE Trans. beach oscillation and rotation Narrabeen Beach, Australia — Geoscience and Remote Sensing, 38(5):2270−2281. DOI: time series, PCA and wavelet analysis. Journal of Coastal 10.1109/36.868884 Research 20(2):523−532. DOI: 10.2112/1551- Niesing, H. (2005) – EUROSION: Coastal erosion measures, 5036(2004)020[0523:DSPIBO]2.0.CO;2 knowledge and results acquired through 60 studies. In: Herrier, Silva, A.J.R.; Abecasis, C.; Leitão, J.C. (2005) – Improving the J.-L.; Mees, J.; Salman, A.; Seys, J.; Van Nieuwenhuyse, H.; navigability and safety conditions in Douro estuary inlet. In: Dobbelaere, I. (eds.), Proceedings ‘Dunes and Estuaries 2005’ – Smith, J.M. (ed.) Coastal Engineering 2004, 19–24 September International Conference on Nature Restoration Practices in 2004, vol 3. pp. 3277–3289, World Scientific, Singapore. ISBN: European Coastal Habitats, pp. 421−431, Koksijde, Belgium, 978-981-256-298-2 19-23 September 2005, VLIZ Special Publication 19. Available on-line at http://ec.europa.eu/environment/life/project/Projects /index. Soares de Carvalho, G.; Granja, H.; Costa, A.L. (2006) – Dois casos cfm?fuseaction=home.showFile&rep=file&fil=LIFE02_NAT_B_00859 de mudanças antrópicas na faixa costeira (praias e dunas) do 1_Seminar.pdf noroeste de Portugal (segmentos costeiros de Leça da Palmeira- Pendleton, E.A.; Thieler, E.R.; Williams, S.J. (2005) – Coastal estuário do Douro e de Aguda-Espinho). Geonovas 20:13–25. Vulnerability Assessment of National Park of American Samoa Souza, C.R. de G.; Luna, G. da C. (2010) – Variação da linha de to Sea-Level Rise. U.S. Geological Survey Open-File Report costa e balanço sedimentar de longo período em praias sob risco 2005-1055. Available on-line at http://pubs.usgs.gov/of/2005/1055/ muito alto de erosão do município de Caraguatatuba (Litoral images/pdf/report.pdf Norte de São Paulo, Brasil). Revista da Gestão Costeira Pereira, C.; Coelho, C. (2013) – Mapas de Risco das Zonas Integrada 10(2):179-199. DOI: 10.5894/rgci176 Costeiras por Efeito da Ação Energética do Mar. Revista da Thieler, E.R., 2000. National Assessment of Coastal Vulnerability to Gestão Costeira Integrada, 13(1):27-43. DOI: 10.5894/rgci325 Future Sea-Level Rise. USGS Fact Sheet, fs-076-100. Available Pinho, J.L.S. (2001) – Mathematical modelling application to on-line at http://pubs.usgs.gov/fs/fs76-00/fs076-00.pdf hydrodynamics and water quality studies of coastal zones. PhD Thieler, E.R.; Hammar-Klose, E.S. (1999) – National Assessment of Thesis. University of Minho, Braga, Portugal (in Portuguese). Coastal Vulnerability to Sea-Level Rise: Preliminary results for Unpublished the U.S. Atlantic Coast. U.S. Geological Survey Open-File Portela, L.I. (2002) – Plano específico de extracção de inertes do Report 99–593. Available on-line at http://pubs.usgs.gov/of/ 1999/of99-593/ rio Douro: caracterização morfodinâmica do estuário do rio Douro. Report 215/02-NEC, 42p., LNEC, Lisbon, Portugal. (in Zhu, Z.; Brilakis, I. (2009) – Comparison of optical-sensor-based Portuguese). spatial data collection techniques for civil infrastructure modelling. Journal of Computing in Civil Engineering, R Development Core Team (2009) – R: A language and envi- 23(3):170–177. DOI: 10.1061/(ASCE)0887-3801(2009) 23:3 ronment for statistical computing. R Foundation for Statistical (170) Computing, Vienna, Austria. ISBN 3-900051-07-0. Available on- line at http://www.R-project.org.

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015)

http://www.aprh.pt/rgci/pdf/rgci-515_Guerreiro.pdf | DOI:10.5894/rgci515

Evolution of the hydrodynamics of the Tagus estuary (Portugal) in the 21st century * @

a @, a a a b Martha Guerreiro , André Bustorff Fortunato , Paula Freire , Ana Rilo , Rui Taborda , Maria Conceição Freitasc, César Andradec, Tiago Silvac, Marta Rodriguesa, d a Xavier Bertin , Alberto Azevedo

Abstract The ongoing rise in sea level affects tidal propagation and circulation in estuaries, and these changes can have far reaching consequences on the sediment dynamics, water quality and extreme water levels. This study aims at anticipating the evolution of the tidal dynamics in the Tagus (Portugal) in the 21st century, in particular due to sea level rise (SLR). The existence of a resonance mode of about 8 hours in this estuary, that selectively amplifies both semi-diurnal and quarter-diurnal tidal constituents, makes the response of the Tagus estuary to SLR unique. The study was conducted with a shallow water model, forced by present and future conditions, namely higher mean sea levels and an extrapolated bathymetry based on present sedimentation rates. Model results showed that SLR will significantly affect tidal asymmetry, in particular because the intertidal area can decrease by up to 40% by the end of the 21st century. As a result, the strong ebb-dominance of this estuary will decrease significantly. This evolution of tidal asymmetry will be counteracted by the effect of sedimentation of the salt-marsh areas. Also, SLR will enhance the resonance in the Tagus estuary. As a consequence, extreme water levels will be higher than the sum of present levels with the SLR. Keywords: tidal asymmetry, resonance, climate change, SELFE, modeling.

Resumo Evolução da hidrodinâmica do estuário do Tejo (Portugal) no século XXI A subida do nível médio do mar afeta a propagação da maré e a circulação em estuários, podendo estas alterações ter conse- quências significativas na dinâmica sedimentar, na qualidade da água e nos níveis de água extremos. O presente estudo visa antecipar a evolução da dinâmica de maré no estuário do Tejo (Portugal) no século XXI, em particular devido à subida do nível médio do mar (NMM). A existência de um modo de ressonância de cerca de 8 horas neste estuário, que amplifica seleti- vamente as constituintes semi-diurnas e quarto-diurnas, torna a resposta do estuário do Tejo à subida do NMM única. O estudo foi efetuado com um modelo hidrodinâmico, forçado por condições presentes e futuras, nomeadamente níveis médios do mar mais elevados e uma batimetria extrapolada com base em taxas de assoreamento atuais. Os resultados do modelo mostram que a subida do NMM irá afetar significativamente a assimetria da maré, em particular porque as áreas intertidais podem diminui até 40% até ao final do século. Em consequência, a forte dominância de vazante deste estuário irá diminuir significativamente. Esta evolução da assimetria da maré será parcialmente compensada pelo efeito da sedimentação nas áreas intertidais. A subida

@ Corresponding author: a National Civil Engineering Laboratory, Av. do Brasil, 101, 1700-066 Lisbon, Portugal. b University of Lisbon, Faculty of Science, LATTEX, IDL., Bloc C-6, 2nd Floor, Campo Grande, 1749-016 Lisbon, Portugal. c University of Lisbon, Faculty of Science, Geology Department, Geology Center, Bloc C-6, Campo Grande, 1749-016 Lisbon, Portugal. d CNRS-Université de La Rochelle, UMR 7266 LIENSS, Institut du Littoral et de l’Environnement, 2 rue Olympe de Gouges, 17 000 La Rochelle, France.

* Submission: 29 MAY 2014; Peer review: 6 JUL 2014; Revised: 19 SEP 2014; Accepted: 22 SEP 2014; Available on-line: 25 SEP 2014

Guerreiro et al. (2015) do NMM irá também amplificar o efeito de ressonância no estuário do Tejo, o que resultará em níveis extremos superiores à soma dos níveis extremos atuais com a subida do NMM. Palavras-chave: assimetria de maré, ressonância, alterações climáticas, SELFE, modelação.

1. Introduction In this context, this study aims at assessing how the hydrodynamics of the Tagus estuary will evolve in the Climate change is affecting the oceans. Rising tempera- 21st century. The assessment was performed using a tures are melting icecaps and glaciers and expanding high-resolution hydrodynamic model, which was cali- the volume of the oceans. As a result, global mean sea brated and validated with field data. The model was level (MSL) is rising, affecting coastal regions in gen- then applied to study likely scenarios of future MSL eral and estuaries in particular. For example, the in- and bathymetries (Table 1). SLR scenarios for the 21st crease in estuarine tidal amplitudes that can result from century were defined from a literature review. Future sea level rise (SLR) can affect the water residence time bathymetries were estimated from the extrapolation of and quality, promote salt-wedge intrusion (Hong & past sedimentation trends, derived from field data. Pos- Shen, 2012) and exacerbate problems of estuarine mar- sible changes in wind patterns associated to climate ginal inundation (Bilskie et al., 2014). Simultaneously, change were briefly assessed and considered negligible. habitats presently found along the estuarine banks may migrate landward (Pethick, 2001). The adequate man- This paper is organized as follows. The Tagus estuary is agement of estuaries requires therefore the anticipation first presented. Then, the methods adopted in the study of the consequences of SLR on their behavior. are described, including the model and its application, The Tagus estuary in Portugal is a particularly relevant and the present and future scenarios considered. Results case. It is one of the largest in Europe and it is included are then presented and discussed. The final section in the territorial unit of Lisbon and Tagus Valley, in- summarizes the major conclusions and proposes direc- volving 18 municipalities in the metropolitan area of tions for further research. Lisbon, with about one million inhabitants directly or 2. The Tagus Estuary indirectly exposed (INE, 2012). It harbors a major port and shipping terminal with about 3000 large vessels The Tagus estuary is located on the Portuguese west entering the harbor annually (Porto de Lisboa, 2013). coast (Figure 1). It covers ca 320 km2, with a deep, long The estuary accommodates economic activities, such as and narrow tidal inlet connecting the Atlantic Ocean to trade, maritime traffic, dredging, industry and fisheries. a shallow, tide-dominated basin, with extensive tidal A substantial part of the estuarine domain is protected flats and marshes that cover about 40% of the inner due to high environmental and ecological values, gen- estuary. About 40 km upstream, the estuary markedly erating conflict with other economically-driven uses. narrows at the bay head. The saline tide reaches about Thus, the future rise in sea level can have severe eco- 50 km upstream from the mouth, near Vila Franca de nomic and environmental implications. Xira. The estuarine bottom is mainly composed of silt

Figure 1 - Location of the Tagus estuary. Figura 1- Localização do estuário do Tejo.

66 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015) and sand, of both fluvial and local origins; marine sands high Courant numbers, and parallelization of the code are confined to the mouth and inlet channel (Freire et offers computational efficiency. SELFE is the core of a al., 2007). Tidal ranges vary between 0.55 and 3.86 m suite of community models, which include modules for in the open coast (Cascais data) but resonance signify- ecosystem dynamics (Rodrigues et al., 2009, 2012), cantly amplifies the semi-diurnal tidal constituents fecal contamination dispersion (Rodrigues et al., 2011), within the estuary (Fortunato et al., 1999). Simulta- oil spill dynamics (Azevedo et al., 2014), sediment neously, the estuary is strongly ebb-dominated due to transport and morphodynamics (Pinto et al., 2012; Do- the large extent of the tidal flats (Fortunato et al., 1999). det et al., 2013) and wave-current interaction (Bruneau et al., 2011, Roland et al., 2012). Table 1 - Summary of the simulations. 3.2. Application, calibration and validation Tabela 1 - Sumário das simulações. Since the late 1980's many hydrodynamic models of the Purpose Bathymetry Boundary conditions Tagus estuary have been implemented (e.g., ADCIRC – Validation Present Real tides Fortunato et al., 1997, 1999; MOHID – Portela & Ne- ves, 1994; ELCIRC – Vargas et al., 2008; SIMSYS – Tidal Asymmetry Present Real tides Dias & Valentim, 2011). In general, barotropic models Present Real tides + SLR have been used, since the estuary is well mixed or par- Future Real tides tially mixed, except in peak discharge situations (Ne- Resonance Present Monochromatic tides ves, 2010). One main limitation of these models is their Extreme levels Present Extreme events poor horizontal resolution, unable to properly resolve all channels (Fortunato & Oliveira, 2000). Hence, a Extreme events + SLR barotropic model was run using a computa-tional grid with significantly higher resolution than used in previ- 3 The average river flow is 368 m /s (Neves, 2010), and ous studies. The fine resolution is also necessary to the estuary is usually well mixed. However, stratifica- properly represent the marginal inundation (e.g., Martyr tion has been observed at high flow rates (Neves, 2010). et al., 2013, Bertin et al., 2014). River discharge may significantly influence water lev- The horizontal domain was represented by an unstruc- els, but only farther than 40 km upstream of the mouth tured mesh with 77,300 nodes that extends from the (Vargas et al., 2008). Downstream, the levels are river (100 km upstream from the mouth) to the open mainly controlled by tide and storm surges. ocean (Figure 2). The horizontal limits of the grid were Ocean waves do not penetrate significantly in the estu- defined based on the limits of the water line at maxi- ary. However, the large extent (fetch) of the estuary mum high spring tide, defined for the Tagus estuary by allows locally-generated waves to develop and rework Rilo et al. (2014), which was further extended inland by the southern embankment (Freire & Andrade, 1999). 50 m in the southern margin of the estuary. This exten- The population growth and development of port activity sion aimed at accommodating a possible extension of has affected both margins, which are extensively inter- the estuary in the SLR scenarios. The grid was not ex- vened by construction to support nautical activities, tended in the northern margin of the estuary because industrial facilities, farming plots, and urban areas. most of this margin is highly urbanized and protected. These factors increased the pressure on the estuarine Hence, it was assumed that retreat will not be socially waterfront and prevented the natural evolution of the acceptable in this area. majority of the margins. The Tagus estuary harbors the The grid resolution is about 25 m in the upper estuary main Portuguese shipping terminal and the nautical and 2000 m in the ocean. The bathymetry combines the activity requires frequent maintenance dredging along latest available data, surveyed between 1964 and 2011. the estuary. Simulations were made for 30 days, with a time step of 30 s. The ocean tidal boundary conditions were taken 3. Methods from the Cascais tidal gauge, located close to the outer 3.1. Model description estuary, through harmonic analysis and synthesis. This study was conducted with the community model The Manning friction coefficient varies between 0.015 SELFE – Semi-implicit Eulerian-Lagrangian Finite- and 0.023 m1/3/s (Figure 3) and was used as a calibra- Element (Zhang & Baptista, 2008). SELFE is a 3D tion parameter, but its spatial variability was deter- baroclinic shallow-water model, although the 2D mined based on the distribution of the bottom sediments barotropic version (Zhang et al., 2011) is used herein. (Freire, 2003). The model was calibrated using the ele- SELFE uses unstructured triangular meshes, which vation data collected in 1972 at 13 stations along the makes it particularly suited to problems with different estuary (Figure 4). The root mean square errors ob- spatial scales and complex geometries. The Eulerian- tained (Figure 4) are smaller than those obtained in Lagrangian semi-implicit algorithm provides stability at previous studies of this estuary.

67 Guerreiro et al. (2015)

Figure 2 - Finite element grid and bathymetry of the upper estuary. The bathymetry is relative to Chart Datum, 2.21 m below present MSL. Figura 2 - Malha de elementos finitos do modelo hidrodinâmico do Tejo e batimetria da área montante do estuário. A batimetria está referida ao Zero Hidrográfico, 2.21 m abaixo do nível médio do mar atual.

Figure 3 - Distribution of the Manning coefficient (m1/3/s). Figura 3 - Distribuição do coeficiente de Manning (m1/3/s).

68 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015)

Figure 4 - a) Tidal gauges and points used in the analysis; b) root mean square errors at the tidal gauges. Figura 4 - a) Marégrafos e pontos usados na análise; b) erros quadráticos médios nas estações maregráficas.

3.3. Scenarios ice sheets. In this context, two mean SLR scenarios were defined: the “best estimate” (0.5 m), which corre- 3.3.1. Sea Level Rise sponds to the central value of the various estimates th During the 20 century, MSL along the Portuguese published after 2007, and a “high-end value” (1.5 m), coast rose by about 0.15 m (Antunes & Taborda, 2009), accepted as suitable for use in coastal zone manage- a value consistent with the global estimates for mean ment, since it satisfies the precautionary principle. It SLR (e.g., Church and White, 2006; Hagedoorn et al., must be noted that the last IPCC assess-ment report, 2007). This agreement stems from negligible glacial- published after this work was concluded, suggests isostatic adjustment and vertical movements in the re- global mean SLR by the year 2100 that vary between gion (Antunes & Taborda, 2009). Hence, future local 0.26 and 0.98 m (IPCC, 2013). The present sea level sea level scenarios for the region can be based on global was estimated as 2.21 m above Chart Datum. estimates. GPS-based methods indicate that this area has an upward vertical movement which is an order of 3.3.2. Wind magnitude smaller than SLR (http://www.sonel.org/spip. Results of the global climate model ECHAM5 in front php?page=gps&idStation=648.php) of the Tagus estuary, for IPCC (2007) scenario A2, The magnitude of future SLR remains controversial reveal modest differences between the statistics of wind among the scientific community and projections still at the end of 20th (1970-2000) and 21st (2070-2100) present a large uncertainty. Most estimates for SLR at centuries (Figure 5). Also, global climate models cannot the end of the 21st century published after the 2007 re- yet reproduce adequately the historical trends, and at- port of the Intergovernmental Panel on Climate Change tempts at downscaling present a variability between (IPCC, 2007) vary between 0.2 and 2 m (Rahmstorf, models higher than the signal associated with climate 2010). The discrepancies are related not only with un- change (Pyor & Barthelmie, 2010). The same authors st certainties in assessing future global temperatures, but consider unlikely that, during the 21 century in Europe, also with difficulties in evaluating the contribution of wind speed and energy density will change more than melting of both glaciers and ice caps and uncertainty the present inter-annual variability (i.e., ± 15%). Thus, concerning the dynamics of Antarctic and Greenland the wind regime was considered invariant. 69 Guerreiro et al. (2015)

Figure 5 - Evolution of the wind regime near the Tagus estuary between the end of the 20th (1970- 2000) and 21st centuries (2070-2100), from the results of the model ECHAM5, for the scenario A2 of the IPCC (2007). Figura 5 - Evolução do regime de vento nas imediações do estuário do Tejo entre final do século XX (1970-2000) e século XXI (2070-2100), dos resultados do modelo ECHAM5 para o cenário A2 do IPPC (2007).

70 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015)

3.3.3. Sedimentation rates components, synthetic annual series were generated by The bathymetric charts available for different areas of combining: the estuary, dated from 1928/32, 1964, 1981, 2009 and 1. The average level determined for a particular year 2011, were analyzed and used to build digital terrain (present or projected, according to a sea level sce- models, which were further compared to determine and nario). map sedimentation rates. The intertidal areas show an 2. The astronomical tide for a set of 19 consecutive average infilling trend ca. 0.3 cm/year. The main chan- years, which allows the consideration of the lunar nels show a slight erosional trend (mean value ca. cycle (ca. 18.6 years). -0.1 cm/year), which partly reflects continuous dredging of navigation channels. 3. Residuals, with a variable phase lag with the tidal signal. These phase lags vary with one hour inter- Sedimentation rates were also inferred for marginal vals between plus and minus 15 days, thereby en- marsh expansions at several locations, by extrapolation 210 137 suring that a given storm surge can occur at any of the isotopic ( Pb, Cs) vertical profiles in sedi- moment in the neap-spring tidal cycle. ment cores. Results indicate that sedimentation rates in these areas vary between 0.7 and 2.2 cm/year with the Like most other approaches to determine extreme sea higher values occurring in the upper estuary (Silva et levels (e.g., Pugh & Vassie, 1980; Tawn, 1992), this al., 2013). A map of sedimentation rates was built by procedure implicitly neglects tide-surge interactions. integrating all these data, and used to project future These interactions, which are still poorly understood, bathymetries (Guerreiro et al., 2013). This projection are believed to be associated to different phenomena. was based on an extrapolation of the present bathyme- First, tidal propagation can be accelerated (decelerated) try, assuming sedimentation rates that are constant in by a positive (negative) surge (Horsburgh & Wilson, time, but variable in space. Since the accretion rates are 2007). This acceleration leads to a phase shift between modest (i.e., below 0.3 cm/year in most of the domain), the measured and predicted tide, which translates into the predicted variations in bottom elevations do not positive (negative) residuals during the rising (falling) exceed a few tens of cm. tide. Secondly, the wind contribution (i.e., the wind stress divided by depth in the shallow water equations) 3.3.4. Extreme water levels is inversely proportional to the water depth. The storm Flooding of the estuarine margins, of both marine and surge will thus tend to be higher at low tide, in particu- fluvial origin, is an increasing concern. Presently, the lar in shallow areas with a large tidal range. European Floods Directive requires all member states to Other mechanisms may be present, such as the increase develop inundation maps for areas at risk. In estuarine in bottom friction at low tide (Rego & Li, 2010) and areas, these floods are often due to the combined effect advection (Idier et al., 2013). Hence, the relation be- of tides and storm surges. SLR will thus aggravate in- tween the tide and the residual is complex and strongly undation hazards and in the case of the Tagus, tidal space-dependent. Yet, it is clear that all these mecha- amplification by resonance can exacerbate this problem. nisms are related to the non-linear terms in the shallow To determine extreme forcing conditions, we followed water equations. Since the continental shelf in front of the approach developed by Fortunato et al. (2013). This the Tagus estuary is very narrow (20-30 km), it is un- approach analyses data from tide gauges to determine likely that tide-surge interactions are significant outside both the extreme levels corresponding to various return of this estuary. periods and the time series that include these extremes. These series are then used to force a hydrodynamic This procedure provides a very large number of hypo- model of the estuary and construct inundation maps. thetical annual series for each MSL scenario: in the present case, 328,320 (i.e., 24x19x30x24) series. The The determination of the time series associated to vari- set of all maxima in all series makes it possible to de- ous return periods starts from a dataset of elevation termine the probability of occurrence (P) of a certain measurements. In this case, we used a 24 year-long data series of hourly water elevation measured at the Cascais extreme level (z). tide gauge, extracted from the 1961 – 2005 series, after Traditional analyses of extreme water levels work with discarding years with significant information gaps. limited numbers of extreme observations (typically a Gaps were considered significant when they included few tens of years). Statistical distribution functions periods with high waves. Waves were determined from must therefore be fit to the data to extrapolate the the hindcast simulations of Dodet et al. (2010). The maximum levels for high return periods. In contrast, the measured signal was split through harmonic analysis in approach used herein leads to a number of annual the following components: the yearly MSL, the astro- maxima large enough to avoid the use of analytic func- nomical tidal level and the residual, which corresponds tions. Finally, the return period associated with a certain to the signal of meteorological origin. From these three level is given by T(z) =1/(1-P(z)) (Figure 6a).

71 Guerreiro et al. (2015)

Figure 6 - Boundary conditions for the inundation simulations: maximum elevations at Cascais for different return periods (a) and time series to force the model (b). ZH stands for Hydrographic Zero, the Portuguese Chart Datum (close to lower low tide). Presently, ZH is about 2.21 m below MSL. Figura 6 - Condições de fronteira para as simulações de inundação: elevações máximas em Cascais para diferentes períodos de retorno (a) e séries temporais (b) para forçar o modelo. ZH corresponde ao Zero Hidrográfico, atualmente cerca de 2.21 m abaixo do NMM.

The values obtained in this study are slightly higher The European Floods Directive recommends that three than those reported by Andrade et al. (2006) (Table 2). cases are studied: a high, a medium and a low probabil- Fortunato et al. (2013) observed similar differences ity scenario. In Portugal, these scenarios were taken as between the extreme values obtained by the method those corresponding to return periods of 20, 100 and presented here and traditional methods, and showed that 1000 years. For consistency with future studies, these the former yields more accurate results. return periods were considered herein, even though the uncertainty on the water levels associated to the 1000- Table 2 - Comparison between the maximum water levels (m year return period is clearly very high. above Chart Datum) obtained in this study and with traditional methods (source: Andrade et al., 2006). 4. Results and discussion Tabela 2 - Comparação entre os níveis extremos (m, acima 4.1. Tidal asymmetry do ZH) obtidos neste estudo com métodos tradicionais Ocean tides are usually symmetrical, i.e. floods and (fonte: Andrade et al, 2006) ebbs have similar durations. Non-linear processes, im- Return period (years) portant in shallow areas, generate high-frequency har- Method 5 10 25 50 100 monic constituents that may distort the tide (e.g., Aubrey & Speer, 1985; Speer & Aubrey, 1985). Typi- This study 4.24 4.28 4.33 4.37 4.42 cally, floods are shorter than ebbs when the ratio be- Traditional 4.1 4.2 4.3 4.3 4.4 tween the tidal amplitude and depth is low. In contrast, extensive intertidal flats favor shorter ebbs (Friedrichs To determine the series used to force the model, the & Aubrey, 1988, Fortunato & Oliveira, 2005), as is the synthetic annual time series (from the set of 328,320 case in the Tagus estuary (Fortunato et al., 1999). series) containing a maximum level corresponding to Tidal asymmetry is an important characteristic of each the selected return period are retained. The initial time estuary, and there is a large body of literature on the for each of these series is adjusted, so that the time of subject (e.g., Brown & Davies, 2010, Jewell et al., maximum elevation coincides, and then the series are 2012). Tidal asymmetry is particularly relevant to averaged to provide the final forcing (Figure 6b) This sediment dynamics (Aldridge, 1997). Shorter ebbs approach only provides the maximum levels associated promote higher average flow velocities on ebb than on with events of marine origin. Vargas et al. (2008) flood because the same volume of water flows in a showed that the Tagus flow has a negligible influence shorter period of time. Under those circumstances, the on the maximum water levels observed in the lower estuary is said to be ebb-dominant. Since the sediment 40 km of the estuary, thus neglecting the effects of this fluxes depend non-linearly on the velocity, an ebb- variable does not significantly influence the results. dominated estuary will tend to export sediments. In

72 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015) contrast, a flood-dominant estuary will tend to silt-up importantly, the extent of the tidal flats will decrease, more rapidly (Lanzoni & Seminara, 2002). further reducing ebb dominance: the intertidal area in Since SLR will affect the generation of tidal harmonics the Tagus estuary decreases by 40% for a SLR of 1.5 m by nonlinear processes, it is likely to affect the pattern of (Figure 8). Note that the reduction of intertidal flats tidal asymmetry. The effects of SLR on tidal asymmetry obtained herein implicitly assumes that the existing were assessed by evaluating the difference between ebb margins would be protected and prevented from flood- and flood durations along the estuary (Figure 7). ing. This choice was dictated by the extensive urbani- Results confirm that the estuary is ebb-dominated in the zation of the margins, and by the dykes that protect the 40 km reach upstream from the mouth (Fortunato et al., agricultural lands (e.g., in the islands in the upper estu- 1999) and shows that it switches to flood-dominated ary). further upstream. The reduction of the ebb-dominance The results obtained in the Tagus cannot necessarily be from km 40 upstream is likely associated to the change extended to other estuaries, since the consequences of in morphology, from a wide bay with extensive tidal SLR will depend on the particular hypsometry of each flats to deep and narrow channels (Figure 2). estuary. Also, the consequences of SLR on the hyp- SLR will shift tidal asymmetry towards flood domi- sometry, hence on tidal asymmetry, will depend on nance. For the 1.5 m SLR scenario, tidal asymmetry be- whether or not the margins are protected or allowed to comes negligible in most of the estuary. This result is flood. In general, allowing the margins to flood will consistent with the known processes that control tidal mitigate the reduction of the intertidal areas, thereby asymmetry (Aubrey & Speer, 1985). Flood dominance avoiding the sharp decrease in ebb-dominance. is fostered by small tidal amplitude to depth ratios, while In contrast to the effect of SLR, morphological changes ebb dominance is promoted by extensive tidal flats. in the Tagus estuary induced by the inferred rates of SLR will increase the depth of the estuary, hence re- sedimentation will increase ebb dominance. This in- ducing the tidal amplitude to depth ratio. As a conse- crease is of the same order of magnitude as the reduc- quence, flood dominance should increase. Perhaps more tion associated to a SLR of 50 cm (Figure 7). The com

Figure 7 - Difference between of the ebb and flood durations along the estuary for different MSL scenarios and the reference bathymetry, and for a future bathymetry calculated using the estimated sedimentation rates. Figura 7 - Diferenças entre a duração de vazante e de enchente ao longo do estuário para diferentes cenários de subida do nível médio do mar e batimetria de referência, e para uma batimetria futura calculada usando as taxas de sedimentação estimadas.

73 Guerreiro et al. (2015) parison between the hypsometries of the estuary for the resonance, this phenomenon results from the overlap- present and future bathymetries does not show a sig- ping of two or more waves with the same frequency, nificant change in the intertidal area (Figure 8), consis- creating a partially stationary wave. Tidal resonance tent with the mostly low sedimentation rates used in the tends to occur in estuaries of large dimensions, since extrapolation of the bathymetry (0.3 cm/year in the tidal the resonance period increases with the length of the flats). For instance, the area that is permanently dry estuary. Tidal amplification increases the tidal prism (roughly 2 m above MSL) only increases by about and the water renewal rate in the estuary (i.e., it reduces 2.5 km2. However, there is a larger increase of the area the residence times). 2 of the estuary just above MSL (about 8 km ). The Due to its physical characteristics, the Tagus estuary strengthening of ebb dominance is thus consistent with presents significant resonance. Fortunato et al. (1999) the findings of Fortunato & Oliveira (2005), which showed that the resonant period is about 8 hours, and showed that intertidal areas slightly above MSL maxi- that the semi-diurnal constituents are significantly am- mized ebb dominance. plified in the upper estuary. In summary, while SLR will significantly reduce ebb- The effect of SLR on tidal resonance was studied by dominance in the Tagus estuary, sedimentation in the simulating the propagation of sinusoidal waves with tidal flats will tend to enhance it. The balance may tend amplitudes of 1 m at the ocean boundary and periods either way, depending on the rate of SLR, the changing between 3 and 19 hours, for all defined scenarios. The sedimentation rates, and how the marginal areas are wave amplitudes predicted at several points along the allowed to flood. estuary (Figure 9) confirm that the resonant period is of about 8 hours, and that the waves with periods between 4.2. Resonance 5 and 19 hours are significantly amplified between Ca- Tides can be strongly amplified as they propagate along cilhas and Vila Franca (Figure 9a). Results show that an estuary when the characteristic period of the tide is SLR will enhance the tidal amplification in the Tagus similar to the resonance period of the basin. Denoted estuary for all frequencies considered (Figure 9a).

