Swell impact on reef sedimentary processes: A case study of theLa Reunion fringing reef

EMMANUEL CORDIER*,� EMMANUEL POIZOT� and YANN MEAR�*IRD/ECOMAR, Universite de La Reunion, 15 avenue Rene Cassin, 97715 Sainte-Denis cedex 9,La Reunion, Indian Ocean, France. (E-mail: emmanuel.cordier@univ-reunion.fr)�Laboratoire Geosciences Reunion, Universite de la Reunion, Institut de Physique du Globe de Paris,Sorbonne Paris-Cite, CNRS UMR 7154, 15 avenue Rene Cassin, 97715 Saint-Denis cedex 9, La Reunion,Indian Ocean, France�Geoceano Group, Laboratoire Universitaire Sciences Appliquees Cherbourg (LUSAC), EA 4253,Cherbourg, France

Associate Editor – Bernhard Riegl

ABSTRACT

Two surface-sediment sampling campaigns were carried out in November and

December 2003, before and after a strong swell event, in the back-reef area of a

microtidal fringing reef on the western coast of La Reunion, Indian Ocean. The

spatial distributions of the mean grain size, sorting and skewness parameters are

determined, and grain-size trend analysis is performed to estimate the main

sediment transport pathways in the reef. The results of this analysis are compared

with hydrodynamic records obtained in the same reef area during fair weather

conditions and during swell events. Sediment dynamics inferred from the

hydrodynamic records show that significant sediment erosion and transport occur

only during swell events and under strongly agitated sea states. Under normal

wave conditions, there is a potential for onshore sediment transport from the reef-

flat to the back-reef, but this transport is episodic and occurs principally during

high-tide stages. Sediment transport trends revealed by the grain-size trend

analysis method show onshore and alongshore low-energy transport processes

that are in agreement with the hydrodynamic records. The grain-size trend

analysis method also provides evidence of an offshore high-energy transport trend

that could be interpreted as a real physical process associated with return flow

from the shore to the reef. The impact of swell on the reef sediment dynamics is

clearly demonstrated by onshore and alongshore transport. Considering different

combinations of the vector transport trends computed through the grain-size trend

analysis approach, more realistic and pertinent results can be obtained by

applying an exclusive OR operation (XOR case) on the vectors. The main results

presented here highlight a trend towards the accumulation of carbonate sands in

the back-reef area of the fringing reef. These sediments can only be resuspended

during extreme events such as storms or tropical cyclones.

Keywords Fringing reef, GSTA, La Reunion, sedimentary dynamics, swellimpact.

INTRODUCTION

The main agent of sediment transport in coral reefecosystems is considered to be linked to wavetransformation and interactions between the fore-

reef, the reef-flat and the back-reef lagoon, basedon a two-dimensional (ocean to lagoon) morpho-logy (Roberts, 1989; Gourlay, 1994, 1996a,b;Kench, 1998; Brander et al., 2004). In shallowand narrow fringing reef environments, the water

Sedimentology (2012) 59, 2004–2023 doi: 10.1111/j.1365-3091.2012.01332.x

2004 � 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists

circulation is driven by wind and waves incombination with the tidal cycle. The way theincident wave is transformed depends on theoverall reef geometry, the width of the shallowreef-flat, the uniformity of depths along andacross the reef and, most importantly, the waterdepth of the reef crest (Lugo-Fernandez et al.,1998, 2004; Storlazzi et al., 2006; Storlazzi & Jaffe,2008).

Sedimentation in reef environments is recog-nized as contributing to the control of coral reefdistribution and growth rates (Hubbard et al.,1981; James & Bourque, 1992; Bruggemann et al.,1996; Anthony, 2000; Larcombe et al., 2001;Chazottes et al., 2008), while high turbidity atten-uates the light, thus reducing the photosyntheti-cally available radiation (Dodge & Vaisnys, 1977;Ogston et al., 2004). As a result, areas with highsediment accumulation rates have relatively fewcoral species, less live coral and reduced netproductivity (Rogers, 1990; Te, 1997). Most stud-ies of sedimentary dynamics in reef environmentshave focussed on terrestrial sediment runoff andits impact on coral health (Storlazzi et al., 2004,2009; Presto et al., 2006).

On the contrary, the fringing reef of the Ermitage-La Saline, off the coast of La Reunion Island,receives no major terrigenous inputs and sedi-mentation appears to be rather poorly marked(Naım, 1993). In addition, this fringing reef showsa low coral diversity and a small area of live coralassociated with increasing algal cover (Naımet al., 2000). The few studies that have focussedon the sedimentology of the Ermitage-La Salinefringing reef described the compositional andgrain-size characteristics of the sedimentarydeposits (Montaggioni, 1978; Gabrie & Montaggioni,1982; Chazottes et al., 2002, 2008). Despite theseprevious studies, there is still a lack of knowledgeon the distribution and dynamics of sediments inthe back-reef areas of the fringing reefs around LaReunion. The present study was carried out tofill this gap and to provide an overall view of thespatial and temporal variability of bottom-sedi-ments and their dynamics. The focus here isplaced on the back-reef area of the Trou d’Eausite, in the southern part of the La Saline reef.

Surficial sediments were collected during twofield campaigns, before and after an annual swellevent causing strong currents and high turbidity(offshore significant wave height exceeding 1Æ5 m,with a period of 15 sec). Mapping was carriedout, as well as a comparison of results from thetwo sediment sampling campaigns to assess thespatial distribution of carbonate sediments in

the back-reef and its modification after a signifi-cant swell event. Hydrodynamic measurementsin the same area were then used to estimate thethreshold of motion of carbonate sediments andassess the sediment dynamics under fair weatherand swell conditions.

Because the two carbonate sediment datasetswere representative of two well-identifiedenvironmental forcings, it was possible to usegrain-size trend analysis (GSTA), as defined byMcLaren (1981) and Gao & Collins (1991), andmodified by Poizot et al. (2006) and Poizot &Mear (2010), to assess the sediment transportpathways. A conceptual model of the sedimentdynamics in the Trou d’Eau fringing reef is thenproposed, which describes sediment characteris-tics and transport trends in a fringing reef undertwo different conditions of sea state – calm andagitated. In this conceptual model, the differentcases of sediment transport vector trends com-puted by the GSTA method are discussed.

STUDY AREA

The Ermitage-La Saline fringing reef is located onthe western coast of La Reunion Island, in theWestern Indian Ocean (Fig. 1A and B). It is thelargest reef of La Reunion, making up 48% of thetotal surface area of the fringing reef. It extendsfor about 9 km alongshore, with a width varyingbetween 100 m and 600 m (Fig. 1C). It is con-nected to the open ocean by two inlets corre-sponding to river mouths: the ‘Ermitage Pass’ inthe middle of the fringing reef and the ‘TroisBassins Pass’ in the south.

The study area is in the southernmost part ofthe fringing reef, at a locality known as ‘Troud’Eau’, with a maximum width of 500 m and acoastline length of roughly 1Æ3 km (Fig. 1D).Three geomorphological zones are identified fromthe shore to the open ocean, namely the back-reef,the reef-flat and the reef crest (Montaggioni &Faure, 1980). The back-reef zone is at a waterdepth of 1Æ2 m below mean sea-level, and isessentially made up of biodetrital sands, occupiedby some scattered coral colonies (Montaggioni,1978; Camoin et al., 1997). The reef-flat has awater depth of about 0Æ30 m below mean sea-leveland is exposed at low tide. It can be subdividedinto an inner reef-flat occupied by branchedcorals and an outer reef-flat made of compactcoral colonies and encrusting algae. The reef crestrepresents the outer limit of the reef-flat and thetransition with the open ocean.

