An AHP-derived method for mapping the physical ... · G. Le Cozannet et al.: An AHP-derived method...

Nat Hazards Earth Syst Sci 13 1209ndash1227 2013wwwnat-hazards-earth-syst-scinet1312092013doi105194nhess-13-1209-2013copy Author(s) 2013 CC Attribution 30 License

EGU Journal Logos (RGB)

Advances in Geosciences

Open A

ccess

Natural Hazards and Earth System

SciencesO

pen Access

Annales Geophysicae

Open A

ccess

Nonlinear Processes in Geophysics

Open A

ccess

Atmospheric Chemistry

and Physics

Open A

ccess

Atmospheric Chemistry

and Physics

Open A

ccess

Discussions

Atmospheric Measurement

Techniques

Open A

ccess

Atmospheric Measurement

Techniques

Open A

ccess

Discussions

Biogeosciences

Open A

ccess

Open A

ccess

BiogeosciencesDiscussions

Climate of the Past

Open A

ccess

Open A

ccess

Climate of the Past

Discussions

Earth System Dynamics

Open A

ccess

Open A

ccess

Earth System Dynamics

Discussions

GeoscientificInstrumentation

Methods andData Systems

Open A

ccess

GeoscientificInstrumentation

Methods andData Systems

Open A

ccess

Discussions

GeoscientificModel Development

Open A

ccess

Open A

ccess

GeoscientificModel Development

Discussions

Hydrology and Earth System

Sciences

Open A

ccess

Hydrology and Earth System

Sciences

Open A

ccess

Discussions

Ocean Science

Open A

ccess

Open A

ccess

Ocean ScienceDiscussions

Solid Earth

Open A

ccess

Open A

ccess

Solid EarthDiscussions

The Cryosphere

Open A

ccess

Open A

ccess

The CryosphereDiscussions

Natural Hazards and Earth System

Sciences

Open A

ccess

Discussions

An AHP-derived method for mapping the physical vulnerability ofcoastal areas at regional scales

G Le Cozannet1 M Garcin1 T Bulteau1 C Mirgon1 M L Yates1 M M endez12 A Baills1 D Idier1 andC Oliveros1

1BRGM 3 Avenue Claude Guillemin Cedex 2 BP 6009 45060 Orleans France2University Diderot-Sorbonne Paris 75205 Paris Cedex 13 France now at Universite Paris-Est Laboratoire drsquoHydraulique Saint-Venant ENPC EDF RampD CETMEF 6 quai Watier BP 4978401 Chatou France now at University Rafael Landivar de Guatemala Institute of Agriculture Natural Resources and the EnvironmentIARNAURL Vista Hermosa 3 Guatemala

Correspondence toG Le Cozannet (glecozannetbrgmfr)

Received 15 October 2012 ndash Published in Nat Hazards Earth Syst Sci Discuss ndashRevised 18 March 2013 ndash Accepted 20 March 2013 ndash Published 16 May 2013

Abstract Assessing coastal vulnerability to climate changeat regional scales is now mandatory in France since the adop-tion of recent laws to support adaptation to climate changeHowever there is presently no commonly recognised methodto assess accurately how sea level rise will modify coastalprocesses in the coming decades Therefore many assess-ments of the physical component of coastal vulnerability arepresently based on a combined use of data (eg digital eleva-tion models historical shoreline and coastal geomorphologydatasets) simple models and expert opinion In this studywe assess the applicability and usefulness of a multi-criteriadecision-mapping method (the analytical hierarchy processAHP) to map physical coastal vulnerability to erosion andflooding in a structured way We apply the method in two re-gions of France the coastal zones of Languedoc-Roussillon(north-western Mediterranean France) and the island of LaReunion (south-western Indian Ocean) notably using the re-gional geological maps As expected the results show notonly the greater vulnerability of sand spits estuaries and low-lying areas near to coastal lagoons in both regions but alsothat of a thin strip of erodible cliffs exposed to waves in LaReunion Despite gaps in knowledge and data the method isfound to provide a flexible and transportable framework torepresent and aggregate existing knowledge and to supportlong-term coastal zone planning through the integration ofsuch studies into existing adaptation schemes

1 Introduction

The attraction of coastal zones has greatly increased in thelast hundred years As a result people and infrastructure havebecome more exposed to coastal hazards including marineerosion and submersion and the costs of protection have in-creased accordingly A European assessment under the Euro-sion project (2004) found that 32 billion euros were spent onstabilising European shorelines in 2001 Given this contextsea level rise is expected to aggravate a wide range of alreadyexisting coastal hazards (Nicholls and Cazenave 2010)

Mitigating and preventing coastal risks often relies on pro-tection measures For example in Languedoc-Roussillonhard defence structures such as groynes and breakwatershave been placed along the coast in order to reduce ero-sion and submersion risks However whether these protec-tive measures can be sustained in the long term remains anopen question The legislation in France therefore requiresregional authorities to develop adaptation strategies designedfor longer timescales than these protective measures whichare expected to be efficient for a few decades This requiresassessments of longer term coastal zone physical vulnerabil-ity to marine erosion and submersion over time frames thatincorporate the effects of climate change ldquoPhysical vulnera-bilityrdquo refers here to the ldquosensitivity of the physical environ-ment of coastal areas themselvesrdquo to erosion and submersionin the context of climate change (Romieu et al 2010) and

Published by Copernicus Publications on behalf of the European Geosciences Union

1210 G Le Cozannet et al An AHP-derived method for mapping the physical vulnerability of coastal areas

mapping it is far from straightforward (eg Preston et al2011)

In shallow coastal waters waves and currents can causesediment transport and bathymetry changes (ie sea floor to-pography) which in turn affect waves and currents (eg Cocoand Murray 2007) On time scales longer than several weeksall of these processes cannot be modelled with sufficient ac-curacy for long-term analysis Therefore longer term studiesattempt to quantify sediment transport processes in a sim-plified way either in the cross-shore andor longshore direc-tion (Hanson et al 2003) as gradients in longshore sedimenttransport may be the result in erosion or accretion It is worthnoting that these morphosedimentary dynamic processes arealso important for anticipating potential future marine sub-mersion because nearshore waves and sea level depend onoffshore atmospheric and marine conditions as well as ofthe bathymetry

However it is still possible to appraise long-term mor-phosedimentary dynamics by analysing a number of coastalfeatures including coastal geomorphology (eg Cooper andJay 2002) exposure to waves and currents local sedimen-tary budgets (potentially affected by human activities) andrelative sea level variations At regional scales Gornitz etal (1991) developed an approach in which data describingthese components of physical vulnerability assessments areaggregated The resulting coastal vulnerability index (CVI)originally took into account the following data historicalerosion rates nearshore slope tidal amplitude observationsand predictions of sea level variations wave climate andcoastal geomorphology (Gornitz et al 1991) Some adapta-tions of the CVI benefited from other data and put more em-phasis on for example coastal subsidence as a component ofrelative sea level rise (eg Doukakis 2005) Beyond qualita-tive assessments of coastal vulnerability (eg Paskoff 2004Lebbe et al 2008 Alpar 2009) the challenge in mappingcoastal vulnerability is to combine all of this heterogeneousinformation into one single map in a structured way and toappropriately integrate the opinion of experts when data areinsufficient to directly solve the problem (eg Fairbanks andJakeways 2006 Vinchon et al 2009 Hanson et al 2010)

In this study we describe a new method for mappingcoastal vulnerability that incorporates expert analysis andqualitative and quantitative data in a more structured wayThe method is based on a multi-criteria decision-mappingmethod (Malczewski 2006 Chakar 2006) ndash the analyti-cal hierarchy process (AHP Saaty 1980 part 2) We in-tegrate AHP into a complete coastal physical vulnerabilitymapping method and apply it in two regions of France theisland of La Reunion in the south-western Indian Ocean andLanguedoc-Roussillon in the Mediterranean (part 3) We pro-vide new maps of coastal vulnerability in the two regions asa result (part 4) We discuss the advantages limitations andtransportability of the proposed method and investigate itspotential to be integrated into emerging adaptation practicesin France (part 5)

Table 1The fundamental scale of absolute numbers (ldquoSaaty scalerdquoadapted from Saaty 2008a)

Intensity ofimportance

Definition

1 Equal importance of both components

3 Judgement slightly favours onecomponent over another (moderatedifference of importance)

5 Judgement strongly favours onecomponent over another (strongdifference of importance)

7 Very strong or demonstratedimportance of one componentwith respect to another

9 Evidence of extreme difference ofimportance of one componentwith respect to another

2 The AHP process

21 The AHP process

The analytical hierarchy process (Saaty 1980) is a multi-criteria decision analysis method that solves decision-makingproblems by ranking alternatives according to several crite-ria The main steps in the method involve (Saaty 2008a)

ndash representing a decision-making problem by organisingits criteria into a hierarchical structure

ndash evaluating the relative importance of the criteria (withina hierarchical tree) and then the alternatives with re-spect to each criterion this is done by constructing pair-wise comparison matrices of criteria and alternatives re-garding each criterion For example a simple decisionproblem with 3 criteria and 5 alternatives will requireone 3times 3 comparison matrix to be built for the crite-ria weighting and three 5times 5 comparison matrices forevaluating the alternative criteria and

ndash synthesising the analysis by calculating the weightedsum of scores for alternatives based on the weights oftheir parent nodes in the hierarchical tree