Figure 8 - Effect of future changes on the hypsometry of the Tagus estuary. The present hypsometric curve (black line) shifts downward due to SLR (blue curve) and upward due to accretion (red curve). The future bathymetry was obtained by extrapolating the present bathymetry using estimated accretion rates. Figura 8 - Efeito de mudanças futuras na hipsometria do estuário. A hipsometria atual (linha a preto) desloca-se para baixo devido à subida do NMM (linha azul) e para cima devido à acreção (linha vermelha). A batimetria futura foi obtida extrapolando a batimetria atual usando taxas de sedimentação estimadas. 74 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015)

Figure 9 - Effect of the resonance: a) along the estuary for the present MSL (2.21 m); b) maximum amplitude for the different SLR scenarios and the different frequencies of the tidal wave. Figura 9 - Efeito da ressonância no estuário: a) ao longo do estuário para o NMM atual (2.21m); b) máxima amplitude para os diferentes cenários de subida do NMM e diferentes frequências de onda de maré.

For the present-day situation, the maximum amplifica- semi-diurnal and quarter-diurnal constituents, this result tion obtained is 65%. For the highest MSL considered, is apparently contradictory with the reduction of tidal i.e., 3.71 m above Chart Datum, the predicted amplify- asymmetry shown above. These results thus suggest cation increases up to 95%, with the amplification vary- that the reduction in tidal asymmetry will occur due to a ing non-linearly with different frequencies. In par- shift in the relative phase lag between the major semi- ticular, for the semi-diurnal constituents the amplifica- diurnal and quarter-diurnal constituents. tion grows asymptotically. For a SLR of 1 m, the 4.3. Extreme water levels maximum amplification will approximately be reached (Figure 9b). Marginal flooding in the Tagus estuary can have ad- verse effects. Some urbanized marginal areas, such as Results indicate that the resonance period of the Tagus Seixal, are low-lying, so that the potential human and estuary does not change significantly with SLR (Fig- material costs of a flood are high. One of the most se- ure 9b). Hence, the amplification of the resonance ef- vere historic episodes described was originated by the fect may be due to smaller frictional losses associated combination of extreme storm surge levels and locally- with the higher mean depth. generated waves during the February 15, 1941, wind- The effects of resonance depend on the period of each storm, causing high human casualties and property constituent. Results indicate that quarter-diurnal con- damages along the estuarine margins (Muir-Wood, stituents will be more amplified than semi-diurnal con- 2011). Recently, the effects of the Xynthia windstorm, stituents by the 1.5 m SLR ((Figure 9b). Since tidal that reached the Portuguese coast on February 27, 2010, asymmetry is mostly due to the interaction between were also observed along the estuary margins, where

75 Guerreiro et al. (2015)

Figure 10 – [caption in the next page]. Figura 10 – [legenda na página seguinte]

76 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):65-80 (2015)

Figure 10 - Maps of extreme water levels for the present MSL (left panels) and considering a 1.5 m SLR (right panels). From top to bottom: high, medium and low probability scenarios, corresponding to return periods of 20, 100 and 1000 years. Levels are in meters, relative to Chart Datum. Figura 10 - Mapas de níveis extremos para o atual NMM (figuras da esquerda) e considerando uma subida do NMM de 1,5m (figuras da direita). De cima para baixo: cenário de probabilidade alta, média e baixa, correspondendo a períodos de retorno de 20, 100 e 1000 anos. Níveis em metros, relativos ao ZH. significant damages in infrastructures occurred. In the relative to the uncertainty in SLR. Simply adding the upper area of the estuary, with extensive agricultural expected SLR to the maximum water levels in the estu- areas, floods may induce salinization and loss of fertile ary appears therefore as an acceptable simplification. land. Raising the MSL implies more frequent floods of 5. Conclusions marine origin. In the particular case of the Tagus estuary, this problem will be exacerbated by the increased The impacts of climate change on estuaries should be tidal amplification due to resonance. anticipated in order to allow for the implementation of adaptation measures, and to inform decision-makers The simulations confirm the importance of SLR on about interventions in the estuary (e.g., construction of extreme water levels (Figure 10). Due to resonance, the infrastructures). This paper contributes to this anticipa- maximum levels grow strongly between Cacilhas and tory procedure in the case of the Tagus estuary. the section of the Trancão River and increase faster than the MSL rise rate. This behavior is illustrated in Fig- The simulations undertaken in this study show that SLR ure 11, which shows, for different return periods and will have significant effects on estuarine hydrodynam- values of the MSL, the difference between the maxi- ics. In the case of the Tagus they will be particularly mum water level and the MSL. Due to increased reso- significant due to the occurrence of resonance, which nance, the maximum water levels at this point grow by amplifies the semi-diurnal constituents of the tide. SLR 4 to 7 cm more for the higher MSL case than for the will trigger two major direct effects: present MSL. However, in spite of this non-linear 1. Tidal asymmetry will decrease significantly. The growth of the maximum water levels with the SLR, due present ebb-dominance will be reduced, and the es- to the resonant effects, these non-linear effects are small tuary may even become flood-dominant. This be 77 Guerreiro et al. (2015)

Figure 11 - Effect of SLR on extreme water levels at point 4 from Figure 3a. Figura 11 - Efeito da subida do NMM nos níveis extremos no ponto 4 da Figura 3a.

havior appears to be mostly due to a significant re- 2009). duction of the intertidal areas (roughly 40% for a 1.5 Using more sophisticated and accurate approaches is SLR) and will be partly compensated by sedimenta- desirable, but entails significant difficulties. First, the tion in the tidal flats. predictability of century-scale estuarine evolutions by 2. The resonance within the estuary will be strength- process-based morphodynamic models remains sketchy, ened, increasing the tidal amplification. As a result, in particular when mixed sediments are involved (e.g., the maximum levels in the estuary will increase Dastgheib, 20112). In the case of the Tagus estuary, the slightly faster than the SLR. lack of extensive bathymetric data also prevents a de- The contrasting effects of SLR and sedimentation on tailed calibration and validation of such a model, just as tidal asymmetry suggest a morphodynamic feedback it provided only coarse estimates of sedimentation rates. that prevents major changes in the estuary. As SLR Also, existing estimates of sediment input into the estu- pushes the estuary towards flood dominance, sedimen- ary are coarse (Vale and Sundby, 1987), and there tation increases and promotes ebb dominance. seems to be a long way before the effect of climate The approach followed herein to assess the interplay change on those inputs can be determined. While the between SLR, sedimentation and hydrodynamics is approach followed herein is a compromise between the simplified. First, the future bathymetries are based on need to assess different processes and the limited accu- simple extrapolation of estimated trends. Ideally, a full racy of the most sophisticated existing methods, it high- morphodynamic model, calibrated and validated with lights the complex interactions between processes and field data, should be applied to better quantify these their contrasting effects. However, further studies using interactions. Secondly, the evolution of the temperature more sophisticated models, supported by more data, are and precipitation associated with climate change is also required to verify the conclusions and quantify how the likely to affect the sediment input into the estuary, thus estuary will evolve under a changing climate. changing the sedimentation rates. Thirdly, tidal asym- Further consequences of SLR, both positive and metry was characterized based on tidal elevations alone. negative, are yet to be investigated in detail. The While this approach is common, a more detailed analy- increase in tidal amplitude will result in larger tidal sis should focus on velocities or even sediment fluxes. prisms. This consequence will reduce the saline Indeed, in very shallow areas, the asymmetry in sedi- stratification (observed in high flow conditions), and ment transport can be the opposite as the one indicated decrease the residence times, thereby improving the by the asymmetry in tidal elevations (e.g., Bertin et al., overall water quality. Salt-wedge intrusion will increase 78 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):733-748 (2015) with potentially negative consequences on water used Church, J.; White, N. (2006) - A 20th century acceleration in global for irrigation and industrial purposes (e.g., cooling of sea-level rise. Geophysical Research Letters, 33: L01602. DOI: 10.1029/2005GL024826 thermal power plants). Higher tidal prisms can also Dastheib, A. (2012) - Long-term process-based morphological reduce the dredging efforts at the inlet channel. modeling of large tidal basins, Ph.D. Dissertation, Balkema, ISBN 978-1-138.00022-3. Unpublished. Acknowledgements Dias, J.M.; Valentim, J.M. (2011) - Numerical modeling of Tagus This work was partially funded by the Foundation for Sci- estuary tidal dynamics. Journal of Coastal Research, ence and Technology, projects AAC-AMB/PTDC/100092/ SI64:1495-1499. 2008, "MorFeed" and PTDC/AAG-MAA/2811/2012, Dodet, G.; Bertin, X.; Bruneau, N.; Fortunato, A.B.; Nahon, A.; “MOLINES”, and research grants SFRH/BPD/87512/2012 Roland, A. (2013) - Wave-current interactions in a wave- (MR) and SFRH/BPD/73089/2010 (AA). This work makes dominated tidal inlet. Journal of Geophysical Research, 118(3): 1587-1605. DOI: 10.1002/jgrc.20146 use of results produced with the support of the Portuguese National Grid Initiative; more information in Dodet, G. (2013) - Morphodynamic modelling of a wave-dominated https://wiki.ncg.ingrid.pt. tidal inlet: the Albufeira lagoon. Ph.D. Dissertation, Université de La Rochelle, 181p., La Rochelle, France. Unpublished. We thank the three anonymous reviewers for their construc- Dodet, G.; Bertin, X.; Taborda, R. (2010) - Wave climate variability tive comments. A preliminary version of this paper was pub- in the North-East Atlantic Ocean over the last six decades. lished as Guerreiro et al., 2013. Ocean Modeling, 31(3-4):120-131. DOI: 10.1016/j.ocemod. 2009.10.010 References Fortunato, A.B.; Baptista, A.M.; Luettich, R.A. (1997) - A three- Aldridge, J.N. (1997) - Hydrodynamic Model Predictions of Tidal dimensional model of tidal currents in the mouth of the Tagus Asymmetry and Observed Sediment Transport Paths in More- Estuary. Continental Shelf Researh, 17(14):1689-1714. DOI: cambe Bay. Estuarine, Coastal and Shelf Science, 44(1):39-56. 10.1016/S0278-4343(97)00047-2 DOI: 10.1006/ecss.1996.0113 Fortunato, A.B.; Oliveira, A.; Baptista, A.M. (1999) - On the effect Andrade, C.; Pires, H.O.; Silva, P.; Taborda, R.; Freitas, M.C. of tidal flats on the hydrodynamics of the Tagus estuary. (2006) – Zonas Costeiras. In: F.D. Santos e P. Miranda (eds.), Oceanologica Acta, 22(1):31-44. DOI: 10.1016/S0399-1784(99) Alterações climáticas em Portugal. Cenários, impactes e medi- 80030-9 das de adaptação. Projecto SIAMII, pp.169- 208, Gradiva, Lis- Fortunato, A.B.; Oliveira, A. (2000) - On the representation of boa. Available on-line at http://siam.fc.ul.pt/siamII_pdf/ SIAMII.pdf bathymetry by unstructured grids. In: L.R. Bentley, J.F. Sykes, Antunes, C.; Taborda, R. (2009) - Sea level at Cascais tide gauge: C.A. Brebbia, and W.G. Gray (eds), Computational Methods in data, analysis and results. Journal of Coastal Research, SI 56: Water Resources XIII, Vol. 2, pp.889-896, Balkema, Rotterdam, Netherlands ISBN: 9058091252. 218–222, Lisbon, Portugal, Available on-line at http://www.cerf- jcr.org/images/stories/218.222_C.Antunes_ICS2009.pdf Fortunato, A. B.; Oliveira, A. (2005) - Influence of intertidal flats on Aubrey, D.G.; Speer, P.E. (1985) - A study of non-linear tidal tidal asymmetry. Journal of Coastal Research, 21(5):1062– propagation in shallow inlet/estuarine systems. Part I: observa- 1067. DOI: 10.2112/03-0089.1 tions. Estuarine, Coastal and Shelf Sc. 21: 185–205. DOI: 10.1016/0272-7714(85)90096-4 Fortunato, A.B.; Rodrigues, M.; Dias, J.M.; Lopes, C.; Oliveira, A. (2013) - Generating inundation maps for a coastal lagoon: a case Azevedo, A.; Oliveira, A.; Fortunato, A.B.; Zhang, J.; Baptista, study in the Ria de Aveiro (Portugal). Ocean Engineering, A.M. (2014) - A cross-scale numerical modeling system for 64:60-71. DOI: 10.1016/j.oceaneng.2013.02.020 mana-gement support of oil spill accidents. Marine Pollution Bulletin, 80(1-2):132-147. DOI: 10.1016/j.marpolbul.2014.01. Freire, P.; Andrade, C. (1999) - Wind-induced sand transport in 028 Tagus estuarine beaches. First results. Aquatic Ecology, 33(3): 225–233. DOI: 10.1023/A:1009911012260 Bertin, X.; Fortunato, A.B.; Oliveira, A. (2009) - A modeling-based analysis of processes driving wave-dominated inlets. Continen- Freire, P. (2003) - Evolução morfo-sedimentar das margens estuari- tal Shelf Research, 29(5-6): 819-834. DOI: 10.1016/j.csr.2008 nas. Estuário do Tejo, Portugal. 380p., LNEC, Lisboa, Portugal. .12.019 ISBN: 9724919595 Bertin, X.; Li, K.; Roland, A.; Zhang, Y.J.; Breilh, J.-F.; Chaumil- Freire, P.; Taborda, R.; Silva, A. M. (2007) - Sedimentary Charac- lon, E. (2014) - A modeling-based analysis of the flooding asso- terization of Tagus Estuarine Beaches (Portugal). A contribution ciated with Xynthia, central bay of Biscay. Coastal Engineering, to the sediment budget assessment. Journal of Soils and Sedi- 94:80-89. DOI: 10.1016/j.coastaleng.2014.08.013 ments, 7 (5), 296-302. DOI: 10.1065/jss2007.08.243 Bilskie, M.; Hagen, S.; Medeiros, S.C.; Passeri, D. (2014) - Dynam- Friedrichs, C.T.; Aubrey, D.G. (1988) - Non-linear tidal distortion in ics of sea level rise and coastal flooding on a changing land- shallow well-mixed estuaries: a synthesis. Estuarine, Coastal scape. Geophysical Research Letters, 41(3):927-934. DOI: and Shelf Science, 27(5):521-545. DOI: 10.1016/0272-7714 10.1002/2013GL058759 (88)90082-0 Guerreiro, M.; Fortunato, A.B.; Freire, P.; Rilo, A.; Taborda, R.; Brown, J.M.; Davies, A.G. (2010) - Flood/ebb tidal asymmetry in a shallow sandy estuary and the impact on net sand transport. Freitas, M.C.; Andrade, C.; Silva, T.; Rodrigues, M. (2013) - Geomorphology, 114(3),431-439. DOI: 10.1016/j.geomorph. Impacte da subida do nível médio do mar na inundação marginal 2009.08.006 do estuário do Tejo. 11º Simpósio de Hidráulica e Recursos Hí- dricos dos Países de Expressão Portuguesa e VII Congresso so- Bruneau, N.; Dodet, G.; Bertin, X.; Fortunato, A.B. (2011) - Devel- bre Planeamento e Gestão das Zonas Costeiras dos Países de opment of a three-dimensional coupled wave-current model for Expressão Portuguesa, 19p., APRH, Portugal. Available on-line at coastal environments, Journal of Coastal Research, SI64:986 - http://www.lnec.pt/organizacao/dha/nec/estudos_id/morfeed/Guerreiro_ 990. et_al_2013.pdf

79 Guerreiro et al. (2015)

Hagedoorn, J.M.; Wolf, D.; Martinec, Z. (2007) - An estimate of Porto de Lisboa (2013) – Publicação Estatística 2013. 17p., Porto global mean sea-level rise inferred from tide-gauge measure- de Lisboa web site, Lisboa, Portugal. Available on-line at ments using glacial-isostatic models consistent with the relative http://www.portodelisboa.pt/portal/page/portal/PORTAL_PORTO_LIS sea-level record. Pure and Applied Geophysics, 164(4):791– BOA/ESTATISTICAS/ACTIVIDADE_PORTUARIA/TAB_PUBLICA 818. DOI: 10.1007/s00024-007-0186-7 COES/publica%E7%E3o2013-sem%20comparativo.pdf Hong, B.; Shen, J. (2012) - Response of estuarine salinity and trans- Pryor, S.C.; Barthelmie R.J. (2010) - Climate change impacts on port processes to potential future sea-level rise in Chesapeake wind energy: A review, Renewable and Sustainable Energy Re- Bay. Estuarine, Coastal and Shelf Science, 104-105(1): 33-45. views, 14(1):430-437. DOI: 10.1016/j.rser.2009.07.028 DOI: 10.1016/j.ecss.2012.03.014 Pugh, D.M.; Vassie, J.M. (1980) - Applications of the joint prob- Horsburgh, K. J.; Wilson, C. (2007) - Tide-surge interaction and its ability method for extreme sea-level computations. ICE Pro- role in the distribution of surge residuals in the North Sea. Jour- ceedings, 69(4):959–975. DOI: 10.1680/iicep.1980.2179 nal of Geophysical Research, 112(C08):003. DOI: 10.1029/ Rahmstorf, S. (2010) - A new view on sea level rise. Nature reports 2006JC004033 Climate Change, 4:44-45. DOI: 10.1038/climate.2010.29 Idier, D.; Dumas, F.; Muller, H., (2012) - Tide-surge interaction in Rego, J.L.; Li, C. (2010) - Nonlinear terms in storm surge predic- the English Channel. Natural Hazards and Earth System Sci- tions: Effect of tide and shelf geometry with case study from ence, 12:3709–3718. DOI: 10.5194/nhess-12-3709-2012 Hurricane Rita. Journal of Geophysical Research, INE (2012) - Census 2011 – Definitive results – Region of Lisbon. 115(C06):020, DOI:10.1029/2009JC005285. 320p., Instituto Nacional de Estatística, Lisboa, Portugal. Rilo, A.; Freire, P.; Mendes, R.N.; Ceia, R.; Catalão, J.; Taborda, IPCC (2007) - Climate Change 2007: The Physical Science Basis. R.; Melo, R.; Andrade, C.; Caçador, M.I.; Freitas, M.C.; Fortu- Contribution of Working Group I to the Fourth Assessment Re- nato, A.B. (2014) - Metodologia para o traçado da Linha de port of the Intergovernmental Panel on Climate Change. [Solo- Máxima Preia-Mar de Águas Vivas Equinociais em ambientes mon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Ave- de transição: aplicação ao estuário do Tejo, Revista Gestão ryt, M. Tignor & H. L. Miller (eds.)], 996p., Cambridge Univer- Costeira Integrada, 14(1):95-107. DOI: 10.5894/rgci450. sity Press, Cambridge, U.K. and New York, NY, USA, Available Rodrigues, M.; Oliveira, A.; Queiroga, H.; Fortunato, A.B.; Zhang, on-line at http://www.ipcc.ch/publications_and_data/ publica- Y.J. (2009) - Three-dimensional modeling of the lower trophic tions_ipcc_fourth_assessment_report_wg1_report_the_physical_science levels in the Ria de Aveiro (Portugal). Ecological Modelling, _basis.htm ISBN: 978-0521795967. 220(9-10):1274-1290. DOI: 10.1016/j.ecolmodel.2009.02.002 IPCC (2013) - Climate Change 2013: The Physical Science Basis. Rodrigues, M.; Oliveira, A.; Guerreiro, M.; Fortunato, A.B.; Contribution of Working Group I to the Fifth Assessment Report Menaia, J.; David, L.M.; Cravo, A. (2011) - Modeling fecal of the Intergovernmental Panel on Climate Change [T. F. contamination in the Aljezur coastal stream (Portugal). Ocean Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Bo- Dynamics, 61(6):841-856. DOI: 10.1007/s10236-011-0392-9 schung, A. Nauels, Y. Xia, V. Bex & P.M. Midgley (eds.)], 1535p., Cambridge University Press, Cambridge, U.K. and New Rodrigues, M.; Oliveira, A.; Queiroga, H.; Brotas, V. (2012) - Sea- sonal and diurnal water quality and ecological dynamics along a York, NY, USA. Available on-line at http://www.ipcc.ch/report/ ar5/wg1/ ISBN: 978-1107661820 salinity gradient (Mira channel, Aveiro lagoon, Portugal). Pro- cedia Environmental Sciences, 13:899-918. DOI: Lanzoni, S.; Seminara, G. (2002) - Long-term evolution and mor- 10.1016/j.proenv.2012.01.084 phodynamic equilibrium of tidal channels. Journal of Geophysi- cal Research, 107(C1):1-13. DOI: 10.1029/2000JC 000468 Roland, A.; Zhang, Y.J.; Wang, H.V.; Meng, Y.; Teng, Y.-C.; Maderich, V.; Brovchenko, I.; Dutour-Sikiric, M.; Zanke, U. Martyr, R.C.; Dietrich, J.C.; Westerink, J.J.; Kerr, P.C.; Dawson, C.; (2012) - A fully coupled 3D wave-current interaction model on Smith, J.; Pourtaheri, H.; Powell, N.; Van Ledden, M.; Tanaka, unstructured grids, Journal of Geophysical Research, 117(C11). S.; Roberts, H.; Westerink, H.; Westerink, L. (2013) - Simulat- DOI: 10.1029/2012JC007952. ing Hurricane Storm Surge in the Lower Mississippi River under Varying Flow Conditions. Journal of Hydraulic Engineer- Silva, T.A.; Freitas, M.C.; Andrade, C.; Taborda, R.; Freire, P.; ing, 139(5):492-501. DOI: 10.1061/(ASCE)HY.1943-7900. Schmidt, S. (2013) - Geormophological response of salt- 0000699. marshes in the Tagus estuary to sea level rise. Journal of Coastal Research (ISSN 0749-0208), SI65:582-587, Lisbon, Muir-Wood, R. (2011) - The 1941 February 15th Windstorm in the Portugal. Iberian Peninsula. Trébol (ISSN 1137-246X), 56:4-13, Fun- dación MAPFRE, Madrid, Spain. Available on-line at Speer, P.E.; Aubrey, D.G. (1985) - A study of non-linear tidal http://www.mapfre.com/mapfrere/docs/html/revistas/trebol/n56/docs/Art propagation in shallow inlet/estuarine systems. Part II: Theory. iculo1en.pdf Estuarine, Coastal and Shelf Science, 21(2):207-224. DOI: Neves, F.S. (2010) - Dynamics and hydrology of the Tagus estuary: 10.1016/0272-7714(85)90097-6 results from in situ observations. 210p,, PhD thesis, Universi- Tawn, J. (1992) - Estimating probabilities of extreme sea-levels. dade de Lisboa, Portugal. Unpublished. Available on-line at Applied Statistics, 41(1):77–93. Article Stable URL: http://www. http://repositorio.ul.pt/handle/10451/2003 jstor.org/stable/2347619 Pethick, J. ( 2001) - Coastal management and sea-level rise. Catena, Vale, C.; Sundby, B. (1987) – Suspended sediment fluctuations in 42(2-4): 307–322. DOI: 10.1016/S0341-8162(00)00143-0 the Tagus estuary on semi-diurnal and fortnightly time scales. Pinto, L.; Fortunato, A.B.; Zhang,, Y.; Oliveira, A.; Sancho, F.E.P. Estuarine, Coastal and Shelf Science, 25(5):495-508. DOI: (2012) - Development and validation of a three-dimensional 10.1016/0272-7714(87)90110-7 morpho-dynamic modelling system for non-cohesive sediments. Vargas, C.I.C.; Oliveira, F.S.B.F.; Oliveira, A.; Charneca, N. (2008) Ocean Modelling, 57-58: 1-14. DOI: 10.1016/j.ocemod. - Análise da vulnerabilidade de uma praia estuarina à inundação: 2012.08.005 aplicação à restinga do Alfeite (estuário do Tejo). Revista de Portela, L.I.; Neves, R., (1994) - Numerical modelling of suspended Gestão Costeira Integrada 8 (1): 25-43. DOI: 10.5894/rgci26 sediment transport in tidal estuaries: a comparison between the Zhang, Y.; Baptista, A.M. (2008) - SELFE: A semi-implicit Tagus (Portugal) and the Scheldt (Belgium-The Netherlands). Eulerian-Lagrangian finite-element model for cross-scale ocean Netherlands Journal of Aquatic Ecology, 28(3-4):329-335. DOI: circulation. Ocean Modeling, 21(3-4):71-96. DOI: 10.1016/ 10.1007/BF02334201 j.ocemod.2007.11.005

80 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):81-90 (2015)

http://www.aprh.pt/rgci/pdf/rgci-521_Maia.pdf | DOI:10.5894/rgci521

Cost-benefit analysis of coastal defenses on the Vagueira and Labrego beaches in North West Portugal*

@, A. Maiaa; C. Bernardes@, b; M. Alvesc

Abstract The Portuguese northwest coast has been suffering for decades from severe erosion and important geomorphological changes, particularly in sandy stretches. This process is related to a number changes which include: a reduction of the supply of sediments, an increase in the mean sea level, an increase of storm frequency and anthropogenic transformations. The Aveiro coastal stretch is morphologically characterized by a sandy barrier-lagoon system often referred to as Ria de Aveiro; the barrier has been submitted to a continuous narrowing and topographic reduction that together potentiate a significant retreat of the coastline, an erosion of the beaches, and over washing. This paper analyses a coastal sector of approximately 3km in length, between the Vagueira and Labrego beaches, located at the south of the jetties that protect the access to the Aveiro lagoon. The study includes the behaviour of the erosion in the last 52 years, the projection of the coastline for the year 2020 and a cost-benefit analysis (ACB) for two scenarios: the present, with the existent coastal defenses, and another one with the absence of defenses. We examined aerial photographs taken in different years in order to determine the different shoreline positions, historical trends and mean shoreline retreat. The analysis of the economic viability of the scenarios was carried out based on the calculation of present value (NPV). The historical (1958- 2010) coastline retreat shows a value of about -3.5m/year, and if the hydrodynamic conditions remain identical to the present, by 2020 a loss of approximately 7ha of territory is expected in semi-natural areas, assuming that the urban areas will continue to be protected. The cost-benefit analysis point out that investment in coastal defenses is not economically viable, particularly in hard defenses. This situation also occurs if nothing is to be done to preserve the natural and built areas. However, in the latter case the economic losses are smaller. Nonetheless, all scenarios and the impact of changes should be based not only on economic factors but also by taking into account the ecological and social importance of the coastal zones. Bearing in mind the present trend, and in order to continue to protect the urban front of Vagueira,, the existing defenses will have to be maintained and reinforced, which will mean more investment in the near future. Keywords: Ria de Aveiro, barrier, coastal defenses, coastal erosion, shoreline.

Resumo Análise de custo benefício de obras de defesa costeira nas praias da Vagueira e do Labrego A costa ocidental portuguesa e, em particular, a zona litoral de Aveiro tem sofrido nas últimas décadas profundas alterações morfológicas, devido à erosão persistente. Este processo está relacionado com o reduzido volume de sedimentos disponível neste segmento costeiro, a ação dos temporais, as ações antropogénicas e a subida do nível médio do mar. Neste contexto, a

@ Corresponding author to whom correspondence should be addressed. a Universidade de Aveiro, Departamento de Ambiente e Ordenamento, 3810-193 Aveiro, Portugal. e-mail: . b Universidade de Aveiro, Departamento de Geociências e CESAM, , 3810-193 Aveiro, Portugal. e-mail: c Universidade de Aveiro, Departamento de Economia, Gestão e Engenharia Industrial e GOVCOPP, 3810-193 Aveiro, Portugal. e-mail:

* Submission: 7 JUN 2014; Peer review: 6 JUL 2014; Revised: 15 OCT 2014; Accepted: 22 OCT 2013; Available on-line: 23 OCT 2014

Maia et al. (2015) barreira arenosa, que define a laguna de Aveiro, tem experimentado um contínuo estreitamento e rebaixamento topográfico que, em conjunto, potenciam um significativo recuo da linha de costa, erosão das praias e galgamentos oceânicos. No presente trabalho é analisado o setor costeiro entre as praias da Vagueira e do Labrego, localizado a sul dos molhes que protegem o canal de acesso à laguna de Aveiro. O estudo incluiu o estudo do comportamento da linha de costa, nos últimos 52 anos, a sua projeção para o ano de 2020, e uma Análise Custo-Benefício (ACB) para duas situações, a existente na atualidade, com obras de defesa costeira e, outra, admitindo a ausência de defesa. Neste contexto, foram examinadas fotografias aéreas de diferentes anos com vista à delimitação das linhas de costa e calculado o recuo médio; a análise da viabilidade económica dos cenários foi realizada a partir do cálculo do Valor Actual Líquido (VAL). O recuo médio da linha de costa foi de -3,5m/ano, com particular expressão nas áreas não protegidas, e se as condições hidrodinâmicas permanecerem idênticas às atuais, até 2020 é esperado um recuo de cerca de 30m a que corresponde uma perda de 7ha de área semi-natural da barreira, admitindo que as áreas edificadas continuarão a ser protegidas. A Análise Custo-Benefício mostra que o investimento em obras de defesa, em particular as de natureza pesada, não é economicamente viável. Situação que também se verifica se nada for feito para preservar o património natural e edificado. No entanto, o cenário sem obras de proteção acarreta menores prejuízos económicos. Convém, contudo, não esquecer que qualquer cenário deve considerar não só fatores económicos mas, também, a importância ecológica e social das zonas afetadas. Tomando em consideração a actual tendência e de modo a continuar a proteger a frente urbana da Vagueira, as obras de defesa existentes terão que ser mantidas e reforçadas o que implicará maior investimento num futuro próximo.

Palavras-chave: Ria de Aveiro, barreira, defesas costeiras, erosão costeira, linha de costa.