Swell impact on fringing reef sedimentary processes 2005

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

The tide propagates in a northward directionand is of mixed semi-diurnal type, showing amicrotidal regime with spring and neap ranges of0Æ90 m and 0Æ10 m, respectively. Two kinds ofswell dominate the offshore waves. The first typeis generated by the predominant south-easterly

trade-winds which prevail throughout the yearand are characterized by a short significant period(5 sec < T sec <10 sec) and significant heights (Hs)rarely exceeding 2 m. The second type occursonly during winter (May to October) and rarelylasts more than 48 h. The waves are characterized

36°E 42°E 48°E 54°E 60°E

28°S

24°S

20°S

16°S

12°S

LA REUNION

Mauritius

MAD

AGAS

CAR

Moz

ambi

que

Curre

nt

Agul

has

Curre

ntSouthEquatorialCurrent

East

Mad

agas

car

Curre

nt

AFRICA

55°E 12’ 24’ 36’ 48’ 56°E

20’

10’

21°S

50’

1000

1000

2000

1000

2000

Saint-GillesReef

Wind record station, Trois BassinsWave record station, Le Port

14·20’ 14·30’ 14·40’ 55°E 14·50’

14·60’ 14·70’ 14·80’

6·30’

6·20’

21°S6·10’

6·00’

5·90’

0 200 400 m

Sediment samplesBefore swellAfter swell

Hydrodynamic stations

13’ 55°E 14·00’

15’ 7’

6’

21°S 5·00’

4’

50

150

250

0 1 km

Reef Flat

Trois BassinsInlet

ErmitageInlet

Saint-GillesHarbor

A B

C

INDIANOCEAN

BR2

BR1

BR3

RF3

RF2

RF1

DD

Reef Flat

Back Reef

Reef Crest

Trou d'Eau Reef area

T1

T2

T4

T5

T6

T3

Fig. 1. Study area, showing: (A) location of La Reunion; (B) location of the Saint-Gilles fringing reef, showing thelocation of the meteorological stations; (C) overall view of the Saint-Gilles fringing reef; and (D) the Trou d’Eau studyarea showing the sediment sampling points. Hydrodynamic measuring stations are indicated by white triangles.

2006 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

by longer periods (12 sec < T sec < 20 sec) andhigher significant heights (Hs ca 3 m, sometimesreaching 8 m), propagating from the south-west(Soler, 1997).

MATERIALS AND METHODS

Sediment sampling and meteorologicalconditions during sampling

The sampling was carried out in late Novemberand December 2003 and was implemented on sixglobal positioning system referenced cross-shoretransects, with a spacing of 200 m, crossing theback-reef of the La Saline fringing reef (Fig. 1D).Along each transect, surface sediment sampleswere collected at approximately 20 m intervals.Samples were collected by hand using a PVC tubeto skim the substrate, providing 300 g of sedimentfor grain-size analysis. A total of 130 sampleswere retrieved. The sampling was conducted intwo sessions, 27/28 November 2003 and 8/9December 2003, before and after a swell eventthat induced strong hydrodynamics in the fring-ing reef, along with high turbidity.

Figure 2 illustrates the meteorological condi-tions between the two sampling periods, showingthe time series of wind velocity, significant waveheight, period (see Fig. 1B for location of waveand wind recording stations) and the wave poweridentified as a good indicator of the swell energyreaching the reef. It is calculated using the deepwater linear wave theory (Dail et al., 2000):

F ¼ ðq � g2 �H2 � TÞ=ð32 � pÞ ð1Þ

where H and T are the significant wave height andperiod, q is the water density and g is the accel-eration due to gravity. The swell event between thetwo sampling periods lasted three days and wascharacterized by a maximum offshore wave energyof 3Æ104 J.m)1sec)1, with H = 1Æ5 m and T = 15 sec.During this period, no persistent wind was notice-able (wind speed <3 m sec)1), suggesting that thehydrodynamic conditions in the reef and theassociated sediment transport processes wereforced only by the oceanic swell.

Sediment analysis

Chazottes et al. (2008) showed that 98% of thesedimentary cover of the studied area is com-posed of medium and coarse carbonate sands.Classical sieve analysis was then performed using

sand class to assess the granulometric parameters.Sediment samples were dried at 50�C for 48 h anddry-sieved over the entire sand class from 2Æ5 to0Æ05 mm for 15 min. A granulometric scale basedon the metric system was chosen so as to yield a /interval of nearly 1/3. The mean grain-size,sorting and skewness parameters were calculatedusing the moment method. Grain-size parameterswere then expressed as / units in order to carryout the grain-size trend analysis.

Spatial distribution of granulometricparameters

The granulometric parameters were mapped on aregular grid to assess the spatial distribution ofcarbonate sands in the back-reef area before andafter the significant swell event. As the distancebetween sediment samples was greater betweentransects than along a transect, the points on theregular grid needed to be spaced by about 30 m inorder to keep the density of points at its initiallevel. A total of 230 grid points were finallyselected to form the grid on which granulometricparameters could be interpolated using a krigingmethod (Poizot et al., 2006). A geostatisticalapproach for interpolation was chosen to beconsistent with the grain-size trend analysisdescribed below.

Hydrodynamic measurements

A recent study focused on hydrodynamic pro-cesses and water circulation in the fringing reef ofLa Saline (Cordier, 2007). As part of this study,current speeds were measured on the Trou d’Eaureef in April, June, October and November 2004(Table 1). Some of these data are integrated intothe present paper in order to document thehydrodynamic conditions of the studied site.

The currents were measured on three transectsduring 24 h periods using a 1 min continuousrecording averaged every 10 min. Burst records at1 Hz and 2 Hz were also performed each hour tomeasure the high-frequency variability of sea-water elevation and flow. The measurementswere performed with a bottom-mounted Nortekacoustic Doppler current meter and a Nortekacoustic Doppler current profiler (Nortek AS,Rud, Norway), positioned on two sites for eachsurveyed transect, that is, in the back-reef zoneand in the inner reef-flat, respectively. The cur-rent meter was programmed so the cell size couldbe adjusted to measure velocities over amaximum depth interval in the water column,

Swell impact on fringing reef sedimentary processes 2007

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

without interference with the surface. The valueobtained in this way is assumed to be a goodestimate of the depth-averaged current velocity.The current profiler was programmed to ensure avertical profile with a minimum of five cells. Asthe velocity of water flowing in shallow water isnot constant throughout a vertical profile, themost common way to compute depth-averagedcurrent speed is related to the velocity profile u(z)through the definition (Soulsby, 1997):

U ¼ 1

h

Zh

0

u zð Þdz ð2Þ

where h is the total water depth, and u(z) is thecurrent velocity recorded at depth z.

Measurements were carried out during a total of10, 24 h periods, which allowed the main phys-ical forcing factors driving water flow in the reefto be characterized (Table 1). All transects were

23Nov 2003

25 27 29 1Dec 2003

3 5 7 9 11 13 150

2

4

6

Wav

e p

ow

erJ.

m–1

.s–1

x 1

04

0

5

10

15

20

Wav

e p

erio

ds

0

0·5

1

1·5

2

Wav

e h

eig

ht

m

0

2

4

6W

ind

sp

eed

m.s

−1A

B

C

D

23Nov 2003

25 27 29 1Dec 2003

3 5 7 9 11 13 15

23Nov 2003

25 27 29 1Dec 2003

3 5 7 9 11 13 15

23Nov 2003

25 27 29 1Dec 2003

3 5 7 9 11 13 15

Fig. 2. Meteorological conditions during sampling: (A) wind velocity; (B) offshore significant wave height; (C) off-shore significant wave period; and (D) offshore significant wave energy. The two periods of sampling are indicated bygrey shading.

2008 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

recorded at least once under fair weather condi-tions. Four periods were characteristic of moreagitated sea states where the reef circulation wasdriven by significant waves or strong winds. Foreach meteorological condition, the mean flowpattern was computed, as well as the princi-pal components of the currents. The low andhigh-frequency variability of the alongshore andcross-shore components of the flow were com-pared with the sea-level variations to assess thelink between the tidal stage and hydrodynamicprocesses in the reef.

Grain-size trend analysis

Fundamental assumptionThe grain-size trend analysis (GSTA) methoduses differences in grain-size statistical parame-ters derived from granulometric distributions ofbottom sediment samples to infer net sedimentpathways and regions of erosion, accretion anddynamic equilibrium (McLaren, 1981; McLaren &Bowles, 1985). The basic assumption is thatdifferences in sediment grain-size distributionscan be linked to sediment transport. According tothe McLaren model, two patterns can be distin-guished which characterize sediment transportbetween two points. In the downstream direction,sediments can become either coarser, bettersorted and more positively skewed (CB+ case),or finer, better sorted and more negatively skewed(FB– case).

Gao & Collins (1991, 1992) improved themethod by developing a 2D model to determinetrend vectors on a grid of sampling sites bycomparing the grain-size parameters of neigh-bouring samples. Neighbouring sampling sites

were identified on the basis of a characteristicdistance that represents the spatial scale of sam-pling. For each sample site where a trend can bedefined, a dimensionless trend vector of unitlength is drawn, in the direction of its neighbour.