The pairwise comparisons of criteria are made using a scaleof absolute judgements commonly called Saatyrsquos scalewhich indicates by how much one item predominates overanother with respect to a given attribute This scale trans-forms qualitative evaluations into numerical values from 1to 9 (Table 1) which are used to fill in the pairwise compari-son matrices The priority scales are then derived by calculat-ing the eigenvector associated with the principal eigenvalueof each comparison matrix (Saaty 1980) When the priority

Nat Hazards Earth Syst Sci 13 1209ndash1227 2013 wwwnat-hazards-earth-syst-scinet1312092013

G Le Cozannet et al An AHP-derived method for mapping the physical vulnerability of coastal areas 1211

Table 2 Pairwise comparison of three zones with respect to their potential to be eroded in 2100 with a hypothesis of no sea level risefollowing Saaty (2008a) The numerical values in the matrix are directly those from ldquoSaatyrsquos scalerdquo (Table 1) Experts filled the boldnumerical values For example the numerical value ldquo5rdquo (line 1 column 2) indicates that the judgement strongly favours that Z1 is morelikely to be eroded than Z2 in 2100 ldquo9rdquo in column 3 indicates that there is extreme evidence that Z3 is more likely to erode than Z1 Theresulting index provides the final scores obtained for each zone (Z1 Z2 and Z3) with respect to the criterion ldquopotential to be eroded in2100rdquo It is calculated by ldquoraising the pairwise comparison matrix to large powers summing each row and dividing each by the total sum ofall rowsrdquo (Saaty 2008a see Sect 23)

Z1 Zone ldquolikely tobe eroded by 2100rdquolocated withinextrapolatedobserved trend

Z2 Zone ldquopossiblyeroded by 2100rdquolocated within twicethe extrapolatedobserved trend

Z3 Zone ldquounlikely tobe eroded by 2100rdquoother areas

Resulting indexusing AHP

Z1 1 5 9 07352Z2 15 1 5 02067Z3 19 15 1 00581

scales in the matrix are consistent the normalised eigenvec-tor is unique and equal to any normalised column of the ma-trix However sometimes and especially when many com-parisons are involved the judgements may be inconsistentIn this case the priority scales can still be derived by solvingthe eigenvalue problem the eigenvector being an approxi-mation of the ideal case (Saaty 2008b Sekitani and Yamaki1999) The method also provides a framework for evaluatingthe consistency of the judgements (Saaty 2008a)

22 AHP in decision mapping problems

In multi-criteria decision-mapping problems alternativesmight be geographical entities or landforms while the crite-ria may be those that help to evaluate for example a level ofrisk or vulnerability There are numerous examples of the useof AHP mapping problems in previous studies In the field ofhazard and risk zoning it has been used to map landslide haz-ard and susceptibility (eg Ayalew et al 2005 Gorsevski etal 2006 Yalcin 2008 Ercanoglu et al 2008) soil erosionhazard mapping (Rahman et al 2009) earthquake hazardzoning (Pal et al 2008) and flood mapping (Nguyen Maiet al 2011 Chen et al 2011) For coastal risks Chang etal (2012) applied the method to rank different coastal protec-tion options However the only application of AHP that wefound for physical coastal vulnerability assessments in thecontext of climate change was the study by Yin et al (2012)which produced a national assessment of coastal vulnerabil-ity in China by adapting a CVI approach in which the vari-ables were weighted with AHP

23 Fundamental principle of applying AHP tomapping coastal vulnerability

In the coastal vulnerability mapping problems there is rela-tively abundant (but non-exhaustive) data (observations sim-ple models and expert opinions) that may describe one com-ponent of physical vulnerability For example a geological

map provides information about lithology and thus sensitiv-ity to erosion similarly past shoreline changes and simpli-fied models may be used to anticipate future erosion Ourapproach simply consists of using the AHP to translate thisinformation from experts simple models and data into com-parable quantitative data and to aggregate this data into a sin-gle multi-criteria mapping framework

In practice the approach consists of using Table 1 as areference for completing pairwise comparisons of differentareas with respect to each other and to each criterion For ex-ample by applying a simple model of shoreline change suchas the Bruun rule or the extrapolation of previously measuredrate of change we can define a mean shoreline change ratefrom now to 2100 and define three zones

ndash (Z1) areas eroded in 2100 according to the simplifiedmodel

ndash (Z2) areas eroded if the actual rate of erosion is twicethe rate according to the simplified model and

ndash (Z3) other locations ie area located farther inland fromzones Z1 and Z2

Here a group of coastal experts made the subjective choiceto use twice the rate of observed erosion to define areas thatwill likely be eroded by 2100 Using Saatyrsquos scale expertsmay agree that their ldquojudgement strongly supports the ideathatrdquo area (Z1) is more likely to be eroded than (Z3) Thisjudgement corresponds to a 5 in Saatyrsquos scale and is insertedin Table 2 Similarly there is ldquoevidence of extreme differenceof importancerdquo between areas (Z1) and (Z3) Therefore thenumerical value 9 is used in Table 2 The resulting indexesare calculated according to the AHP method

The next section explains how we have incorporated thissimple idea into a complete approach by replicating thesepairwise comparisons for all components accounting for thephysical vulnerability

wwwnat-hazards-earth-syst-scinet1312092013 Nat Hazards Earth Syst Sci 13 1209ndash1227 2013


33

820

821

Figure 1 Sequential process used in this study to produce and validate physical vulnerability 822

maps 823

824

Fig 1 Sequential process used in this study to produce and validate physical vulnerability maps

3 Method application at the study sites

We summarise the sequential method used in this paper forevaluating coastal physical vulnerability in Fig 1 We startwith preliminary investigations at the coastal sites and globalchange consequences on existing coastal processes (31) andthen apply a complete approach that integrates a number ofimportant criteria contributing to physical coastal vulnerabil-ity (32)

31 Preliminary investigations in the study sites

311 Site description

We apply the method in two regions of France with differ-ent climatic conditions geomorphology and data availabil-ity La Reunion in the south-western tropical Indian Oceanand Languedoc-Roussillon along the western Mediterraneancoast



Table 3Summary of data sources used in this study

Data Languedoc-Roussillon La Reunion

Data source DescriptionComment Data source DescriptionComment

Coastalgeomorphology

Eurosion (2004) 1100 000 descriptionof geomorphology anderosion trends

BRGM De LaTorre (2004) De La Torre etal (2006) Blangy etal (2010)

Local description ofgeomorphology anderosion trends

Observederosion trends

DREAL datareinterpreted byYates et al (2011)

Local rates of observedpast erosion

De La Torre etal (2006)

Sparse data from somerepresentative sites

Geology BRGM 1 50 000 Vectorisedgeological map

BRGM 1 50 000 Vectorisedgeological map

Digitalelevation model

Geographicinstitute (IGN)

50 m horizontalresolution grid

IGN data 10 m horizontalresolution grid

Land use Corine Land Cover(European EnvironmentAgency)

1 100 000 land coverdescription

Hydrographicnetwork

BD Carthage (IGN) 150 000 description ofwatersheds

BD Carthage

Exposure towaves climatesand eventuallycyclones

DREAL LanguedocRoussillon

Recommendations forsetting coastal risksprevention plans

Lecacheux et al (2012) Exposure to cyclonicsouthern and easternwaves

La Reunion is a volcanic island whose base lies 4000 mdeep on the ocean floor Its coastline is 250 km in length andits coastal geomorphology is dominated by volcanic rocksand pebble beaches (De La Torre 2004) The coastal zone isnourished by alluvial sediments particularly near the estuar-ies of large rivers The island is located in a region impactedby cyclones as well as southern swell Finally especially inthe east the islandrsquos very high precipitation enhances sed-iment transport in rivers nourishing the coast with alluvialsediments in the vicinity of some estuaries (Garcin et al2005) Beaches are thus composed of volcanic materials butalso of biodetritic sand in the south-west where a fringingreef has developed

The coast of Languedoc-Roussillon mainly consists ofsandy beaches with some rocky outcrops of Cenozoicbasalts and limestone (Agde Cap Leucate) or Mesozoiclimestone (Sete) To the south near Spain the coast fea-tures Paleozoic rock cliffs with pocket beaches Earlier stud-ies (Vinchon et al 2009) have concluded from a previousanalysis that coastal risks in this region would worsen withclimate change

312 Available data

The two regions differ considerably in terms of data avail-ability (Table 3) in Languedoc-Roussillon public data arevery heterogeneous with some inconsistencies in variousshoreline datasetsrsquo positioning (Yates-Michelin et al 2011)We used the data collected and acquired by the DREAL-LR (the regional agency of the national Environment Min-istry in Languedoc Roussillon) to calculate shoreline changerates over the entire region On the other hand while shore-line positioning in La Reunion was found to be mostly con-sistent between the various geographical data sources in-formation on erosion rates only exists at a few particularsites (De la Torre et al 2006)