1. Introduction northern rivers basins, particularly the Douro River be- The Portuguese coast, and in particular the Aveiro cause of its importance in sedimentary supply, and de- littoral zone, has been suffering for decades from strong fense works carried out along the coast which retain a erosion and coastline retreat. This process, induced by significant volume of sediment. Due to the problems human actions that change the coastal dynamics, is a that affect the Aveiro area, several authors have challenge for planning and coastal zone management dedicated studies to the erosion and coastal defense and for the need to minimize impacts, whether through (Dias et al, 1994; Boto et al, 1997; Taborda et al., 2005; protective measures or loss of territory, natural and Santos, 2008; Coelho et al, 2009;. Maia, 2012, among urban, as well as social and economic implications. others), as well as issues related to the cost-.benefit of Coastal defense work involves high costs, and it is coastal protection investment (Alves et al, 2009; therefore necessary to assess the economic viability of Roebeling et al, 2011; Roebeling et al, 2013). protective measures that impact at both a local and The objective of this paper is to evaluate the behavior regional level. of the shoreline in the last 52 years, in the sector be- The coastal zone of Aveiro is the most problematic of tween the beaches of Vagueira and Labrego, and to the north central region of Portugal, in terms of evaluate the cost-benefit, and the impact of, coastal vulnerability and risk. This situation results from the protection options (including groins, seawalls and also geomorphologic features and the oceanographic con- designated revetments) at local level. text: a wide area with a pronounced topographic unif- ormity, exposed to wind action and to a strong maritime 2. Study area climate. The most important morphological element is The study area is located in the south of the Aveiro la- the presence of a sandy barrier with a variable width. goon entrance (tidal channel), usually designated by Ria For decades, the barrier has presented transgressive de Aveiro (Figure 1). The man-made tidal channel is characteristics, e.g., the rate of accumulation of sedi- sheltered by two large jetties which have undergone ment is less than the carrying capacity, so the coastline successive extensions, in particular the north jetty; these has been suffering a significant migration towards the structures are part of a set of hard coastal defenses ex- land, induced by reduced sediment supply and wash- tending to the south. over processes. Similar situations observed in other This coastal stretch is morphologically characterized by systems are attributed to a rising mean sea level, a a sandy barrier extending NNE-SSW that has been decrease in sediment delivery to the coast, anthropo- subjected to critical coastal erosion during the last few genic influence and, in some cases, changes in the decades. The sector considered, the beaches from storm regime (Pethick 2001; Zhang et al. 2004). Vagueira to Labrego, with a length of approximately The reduced sediment supply in coastal transport that 3km, is backed by a degraded foredune ridge. The fore- affects the north and central Portuguese coast is referred dune shows dune morphologies poorly defined and par- to by several authors. The problem has been gathering tially destroyed by human occupation or erosive proc- over the last two decades, due to the changes in the esses, and was replaced by sand dykes (Figure 1). The

82 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):81-90 (2015)

Figure 1 - Location of the study area. Figura 1 - Localização da área de estudo. dykes are man-made structures resulting from the set- storms (significant wave heights often exceed 7m and tlement of sand from other places or sources, and play a wave periods of 13 s), particularly between October and protective role after the erosion of natural defenses. March, with storms of greater intensity from the south- The sandy barrier also individualizes two morphody- west (Costa, 1994). The mean wave direction is from namic domains: an internal area, protected from direct NW and induces an important littoral drift from north to action of marine processes which integrate the lagoon south. The semidiurnal mesotidal regime in the region and the associated tidal flat, sometimes lower than the presents a maximum tidal range of 3.2 m and a mini- mean sea level, and an external domain, exposed to the mum tidal range of 0.9 m; extreme water level fluctua- ocean processes, that comprises the beach and the fore- tions induced by storm surges usually exceed 40 cm dune or sandy dykes. (Gama et al., 1994). The whole area is characterized by high-energy wave The problems associated with beach erosion, washover conditions. Significant waves are smaller during sum- and coastline retreat, result in a significant reduction in mer (offshore significant wave heights of 1 to 3 m and the mean width of the barrier, in particular, in the wave periods of 11 to 13 s) and larger during winter southern area of the entrance to the Aveiro lagoon be- 83 Maia et al. (2015)

tween Barra and Praia de Mira (Bernardes & Baptista, 2002. The infrastructure, used seasonally, comprises: a 2011). As a result, episodes of over washing are some- water fun park, a bungalow park and some beach facili- times accompanied by the opening of a temporary chan- ties (Figures 1 and 2). The sandy dykes are present at nel between the sea and the lagoon (Mira channel), as the south of the Vagueira and Labrego groins; down- observed in 2001 and 2011, which destabilizes social drift the groin and dyke have been reconstructed and and economic activity. reinforced several times as a result of critical erosion Vagueira is a very popular beach during the summer, during winter storms. The last major reconstruction work particularly as the urban population has increased signifi took place in 2011 following episodes of over washing cantly in recent years. Currently, the urban front is and the opening of a temporary tidal channel; in Febru- about 650 m long and is sheltered by a seawall, a longi- ary 2014, the sandy dyke was reinforced after severe tudinal and adherent structure built in 1978. In 1979 storms. this protection was reinforced and was accomplished by 3. Materials and methods the construction of a groin in the southern area (Figures 1 and 2). These coastal protection measures have been 3.1. Definition and calculation of shoreline retreat subject to frequent maintenance interventions. The analysis of aerial photography, at scales 1/25.000 The Labrego beach, located about 1km south of Va- and 1/28.000, was the basis for shoreline definition cov- gueira, has a groin which was built between 1998 and ering different years (1958, 1970, 1998, 2002 and 2010),

Figure 2. Land cover of study area. Figura 2. Tipo de ocupação do solo na área de estudo. Escarpa de erosão

84 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):81-90 (2015) supported by a geographic information system (ArcGIS). since the houses in the water front are essentially sea- Due to the variety of situations, the coastline reference sonal residences and the factor of choice in the summer was defined from period is the proximity to the beach, as well as the i) dune base when present, socio-economic component and touristic activities that ii) erosion escarpment carved on dune, may be affected in the near future. iii) base of sandy dyke when it replaces the frontal The costs and benefits calculated in the CBA were dune, based on the values associated with the defense works and the natural and urban areas, estimated by Roebeling iv) base of longitudinal structure (fixed coastline). et al. (2011) and updated for the year 2012, using the For the years 2002 and 2010 the beach width, with inflation GDP deflator (World Bank, s/d) (Table 1). potential use in the summer, was determined as the distance from coastline reference to the seaward limit of Table 1 - Values associated with costs and benefits of coastal the subaerial beach in low water spring tides. An update defenses. of the built areas was also performed in Vagueira. Tabela 1 - Valor associado aos custos e benefícios da The rates of shoreline migration were calculated using protecção costeira. the extension digital shoreline analysis system (DSAS). Investment coasts 10.137 €/m a This data was the basis for the projection of the coast- Seawall Maintenance costs 2.282 €/m line for the years 2015 and 2020, according to that pro- (yr 3,6 …) posed by Ferreira et al. (2006). The projection took into Investment coasts 12.671 €/m account only the evolution trend of the shoreline and Groins a not any other factor, climatic, morphological change or Maintenance costs 2.534 €/m (yr 3,6 …) anthropogenic action that modified the sediment dynamics in the study area. The shoreline position was Investment coasts 867 €/m Sand dyke b calculated using the formula Sa=S0+SERxa, where a is Maintenance costs 867 €/m the time horizon to be analyzed; Sa is the shoreline (yr 3,6 …) position after a year’s; S0 is the position of the baseline Natural 28.781 €/m considered (in this study is 2010); SER corresponds to Protected area (beach and dune) c the erosion rate. and lost area d Urban 623.048 €/m 3.2. Cost-benefit analysis a Coastal defense intervention investment and maintenance costs in The cost-benefit analysis (CBA) has been realized in central Portugal (Roebellinget al., 2011) b Sand dyke investment and maintenance costs in southern Vagueira order to evaluate the costs and the benefits in the pres- beach ence (with protection) and absence (no protection) of c Coastal ecosystem values (Roebellinget al., 2011) coastal defense works. d From Almeida (2011) In the case of the presence of coastal defense, were calculated: i) the costs of construction and maintenance; The economic viability of the scenarios was carried out ii) the costs associated if the defenses were replaced, from the equation proposed by Roebeling et al. (2011), called opportunity cost, in particular the longitudinal which allowed us to calculate the Net Present Value revetment of Vagueira; iii) potential loss due to the (NPV). presence of structures such as groins that accentuate erosion downdrift of the same area. The benefits arising from the protection calculated were: the protected area (not lost) due to the presence of where t = 0, n (n is the number of years considered: 0 is the works and the increase of the beach width near the the first year, in this case 1978, and n = 2020); r = time groins. discount rate in the year t; Bt = total benefits expected In the case of the absence of defense works, the costs of in the year t; Ct = total costs, expected in year t. area potentially lost due to coastal erosion were esti- The time discount rate allowed converting the costs and mated; the economic estimative, include natural areas benefits to present values that will be compared. The of beach, dunes, inland adjacent zones, and urban areas decreasing discount rate was used because it permits a based on the property value of homes. The benefit was better valorization of the costs and benefits over major calculated taking into account that any area lost due to time periods (HM Treasury, 2003). the presence of hard coastal protection, including areas The benefits of coastal defense works are not immedi- down-drift the groins. ate but only evident in the medium and long term; for In the study, the devaluation of the urban area situated that reason a higher value is given to the benefits in the in areas with risk of coastal erosion are not considered, CBA when compared with a constant discount rate. In

85 Maia et al. (2015)

Figure 3 - A. Evolution of shoreline between 1958 and 2010. B. Beach area in 2002 and 2010 in low tide level (Maia, 2012). Figura 3 - A. Evolução da linha de costa entre 1958 a 2010. B. Área útil de praia em 2002 e 2010 em situação de baixa-mar, 2002 e 2010 (Maia, 2012). this case, for the period of 30 years, from 1970 to 2000, ban occupation; the density of built area has increased a discount rate of 3.5% was considered, and for the foll- significantly over the past few decades (Figure 4). owing period, from 2001 to 2020, a rate of about 3%. In 1998 the coastline, influenced by the adherent structures, showed a "fixed" position along the longitudinal 4. Results and discussion seawall and a rotation tendency immediately downdrift 4.1. Evolution of the shoreline of the groins, accompanied by an important coastline The shoreline evolution for the period between 1958 retreat. From 1998 to 2002 the shoreline remained rela- and 2010 shows a general retreat, which was accentu- tively stable, except to the south of the Labrego groin, ated after the coastal protection works, which were with an increasing erosion area (Figure 3.A). started in 1978 (Figure 3.A). Between 2002 and 2010, the position of the shoreline The aerial photography of 1958 did not reveal situations remained constant between the Vagueira and Labrego of erosion in the Vagueira zone: the beaches were wide beaches, recording some retreat to the south of the and the foredune well developed. However, in 1970 and Labrego groin (Figure 3.A). without defense structures, some erosion in the fore- The subaerial beach area between the mean low-water dune system was recognised, which may return the ef- level and the base of the seawall or dune/sand dyke, has fects of the interventions in the north region, specifi- been being reduced since 2002 (Figure 3.B). Until 2010 cally in Costa Nova and nearby the main entrance to the there was a mean decrease of about 8,8 ha of total sur- Aveiro lagoon (Figures 1 and 3). face area (approximately 1 ha/year, from 2002 to 2010), Constructions of coastal defense structures in Vagueira which was felt along the beach, including in the area (APA, 2012) began at the end of the 70s and were ac- 200 m immediately updrift of the groins. This trend had companied almost simultaneously by an increase of ur already been observed in 2010 and associated with the

86 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):81-90 (2015)

Figure 4 - Evolution of Vagueira urban area, between 1958 and 2010 (Maia, 2012). Figura 4 - Evolução da área urbana da Vagueira entre 1958 e 2010 (Maia, 2012). decrease of the average width of the sand barrier (beach destruction of the frontal dunes and their reinforcement and dune system) with a value of about 250 m in the or total replacement by sandy dykes. Vagueira area (Bernardes & Baptista, 2011). The shoreline projection for the years 2015 and 2020 The coastline between 1958 and 2010 shows an average was performed by taking into account the present situa- retreat of -3,5 m/year, which reflects the trend of the tion and a hypothetical scenario without protection of entire littoral sector. Before the construction of coastal either fixed (adherent defenses) or dynamic (sandy defenses a mean value of -3,2 m/year had been re- dykes and beach nourishment). The results show, as ex- corded, and after building this value reached 2,5 m/year pected, that the coastline retreat in the Vagueira water- (Maia, 2012). The decrease in the shoreline retreat rates front and updrift of the Labrego groin would be greater that was not verified to the north or south of the without defense structures, while in areas protected by Vagueira stretch was due to the coastline fixation in the groins the retreat would be less significant in the ab- urban front, balancing the retreat recorded before, and sence of the protective structures. to the accretion which became to occur updrift of the For the period 2010-2015 it is estimated that the present Labrego groin (Figure 5). Although the retreat has coastal defenses avoid an erosion of 13,5 ha of total diminished in relative terms, there was an increase of area (9 ha of dunes and sandy dyke, 4,5 ha of urban erosion downdrift of the groins that caused the partial area and 6 ha of area not protected). By 2020, it is

87 Maia et al. (2015)

Figure 5 - Erosion and accretion rates of shoreline (in m/year) in Vagueira and Labrego. A- Before the implementation of the coastal defenses. B- After the implementation of the coastal defenses (Maia, 2012). Figura 5 - Taxas de erosão e acreção da linha de costa (m/ano) na Vagueira e na praia do Labrego. A- Antes da construção das obras. B- Após a construção das obras (Maia, 2012).

Figure 6 - Projection of shoreline position for the years 2015 (A) and 2020 (B) (Maia, 2012). Figura 6 - Projecção da linha de costa. A- 2015; B- 2020 (Maia, 2012). 88 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):81-90 (2015) estimated that 7 ha will be eroded and 16 ha protected presence of structures, largely penalize the benefits; (10,4 ha of semi-natural area and 5,6 ha of urban area) amounts relating to the loss of territory without protec- (Figure 6). tion are considerably lower, which raises some questions regarding the value assigned to the intrinsic value 4.2. Cost-benefit analysis (CBA) of ecosystem components. In Vagueira, the results of ACB show for the situations with and without protection a negative NPV in the 5. Conclusions periods considered, indicating the unviability of the The coastal defenses on the Vagueira and Labrego defense structures, taking into account only economic beaches have contributed to the mitigation of negative reasons (Table 2). Coastal protection involves high impacts caused by coastal erosion. In Vagueira, the construction costs and also ongoing maintenance costs; shoreline has remained "fixed" since the early 80s, due the total costs outweigh the benefits provided to natural to the construction of the seawall; the beach will still and urban areas (Table 1). For a hypothetical scenario benefit from the presence of the groin in the southern without the defense works the VAL, although negative, zone, however, in the urban front the subaerial beach is much lower even in the near future. area has been decreasing, and even nonexistent during For the Labrego beach, the results of the CBA are high tide conditions. Despite the reinforcement of the similar to the situation in Vagueira. In both scenarios, seawall, under storm conditions the structure suffers with and without protection, the NPV is always nega- frequent overtopping and both the areas to north and tive, though lower costs are observed in a scenario of no south are seriously affected by breaking wave and wave protection (Table 3). set up, necessitating emergency repair work (Figure 2). In the period 1998-2002, the benefits created by the The projection of the shoreline based on the most simp- groins were negative because this period coincided with listic formulation, that excluded the increase of mean the construction of the structure and its effects still have sea level, the impact of storms and the storm surge, the not yet been felt. For a longer period of analysis, the morphodynamic characteristics of the coastal stretch, or costs are greater than the benefits. In the period 1998- anthropogenic intervention, shows that the short term 2010, the costs associated with the eroded area have coastal retreat, in the absence of defense works, would increased which has caused a drastic increase in NPV. be significant and can induce considerable damage. The The projections for 2015 and 2020, both for Vagueira position of the projected coastline for 2015 has already and Labrego as well as the costs associated with the been reached and slightly exceeded in February 2014;

Table 2 - Cost-benefit analysis of the coastal defense in Vagueira. Tabela 2 - Análise Custo-Benefício (ACB) das obras da Vagueira. VAGUEIRA 1970-1998 1970-2002 1970-2010 1970-2015 1970-2020 Total benefits (€) 46.749 41.773 29.509 36.577 43.357 Without protection Total costs (€) 1.411.039 1.728.628 2.444.657 3.091.508 3.697.639 Net Present Value (€) -1.364.291 -1.686.855 -2.415.148 -3.054.932 -3.654.282 Total benefits (€) 1.308.717 1.605.988 2.368.138 2.903.120 3.482.609 With protection Total costs (€) 14.561.992 16.155.154 18.086.819 19.388.210 20.271.067 Net Present Value (€) -13.253.274 -14.549.166 -15.718.681 -16.485.089 -16.788.458

Table 3 - Cost-benefit analysis of the coastal defense in Labrego beach. Tabela 3 - Análise Custo-Benefício (ACB) das obras da praia do Labrego. LABREGO 1998-2002 1998-2010 1998-2015 1998-2020 Total benefits (€) 136.086 136.086 187.215 242.626 Without Total costs (€) 82.057 196.547 289.304 450.447 protection Net Present Value (€) 54.029 -60.461 -102.089 -207.821 Total benefits (€) 0 4.110 53.511 107.468 With protection Total costs (€) 3.050.179 6.851.946 8.526.151 10.360.651 Net Present Value (€) -3.050.179 -6.847.837 -8.472.640 -10.253.183

89 Maia et al. (2015) this situation took place as a consequence of the highly (eds.), Colectânea de Ideias sobre a Zona Costeira de Portugal, energetic and persistent storms that affected the Portu- pp.449-467, Associação Eurocoast – Portugal, Porto, Portugal. guese coast last winter. In the study area, it was neces- Coelho, C.; Silva, R.; Veloso-Gomes, F.; Taveira-Pinto, F. (2009) - Potential effects of climate change on northwest Portuguese sary to proceed with the partial nourishment of the coastal zones. ICES Journal of Marine Science 66(7):1497- beach in the area immediately north of the Vagueira 1507. DOI: 10.1093/icesjms/fsp132 seawall and the reinforcement of sandy dykes (Fig. 2). Emergency action of this type has been relatively Costa, C.L. (1994) - Final report of sub-project. A “wind wave effective in maintaining shoreline position and prevent- climatology of the Portuguese coast”. 80p., Instituto Hidro- gráfico/LNEC, Report 6/94-A, Lisboa, Portugal. Não publicado. ing coastline erosion trends. The CBA shows that Dias, J.A.; Ferreira, Ó.; Pereira, A.R. (1994) - Estudo sintético de coastal defenses on the Vagueira and Labrego beaches diagnóstico da geomorfologia e da dinâmica sedimentar dos are not economically viable. However, this situation troços costeiros entre Espinho e Nazaré. 261p., ESAMIN- also occurs if nothing is done to preserve the natural Estudos de Ambiente e Informática / ICN - Instituto de and built heritage. Conservação da Natureza, Lisboa, Portugal. Available on-line at http://w3.ualg.pt/~jdias/JAD/eb_EspinhoNazare.html The sandy dykes have contributed to the maintenance of the coastline position mainly downdrift of the groins, Ferreira, Ó.; Garcia, T.; Matias, A.; Taborda, R.; Dias, J.A. (2006) - although each emergency action in order to repair the An integrated method for the determination of set-back lines for coastal erosion hazards on sandy shores. Continental Shelf dykes is further inland. Nevertheless, the retention ca- Research, 26(9):397–407. DOI: 10.1016/j.csr.2005.12.016 pacity of the Vagueira groin has vanished due to a Gama, C.; Dias, J.A.; Ferreira, Ó.; Taborda, R. (1994) - Analysis of significant decrease in the supply of sediments in the storm surge in Portugal, between June 1986 and May 1988. littoral drift; a similar situation is present in Labrego. Proceedings of Littoral´94, pp.381-387, EUROCOAST. Lisboa, Portugal. Availabe on-line at http://w3.ualg.pt/~jdias/JAD/papers Taking the present critical situation into account, the /CI/94_Lit94_381_CG.pdf option of defending beaches is both relevant and impor- HM Treasury (2011) - Green Book: Appraisal and Evaluation in tant in order to refocus on coastal defenses and, given Central Government. 118p., HM Treasury (Her Majesty's the current conditions, the progressive additions to the Treasury) / TSO (The Stationery Office), London, U.K.. Available on-line at: https://www.gov.uk/government/uploads/system length of the groins and the reinforcement of hard defen- /uploads/attachment_data/file/220541/green_book_complete.pdf ses. Soft interventions, like sand dykes and beach nouri- Maia, A. (2012) - Evolução Litoral entre a Vagueira e a P. de Mira: shment, have proved effective from the point of view of análise geoeconómica. 94p., Dissertação de Mestrado em protection and maintenance costs, although the challen- Ciências do Mar e das Zonas Costeiras, Universidade de Aveiro, ge in obtaining deposits of sand can be a limiting factor. Aveiro, Portugal. Available on-line at http://hdl.handle.net/10773 /9749. The investment and maintenance costs estimates should Pethick, J. (2001) - Coastal management and sea-level rise. Catena, not only take into account the single market values, but 42(2-4):307–322. DOI: 10.1016/S0341-8162(00)00143-0 also “non-market” coastal ecosystem values and should Roebeling, P.; Coelho, C.; Reis, E. (2011) - Coastal erosion and be included in the impact assessment and cost-benefit coastal defense interventions: a cost-benefit analysis. Journal of analysis (Alves et al., 2009; Roebeling et al., 2011). It Coastal Research (ISSN: 0749-0208), SI64:1415-1419. should also be noted that any scenario should be based Szczecin, Poland Availabe on-line at http://www.ics2011.pl/ on the social importance of affected areas and costs artic/SP64_1415-1419_P.C.Roebeling.pdf involved, whichever option is considered in the future. Roebeling, P.; Costa, L., Magalhães-Filho, L.; Tekken, V. (2013) - Ecosystem service value losses from coastal erosion in Europe: historical trends and future projections. Journal Coastal Con- References servation, 17(3):389–395. DOI: 10.1007/s11852-013-0235-6 Alves, F.; Roebeling, P.; Pinto, P.; Batista, P. (2009) - Valuing Santos, L. (2008). Comunicação de riscos associados às zonas ecosystem service losses from coastal erosion using a benefits costeiras – o caso da Praia da Vagueira. 298p., Dissertação de transfer approach: a case study for the Central Portuguese coast. mestrado em Planeamento do Território–Riscos Naturais e Journal of Coastal Research (ISSN: 0749-0258), SI56:1169- Tecnológicos. Universidade de Aveiro, Aveiro, Portugal. 1173, Lisboa, Portugal. Availabe on-line at http://egeo.fcsh.unl.pt Available on-line at http://hdl.handle.net/10773/577 /ics2009/_docs/ICS2009_Volume_II/1169.1173_F.Alves_ICS2009.pdf Taborda, R.; Magalhães, F.; Ângelo, C. (2005) - Evaluation of coastal defence strategies in Portugal. In: B. Dean & C. APA (2012) – Plano de acção de protecção e valorização do litoral Zimmermann (eds.), Environment Friendly Coastal Protection 2012-2015. 86p., Agência Portuguesa do Ambiente, Lisboa, Structures, pp.255-265, Kluwer, Netherlands. ISBN: 978- Portugal. Available on-line at http://www.apambiente.pt/_zdata/ 1402033018. destaques/2012/papvl_2012-2015-junho.pdf The World Bank (s/d) - Inflation, GDP deflator (anual%). Web Bernardes, C.; Baptista, P. (2011) - Evolução recente da barreira page, The World Bank , Washington, DC, U.S.A. Available on- arenosa de Aveiro. Actas das Jornadas da Ria de Aveiro, pp.28- line at http://data.worldbank.org/indicator/NY.GDP.DEFL.KD.ZG 33, Universidade de Aveiro, CESAM – Centro de Estudos do Ambiente e do Mar, Aveiro, Portugal. ISBN: 978-9727893379. Zhang, K.; Douglas, B.C.; Leatherman, S.P. (2004) - Global war- Available on-line at http://la.cesam.ua.pt/documentos/LivroActas ming and coastal erosion. Global Warming and Coastal Ero- JornadasRiaAveiro2011_Cores.pdf sion, 64(1-2):41–58. DOI: 10.1023/B:CLIM.0000024690 .32682.48

Boto, A.; Bernardes, C.; Dias, J.A. (1997) - Erosão litoral e recuo da linha de costa entre a Costa Nova e a praia do Areão, Portugal. In: G. Soares de Carvalho, F. Veloso Gomes & F. Taveira Pinto 90 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

http://www.aprh.pt/rgci/pdf/rgci-555_Verocai.pdf | DOI:10.5894/rgci555

Addressing climate extremes in Coastal Management: The case of the Uruguayan coast of the Rio de la Plata System* @, José E. Verocai @, a, b; Monica Gómez-Erachea, c; Gustavo J. Nagya; Mario Bidegainc, d

Abstract This paper deals with current storm surges (extremes) in the Uruguayan coast of the Rio de la Plata system (Argen- tina-Uruguay), a large (38,000 km2) microtidal (<0.5 m) river estuary. Extremes are associated with two synoptic situations able to increase water level above mean sea level (MSL) at Montevideo (≥102 cm): the passage of cold fronts and extra-tropical cyclones over Atlantic coast. The assessment of coastal hazards and the reduction of the associated risks can be embedded within Coastal Management (ICZM). Infrastructure frequently does not comply with the legal 250 m setback. Thus, possible impacts due to inundation and flooding related to sea-level rise (SLR) and extremes respectively are high. Also, rapid-onset high river inflow often associated with El Niño events, are relevant due to the positive water level anomaly (> 10 cm) related to them. The existence of differences in the monthly distribution of both MSL and extremes occurrence suggest that both wind climate and river inflow (QRP) explain them. The most severe extremes (≥ 350 cm) at Montevideo have occurred before 1935. The frequency of events ≥200, 250 and 280 cm has slightly increased over 2003-2012, which is not attributed neither to SLR nor to increased inflow, but to wind regime changes. Two extreme events (March 1998 and August 2005) are presented which serve to enhance our understanding of the causality and timing of extremes and as analogues to developing scenarios under a changing climate. Some impacts are as follows: coastal stormwater and sewage systems become useless affecting beach quality, artisanal fisheries income is reduced, and sandy beaches and bars are eroded. Among the several obstacles to achieving a comprehensive ICZM in regards to extreme events is the lack of an effective monitoring network. This paper aims to increase awareness, research and application of the knowledge of extreme events on ICZM and risk-management in the face of SLR and flooding. Keywords: water level, storm surges, river flow, climate adaptation, ENSO, Montevideo.

Resumo Sobre extremos climáticos em gestão costeira: o caso do sistema de Rio de la Plata na costa do Uruguai. Neste artigo analisam-se os extremos de maré meteorológica (elevação do nível do mar de índole meteorológica) na costa u- ruguaia do sistema de Rio de la Plata (Argentina-Uruguai), um grande (38.000 km2) corpo estuarino com características de micro-maré (<0,5 m). Os extremos estão associados a duas situações sinóticas capazes de aumentar o nível de água acima do nível médio do mar (MSL) em Montevidéu (≥102 cm): a passagem de frentes frias e de ciclones extra-tropicais sobre a costa

@ Corresponding author to whom correspondence should be addressed. a Universidad de la República, Grupo de Cambio Ambiental y Gestión Costero Marina, Oceanografía y Ecología Marina, Instituto de Ecología y Ciencias Ambientales (IECA), Montevideo, Uruguay. e-mails: Verocai ; Nagy ; Gómez-Erache b Armada Nacional, Servicio de Oceanografía, Hidrografía y Meteorología (SOHMA), Departamento de Oceanografía, Montevideo, Uruguay c GEF Project “Implementing Pilot Adaptation Measures in Coastal Areas of Uruguay” d Instituto Uruguayo de Meteorología (INUMET), Montevideo, Uruguay. e-mail: Bidegain:

* Submission: 16 SEP 2014; Peer review: 16 OCT 2014; Revised: 17 NOV 2014; Accepted: 23 NOV 2014; Available on-line: 25 NOV 2014

Verocai et al. (2015) atlântica. A avaliação de riscos costeiros e da redução dos riscos associados a tais eventos pode ser incorporado dentro de Gerenciamento Costeiro (GIZC). Com frequência, as infra-estruturas não cumprem os 250 m legais de afastamento da linha de costa. Assim, os possíveis impactos devidos a inundações relacionadas com a subida do nível do mar (SLR) e com extremos da maré meteorológica têm elevada magnitude. Além disso, há que ter em consideração as cheias fluviais relacionadas com eventos de El Niño, pois que provocam também anomalias positivas do nível de água (> 10 cm). A existência de diferenças na distribuição mensal de ambos (MSL e ocorrência de extremos de maré meteorológica) permite deduzir que tanto o vento como os fluxos fluviais (QRP) permitem explicá-las. Os maiores extremos (≥ 350 centímetros) registrados em Montevideo ocorreram antes de 1935. A freqüência de eventos ≥200, 250 e 280 centímetros aumentou ligeiramente no período 2003-2012, o que não é atribuído ao SLR nem ao aumento de fluxos fluviais, mas sim a mudanças no regime de ventos. São apresentados dois eventos extremos (Março de 1998 e Agosto de 2005) que servem para melhorar a nossa compreensão da causalidade e da tempo- rização dos eventos extremos, constituindo bons análogos para o desenvolvimento de cenários num clima em mudança. Al- guns impactos são os seguintes: sistemas costeiros de águas pluviais e de esgotos tornam-se inúteis, o que afeta a qualidade das praias, redução da pesca artesanal, e erosão das praias e bares. Entre os vários obstáculos para alcançar uma GIZC a- brangente no que diz respeito a eventos extremos é a ausência de uma rede de monitoramento eficaz. Este trabalho tem como objetivo aumentar a conscientização, pesquisa e aplicação do conhecimento de eventos extremos na GIZC e da gestão de riscos perante as inundações e o aumento do nível médio do mar (SLR). Palavras-chave: nível da água, mare meteorological, fluxo fluvial, adaptação ao clima, ENSO, Montevideo.

1. Introduction Uruguay). The impacts of coastal inundation related to storm surges are greater than those related to SLR. This article continues a series of five papers written by As SLR accelerates, it will become increasingly neces- the Team of Environmental Change and Coastal-Marine sary and useful to distinguish coastal “flooding” from Management of the School of Sciences (FC-UdelaR), “inundation” (Flick et al., 2012). These authors propose Uruguay for the GEF project “Implementing Pilot Ad- that the term flooding” be used when dry areas become aptation Measures to Climate Change in Coastal Areas wet temporarily and that “inundation” be used to denote of Uruguay”, from now on “the Project” (PRODOC, the process of a dry area being permanently drowned or 2008). The four other works (Nagy et al. 2014a, 2014b, submerged. Thus, the former is related to episodic ex- 2014c, in press), discuss the methodological evolution tremes and the latter to SLR. to cope with observed and current variability, to adapt In most cases, the infrastructure in Uruguay does not to future climate change, and the development of alter- comply with the 250 m setback established by the native future scenarios to better understand and manage Water Code. Thus, current or possible territorial im- the coastal areas of Uruguay focused on climate threats pacts due to flooding associated with SLR and/or wind such as episodic extreme river flows and storm surges, storm surges are high. In view of this, sections have and gradual sea level rise (SLR). been identified for coastal management, in an attempt to The coast line can adopt a stable profile or shape when set up a coastal management and planning unit in each the processes for the input and removal of sediments are section, in the light of current and applicable legislation balanced. However, external factors such as storms of- (Medina, 2009; Medina & Gómez-Erache, 2014a, ten induce morphodynamics changes which disrupt the 2014b, 2014c). state of balance. Climate change and the increase of the This article emphasizes on the occurrence of extreme mean sea-level (MSL) affect the transfer of sediment in events, mainly storm surges, in the Uruguayan coast of complex ways; non-linear and abrupt changes can occur the Río de la Plata (RdlP) river estuary from 1983- when certain thresholds are exceeded. If the increase of 2012, regardless of the associated flooding and inunda- the MSL is gradual and slow, the balance can be main- tion. This period was chosen to analyze the current cli- tained, even with morphological changes, but accelera- mate baseline, including the reported changes in hydro- tion in the rate of increase can make it difficult to main- climatic variables, winds, and SLR observed in the tain, particularly where sedimentary input is limited, RdlP basin, river estuary, and the coastal areas of Uru- such as in coastal lowlands with a tendency to flood guay (Escobar et al. 2004; Barros et al. 2005; Bidegain (Solomon et al., 2007). et al. 2005; Nagy et al. 2008a,b, in press). According to ECLAC (2011) the 50-years return inundation level is greater in Chile, Argentina and Uruguay. Extreme Events and Integrated Coastal Zone Man- The trends of coastal inundation levels in Latin agement America have increased to ≤ 0.5 cm over the last six The assessment of coastal hazards and the mitigation of decades due to SLR, waves and storm surges. The re- the risks in respect of those hazards can be embedded gion where this increase - excluding hurricanes -was the within the Integrated Coastal Zone Management highest (> 1cm/year) is the Rio de la Plata (Argentina- (ICZM) (IOC, 2009).