More recently, Poizot et al. (2006, 2008) andPoizot & Mear (2010) developed GiSedTrend (seeSupporting Information), a plugin of the QGIS freegeographical information system (GIS) software. Itimplements the GSTA approach with the use of thestatistical software package ‘R’ and the ‘geoR’package (Paulo & Diggle, 2001).

Step 1 of GSTA: Creation of a regular gridAs noted by Le Roux (1994a,b), sampling sitescan rarely be collected on a regular grid pattern(the same distance between sampling stations).When the sampling scheme shows many discon-tinuities, edge effects may be very noticeable andbecome dominant in some parts of the surveyarea (Rıos et al., 2003; Friend et al., 2006; Poizotet al., 2006; Plomaritis et al., 2008). To avoid thiseffect, the samples were interpolated on a regulargrid with a constant cell resolution of 30 m,using an ordinary kriging method. This grid isthe same as the one presented in the precedingsection detailing spatial distribution of granulo-metric parameters.

Step 2 of GSTA: Determination of characteristicdistanceThe grain-size parameters of each sampling sta-tion are compared with those of the neighbouringstations. The adjoining stations are identified by acharacteristic distance which is generally taken asthe maximum sampling interval (Gao & Collins,1992, 2001; Gao, 1996). This distance is a key

Table 1. Period of hydrodynamic measurements and associated meteorological conditions. The bold text highlightsthe strong meteorological conditions associated with an agitated sea-state in the reef.

Stations DatesMean offshorewave height Hs (m)

Mean windspeed (m.sec)1)

Maximum windspeed (m.sec)1)

Meteorologicalconditions

BR1 and RF1 2004-10-20 0Æ49 2 3 Calm weather2004-10-28 0Æ48 1Æ7 3Æ1 Calm weather2004-11-12 0Æ60 2Æ7 5Æ7 Calm weather

BR2 and RF2 2004-05-26 1Æ13 5 11 Agitated sea2004-06-02 1Æ33 3Æ4 11Æ6 Agitated sea2004-06-08 0Æ43 2Æ4 4Æ2 Calm weather2004-06-16 0Æ33 1Æ5 2Æ9 Calm weather

BR3 and RF3 2004-04-26 0Æ91 6Æ6 12 Agitated sea2004-05-03 0Æ53 1Æ8 4Æ3 Calm weather2004-05-18 1Æ01 8 13Æ3 Agitated sea

Swell impact on fringing reef sedimentary processes 2009

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

parameter for computing trend vectors and itschoice depends on the sampling interval, whichitself results from previous knowledge of envi-ronmental factors over the studied area (sedimentsources, sediment type, hydrodynamics, etc.) aswell as economic considerations. To overcomethese constraints Poizot et al. (2006) used ageostatistical approach based on the analysis ofsemi-variograms to define the characteristicdistance. The study of the semi-variogram forthe three granulometric parameters allowed adetermination as to whether or not there is aspatial correlation. The sill reached describes themaximum of variance above which the samplesare spatially independent (Fig. 3). The distance atwhich this occurs is then used to determine thecharacteristic distance, denoted ‘Dg’ (geostatisti-cal distance), taken here as 170 m (Fig. 3). Thisdistance takes into account the spatial structureof the granulometric data, based on the under-lying physical mechanisms controlling the sedi-ment distribution.

Step 3 of GSTA: Computation of vector trendsThe vector trends are computed by comparingadjoining stations. From a central grid point, aunit vector is drawn in the direction of eachneighbour (the points distant of Dg or less) if theCB+ case, the FB– case and the CB+/FB– com-bined case of transport trend are encountered(McLaren & Bowles, 1985; Gao & Collins, 1992).The exclusive ‘OR’ operation (XOR) as introduced

by Poizot & Mear (2010) was also applied to thesame datasets, while considering the two trendcases CB+ and FB–. Using GiSedTrend allows thevector trends computation to take into accountthe complex morphology of the studied site. Incase of a morphological obstacle that wouldprevent sediment transport (for example, thepresence of a reef-flat) between two neighbouringstations, the vector trend is not computed in thatdirection, allowing a more realistic response ofthe method.

The computed trend vector maps show vectorsof different relative magnitudes. This variationresults from the summation of the unit vectorscomputed in one point. When unit trend cases arefound globally in the same direction away fromthe studied sample point, the resulting vector hasa higher modulus. On the contrary, when unittrend cases are found in many different directionsaround the studied sample point, the resultingvector has a smaller modulus. In any case thismodulus encodes the underlying sediment trans-port rate but it highlights the degree of confidencethat can be applied to the computed resultanttrend vector; the higher the magnitude of thevector, the better the reliability of the computedtrend.

RESULTS AND INTERPRETATIONS

Spatial distribution of sediments

The back-reef area was mainly characterized bycoarse and medium sand, with a poorly to mod-erately sorted and fine-skewed granulometry(Fig. 4). Before the swell event, the sedimentarycover exhibited cross-shore grading, with the finercoarse-skewed sediments near the coast and thecoarser fine-skewed sediments concentrated onthe inner reef-flat. The spatial distribution of thesorting parameter reflected a decreasing gradienttowards the coast. According to these observa-tions, three main compartments (or cells) weredistinguished. A first cell contained transects 1and 2, with finer sands and a coarse-skeweddistribution, spatially arranged according to analongshore gradient. The second larger cell, con-taining transects 3 to 5, occupied two-thirds of thetotal surface area and was characterized by a fineand medium sand patch with offshore coarsening.The spatial distribution of skewness highlighted across-shore gradient with very fine-skewed sedi-ments on the inner reef-flat and coarse-skewednear-shore. The central part of this cell was

Distance (m)

Gam

me

(h)

0·05

0·10

0·15

50 100 150 200 250

Variogram of the mean grain size

Fig. 3. Variogram of mean grain size parameter.

2010 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

−0·

5

00·5

11·5

2

14·

30’

55°

E 1

4·50

’ 1

4·70

’

6·3

0’

6·2

0’

21°

S 6

·10’

6·0

0’

5·9

0’

ME

AN

0·8

11·2

1·4

1·6

1·8

2

14·

30’

55°

E 1

4·50

’ 1

4·70

’

6·3

0’

6·2

0’

21°

S 6

·10’

6·0

0’

5·9

0’

SO

RT

ING

−2

−1·

5

−1

−0·

5

00·5

11·5

2

14·

30’

55°

E 1

4·50

’ 1

4·70

’

6·3

0’

6·2

0’

21°

S 6

·10’

6·0

0’

5·9

0’

SK

EW

NE

SS

−0·

5

00·5

11·5

2

14·

30’

55°

E 1

4·50

’ 1

4·70

’

6·3

0’

6·2

0’

21°

S 6

·10’

6·0

0’

5·9

0’

0·8

11·2

1·4

1·6

1·8

2

14·

30’

55°

E 1

4·50

’ 1

4·70

’

6·3

0’

6·2

0’

21°

S 6

·10’

6·0

0’

5·9

0’

−2

−1·

5

−1

−0·

5

00·5

11·5

2

14·

30’

55°

E 1

4·50

’ 1

4·70

’

6·3

0’

6·2

0’

21°

S 6

·10’

6·0

0’

5·9

0’

500

m

500

m

500

m50

0 m

500

m50

0 m

AF

TE

R S

WE

LL

BE

FO

RE

SW

ELL

ME

AN

SO

RT

ING

SK

EW

NE

SS

AB

C

EF

D

Fig

.4.

Sp

ati

al

dis

trib

uti

on

of

mean

gra

insi

ze,

sort

ing

an

dsk

ew

ness

para

mete

rsbefo

reth

esw

ell

even

t(A

)to

(C)

an

daft

er

the

swell

even

t(D

)to

(F).

Swell impact on fringing reef sedimentary processes 2011

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

covered by sediments with a symmetrical grain-size distribution. Finally, the third compartmentwas represented by transect 6, with mediumsands and quasi-symmetrical grain-size distribu-tion in the near-shore zone.

After the swell event the extent of the firstcompartment remained identical, but theobserved gradients became cross-shore (Fig. 4D).The fine-skewed sediments located at the innerreef-flat limit were more abundant, showing astronger skewness gradient shoreward (Fig. 4F).