Using lidar data available along a 20 km stretch of shore-line in the northern part of the Gulf of Lion Yates etal (2011) evaluated the vertical precision (127 m) and ac-curacy (016 m) of available digital elevation model (DEM)In La Reunion this could not be quantified but a comparisonwith topographical maps suggests that the available data areof similar quality

Geological maps represent the superficial geological for-mations based on an interpretation of field observations Thelimits of use of digital vectorised geological maps are re-lated to their scale (150 000) and in addition they do not



34

825

Figure 2 Simplified diagram of the interactions between various factors controlling the 826

physical vulnerability of coastal areas 827

828

829

Fig 2Simplified diagram of the interactions between various factors controlling the physical vulnerability of coastal areas

highlight if a formation has been observed or interpreted by ageologist In the two regions the late Quaternary (Holocene)deposits are sufficiently well detailed to use this informationin coastal areas

313 Regional consequences of global change and sealevel scenarios

While the future consequences of global change in both re-gions are still largely uncertain it is possible to identifya number of trends that suggest some possible future out-comes In both regions human use of the coastal area in-creased drastically during the second half of the 20th cen-tury (eg Deboudt 2010 Cazes-Duvat and Paskoff 2004)However current regulations and coastal risk preventionplans prevent unlimited urbanisation of the seafront In this

study we assumed that any further coastal urbanisationwould occur farther inland and did not consider their poten-tial effects on shoreline change (eg through the constructionof new coastal defences)

In this study we found that there is insufficient knowledgeabout future changes in waves storms and precipitation atthe study sites to take into account their effects on coastalprocesses Therefore we only considered changes due to fu-ture sea level rise Estimates of sea level rise in 2100 rangefrom about 05 m (Meehl et al 2007) to roughly 1 m (egRahmstorf 2007) with the main source of uncertainty be-ing the future contribution of melting processes in Green-land and the western Antarctic ice sheets Sea level changesvary regionally not only because of non-uniform warmingof the ocean and variations in salinity (eg Lombard et al2005 Meyssignac and Cazenave 2012) but also because of



changes in the gravitational field due to the redistributionof ice and water masses (Slangen et al 2012) and subsi-dence or uplift potentially affecting the coastal zone at var-ious scales (eg Woppelmann et al 2007) Therefore al-though Slangen et al (2012) developed a method for pre-dicting all of the components of future regional sea levelrise local projections are still uncertain In particular in ourcase studies the potential vertical ground motions are insuf-ficiently known (data available fromwwwsonelorg) More-over in the particular case of the Mediterranean Sea it isworth noting that using a single climate scenario Tsimpliset al (2008) showed that thermosteric effects may not causeany sea level rise along the French Mediterranean coast

These few examples support the idea that several sea levelscenarios should be considered in regional assessments of theeffects of future sea level rise In this study we consideredtwo extremely contradictory sea level scenarios for the endof the 21st century

ndash a 1 m sea level rise scenario by 2100 (called the ldquosealevel rise scenariordquo hereafter)

ndash a scenario in which current processes (eg erosion) areassumed to continue as in the recent past without be-ing significantly affected by sea level rise (ldquotrend sce-nariordquo)

The ldquosea level riserdquo scenario is consistent with results ofsemi-empirical studies (eg Rahmstorf 2007) It may also re-sult from a combination of moderate global sea level changesuperimposed with regional variability The second scenariomay result of regional variability in future relative sea levelrise but it is primarily designed for analysing how coastalvulnerability may change even without sea level rise (seeSect 51) These two scenarios are intended to provide con-tradictory hypotheses with which to test the robustness andreversibility of adaptation measures

314 Identification of important processes and the rela-tions between them

Using the knowledge and data described above we have builtup a scheme (Fig 2) of the various factors that control vul-nerability in the coastal system and their interactions Thesefactors were grouped into the following

ndash state factors describing how coastal erosion and ma-rine submersion relates to the local geomorphologyie lithology along coastal strips elevation and currentsedimentary balance as revealed by observations of pasterosion

ndash marine factors describing how each location is exposedto sea waves and surges eg during an easterly storm inLanguedoc-Roussillon or during a cyclone or southernwave event in La Reunion

ndash context factors describing how physical vulnerability ina specific location may be increased or decreased due tothe influence of nearby geomorphological systems forexample a beach located near to a major estuary may benourished with sediments thus preventing or reducingerosion

Figure 2 is a simplified scheme of the functioning of thecoastal zone which can be applied to a wide range of coastsat least those considered in this study

32 AHP application

This section describes how we set up the method in LaReunion and in Languedoc-Roussillon (Fig 1) We use Ar-cGIS for the geoprocessing and Matlab for the AHP calcula-tions

321 Building up the AHP framework

Using the preliminary investigations we convert the gen-eral conceptual tree developed during the preliminary in-vestigations (Fig 2) into an AHP-suitable hierarchicalscheme (Fig 3) which organises the various criteria thatare used to evaluate the physical vulnerability Then we de-termine the approximate radius of influence of coastal ero-sion and submersion (Eurosion 2004) which we divide intoelementary geographical entities ensuring that each entityis related to a single combination of factors accounting forits physical vulnerability We determine these elementarygeographical entities by intersecting existing data simplelaws and hypotheses (eg by cross-tabulating reference stormsurge levels with a DEM inferring susceptibility to erosionfrom a geological map etc)

322 Evaluating state marine and context factors

Replicating the initial principle presented in Sect 213 weuse Saatyrsquos table to compare the various geographical entitieswith respect to each criterion accounting for physical vulner-ability We apply the approach presented in Fig 3 as follows

State factors

For marine submersion we determine the approximate ra-dius of influence of marine submersion by crossing a digitalelevation model with water levels or reference storm surgelevels In Languedoc-Roussillon and La Reunion we use 2 mand 4 m respectively as the reference water levels in the lo-cal terrestrial framework (LTF) At La Reunion we obtainedthese values by summing the maximum tidal level (up to04 m in the LTF) storm surge (including pressure and windsurges (085 m)) and wave set-up and run-up (up to 2 m to 3 mfor typical wave heights during a cyclone) In Languedoc-Roussillon the centennial water level used for the regulatorycoastal risks plan is 2 m For the sea level rise scenario we



35

830

Figure 3 Hierarchical tree and related elementary zoning used for the ldquosea level riserdquo scenario 831

in Languedoc-Roussillon A very similar tree was used for La Reacuteunion (See table 5 for a 832

comparison of the two sites) In practice most of the geoprocessing work consisted in 833

defining elementary geographical spatial entities and rating them for each criterion 834

835

Fig 3 Hierarchical tree and related elementary zoning used for the ldquosea level riserdquo scenario in Languedoc-Roussillon A very similar treewas used for La Reunion (see Table 5 for a comparison of the two sites) In practice most of the geoprocessing work consisted of definingelementary geographical spatial entities and rating them for each criterion

raise these values by 1m Then we introduce a scaling of un-certainties to take account of the uncertainties in DEM andlow-lying area submersion processes

To estimate coastal shoreline change we apply simplifiedmodels and the associated scaling of uncertainties takinginto account the susceptibility of the coast to erode We de-rive this information from the geological map For examplein Languedoc-Roussillon shoreline retreat will not occur in

regions with Cretaceous rocks (as in the town of Sete) butwill occur in regions with easily erodible sediments such asHolocene sands or mud deposits (Table 4)

The simplified model of beach erosion is the formula sug-gested by the Eurosion (2004) project which combines ex-pected erosion due to sea level rise (Bruun 1962) with shore-line change due to ongoing processes as calculated from past



36

836

Figure 4 Baseline geographical data used to estimate state factors example in La Reacuteunion 837

near Saint-Paul (North-West) Submersion hazard risk is estimated in two scenarios by 838

crossing reference water levels with a DEM (Maps A and B) Observed shoreline change is 839

used to estimate how the shoreline may retreat in the future in the trend scenario (C) and the 840

Fig 4 Baseline geographical data used to estimate state factors example in La Reunion near Saint-Paul (north-west) Submersion hazardrisk is estimated in two scenarios by crossing reference water levels with a DEM (MapsA andB) Observed shoreline change is used toestimate how the shoreline may retreat in the future in the trend scenario (C) and the sea leve rise scenario (D) Erodibility is estimated fromgeological maps (E) Another example of erodibility mapping to the south-west of Saint-Paul is given in map (F) showing erodible beachessurrounded by rocky spurs (Data BRGM IGN)



Table 4A selection of geological classes and their associated erodibility and topography as derived for preparing the zonation In Languedoc-Roussillon 170 different geological classes have been classified according to their susceptibility to erosion and typical topography (18 in LaReunion island)

Description of geology in the 1 50 000vectorised geological map

Typical topography (elicited) Typical sensitivity to erosion (elicited)

Languedoc-Roussillon

Alluvium (Recent or active) Very low lying areas Highly erodibleAeolian decarbonated sandy dunes ofthe Vendres plateau

Low lying areas Highly erodible

Lacustrine limestones (late Oligocene) Hilly areas Slightly erodiblePresent barrier beach Very low lying areas Highly erodible