92 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

Of all the impacts from climate change, the projected 2. The Rio de la Plata River Estuary and the Vul- rise in mean sea level is one of the most significant nerability of the Uruguayan Coast concerns for ICZM. In addition to projected higher fre- The complex geophysical environment of the RdlP ba- quency of storm surge and inundation levels, SLR will sin and river estuary is under stress from existing pres- also result in shoreline recession of unconsolidated sures such as changing hydro-climatic and wind re- sandy shorelines (DECCW, 2010). gimes, sea-level rise, extreme events, growing popula- ICZM is an acknowledged tool to deal with current and tion and associated increases in development over the long-term coastal challenges, including climate change last few decades (Barros et al. 2005; Nagy et al. 2008a, and its impacts (for instance SLR, changes in storm fre- 2008b, in press). The coastal areas in Argentina and quency, strength and patterns, and increased coastal Uruguay are vulnerable to these trends over the last few erosion and flooding; Ballinger & Rhisiart, 2011). decades, especially to storm surges (Escobar et al. Notwithstanding, the increasing experience in coastal 2004; Barros et al. 2005; Bischoff, 2005; Magrin et al. management education, research, policies and actions in 2007; Nagy et al. 2007; 2008a, 2008b, 2013; ECLAC, Uruguay, the country still has to increase current and 2011). future climate and meteorological knowledge and The RdlP river estuary shared by Argentina and Uru- threats. Recently, this approach has been conducted by guay (Figure 1) is 290 km long, 40 to 220 km wide, has “The Project” which has cooperated and developed an average depth ≤10 m, and 38,000 km2, where it common goals related to climate adaptation with Eco- mixes the continental freshwater (see Figure 1 below) Plata Program (Gómez-Erache et al. 2010; Echevarría from the RdlP basin (Paraná and Uruguay rivers, which et al. 2013; Nagy et al. 2014a, c, in press). supply 75 and 25% of the total river inflow 3 The baseline of storm surges (1983-2012) is presented (25,000 m /s) respectively, and the Atlantic sea waters focusing on their occurrence and on examples of two (Nagy et al., 1997, 2008a; Lappo et al., 2005). Al- events. A plausible increase in the frequency and/or in- though the Uruguayan riverside of the RdlP is less ex- tensity of extremes under projected SLR scenarios is posed to winds than the Argentinean one, the shallow expected to be the main threat to coastal stability until it inner fluvial and the middle estuarine front regions suf- reaches at least 20-30 cm plus, by around 2040-60 fer the effects of southern and southeastern winds (Nagy et al., 2006). Thus, a better knowledge of the oc- which drag water into the system increasing water level currence, causality, timing, and impacts of storm surges by 100 to 300 cm (Balay, 1961; Escobar et al. 2004; is needed to planning current and near-future coastal Barros et al. 2005; Bidegain et al. 2005; Bischoff, management options. 2005). Montevideo is located in the estuarine front. Storm surge is defined as an increase in coastal water The Uruguayan coast has a mixed micro-tidal regime level caused by the effects of storms. Storm surge con- (< 0.5 m), semidiurnal with diurnal inequality. The tide sists of two components: an increase in water level wave comes from the Atlantic Ocean, being deformed caused by a reduction in barometric pressure (baromet- within the estuary due to the shape, banks, channels, ric setup) and an increase in water level caused by the depth, and Coriolis deflection towards Montevideo, action of wind blowing over the sea surface (wind where a strong decrease of the cross-section occurs. setup) (DECCW, 2010). The most common impacts due Thus, an increase of isoamplitudes occurs towards the to storm surges are damages to the: inner river (upstream the estuarine front) which facili- tates the occurrence of “storm surges”, that is to say the • Population, i.e., deaths, injuries, evacuated, tempo- positive anomaly between the astronomic tide and the ral loss of services such as electricity and drinking observed level, which is attributed to residual effects of water, sewage; winds, waves, atmospheric pressure, and the flood of • Real estate and infrastructures; freshwater (Balay, 1961; Nagy et al., 1997; Verocai et • Coastal morphology, i.e., sandy beaches, barriers al., 2014). and dunes erosion; There are two typical weather development situations • Ecosystems and biodiversity. able to produce strong positive water level anomalies (+ 100 to 300 cm): The goals of this article are to: i) Identify water level i) the strengthening of an extra tropical cyclone over extremes defined as those exceeding 200 cm at Monte- video; ii) Explain the relationship between hydrocli- Atlantic coast and matic variability (river inflow to the RdlP) and surface ii) the passage of cold fronts coming from the South. wind behavior with water level, and iii) Present two ex- The development of cyclones near the Argentinean- amples of observed storm surges, focused on the cli- Uruguayan coast is usually associated with strong matic system over the Rio de la Plata region. southeastern winds (35 – 50 km/h) called “Sudestadas”

93 Verocai et al. (2014)

Figure 1 - Above: Rio de la Plata basin and river estuary, Southeastern South America (accordingly to Nagy et al., 2014a). Below: Location of the estuarine front (March 30, 2011) under typical river inflow conditions. Figura 1 – Em cima: bacia e estuário do Rio de la Plata, no Sudeste da América do Sul (segundo Nagy et al., 2014a). Em baixo: localização da frente estuarina (30 de março de 2011) em condições típicas de fluxo fluvial. in the RdlP region. These events last around 48 hours. et al. 1992; Norbis, 1995; Nagy et al., 2008b, 2014b), The arrival of cold fronts originates southwest winds and the coastal populations, real estate, infrastructure, called “Pamperos” which duration is usually less than and biodiversity (Volonte & Nicholls, 1995; Saizar, the one of “Sudestadas”. These winds drag waters to- 1997; Nagy et al., 2005, 2007; ECLAC, 2011). wards the inner, narrower, and shallower fluvial and An estimated 70 per cent of Uruguayans live in coastal frontal zones (4-10 m depth) piling-up water at both areas, 34 percent of which is urbanized, including the riversides. The maximum water levels vary depending metropolitan area of Montevideo. Over two thirds of on the intensity and persistence of the wind, and the si- economic activities and income generated in Uruguay multaneous occurrence of the peak of the storm surge are directly or indirectly related to the coast. Also, and the astronomic tide. Several papers explain how globally relevant biodiversity sites are located in coastal these winds affect water circulation (Simionatto et al. areas. Consequently, coastal processes coexist along- 2005), frontal location and displacement (Framiñan & side a variety of activities that compete for space and Brown, 1996; Simionatto et al. 2001; Nagy et al. resources, the subsequent result being coastal degrada- 2008a), and water level and coastal inundations (Balay, tion (Gómez-Erache et al., 2010). 1961; Jaime & Menéndez, 1999; Bischoff, 2005; Uruguay has been identified as one of the most exposed D´Onofrio et. al. 1999, 2008). Latin American countries to coastal climate change These events negatively impact the aquatic environment (Magrin et al., 2007; Dasgupta et al., 2007) in terms of and resources, i.e. the artisanal fishing activity (Acuña exposed population to a SLR of one meter, namely

94 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

30 per cent of the country’s population. The estimated above pre-existing high thresholds. Such events cost of climate change on coastal resources for SLR and are generally rare; for example, extreme wind wind-induced flooding + 0.3, 0.5 and 1,0 meter repre- speeds exceeding the 100-year return value, sents two, four and twelve per cent of 2008 GDP re- which have a probability of only 0.01 of occur- spectively (ECLAC, 2011; Nagy et al., in press). ring in any particular year. Usually, ICZM in Uruguay deals with current issues, • High-impact (severe) events: are those that can be including climatic variability and extremes, and vulner- either short-lived weather systems (e.g. severe ability (Nagy et al. 2014a, in press). Notwithstanding, storms) or longer-duration events such as block- there is a lack of a comprehensive study on the vulner- ing episodes that can lead to prolonged heat ability, the risks, and the overall impacts of storm waves and droughts. surges and flooding in the RdlP region. This issue is Extreme events have attributes such as: increasingly considered by the climate adaptation ini- • rate (probability per unit time) of occurrence tiative “the Project”, but it needs yet to be fully integrated by both ICZM and Emergency plans. • magnitude (intensity) • temporal duration and timing 3. Clarifying concepts on extreme events on coastal • spatial scale (footprint) areas • multivariate dependencies Assuming that many readers are more familiar with In- Acute extremes are events that have a rapid onset and tegrated Coastal Management than with weather and follow a short but severe course. Examples are short- climate extremes, and because some terms have differ- lived weather systems such as tropical and extra tropi- ent meanings, some definitions and concepts based on cal cyclones, which are generally referred to as ‘‘wind- recent literature and our experience used in this paper storms’’, and convective storms with extreme values of are presented here. meteorological variables such as wind speed, precipita- A changing climate leads to changes in the frequency, tion, flash or coastal floods that can lead to devastating intensity, spatial extent or duration of weather and cli- wind and flood damage. mate extremes, and can result in unprecedented ex- Here, the use of the term extreme event refers to storm tremes (Field et al., 2012). Extreme weather includes surges measured as water/sea levels at Montevideo unusual, severe or unseasonal weather phenomena at greater than 200 cm associated with cold frontal sys- the extremes of the climatological or historical distribu- tems passages and the strengthening of a cyclone over tion (the range that has been seen in the past) on a geo- Atlantic coast, and strong values of wind speed. The graphical location (Solomon et al., 2007). Extreme 200 cm threshold was selected because this level is weather occurs only 5 to 10 % or less (1 %) of the time. reached at least once a year since the beginning of wa- In recent years some extreme weather events have been ter level records at Montevideo Harbor in 1902 (Pshen- attributed to human-induced global warming, and cli- nikov et al., 2003; Bidegain et al., 2005). Most years mate models and observed trends show that with cli- this value is exceeded several times. Thus, the op- mate change, the planet will experience more extreme erational definition is based on recurrence, regardless of weather (Hansen et al., 2012). flooding and damage. Usually this threshold is not as- According to Stephenson (2008), meteorological events sociated with risks or losses. However, the sedimentary can be classified according their rarity, rapidity and se- balance of sandy beaches may be affected by water lev- verity as: els below the threshold, usually associated with south- • Rare events: are those that have a low probability ern wind anomalies (Gutiérrez et al., 2013, 2015; Or- of occurrence. Because of the rarity of these tega et al., 2013). Also, acute rapid-onset river inflow - events, human societies and ecosystems are often flash floods - or chronic climatic high precipitations in not well adapted to them and so suffer large the La Plata basin, often associated with El Niño events, amounts of damage when they do occur. Hence, are relevant extreme events for the Uruguayan coast of despite their rarity, the large vulnerability associ- the RdlP, due to the positive water level anomaly re- ated with such events can often lead to large lated to the freshwater input exceedance (Nagy et al., mean losses. 2013). • Extreme events: are those that have extreme values of certain important meteorological variables. 4. Methodological approach Damage is often caused by extreme values of All the events exceeding 200 cm at Montevideo on precipitation (e.g. floods), high wind speeds (e.g. hourly basis from 1983-2012 have been considered. cyclones), etc. Extreme is generally defined as ei- Monthly and yearly frequencies were calculated for this ther taking maximum values or exceedance period, computing the month of occurrence. Two case

95 Verocai et al. (2014) studies were selected to compare storm surges at three ones) and November, whereas the ones with lower oc- sites of the Uruguayan coast of the RdlP: i) Colonia currence were July (austral winter), October and De- (tidal freshwater upper estuary, see Figure 1 below), cember. Extremes up to 260-280 cm occurred all over ii) Montevideo (brackish waters within the middle es- the year, whereas those greater than 300 cm occurred tuarine front), and iii) Punta del Este (marine lower es- only in 5 months (January, February, April, June and tuary). The event of March 24-26 1998 was selected August). because the water level reached 300 cm (over the zero reference) and the long duration - 3 days - of the storm Figure 2 - Monthly distribution of storm surges from 1983- surge. The event of August 23-24, 2005 was selected 2012. A) storm surges ≥200 cm and B) storm surges because the water level exceeded 300 cm at Montevi- above 250 cm. deo due to an extremely rapid onset and the intensity of Figura 2 –Distribuição mensal das marés meteorológicas en- the storm surge, associated with an extra tropical cy- tre 1983 e 2012. A) marés meteorológicas ≥200 cm; clone, as well as because it was the only hydro mete- B) marés meteorológicas superiors a 250 cm orological disaster in Uruguay which caused a toll of 10 deaths (Magrin et al., 2007). Notwithstanding, none of deaths was attributed to the storm surge but to very strong winds. Cross correlations (r Pearson) between the monthly Paraná and Uruguay rivers inflows (measured at the basin about 500 km to the northwest of Montevideo) to the RdlP and monthly freshwater/sea levels at Colonia (220 couples of data) and Montevideo (530 couples of data) were calculated. The selection of a monthly time window is due to the lag between river flows and their arrival into the RdlP, estimated as varying from 1 week for very high discharges to 2 months for very low discharges (Nagy, 2000). Thus, all moon phases were taken into account and a 5% error is accepted. More than 1000 couples of data were recorded for southern and southeastern winds and water level at Montevideo. Also, water/sea level at Colonia and Montevideo were related to the river inflow to the RdlP (QRP: Paraná River + Uruguay River) on monthly average basis.

5. Results A slight increase of extreme events was observed over Wind, river flow and water/sea levels the last decade 2003-2012. The maximum level The number of events exceeding 200, 250, 280, and achieved was 320 cm, which is well below the three 300 cm at Montevideo from 1983 – 2012 were 356 historical maxima of 430 cm in 1923 (July), 400 cm in (11,4/year), 52 (1,7/year), 17 (0,56/year), and 6 1914 (April), and 350 cm in 1935 (November). Accord- (0,19/year), respectively (Table 1). ing to ECLAC (2011) and based on larger spatial scales The monthly distribution of events from 1983-2012 is than the ones used here, coastal inundations in the RdlP unequal (Figure 2 and Table 2). The months with show no marked seasonality. greater occurrence of extreme events were January, Table 2 shows the percentage (%) of occurrence of February, March, April, May, June, August (the highest storm surges (i.e. > 200 or 280 cm) for each month.

Table 1 - Decadal frequency of storm surges higher than 200 cm, 250 cm and 280 cm at Montevideo, from 1983 to 2012. Tabela 1 - Frequência decadal de marés meteorológicas maiores do que 200 cm,250 cm e 280 cm, em Montevidéu, entre 1983 e 2012. Decade N° of events (≥200cm) N° of events (≥250cm) N° of events (≥280cm) 1983 – 1992 115 13 1 1993 – 2002 104 19 6 2003 – 2012 137 20 10

96 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

Table 2 - Percentage (%) of occurrence of storm surges greater than a given freshwater/sea level (i.e. > 200 or 280 cm) for each month (January= 1; December= 12) from 1983-2012. The highest monthly occurrences are highlighted as follows: Higher monthly occurrence (dark grey), second higher occurrence (grey), and third higher monthly occurrence (light gray). Tabela 2 - Percentual (%) de ocorrência de marés meteorológicas maiores do que um dado nível de água doce / mar (ou seja, > 200 ou 280 cm) para cada mês (janeiro = 1; dezembro = 12) entre 1983 e 2012. As maiores ocorrências mensais estão realçadas da forma seguinte: maior ocorrência mensal (cinza escuro), segunda maior ocorrência (cinza), e terceira ocorrência mensal mais elevada (cinza claro). Months Level 200 Level 220 Level 240 Level 260 Level 280 Level 300 1 8,3 9,6 11,9 16,7 17,7 14,3 2 10,8 10,1 8,3 5,6 11,7 28,6 3 9,1 10,1 8,3 5,6 5,9 0,0 4 10,2 10,1 9,5 8,3 11,7 14,3 5 9,4 6,9 9,5 11,1 17,7 0,0 6 6,6 8,5 9,5 11,1 5,9 14,3 7 7,2 6,9 3,6 2,8 0,0 0,0 8 7,5 6,9 8,3 11,1 11,7 28,6 9 9,1 9,0 6,0 8,3 11,7 0,0 10 5,5 5,3 7,1 2,8 0,0 0,0 11 7,7 8,5 10,7 11,1 5,9 0,0 12 8,6 8,0 7,1 5,6 0,0 0,0

All the correlations (r Pearson) between river flow nency of water levels higher than MSL is approxi- (River Uruguay-QU or the accumulated of Rivers Uru- mately 50%, whereas it is less than 3, 1 and 0,5% for guay and Paraná-QRP) against sea level at Montevideo water levels higher than 200, 260 and 280 cm respec- were significant (r= p<0.05). The highest correlations tively. Most positive deviations are attributable to non- were found for both Qu and QRP at yearly time-scale. tidal or residual oscillations due to southern winds, low The distribution of data is concentrated around the sea-level pressure (SLP), and high river flow. mean water level for Colonia (73 cm.) and QRP (22,500 m3/s). The lowest values at Colonia, < 60 cm 3 were never observed for high QRP > 40,000 m /s. A positive relationship between river inflow and water level was observed at Colonia (Table 3) and Montevi- deo respectively (Figure 3), showing that river flow is a climatic forcing of water/sea level. The unequal monthly distribution of sea-level at Mon- tevideo is related to prevailing winds (Nagy et al., 1997) and river inflow, especially the Paraná River (QP), the main input of freshwater (75% on average to the RdlP). The latter is shown in Figure 4, where a good agreement is found with this river. When the agreement is lower, this is due to low QP and wind effect, i.e. during December-January, or the periodic Uruguay River floods, during May-June and September-November. 3 The relationship between winds and sea level at Monte- Figure 3 - Observed monthly Rio de la Plata inflow (m /s) video shows that southern and southeastern winds have and water level at Colonia. Dispersion and linear regres- sion, with 95% confidence level limits of data adjusted for a positive and significant correlation when a strong the correlation (lag 0) are shown. positive level anomaly occurs (Figure 5). Figura 3 - Fluxos mensais observados em Rio de la Plata Table 4 summarizes the causes and the magnitude of (m3/s) e níveis da água em Colonia. Representou-se a water/sea level fluctuations at Montevideo around the dispersão e regressão lineares com 95% de intervalo de current mean sea level (MSL) of ≥102 cm. The perma confiança dos dados ajustados à correlação (lag 0).

97 Verocai et al. (2014)

Table 3 - Cross correlations between monthly and yearly river discharges (Uruguay River and RdlP) and freshwater level at Colonia and sea-level at Montevideo for lags 0 and -1 (month/year). All correlations are significant (p<0.05). Tabela 3 - Correlações cruzadas entre as descargas fluviais mensais e anuais (Rio Uruguai e RdlP), níveis de água doce em Colonia e do nível do mar em Montevidéu para defasagens 0 e -1 (mês / ano). Todas as correlações são significativas (p <0,05). Montevideo ( 0 ) Montevideo (-1 ) Colonia ( 0 ) Colonia (-1 ) Monthly River Uruguay 0.09 0.10 0.24 0.15 Monthly RdlP 0.22 0.23 0.34 0.25 Yearly river Uruguay 0.39 0.05 0.48 0.07 Yearly RdlP 0.42 0.20 0.47 0.16

Table 4 - Causes and magnitude of sea level fluctuations at Montevideo. Tabela 4 - Causas e magnitude das flutuações do nível do mar em Montevidéu.

Cause Level and approximate fluctuation (cm) Current MSL ≥102 Typical yearly fluctuations 40 to 200 Tidal oscillation +/- 30 to 40 Residual oscillations Typically from 100 to 200 cm and up to 350 cm River inflow +/- 0 to 25 Barometric setup +/- 0-30 Wind setup +/- 0-300

Figure 4 - Monthly Paraná River flow (light bar, right axis, m3/s) and sea level at Montevideo (green bar, left axis, cm) since 1961. Figura 4 - Fluxos mensais do Rio Paraná (barras mais claras, eixo da direita, m3/s) e níveis do mar em Monte- video (barra verde, eixo esquerdo, cm) desde 1961.

98 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

Figure 5 - Relationship between southeastern (SE) winds and sea-level at Montevideo. Dispersion and linear re- gression of 1294 couples of wind-water level data are shown. Water level for zero SE wind is very close to current mean sea-level (≥ 102 cm) and the interception wind speed 200 cm occurs around 38 knots (19 m/s). Figura 5 - Relação entre ventos do sudeste (SE) e do nível do mar em Montevidéu, com base na dispersão e re- gressão linear de 1.294 pares de dados de nível de vento-água. O nível de água para vento zero de SE está muito perto do nível médio do mar atual (≥ 102 cm) e a intercepção do vento com 200 cm ocorre a cerca de 38 nós (19 m/s).

Case Studies Southwestern and Southern southwestern winds reaching gusts > 40 knots (> 20 m/s) by March 24 which af- Herein the analysis is focused on the synoptic time- fected the Uruguayan coast from Montevideo to Punta scale of two extreme storm surges which serve as ana- del Este. Freshwater inputs (Q ) before this event were logues useful for developing scenarios of extreme RP well above (52,000 m3/s) than usual for March, the events under a changing climate. These events are 1998 average (47,341 m3/s) and the long-term (1961- March 24-26, 1998 and August 23-24, 2005. 2008) average (24,608 m3/s) which was associated with

the strong El Niño 1997-98 event (Figure 7). Yearly Case 1: March 24-26, 1998 1998 sea-level average was around 10 cm above the From March 24-26, 1998 water-level exceeded 300 cm 1971-2003 trend-line (Bidegain et al., 2005; Nagy et during several hours (Figure 6) because of strong al., 2005).

Figure 6 - Water level (WL) at Montevideo: forecasted (clear grey) and observed (dark grey) for the period March 22-28, 1998. The storm surge (residue > 200 cm) develops by the end of March 24 and astronomical spring-tides are observed from March 25. Figura 6 - Nível de água (WL) em Montevideo: previsão (cinza claro) e observado (cinza escuro) para o período de 22-28 março de 1998. A sobreelevação (resíduo > 200 cm) desenvolve-se até ao final de 24 de Março e as maré vivas astronómi- cas são observadas a partir de 25 de março.

99 Verocai et al. (2014)

Figure 7 - Monthly River flows of the Paraná and Uruguay Rivers, and their combined inflow to the Rio de la Plata from 1983-2013. Part of the observed variability is ENSO-related, especially the extremes: La Niña low discharge and El Niño high discharge (i.e. 1997-98). Source: Instituto Nacional del Agua y el Ambiente, INAA, Argentina. Figura 7 - Caudais fluviais mensais dos rios Paraná e Uruguai, e fluxos combinados debitados para o Rio de la Plata no perí- odo 1983-2013. Parte da variabilidade observada está relacionada com a ENSO, especialmente os extremos: La Niña - baixa descarga; El Niño - alta descarga (ou seja, 1997-1998). Fonte: Instituto Nacional del Agua y el Ambiente, INAA, Argentina

The forecasted astronomical high tides were greater An estimate of the impacts is not available, but at Mon- than 100 cm over the zero reference at Montevideo tevideo, the addition of strong winds, storm surge, and (about 30 cm above MSL), which accounted for at least heavy rains impacted the coastal infrastructure and + 20 cm. The wind-induced sea-level increase happened maintenance works of the seawall (BPL, 2009). to coincide with high astronomical tides and very high El Niño-related River flows. This case is an excellent Case 2: August 23-24, 2005 analogy for likely near future threats due to the combination of natural variability and increased climatic pres- From August 23-24, 2005 the passage of a low pressure sures. system or extra tropical cyclone with winds exceeding At Colonia, freshwater level reached a maximum of 170 km/h (up to 190-200 km/h) impacted the RP (Fig- 225 cm on March 25, and the predicted astronomical ure 8), particularly the Uruguayan riverside (Magrin et high-tide was near 100 cm, thus a storm surge al., 2007; Possia et al., 2011). Freshwater inputs (QRP) ≥ 125 cm. At Punta del Este sea-level showed two before this event were below average. peaks ≥ 200 cm recorded on March 25 (separated by the The death toll was 10 and the physical losses were the astronomical low tide). When comparing the three sites, highest ever recorded in Uruguay. For instance, at Mon- the storm surges, that is to say the observed minus fore- tevideo 30% of homes suffered electrical shutdowns for casted water/sea level, was greater at Montevideo and 2 days at least and the supply of drinking water was af- lower at Colonia. fected at several locations. All the harbors,

Figure 8 - Storm surge of August 23, 2005. Predicted heights (red line) and observed heights (blue line). The difference between the both is due to the barometric and wind set-up. Figura 8 – Sobreelevação de 23 de agosto de 2005. Linha vermelha: alturas previstas; Linha azul: alturas observadas. A dife- rença entre os dois é devida ao vento barométrico e à distância de atuação.

100 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015) especially Montevideo, suffered physical losses, in- Thus, a strong pressure gradient was observed to the cluding damaged and sunk ships, and the services were southwest of the low pressure center, eastern and south- affected for several days. The eastern coast of the RdlP, eastern winds blown over the Argentinean Riverside, close to Punta del Este, suffered the loss of at least and eastern winds blown over the RdlP increasing water 160,000 tons of sand and the dunes retreated 3- level at the Argentinean coast, together with strong pre- 5 meters. The quantified losses exceeded U$S cipitations (138 mm on August 23, for a monthly cli- 37 million (BPL, 2009), which seems to be highly un- matic normal 1961-90 = 70 mm). This precipitation was derestimated. a daily record for August and the fourth daily maximum since 1882 (Bidegain et al. 2006). The atmospheric Synoptic situation: August 23, 2005. pressure decreased to < 993 hPa over the RdlP (1800 Z) whereas an anticyclone (1018 hPa) developed to the The evolution of horizontal pressure fields gradients at West over Argentina and southeastern winds blown sea level is a good estimate of the hourly mean surface over the RdlP. wind velocities. A strong low atmospheric pressure or extra tropical cyclone developed on August 22 over Ar- Finally, the atmospheric field pressure over the Atlantic gentina, which is usual in the region, especially during ocean was < 992 (0000 Z, August 24), high pressure austral autumn and spring time. However, the atmos- achieved 1018 hPa over Argentina, southwestern winds pheric perturbation developed in August 2005 was rare blown over the RdlP, the gradient pressure increased due to its intensity. A cold atmospheric frontal system from the Atlantic, and very strong winds blown over the displaced over southern Uruguay on August 23, Uruguayan coast near Montevideo (gusts over 170 km/h whereas another warm front evolved over the Atlantic reaching 190-200 km/h). Water level reached 316 cm. Ocean at 39° south. After the passage of the cold one, a The lag between the observed storm surge at La Paloma high pressure system entered the RdlP from the south- (200 km to the East of Montevideo) and Montevideo west. The surface pressure field (0600 Z) on August 23 from 1988-2008 showed the best cross correlation is showed the entrance of a low pressure ≤ 1002 hPa (Fig- 4,30-5,00 hours. Thus, when storm surges associated ure 9). with eastern and southeastern winds develop, the in-

Figure 9 - Sea Level Pressure forecast for 00 Z August 24, 2005. Figura 9 - Previsão de pressão atmosférica ao nível do mar para 00 Z 24 de agosto de 2005.

101 Verocai et al. (2014) crease in sea level is first observed at La Paloma (LP), (26,798 and 27,846 m3/s respectively). Southern and then at Punta del Este (2 hours after LP), Montevideo southeastern winds exert a strong effect raising water (4-5 hours after LP) and Colonia (6-9 hours after LP). level because they drag water, block or decrease river discharge, generate extraordinary high-tides, waves, 6. Discussion and residual currents stronger than normal. The analysis and monitoring of wave climate, wind and If the events are grouped by season, 29,5% and 20,9% tides system, and the sedimentary transfer process over occurred during Summer and Winter respectively (see time are essential for the implementation of coastal cli- Figure 2). However, QRP is above the average in July, mate change adaptation measures (EcoPlata, 2000; below the average in August, and well below in Sep- Gómez-Erache, 2012, 2013). Unfortunately, the public tember. Thus, river flow is not a driver in July but it institutions and local governments in charge of moni- may be in September. toring, forecasting, early warning system, emergency Monthly occurrence of severe extremes > 280 cm from management, and/or climate adaptation lack an effec- 1983-2012 (< 0,5% of permanency) suggests risks are tive coordination of actions to reduce vulnerability, concentrated in February and August and in a less de- risks, and damages. In this regard, the occurrence of gree in January and April. This occurrence has in- 356, 52 and 17 events exceeding 200, 250 and 280 cm creased over the last three decades from only 1 to 6 and (11,4, 1,7 and 0,56 /year) respectively at Montevideo, 10. Extremes ≥ 300 cm have been only 6 and those that is to say 109, 159 and 189 cm above the historical events >320 cm have not occurred from 1983-2012. reference (91 cm) and 97, 147, and 177 cm above the Notwithstanding, they were relatively frequent before current MSL (≥102 cm) respectively, are useful indi- 1935 (i.e., July 1923, April 1914 and November 1935). cators of risk and are relevant to understand the coast Two of these maxima occurred in July and November, line evolution. which would not be expected from the current climatol- Due to the small tidal amplitude (< 0.5 m), the coast ogy. This observed seasonal pattern of storm surges is line should not be highly resilient to extremes within in disagreement with the lack of seasonality of coastal the range 200-250 cm. The distribution of the events inundations reported by ECLAC (2011) which may be exceeding 200, 250 and 280 cm was not uniform over attributed to the large scale of their study. the last three decades, reaching 13,7, 2 and < 1 From the two case studies (March 1998 and August events/year respectively over 2003-2012. This slight 2005), the former occurred during the peak of local im- increase could be attributed to both the increase in pacts of the ENSO warm phase 1997-98, a strong El southern winds and/or river inflow. Even if precipi- Niño event which caused strong positive anomalies in tations over the RdlP basin and river discharges have river discharges and wind regime changes. Thus, the increased since early 1970s (García & Vargas, 1998; water/sea level was increased depending on the site, Barros et al., 2005), showing strong interannual fluctu- more at the tidal river region (Colonia) and less at the ations partly associated with ENSO variability (Barros marine one (Punta del Este). et al., 2005; Nagy et al., 2002; 2008a), the yearly river Previous works (Bidegain et al., 2005; Nagy et al., inflow average stabilized over the last decade - not the 2005, 2013) have stated that during extreme anomalies variability -, whereas the frequency of southeastern associated with strong El Niño events, sea level at winds increased (Bidegain et al., 2011; Nagy et al., Montevideo fluctuates around 110 cm and plus. During 2013; Gutierrez et al., 2013; Ortega et al., 2013). Fur- the event of August 2005 (almost neutral ENSO) river thermore, the acceleration of SLR from 1971-2002 flows below the average did not account for an increase (Nagy et al., 2005; Magrin et al., 2007) almost stopped in water level. Nevertheless, the 300 cm level was since 2004 which was attributed to less river inflow reached at both case studies. (Nagy et al., 2013, 2014a, b). The influence of river flow and ENSO-related variability on SLR in the Uru- During the event of August 2005 sea-level rose 171 cm guayan coast is explained by the local effect of a great sharply in 10 hours (see Figure 8) and lasted 23 hours to river (Nicholls et al., 2011) such as the RdlP tributaries back to the initial level. Wind velocities were stronger (Nagy et al., 2014a). Thus, it is plausible that during El on the previous day, August 23. Southern winds gusts Niño years, the positive anomalies of both freshwater exceeded 170 Km/h, with a maximum in the range 190- inflow and southeastern winds lead to water/sea-level 200 km (Possia et al. 2011). A one day lag between increases. southern wind velocities and water level increase has Also, the higher frequency of extreme water levels from been frequently observed, but it is not discussed here. February-April can be partly associated with high QRP This event was a clear evidence of the lack of prepared- during March and April, which are the third and first ness of Uruguay against extreme events. The death toll higher recorded monthly QRP over the last 52 years was 10, which is considered a hydro meteorological

102 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015) disaster according to the United Nations International Table 2), beach quality and tourism are impacted during Strategy for Disaster Reduction (UNISDR) criteria. 24-48 hours because the beaches are closed to bath Despite the availability of abundant physical informa- (Nagy et al., 2014b). tion and the improvement over the last few years (since Existing coastal gravity drainage, storm water infra- 2009-11), there is no proper coordination between pub- structure and sewerage systems may become gradually lic institutions in charge of the survey, prognosis and affected over time as the mean sea level and/or storm early warning of extreme events in coastal areas. The surges increase. Some of the common impacts are as expected impacts of storm surges were classified and an follows: associated risk was assigned according to their proba- • The maritime channel signalization buoys can be re- bility of occurrence and to the consequent sea level at leased from its anchorage, displaced, or may suffer Montevideo (Table 5). light disconnection, becoming useless and even a Three questions emerge in regards to these coastal ex- threat for navigation instead of an aid. treme events: • The artisanal fisheries located at Pajas Blancas to i) What would have happened on August 2005 if the west of Montevideo, which exploit the croaker river flows had been twice or three times greater? Micropogonias furnieri during the spawning pe- ii) Was the event of 1998 so strong mainly because riod from October to March, are affected by south- the coincidence of very high river inflow and ern winds (Norbis, 1995; Nagy et al., 2008b, maximum spring-tide? 2014b), and during an extreme event, the currents iii) What would be the impact if events in the range and waves drag their fishing nets. Thus, besides 350-430 cm occur again? the economic and environmental impacts, fisher- When storm surges occur at Montevideo, even the more men’s lives are at risk because they try to recover their nets before the storm becomes very strong. frequent ones, in the range 200-250 cm, sewage outfalls are assisted by hydraulic pumps to redirect surplus • The sandy beaches, dunes and bars are highly vul- water to overflow channels, and when water level ex- nerable and threaten by both storm surges and SLR ceeds 250 cm, they become useless and sewage is dis- (Panario & Gutiérrez, 2006; Gutierrez et al., 2013, persed in the coastal waters and beaches, contaminating 2015; Ortega et al., 2013; Nagy et al., 2014a). them with fecal coliforms. Since these events are more Storm surges ≥ 260-280 cm at Montevideo are able frequent during the austral summer (see Figure 2 and to flood most sandy beaches and bars.