The configuration of compartments 2 and 3 wasmodified by a northward migration of the finesand patch by about 200 m. The second compart-ment only included transect 3, and was charac-terized by a patch of coarse sands (Fig. 4D); thiscould be explained either by an input of coarsermaterial coming from the reef-flat, or by theloss of a finer fraction. The third compartmentincluded the area between transects 4 and 6, andwas characterized by a near-shore fine sand patchwith a cross-shore gradient steeper than beforethe swell event. The studied swell event did notdrastically modify the spatial distribution of thesediments, but was responsible for a slight shift of

the identified sedimentary cells along the shoreaxis.

Grain-size distributions

Changes in the sedimentary cover were alsoreflected in the grain-size distributions of eachsample (Fig. 5). In the near-shore, before the swellevent, most of the distributions were bimodal,either with a pronounced mode in the fine sands(Fig. 5A and E), or with two identical modes inthe fine and coarse sands (Fig. 5C and G). Theswell event caused strong modifications of thedistributions for transects 1 and 6, highlighted byan important gain of fine sands. Transects 3 and 4were modified with a loss of fine sands and a gainof coarser material; this suggested a transport ofsands parallel to the shore, from the central partof the studied area towards its southern andnorthern extremities.

In the inner reef-flat, the distributions before theswell event were unimodal in the coarse sands.The impact of the swell caused fewer modifica-tions than in the near-shore. The distributions inthe extremities of the studied area showed a gain

10

20

30

10

20

30

−1 0 1 2 3 40

10

20

30

Sand size (phi)

Wei

ght r

etai

ned

(%)

−1 0 1 2 3 4Sand size (phi)

Before swellAfter swell

A B

C D

G H

E F

10

20

30

Near-shore Inner reef-flat

Transect 1 Transect 1

Transect 3

Transect 4

Transect 6 Transect 6

Transect 4

Transect 3

10

20

30

10

20

30

0

10

20

30

Wei

ght r

etai

ned

(%)

10

20

30

Fig. 5. Granulometric distributions of samples collected near-shore and on the inner reef-flat, before (white) andafter (black) the swell event: (A) transect 1 near-shore and (B) inner reef-flat; (C) transect 3 near-shore and (D) innerreef-flat; (E) transect 4 near-shore and (F) inner reef-flat; (G) transect 6 near-shore and (H) inner reef-flat.

2012 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

of coarse sand and a loss of fine sands (Fig. 5B andH). The distributions in transect 3 (Fig. 5D)exhibited a low gain of medium and fine sandsand a loss of coarse sands. Transect 4 was notsignificantly affected by the swell event (Fig. 5F).

Circulation patterns

General flow patterns of circulationDuring fair weather conditions (Fig. 6A), thereef-flat current ellipses highlighted a mean flowweaker than its variability, implying frequentreversals. In the back-reef, the current was morecommonly unidirectional with a mean flowstronger than its variability. At station BR2 adivergence of the flow was characterized by amean cross-shore current of 2Æ1 cm sec)1 direc-ted offshore. The orientation of the meancurrent for BR1 and BR3 was constrainedby the north and south inlets, respectively,located several kilometres away at the end ofthe reef.

During severe weather conditions (Fig. 6B), thewater flow was mainly directed onshore over thereef-flat, with a mean current flow stronger thanits variability. The BR2 and RF2 stations weremeasured during swell forcing, and showedstronger currents on the reef-flat (U = 15Æ1cm.s)1) than in the back-reef (U = 1Æ3 cm.s)1).The low mean current in the back-reef was dueto frequent flow reversals that reduced the netwater flux. The BR2 station always showed adiverging flow, with an enhanced magnitude

caused by the subsequent mass of water enteringthe reef due to wave breaking. On the contrary,records from the BR1 and RF1 stations wereobtained during wind forcing, and highlightedweaker currents on the reef-flat (U = 10Æ9 cmsec)1) than in the back-reef (U = 18Æ4cm sec)1).This could be due to the shallowness of the reef-flat inducing stronger bottom friction, thus pre-venting the generation of local wind-waves andwind-driven flow (Brander et al., 2004; Cordier,2007).

Low frequency variability of flowThe general pattern of water circulation wasconfirmed by the 24 h time series record of thesea-level variations, the cross-shore and thealongshore current components at the BR2 andRF2 stations (onshore and northward componentsof the flow were identified by positive values).Cross-shore, BR2 was affected by an offshorecurrent during fair weather conditions, while RF2showed a net onshore current (Fig. 7B). Along-shore, many flow reversals highlighted the diverg-ing circulation at BR2 and RF2 (Fig. 7C). The flowreversals seemed to be unrelated to the tidalvariation (Fig. 7A), suggesting that the tideplayed a minor role in the mean circulationpattern on the inner reef. A similar result washighlighted by Coronado et al. (2007) in a narrowand shallow Caribbean fringing reef. The situa-tion remained almost the same during severeweather conditions, but with strengthened cur-rents (Fig. 7D to F).

14·20’ 14·30’ 14·40’ 55°E 14·50’

14·60’ 14·70’ 14·80’

6·30’

6·20’

21°S 6·10’

6·00’

5·90’

0 200 400 m

10 cm.s–1

Fair weather conditions

BR1

BR2

BR3

RF1

RF2

RF3

RF1

RF2 BR1

BR2

14·20’ 14·30’ 14·40’ 55°E 14·50’

14·60’ 14·70’ 14·80’

6·30’

6·20’

21°S 6·10’

6·00’

5·90’

0 200 400 m

20 cm·s–1

Strong wave events

RF1

RF2

BA

Fig. 6. Hydrodynamic records in the Trou d’Eau area. Ellipses indicate the direction of the major and minor semi-axes of variability, the arrow indicates the mean current vector and the grey ticks represent the complete set ofcurrent data.

Swell impact on fringing reef sedimentary processes 2013

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

High-frequency variability of the flowThe high-frequency variability of the flow(Figs 8 and 9) was mainly representative ofcirculation due to wave processes during fairweather conditions and stronger sea states.During fair weather conditions, strong surge-currents were characterized by an acceleration ofthe flow lasting a few seconds. At low tide, thecurrent speed varied between ca )20 cm sec)1

(offshore and southward) and ca 20 cm sec)1

(onshore and northward on Fig. 8B and C). Athigh tide, extreme values were more frequent inthe back-reef and exceeded 40 cm sec)1 in bothdirections (Fig. 8D and E). Although the tideappeared to have a minor influence on the meanflow, the strength of the surge-currents wasgreater during high-tide stages. These mecha-

nisms remained similar in the reef-flat (Fig. 8Gto J).

For stronger sea states, the cross-shore surge-currents in the back-reef at high tide were inten-sified and lasted longer, from several tens ofseconds up to a minute (Fig. 9B). Reversals weremore frequent, occurring about every minute,with stronger current velocities sometimes reach-ing 75 cm sec)1 in absolute terms. The alongshorecomponent also underwent flow reversals, alter-nating between a southward and a northwarddirection, with current intensities rarely exceed-ing 40 cm sec)1 in either direction. At low tide(Fig. 9D and E), the surge-currents were weaker,enhancing the close link between sea-level andthe amount of energy transferred from the openocean. Similar results were obtained on the inner

09 14 19 00 04

1·3

1·4

1·5

1·6m

Tide

09 14 19 00 04

−0·2

−0·1

0

0·1

0·2

m.s

−1

Cross−shore velocity

09 14 19 00 04

−0·2

0

0·2

Hours starting on 08−Jun−2004, 08:30:00

m.s

−1

Alongshore velocity

09 14 19 00 04

1·4

1·6

1·8

2

2·2

m

Tide

09 14 19 00 04

−0·5

0

0·5

m.s

−1

Cross−shore velocity

09 14 19 00 04−0·4

−0·2

0

0·2

0·4

Hours starting on 02−Jun−2004, 08:30:00

m.s

−1

Alongshore velocity

STRONGER WAVE EVENTFAIR WEATHERA

B

C

D

E

F

Backreef Reef−flat

Backreef Reef−flat

Backreef Reef−flat

Backreef Reef−flat

Fig. 7. 24 Hours hydrodynamic records of sea-level variation, cross-shore and alongshore flow velocities, for stationsBR2 (back-reef in blue) and RF2 (reef-flat in red), under fair weather conditions (A) to (C) and for a stronger waveevent (D) to (F). Positive values of the cross-shore component of the flow are directed towards the shore, whilepositive values of the alongshore component refer to a northward flow.