La Reunion island

Sandy beaches and dunes ndash Highly erodibleAlluvium ndash ErodiblePyroclastic materials lahars tuffscrees

ndash Moderately erodible

Basaltic ldquoflowsrdquo ndash Not erodible

shoreline observations

Rfuture = Rhistorical+Sfutureminus Shistorical

tan(α) (1)

where tan(α) is the beach slope and S is the sea level rise Forcliff erosion we combine spatially continuous knowledge ofpast erosion local knowledge of erosion rates calculated fora few representative sites and expert opinions (De La Torre etal 2006) The use of these models is arguable (eg Cooperand Pilkey 2004) so we consider a scaling of uncertaintiesin potential future erosion instead of a single future shore-line position (Table 2) We obtain the areas potentially af-fected by coastal erosion by cross-tabulating susceptibility toerosion with the probabilistic radius of influence of coastalerosion (Fig 4)

Marine factors

We elaborate complementary marine factors in order to high-light the more important physical vulnerability in low-lyinggeographical entities in areas exposed to waves and surgesIn Languedoc-Roussillon this leads to increasing the vulner-ability index where low-lying entities are close to lagoonsor rivers In La Reunion we use the previous results ofLecacheux et al (2012) to identify areas that can be affectedby extreme waves from cyclones or southerly waves

Context factors

Finally to further refine the mapping of coastal vulnerabilitywe considered how the various coastal systems may be af-fected by changes in sediment supply from adjacent units InLanguedoc-Roussillon we use watershed land use as an in-dicator of how human activities may have affected sedimenttransport from the watershed to the coastal zone We assume

that higher levels of urbanisation and to a lesser extent agri-culture reduce the transport of coarse sediments to coastal ar-eas In this study the Rhone delta is the most affected water-shed which is consistent with feedback from more quantita-tive analyses (Delmas et al 2012) Similarly in La Reunionwe consider the fact that the larger rivers are the main sourceof sediments but that coral reefs also nourish some westernbeaches This simplified approach produces an initial proxyfor evaluating the availability of coarse sediments but doesnot consider the subsequent coastal sediment transport pro-cesses eg the effects of longshore northeasterly sedimenttransport induced by easterly trade waves in La Reunion is-land This approach defines entities that are more likely tobe resilient to worsening erosion due to the role of adjacentsedimentary units

All of the qualitative assessments are evaluated usingSaatyrsquos correspondence table (Table 1) and calculations ofscores from pairwise comparison matrices (similar to Ta-ble 2) Here we calculated between 5 and 7 matrices perscenario and per study site with size ranging from 3times 3 to7times7 depending on the number of classes considered for eachcriterion We provide an example in Table 4 which illus-trates how we classify the initial 170 geological classes in the1 50 000 geological map of the coastal zones of Languedoc-Roussillon into 7 bins leading to a 7times 7 matrix defining anerodability index at each location

323 Weighting process role of user groups andexploitation of the results

Finally we integrate and aggregate the geographical featuresand the basic indicators defined above into an AHP hierarchi-cal tree Table 5 shows how we weighted the various criteria



in the hierarchical tree with the factor weighting resultingfrom an application of AHP

One user group and two expert groups critically reviewedthe resulting maps and weights (Fig 1) The French Ministryof the Environment set up and coordinated the user group toensure that the work undertaken to map coastal vulnerabilitywould comply with the needs of coastal stakeholders We setup the expert groups to help in the preliminary investigationsto monitor the adaptation of the AHP method to review thevulnerability maps and the associated uncertainties and ulti-mately to approve their dissemination The experts comparedthe preliminary maps with their subjective understanding ofcoastal vulnerability in the region

In order to estimate the potential social and economic im-pacts previous studies developed indicators that combinedphysical and societal vulnerability using for example pop-ulation density data (eg McLaughlin et al 2002 Hegde andReju 2007) or more detailed social surveys including factorssuch as employment income poverty age etc (Boruff et al2005) Here we simply identify a selection of ldquohotspotsrdquo byoverlaying the physical vulnerability with the existing assets(current residential tourist harbour and industrial assets) asan illustrative example of a potential application of the phys-ical vulnerability maps

324 Uncertainties

Although uncertainties of multi-criteria mapping are diffi-cult to quantify accurately it is still possible to list them andto estimate qualitatively their significance (Bell and Glade2004) We provide this analysis in Table 6 for the two appli-cations in Languedoc-Roussillon and La Reunion One im-portant source of uncertainty is the difficulty that the expertshad in judging the relative importance of different factors atthe highest levels of the AHP decision tree (eg relative im-portance of context factors with respect to marine or statefactors) This difficulty is also acknowledged in other coastalvulnerability multi-criteria analyses For example Gornitz etal (1991) chose a geometric aggregation in the coastal vul-nerability index because it was considered to be the least sen-sitive to gaps in data Here the state factors were the mostimportant criteria with respect to marine and context fac-tors with a difference of importance of intensity 4 on theSaaty scale Nevertheless similarly to Coelho et al (2006)we tested other weighting options in a sensitivity analysis tohelp evaluate the significance of these uncertainties (Fig 5)The results are less affected by a change of the weightingsof the decision tree than by a change the initial sea level risehypothesis Table 6 also shows that many uncertainties arisefrom the limitations of the conceptual models used for eval-uating future shoreline retreat and submersion Finally thequestion of spatial and temporal multi-scale dynamics wasnot addressed in this study

4 Results

41 Characteristics of resulting maps

We provide examples of the resulting maps in Fig 6 Thesemaps represent the ranking obtained after applying the AHPto each geographical entity considered as geomorphologi-cally and morphodynamically homogeneous The results canbe interpreted as a physical vulnerability index Since theranking is performed independently for the two sea level sce-narios the scales cannot be compared in Figs 6a and b (re-spectively 6d and e) Therefore we produced a third seriesof maps to show where physical vulnerability is expectedto increase with sea level rise compared to the trend sce-nario (Figs 6c and f) Taking into account the uncertainties(Table 6) we estimated that the map resolution should be1 50 000 for Languedoc-Roussillon and under 1 25 000 forLa Reunion

Regarding the sea level rise scenarios in both regionsthese maps reflect the fact that sea level rise is expected toincrease erosion of beaches and soft cliffs and submersionof low-lying areas especially those located close to lagoonsComparing maps 6a and b (respectively 6d and e) shows thateven without sea level rise some portions of coastal areasare vulnerable to erosion and flooding These conclusions areconsistent with an intuitive analysis but the added value ofthe maps presented here is that each location is assigned aphysical vulnerability index which can be used for land useplanning For example when considering establishing newinfrastructure with a multi-decadal design life it is preferableto choose an area with a lower physical vulnerability indexnotwithstanding other considerations (eg the economic orenvironmental advantages of a given location) and future im-provements in the understanding of coastal hazard changes

Depending on the implementation of adaptation strategiessome possible consequences of sea level rise in the mostvulnerable areas are the loss of land changes in land useor increased costs for risk prevention For example someparts of the harbour in Sete (in Languedoc-Roussillon) andLe Port (in La Reunion) are shown to be highly vulnerableHowever in urban industrialised or harbour environmentsthe main changes are expected to be caused by human ac-tivities and not climate change The maps presented here in-dicate that according to current knowledge these areas willlikely be affected by more frequent degradations (eg due toflooding erosion scouring sea wave pressure or the collec-tion of debris in between them) resulting in higher mainte-nance costs

42 La Reunion island

For both of the La Reunion scenarios the results show thehighest vulnerability in low-lying areas and areas immedi-ately adjacent to the sea particularly caused by coastal beachand cliff erosion Compared to the trend scenario the sea



Table 5 Hierarchical tree and associated weighting used for Languedoc-Roussillon The criteria weighting are obtained by applying AHPsimilar to the example in Table 2

Factors Factor weighting Criteria Criteria weighting


State factors 0667 Area potentially affected by erosion 05Area potentially affected by flooding 05

Marine factors 0167 Exposure to waves during storms 1Context factors 0167 External sediment supply 1

La Reunion


Marine factors 0167 Extreme events 084Non-extreme coastal dynamics 016

Context factors 0167 External sediment supply 1

37

sea leve rise scenario (D) Erodibility is estimated from geological maps (E) Another 841

example of erodibility mapping to the south-west of Saint-Paul is given in map (F) showing 842

erodible beaches surrounded by rocky spurs (Data BRGM IGN) 843

844

845

Figure 5 Sensitivity analysis performed for La Reacuteunion to compare the redistribution of 846

entity rankings when the criteria weightings in Table 5 change according to the sea level rise 847

(right) and trend scenarios (left) On the right the criteria are considered either of equal 848

weight or as in Table 5 This corresponds to the maximum acceptable range of criteria 849

weighting The largest spread in the left-hand histogram (Dgtd) shows that considering two 850

scenarios causes more changes in entity ranking than modifying the criteria weights in 851