Table 5 - Water/Sea Level (cm), consequences, and probability of occurrence of storm surges (> 200 cm). The extreme coastal events painted in light gray were observed in this study (modified from Verocai et al., 2014). Tabela 5 - Nível da água / do mar (cm), consequências e probabilidade de ocorrência de sobreelevações (> 200 cm). Os eventos extremos costeiros destacados em cinza claro foram observados neste estudo (modificado de Verocai et al., 2014). IMPACTS Sea Level (cm) Slightly harmful Moderately harmful Highly Harmful Extremely Harmful (>200) (>250) (>300) (>350) High: 1 each 1 month Moderate Risk High Risk Very high Risk Intolerable Risk P Medium: 1 each 6 months Tolerable Risk Moderate Risk High Risk Very high Risk R O Low: 1 each 5 years Trivial Risk Tolerable Risk Moderate Risk High Risk B A B I Very Low: 1 each > 30 years Negligible Risk Trivial Risk Tolerable Risk Moderate Risk L I T Y

103 Verocai et al. (2014)

7. Conclusions ticularly in regards to climatic and meteorological threats is the deficit of an effective and integrated hy- Main findings drological, meteorological and oceanographic, moni- Sea level rise (SLR) and storm surges are not strongly toring network. The three of them exist, are not coordi- present in Uruguay’s public agenda. The main reasons nated yet, resulting ineffective to prevent an eventual for this are that long (i.e., > 12 hours) > 320-350 cm event. i) SLR does not represent a serious risk for human set- The occurrence of extremes greater than 280 and tlement and infrastructure (yet); 300 cm from 1983-2012 shows that risks are concen- ii) the occurrence of storm surges which cause moder- trated in February, August, and less in January and ate to high impacts, i.e., > 250 and > 280 cm, and April. Nevertheless, the historical maximum occurred particularly > 300 cm are not very frequent; and in July 1923, which would not to be expected under current climate patterns. iii) those events able to harm many coastal services The public institutions and local governments in charge and built environment, i.e., > 350 cm, have not occurred over the last eight decades. of monitoring, forecasting, modelling, early warning system (EWS), and emergency management lack an ef- Notwithstanding, the high number of events > 200 cm fective coordination of actions to reduce vulnerability, and mainly those reaching > 250 and > 280 cm ob- risks, and damages. Despite a significant progress over served during the studied period (1983-2012) impact the last few years (since 2009-11), there is a less than the coastal services, morphology, ecosystem resources, ideal coordination between monitoring, preparedness, and even the infrastructure and housing. Thus, there is a alert, and reactive phases in order to face extremes need to include storm surges in coastal management, greater than those usual. climate adaptation and risk-management initiatives, plans and programs. Recommendations

Case studies A few easy, low cost and viable useful improvements are to: Two case studies of storm surges which serve as ana- • Enhance coordination of all public institutions in logues useful for developing scenarios of extreme charge of observing and monitoring climatic, hy- events under a changing climate were presented. These drological, and oceanographic variables, as well as examples show evidence supporting a plausible alter- those in charge of preparedness, disaster manage- native scenario of extreme wind storm, very high river ment, and EWS, such as the National System of Re- discharge, spring high tide, and increasing SLR which sponse to Climate Change and Variability (SNRCC). could favour the occurrence of water levels ≥ 300 cm • Concentrate monitoring and EWS efforts when the reaching ≥ 350 cm. The latter could be accepted as a first signals are forecasted and observed, some of rough estimation impact threshold based on observed which have been identified in this article, including impacts (i.e., > 400 cm in July 1923), and the recent the analysis of past extreme weather causality, stat- ones (i.e., 300-320 cm). A scenario including SLR istical occurrence, climatic forecasts, SST 3.4 evo- and/or – SLP + wind + river induced flooding - 20- lution, modeling, hydrologic observations, and/or 30 cm and plus could reach this threshold at any time weather forecasts. over the next few decades. • Improve the tactics and procedures to achieving effective preparedness for coastal extreme events Application of the knowledge on extreme events on based on past experience and failures. In this regard, ICZM and Early Warning Systems a more detailed analysis of the causality and impacts The coastal threats described in this paper call for of events > 280 cm over the last 30 years should coastal management and planning which take into ac- serve to better understanding and simulate the extra- count the vulnerability of major systems and sectors in ordinary extremes > 350 cm which occurred before the face of climate change, SLR and wind-induced 1935. The frequency and permanency of events flooding. Emphasis is put on the plausible occurrence of > 280 cm, which have increased over the last three extreme events greater (i.e., > 320-350 cm) than those decades should be closely followed as an indicator usual over the last 30 years (i.e., 280-320 cm), under of change, impact and risk. slow and gradual SLR, and especially strong El Niño • To update the estimation of potential economic, years, when increased river flows and southeastern human and environmental impacts associated with a winds increase the magnitude and impact of flooding. long (> 12 hours) > 350 cm storm surge at Monte- One of the obstacles to achieving a comprehensive video useful to be incorporated in coastal zone use, integrated coastal management in Uruguay and par infrastructure, regulations and governance.

104 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

References 3P/IB/2007/02/09/000016406_20070209161430/Rendered/PDF/wps413 6.pdf Acuña, A.; Verocai, J.E.; Márquez, S. (1992) - Aspectos biológicos DECCW (2010) – Coastal Risk Management Guide. Incorporating de Micropogonias furnieri (Desmarest, 1823) durante dos zafras sea level rise benchmarks in coastal risk assessments. 6p., De- en una pesquería artesanal al oeste de Montevideo. Revista de partment of Environment, Climate Change and Water NSW. Biología Marina (ISSN: 0717-3326), 27(1):113-132, Val- ISBN: 978-1742329222. Avaliable on-line at http://www. environ- paraíso, Chile. Available on-line at http://www.revbiolmar.cl/ ment.nsw .gov.au. escaneados/271-113.pdf, http://www.undp.adaptationlearning.net/ spa_uruguay. Echevarría, L.; Gómez, A.; Piriz, C.; Quintas, C.; Tejera, R.; Conde, Balay, M.A. (1961) - El Río de la Plata entre la atmósfera y el mar. D. (2013) - Capacity building for local coastal managers: a par- 153p., Publicación H-621, Servicio de Hidrografía Naval ticipatory approach for Integrated Coastal and Marine Zones (SHN), Secretaria de Argentina.Marina, Buenos Aires. Management in Uruguay. Journal of Integrated Coastal Zone Management, 13(4):445-456. DOI: 10.5894/rgci402 Ballinger, R.; Rhisiart, M. (2011) - Integrating ICZM and futures approaches in adapting to changing climates. MAST - Maritime ECLAC (2011) - Effects of climate change on the coast of Latin America and the Caribbean: dynamics, trends and climate vari- Studies (ISSN: 2212-9790), 10(1):115-138, Amsterdam. Avail- able on-line at http://www.marecentre.nl/mast/documents/Pagesfrom ability. LC/W.447. Economic Commission for Latin America MAST10.1_Ballinger_Rhisiart.pdf. and the Caribbean (ECLAC), United Nations, Santiago de Chile, Chile. Available on-line at http://www.cepal.org/ publica- Barros, V.; Menéndez, A.; Nagy, G.J. (2005) - El Cambio Climático ciones/xml/2/45542/W.447.pdf. en el Río de la Plata. 200p., Project assessments of impacts and adaptation to climate change (AIACC), CIMA/CONICET, Bue- ECOPLATA, (2000) - Diagnóstico Ambiental y Socio-Demográfico nos Aires, Argentina. Available on-line at http://www.cima.fcen.uba de la Zona Costera Uruguaya del Río de la Plata. Recopilación .ar/~lcr/libros/Cambio_Climatico-Texto.pdf. de informes técnicos. 991p., López Laborde, J., Perdomo, A., Gómez-Erache, M. (eds.) CIID/PNUD/MVOTMA/UNESCO Bidegain, M.; Caffera, R.M.; Blixen, F.; Pshennikov, V.; Lago- /EcoPlata, Montevideo, Uruguay. Available on-line at marsino, J.J.; Forbes, E.A.; Nagy, G.J. (2005) - Tendencias cli- http://www.pnuma. org/gobernanza/cd/Biblioteca/Zonas Costeras/3 Di- máticas, hidrológicas, y oceanográficas en el Río de la Plata y agnostico_Ambiental _y_Socio-Demografico_de_la_Zona_Costera.pdf Costa Uruguaya. In: V. Barros, Á. Menéndez, G.J. Nagy (Eds.), El Cambio Climático en el Río de la Plata, pp.137–143, Project Escobar, G.; Vargas, W.; Bischoff, S. (2004) - Wind tides in the Rio assessments of impacts and adaptation to climate change de la Plata estuary: meteorological conditions. International Journal of Climatology, 24(9):1159–1169. DOI: 10.1002/ (AIACC), CIMA/CONICET, Buenos Aires, Argentina. Available on-line at http://www.cima.fcen.uba.ar/~lcr/libros/Cambio_ Climatico- joc.1026. Texto.pdf. Field, C.B.; Barros, V.; Stocker, T.F.; Qin, D.; Dokken, D.J.; Ebi, Bidegain, M.; Caffera, R.M.; de los Santos, B.; Castellazzi, P.; K.L.; Mastrandrea, M.D.; Mach, K.J.; Plattner, G.K.; Allen, Gómez-Rivera, J. (2006) – Performance of the WRF Regional S.H.; Tignor, M.; Midgley, P.M. (2012) - Managing the Risks of Model over SouthEastern South America during an Extreme Extreme Events and Disasters to Advance Climate Change Ad- Event. In: Proceedings of 8 ICSHMO, pp.1655-1658, INPE, Foz aptation. A Special Report of Working Groups I and II of the In- do Iguaçu, Brazil. Available on-line at http://mtc-m15.sid.inpe tergovernmental Panel on Climate Change. 582p., Cambridge .br/col/cptec.inpe.br/adm_conf/2005/11.01.19.21/doc/1655-1658.pdf. University Press, Cambridge, UK, and New York, NY, USA. Bidegain, M.; de los Santos, B.; Skusunosky, S.; Kitoh, A. (2011) - ISBN: 978-1-107-02506-6. Available on-line at https://www.ipcc- wg1. unibe.ch/srex/downloads/SREX-All_FINAL.pdf Escenarios Climáticos futuros con el modelo japonés MRI- CGCM2.3. In: Previsão e Modelação em Ciências Geofísicas, Flick, R.E.; Chadwick, D.B.; Briscoe, J.; Harper, K.C. (2012) - Meteorología e Tema, Anais do 7° Simpósio de Meteorologia e Flooding” versus “inundation”. Eos, 93(38):365-366. American Geofísica da APMG and XIV Congresso Latino-Americano e Geophysical Union. DOI: 10.1029/ 2012EO380009. Ibérico de Meteorologia, Setubal, Portugal. Available on-line at Framiñan, M.B.; Brown, O.B. (1996) – Study of the Río de la Plata http://www.apmg.pt/?page_id=323. turbidity front, Part I: spatial and temporal distribution. Conti- Bischoff, S. (2005) – Sudestadas. In: V. Barros, Á. Menéndez, G.J. nental Shelf Research, 16(10):1259-1282. DOI: 10.1016/ 0278- Nagy (Eds.), El Cambio Climático en el Río de la Plata, pp.55– 4343(95)00071-2 67, Project assessments of impacts and adaptation to climate García, N.O.; Vargas, W.M. (1998) - The temporal climatic vari- change (AIACC), CIMA/CONICET, Buenos Aires, Argentina. ability in the ‘Río De La Plata’ basin displayed by the river dis- Available on-line at http://www.cima.fcen.uba.ar/~lcr/libros/Cambio_ charges. Climatic Change, 38(3):359-379. DOI: 10.1023/ Climatico-Texto.pdf. A:1005386530866 BPL (2009) - Archivos de Prensa escrita (Diarios consultados: El Gómez-Erache, M. (2013) - Variabilidad y Cambio Climático en la País, La República, El Observador, Ultimas noticias, La Zona Costera. Evaluación de los impactos factibles y com- Mañana, El Diario). Biblioteca del Palacio Legislativo, Mon- probados. Informe del Proyecto GEF URU/07/G32. Imple- tevideo, Uruguay. Available on-line at http://biblioteca.parlamento. mentación de Medidas de Adaptación al Cambio Climático en gub.uy:8080/ áreas costeras del Uruguay. Unidad de Cambio Climático, D'Onofrio, E.E.; Fiore, M.M.E.; Pousa, J.L. (2008) - Changes in the MVOTMA, 32p., Montevideo, Uruguay. Available on-line at Regime of Storm Surges at Buenos Aires, Argentina. Journal of http://www.undp.adaptationlearning.net/spa_uruguay Coastal Research, 24(1A):260-265. DOI: 10.2112/05-0588.1 Gómez-Erache, M. (2012) - Análisis de la vulnerabilidad de la D’Onofrio, E.E.; Fiore, M.M.E., Romero, S.I., (1999) - Return peri- costa uruguaya ante el aumento del nivel medio del mar. ods of extreme water levels estimated for some vulnerable areas Informe del Proyecto GEF URU/07/G32. Implementación de of Buenos Aires. Continental Shelf Research, 19(13):1681- Medidas de Adaptación al Cambio Climático en áreas costeras 1693. DOI: 10.1016/S0278-4343(98)00115-0. del Uruguay. Unidad de Cambio Climático, MVOTMA, 43p., Dasgupta, S.; Laplante, B.; Meisner, C.; Wheeler, D.; Yan, J. Montevideo, Uruguay. Available on-line at http://www.undp. (2007) - The impact of sea level rise on developing countries: A adaptation learning.net/spa_uruguay comparative analysis. Washington, DC: World Bank Policy Re- Gómez-Erache, M.; Conde, D.; Villarmarzo, R. (2010) - Sustain- search Working Paper – WPS 4136, online. Available on-line at ability of integrated management in the coastal zone of Uru- http://www-wds.worldbank.org/external/default/WDSContentServer/IW guay: Connecting knowledge to action. Programa EcoPlata /

105 Verocai et al. (2014)

IDRC-CRDI, Montevideo, Uruguay. ISBN: 978-9974-7646-8-2. Vivienda, Ordenamiento Territorial y Medio Ambiente. August, Available on-line at http://www.ecoplata.org/wpcontent/files_mf/ 2010 Montevideo, Uruguay. ecoplata_thesustainabilityofimczingles.pdf. Medina, M.; Gómez-Erache, M. (2014c) - Formulación integral del Gutiérrez, O.; Panario, D.; Bidegain, M.; Montes, C. (2013) – Tele- Programa Estratégico de Adaptación Costera. Informe técnico conexiones El Niño - La Niña con retrocesos - avances de Nº 6. 85p. Convenio específico entre la Universidad de la playas del estuario medio del Río de la Plata. In: XV Congreso República y el Ministerio de Vivienda, Ordenamiento Territorial Latinoamericano de Ciencias del Mar (COLACMAR), Resumen y Medio Ambiente. August, Montevideo, Uruguay. N° 2668-3, 27-31 october 2013, Punta del Este, Uruguay. Nagy, G.J.; Gómez-Erache, M.; Kay, R. (In Press) - A risk-based Gutiérrez, O.; Panario, D.; Nagy, G.J.; Piñeiro, G.; Montes, C. and participatory approach to assessing climate vulnerability (2015) - Long-term morphological evolution of urban pocket and improving governance in coastal Uruguay. In: B. Glavovic, beaches in Montevideo (Uruguay): impacts of coastal interven- R. Kay, M. Kelly, A. Travers (Eds.), Climate change and the tions and links to climate forcing. Journal of Integrated Coastal coast. Building resilient communities. CRC Press / Taylor & Zone Management, in press. Francis, London. ISBN: 978-0-415-46487-1 Hansen, J.; Sato, M.; Ruedy, R. (2012) - Perception of climate Nagy, G.J.; Seijo, L.; Verocai, J.E.; Bidegain, M. (2014a) - change. PNAS Plus_109(37): E2415-E2423. DOI: 10.1073/ Stakeholders’ climate perception and adaptation in coastal Uru- pnas.1205276109. guay. International Journal of Climate Change Strategies and IOC (2009) - The International Oceanogragaphic Commission Management, 6(1):63-84. DOI: 10.1108/IJCCSM-03-2013- (IOC) of UNESCO and its role in ICAM. 1-3 December 2009, 0035. Montevideo, Uruguay. Available on-line at http://www.unesco.org. Nagy, G.J.; Munoz, N.; Verocai, J.E.; Bidegain, M.; de los Santos, uy/ci/fileadmin/ciencias%20naturales/COI/Seminario%202009/PDFs/ B.; Seijo, L.; Feola, G.; Brena, B.; Risso, J.; Martin, J. (2014b) - StephanoBelfiore.pdf. Integrating Climate Science, Monitoring, and Management in Jaime, P.R.; Menéndez, A.N. (1999) - Modelo Hidrodinámico “Río the Rio de la Plata Estuarine Front (Uruguay). In: Leal Filho,W.; de la Plata 2000”. Informe LHA 01-183-99, Instituto Nacional Alves, F.; Azeiteiro, U.; Caeiro, S. (eds.), International perspec- del Agua y el Ambiente (INA), Secretaría de Recursos Naturales tives on climate change: Latin America and Beyond, pp.79-91. y Desarrollo Sustentable, Ezeiza, Buenos Aires. Available on-line Springer, Heidelberg. ISBN:978-3-319-04488-0. DOI: at http://www.ina.gov.ar/pdf/LH-Info%20RDP%20LHA-01-183- 10.1007/978-3-319-04489-7. 99%20-%20RioDeLaPlata-RP2000%20-%20Sep-1999.pdf. Nagy, G.J.; Muñoz, N.; Verocai, J.E.; Bidegain, M.; Seijo, L. Lappo, S.S.; Morozov, E.G.; Severov, D.N.; Sokov, A.V.; Klyuvit- (2014c) - Adjusting to current climate threats and building alter- kin, A.A.; Nagy, G. (2005) - Frontal mixing of river and sea wa- native future scenarios for the Rio de la Plata coast and estuarine ters in Rio de La Plata. Doklady Earth Sciences C/C of Doklady front, Uruguay. Journal of Integrated Coastal Zone Manage- - Akademiia Nauk (ISSN: 1028-334X), 401(2):226-228. ment, 14(4): published on-line at 20 June 2014. DOI: Magrin, G.; García, C.G.; Choque, D.C.; Giménez, J.C.; Moreno, 10.5894/rgci472. A.R.; Nagy, G.J.; Nobre, C.; Villamizar, A. (2007) - Latin Ame- Nagy, G.J.; Seijo, L.; Verocai, J.E.; Brugnoli, E.; Bidegain, M. rica. Climate Change 2007: impacts, adaptation and vulnerabil- (2013) - Enfoque, conocimiento y medidas para enfrentar las ity. In: Parry, M.L.; Canziani, O.F.; Palutikof, J.P.; van der Lin- amenazas del clima presente en la zona frontal del Río de la den, P.J.; Hanson, C.E. (eds.), Contribution of Working Group II Plata, Uruguay. Costas, Revista Iberoamericana de Manejo to the Fourth Assessment Report of the Intergovernmental Panel Costero Integrado (ISNN: 2304-0963), 2(2):69-87, Montevideo, on Climate Change, pp. 581–615, Cambridge University Press, Uruguay. Available on-line at http://unesco.unesco.org/images/0022/ Cambridge. ISBN: 978-0521 88010-7. Available on-line at 002270/227017m.pdf - 227739 http://www. ipcc.ch/publications_and_data/ar4/wg2/en/ch13.html. Nagy, G.J.; Severov, D.N.; Pshennikov, V.A.; De los Santos, M.; Medina, M. (2009) - Diagnóstico y evaluación de infraestructuras e Lagomarsino, J.J.; Sans, K.; Morozov. E.G. (2008a) - Rio de la la zona costera uruguaya (Colonia – Rocha). Segundo Informe: Plata estuarine system: Relationship between river flow and Estado de situación de la infraestructura existente en la zona frontal variability. Advances in Space Research, 41(11):1876- costera. EcoPlata conectando el conocimiento con la acción 1881. DOI: 10.1016/j.asr.2007.11.027. para la gestión integrada de la zona costera uruguaya del Río de Nagy, G.J.; Bidegain, M.; Norbis, W.; Ponce, A.; Pshennikov, V.; la Plata, Proyecto URU/06/016. August, Montevideo, Uruguay. Severov, D.N. (2008b) - Fishing strategies for managing climate Available on-line at http://www.ecoplata.org/adjuntos/2010/08/ Diag- variability and change in the Estuarine Front of the Río de la nostico-y-Evaluacion-de-Infraestructuras-en-la-Zona-Costera- Uruguaya.pdf Plata. In: N. Leary, J. Adejuwon, V. Barros, I. Burton, J. Kul- karni & R. Lasco, (eds.), Climate Change and Adaptation, Medina, M.; Gómez-Erache, M. (2014a) - Borrador inicial para la Chapter 20, pp.353-370, Earthscan, London. ISBN: 978- discusión en el Sistema Nacional de Respuesta al Cambio Cli- 1844074709. mático de la línea de base actualizada de acciones de las Inten- Nagy, G.J.; Gomez-Erache, M.; Fernandez, V. (2007) - El aumento dencias y relevamiento de capacidades y necesidades de forta- lecimiento institucional para la adaptación costera al cambio del nivel del mar en la costa uruguaya del Río de la Plata: climático. Informe técnico Nº 3 y Nº 4. 115p. Convenio espe- Tendencias, vulnerabilidades y medidas para la adaptación. Medio Ambiente y Urbanizacion (ISSN: 0328-0306), 67(1):77-93, Bue- cífico entre la Universidad de la República y el Ministerio de nos Aires, Argentina. Available on-line at http://www.ingentaconnect Vivienda, Ordenamiento Territorial y Medio Ambiente. August, .com/content/iieal/meda/2007/00000067/00000001/art00006 Montevideo, Uruguay. Nagy, G.J.; Caffera, R.M.; Aparicio, P.; Barrenechea, P.; Bidegain, Medina, M.; Gómez-Erache, M. (2014b) - Mapeo de infraes- M.; Giménez, J.C.; Lentini, J.; Magrin, G.; Murgida, A.; Nobre, tructuras urbanas vulnerables ante la variabilidad y el cambio C.; Ponce, A.; Travasso, M.; Villamizar, A.; Wehbe, M. (2006) - climático en la franja costera. Identificación de sitios de vulne- Understanding the Potential Impacts of Climate Change and rabilidad significativa por cada Departamento y formulación Variability in Latin America and the Caribbean. Report prepared del conjunto de medidas que se deberían implementar en cada for the Stern Review on the Economics of Climate Change. De- sitio para incrementar la resiliencia a los efectos del cambio partment for International Development-DFID, HM Treasury, climático. Informe técnico Nº 5. 235p. Convenio específico United Kingdom. Available on-line at http://www.dfld.de/ Presse/2006 entre la Universidad de la República y el Ministerio de /061030cA.pdf.

106 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):91-107 (2015)

Nagy, G.J.; Ponce, A.; Pshennikov, V.; Silva, R.; Forbes, E.A.; 36(2):95-110, Centro Argentino de Meteorólogos, Buenos Ai- Kokot, R. (2005) - Desarrollo de la capacidad de evaluación de res, Argentina. Available on-line at http://www.cenamet.org.ar/ ar- la vulnerabilidad costera al Cambio Climático: Zona Oeste de chivos/Met.Vol 36 N2 2011.pdf. Montevideo como caso de estudio. In: B. Barros, A. Menéndez, G.J. Nagy, (eds.), El Cambio climático en el Río de la Plata, PRODOC (2008) - Pilot Sites of Adaptation Measures to Climate pp.173–180, Project assessments of impacts and adaptation to Change in the Uruguayan Coastal Areas. 131p., project docu- climate change (AIACC), CIMA/CONICET, Buenos Aires, ment prepared for UNDP/GEF Trust Fund. Available on-line at Argentina. Available on-line at http://www.cima.fcen.uba.ar/ http://www.undp.adaptationlearning.net/spa_uruguay ~lcr/libros/ Cambio_Climatico-Texto.pdf. Pshennikov, V.; Bidegain, M.; Blixen, F.; Forbes, E.A.; Lagomars- Nagy, G.J.; Gómez-Erache, M.; López, C.H.; Perdomo, A.C. (2002) ino, J.J.; Nagy, G.J. (2003) – Climate extremes and changes in - Distribution patterns of nutrients and symptoms of eutrophica- precipitation and wind-patterns (1971-2002) in the vicinities of tion in the Rio de la Plata river estuary system. Hydrobiología, Montevideo, Uruguay. paper presented at the Session 3: Ob- 475/476:125-139. DOI: 10.1007/978-94-017-2464-7_10. served Climate Change and Climate Variability, in the First AIACC Regional Workshop for Latin America and Caribbean, Nagy, G.J. (2000) - Uruguay estuarine systems: The frontal zone of José de Costa Rica, Available on-line at http://www.start. the Río de la Plata system, Uruguay/Argentina. In: S. V. Smith, org/Projects/AIACC_Project/meetings/meetings.html V. Dupra, J. I. Marshall-Crossland & C. J. Crossland, C. (eds.), Estuarine systems of the South American region: carbon, nitro- Saizar, A. (1997) - Assessment of impacts of a potential sea-level gen and phosphorus fluxes, pp.40-43, LOICZ Reports & Studies rise on the coast of Montevideo, Uruguay. Climate Research, No. 15, LOICZ-UNEP, Texel, The Netherlands. ISSN: 1383- 9(1-2):73–79. DOI: 10.3354/cr009073. 4304. Available on-line at http://www.loicz.org/imperia/md/content/ Simionatto, C.G.; Vera, C.S.; Siegismund, F. (2005) - Surface wind loicz/print/rsreports/15report.pdf. variability on seasonal and interannual scales over Río de la Nagy, G.J.; Martínez, C.M.; Caffera, R.M.; Pedrosa, G.; Forbes, Plata area. Journal of Coastal Research, 21(4):770–783. DOI: E.A.; Perdomo, A.C.; López Laborde, J. (1997) - Marco hidro- 10.2112/008-NIS.1. lógico y climático del Río de la Plata (The Hydrological and Simionatto, C.G.; Núñez, M.N.; Engel, M. (2001) – The salinity Climatic Setting of the Rio de la Plata). In: P.G. Wells & G. R. front of the Rio de la Plata - A numerical case study for Winter Daborn (eds.), The Río de la Plata: An Environmental Over- and summer conditions. Geophysical Research Letters, view. An EcoPlata Project Background Report. pp.17-70, Dal- 28(13):2641-2644. DOI: 10.1029/2000GL012478. housie University, Halifax, Nova Scotia, Canada. ISBN: Solomon, S.; Qin, D.; Manning, M.; Chen, Z.; Marquis, M.; Averyt, 0770328520. Available on-line at http://www.ecoplata.org/wp-content/ K.B.; Tignor, M.; Miller H.L. (2007) - Climate Change 2007: files_mf/el_rio_de_la_plata_una_revision_ambiental78.pdf. The Physical Science Basis. Contribution of Working Group I to Nicholls, R.J.; Hanson, S.E.; Lowe, J.A.; Warrick, R.A.; Lu, X.; the Fourth Assessment Report of the Intergovernmental Panel Long, A.J.; Carter, T.R. (2011) - Constructing Sea-Level Sce- on Climate Change. Cambridge University Press, Cambridge, narios for Impact and Adaptation Assessment of Coastal Area: UK and New York, NY, USA. ISBN 978-0-521-88009-1. Avail- A Guidance Document. Supporting Material, Intergovernmental able on-line at http://www.ipcc.ch/pdf/assessment-report/ ar4/wg1/ar4 Panel on Climate Change Task Group on Data and Scenario _wg1_full_report.pdf Support for Impact and Climate Analysis (TGICA), 47p., Avail- Stephenson, D.B. (2008) - Definition, diagnosis, and origin of ex- able on-line at http://www.ipcc-data.org/docs/Sea_Level_Scenario treme weather and climate events. In: Diaz, H.F.; Murnane, R.J. Guidance_Oct2011.pdf. (Eds.), Climate Extremes and Society, pp.11-23, Cambridge Norbis, W. (1995) – Influence of wind, behaviour and characteris- University Press, Cambridge, UK. ISBN: 978-0521870283. tics of the croaker (Micropogonias furnieri) artisanal fishery in DOI: 10.1017/CBO9780511535840.00410.1017/CBO9780511 the Rio de la Plata (Uruguay). Fisheries Research, 22(1-2):43- 535840.004. 58. DOI: 10.1016/0165-7836(94)00310-S Verocai, J.; Bidegain, M.; Nagy, G.J. (2014) - Nivel del mar y Ortega, L.; Celentano, E.; Finkl, C.; Defeo, O. (2013) - Effects of eventos extremos en las aguas costeras del Río de la Plata y climate variability on the morphodynamics of Uruguayan sandy Costa Oceánica Uruguaya. In: Goso, C. (Ed.), Problemática de beaches. Journal of Coastal Research, 29(4):747–755. DOI: los Ambientes Costeros, Capítulo 6, pp.131-146, DIRAC, 10.2112/JCOASTRES-D-13-00003.1 Facultad de Ciencias, Universidad de la República, Montevideo, Panario, D.; Gutiérrez, O. (2006) - Dinámica y fuentes de sedi- Uruguay. mentos de las playas uruguayas. In: R. Menafra, L. Rodríguez- Volonte, C.R.; Nicholls, R.J. (1995) – Uruguay and sea level rise: Gallego, F. Scarabino & D. Conde (Eds.), Bases para la con- potential impacts and responses. Journal of Coastal Research servación y el manejo de la costa uruguaya, pp.21–34, Vida Sil- (ISSN: 0749-0208), SI14:262–284. Article Stable URL: vestre, Montevideo, Uruguay. ISBN: 9974758920. Available on- http://www.jstor.org/stable/25735712 line at http://vidasilvestre.org.uy/wp-content/uploads/2012/09/03_ Dinámicas-y-fuentes-de-sedimentos-de-las-playas-uruguayas-Panario .pdf Possia, N.E.; Vidal, L.; Campetella, C.C. (2011) – Un temporal de viento en el Río de la Plata. Meteorologica (ISSN: 1850-468X),