2014 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

reef-flat, with onshore surge-currents stronger athigh tide and reaching 1 m sec)1.

For the two other stations (BR1-3 and RF1-3),similar surge-currents were identified during fairweather as well as during agitated sea states. Asevidenced by numerous authors (Roberts, 1989;Gourlay & Colleter, 2005; Presto et al., 2006;Coronado et al., 2007; Taebi et al., 2011) thegeomorphology of the narrow and shallow fring-ing reef constrained the water circulation. In thiscase, the proximity of the ‘Trois Bassins Pass’ atthe southern end of the reef induced strongeraccelerations at the BR1 and RF1 stations. Asa result, hydrodynamic conditions are moreenergetic in the southern part of the studied areathan in its northern part.

Implications on sediment transport conditions

Using the sediment characteristics and the resultsof hydrodynamic measurements, the threshold ofmotion for sediments could be estimated usingthe Soulsby (1997) formula:

Ucr ¼ 7ðh=d50Þ1=7 � ½gðs� 1Þd50 � f ðD�Þ�1=2 ð3Þ

with:

f ðD�Þ ¼ ½0�3=ð1þ 1�2D�Þ�þ 0�055½1� expð�0�02D�Þ� ð4Þ

where Ucr is the mean velocity (m sec)1), h is thedepth, d50 is the mean grain size, g is the

1·4

1·6

m

Tide level

0 1 2 3 4 5 6 7 8−0·4

0

0·4

−0·4

0

0·4

−0·4

0

0·4

m.s

−1

BR2Cross−shore H1

0 1 2 3 4 5 6 7 8−0·4

0

0·4

m.s

−1

RF2

0 1 2 3 4 5 6 7 8

m.s

−1

Alongshore H1

0 1 2 3 4 5 6 7 8−0·4

0

0·4

m.s

−1

0 1 2 3 4 5 6 7 8

m.s

−1

Cross−shore H2

0 1 2 3 4 5 6 7 8−0·4

0

0·4m

.s−

1

0 1 2 3 4 5 6 7 8−0· 4

0

0· 4

Time in minutes

m.s

−1

Alongshore H2

0 1 2 3 4 5 6 7 8−0·4

0

0·4

Time in minutes

m.s

−1

H1

H2

Cross−shore H1

Alongshore H1

Cross−shore H2

Alongshore H2

A

D

C

B

G

F

E I

H

Fig. 8. High-frequency variability of flow at BR2 and RF2 under fair weather conditions, on 8 June 2004. The first 8min of a high tide (H1) and a low tide (H2) burst record are shown. Cross-shore (onshore >0) and alongshore(northward >0) are illustrated.

Swell impact on fringing reef sedimentary processes 2015

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

acceleration due to gravity, s = qs/q is the ratiobetween the sea water density q and sedimentdensity qs, D* is a non-dimensional grain diam-eter defined as:

D� ¼ d50½gðs� 1Þ=m2�1=3 ð5Þ

where m is the kinematic viscosity.

The settling velocity of sediment was also com-puted using the following formula (Soulsby, 1997):

Ws ¼ ðm=d50Þ � ½ð10�362þ 1�049D�3Þ1=2 � 10�36�ð6Þ

The sediment density was assumed to be 2Æ0 gcm)3, which appears to be a good estimate for the

carbonate sediments of the Ermitage-La Salinereef. In fact, densities ranging from 1Æ4 to2Æ4 g.cm)3 have been obtained for the sedimentsof the same reef and, by comparison, Kench &McLean (1997) estimated the carbonate sedimentdensity as 1Æ85 g cm)3 for Cocos (Keeling) Island,an atoll in the Indian Ocean.

Following Eqs 3 to 6, the estimated thresholdmean velocity required to erode carbonate sandsranged from 25 cm sec)1 for fine particles (meangrain size of 0Æ063 mm) to 50 cm sec)1 for coarseparticles (mean grain size of 2 mm); this explainsthe low capacity of the mean flow to erodesediments in the back-reef under fair weatherconditions as much as under swell conditions.On the other hand, the settling velocity of thesesame particles is slow, and deposition occurs at

1·5

2

–1

0

1

–1

0

1

–1

0

1

–1

0

1

–1

0

1

–1

0

1

m

Tide level

0 1 2 3 4 5 6 7 8

m.s

−1

m.s

−1

m.s

−1

m.s

−1

BR2

Cross−shore H1

0 1 2 3 4 5 6 7 8–1

0

1

m.s

−1

RF2

0 1 2 3 4 5 6 7 8

Alongshore H1

0 1 2 3 4 5 6 7 8m

.s−

1

0 1 2 3 4 5 6 7 8

Cross−shore H2

0 1 2 3 4 5 6 7 8

m.s

−1

0 1 2 3 4 5 6 7 8Time in minutes

Alongshore H2

0 1 2 3 4 5 6 7 8Time in minutes

m.s

−1

H1

H2

Cross−shore H1

Alongshore H1

Cross−shore H2

Alongshore H2

A

D

C

B

G

F

E I

H

–1

0

1

Fig. 9. Same legend as for Figure 8, but for a stronger wave event on 2 June 2004.

2016 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

velocities of less than 15 cm sec)1 for the verycoarse sands (2 mm), and less than 0Æ5 cm sec)1

for the very fine sands (0Æ063 mm).While the mean flow appeared insufficient to

erode sediments off the substrate, the instanta-neous surge-currents had the capacity to resus-pend sediments temporarily off the bottom, evenin the back-reef or in the inner reef-flat. Thesesediments could be partially transported in thedirection of the current, until the flow slackenedbelow the threshold velocity for deposition.

Under fair weather conditions, a current speedof 5 cm sec)1 was able to transport particles witha maximum size of 0Æ5 mm. However, theseparticles required a minimum speed of 28 cmsec)1 to be eroded, a pre-condition that onlyoccurred during high-tide stages (Fig. 8D and E).Consequently, carbonate sediments always ap-peared to be in a transient state during fairweather conditions, corresponding to a temporarytransport stage while the sea-level dropped to aminimum position (i.e. low-tide periods). Thesediment transport during such stages mainlyoccurred towards the shore on the inner reef-flat,but could be in any cross-shore or alongshoredirection in the back-reef.

Under severe weather conditions, with wave-induced circulation driven by waves breaking atthe reef-crest, a mean current speed of 5 cm sec)1

was associated with frequent surge-currentsexceeding 25 cm sec)1 (Fig. 9B to E). Under suchconditions, the medium sands and smallerparticles were transported easily in suspension.Coarser materials, such as very coarse sands andeventually some coarser fragments, were alsotransported, and these processes occurred mostlyon the sea bed rather than in the water column.

Grain-size trend analysis

Before swell eventThe combined CB+/FB– vectors were mainlyoriented from coast to reef (Fig. 10A). In theinner reef-flat, the sediment transport trendswere progressively modified and showed succes-sive converging and diverging points locatednear the reef. The same three compartmentspreviously identified can be observed here. Thefirst transport cell showed vectors from the reef-flat and the coast converging in the central partof the back-reef area and oriented towards thesouth. The second transport cell was composedof two branches converging at the inner bound-ary of the reef, flowing eastward and southward,respectively. The third transport cell showed

vector trends converging toward the inner reef-flat. The CB+ case alone (Fig. 10B) showed awell-defined pattern of offshore transport trends,without reflecting the presence of transport cells.The FB– case (Fig. 10C) showed a distribution oftransport trends according to two cells concen-trated in the southern part of the studied area.The first cell (transects 1 and 2) was character-ized by onshore and southward transport, with aconvergence of vectors at the coast. The secondtransport cell (transects 3 and 4) reflected anoverall onshore and northward transport trend,converging at the coast near the fourth transect.The area between these two cells was a zone ofdiverging vectors, suggesting a preferential areafor sediment erosion. The small size of thevectors in the north expressed a low significanceof the FB– transport trends in this part of thereef.

Using the XOR operation, an almost balancedspatial distribution of trends was obtained, with55% of CB+ cases and 45% of FB– cases (Fig.10D). CB+ trends dominated in the northern partof the studied area, associated with an offshoretransport direction, while FB– transport trendswere predominant in the south part and directedtowards the south.