Table 5 This sensitivity analysis is used to prepare the qualitative assessment of uncertainties 852

in Table 6 853

854

Fig 5Sensitivity analysis performed for La Reunion to compare the redistribution of entity rankings when the criteria weightings in Table 5change according to the sea level rise (right) and trend scenarios (left) On the right the criteria are considered either of equal weight or asin Table 5 This corresponds to the maximum acceptable range of criteria weighting The largest spread in the left-hand histogram (Dgtd)shows that considering two scenarios causes more changes in entity ranking than modifying the criteria weights in Table 5 This sensitivityanalysis is used to prepare the qualitative assessment of uncertainties in Table 6

level rise scenario is characterised by increases in floodinghazards in these three areas as well as by greater erosion oferodible cliffs and beaches (Fig 6c)

We found that physical vulnerability is highest in the fol-lowing areas (Fig 7) Sainte Suzanne (north-east) Saint-Paul Le Port (north-west) the Etang-Sale (south) and in anumber of more localised places adjacent to beaches (egin Saint Pierre) or erodible cliffs (eg Saint Philippe) How-ever a first analysis of vulnerable assets highlights Le Portand Saint-Paul as the most impacted areas concerning the po-tential losses In Saint-Paul relatively low-lying and erodi-ble areas are exposed to storms and cyclones because thecoast is not protected from waves by an offshore coral reef

However the existing assets are presently considered to belocated in sufficiently elevated areas to prevent flooding inmost cases In Le Port the industrial assets are located in anarea affected by high rates of coastal erosion Field surveysindicated that the nourishment of this area by river sedimentsis prevented by river and coastal work (Aubie and Oliv-eros 1999) Other hotspots are distributed along the coastincluding assets such as crops buildings roads recreationalpathways and beaches (De La Torre et al 2006 Blangy etal 2010)



Table 6Qualitative assessment of uncertainties for physical vulnerability assessment in Languedoc-Roussillon (LR) and La Reunion (REU)(following the approach of Bell and Glade 2004)

Source of uncertainty Uncertainty Reason Significance

Building the AHPhierarchical tree

Medium The hierarchical tree is a simple but limited representa-tion of the coastal system and its interactions (Figs 2and 3)

Medium

Elementary zoning High in LRVery highin REU

Limitations of models rough estimations sparse data inREU

Very high

Evaluating scores for eachcriteria for each geographicalentity

Medium in LRLow in REU

Insufficient resolution accuracy and precision of somedatasets (eg MNT in LR) inaccurate relative position-ing of data heterogeneity of acquisition methods

Medium

Defining the weighting ofcriteria and factors in the AHPtree

Medium Some subjectivity However its significance remainscontained according to sensitivity analysis (Fig 5)

Low to medium

Expert validation High Subjectivity However its significance remains limitedsince the role of experts remains to provide rules anddata for the evaluation

Medium

Interactions between local andregional processes

High Not taken into account Medium

38

855

Figure 6 Examples of results for La Reacuteunion (Saint Paul) and Languedoc-Roussillon (near 856

Palavas-les-Flots) The maps are provided for 2 scenarios sea level rise (A D) and trend 857

scenarios (B E) Maps C F show areas where the physical vulnerability is expected to 858

increase the most in the sea level rise scenario (Data BRGM IGN) 859

Fig 6 Examples of results for La Reunion (Saint-Paul) and Languedoc-Roussillon (near Palavas-les-Flots) The maps are provided for2 scenarios sea level rise(A D) and trend scenarios(B E) MapsC F show areas where the physical vulnerability is expected to increasethe most in the sea level rise scenario (Data BRGM IGN)



39

860

Figure 7 Illustrative example of potential use of the resulting maps cross-tabulation of 861

physical vulnerability with major assets in La Reacuteunion and Languedoc-Roussillon 862

863

Fig 7 Illustrative example of potential use of the resulting maps cross tabulation of physical vulnerability with major assets in La Reunionand Languedoc-Roussillon

43 Languedoc-Roussillon

In Languedoc-Roussillon the trend and sea level rise sce-narios both show the high physical vulnerability of low-lying areas (eg the shores of the Sigean and Leucate la-goons) sandbars (eg from Carnon to Frontignan) anddeltas (Camargue) The maps also show that the river flood-plains (ie recent sedimentary deposits) are most exposed toflash flood hazards especially during heavy rainfall eventscombined (floods) with marine storm surges

In many sand spits (eg east of Carnon-Plage) the mapsshow that lagoon and sea shores are vulnerable for two differ-ent reasons The coastline facing the sea is most impacted byerosion while the lagoon shoreline lying no more than 1mabove sea level is more impacted by submersion The cen-tral part of sand spits (low sand dunes sometimes affected byhuman activity and typically lying 2 to 3 m above mean sealevel) appears to be the least vulnerable feature

Figure 7 highlights the locations where high levels ofphysical vulnerability coincide with assets The maps arequite similar to the earlier findings of Vinchon et al (2009)with some differences in the identification of hotspots andtheir relative importance For example in this study we onlyconsidered hotspots with major human infrastructure whileVinchon et al (2009) also considered assets of major envi-ronmental value (eg in the Camargue south of Le Grau duRoi) Regardless both maps show high coastal vulnerabil-

ity in this region whatever the future implications of climatechange may be

5 Discussion

In order to analyse the benefits limitations and potentialuses of the proposed approach we evaluated the method us-ing the criteria of Preston et al (2011) in their generic rec-ommendations for undertaking vulnerability mapping stud-ies (Table 7)

51 Benefits of the method

511 Integration of expert opinion

When attempting to evaluate potential consequences of cli-mate change in coastal areas a first approach consists of di-rectly assessing potential damages due to changing coastalerosion and flooding processes under the hypothesis of a cer-tain rate of sea level rise (eg Hinkel and Klein 2009) Suchapproaches are commonly expected to produce assessmentsof adaptation measures in terms of costs and benefits Whensufficiently high-value assets are at stake and when futurechanges in coastal processes are well understood such cost-benefit analyses are useful (eg Hallegatte et al 2011) How-ever in other cases eg the coastal sites of this study the un-certainties resulting from climate modelling hazard and risk



assessments and economic estimations of potential damagecan make these evaluations unreliable (Yates-Michelin et al2011)

In these cases integrating expert opinions in the approachused to map future coastal vulnerability is highly impor-tant (eg Hanson et al 2010) Here AHP makes it pos-sible to convert qualitative assessments of groups of ex-perts into quantitative information that can be used to definethe weighting of various criteria in multi-criteria analysesAlthough achieving a consensus amongst members of userand expert groups is not easy their involvement in ourstudy (Fig 1) was useful for identifying the remaining in-consistencies in weighting criteria and alternatives and fur-ther improving the maps In practice this step does not allowthe experts to create a new map as they wish because the re-sults remain constrained by the basic data and the referencescale of the AHP method (Table 1) In addition they need tosupport their opinions with data observations or models Fi-nally this work with expert and user groups helps to identifymore accurately the most important uncertainties (Table 6)The ability of AHP to manage different degrees of convictionfacilitates the integration of uncertainties associated for ex-ample with the future position of shorelines To summarisethe main benefit of the proposed approach is its capacity togather existing knowledge and existing datasets in a struc-tured consistent manner in order to map future coastal phys-ical vulnerability

512 Transportability of the method

We tested the method in two different contexts in terms ofgeodynamics and data availability In practice we were ableto adapt the method to these two different contexts The min-imum data requirements for applying this method are theexistence of coastal geomorphology data historical erosionand accretion rates and topographic data in the coastal zoneThe geological map improves the final results once the ge-ological information is converted into information about thelithology and thus the susceptibility to erosion the final mapinherits the geometry of the geological layer However thisinterpretation of the geological notations is not straightfor-ward because it depends on the geological map and requiresgeological expertise In particular Holocene deposits mustbe accurately mapped Therefore we think that the approachcan be transported to other coastal locations provided thereis sufficient data and the users accept its limitations and un-certainties

52 Limitations of the method

We discuss below the limitations of this application of AHPin coastal vulnerability mapping by referring to the assess-ments of uncertainties (Table 6) and to the method as awhole (Table 7)

521 Representation of the coastal system in anAHP-hierarchical tree

First the hierarchical framework (Fig 3) only offers a lim-ited representation of the conceptual graphs presented inFig 2 the hierarchical framework cannot handle interactionsbetween criteria because it uses each criterion as an indica-tion of the physical vulnerability of each geographical en-tity regardless of potential relationships with other criteriaFor example in Languedoc-Roussillon the context factorseffectively provide a regional indication of where the sedi-ment supply can compensate for erosion This indicator isimportant as it links watersheds to coastal zones Howeverwe simply added it to the state factors which use more de-tailed and local information on erosion and no interactionbetween scales is modelled here Implicitly it is postulatedthat the combination of simple maps of criteria would av-erage these interactions and produce a consistent picture ofvulnerability at the regional scale This limitation is inher-ent to the multi-criteria approach with additive aggregationschemes