107

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):109-120 (2015)

http://www.aprh.pt/rgci/pdf/rgci-492_Silva.pdf | DOI:10.5894/rgci492

Análise da estabilidade da Praia do Janga (Paulista, PE, Brasil) utilizando ferramenta computacional *

Elida Regina de Melo e Silva@, a; Daniele Laura Bridi Mallmanna; Pedro de Souza Pereiraa

Resumo Este artigo tem por objetivo a análise de estabilidade da forma em planta das baías formadas entre os quebra-mares da praia do Janga (Paulista, Pernambuco, Brasil). No local, onde a erosão se intensificou após os anos 90, foi construída uma série de nove quebra-mares destacados no intuito de conter o processo, dando origem a uma sequência de baías artificiais. Assim, cada baía foi tratada em duas metades, de modo a estudar a estabilidade de cada enseada. Praias de enseada se caracterizam por apresentar uma linha de costa arenosa bordejada por afloramentos ou promontórios rochosos de origem natural ou artificial (como quebra-mares), a qual é modelada por processos de difração e refração de ondas. O modelo parabólico permite, por meio de uma equação empírica, que seja determinada a situação de equilíbrio morfodinâmico das praias de enseada. No intuito de permitir a aplicação do modelo parabólico de modo manual com o uso de mapas, fotografias aéreas verticais, ortofotocartas ou imagens de satélite a praias de enseadas, a partir das quais seja possível o estabelecimento dos parâmetros iniciais do modelo, foi desenvolvido o software Mepbay. Esse software foi utilizado na área de estudo para analisar a estabilidade das enseadas formadas junto à sequência de quebra-mares da Praia do Janga. Para tanto, foram utilizadas imagens verticais obtidas gratuitamente por meio do programa GoogleEarth® e de seu banco de dados associado. As imagens selecionadas foram aplicadas ao modelo com o uso do software, e os resultados obtidos foram comparados aos resultados da aplicação de relações empíricas desenvolvidas para estimar a resposta da linha de costa frente à construção de quebra-mares, a partir da geometria da obra e da sua distância até a costa. Os resultados demonstraram que a Praia do Janga pode ser classificada como instável em sua maior parte, uma vez que a configuração da linha de costa nas diversas baías geradas pela difração de ondas entre os quebra-mares ainda não alcançou a configuração da projeção gerada pelo modelo parabólico em sua totalidade. Naquelas em que a praia já está quase ajustada à projeção, pode-se dizer que está próxima do seu equilíbrio estático. A aplicação do modelo parabólico através do software Mepbay mostrou-se uma ferramenta de fácil aplicação e rápida resposta, que pode ser utilizada para predizer o ajuste da linha de costa mediante a construção de uma estrutura de proteção costeira. No entanto, o aplicativo não realiza predições acerca da morfologia final em planta (tômbolo ou saliência) que se desenvolverá a partir da difração sofrida pela onda junto à zona de sombra da estrutura de proteção. Por essa razão, uma alternativa interessante e que foi adotada neste estudo é a combinação da aplicação do software com o uso de relações empíricas. As análises aqui apresentadas, feitas de maneira visual, qualitativa e com poucos recursos, permitem concluir que o Mepbay é potencialmente útil para estudos rápidos voltados ao manejo da linha de costa. Palavras-chave: Enseada, quebra-mares, modelo parabólico, saliência, tômbolo.

Abstract Stability assessment of Janga Beach (Paulista, PE, Brazil) using computational tool This article aims to analyze the stability of the planform shape of the bays formed between the breakwaters located at Janga Beach (Paulista, Pernambuco, Brazil). On site, where erosion has intensified in the 90’s, were constructed a series of nine detached breakwaters in order to contain the process, creating a sequence of artificial bays. Thus, each bay was treated in

@ Corresponding author: [email protected] a Universidade Federal de Pernambuco, LABOGEO – Laboratório de Oceanografia Geológica, Departamento de Oceanografia, Avenida Acadêmico Hélio Ramos s/n° CEP: 50740-530, Pernambuco, Brasil.

* Submission: 16 JAN 2014; Peer review: 13 FEB 2014; Revised: 12 MAY 2014; Accepted: 26 JUL 2014; Available on-line: 24 SEP 2014

Silva et al. (2015) two halves, in order to study the stability of each bay-shaped beach. Bay-shaped beaches are characterized by having a sandy shoreline bordered by rocky headlands or outcroppings of natural or artificial origin (such as breakwaters), which is modeled by processes of wave diffraction and refraction. The parabolic model allows, by means of an empirical equation determine the type of morphodynamic stability of the beaches. In order to allow the application of the parabolic model manually with the use of maps, aerial photographs, orthophotos or satellite images from which it is possible to establish the initial parameters of the model, were developed the software Mepbay. This software was used in the study area to analyze the stability of bay-shaped beaches formed by the sequence of breakwaters at Janga Beach. For this purpose, vertical images obtained free through GoogleEarth ® software and its associated database were used. The selected images were applied to the model using the software and the results were compared to the results of applying empirical relationships developed to predict the response of the shoreline opposite the construction of breakwaters, based on the structure geometry and their distance from the coast. The results showed that Janga Beach can be classified as unstable for the most part, once that the shoreline configuration in several bays generated by wave diffraction of between the breakwaters has not yet reached the setting of the projection generated by the model parabolic in its entirety. Those where the beach is almost adjusted to the projection, we can say that is close to its static equilibrium. The application of the parabolic model by Mepbay software proved to be a tool for easy application and quick response that can be used to predict the adjustment of the coastline by building a coastal protection work. However, it does not do predictions about the final morphology in plan (tombolo or salience) that will develop from the diffraction suffered by the wave near the shadow of the protective zone. For this reason, an interesting alternative, which was adopted in this study is the combination of the software application using empirical relationships. The analyzes presented here, made visually, qualitative way and with few resources, we can conclude that Mepbay is potentially useful for quick studies related to the management of the shoreline.

Keywords: pocket beach, breakwaters, parabolic model, salient, tombolo.

geometria da praia e dados sobre a dinâmica da 1. Introdução antepraia (por exemplo, o ponto de difração e a direção Praias de enseada se caracterizam por apresentar uma preferencial de incidência de ondas) (Hsu et al., 2008), linha de costa arenosa bordejada por afloramentos ou sendo bastante adequado para a análise da estabilidade. promontórios rochosos, a qual é modelada por No intuito de facilitar a aplicação do modelo parabólico processos de difração e refração de ondas, que as em praias de enseada, foi desenvolvido o software tornam curvilíneas e, em geral, assimétricas. Embora Mepbay (Klein et al., 2003), que foi criado com a sejam comuns praias naturais com tais características, finalidade de definir qual a condição de uma inúmeras têm sido criadas de forma artificial por meio determinada praia de enseada, em termos da de projetos de engenharia que incluem estruturas de estabilidade da sua linha de costa, e qual forma esta virá proteção costeira e portos (Klein et al., 2003). a assumir se continuar exposta à mesma dinâmica. No que tange à estabilidade, as praias de enseada Desde então, esse software tem se mostrado uma podem apresentar equilíbrio estático, dinâmico ou importante ferramenta para auxílio de profissionais das instabilidade (Silvester & Hsu, 1993, 1997; Hsu et al., áreas de Engenharia Costeira e Geociências na análise 2000). No equilíbrio estático, as ondas incidentes morfológica das praias de enseada e suas alterações atingem simultaneamente toda a periferia da baía, de tal (Klein et al., 2003; Raabe et al., 2010). forma que a deriva litorânea é praticamente inexistente. O presente artigo apresenta resultados da análise de Nessas condições, não há erosão ou acresção de longo- estabilidade da forma em planta das baías formadas termo, apenas por ocasião de tempestades. Para baías entre os quebra-mares da praia do Janga (município de em equilíbrio dinâmico, a manutenção da linha de costa Paulista, Pernambuco, Brasil), realizada através da em uma mesma posição depende do equilíbrio no aplicação do modelo parabólico a imagens de satélite balanço sedimentar, e a linha de costa pode mudar de com uso do software Mepbay. Adicionalmente, são posição em função de variações no suprimento de apresentadas análises geradas a partir da comparação sedimentos. Finalmente, enseadas instáveis são aquelas desses resultados com dados oriundos da aplicação de nas quais ocorre deslocamento da linha de costa, em equações empíricas para predizer o comportamento da geral, erosão em um dos extremos e acresção no outro. linha de costa frente à zona de sombra dos referidos Tal deslocamento pode ser observado e medido por quebra-mares. meio de levantamentos em campo ou a partir de fotografias aéreas multitemporais (Klein et al., 2003; 2. Área de Estudo Silveira et al., 2010). O município de Paulista, no qual se localiza a Praia do Inúmeros modelos empíricos têm sido elaborados para Janga, dista cerca de 20km a nordeste de Recife, capital representar a forma em planta das praias de enseada. do estado, estando inserido na Região Metropolitana do Dentre esses modelos, o modelo parabólico (Hsu & Recife e sendo delimitado pelos rios Timbó, a norte, e Evans, 1989) incorpora como parâmetros diretos a Paratibe, a sul. A planície costeira do município segue

110 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):109-120 (2015) os padrões da região, com menos de 10km de largura e imobilizada pela presença de um enrocamento, não altitude média de 4m, é formada por depósitos permitindo a aplicação do modelo. As imagens sedimentares e alinhada na direção NNE-SSO. Extensos utilizadas para aplicação do modelo e verificação da alinhamentos de beachrocks constituem as feições mais resposta da costa datam, respectivamente, de conspícuas da linha de costa (Neves & Muehe, 1995). 01/07/2003 e 12/01/2009. Os ecossistemas presentes na região incluem mangue- zais, recifes de coral e restingas. A região apresenta temperatura média anual em torno de 27°C e pluviosidade aproximada de 2.000 mm/ano, distribuídos de modo desigual entre períodos secos e chuvosos. Predominam os ventos de SE. As ondas respondem ao regime dos ventos, predominando de SE. As marés registradas na região são semi-diurnas, sendo classificadas em termos de amplitude como mesomarés. A altura significativa de ondas varia de acordo com a localidade, apresentando valores entre 0,27/0,29m para o município de Paulista (FINEP/ UFPE, 2009). A praia do Janga localiza-se no município de Paulista, na Região Metropolitana do Recife, estando situada entre as coordenadas 7° 57' 8.64'' e 7° 55' 50.88'' S e 34° 48' 56.16'' W e 34° 49' 22.08'' W, apresentando uma extensão de 3,8km (Fig. 1). A área possui um conjunto de nove quebra-mares, além de obras de contenção da erosão que incluem enrocamento, muros de contenção e aterro hidráulico. Os quebra-mares possuem comprimento variável de 120 a 320m, com ênfase para o último, que, após a conclusão da obra, passou a ter 650m (CPRH/GERCO- PE, 2001). A distância média entre os quebra-mares e a costa é de 150m. O conjunto de obras de contenção do processo erosivo foi construído a partir da demanda gerada pela transferência de erosão do município de Olinda, situado a sul. A erosão atingiu a área a partir da década de 80, após a construção dos molhes do Rio Paratibe (também Figura 1 - Localização da área de estudo indicando os quebra-mares nomeados de Q1 a Q8 de sul para norte conhecido como Rio Doce), tendo evoluído na década (Imagem: GoogleEarth). de 90 para uma situação catastrófica. A partir de então, Figure 1 - Study area indicating the breakwaters which are foram construídos os quebra-mares, cuja obra foi named from Q1 to Q8 from south to north. (Image: concluída no ano de 2001 (Costa, 2002). GoogleEarth).

3. Metodologia (ii) Aplicação do modelo A metodologia aplicada neste estudo foi dividida em três etapas: (i) seleção das imagens; (ii) aplicação do O modelo parabólico consiste num modelo empírico modelo; e (iii) análise da estabilidade das praias, que permite determinar a situação de equilíbrio conforme descrito a seguir. morfodinâmico das praias de enseada. Schiaffino et al. (2012) descrevem que, para o referido modelo (Fig. 2), (i) Aquisição das imagens o sistema de coordenadas tem início em (X0, Y0); (X1, As imagens verticais utilizadas foram obtidas Y1) é definido como ponto cuja tangente à linha de gratuitamente por meio do programa GoogleEarth® e costa é paralelo à crista de onda. Cada ponto em (X,Y) de seu banco de dados associado. Foi utilizado um pertence à linha de costa definida por coordenadas recorte da imagem de satélite para cada baía formada polares (R,θ). Os coeficientes (C1, C2 e C3) dependem por um conjunto de dois quebra-mares. Uma lacuna de β (ângulo formado entre as linhas de crista de onda entre dois quebra-mares, localizada ao norte na área de predominantes e a linha de controle Rβ, sendo esta a estudo, entre os quebra-mares Q8 e Q9, não pôde ser linha que une o ponto de controle localizado no analisada, uma vez que sua linha de costa se encontra promontório rochoso – onde se inicia o processo de

111 Silva et al. (2015)

Para a aplicação do modelo parabólico com uso do Mepbay, as imagens selecionadas foram carregadas no software, no qual foi feito o seu escalonamento. Tal procedimento é realizado através da associação manual de uma feição observada na imagem à sua medida real conhecida em campo. Posteriormente, foram inseridos os pontos de controle. Para o adequado estudo do comportamento de uma praia de enseada através da metodologia aqui aplicada, é necessário saber, ainda, a localização do promontório rochoso (natural ou artificial), o final da praia e a direção de incidência predominante das ondas (Fig. 3). É através da indicação desses pontos que o Mepbay calcula a posição da linha de costa teórica através do modelo parabólico de Hsu & Evans (1989) e a projeta sobre a imagem base. Figura 2 - Esquema que ilustra a definição dos parâmetros principais do modelo parabólico para praias de enseada. Segundo Klein et al., 2003. Figure 2 - Sketch for definition of major parameters of parabolic bay shape model. According to Klein et al., 2003. difração de ondas – até a extremidade final da praia). O modelo parabólico foi desenvolvido por Hsu & Evans (1989), sendo descrito pela seguinte equação (Eq. 1). !"!#=$0+$1#%"+ $2#%" 2(1) O software Mepbay foi desenvolvido para permitir a aplicação do modelo parabólico de modo manual com o uso de mapas, fotografias aéreas verticais, ortofotocartas ou imagens de satélite a praias de enseadas a partir das quais seja possível o estabelecimento dos parâmetros iniciais do modelo (β e Rβ). Os resultados Figura 3 - Definição dos pontos de controle. Adaptado de dependem da resolução da imagem utilizada (Raabe et Klein et al., 2003. al., 2010) e da habilidade do operador em estabelecer os Figure 3 - Definition of the control points. Adapted from pontos que determinam os dados de entrada. Klein et al., 2003). .

Tabela 1 - Resultados da aplicação de equações empíricas para predizer o comportamento da linha de costa frente à construção de quebra-mares (Fonte: Imambaks et al., 2011; Marinho, 2012). Table 1 - Results of empirical equations aplication to predict the shoreline behavior after the breakwaters construction (Source: Imambaks et al., 2011; Marinho, 2012). Classificação Quebra- Comprimento Distância da Relação Is=EXP(1,72- Classificação (Ahrens & mar (L) (m) costa (D) (m) D/L 0,41*L/D) (Bosboom & Stive, 2011) Cox, 1990) Q1 127 143 1,13 3,88 Saliência Saliência pouco desenvolvida Q2 250 118 0,59 2,79 Tômbolo Saliência bem desenvolvida Q3 204 172 0,84 3,43 Saliência Saliência bem desenvolvida Q4 240 170 0,71 3,13 Tômbolo Saliência bem desenvolvida Q5 310 180 0,58 2,76 Tômbolo Saliência bem desenvolvida Q6 240 210 0,88 3,50 Saliência Saliência pouco desenvolvida Q7 180 192 1,07 3,80 Saliência Saliência pouco desenvolvida Q8 270 130 0,48 2,38 Tômbolo Tômbolo periódico

112 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):109-120 (2015)

Definidos os pontos de controle, o modelo foi aplicado, modelos empíricos serão referidos como (AC90) para resultando na geração de linhas teóricas calculadas pelo Ahrens & Cox (1990) e (BS11) para Bosboom & Stive programa sobre a imagem, a partir das quais é possível (2011). inferir sobre o tipo de equilíbrio em que a praia se encontra. Cabe observar que, para a aplicação da equação 4. Resultados e Discussão parabólica às baías, foi necessária a divisão da baía em Os resultados da aplicação do modelo parabólico duas partes, criando duas enseadas, visto que o modelo através do software Mepbay são apresentados a seguir, foi desenvolvido para praias com essa configuração. com uma breve interpretação para cada uma das baías (iii) Análise da estabilidade das praias analisadas. As pranchas mostram as projeções feitas pelo modelo (com base em imagens datadas de 2003) e A análise da estabilidade da praia do Janga foi feita observadas em imagens mais recentes (2009). através da comparação da linha de costa teórica calculada pelo programa Mepbay e a linha de costa BAÍA 1 mais atual obtida nas imagens do GoogleEarth, Para o segmento de linha de costa da praia do Janga que possibilitando a discussão sobre o tipo de equilíbrio em sofre influência do primeiro quebra-mar (Q1) (Fig. 4), a que a enseada se encontra. Foram, ainda, utilizados tendência projetada pelo modelo parabólico utilizando a relatórios e monografias (Imambaks et al., 2011; imagem de 2003 indica erosão na extremidade sul da Marinho, 2012) como material de apoio para embasar a área sob influência do quebra-mar e acúmulo de análise. sedimentos localizado imediatamente à zona de sombra Para fins de discussão, são apresentados na tabela 1, (zona abrigada da ação direta de ondas) da obra. De conforme Imambaks et al. (2011), os resultados da acordo com os resultados das relações empíricas aplicação de relações empíricas baseadas nos modelos (Imambaks et al., 2011), há tendência para a formação desenvolvidos por Ahrens & Cox (1990) e Bosboom & de saliência (AC90) ou saliência pouco desenvolvida Stive (2011) para aplicar à geometria dos quebra-mares (BS11) junto à zona de sombra do quebra-mar. A do Janga e predizer o comportamento da linha de costa análise da imagem de 2009 do mesmo segmento revela junto à zona de sombra da obra. A interpretação permite que houve erosão na extremidade sul, indicada pela inferir sobre a morfologia final da costa, se uma presença de um enrocamento e deposição de sedimentos saliência ocorrerá, se esta será pouco ou bem configurando um tômbolo periódico, isto é, um tômbolo desenvolvida, se ocorrerá um tômbolo periódico que somente é visível durante a baixamar na área de (apenas na maré baixa) ou um tômbolo permanente. Os sombra formada por Q1. A projeção do software sobre

Guia Corrente – Q1

Tendência projetada pelo modelo Tendência observada (Imagem de Tendência observada (Imagem 2009) parabólico (Imagem de 01/07/2003) 12/01/2009) sobreposta pela projeção do modelo parabólico (2003) Figura 4 - Análise da estabilidade da linha de costa para a enseada protegida pelo quebra-mar Q1. (Imagens: GoogleEarth). Figure 4 - Shoreline stability analysis made for bay protected by the breakwater Q1. (Images: GoogleEarth).

113 Silva et al. (2015) a erosão no segmento foi confirmada. Entretanto, de acordo com observação é possível dizer que houve um acúmulo de sedimentos maior que o previsto pela aplicação das equações empíricas, cuja previsão seria de saliência pouco desenvolvida (BS11) ou saliência (AC90), indicando que no processo de sedimentação há outros fatores interferindo na configuração em planta da zona de sombra do quebra-mar que não são avaliadas pela metodologia utilizada (isto é, da geometria da obra). A previsão para a feição morfológica da zona de sombra se ajustou àquela prevista, no entanto, o processo erosivo do centro da baía estimado pelo Mepbay excede o observado. Desta forma, pode-se dizer que este trecho da linha de costa se encontra em equilíbrio dinâmico, ainda se ajustando ao seu estado final de equilíbrio. Tendência projetada pelo modelo parabólico (Imagem de 01/07/2003) BAÍA 2 Na imagem de 2003, observou-se que a projeção do modelo parabólico para a configuração da linha de costa apresentada diante da difração de onda que ocorre nos quebra-mares Q1 e Q2 (Fig. 5) é de meia-lua, com erosão no centro da baía e acresção nas laterais. Os resultados obtidos através das relações empíricas

apontam para a formação de saliência (AC90) ou

saliência bem desenvolvida (BS11) para o quebra-mar 1 Q2 - e tômbolo (AC90) ou saliência bem desenvolvida Q1 (BS11) para o quebra-mar 2. As imagens de 2009 permitem observar a formação de um tômbolo periódico para o Q1 e de um tômbolo para Q2. A análise comparativa entre a projeção gerada pelo modelo e a configuração em planta observada na imagem de 2009 foram confirmadas, com exceção da erosão no centro da baía. Diante do cenário observado pode-se dizer que Tendência observada (Imagem de 12/01/2009) esta praia se encontra em equilíbrio dinâmico, tendendo a erodir o fundo da baía até chegar ao equilíbrio previsto. BAÍA 3 Com base nas informações da imagem de 2003, a baía projetada pelo Mepbay como resultado da difração causada pelo conjunto de quebra-mares Q2 e Q3 (Fig. 6) indica uma tendência erosiva no centro da baía e progradante nas extremidades. Tal projeção está de acordo com a análise semântica da imagem mais recente, que mostra acresção na extremidade sul com formação de tômbolo e uma modesta acresção na extremidade norte, além de erosão no centro da baía. Entretanto, como a saliência observada na zona de sombra de Q3 ainda não atingiu totalmente a configuração prevista pela aplicação das relações Tendência observada (Imagem 2009) sobreposta pela empíricas, que apontam para o desenvolvimento de uma projeção do modelo parabólico (2003) saliência ou saliência bem desenvolvida, e o contorno Figura 5 - Análise da estabilidade da linha de costa para a da baía também não alcançou desenho indicado pelo baía, formada entre Q1 e Q2 (Imagens: GoogleEarth). modelo, pode-se dizer que a praia se encontra em Figure 5 - Shoreline stability analysis made for the bay instabilidade. formed between Q1 and Q2 (Images: GoogleEarth).

114 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):109-120 (2015)

Tendência projetada pelo modelo parabólico Tendência projetada pelo modelo parabólico (Imagem de 01/07/2003) (Imagem de 01/07/2003)

Q3 Q4 - - Q2 Q3

Tendência observada (Imagem de 12/01/2009) Tendência observada (Imagem de 12/01/2009)

Tendência observada (Imagem 2009) sobreposta pela Tendência observada (Imagem 2009) sobreposta pela projeção do modelo parabólico (2003) projeção do modelo parabólico (2003) Figura 6 - Análise da estabilidade da linha de costa para a Figura 7 - Análise da estabilidade da linha de costa para a baía 3, formada entre Q2 e Q3. (Imagens: GoogleEarth). baía 4, formada entre Q3 e Q4. (Imagens: GoogleEarth). Figure 6 - Shoreline stability analysis made for the bay 3 Figure 7 - Shoreline stability analysis made for the bay 4 formed between Q2 and Q3. (Images: GoogleEarth). formed between Q3 and Q4. (Images: GoogleEarth).

115 Silva et al. (2015)

BAÍA 4 O contorno da baía projetado pelo modelo parabólico para o trecho entre os quebra-mares Q3 e Q4 (Fig. 7) indica erosão no centro da baía e acresção nas laterais. A imagem mais recente mostra a acresção nas laterais, confirmando parcialmente a projeção gerada pelo modelo parabólico, já que não ocorreu erosão no setor indicado. No local, pode-se observar inclusive vegetação incipiente na pós-praia. As previsões baseadas nas relações empíricas indicam que, na área de sombra de Q3, poderá se formar uma saliência (AC90) ou uma saliência bem desenvolvida (BS11), e na zona de sombra de Q4, um tômbolo (AC90) ou uma saliência bem desenvolvida (BS11). As projeções não coincidem Tendência projetada pelo modelo parabólico com o observado na imagem de 2009, e a linha de costa (Imagem de 01/07/2003) se encontra à retroterra em relação ao previsto, configurando instabilidade. BAÍA 5 A projeção do modelo parabólico baseado na imagem de 2003 indica tendência erosiva no centro da baía e progradante nas laterais. Utilizando a imagem de 2009 entre os quebra-mares Q4 e Q5 (Fig. 8), observou-se Q5 que a projeção do modelo foi confirmada pela formação - de saliências na zona de sombra dos quebra-mares 4 e Q4 5. Tais observações estão de acordo, ainda, com as predições feitas a partir da geometria dos quebra-mares, as quais indicam, para ambos os quebra-mares, a formação de saliências bem desenvolvidas (AC90) ou tômbolos (BS11). Esta baía foi classificada como instável, porque a configuração em planta das Tendência observada (Imagem de 12/01/2009) extremidades ainda não atingiu a conformação projetada pelo Mepbay, indicando que o processo erosivo do centro da baía pode ter continuidade até atingir o limite indicado na projeção. BAÍA 6 Para a baía formada entre os quebra-mares 5 e 6 (Fig. 9), a partir da imagem base de 2003, o modelo parabólico projetou uma configuração indicativa de erosão no centro da baía com acresção nas extremidades. A observação da imagem de satélite datada de 2009 confirmou as previsões feitas a partir das relações empíricas, as quais indicam o acúmulo de sedimentos na zona de sombra dos quebra-mares com a consequente formação de saliência bem desenvolvida (AC90) e saliência pouco desenvolvida (BS11), em Q5 e Q6, respectivamente. Sob o ponto de vista da Tendência observada (Imagem 2009) sobreposta pela estabilidade praial, a baía foi classificada como instável, projeção do modelo parabólico (2003) devido a o desenho da forma em planta não ter Figura 8 - Análise da estabilidade da linha de costa para a alcançado a configuração indicada na projeção do baía 5, formada entre Q4 e Q5. (Imagens: GoogleEarth). modelo. Figure 8 - Shoreline stability analysis made for the bay 5 formed between Q4 and Q5. (Images: GoogleEarth).

116 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):109-120 (2015)

Tendência projetada pelo modelo parabólico

(Imagem de 01/07/2003) Tendência projetada pelo modelo parabólico (Imagem de 01/07/2003)

Q6 Q7 - - Q5 Q6

Tendência observada (Imagem de 12/01/2009) Tendência observada (Imagem de 12/01/2009)

Tendência observada (Imagem 2009) sobreposta pela Tendência observada (Imagem 2009) sobreposta pela projeção do modelo parabólico (2003) projeção do modelo parabólico (2003) Figura 9 - Análise da estabilidade da linha de costa para a Figura 10 - Análise da estabilidade da linha de costa para a baía 6, formada entre Q5 e Q6. (Imagens: GoogleEarth). baía 7, formada entre Q6 e Q7. (Imagens: GoogleEarth). Figure 9 - Shoreline stability analysis made for the bay 6 Figure 10 - Shoreline stability analysis made for the bay 7 formed between Q5 and Q6. (Images: GoogleEarth). formed between Q6 and Q7. (Images: GoogleEarth).