After the swell eventThe combined CB+/FB– trend vectors were ar-ranged according to three converging vector setscorresponding to three transport cells, as ob-served before the swell (Fig. 10E). An analysis ofthe separate cases CB+ (Fig. 10F) and FB– (Fig. 10G)showed that the CB+ case was associated withoffshore transport, whereas the FB– case repre-sented onshore transport. The low number of CB+vectors suggested that they were less significantthan the FB– transport trends, confirmed by theapplication of the XOR operation, with 42% ofCB+ and 58% of FB– cases (Fig. 10H).

DISCUSSION

Grain-size parameters derived from skeletal sed-iments have sometimes been used as empiricalindicators of changes in energy of the hydrody-namic regime across reef systems (for instance,see Maiklem, 1968; Braithwaite, 1973; Montag-gioni, 1978). In addition, grain-size trend analy-sis (GSTA) has been applied in many differentsedimentary environments, such as tidal inlets(Gao & Collins, 1992), tidal sand banks (Van-wesenbeeck & Lanckneus, 2000) and tidal flats

Swell impact on fringing reef sedimentary processes 2017

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

6·30’

6·20’

6·00’

5·90’ CombinedCB+/FB−

21°S6·10’

BEFORE SWELL

T2

T4

T3

T6T5

T1

FB−

T2

T4

T3

T6T5

T1

14·20’ 14·30’ 14·40’ 14·60’ 14·70’ 14·80’

CB+ XOR FB−

55°E14·50’

CB+FB–

T2

T4

T3

T6T5

T1

CombinedCB+/FB−

AFTER SWELL

T2

T4

T3

T6T5

T1

CB+

T2

T4

T3

T6T5

T1

14·20’ 14·30’ 14·40’ 55°E14·50’

14·60’ 14·70’ 14·80’

6·30’

6·20’

6·00’

5·90’ CB+ XOR FB−

21°S6·10’

CB+FB–

T2

T4

T3

T6T5

T1

6·30’

6·20’

6·00’

5·90’ FB−

21°S6·10’ T2

T4

T3

T6T5

T1

6·30’

6·20’

6·00’

5·90’ CB+

21°S6·10’ T2

T4

T3

T6T5

T1

A E

B F

C G

D H

Fig. 10. Sediment transport trend vectors computed using the GSTA method before and after the swell event.Different cases of combination (or not) of the FB– and CB+ vector trends are represented: (A) and (E) combined CB+and FB– transport cases; (B) and (F) CB+ transport cases alone; (C) and (G) FB– transport cases alone; (D) and (H)exclusive OR operation on the CB+ and FB– transport cases.

2018 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

(Pedreros et al., 1996). The present study isthought to be the first application of the GSTAprocedure in a microtidal fringing reef environ-ment on a substrate mainly composed of biode-trital carbonate sediments. A temporal scalecould be introduced into the GSTA method byusing the sampling depth of sediments (Gao &Collins, 1992). In the present study, the sam-pling depth, around 2 to 5 cm, could be associ-ated with a time scale of one week to one month.Hence, the vectors computed before the swellevent would be related to several previoustransport processes, combined with the over-printing of biological activity (Chazottes et al.,2002). The vector trends computed after theswell event would be representative of contem-poraneous transport processes, strictly related tothis event.

The results of this study, derived from hydro-dynamic measurements and GSTA procedure,suggested that sediment movements would becontrolled by the position of sea-level on the reef-flat, the offshore wave climate and the reefmorphology. Independently of the meteorologicalconditions, the GSTA procedure highlighted FB–onshore and alongshore transport trends (low-energy transport process according to McLaren,1981), and CB+ offshore trends (high-energytransport process). Following Roberts (1989) andStorlazzi et al. (2004, 2009), the present studyleads to a conceptual model of spatialized car-bonate sand transport. This model applies to theback-reef area of a fringing reef laterally boundedby two inlets, and under various different meteo-rological conditions (Fig. 11). This conceptualmodel supports the hypothesis that the reef actsas a reservoir of carbonate sands for the back-reefarea (Halley, 2000), which appears as a sink, andfavours the low-energy transport processes(Roberts, 1989; Gourlay & Colleter, 2005).

Under fair weather conditions

In this case, transient sediment transport resultsfrom surge-currents that are only active duringhigh tide. When a critical water depth is reachedon the reef-flat, the transfer of offshore waveenergy into the reef is at its maximum (Fig. 8).However, these events are of short duration (fewseconds) and the mean flow does not maintain thesediments in suspension, favouring the settlingout of the transported sediments on the inner partof the reef-flat (Fig. 11A). This mechanism couldexplain the results obtained for the onshoreFB– case (Fig. 10C). The accumulated sediments

were produced on the reef-flat by biologicalprocesses (bioerosion) due to organisms such asthe sea urchin Echinometrix mathaei and parrot-fish (Bruggemann et al., 1996; Chazottes et al.,2004). These sediments represent a source ofmedium to very fine sands available for supplyingthe back-reef and the near-shore zone.

As a result, sediments of the inner reef-flat arecharacterized by coarse sands, which were al-ready present in situ, with a marked granulomet-ric asymmetry towards fine particles deposited bythese temporary mechanisms (see distribution S2on Fig. 11A). The sediments of the near-shorezone are characterized by fine sands with a stronggranulometric asymmetry towards coarse sands,and a tendency to bimodality. This trend, reflect-ing the existence of two sedimentary supplies, isthe inheritance of historic processes of transportand sedimentation. As a result, alongshore FB–and offshore CB+ cases, obtained before the swellevent (Fig. 10B and C), are not representative ofcontemporaneous dynamics, as the sediments inthe near-shore originate from the superposition ofanterior transport processes.

Under severe weather conditions

In this case, with circulation forced by offshorewaves breaking at the reef-crest, sediment trans-port is more active and follows three mainpathways (Fig. 11B). The transport of suspendedsediment takes place from the reef-flat to the back-reef area, due to advection by the mean onshoreflux (onshore FB– case, Fig. 10G). In the back-reef, the whole reef morphology conditions thenear-shore circulation. It induces a divergence ofthe flow due to the presence of a pass 1Æ5 km tothe south end of the reef. These fine sandsadvected in the near-shore are trapped in thelittoral drift and then transported alongshore (FB–case, Fig. 10G and H). In reef environments, manyauthors described a stratification of circulation inthe water column due to the slope of the seasurface induced by the accumulation of waternear the shore, and related to the relative sea-level height (Kench, 1998; Hearn, 1999; Branderet al., 2004; Lugo-Fernandez et al., 2004;Storlazzi et al., 2004, 2009). In this case, at hightide, the return current that takes place on the seabed to balance the wave-induced flow can pro-duce significant bedload transport, thus carryingthe coarser sands further offshore (offshore CB+case, Fig. 10F and H). However, these backwashcurrents are episodic and occur mainly duringhigh-tide sea states. Consequently, there is no

Swell impact on fringing reef sedimentary processes 2019

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

reason why CB+ offshore bedload transportshould have a significant impact on the spatialsediment distribution in the back-reef area. More-over, the predominance of FB– trends computedjust after the swell event suggests that this type oftransport case is prevalent during severe weatherconditions. This result is emphasized by the useof the XOR operation in the vector trends com-putation (Poizot & Mear, 2010).

The resulting granulometric distributions showfiner sands near the shore at both extremities ofthe diverging flow (distribution S4 on Fig. 11B),whereas the divergence zone is characterizedby a loss of fine sands and a gain of coarsermaterials (distribution S3 on Fig. 11B). In the

inner reef-flat, sediment distributions are almostunchanged, slightly finer in the divergence zoneand coarser at the extremity of the studied area.