522 Lack of data

The second limitation relates to random uncertaintiesie those associated with a lack of data or with data qual-ity In this study the existing datasets are heterogeneous interms of quality and their geometry (spatial precision and ac-curacy) is sometimes insufficient for detailed zoning Someinformation is lacking for example the geological map doesnot provide any geotechnical information which may be im-portant for quantifying cliff susceptibility to erosion (Benu-mof and Griggs 1999) We were also lacking data to anal-yse the consequences of fringing coral reef degradation andsea level rise on lagoon waves currents and shoreline dy-namics (eg Storlazzi et al 2011) In some cases existingdata presented some spatial inconsistencies for example theactual shoreline position differs between various GIS shore-line datasets in Languedoc-Roussillon In this region the dif-ferences in shoreline position according to the various geo-graphical layers (eg geology DEM land cover etc) requiresubstantial geoprocessing and can result in errors in physi-cal vulnerability assessments when zooming to local scalesA similar difficulty arises from the fact that the vectorisedgeological map legends are inconsistent in different admin-istrative departments These random uncertainties could bereduced with new datasets



Table 7Evaluation of the proposed method

Evaluation criteria (after Preston et al 2011) Application in this study

Objective of the approach Mapping physical vulnerability to erosion and floodingin the context of future sea level rise

Utility associated with spatial analysis of vulnerability Local and regional vulnerability assessments arerequested by an environmental regulation framework

Benefits for stakeholders Planning long-life infrastructures forecasting long-term development

Potential risk associated with representing information spatially Misevaluation of vulnerability due to imprecise vulner-ability scale use of maps at too local scales (eg under1 5 000 in La Reunion )

Determinants of vulnerability See Fig 2

Representation of spatial temporal and multi-scale dynamics Not satisfactorily addressed because of lack of data andmethods

Methods for vulnerability assessment and mapping Multi-criteria mapping structured with AHP

Uncertainties assessment Qualitative assessment completed with sensitivity anal-ysis (See Sect 324)

Validation of maps Interactions with 2 expert groups of geomorphologists

Intended audience Regional authorities in charge of implementing regionalscenarios

Communicability of the results Interactions with a user group (coastal managementstakeholders of state and regional authorities) gatheredby the Ministry of Environment

Translation of vulnerability assessment into actions Existing environmental framework to support adapta-tion to climate change existing regulations onadaptation and land use planning

523 Lack of knowledge

The third limitation relates to the epistemic uncertaintiesie those due to unknowns in physical processes First themaps presented here remain regional evaluations that do notaddress the complexity of scale interactions and local pro-cesses such as the regional consequences of localised sandspit changes Secondly our maps rely on a number of fun-damental laws that produce a spatial representation of futurecoastal erosion and submersion In this study we partly com-pensated for the indiscriminate nature of these laws by intro-ducing degrees of uncertainty and by taking susceptibility toerosion and submersion into account for each geographicalentity However using these laws remains arguable and in-troduces spatial zoning uncertainties that are difficult to es-timate For example recent studies propose more sophisti-cated approaches than the use of the Bruun rule but they re-quire more data and have presently been developed only forspecific locations (eg Ranasinghe et al 2012 2013) As asummary this limitation is related to the lack of commonlyrecognised methods to properly estimate the impacts of sealevel rise on coastal processes

524 Limited range of hazards considered in this study

Finally this study uniquely focuses on coastal threats whileusers also need information on other threats Especially inLa Reunion ldquomulti-hazardrdquo or ldquomulti-riskrdquo approaches (egVecchia 2001 Grunthal et al 2006 Douglas 2007) arenecessary to favour the establishment of infrastructure andhomes in areas where the overall threat is low whether re-lated to coastal erosion marine flooding or other completelyindependent risks such as landslides river flooding or evenvolcanism Such approaches are currently applied in regula-tory risk prevention plans with a focus on present hazards

53 Potential use in adaptation policies

Our resulting maps must be considered critically and couldbe improved upon with new knowledge or data How-ever the expert groups judged them informative and rele-vant for stakeholders involved in long-term coastal planningdevelopment and adaptation since they incorporate moreknowledge about sea level rise vulnerability than many ofthe existing shorter-term coastal prevention plans They may



also offer additional support for local authorities making de-velopment decisions particularly in low-lying areas

Intergovernmental and European commitments requireEuropean countries to introduce climate mitigation andadaptation strategies In France ldquoclimate and energy plansrdquoare implemented at the regional level including adaptationmeasures that should be based on assessments of vulnera-bility to climate change These plans should then feed intoregulatory regional land use policies One of the challengesof this process is to generate realistic scenarios in a contextof high uncertainty

From the decision-making perspective one advantage ofthe approach presented here is its ability to generate mul-tiple scenarios that can then be used in participatory ap-proaches to develop adaptive measures Coastal stakeholdersmay choose one or more of the scenarios produced here oruse the proposed framework for creating their own scenar-ios taking account of additional relevant information Basedon the different assessments of vulnerability they can thendecide whether the proposed adaptation strategies are themost appropriate from the point of view of their robustnessreversibility and potential short-term benefits (Hallegatte2009)

The short-term benefits of proposed adaptation strategieslink adaptation with risk prevention Our results support pre-vious studies that stress the importance of introducing riskreduction strategies in low-lying coastal areas especially forsand spits and around estuaries In this context an exampleof a ldquono regrets strategyrdquo could be to limit new peri-urbansettlements and tourist infrastructure in these areas This pro-cess is complementary to regulatory coastal risk preventionplans that address shorter term timescales

As well as ldquono regretsrdquo strategies exchanges with theuser group highlighted the difficulties of imagining adequateadaptation measures for long-term planning For examplein highly vulnerable and already urbanised areas such asPalavas-Les-Flots in Languedoc-Roussillon or Saint Paul inLa Reunion removing exposed assets is often not an accept-able option for stakeholders and citizens It therefore seemsthat in many cases the only remaining option is to maintainshorelines and protect them from submersion and erosionwhatever the cost

6 Conclusions

Using an AHP-derived method we evaluated and mappedphysical vulnerability to erosion and submersion at regionalscales in two different settings The main strength of themethod is its ability to convert expert opinions into numeri-cal values and to integrate quantitative and qualitative knowl-edge and data in a structured way This is very useful fordefining weightings in multi-criteria approaches when theoverall problem is not well formalised which is the casefor assessments of coastal physical vulnerability While the

weaknesses of the approach as well as the epistemic and ran-dom uncertainties are recognised the resulting maps wereconsidered to be informative and useful for coastal managersresponsible for long-term planning

In practice such approaches can provide a baseline forphysical vulnerability assessments to support adaptationstrategies This study has shown that we are able to pro-duce vulnerability maps that satisfy geomorphology ex-pert groups Along with other incentives and regulations(eg coastal risk prevention plans) this may result in furtherinforming spatial planning policies in coastal areas and con-tribute to ongoing forward studies on climate change adapta-tion

AcknowledgementsThis study was funded by the French Environ-ment Ministry through the Explore 2070 project with additionalsupport from the French Research National Agency (ANR) forthe masterrsquos thesis of M Mendez and the publication of this work(CECILE project lsquoCoastal Environmental Changes Impact of seaLEvel risersquo under grant number ANR-09-CEP-001-01) DREAL-LR the French lsquoCoastal Observatoryrsquo the Herault Departementand EID Mediterranee provided data for this study We also usedthe SONEL database We thank the three groups of experts andusers for their useful inputs and comments as well as RodrigoPedreros Yann Balouin Alexis Stepanian Arnaud Blangy andYwenn De La Torre for their support We also thank the twoanonymous reviewers for their comments that helped to improvethis paper

Edited by O KatzReviewed by two anonymous referees

References

Alpar B Vulnerability of Turkish coasts to accel-erated sea-level rise Geomorphology 107 58ndash63doi101016jgeomorph200705021 2009

Aubie S and Oliveros C Evolution du trait de cote de 1950a 1997de la baie de la Possessiona lrsquoembouchure de lrsquoEtang de SaintPaul Ile de la Reunion [Shoreline evolution from 1950 to 1997in Possession Baya at Saint Paul swamp estuary La Reunion is-land] BRGMR 40780 1999

Ayalew L Yamagishi H Marui H and Kanno T Landslidesin Sado island Japan Part ii GIS-based susceptibility mappingwith comparisons of results from two methods and verificationsEng Geol 81 432ndash445doi101016jenggeo2005080042005

Bell R and Glade T Quantitative risk analysis for landslides ndash Ex-amples from Bıldudalur NW-Iceland Nat Hazards Earth SystSci 4 117ndash131doi105194nhess-4-117-2004 2004

Benumof B T and Griggs G B The Dependence of Seacliff Ero-sion Rates on Cliff Material Properties and Physical ProcessesSan Diego County California Shore amp Beach 67 29ndash41 1999

Blangy A De La Torre Y Vaslet E Mallet C and Dewez TMorphodynamique des littoraux de La Reunion Phase 3 Suiviet gestion de lrsquoerosion cotiere sur 12 sites identifies comme sensi-bles [Morphodynamics of coasts of La Reunion island phase 3



Monitoring and management of coastal erosion over 12 sensitivesites] BRGMRP-57431-FR 2010

Boruff B J Emrich C and Cutter S L Erosion hazard vul-nerability of US coastal counties J Coast Res 21 932ndash942doi10211204-01721 2005

Bruun P Sea level rise as a cause of shore erosion Journal Water-ways and Harbours Division 88 117ndash130 1962