117 Silva et al. (2015)

BAÍA 7 A projeção baseada na imagem de 2003, entre os quebra-mares Q6 e Q7 (Fig. 10), resultou em um arco mais fechado, indicando erosão no centro da baía e acresção nas laterais, sendo mais acentuada a norte. Por sua vez, os resultados da aplicação das relações empíricas apontam para a formação de saliência (AC90) ou saliência bem desenvolvida (BS11) nas zonas de sombra de Q6 e Q7. A imagem de 2009 mostra que, salvo o aumento da saliência na zona de sombra de Q6 e uma pequena acresção no centro da baía, não houve grande alteração na configuração em planta da linha de costa. Cabe ressaltar que, nesta área, parte da ponte de acesso ao quebra-mar, utilizada na época da construção da obra, foi mantida, o que modifica os efeitos de difração e refração na área abrigada. Como a Tendência projetada pelo modelo parabólico (Imagem de 01/07/2003) configuração em planta projetada não foi alcançada e há acúmulo de sedimento junto aos quebra-mares, indicando uma possível reformulação do desenho dessas feições em planta, a baía foi classificada como instável. BAÍA 8

As predições realizadas por Imambaks et al. (2011), 8

através da aplicação das relações empíricas, indicam a Q - formação de saliência ou saliência bem desenvolvida na 7 Q zona de sombra de Q7 e tômbolo (AC90) ou tômbolo periódico (BS11) na zona de sombra de Q8. Analisando a imagem de 2009, observou-se que as projeções do modelo parabólico e as predições empíricas previstas para o quebra-mar 8 ocorreram, entretanto não houve modificações notáveis no contorno da linha de costa na área de influência do quebra-mar 7. Outro fato Tendência observada (Imagem de 12/01/2009) observado foi que o contorno da linha de costa de toda a baía não se ajustou à projeção do modelo parabólico, e a linha de costa está à retroterra da linha estimada. Por esse motivo, a baía foi classificada como instável (Fig. 11). Diante dos resultados acima, podemos classificar a praia do Janga como instável em sua maior porção, pois a configuração da linha de costa nas diversas baías, gerada pela difração de ondas entre os quebra-mares, ainda não alcançou a configuração da projeção gerada pelo modelo parabólico desenvolvido por Hsu & Evens (1989) em sua totalidade. Naquelas em que a praia já está quase ajustada à projeção, pode-se dizer que está próxima do seu equilíbrio estático (Klein et al., 2003). Embora as tendências indicadas pela aplicação das relações empíricas à geometria dos quebra-mares Tendência observada (Imagem 2009) sobreposta pela (Imambaks et al., 2011), as tendências projetadas pelo projeção do modelo parabólico (2003) Mepbay e as observadas nas imagens de satélite Figura 11 - Análise da estabilidade da linha de costa para a estejam, na sua maior parte, de acordo, cabe ressaltar baía 8, formada entre Q7 e Q8. (Imagens: GoogleEarth). que existe uma gama de parâmetros que definem a Figure 11 - Shoreline stability analysis made for the bay 8 forma em planta de uma baía e que não foram aqui formed between Q7 and Q8. (Images: GoogleEarth). abordados, motivo pelo qual os resultados devem ser tomados com critério. Benedet et al. (2005) mencionam 118 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):109-120 (2015) abordados, motivo pelo qual os resultados devem ser Referências bibliográficas tomados com critério. Benedet et al. (2005) mencionam Ahrens, J.P.; Cox, J. (1990) - Design and Performance of Reef como alguns desses parâmetros a orientação da praia, a Breakwaters. Journal of Coastal Research (ISSN 0749-0208) configuração batimétrica e o suprimento sedimentar. SI7:61-75, Fort Lauderdale, FL, U.S.A. Article Stable URL: Ademais, vale ressaltar que tanto o modelo parabólico http://www.jstor.org/stable/25735389. quanto os modelos de AC90 e BS11 não indicam o Benedet, L.; Klein, A.H.F.; Hsu, J.R.C. (2005) - Practical insights and applicability of empirical bay shape equations. Coastal tempo necessário para que a praia atinja o seu equilíbrio Engineering 2004, 2:2181-2193. DOI: 10.1142/97898127 ou a formação das saliências e dos tômbolos, 01916_0175. respectivamente. Tal limitação sugere a inclusão em Bosboom, J.; Stive , M.J. (2011) - Coastal Dynamics I, Lecture estudos futuros de uma análise multitemporal com mais notes - version 0.4 2013. 573p., Delft Academic Press, Delft, imagens que permitam estabelecer uma tendência e Netherlands. ISBN: 978-9065622860. comparar essa tendência às predições dos modelos. Costa, J.E.R. (2002) - Morfodinâmica praial do município de Paulista-PE. 40p., Monografia de Especialização em 5. Conclusões Oceanografia, Universidade Federal de Pernambuco, Recife, PE, Brasil. Não publicado. A construção de estruturas rígidas como alternativa de CPRH/GERCO-PE (2001) - Obras de Proteção do Litoral de proteção costeira deve ser realizada com cautela, Paulista: Histórico do Processo. 150p. Relatório técnico, sempre acompanhada de estudos que mostrem o seu Volume único, Agência Estadual de Recursos Hídricos(CPRH) / Gerenciamento Costeiro Recife(GERCO), impacto sobre a dinâmica costeira. A aplicação do Recife, PE, Brasil. Não publicado. modelo parabólico através do software MepBay FINEP/UFPE (2009) - Monitoramento Ambiental mostrou-se uma ferramenta de fácil aplicação e rápida Integrado/Pernambuco – MAI/PE. 485p., Relatório Final, 3 resposta, que pode ser utilizada para predizer o ajuste Volumes, Financiadora de Estudos e Projetos (FINEP) / da linha de costa mediante a construção de uma Universidade Federal de Pernambuco (UFPE), Recife, PE, estrutura, como no caso dos quebra-mares. No entanto, Brasil. Não publicado. o aplicativo não realiza predições acerca da morfologia Hsu, J.R.C.; Benedet, L.; Klein, A.H.F.; Raabe, A.L.A.; Tsai, C.P.; HSU, T.W. (2008) - Appreciation of Static Bay Beach final em planta (tômbolo ou saliência), que se Concept for Coastal Management and Protection. Journal of desenvolverá a partir da difração sofrida pela onda Coastal Research, 24(1):198-215. DOI: 10.2112/05-0579.1. junto à zona de sombra da estrutura de proteção. Por Hsu, J.R.C.; Evans, C. (1989) - Parabolic bay shapes and essa razão, uma alternativa interessante e que foi applications. ICE Proceedings, 87(4):557–570. DOI: 10.1680 adotada neste estudo é a combinação da aplicação do /iicep.1989.3778. software com o uso de relações empíricas, como as de Hsu, J.R.C.; Uda, T.; Silvester, R. (2000) - Shoreline protection Ahrens & Cox (1990) e Bosboom & Styve (2011). methods—Japanese experience. In: J.B. Herbich (ed.), Handbook of Coastal Engineering, pp. 9.1–9.77, McGraw- O modelo foi aplicado às baías formadas pelo campo de Hill, Nova York, NY, U.S.A. ISBN: 978-0071344029. quebra-mares da praia do Janga para estimar o seu Imambaks, R.; Anijs, M.; Kibrit, B.; Sedaghat, A.; Mungar, S. estágio de equilíbrio, e sua aplicação permite concluir (2011) – Recife: Coastal Protection Plan. 275p. Relatório que as baías analisadas, em sua maior parte, ainda não Técnico, volume 1 e appendix. Delft University of alcançaram seu equilíbrio estático, indicando que o Technology (DUT), Delft, Zuid-Holland, Netherlands. Report Stable URL: http://repository.tudelft.nl/view/ir/uuid% processo erosivo no centro das baías deverá permanecer 3A70277a41-5591-4c98-9633-06542a0a2ac1/. até que esse processo alcance o limite projetado pelo Klein, A.H.F.; Vargas, A.; Raabe, A.L.A.; Hsu, J.R.C. (2003) - Mepbay. O sedimento oriundo da erosão no centro da Visual assessment of bayed beach stability with computer baía possivelmente migrará para as extremidades, software. Computer & Geosciences, 29(10):1249 -1257. DOI: complementando a configuração projetada pelo 10.1016/j.cageo.2003.08.002. software. No caso das baías instáveis, é necessário que Marinho, J.S. (2012) - Avaliação dos efeitos da presença dos haja suprimento sedimentar superior à retirada, para que quebra-mares sob a linha de costa da praia do Janga - Pernambuco - Brasil. 46p., Trabalho de conclusão de curso, a praia alcance a conformação prevista. Faculdade Maurício de Nassau, Recife, Pernambuco, Brasil. A ferramenta permite a aplicação da equação parabólica Não publicado. para prever a curvatura de praias de enseada como Neves, F.N.; MUEHE, D. (1995) - Potential impacts of sea level- resposta da linha de costa à existência de dois pontos de rise on the Metropolitan Region of Recife, Brazil. Journal of Coastal Research (ISSN: 0749-0208) SI14: 116-131, Fort controle decorrentes da presença de obras costeiras Lauderdale, FL, U.S.A. Article Stable URL: http://www.jstor. como quebra-mares. Outro fato importante a ser org/stable/25735704. destacado é a possibilidade de uma análise consistente Raabe, A.L.A.; Klein, A.H.F.; González, M.; Medina, R. (2010) - dos efeitos promovidos por modificação da MEPBAY & SMC: Software tools to support different configuração em planta da baía mediante alteração da operational levels of headland bay beach coastal engineering localização dos pontos de difração das ondas. Tais projects. Coastal Engineering, 57(2):213–226. DOI: 10.1016/ j.coastaleng.2009.10.008. aplicações permitem concluir que o Mepbay é Schiafﬁno, C.F.; Brignone, B.; Ferrari, M. (2012) - Application of potencialmente útil para estudos rápidos voltados ao the parabolic bay shape equation to sand and gravel beaches manejo da linha de costa. 119 Silva et al. (2015)

on Mediterranean coasts. Coastal Engineering, 59(1):57–63. Silvester, R.; Hsu, J. R.C. (1993) - Coastal stabilization: Innovative DOI: 10.1016/j.coastaleng.2011.07.007. concepts. 578p., Prentice-Hall, Englewood Cliffs, NJ, U.S.A. Silveira, L.F.; Klein, A.F.; Tessler, M.G. (2010) - Headland-bay ISBN: 0131403109. beach planform stability of Santa Catarina state and of the Silvester, R.; Hsu, J. R.C. (1997) - Coastal Stabilization - Advanced northern coast of São Paulo state. Brazilian Journal of Series on Ocean Engineering vol.14. 578 p., World Scientific, Oceanography, 58(2):101-122. DOI: 10.1590/s1679- Singapore. ISBN: 9810231377. 87592010000200003.

120 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):121-131 (2015)

http://www.aprh.pt/rgci/pdf/rgci-540_Germani.pdf | DOI:10.5894/rgci540

Vulnerabilidade costeira e perda de ambientes devido à elevação do nível do mar no litoral sul do Rio Grande do Sul *

@, Yana Friedrich Germani@, a,; Salette Amaral de Figueiredob; Lauro Júlio Calliarib; Carlos Roney Armanini Taglianib

Resumo No atual cenário de mudanças climáticas globais, em que alguns dos efeitos observados são a elevação do nível médio do mar, e o aumento na frequência e intensidade de eventos extremos, regiões costeiras de todo o mundo deverão tornar-se, de forma geral mais vulneráveis. Neste contexto, a costa do Rio Grande do Sul, devido as suas características morfodinâmicas intrínse- cas, tais como o alto grau de exposição à dinâmica oceânica e baixa declividade, coloca os vários ecossistemas a ela associados, como praia, dunas, marismas, campos, banhados e sangradouros, numa posição mais vulnerável às variações do nível médio do mar. Visando a um planejamento mais adequado destas áreas no cenário de mudanças climáticas globais, e a fim de determinar uma possível perda de ambientes, no caso de uma elevação do nível do mar em longo prazo (cenário 2100), os ambientes costeiros foram inicialmente classificados através da utilização do programa IDRISI, onde foram calculados os percentuais relativos de perda de cada um, para três possíveis cenários de linha de costa, projetados para 2100. Com relação aos resultados obtidos nos cálculos de perda ambiental, considerando as linhas de costa projetadas para 2100, percebemos que todos os ambientes teriam um percentual de área perdida, mesmo no melhor cenário considerado. Neste mesmo cenário onde a linha de costa projetada para 2100 encontra-se recuada 561m da posição atual e tem 70% de probabilidade de ser ultrapassada, observamos também, o maior percentual relativo de perda de área urbana. Para o cenário considerado como intermediário, onde a linha de costa projetada encontra-se recuada 695 m da atual, e tem 50% de probabilidade de ser ultrapassada, dunas representam a maior parte da área ambiental perdida, porém, dividem espaço com áreas de campos e em menor escala, áreas de matas. Já, para o pior cenário considerado, onde a linha de costa projetada encontra-se recuada 1032m da atual e com 10% de probabilidade de ser ultrapassada, houve uma inversão na perda ambiental com relação ao cenário anterior. Neste cenário, campos litorâneos e dunas aproximaram-se de 50 % e 35 % de perda ambiental, respectivamente, representando as maiores áreas impactadas. A vulnerabilidade costeira também foi avaliada através do cálculo do Índice de Vulnerabilidade Costeira (IVC), para os cenários atual e futuro (projetado para 2100). Com relação à vulnerabilidade costeira, de forma geral, os valores do IVC calculados para a região foram de 14,4 (atual) e 51,03 (futuro), o que faria com que a mesma deixasse de ser caracterizada como uma região moderadamente vulnerável, e passasse a ser considerada como uma região costeira de vulnerabilidade muito alta à elevação do nível do mar. Dessa forma, as informações aqui obtidas são úteis para planejar adequadamente a ocu- pação, o uso e a adaptação desta região costeira, bem como, possibilitar estudos futuros acerca do manejo dos ambientes nela encontrados, face ao atual cenário de mudanças climáticas globais. Palavras-chave: Mudanças Climáticas Globais, Índice de Vulnerabilidade Costeira (IVC), Rio Grande do Sul.

@ Corresponding author, to whom correspondence should be addressed. a Nav Oceanografia Ambiental Ltda. Rua Gildásio Amado,55/ 50, CEP 22631-020, Barra da Tijuca, Rio de Janeiro, RJ, Brasil. e-mail: b Universidade Federal do Rio Grande (FURG), Laboratório de Oceanografia Geológica, Caixa Postal 474, CEP: 96201-900, Rio Grande, RS, Brasil. e-mails: Figueiredo ; Calliari ; Tagliani

* Submission: 29 JUL 2014; Peer review: 3 SEP 2014; Revised: 19 SEP 2014; Accepted: 12 FEB 2015; Available on-line: 16 FEB 2015

Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):121-131 (2015)

Abstract In the current scenario of global climate change, in which some of the observed effects are: the mean sea level rise, and an increased frequency and intensity of extreme events, coastal regions around the world should in general become more vulnerable. In this context, the Rio Grande do Sul coastline, due to its intrinsic morphodynamic characteristics, such as, a low lying coastal gradient combined with a high degree of exposure to ocean dynamics, places several associated ecosystems, such as beach, dunes, marshes, fields, wetlands and washouts in a vulnerable position to changes in sea level. Model simulations of coastal response to sea level rise for the Rio Grande do Sul pointed out Cassino beach sector as being the most affected with the largest shoreline recession along the entire coast. Aiming for adequate planning of these areas under global climate change scenarios, and in order to determine the possible loss of habitats in case of a rise of mean sea level in the long term (2100), present day shoreline was digitalized at the foredune toe limit using in ArcGis, and coastal environments were initially classified using IDRISI program. In this environment, the percentages for each habitat loss were calculated for three possible coastline scenarios, designed for 2100. It was noticed that all environments would display some lost area, even at the best case scenario. In this scenario where the projected shoreline is located at 561m inland form the present day shoreline with 70% probability of being exceeded by 2100, more than half of the dune area would be lost, followed by a smaller fraction of fields. It can also be observed that in this scenario the urban area had the highest percentage loss. For the intermediate scenario where the projected shoreline is located at 695m inland from the present day shoreline with 50 % probability of being exceeded, dunes represent the largest loss in habitat, sharing space, with surrounding green fields and to a lesser extent, to forests. Yet for the worst case scenario where the projected shoreline is located at 1032m inland from the present day shoreline with10 % probability of being exceeded, there was a reversal in habitat loss with respect to the intermediate scenario. In this scenario, green fields and coastal dunes approached 50 % and 35 % loss in habitat, respectively, representing the largest impacted areas. It should be highlighted that the fresh water creeks, called washouts, in the best case scenario had its entire area lost. However, in the long term they probably should re-establish its course if topography allows. On the one hand, considering any of the simulated scenarios, as the shoreline retreats the salt marsh area would become closer to the foredune areas. On the other hand, even though not simulated here, there is also the possibility that with sea-level rise it could be inun- dated through the connecting channel to Patos Lagoon resulting in an increased habitat area. Coastal vulnerability was evaluated through the use of Coastal Vulnerability Index (CVI) for the present and future scenarios for 2100. Looking at coastal vulnerability in general, calculated CVI values for the study area were 14.4 and 51.03, designed for the present and for the year 2100, respectively. The results indicate that a region currently characterized as a moderately vulnerable, should become highly vulnerable in the future due to the rising of mean sea level as projected. This change would take place basically influenced by two of the index variables: shoreline erosion/accretion rates; and relative sea-level change, considered here as the controlling variables for coastal vulnerability under climate change in this area. Thus, the information obtained is critical to properly plan, land use and adaptation of this coastal region under the current scenario for global climate change, as well as enable future studies on habitat management. Keywords: Global Climate Changes, Coastal Vulnerability Index (CVI), Rio Grande do Sul.

1. Introdução al até decadal, e a longo prazo, constataram que a intensidade de tempestades tropicais e extratropicais aumen- Atualmente vivemos em um período marcado por mu- tarão, o que implica em mais impactos na costa do que danças climáticas globais, o qual vem sendo estudado apenas aos atribuídos a elevação do nível médio do com especial interesse pela comunidade científica que mar, especialmente para sistemas costeiros tropicais e se dedica às ciências da natureza e da sociedade. Alte- de médias latitudes. Assim sendo, a região costeira do rações climáticas são alguns dos efeitos já visíveis da Estado do Rio Grande do Sul, por ser uma extensa pla- grande modificação que o planeta está sofrendo devido nície arenosa e de baixo gradiente topográfico, ficará ao fenômeno do aquecimento global (Andrade & La- ainda mais vulnerável a eventos erosivos com o aumen- cerda, 2009; Bindoff et al., 2007). Outros efeitos obser- to do nível médio do mar, quando comparada a áreas vados, são a acentuação do efeito El Niño, catástrofes mais íngremes, especialmente durante eventos extre- de grandes proporções, desequilíbrio das chuvas e mos. grandes enchentes. Somente na década de 90, o enfoque nas possíveis alte- A zona costeira que se destaca pelos seus recursos natu- rações da linha de costa e seus efeitos na erosão costei- rais e diversidade ambiental, bem como pelo seu grande ra, devido a uma elevação do nível médio do mar, des- potencial para o desenvolvimento de atividades econô- pertaram maior atenção dos cientistas, da mídia e do micas múltiplas (Carter, 1990; Diehl et al., 2010) pode- público em geral. Dentre os principais indicadores de rá ter os efeitos das mudanças climáticas agravados pela tais elevações, destacam-se: o recuo da linha de costa, a intensa antropização dos espaços urbanos, trazendo erosão costeira, o desaparecimento de ilhas, a destrui- situações que caracterizam vulnerabilidades e adapta- ção de ecossistemas costeiros e áreas urbanizadas, den- ções para as populações e instalações (IPCC, 2001). tre outros. Assim, desde o avanço desses estudos, várias Resultados obtidos a partir de modelagem (Meehl et al., tentativas para prever este comportamento durante os 2007; Marengo, 2006), em escalas que vão de interanu- próximos séculos têm surgido. 122 Germani et al. (2015)

No que se refere à média das mudanças do nível médio do mar com valor mínimo de 0,18 m, modal de 0,89 m do mar feitas no Brasil, os dados sobre o comportamen- e máximo de 1,1 m para o ano de 2100. Estes são valo- to atual do nível médio do mar possuem disponibilidade res combinados de projeções de Rahmstorf (2007) e do limitada, o que, eleva a incerteza para os modelos de Ipcc (2007). Devido ao fato de que a incerteza na pró- resposta costeira para o futuro aumento do nível do pria modelagem de clima é inerente a todas as previsões mar. No Brasil, poucos locais apresentam registros sufi- e está intimamente relacionada o grau de contribuição cientes para permitir comparações climáticas (Mesqui- antropogênica e em relação à dinâmica do derretimento ta, 2003). Apenas no ano de 2004 houve a implantação das placas de gelo (IPCC, 2007), os resultados da modo Sistema Global de Observação do Nível do Mar delagem de resposta costeira sob condições de mudan- (GLOSS), coordenado pelo Centro de Hidrografia da ças climáticas é apresentado na forma de probabilidade Marinha do Brasil (CHM). O principal objetivo do pro- de serem excedidas, e.g., a linha de costa localizada a jeto é unir esforços de instituições brasileiras que de- 695 m da linha de costa atual tem 50% de probabilidade pendem das observações do nível do mar para seus de ser ultrapassada em 2100. monitoramentos e pesquisas. Neste trabalho, as projeções de linha de costa obtidas Neste trabalho, destacaremos os impactos causados pela por Figueiredo (2013) para o ano de 2100, são resultan- elevação do nível médio do mar, ocasionada pelas mu- tes da aplicação do modelo proposto por Cowell et al. danças climáticas, para a região costeira adjacente a (2006), onde os impactos relacionados aos efeitos das desembocadura da Lagoa dos Patos. Assim consideran- mudanças climáticas incluem aumento do nível médio do as características geológicas e morfodinâmicas, prin- do mar, alterações no clima de ondas e distúrbios no cipalmente o caráter oceânico aberto; a orientação Su- balanço de sedimentos ao longo da costa, podendo re- doeste-Nordeste da linha de costa, que a expõe às tem- sultar em erosão e recuo da mesma. O autor também pestades mais intensas provenientes do quadrante Sul- considera que, áreas mais vulneráveis à erosão costeira Sudeste; e o baixo gradiente topográfico da antepraia; a podem ser identificadas em termos de distância conti- costa gaúcha –atualmente inserida numa faixa de alto nente adentro a partir da atual linha de dunas, à medida risco a qualquer elevação do nível do mar, ficará nos que a linha de costa recua num determinado período de cenários de mudanças climáticas globais relatadas no tempo no futuro (estabelecido na modelagem). O recuo Ipcc (2007) e Church et al. (2013), ainda mais vulnerá- total em função do tempo leva em consideração os se- vel à erosão. guintes termos: tendência média de recuo devido ao Projeções de elevação do nível médio do mar apresen- desequilíbrio no suprimento de sedimentos litorâneos tadas no terceiro e quarto relatórios de avaliação no (no caso de balanço negativo), tendência média de re- Painel Intergovernamental sobre Mudanças Climáticas cuo devido à elevação acelerada do nível do mar, e (IPCC, 2001; IPCC, 2007) mostram taxas similares de aumento da erosão devido à ocorrência de tempestades aumento do nível do mar, no entanto, este último de- (ver Eq. 1). Assim, as previsões de recuo da linha de pende de um maior número de modelos climáticos de costa podem ser utilizadas para estimar o volume de crescente complexidade e realismo, resultando em ta- perda de areia e futuramente prever custos da realimen- xas mais confiáveis de aumento do nível do mar global tação artificial necessária para preservar praias e prote- que variam entre 0,18 m a 0,59 m em 2100. Com base ger os recursos imediatamente interiores a elas. nos resultados das projeções de elevação acelerada do nível médio do mar, foram desenvolvidos modelos (1) capazes de prever a resposta costeira em diferentes situações de comportamento do nível médio do mar. onde, é o recuo total, é o recuo médio devi- Historicamente, a Praia do Cassino vem sofrendo acre- do ao desequilíbrio do suprimento sedimentar, é o recuo médio devido à elevação acelerada do nível do ção (Lélis & Calliari, 2004) devido, em parte, à cons- trução dos molhes da Barra do Rio Grande, a qual inter- mar e é a alteração do balanço sedimentar devi- feriu no balanço sedimentar da área ao obstruir o trans- do ao aumento da intensidade das tempestades. porte de sedimentos para Nordeste, acarretando em um Também, a fim de prever os impactos da elevação do aumento localizado na largura da praia na região do nível do mar, uma metodologia para identificar as áreas Cassino. Entretanto, Figueiredo (2013), ao simular ce- que podem ser mais vulneráveis a essa elevação no nários futuros de resposta costeira a elevação do nível futuro foi desenvolvida (Gornitz, 1991). Esta técnica do mar para a costa no Rio Grande do Sul, utilizando o será descrita com detalhes na subseção 3.2. STM (Shoreline Translation Model) de Cowell et al. Transgressões e regressões marinhas ocorreram anteri- (1992), constatou os maiores valores de recuo da linha ormente durante o Pleistoceno e Holoceno moldando a de costa para o setor que inclui a Praia do Cassino al- costa como a conhecemos hoje. Estimativas indicam cançando distâncias médias de recuo de até 695m. A que durante o Holoceno no Rio Grande do Sul, o nível autora utiliza uma faixa de valores de elevação do nível médio do mar atingiu altitudes máximas entre 2 e 4 m

123 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):121-131 (2015) acima do nível atual (Dillenburg et al., 2000; Dillen- nantes: de leste-nordeste e de sudeste (Martinho et al., burg et al., 2005). Entretanto, os impactos de uma su- 2009; Pianca et al., 2010). bida acelerada do nível do mar na presente configura- A ação morfogênica na área estudada, relacionada com ção costeira, ainda são em grande parte desconhecidos. a atividade das correntes, é controlada basicamente pelas correntes litorâneas induzidas pelas ondas e ven- 2. Área de estudo tos e, muito secundariamente, por correntes fluviais. Do A costa do Rio Grande do Sul (RS), de orientação su- mesmo modo, Fontoura (2013) descreve que o transpor- doeste-nordeste (Fig. 1), possui características oceâni- te longitudinal na região é bidirecional, apresentando cas, devido à ação de ondas, correntes litorâneas e o resultante positiva para nordeste. vento, que são os principais fatores responsáveis por alterar a sua morfologia. A plataforma continental adja- 3. Material e métodos cente é larga, com 150 a 200 km de extensão, apresen- 3.1. Análises de Perda Ambiental tando profundidades máximas variando entre 100 e 140 m, e declividade suave da ordem de 0,5 a 1,5 m/km O mapeamento das unidades ambientais presentes na (Dillenburg et al., 2000). Ao longo dos 640km do lito- área de estudo foi realizado em meio digital, através da ral do estado observam-se poucas descargas flúvio- técnica de classificação supervisionada (Maxlike) dis- lagunares importantes (Figueiredo & Calliari, 2005), ponível no SIG Idrisi, utilizando imagens do satélite porém, inúmeros cursos de água doce, conhecidos como ALOS de 2010, com resolução espacial de 10 metros. A sangradouros, estão presentes. linha de costa atual foi digitalizada no SIG ArcGis se- guindo a base das dunas frontais. As seguintes unidades foram identificadas e mapeadas: marismas, dunas, sangradouros, banhados, campos litorâneos, matas e áreas com a infraestrutura urbana. Figueiredo (2013) estimou através de modelagem o recuo da linha de costa na área de estudo para o ano de 2100 em condições de mudanças climáticas. Suas pro- jeções de linha de costa foram apresentadas na forma de distribuição probabilidade, onde cada valor projetado para futura posição da linha de costa em metros está associado a uma probabilidade (de 0% a 100%) de ser ultrapassado num dado horizonte temporal, neste caso o ano 2100. A partir destas projeções, foram selecionadas para este estudo três linhas de costa projetadas com as respectivas probabilidades de serem ultrapassadas em 2100: linha azul, com 70% de probabilidade de ser ultrapassada (1:1,44); linha laranja, 50% de probabilidade de ser ultrapassada (1:2); linha verde, 10% de probabilidade (1:10) (Fig. 2). Para cada situação/posição da linha de costa projetada foram calculadas as perdas ambientais (áreas e percentuais relativos) utilizando os métodos overlay e area disponíveis no programa Idrisi. A Praia do Cassino, no município de Rio Grande, RS, encontra-se ao sul da desembocadura da Laguna dos Patos, a partir da qual se desenvolve uma extensa restinga (barreira arenosa) com 220 km de extensão, limitada ao sul pelo Arroio Chuí. A área de estudo compre- Figura 1 - Localização da área de estudo - Região a partir dos ende um embaiamento sutil de aproximadamente 60km Molhes da Barra de Rio Grande até o Farol do Sarita. de extensão (zona UTM 22S, entre 396383/6442021 e Figure 1 - Location of the study area - Region from Rio 364937/6382475), desde a raiz do Molhe Oeste até a Grande jetties to Sarita Lighthouse. região do Farol do Sarita (Fig. 1). A região possui ca- A região é incluída no regime de micromarés, onde as racterísticas semelhantes, representando uma unidade oscilações têm amplitude máxima de 0,5 m. As maiores geomorfológica da barreira arenosa holocênica (Toma- alterações do nível médio do mar percebidas nesse local zelli & Villwock, 2005) apresentando sistemas praiais devem-se às chamadas marés meteorológicas ou de dissipativos, campos de dunas, marismas e sangradou- tempestades. Já as ondulações que atingem a região ros associados, e suas praias são compostas de areia costeira da área de estudo têm duas direções predomi- quartzosa de granulometria fina. 124 Germani et al. (2015)

3.2. Índice de Vulnerabilidade Costeira à Elevação do Nível do Mar Para descrever a vulnerabilidade a uma mudança física como o aumento do nível médio do mar na costa em estudo, foi utilizada uma metodologia desenvolvida por Gornitz (1991). Esta técnica utiliza diferentes intervalos de vulnerabilidade (muito baixa a muito alta), asso- ciando a cada um deles um valor (Tabela 1). Valendo-se de seis variáveis (físicas e geológicas), e conhecendo seus valores referentes à área estudada, os valores dos intervalos de vulnerabilidade corresponden- tes a cada parâmetro são incorporados a uma equação (ver Eq. 2). Esta produz um índice de vulnerabilidade costeira (IVC), onde: valores de IVC abaixo de 8,7 são incluídos na categoria de baixa vulnerabilidade; valores entre 8,7 - 15,6 são considerados de vulnerabilidade moderada; valores entre 15,6 e 20 indicam alta vulnerabilidade; valores de IVC acima de 20 são classificados como de vulnerabilidade muito alta. A partir dessa téc- nica, foi calculado o índice para a costa atual, bem como o projetado para o ano de 2100, e comparados os valores obtidos.

Figura 2 - Contornos representando as possíveis futuras linhas de costa para a Praia do Cassino, com suas res- (2) pectivas probabilidades de serem ultrapassadas, ou diferentes níveis de risco, para 2100. Adaptado de Figueiredo (2013). Onde: A- geomorfologia; B - taxas de erosão/acreção da -1 Figure 2 - Contours representing possible future shorelines linha de costa (m.ano ); C - declividade da costa (%); -1 to Cassino’s Beach, with their respective probabilities of D - mudanças relativas do nível do mar (mm.ano ); E - exceedence or different risk levels by 2100. Adapted from altura significativa de onda (m); e F - amplitude de Figueiredo (2013). maré (m).

Tabela 1 - Valores das variáveis para o cálculo do Índice de Vulnerabilidade para a Costa Atlântica. Fonte: Adaptada de Gor- nitz (1991). Table 1 - Variables values for calculating the Vulnerability Index for the Atlantic Coast. Source: Adapted from Gornitz (1991).