CONCLUSION

In classical grain-size trend analysis (GSTA), acombined trend type is computed (as the vectorialsum of the occurrence of either CB+ or FB–) thatis supposed to yield a better correlation thanusing the trends individually. In the presentstudy, the use of GSTA with an exclusive ORoperation leads to better results, since the netsediment transport pathways are more consistent

BACK REEF

CORAL BEACH

REEF FLAT

OFFSHORE

CHENAL

BACK REEF

CORAL BEACH

REEF FLAT

OFFSHORE

CHENAL

CHENAL

Suspended sediment transport(onshore and alongshore)

Area of sediment deposition

Bedload sediment transportoccurring during high tide (offshore)

S1

S2

–2 0 2 4 60

10

20

%

phi units

–2 0 2 4 60

10

20

%

phi units

S1

Sediment distribution in the inner reef-flat

S2

Sediment distributionnear-shore

–2 0 2 4 60

10

20

%

phi units

–2 0 2 4 60

10

20

%

phi units

–2 0 2 4 60

10

20

%

phi units

S3 Sediment distribution near-shorediverging zone

S5 Sediment distributionin the inner reef-flat

S4 Sediment distributions near-shore extremity of the diverging zone

S3

S4

S4

S5

Exiting flow through the pass

A

Fair weatherconditions

B

Strong weather conditions

CHENAL

Fig. 11. Conceptual model of sediment dynamics in a discontinuous fringing reef located between two outlets orchannels: (A) under fair weather conditions; and (B) for a more agitated sea state due to wind or wave forcings. Thismodel indicates the main pathway of sediment transport, the main area of sediment deposition in the back-reef andthe granulometric characteristics of the deposited sediment.

2020 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

with the hydrodynamic records and physicalprocesses in fringing reefs. However, the resultsof transport trends have to be handled with carebefore the swell event because the sedimentarydeposits result from the superposition of severalprevious transport processes. On the other hand,the use of GSTA after the swell event is relevantsince it highlights transport trends strictly relatedto the transport process under consideration.

The following further studies and applicationsof GSTA in fringing reef environments are rec-ommended:

1 The use of GSTA analysis including an XORoperation on the computed vectors. A good levelof confidence can be assumed when using theFB– trends, but care needs to be taken in theinterpretation of CB+ transport cases.

2 Taking other sediment transport cases intoaccount, such as poorly sorted sediments, wouldimprove the use of the GSTA method in coral reefenvironments. Because of the different possiblesources of sediments, most of the sediments incoral reef ecosystems are poorly sorted and showbimodal grain-size distributions in the back-reefarea.

The results of this study support the hypothesisthat the back-reef zone of fringing reefs is an areaof net sedimentation and a sink for carbonatesands derived from the reef-flat. Moderate off-shore wave climate and wind forcing inducewater flows that cannot remove all the sedimentsproduced by reef organisms. Only extreme events,such as tropical cyclones or strong south-westerlydeep-water waves during the winter season,could lead to significant sediment transport tobalance the sediment production, which wouldprevent the reef being buried under its ownaccumulated sediments.

ACKNOWLEDGEMENTS

The authors wish to thank the French Institute forResearch and Development (IRD) and the Geo-sciences Laboratory of the University of LaReunion for their support. Emmanuel Cordierwas funded by the Conseil Regional of LaReunion Island, and by the Run Sea Scienceprogramme, under the European Union’s SeventhFramework Programme (FP7) for Research andTechnological Development. The ANR INTER-FACE programme made it possible to finalize thisproject. Dr. M.S.N. Carpenter and Dr. T. Catrypost-edited the English style.

NOMENCLATURE

d50 Mean grain size of sedimentD* Non-dimensional grain diameterF Incident wave powerg Acceleration due to gravityH Significant wave heighth Water depthv Kinematic viscosityq Sea water densityqs Sediment densitys Ratio between the sea water density q and

sediment density qs;

T Significant wave periodUcr Critical current velocity, threshold of

motion for sedimentsWs Settling velocity of sediment

REFERENCES

Anthony, K.R.N. (2000) Enhanced particle-feeding capacity of

coral on turbid reefs (Great Barrier Reef, Australia). Coral

Reefs, 19, 59–67.

Braithwaite, C.J.R. (1973) Settling behaviour related to sieve

analysis of skeletal sands. Sedimentology, 20, 251–262.

Brander, R., Kench, P.S. and Hart, D. (2004) Spatial and

temporal variations in wave characteristics across a reef

platform, Warraber Island, Torres Strait, Australia. Mar.

Geol., 207, 169–184.

Bruggemann, J.H., van Kessel, A.M., van Rooij, J.M. and

Breeman, A.M. (1996) Bioerosion and sediment ingestion

by the Caribbean parrotfish Scarus vetula and Sparisoma

viride: implications of fish size, feeding mode and habitat

use. Mar. Ecol. Prog. Ser., 34, 59–71.

Camoin, G., Colonna, M., Montaggioni, L.F., Casanova, J.,Faure, G. and Thomassin, B.A. (1997) Holocene sea level

changes and reef development in the southwestern Indian

Ocean. Coral Reefs, 16, 247–259.

Chazottes, V., Le Campion-Alsumard, T., Peyrot-Clausade, M.and Cuet, P. (2002) The effects of eutrophication-related

alterations to coral reef communities on agents and rates of

bioerosion (Reunion Island, Indian Ocean). Coral Reefs,

21(4), 375–390.

Chazottes, V., Chevillotte, V. and Dufresne, A. (2004) Char-

acteristics of the particulate production by the sea urchin

Echinometra mathaei on Indo-Pacific reefs: influence of

coral communities and implications on the sedimentary

dynamics. Geobios, 37(1), 13–22.

Chazottes, V., Reijmer, J.J.G. and Cordier, E. (2008) Sediment

characteristics in reef areas influenced by eutrophication-

related alterations of benthic communities and bioerosion

processes. Mar. Geol., 250, 114–127.

Cordier, E.. (2007) Dynamique hydrosedimentaire du reciffrangeant de l’Ermitage / La Saline (La Reunion): processus

physiques et flux sedimentaires. These de doctorat, Uni-

versite de La Reunion, 193 pp.

Coronado, C., Candela, J., Iglesias-Prieto, R., Sheinbaum, J.,Lopez, M. and Ocampo-Torres, F.J. (2007) On the circula-

tion in the Puerto Morelos fringing reef lagoon. Coral Reefs,

26(1), 149–163.

Swell impact on fringing reef sedimentary processes 2021

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

Dail, H.J., Merrifield, M.A. and Bevis, M. (2000) Steep beach

morphology changes due to energetic wave forcing. Mar.

Geol., 162, 443–458.

Dodge, R.E. and Vaisnys, J.R. (1977) Coral populations

and growth pattern: responses to sedimentation and tur-

bidity associated with dredging. J. Mar. Res., 35, 715–730.

Friend, P.L., Velegrakis, A.F., Weatherston, P.D. and Collins,M.B. (2006) Sediment transport pathways in a dredged RIA

system, southwest England. Estuar. Coast. Shelf Sci., 67,491–502.

Gabrie, C. and Montaggioni, L.F. (1982) Sediments from

fringing reefs of Reunion Island, Indian Ocean. Sediment

Geol, 31, 281–301.

Gao, S. (1996) A Fortran program for grain-size trend analysis

to define net sediment transport pathways. Comput. Geosci.,22, 449–452.

Gao, S. and Collins, M. (1991) A critique of the ‘‘McLaren

Method’’ for defining sediment transport paths. J. Sediment.

Petrol., 61, 143–146.

Gao, S. and Collins, M. (1992) Net sediment transport patterns

inferred from grain-size trends, based upon definition of

‘‘transport vectors’’. Sediment. Geol, 81(1), 47–60.

Gao, S. and Collins, M. (2001) The use of grain size trends in

marine sediment dynamics: a review. Chin. J. Oceanol.

Limnol., 19, 265–271.

Gourlay, M.R. (1994) Wave transformation on a coral reef.

Coast. Eng., 23, 17–42.

Gourlay, M.R. (1996a) Wave set-up on coral reefs: 1. Set-up

and wave-generated flow on an idealized two dimensional

horizontal reef Coast. Eng., 27, 161–193.

Gourlay, M.R. (1996b) Wave set-up on coral reefs: 2. Set-up on

reefs with various profiles Coast. Eng., 28, 17–55.

Gourlay, M.R. and Colleter, G. (2005) Wave-generated flow on

coral reefs – an analysis for two-dimensional horizontal

reef-tops with steep faces. Coast. Eng., 52(4), 353–387.

Halley, R.B.. (2000) 11 Things a geologist thinks an engineer

should know about carbonate beaches. in: Carbonate Bea-ches 2000, Proceedings of First International Symposium on

Carbonate Sand Beaches, (Lisa.L. Robbin, Orville.T.

Magoon and Lesley. Ewing, Eds, ASCE), pp 1–14.