Cazes-Duvat V and Paskoff R Les littoraux des Mascareignesentre nature et amenagement [Mascarene coasts between natureand development] LrsquoHarmattan 186 pp 2004

Chakar S Cartographie decisionnelle multicriteres formalisationet implementation informatique [Multicriteria decision map-ping Formalization and computer implementation] Doctoralthesis University of Paris Dauphine 2006

Chang H-K Liou J-C and Chen W-W Protection priorityin the coastal environment using a hybrid ahp-topsis methodon the Miaoli coast Taiwan J Coast Res 28 369ndash374doi102112jcoastres-d-10-000921 2012

Chen Y-R Yeh C-H and Yu B Integrated application of theanalytic hierarchy process and the geographic information sys-tem for flood risk assessment and flood plain management inTaiwan Nat Hazards 59 1261ndash1276doi101007s11069-011-9831-7 2011

Coco G and Murray A B Patterns in the sand From forcingtemplates to self organization Geomorphology 91 271ndash290httpdxdoiorg101016jgeomorph200704023 2007

Coelho C Silva R Veloso-Gomes F and Pinto F T A vul-nerability analysis approach for the Portuguese West Coast WitTrans Ecol Envir 91 251ndash262doi102495risk060241 2006

Cooper N J and Jay H Prediction of large scale coastal ten-dency development and application of a quantitative behaviorbased methodology Journal of Coastal Research ICS 2002 pro-ceedings Special Issue 36 173ndash181 2002

Cooper J A G and Pilkey O H Sea-level rise and shoreline re-treat Time to abandon the bruun rule Global Planet Change 43157ndash171doi101016jgloplacha200407001 2004

Deboudt P Towards coastal risk management inFrance Ocean Coast Manag 53 366ndash378doi101016jocecoaman201004013 2010

Delmas M Cerdan O Cheviron B Mouchel J-M and EyrolleF Sediment export from French rivers to the sea Earth SurfProc Landf 37doi101002esp3219 2012

De La Torre Y Synthese morphodynamique des littoraux de LaReunionetat des lieux et tendances drsquoevolution a lrsquoechelle delrsquo ıle [Synthesis of morphodynamics of La Reunion coasts inven-tory and trends across the island] BRGMRP53307-FR 2004

De La Torre Y Balouin Y and Dewez T Morphodynamiquedes littoraux de La Reunion Phase 2 estimation de lrsquoerosioncotiere sur les sites identifies comme sensibles [Morphodynam-ics of coasts of La Reunion island phase 3 Monitoring and man-agement of coastal erosion over 12 sensitive sites] BRGMRP-55014-FR 2006

Douglas J Physical vulnerability modelling in natural hazardrisk assessment Nat Hazards Earth Syst Sci 7 283ndash288doi105194nhess-7-283-2007 2007

Doukakis E Identifying coastal vulnerability to climate changes JMarine Environ Eng 8 155ndash160 2005

Ercanoglu M Kasmer O and Temiz N Adaptation and compar-ison of expert opinion to analytical hierarchy process for land-

slide susceptibility mapping Bull Eng Geol Environ 67 565ndash578doi101007s10064-008-0170-1 2008

Eurosionwwweurosionorg last access 16062012 2004Fairbank H and Jakeways J Mapping coastal risk in a changing

climate Isle of Wight Centre for Coastal Environment 2006Garcin M Poisson B and Pouget R High rates of geomor-

phological processes in a tropical area the Remparts Rivercase study (Reunion Island Indian Ocean) Geomorphology 67doi101016jgeomorph200411002 2005

Gornitz V Global coastal hazards from future sea-level riseGlobal Planet Change 89 379ndash398 1991

Gorsevski P V Jankowski P and Gessler P E An heuristic ap-proach for mapping landslide hazard by integrating fuzzy logicwith analytic hierarchy process Control Cybern 35 121ndash1462006

Grunthal G Thieken A H Schwarz J Radtke K S SmolkaA and Merz B Comparative risk assessments for the city ofcologne - storms floods earthquakes Nat Hazards 38 21ndash44doi101007s11069-005-8598-0 2006

Hallegatte S Strategies to adapt to an uncertain cli-mate change Global Environ Chang 19 240ndash247doi101016jgloenvcha200812003 2009

Hallegatte S Ranger N Mestre O Dumas P Corfee-MorlotJ Herweijer C and Wood R M Assessing climate changeimpacts sea level rise and storm surge risk in port cities acase study on Copenhagen Climatic Change 104 113ndash137doi101007s10584-010-9978-3 2011

Hanson H Aarninkhof S Capobianco M Jimenez J A Lar-son M Nicholls R J Plant N G Southgate H N SteetzelH J Stive M J F and de Vriend H J Modelling of coastalevolution on yearly to decadal time scales J Coast Res 19190ndash811 2003

Hanson S Nicholls R J Balson P Brown I French J RSpencer T and Sutherland W J Capturing coastal geomor-phological change within regional integrated assessment Anoutcome-driven fuzzy logic approach J Coast Res 26 831ndash842doi102112jcoastres-d-09-000781 2010

Hegde A V and Reju V R Development of coastal vulnerabilityindex for the Mangalore coast India J Coast Res 23 1106ndash1111doi10211204-02591 2007

Hinkel J and Klein R J T Integrating knowledge to as-sess coastal vulnerability to sea-level rise The developmentof the DIVA tool Global Environ Chang 19 384ndash395doi101016jgloenvcha200903002 2009

Lebbe L Van Meir N and Viaene P Potential implicationsof sea-level rise for Belgium J Coast Res 24 358ndash366doi10211207a-00091 2008

Lecacheux S Pedreros R Le Cozannet G Thiebot JDe La Torre Y and Bulteau T A method to characterize thedifferent extreme waves for islands exposed to various waveregimes a case study devoted to Reunion Island Nat HazardsEarth Syst Sci 12 2425ndash2437doi105194nhess-12-2425-2012 2012

Lombard A Cazenave A DoMinh K Cabanes C andNerem R S Thermosteric sea level rise for the past 50years comparison with tide gauges and inference on watermass contribution Global and Planetary Change 48 303ndash312doi101016jgloplacha200502007 2005



Malczewski J GIS-based multicriteria decision analysis A surveyof the literature Int J Geogr Information Sci 20 703ndash726doi10108013658810600661508 2006

McLaughlin S McKenna J and Cooper J A G Socio-economicdata in coastal vulnerability indices constraints and opportuni-ties J Coast Res ICS 2002 proceedings Special Issue 36 487ndash497 2002

Meehl G A Stocker T F Collins W D Friedlingstein P GayeA T Gregory J M Kitoh A Knutti R Murphy J M NodaA Raper S C B Watterson I G Weaver A J and Zhao Z-C Global Climate Projections in Climate Change 2007 ThePhysical Science Basis Contribution of Working Group I to theFourth Assessment Report of the Intergovernmental Panel onClimate Change edited by Solomon S Qin D Manning MChen Z Marquis M Averyt K B Tignor M and Miller HL Cambridge University Press Cambridge UK and New YorkNY USA 747ndash846 2007

Meyssignac B and Cazenave A Sea level A review of present-day and recent-past changes and variability J Geodynam 5896ndash109 2012

Nicholls R J and Cazenave A Sea-Level Rise and Its Impacton Coastal Zones Science 328doi101126science11857822010

Nguyen Mai D Babel M S and Luong H T Evaluation of foodrisk parameters in the day river flood diversion area Red Riverdelta Vietnam Nat Hazards 56 169ndash194doi101007s11069-010-9558-x 2011

Pal I Nath S K Shukla K Pal D K Raj A Thingbai-jam K K S and Bansal B K Earthquake hazard zonation ofSikkim Himalaya using a GIS platform Nat Hazards 45 333ndash377doi101007s11069-007-9173-7 2008

Paskoff R P Potential implications of sea-level rise forFrance J Coast Res 20 424ndash434doi1021121551-5036(2004)020[0424piosrf]20co2 2004

Preston B L Yuen E J and Westaway R M Putting vulnerabil-ity to climate change on the map A review of approaches ben-efits and risks Sustain Sci 6 177ndash202doi101007s11625-011-0129-1 2011

Rahman M R Shi Z H and Chongfa C Soil erosionhazard evaluation-an integrated use of remote sensing GISand statistical approaches with biophysical parameters to-wards management strategies Ecol Model 220 1724ndash1734doi101016jecolmodel200904004 2009

Rahmstorf S A semi-empirical approach to projecting future sea-level rise Science 315 368ndash370doi101126science11354562007

Ranasinghe R Callaghan D and Stive M J F Estimatingcoastal recession due to sea level rise beyond the Bruun ruleClimatic Change 110 561ndash574doi101007s10584-011-0107-8 2012

Ranasinghe R Duong T M Uhlenbrook S Roelvink Dand Stive M Climate-change impact assessment for inlet-interrupted coastlines 3 83ndash87 2013

Romieu E Welle T Schneiderbauer S Pelling M and VinchonC Vulnerability assessment within climate change and naturalhazard contexts revealing gaps and synergies through coastalapplications Sustain Sci 5 159ndash170doi101007s11625-010-0112-2 2010