Variáveis Muito Baixa Baixa Moderada Alta Muito Alta 1 2 3 4 5 Geomorfologia Costas de falésia Falésias médias, Falésia baixa, Praias cas- Praias arenosas, rochosa, fiord costas recortadas costas de deriva calhosas, pântanos salgados, glacial, planície estuários, lagoas planos lamosos, aluvial deltas, mangues, recifes de coral Taxas de erosão / >2,0 1,0 – 2,0 -1,0 – 1,0 -2,0 - -1,0 <2,0 acreção (m/a) Declividade da >1,20 1,20 – 0,90 0,90 – 0,60 0,60 – 0,30 <0,30 costa (%) Variação relativa <1,8 1,8 – 2,5 2,5 – 3,0 3,0 – 3,4 >3,4 do nível médio do mar (mm/a) Altura signi- <0,55 0,55 – 0,85 0,85 – 1,05 1,05 - 1,25 >1,25 ficativa da onda (m) Amplitude da >6,0 4,0 – 6,0 2,0 – 4,0 1,0 – 2,0 <1,0 maré (m)

125 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):121-131 (2015)

4. Resultados e discussão de área ambiental perdida. Porém, dividindo espaço com áreas de campos e, em menor escala, de matas. 4.1. Perda ambiental para 2100 Tabela 3 - Área perdida e respectivo percentual relativo para Nas Tabelas 2, 3 e 4 são apresentados os resultados cada ambiente, considerando a possível linha de costa dos cálculos de perda ambiental devido à elevação do projetada para 2100, com 50% de probabilidade de ser ul- nível do mar, em km² e % relativo. Foi considerada trapassada. como máxima área perdida, apenas a contida entre a Table 3 - Lost area and its respective relative percentage for linha da base das dunas frontais e a linha de costa mo- each environment, considering the possible shoreline pro- delada com 10% de probabilidade de ser ultrapassada jected for 2100, with 50% probability of being exceeded. (recuada 1032m da atual linha de costa). Entretanto, para fins de classificação inicial dos ambientes, foi in- Ambientes Área Perdida (km²) % relativo cluída toda a área da barreira arenosa holocênica (Fig. Dunas 21,31 46,93 3). Para um melhor entendimento, os resultados obtidos foram separados em três cenários. Matas 3,07 6,76 1) Melhor cenário Banhados 1,45 3,19 Este cenário utiliza a linha de costa modelada que recu- Campos 16,6 36,56 ou 561m da posição da atual linha de costa e tem 70% Marismas 0,09 0,20 de chance de ser ultrapassada até o ano 2100. Nesta Sangradouros 0,12 0,26 situação, observa-se que mais da metade da área perdida é representada por dunas, seguida de uma fração Área Urbana 2,77 6,10 menor, porém significativa, de campos (Tabela 2). TOTAL 45,41 100 Comparando os três cenários (Tabelas 2, 3 e 4), percebe-se com este cenário o maior percentual de perda de área urbana, além da totalidade da área de sangradou- 3) Pior cenário ros, já que estes se limitam à região mais próxima à Este cenário utiliza a linha de costa modelada que costa. recuou 1032m da posição da atual linha de costa e tem 10% de chance de ser ultrapassada até o ano 2100. Este Tabela 2 - Área perdida e respectivo percentual relativo para cada ambiente, considerando a possível linha de costa cenário mostra uma inversão no percentual dos ambien- projetada para 2100, com 70% de probabilidade de ser ul- tes perdidos (Tab. 4). Pode-se observar que a área de trapassada. campos foi a mais afetada com a máxima incursão ma- Table 2 - Lost area and its respective relative percentage for rinha considerada, chegando próximo a 50% de perda. each environment, considering the possible shoreline pro- Com aproximadamente 35% de área perdida estão as jected for 2100, with 70% probability of being exceeded. dunas, e apresentando pouco mais de 10%, as matas. Ambientes Área Perdida (km²) % relativo Tabela 4 - Área perdida e respectivo percentual relativo para cada ambiente, considerando a possível linha de costa Dunas 20,71 55,40 projetada para 2100, com 10% de probabilidade de ser ultrapassada. Matas 1,39 3,72 Table 4 - Lost area and its respective relative percentage for Banhados 1,04 2,78 each environment, considering the possible shoreline projected for 2100, with 10% probability of being exceeded. Campos 11,70 31,3 Marismas 0,08 0,21 Ambientes Área Perdida (km²) % relativo Sangradouros 0,12 0,32 Dunas 23,61 34,21 Área Urbana 2,34 6,26 Matas 7,08 10,26

TOTAL 37,38 100 Banhados 2,25 3,26 Campos 31,94 46,28 2) Cenário intermediário Marismas 0,14 0,20 Este cenário utiliza a linha de costa modelada que Sangradouros 0,12 0,17 recuou 695m da posição da atual linha de costa e tem 50% de chance de ser ultrapassada até o ano 2100. Na Área Urbana 3,87 5,61 Tabela 3 observa-se que, do mesmo modo que no con- TOTAL 69,01 100 texto anterior, as dunas ainda representam a maior parte

126 Germani et al. (2015)

Figura 3 - Classificação dos ambientes da área de estudo e representação das possíveis futuras linhas de costa. Figure 3 - Environmental classification of the study area and representation of possible future shorelines.

Os resultados aqui apresentados assumem um cenário impactos sobre as diversas comunidades da fauna e de subida acelerada de nível do mar numa região cos- flora da zona costeira. teira de gradiente muito suave, com declividade da No caso específico dos sangradouros, mesmo no melhor antepraia em torno de 0,08° e dunas que não ultra- cenário considerado havendo perda total de área, foi passam 3,5 m. concluído que, em longo prazo eles restabeleceriam sua Entretanto, devemos ter em mente que a reorganização drenagem em direção ao mar. Isso pode ser comprova- morfodinâmica em larga escala temporal, de dezenas a do através de observações do interior da barreira areno- centenas de anos se dá de forma lenta e que não sa holocênica, onde ficaram evidentes interrupções nos ocorrerá uma inundação abrupta dos ambientes pela cordões litorâneos que dão origem aos canais. Este fato subida do nível do mar em longo prazo. Esta reorga- também foi observado por Hesp et al. (2007) próximo à nização dos ambientes está relacionada com a trans- Praia de Curumim, no litoral norte do RS, onde inúme- ladação da linha de costa em função de trocas sedi- ros paleo-sangradouros corriam através da barreira ho- mentares transversais entre a região emersa da barreira, locênica em direção ao mar, estando frequentemente incluindo dunas frontais e áreas posteriores e o limite ligados aos canais atualmente ativos. Estes paleo- inferior da antepraia como resultado da subida do nível sangradouros provavelmente foram formados durante o médio do mar (Bruun, 1988), além de trocas longi- Holoceno, quando o nível médio do mar na região caía tudinais que influenciam o balanço sedimentar costeiro progressivamente, após ter atingido um máximo de no setor como um todo. elevação por volta de 5100 anos atrás (cerca de 1 - 3m Por outro lado, mesmo que ambientes como o pós-praia acima do atual), atuando como um sistema organizado e campos de dunas migrem gradativamente continente de drenagem (Martin et al., 1979). adentro, regiões mais interiorizadas e estáveis, como as Com relação à única ocorrência de marisma na área de marismas, campos litorâneos, banhados e até áreas estudo, restrita às adjacências do Molhe Oeste, obser- urbanizadas, ficarão de qualquer forma mais próximas vou-se que em todos os cenários possíveis há alguma da linha de costa e assim mais vulneráveis aos impactos perda de área deste ambiente. Embora não represente de eventos extremos. Adicionalmente, teremos grande significância, em termos de percentual perdido, 127 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):121-131 (2015) a marisma é um dos ambientes de maior relevância 4.2. Vulnerabilidade Costeira à Elevação do Nível do ecológica e sensibilidade ambiental da área, com rela- Mar: Cenários Atual e Futuro ção à fauna e flora. Por ser uma região estável, em ter- Como forma de quantificar a vulnerabilidade da Praia mos morfodinâmicos, diferentemente da região de du- do Cassino à elevação acelerada do nível do mar, foram nas, a princípio, este ambiente não se deslocaria conti- inseridos os valores da região, relativos a cada uma das nente adentro devido à uma elevação acelerada do nível variáveis, físicas e geológicas, na tabela utilizada para o do mar. Assim, concluímos que, as marismas ficarão, cálculo do Índice de Vulnerabilidade para a Costa A- em qualquer um dos três possíveis cenários projetados, tlântica de Gornitz (1991). Foi obtido um valor de IVC localizadas mais próximas dos cordões de dunas do que de 14,4 (ver Equação 3), classificando a praia como atualmente. Dessa forma, tornando-se mais vulneráveis moderadamente vulnerável a uma elevação do nível do aos impactos da transgressão marinha. Conforme Huis- mar na atualidade, conforme mostrado na Tabela 5. kes (1990), um pequeno aumento do nível médio do mar seria suficiente para acarretar mudanças na zonação Tabela 5 - Valores dos parâmetros físicos e geológicos para classificação de vulnerabilidade ao aumento do nível do desse ecossistema, até sua total eliminação. mar para a região de estudo. Taxas de erosão/acreção da Por outro lado, foi considerado que, por haver conexão linha de costa segundo Lélis & Calliari (2004); declivida- entre o estuário da Lagoa dos Patos e a região de made da costa segundo Dillenburg et al. (2000); mudanças rismas em questão, existe a possibilidade de que haja relativas do nível do mar segundo Church & White um aumento de área deste ambiente. Com a subida do (2006); altura significativa de onda segundo Pianca et nível do mar, o nível da lagoa também deve subir, in- al.(2010); amplitude de maré segundo Calliari & Klein fluenciando no acréscimo da área de marismas. (1993). Segundo Costa (1998), existem cerca de 70 espécies de Table 5 - Values of physical and geological parameters for classification of vulnerability to rising sea level to study plantas que habitam a região de marismas no estuário area. Rates of erosion / accretion of shoreline according inferior. Estas possuem características muito específi- to Lelis & Calliari (2004). Slope of the coast from Dillen- cas, com diferentes gradientes de tolerância à inunda- burg et al. (2000), relative changes in sea level following ção, dessecação e salinidade. Assim, em qualquer uma Church & White (2006), significant wave height accor- das duas situações, seja de perda ou acréscimo de área, ding to Pianca et al.(2010); tidal range from Calliari & elas sofreriam com o estresse às mudanças da dinâmica Klein(1993). ambiental. Dessa forma, poderiam facilmente desapare- Geomorfologia Valores na Vulnera- cer, caso não forem capazes de se adaptar a essas alte- Parâmetros classi- bilidade rações ocasionadas pela elevação do nível médio do ficação mar. Taxas de erosão / Praias are- Muito Por fim, o último ambiente a ser considerado neste es- 5 acreção (m/a) nosas, Alta tudo, também sujeito aos impactos da subida do nível marismas do mar - o Urbano - é o que oferece maior preocupação Declividade da Muito à população que vive na região costeira. Por afetar dire- costa (%) +3,2 1 tamente a segurança da vida humana e seus bens mate- Baixa riais, o risco de uma inundação provocando perdas ou Variação relativa Muito 0,025 5 danos as suas propriedades mantém as pessoas em estado nível médio Alta do de alerta. Embora a diferença seja pouco expressiva do mar (mm/a) Altura signi- em nossa área de estudo, o percentual representado pela ficativa da onda 2 2 Baixa área urbana é maior quanto mais próximo à costa. Isto (m) representa uma tendência nas regiões costeiras de todo Amplitude da 1 - 2 5 Muito o mundo, tornando os riscos de uma elevação do nível maré (m) Alta Geomorfologia Muito médio do mar ainda mais preocupantes. Foi considera- 0,5 5 do ainda, que a elevação média do nível do mar, quando Alta associada com tempestades costeiras e ressacas (marés meteorológicas), potencializará ainda mais os impactos previstos para essa área. (3) Assim sendo, para os futuros cenários projetados, cer- tamente serão necessárias estratégias de adaptação, que Ao observar a Tabela 5, nota-se que, se consideradas podem incluir obras de proteção costeira para conter o constantes as variáveis: geomorfologia, declividade da avanço marinho, e realocação, mas principalmente um costa, altura significativa de onda e amplitude de maré; melhor planejamento da ocupação em zonas costeiras e utilizando os parâmetros das duas variáveis restantes, vulneráveis aos impactos diretos das mudanças climáti- para o cenário projetado (2100), a região de estudo seria cas globais. classificada como muito altamente vulnerável a uma

128 Germani et al. (2015) elevação do nível do mar, com um IVC de 51,03 (ver Com relação às Taxas de Erosão/Acreção da Linha de Equação 4). Costa, estudos de Dillenburg et al. (2000); Martinho et Segundo Toldo Jr. et al. (2005), a retração e prograda- al. (2009) concluíram que, o volume de sedimentos ção das zonas costeiras arenosas submetidas ao regime transportados ao longo da costa, bem como a energia de de micromarés, caso da região em estudo, resultam ondas, tende a aumentar conforme a costa torna-se con- principalmente da dinâmica entre a quantidade e tipo de vexa (projeções) e diminuir conforme a costa torna-se suprimento sedimentar, energia física de ondas e mu- côncava (embaiamentos). A causa desse padrão está danças relativas do nível médio do mar. provavelmente relacionada aos processos de refração e dissipação, que as ondas de águas profundas encontram Neste sentido, a Tabela 6 apresenta os valores dos pa- ao se aproximarem da costa. Quando a costa é côncava, râmetros físicos e geológicos para classificação de vul- a plataforma interna é larga e suave, esses processos nerabilidade ao aumento do nível médio do mar, proje- ocorrem lentamente, e ao atingirem a costa, as ondas já tados para 2100 por Figueiredo (2013), enquanto a E- perderam grande parte da sua energia. Assim, a depo- quação 4 resulta no IVC calculado para o mesmo cená- sição é favorecida. Por outro lado, quando a costa é rio futuro. convexa, a plataforma interna é estreita e íngreme, os Tabela 6 - Valores dos parâmetros físicos e geológicos para processos de refração e dissipação ocorrem menos, e as classificação de vulnerabilidade ao aumento do nível do ondas atingem a costa com alta energia. Portanto, o mar para a região de estudo, para um cenário futuro (2100). Taxas de erosão/acreção da linha de costa, e volume de sedimentos retirado destes locais é maior. mudanças relativas do nível do mar, adaptadas de Somada à dinâmica natural, historicamente, a linha de Figueiredo (2013); declividade da costa segundo Dil- costa adjacente a Praia do Cassino, vem sofrendo uma lenburg et al.(2000); altura significativa de onda segundo acreção relevante. Parte desta acreção se deve ao Pianca et al.(2010); amplitude de maré segundo Calliari & período que sucede a construção dos molhes da Barra Klein (1993). do Rio Grande, a qual interferiu no balanço sedimentar Table 6 - Physical and geological parameters for classificada área, ao bloquear o transporte de sedimentos para tion of coastal vulnerability to sea level rise, for a future nordeste, intensificando a acreção e acarretando um scenario (2100). Rates of shoreline erosion/accretion and aumento localizado na largura da praia. Dessa forma, relative changes in sea level, modified from Figueiredo com uma acreção da linha de costa atual de aproxi- (2013); coastal slope from Dillenburg et al. (2000), significant wave height according to Pianca et al.(2010); madamente 3,2m/ano (Lélis & Calliari, 2004), a região tidal range from Calliari & Klein (1993). está inserida num setor de barreira progradante. Porém, considerando o cenário de elevação acelerada do nível Variáveis Parâmetros Valores Vulnera- do mar prevista para os próximos anos, esta poderá se na bilidade tornar retrogradante, com taxas de erosão da linha de classi- costa estimadas em 11,47 m/ano no pior cenário pro- ficação jetado (com 10% de probabilidade) para 2100 Geomorfologia Praias are- 5 Muito Alta (Figueiredo, 2013). nosas, Com relação às Mudanças Relativas do Nível do Mar, a marismas região de estudo possui uma taxa atual de 2mm/ano de Taxas de -11,47 1 Muito Alta mudança relativa do nível do mar. Observa-se que o Erosão/Acreção nível atual sobe lentamente, colocando a região costeira da Linha de em um cenário de baixa vulnerabilidade. No entanto, Costa (m/ano) quando consideradas as taxas para 2100 (1,1m/90 anos), Declividade da 0,025 5 Muito Alta percebemos que a vulnerabilidade é considerada muito Costa (%) alta, quando o nível do mar atinge taxas de aproxi- Mudanças 12 2 Muito Alta madamente 12 mm/ano. Relativas do Através do exposto, percebe-se que, alterando os Nível do Mar valores de somente duas variáveis, a Praia do Cassino (mm/ano) deixaria de ser caracterizada como uma região modera- Altura Signi- 1 - 2 5 Muito Alta damente vulnerável, atualmente, e passaria a ser consi- ficativa da derada como uma praia de vulnerabilidade muito alta às Onda (m) elevações do nível do mar, apresentando um IVC futuro Amplitude de 0,5 5 Muito Alta de 51,03. Maré (m) 5. Conclusões A região costeira do Estado do Rio Grande do Sul, na (4) qual a Praia do Cassino está inserida, possui caracterís-

129 Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management, 15(1):121-131 (2015) ticas geomorfológicas herdadas que, de forma geral, a Bindoff, N.L.; Willebrand, J.; Artale, V.; Cazenave, A.; Gregory, J.; torna bastante vulnerável às variações do nível médio Gulev, S.; Hanawa, K.; Le Quéré, C.; Levitus, S.; Nojiri, Y.; Shum, C.K.; Talley, L.D.; Unnikrishnan, A. (2007) - Observa- do mar. Considerando o atual cenário de mudanças tions: Oceanic Climate Change and Sea Level. In: Solomon, S.; climáticas globais, é de extrema importância o estudo Qin, D.; Manning, M.; Chen, Z.; Marquis, M.; Averyt, K.B.; detalhado da área citada. Este se deu no presente estudo Tignor, M.; Miller, H.L. (org.), Climate Change 2007: The de forma qualitativa e quantitativa, de modo a determi- Physical Science Basis. Contribution of Working Group I to the nar a vulnerabilidade costeira da região em questão, Fourth Assessment Report of the Intergovernmental Panel on Climate Change, pp.385-432, Cambridge University Press, frente ao aspecto acima mencionado, com base nos Cambridge, U.K. / New York, NY, U.S.A. ISBN: 978- resultados obtidos por Figueiredo (2013). 0521705967. Disponível on-line em http://www.ipcc.ch/pdf/ assess- Neste contexto, como principais conclusões do trabalho ment-report/ar4/wg1/ar4-wg1-chapter5.pdf efetuado destacam-se as seguintes: Bruun, P. (1988) - The Bruun Rule of erosion by sea-level rise – a discussion on large-scale two-dimensional and 3-dimensional • Conforme os cálculos de perda de área, considerando usages. Journal of Coastal Research (ISSN: 0749-0208), as prováveis futuras linhas de costa projetadas para 4(4):627-648, Coastal Education & Research Foundation, Char- lottesville, VA, U.S.A. Article Stable URL: 2100, todos os ambientes da região de estudo apresen- http://www.jstor.org/stable/4297466 taram um percentual perdido de ambientes costeiros Carter, R.W.G. (1990) - Coastal Environments: An Introduction to emersos, mesmo no melhor cenário considerado; the Physical, Ecological and Cultural systems of coastlines. • Campos litorâneos e dunas representaram os maiores 617p., Academic Press, San Diego, CA, U.S.A. ISBN: 978- percentuais absolutos de perda ambiental quando con- 0121618568. siderado o pior cenário projetado (linha de costa mo- Calliari, L.J.; Klein, A.H.F. (1993) - Características morfo- delada com 10% de probabilidade de ser ultrapassada dinâmicas e sedimentológicas das praias oceânicas entre Cas- sino e Chuí, RS. Pesquisas (ISSN: 0100-5375), 20(1):48-56, em 2100); Instituto de Geociências, UFRGS, Porto Alegre, RS, Brasil. Dis- • Sangradouros tiveram a totalidade de sua área perdi- ponível on-line em http://www.praia.log.furg.br/Publicacoes da, mesmo no melhor cenário projetado (linha de cos- /1993/1993a.pdf ta modelada com 70% de probabilidade de ser ultra- Church, J.A.; White, N.J. (2006) - A 20th century acceleration in global sea-level rise. Geophysical Research Letters, 33(1):1-4. passada em 2100); DOI: 10.1029/2005GL024826 • Os valores obtidos nos cálculos de IVC indicam que, Church, J.A.; Clark, P.U.; Cazenave, A.; Gregory, J.M.; Jevrejeva, a Praia do Cassino é classificada atualmente como S.; Levermann, A.; Merrifield, M.A.; Milne, G.A.; Nerem, R.S.; moderadamente vulnerável às elevações do nível do Nunn, P.D.; Payne, A.J.; Pfeffer, W.T.; Stammer, D.; Un- mar, enquanto que nas situações projetadas para 2100 nikrishnan, A.S. (2013) - Sea Level Change. In: Stocker, T.F.; (Figueiredo, 2013), a área de estudo seria caracteriza- Qin, D.; Plattner, G.-K.; Tignor, M.; Allen, S.K.; Boschung, J.; Nauels, A.; Xia, Y.; Bex, V.; Midgley, P.M. da como muito altamente vulnerável à subida do nível (org.), Climate Change 2013: The Physical Science Basis. Con- do mar. tribution of Working Group I to the Fifth Assessment Report of Os dados obtidos na pesquisa são considerados extre- the Intergovernmental Panel on Climate Change, pp.1137- 1216. Cambridge University Press, Cambridge, United King- mamente úteis para a realização de adaptações na regi- dom, New York, NY, USA. ISBN: 978-1107057991. Disponível ão de estudo, como obras de contenção ao avanço mari- on-line em http://www.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5 nho, bem como de um manejo minucioso dos ambientes _Chapter13_FINAL.pdf nela encontrados, no cenário de mudanças climáticas Costa, C.S.B (1998) - Plantas de marismas e terras alagáveis. In: globais. Assim, será possível planejar adequadamente o Seeliger, U.; Odebrecht, C.; Castello, J.P. (org.), Os Ecos- sistemas Costeiro e Marinho do Extremo Sul do Brasil, pp.25- uso, a ocupação e a adaptação desta região costeira, 29, Editora Ecoscientia, Rio Grande, RS, Brasil. ISBN: 85- valendo-se, sobretudo, das individualidades de cada 87167014. Disponível on-line em http://www2.furg.br/instituto setor da costa do Rio Grande do Sul. /io/ecoveco/ecomidia/livros/Os Ecossistemas Costeiro e Marinho do Extremo Sul do Brasil.pdf Agradecimentos Cowell, P.J.; Roy, P.S.; Jones, R.A. (1992) - Shoreface translation model: computer simulation of coastal-sand-body response to Agradecemos aos revisores deste artigo; à Coordenação de sea level rise. Mathematics and Computers in Simulation, 33(5- Aperfeiçoamento de Pessoal de Nível Superior (CAPES); a 6):603-608. DOI: 10.1016/0378-4754(92)90158-D Agência Nacional do Petróleo (ANP), a qual subsidiou o traba- Cowell, P.J.; Thom, B.G.; Jones, R.A.; Everts,C.H.; Simanovic, D. lho original, base para este artigo; e a todas as pessoas e institu- (2006). Management of Uncertainty in Predicting Climate- ições que contribuíram para as experiências e aprendizado em change Impacts on Beaches. Journal of Coastal Research gestão costeira. 22(1):232-245. DOI: 10.2112/05A-0018.1 Diehl, F.P.; Spinola, J.L.; De Azevedo, N.T. (2010) - Mudanças References climáticas globais e os impactos físicos e biológicos na zona costeira: Uma proposta de classificação. In: Dantas, M.B.; Andrade, E.S.M.; Lacerda, G.B.M. (2009) - Mudanças Climáticas e Séguin, E.; Ahmed, F. (org.), O Direito Ambiental na atua- Análise de Risco da Indústria de Petróleo no Litoral Brasileiro. lidade, pp.127-134,Lumen Juris Ltda., Rio de Janeiro, RJ, 122p., Fundación Mapfre, Madrid, Espanha. ISBN: 978- Brasil. ISBN: 9788537507087. Disponível on-line em http://www. 8498441642. Disponível on-line em http://www.mapfre.com/ docu- ambito-juridico.com.br/site/index.php?n_link=revista_artigos_leitura& mentacion/publico/118n/catalogo_imagenes/grupo.cmd?path=1053958. artigo_id=9162 130 Germani et al. (2015)

Dillenburg, S.R.; Roy, P.S.; Cowell, P.J.; Tomazelli, L.J. (2000) - Lélis, R.J.F.; Calliari, L.J. (2004) - Historical shoreline changes near Influence of antecedent topography on coastal evolution as lagoonal and river stabilized inlets in Rio Grande do Sul State, tested by the shoreface translation-barrier model (STM). Journal southern Brazil. Journal of Coastal Research (ISSN: 0749- of Coastal Research (ISSN: 0749-0208), 16(1): 71-81, Coastal 0208), SI(39): 301-305, Coastal Education & Research Founda- Education & Research Foundation, Charlottesville, VA, U.S.A. tion, Charlottesville, VA, U.S.A. Disponível on-line em Article Stable URL: http://www.jstor.org/stable/4300012 http://www.cerf-jcr.org/images/stories/59_lelis.pdf Dillenburg, S.R.; Tomazelli, L.J.; Martins, L.R.; Barbosa, E.G. Marengo, J.A. (2006). Mudanças Climáticas Globais e Efeitos sobre (2005) - Modificações de longo período da linha de costa das a Biodiversidade - Caracterização do clima atual e definição barreiras costeiras do Rio Grande do Sul. Gravel (ISSN: 1678- das alterações climáticas para o território Brasileiro ao longo 5975), 3(1):9-14, Instituto de Geociências, UFRGS, Porto do Século XXI. Subprojeto do Programa de Conservação e Uti- Alegre, RS, Brasil. Disponível on-line em lização Sustentável da Diversidade Biológica Brasileira http://www.ufrgs.br/gravel/3/Gravel_3_02.pdf (PROBIO). 212 p., Instituto Brasileiro do Meio Ambiente e dos Figueiredo, S.A. (2013) - Modelling climate change effects in Recursos Naturais Renováveis, Brasília, DF, Brasil. ISBN: 85- southern Brazil. Journal of Coastal Research, SI (65):1933- 77380386. Disponível em http://www.mma.gov.br/estruturas/imprensa/ 1938. DOI: 10.2112/SI65-327.1 _arquivos/livro completo.pdf Figueiredo, S.A.; Calliari, L.J. (2005) - Sangradouros: Distribuição Martin L, Flexor J-M, Vilas Boas GS, Bittencourt ACSP, Guimaraes espacial, variação sazonal, padrões morfológicos e implicações MMM. (1979). Courbe de variations du niveau relative de la no gerenciamento costeiro. Gravel (ISSN: 1678-5975), 3(1):47- mer aucours des 7000 dernieres annees sur un secteur homogene 57, Instituto de Geociências, UFRGS, Porto Alegre, RS, Brasil. du littoral bresilien (nord de Salvador, Bahia). In: Sugio K.; Disponível on-line em http://www.ufrgs.br/gravel/3/Gravel_3_06.pdf Fairchild R.R.; Martin L.; Flexor J.M. (Eds.) Proceedings of the International Symposium on Coastal Evolution in the Quater- Figueiredo, S.A. (2011) - External forcing and internal controls on nary, pp.264 –274, Editora da Universidade de Sao Paulo, São coastal response to climate change at Rio Grande do Sul, south- Paulo, SP, Brasil. ISBN: 978-8531405297. Disponível on-line em ern Brazil. 164p., Tese de Doutorado, Universidade de Sidnei, http://horizon.documentation.ird.fr/exl-doc/pleins_textes/pleins_textes_6 Sidnei, Austrália. Disponível on-line em http://www.praia. /b_fdi_35-36/41921.pdf log.furg.br/Publicacoes/2011/2011b.pdf Martinho, C.T.; Dillenburg, S.R.; Hesp, P. (2009) - Wave energy Fontoura, J.A.S.; Almeida L. E.; Calliari, L. J.; Cavalcanti, A. M.; and longshore sediment transport gradientes controlling barrier Möller, O. Jr., Romeu, M.A.R. and Christófaro, B.R. (2013) - evolution in Rio Grande do Sul, Brazil. Journal of Coastal Re- Coastal Hydrodynamics and Longshore Transport of Sand on search, 25(2):285-293. DOI: 10.2112/06-0645.1 Cassino Beach and on Mar Grosso Beach, Southern Brazil. Journal of Coastal Research, (4):855-869. DOI: Meehl, G.A.; Stocker, T.F.; Collins, W.D.; Friedlingstein, P.; Gaye, 10.2112/JCOASTRES-D-11-00236.1 A.T.; Gregory, J.M.; Kitoh, A.; Knutti, R.; Murphy, J.M.; Noda, A.; Raper, S.C.B.; Watterson, I.G.; Weaver, A.J.; Zhao, Z.-C. Gornitz, V. (1991) – Global sea level hazards from future sea level (2007) - Global Climate Projections. In: Solomon, S.; Qin, D.; rise. Palaeogeography, Palaeoclimatology, Palaeoecology Manning, M.; Chen, Z.; Marquis, M.; Averyt, K.B.; Tignor, M.; (Global and Planetary Change Section), 89(1):379-398. Else- Miller, H.L. (org.), Climate Change 2007: The Physical Science vier Science Publishers B.V., Amsterdam, The Netherlands. Dis- Basis. Contribution of Working Group I to the Fourth Assess- ponível on-line em http://pubs.giss.nasa.gov/docs/1991/ ment Report of the Intergovernmental Panel on Climate 1991_Gornitz_1.pdf Change, pp. 749-844, Cambridge University Press, Cambridge, Hesp, P.A.; Dillenburg, S.R.; Barbosa, E.G.; Clerot, L.C.P.; U.K. / New York, NY, U.S.A. ISBN: 978-0521705967. Dis- Tomazelli, L.J.; Ayup-Zouain, R.N. (2007) - Morphology of the ponível on-line em http://www.ipcc.ch/pdf/assessment- Itapeva to Tramandai transgressive dune field barrier system and report/ar4/wg1/ar4-wg1-chapter10.pdf mid-to late Holocene sea level change. Earth Surface Processes Mesquita, A. R. (2003) - Sea level variations along the Brazilian and Landforms, 32(3):407-414. DOI: 10.1002/esp.1408 coast: a short review. Journal of Coastal Research, (ISSN: Huiskes, A.H.L. (1990) - Possible effects of sea level changes on 0749-0208), SI(35): 21-31, Coastal Education & Research salt-marsh vegetation. In: Beukema, J. J.; J. J. W. M. Brouns Foundation, Charlottesville, VA, U.S.A. Article Stable URL: (Eds.), Expected effects of climatic change on Marine Coastal http://www.jstor.org/stable/40928745 Ecosystems, pp.167-172. Springer Netherlands, Netherlands. Pianca, C.; Mazzini, P. L. F.; Siegle, E. (2010) - Brazilian offshore ISBN: 978-9400920033. Disponível on-line em 10.1007/978- wave climate based on NWW3 reanalysis. Brazilian Journal of 94-009-2003-3_20 Oceanography, 58(1):53-70. DOI: 10.1590/S1679- IPCC (2001) - Climate Change 2001: The scientific basis. (Contri- 87592010000100006 bution of Working Group I to the Third Assessment Report of Rahmstorf, S. (2007) - A semi-empirical approach to projecting the Intergovernmental Panel on Climate Change). [Houghton, future sea-level rise. Science, 315(5810):368-370. DOI: J.T., Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. 10.1126/science.1135456 Dai, K. Maskell, and C.A. Johnson (eds.)]. 873p., Cambridge University Press, Cambridge, United Kingdom / New York, NY, Toldo Jr., E. E.; Almeida, L.E.S.B.; Nicolodi, J.L.; Martins, L.R. (2005) - Retração e progradação da zona costeira do estado do U.S.A. ISBN: 0521014956. Disponível on-line em http://www.ipcc.ch/ipccreports/tar/wg1/pdf/WG1_TAR-FRONT.PDF Rio Grande do Sul. Gravel (ISSN: 1678-5975), 3(1): 31-38. Ins- tituto de Geociências, UFRGS, Porto Alegre, RS, Brasil. Dis- IPCC (2007) - Climate Change 2007: The physical science basis. ponível on-line em http://www.ufrgs.br/gravel/3/Gravel_3_04.pdf (Contribution of Working Group I to the Fourth Assessment Re- port of the Intergovernmental Panel on Climate Change). Tomazelli, L.J.; Villwock, J.A. (2005) – Mapeamento geológico de [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. planícies costeiras: o exemplo da costa do Rio Grande do Sul. Averyt, M.Tignor and H.L. Miller (eds.)]. 996 p., Cambridge Gravel (ISSN: 1678-5975),3(1):109-115. Instituto de Geo- University Press, Cambridge, United Kingdom and New York, ciências, UFRGS, Porto Alegre, RS, Brasil. Disponível on-line em http://www.ufrgs.br/gravel/3/Gravel_3_11.pdf NY, USA. Disponível on-line em http://www.ipcc.ch/pdf/assessment- report/ar4/wg1/ar4_wg1_full_report.pdf

131