Hearn, C.J. (1999) Wave-breaking hydrodynamics within coral

reef systems and the effect of changing relative sea level.

J. Geophys. Res., 104, 30007–30019.

Hubbard, D.K., Sadd, J.L. and Roberts, H.H. (1981) The role of

physical processes in controlling sediment transport pat-

terns on the insular shelf of St. Croix, US Virgin Islands.

Proc. 4th Int. Coral Reef Symp., Manila, 1, 399–404.

James, N.P. and Bourque, P.A.. (1992) Reefs and mounds. In:

Walker. R.G. and N.P. James., Facies models, Geological

Assoc. Canada, 323–348.

Kench,P.S. (1998)Acurrentsofremovalapproachforinterpreting

carbonate sedimentary processes. Mar. Geol., 145, 197–223.

Kench, P.S. and McLean, R.F. (1997) A comparison of settling

and sieve techniques for the analysis of bioclastic sedi-

ments. Sediment Geol, 109, 111–119.

Larcombe, P., Costen, A. and Woolfe, K.J. (2001) The hydro-

dynamic and sedimentary setting of near-shore coral reefs,

central Great Barrier Reef shelf, Australia: Paluma Shoals, a

case study. Sedimentology, 48, 811–835.

Le Roux, J.P. (1994a) An alternative approach to the identifi-

cation of net sediment transport paths based on Grain size

trends. Sediment Geol, 94, 97–107.

Le Roux, J.P. (1994b) Net sediment transport patterns inferred

from grain-size trends, based upon definition of ‘‘transport

vectors’’ – comment. Sediment Geol, 90, 153–156.

Lugo-Fernandez, A., Roberts, H.H. and Suhayda, J.N. (1998)

Wave transformations across a Caribbean fringing-barrier

coral reef. Cont. Shelf Res., 18, 1099–1124.

Lugo-Fernandez, A., Roberts, H.H. and Wiseman Jr, W.J.(2004) Currents, water levels, and mass transport over a

modern Caribbean coral reef: Tague Reef, St. Croix, USVI.

Cont. Shelf Res., 24(17), 1989–2009.

Maiklem, W.C. (1968) Some hydraulic properties of bioclastic

carbonate grains. Sedimentology, 10, 101–109.

McLaren, P. (1981) An interpretation of trends in grain size

measures. J. Sediment. Petrol., 51, 611–624.

McLaren, P. and Bowles, D. (1985) The effects of sediment

transport on Grain size distributions. J. Sediment. Petrol.,

55, 457–470.

Montaggioni, L.F.. (1978) Recherches geologiques sur lescomplexes recifaux de l’Archipel des Mascareignes (Ocean

Indien Occidental). Unpublished DSc Thesis, University of

Marseilles 2, 524 p.

Montaggioni, L.F. and Faure, G.. (1980) Les recifs coralliensdes Mascareignes (Ocean Indlen). Coll. travaux du Centre

universltaire, Univ. Reunion. 151 pp.

Naım, O. (1993) Seasonal responses of a fringing reef com-

munity to eutrophication (Reunion Island, Western Indian

Ocean). Mar. Ecol. Prog. Ser., 99, 137–151.

Naım, O., Chabanet, P., Done, T., Tourrand, C. and

Letourneur, Y. (2000) Regeneration of a reef-flat ten years

after the impact of the cyclone Firinga (Reunion, SW Indian

Ocean). Proc. 9th Int. Coral Reef Symp., Bali, Indonesia, 1,547–553.

Ogston, A.S., Storlazzi, C.D., Field, M.E. and Presto, M.K.(2004) Sediment resuspension and transport patterns on a

fringing reef-flat, Molokai, Hawaii. Coral Reefs, 23, 559–569.

Paulo Jr, J.R. and Diggle, P.J. (2001) GeoR: A Package for

Geostatistical Analysis. R news, The newsletter of the R

project, 1 2, 15–18.

Pedreros, R., Howa, H.L. and Michel, D. (1996) Application of

grain size trend analysis for the determination of sediment

transport pathways in intertidal areas. Mar. Geol., 135(1),35–49.

Plomaritis, T., Paphitis, D. and Collins, M. (2008) The use of

grain size trend analysis in macrotidal areas with break-

waters: implications of settling velocity and spatial sam-

pling density. Mar. Geol., 253, 132–148.

Poizot, E. and Mear, Y. (2010) Using a GIS to enhance grain

size trend analysis. Environ. Model Softw, 25, 513–525.

Poizot, E., Mear, Y., Thomas, M. and Garnaud, S. (2006) The

application of geostatistics in defining the characteristic

distance for grain size trend analysis. Comput. Geosci., 32,360–370.

Poizot, E., Mear, Y. and Biscara, L. (2008) Sediment Trend

Analysis through the variation of granulometric parameters:

a review of theories and applications. Earth Sci. Rev., 86,15–41.

Presto, M.K., Ogston, A.S., Storlazzi, C.D. and Field, M.E.(2006) Temporal and spatial variability in the flow and

dispersal of suspended-sediment on a fringing reef-flat,

Molokai, Hawaii. Estuar. Coast. Shelf Sci., 67, 67–81.

Rıos, F., Ulloa, R. and Correa, I.C.S. (2003) Determination of

net sediment transport patterns in Lirquen Harbor, Chile,

through grain-size trend analysis: a test of methods.

Pesquisas em Geosciencias, 30, 65–81.

Roberts, H.H. (1989) Physical processes as agents of sediment

transport in carbonate systems: examples from St. Croix

USVI. Terrestrial and Marine Geology of St. Croix, US Virgin

Islands, WIL Special Publication, 8, 95–104.

2022 E. Cordier et al.

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023

Rogers, C.S. (1990) Responses of coral reefs and reef organisms

to sedimentation. Mar. Ecol. Prog. Ser., 62, 185–202.

Soler, O.. (1997) Atlas Climatique de la Reunion. Meteo-

France ed.., La Reunion, France 80 p.

Soulsby, R.. (1997) Dynamics of Marine Sands. Thomas Tel-

ford Ltd, London, 249 pp.

Storlazzi, C.D. and Jaffe, B.E. (2008) The relative contribution

of processes driving variability in flow, shear, and turbidity

over a fringing coral reef: West Maui, Hawaii. Estuar. Coast.

Shelf Sci., 77, 549–564.

Storlazzi, C.D., Ogston, A.S., Bothner, M.H., Field, M.E. and

Presto, M.K. (2004) Wave- and tidally-driven flow and

sediment flux across a fringing coral reef: Southern Molokai,

Hawaii. Cont. Shelf Res., 24, 1397–1419.

Storlazzi, C.D., McManus, M.A., Logan, J.B. and McLaughlin,B.E. (2006) Cross-shore velocity shear, eddies and hetero-

geneity in water column properties over fringing coral reefs:

West Maui, Hawaii. Cont. Shelf Res., 26, 401–421.

Storlazzi, C.D., Field, M.E., Bothner, M.H., Presto, M.K. and

Draut, A.E. (2009) Sedimentation processes in a coral reef

embayment: Hanalei Bay, Kauai. Mar. Geol., 264, 140–151.

Taebi, S., Lowe, R.J., Pattiaratchi, C.B., Ivey, G.N., Symonds,G. and Brinkman, R. (2011) Nearshore circulation in a

tropical fringing reef system. J. Geophys. Res., 116, 15.

Te, F.T. (1997) Turbidity and its effect on corals: a model using

the extinction coeficient (k) of photosynthetic active radi-

ance (PAR). Proc. 8th Int. Coral Reef Symp, 2, 1899–1904.

Vanwesenbeeck, V. and Lanckneus, J. (2000) Residual sedi-

ment transport paths on a tidal sand bank: a comparison

between the modified McLaren model and bedform analy-

sis. J. Sediment. Res., 70, 470–477.

Supporting Information

Additional Supporting Information may be foundin the online version of this article:

Figure S1. Workflow of the GSTA methoddescribing the four steps used to compute vectortrends.

Please note: Wiley-Blackwell are not responsi-ble for the content or functionality of any sup-porting materials supplied by the authors. Anyqueries (other than missing material) should bedirected to the corresponding author for thearticle.

Manuscript received 3 June 2011; revision accepted 8March 2012

Swell impact on fringing reef sedimentary processes 2023

� 2012 The Authors. Journal compilation � 2012 International Association of Sedimentologists, Sedimentology, 59, 2004–2023