Saaty T L The Analytic Hierarchy Process New York McGrawHill International Translated to Russian Portuguese and Chi-nese Revised editions Paperback (1996 2000) Pittsburgh RWSPublications 1980

Saaty T L Decision making with the analytic hierarchy processInt J Services Sciences 1 1 2008a

Saaty T L Relative Measurement and Its Generalization in Deci-sion Making Why Pairwise Comparisons are Central in Math-ematics for the Measurement of Intangible Factors The Ana-lytic HierarchyNetwork Process RACSAM 102 2008 251ndash318 2008b

Sekitani K and Yamaki N A logical interpretation for the eigen-value method in AHP Journal of the Operations Research SocJapan 42 219ndash232 1999

Storlazzi C D Elias E Field M E and Presto M K Numer-ical modeling of the impact of sea-level rise on fringing coralreef hydrodynamics and sediment transport Coral Reefs 30doi101007s00338-011-0723-9 2011

Slangen A B A Katsman C A van de Wal R S W Ver-meersen L L A and Riva R E M Towards regional pro-jections of twenty-first century sea-level change based on IPCCSRES scenarios Climate Dynamics 38 101007s00382-011-1057-6 2012

Tsimplis M N Marcos M and Somot S 21st century Mediter-ranean sea level rise Steric and atmospheric pressure contribu-tions from a regional model Global Planet Change 63 105ndash111doi101016jgloplacha200709006 2008

Vecchia A V A unified approach to probabilistic risk assessmentsfor earthquakes floods landslides and volcanoes Proceedingsof a multidiscipilinary workshop held in Golden Colorado 16ndash17 November 1999 Open-File Report 01-324 US GeologicalSurvey 2001

Vinchon C Aubie S Balouin Y Closset L Garcin MIdier D and Mallet C Anticipate response of climatechange on coastal risks at regional scale in Aquitaine andLanguedoc-Roussillon (France) Ocean Coast Manag 52 47ndash56doi101016jocecoaman200809011 2009

Woppelmann G Miguez B M Bouin M N and Altamimi ZGeocentric sea-level trend estimates from GPS analyses at rel-evant tide gauges world-wide Global Planet Change 57 396ndash406doi101016jgloplacha200702002 2007

Yalcin A GIS-based landslide susceptibility mapping using an-alytical hierarchy process and bivariate statistics in Ardesen(Turkey) Comparisons of results and confirmations Catena 721ndash12doi101016jcatena200701003 2008

Yates M L Le Cozannet G and Lenotre N Quantifying errorsin long-term coastal erosion and inundation hazard assessmentsJournal of Coastal Research Proceedings of the 11th Interna-tional Coastal Symposium Special Issue 64 260ndash264 2011

Yates-Michelin M Le Cozannet G Krien Y and Lenotre NAmelioration de la methode RNACC caracterisation des incer-titudes relativesa la quantification des impacts de lrsquoelevation duniveau marin [improving the RNACC method characterizationof uncertainties in the quantification of the impacts of sea levelrise] rapport final BRGMRP 59405-FR 2011

Yin J Yin Z Wang J and Xu S National assessment ofcoastal vulnerability to sea-level rise for the Chinese coastJ Coast Conservation 16 123ndash133doi101007s11852-012-0180-9 2012



mapping it is far from straightforward (eg Preston et al2011)

In shallow coastal waters waves and currents can causesediment transport and bathymetry changes (ie sea floor to-pography) which in turn affect waves and currents (eg Cocoand Murray 2007) On time scales longer than several weeksall of these processes cannot be modelled with sufficient ac-curacy for long-term analysis Therefore longer term studiesattempt to quantify sediment transport processes in a sim-plified way either in the cross-shore andor longshore direc-tion (Hanson et al 2003) as gradients in longshore sedimenttransport may be the result in erosion or accretion It is worthnoting that these morphosedimentary dynamic processes arealso important for anticipating potential future marine sub-mersion because nearshore waves and sea level depend onoffshore atmospheric and marine conditions as well as ofthe bathymetry

However it is still possible to appraise long-term mor-phosedimentary dynamics by analysing a number of coastalfeatures including coastal geomorphology (eg Cooper andJay 2002) exposure to waves and currents local sedimen-tary budgets (potentially affected by human activities) andrelative sea level variations At regional scales Gornitz etal (1991) developed an approach in which data describingthese components of physical vulnerability assessments areaggregated The resulting coastal vulnerability index (CVI)originally took into account the following data historicalerosion rates nearshore slope tidal amplitude observationsand predictions of sea level variations wave climate andcoastal geomorphology (Gornitz et al 1991) Some adapta-tions of the CVI benefited from other data and put more em-phasis on for example coastal subsidence as a component ofrelative sea level rise (eg Doukakis 2005) Beyond qualita-tive assessments of coastal vulnerability (eg Paskoff 2004Lebbe et al 2008 Alpar 2009) the challenge in mappingcoastal vulnerability is to combine all of this heterogeneousinformation into one single map in a structured way and toappropriately integrate the opinion of experts when data areinsufficient to directly solve the problem (eg Fairbanks andJakeways 2006 Vinchon et al 2009 Hanson et al 2010)

In this study we describe a new method for mappingcoastal vulnerability that incorporates expert analysis andqualitative and quantitative data in a more structured wayThe method is based on a multi-criteria decision-mappingmethod (Malczewski 2006 Chakar 2006) ndash the analyti-cal hierarchy process (AHP Saaty 1980 part 2) We in-tegrate AHP into a complete coastal physical vulnerabilitymapping method and apply it in two regions of France theisland of La Reunion in the south-western Indian Ocean andLanguedoc-Roussillon in the Mediterranean (part 3) We pro-vide new maps of coastal vulnerability in the two regions asa result (part 4) We discuss the advantages limitations andtransportability of the proposed method and investigate itspotential to be integrated into emerging adaptation practicesin France (part 5)

Table 1The fundamental scale of absolute numbers (ldquoSaaty scalerdquoadapted from Saaty 2008a)

Intensity ofimportance

Definition

1 Equal importance of both components

3 Judgement slightly favours onecomponent over another (moderatedifference of importance)

5 Judgement strongly favours onecomponent over another (strongdifference of importance)

7 Very strong or demonstratedimportance of one componentwith respect to another

9 Evidence of extreme difference ofimportance of one componentwith respect to another

2 The AHP process

21 The AHP process

The analytical hierarchy process (Saaty 1980) is a multi-criteria decision analysis method that solves decision-makingproblems by ranking alternatives according to several crite-ria The main steps in the method involve (Saaty 2008a)

ndash representing a decision-making problem by organisingits criteria into a hierarchical structure

ndash evaluating the relative importance of the criteria (withina hierarchical tree) and then the alternatives with re-spect to each criterion this is done by constructing pair-wise comparison matrices of criteria and alternatives re-garding each criterion For example a simple decisionproblem with 3 criteria and 5 alternatives will requireone 3times 3 comparison matrix to be built for the crite-ria weighting and three 5times 5 comparison matrices forevaluating the alternative criteria and

ndash synthesising the analysis by calculating the weightedsum of scores for alternatives based on the weights oftheir parent nodes in the hierarchical tree

The pairwise comparisons of criteria are made using a scaleof absolute judgements commonly called Saatyrsquos scalewhich indicates by how much one item predominates overanother with respect to a given attribute This scale trans-forms qualitative evaluations into numerical values from 1to 9 (Table 1) which are used to fill in the pairwise compari-son matrices The priority scales are then derived by calculat-ing the eigenvector associated with the principal eigenvalueof each comparison matrix (Saaty 1980) When the priority








Z1 1 5 9 07352Z2 15 1 5 02067Z3 19 15 1 00581















33

820

821


maps 823

824





























Hydrographicnetwork


BD Carthage







312 Available data






34

825



828

829




















32 AHP application






State factors




35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References
















































































Z1 1 5 9 07352Z2 15 1 5 02067Z3 19 15 1 00581















33

820

821


maps 823

824





























Hydrographicnetwork


BD Carthage







312 Available data






34

825



828

829




















32 AHP application






State factors




35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References











































































33

820

821


maps 823

824





























Hydrographicnetwork


BD Carthage







312 Available data






34

825



828

829




















32 AHP application






State factors




35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References































































































Hydrographicnetwork


BD Carthage







312 Available data






34

825



828

829




















32 AHP application






State factors




35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References











































































34

825



828

829




















32 AHP application






State factors




35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References






















































































32 AHP application






State factors




35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References











































































35

830





835








36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References











































































36

836















La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References


















































































La Reunion island






tan(α) (1)


Marine factors


Context factors











324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References














































































324 Uncertainties


4 Results














La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References
















































































La Reunion




37




844

845








in Table 6 853

854











Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References















































































Medium



Very high




Medium



Low to medium


Medium



38

855








39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References











































































39

860



863







5 Discussion















522 Lack of data































6 Conclusions






References



















































































522 Lack of data































6 Conclusions






References






































































































6 Conclusions






References










































































An AHP-derived method for mapping the physical ... · G. Le Cozannet et al.: An AHP-derived method...

Documents

Transcript of An AHP-derived method for mapping the physical ... · G. Le Cozannet et al.: An AHP-derived method...