Hydrodynamics and Sediment Fluxes across an Onshore Migrating Intertidal Bar

Journal of Coastal Research 22 2 247–259 West Palm Beach, Florida March 2006

Hydrodynamics and Sediment Fluxes across an OnshoreMigrating Intertidal BarTroels Aagaard†, Michael Hughes§, Regin Møller-Sørensen†, and Steffen Andersen†

†Institute of GeographyUniversity of CopenhagenOster Voldgade 10DK-1350 Copenhagen,

[email protected]

§School of GeosciencesUniversity of SydneySydney NSW 2006, Australia

ABSTRACT

AAGAARD, T.; HUGHES, M.; MØLLER-SØRENSEN, R., and ANDERSEN, S., 2006. Hydrodynamics and sedimentfluxes across an onshore migrating intertidal bar. Journal of Coastal Research, 22(2), 247–259. West Palm Beach(Florida), ISSN 0749-0208.

Detailed hydrodynamic and morphological data are presented from a field deployment spanning 2 days (four tidecycles). The data include bed-elevation changes measured at each low tide and continuous records of water-surfaceelevation, cross-shore and long-shore current velocities, and suspended sediment concentrations all measured within20 cm of the bed. During the deployment, an intertidal bar migrated onshore and infilled a runnel on its landwardside. The depth of this runnel was initially 0.6 m. During the migration of the bar, the significant wave height indeep water was ca. 2 m and wave period was 7 seconds. The significant wave height over the intertidal bar crest wasabout 0.25 m. Suspended sediment fluxes were estimated (product of current velocity and suspended sediment con-centration profile) and partitioned between mean and oscillatory components with the latter further partitioned be-tween short and long wave contributions. When the bar was migrating shoreward and infilling the runnel, estimatedsuspended sediment flux for all components was directed landward on the bar crest. Once the migrating bar hadinfilled the runnel, however, the suspended sediment fluxes for the mean component were directed seaward, whereasthe short wave-driven flux was still directed landward. These results represent a clear example of morphodynamicinteractions—(a) as waves cross the intertidal bar the onshore mean and oscillatory components transport sedimentshoreward, (b) the presence of the runnel reduces the offshore component of oscillatory transport by channeling theflow alongshore, (c) the runnel rapidly infills due to the strong transport asymmetry, (d) once the runnel has infilled,the mean cross-shore current and mean sediment flux reverse direction. When the runnel is present, the generalintertidal circulation is a horizontal cell circulation with rip currents, whereas it becomes a vertical undertow circu-lation when the runnel has infilled.

ADDITIONAL INDEX WORDS: Swash bar, ridge and runnel, cell circulation, morphodynamics.

INTRODUCTION

The intertidal zone of micro/mesotidal beaches in semien-closed seas are often characterized by the existence of one ormore intertidal bars. Such intertidal bars can take on variousforms and display different types of dynamic behavior. Fol-lowing GREENWOOD and DAVIDSON-ARNOTT (1979), WIJN-BERG and KROON (2002) distinguished between two maintypes of intertidal bars, (a) slip-face ridges that are asym-metric forms and relatively mobile and (b) low-amplituderidges that are more symmetric in form and largely staticfeatures. These two bar types also correspond to the termsswash bars and ridge and runnels (ORFORD and WRIGHT,1978), respectively.

Despite the accessibility of the intertidal zone, the physicalprocesses governing intertidal bar behavior are not well un-derstood, mainly because of the difficulties in measuring hy-drodynamics and sediment transport in very shallow waterdepths. In the case of slip-face ridges, it is unclear whether

DOI:10.2112/04-0214.1; received 3 May 2004; accepted in revision 20September 2004.

these bars are in fact generated and maintained by swash(WIJNBERG and KROON, 2002) or surf-zone processes (AA-GAARD et al., 1998a; KROON and MASSELINK, 2002). Theytend to form in the mid to lower intertidal zone and migrateonshore under low/moderate-energy conditions, whereas theymay be eroded during high-energy situations. The water levelacross the bar crest would seem to be an important param-eter in determining the bar behavior. KROON and MASSE-LINK (2002) observed landward migration associated withmean onshore flows when water depths were less than 0.2 mat the bar crest, while DAWSON, DAVIDSON-ARNOTT, andOLLERHEAD (2002) observed a critical depth of 0.1 m. Similaronshore-directed mean flows were observed at bar crests byAAGAARD et al. (1998a, 1988b) and by KROON and DE BOER

(2001).The origin and detailed dynamics of the ridge and runnel

type of intertidal bars may be even more obscure. This bartype mainly occurs with relatively large tidal ranges andsmall waves (KING and WILLIAMS, 1949; ORFORD andWRIGHT, 1978), i.e., for larger values of the relative tide range(MASSELINK and SHORT, 1993). VOULGARIS et al. (1998) mea-

248 Aagaard et al.

Journal of Coastal Research, Vol. 22, No. 2, 2006

Figure 1. Cross-shore profile at Skallingen, August 25, 2002. The profilecomprises an intertidal bar and two nearshore (subtidal) bars. Positionsof the four instrument stations across the intertidal bar are indicated bythe vertical lines. Mean annual sea level is at 1 0.14 m DNN (DanishOrdnance Datum).

sured hydrodynamics and sediment transport across a ridge-and-runnel system under low-energy conditions, but theywere unable to reconcile their measurements (and derivedmodeling efforts) to the observed landward bar migration andtracer movement. As their measurements were restricted toshoaling wave conditions in water depths .0.8 m, they con-cluded that measurements in shallow water depths (swashand inner surf zones) are required to quantify the sand trans-port and morphological development of intertidal bars. Sim-ilar problems affected the results of STEPANIAN et al. (2001),who also measured hydrodynamics in the shoaling wave zoneand were unable to relate onshore tracer movements andridge migration to observed processes.

One distinguishing feature about intertidal bars is thatthey are often dissected by rip channels, and onshore flowshave been recorded at bar crests in association with offshoreflows in rip channels. The currents thus form a cell circula-tion system with longshore rip feeder flows in the runnels,and this type of circulation is typical for moderate-energyconditions when waves are breaking over the landward mi-grating bars. While some numerical models for such three-dimensional flows have started to emerge (e.g., SHORECIRC;HAAS et al., 2003), these models have not yet been extendedto simulate sediment transport and morphological change.Consequently, morphodynamics and bar behavior in three-dimensional bathymetric settings are not well predicted bynumerical models. Indeed, most available models are two-di-mensional and simulate hydrodynamics and sediment trans-port in a cross-shore profile; they cannot simulate the land-ward movement of intertidal bars under breaking waves.More field data on hydrodynamics and sediment transportfrom three-dimensional morphological settings are thereforerequired to constrain future model development (SOULSBY,1999).

The present study obtained such measurements under low-to moderate-energy conditions across an intertidal bar of theslip-face ridge type. In the course of three tidal cycles, a largeintertidal bar migrated landward and welded to the beach.The processes responsible for this behavior are documentedthrough measurements of flow velocities and sediment con-centration obtained close to the bed at four measurement po-sitions. As the tide rose and fell, the instruments were sub-jected to various hydrodynamic regimes and the relative ef-fects of swash, surf, and shoaling wave processes to intertidalbar dynamics are evaluated. Furthermore, the morphodyn-amic feedbacks between morphology and hydrodynamic pro-cesses are elucidated; as the bar moved onshore and closedthe landward runnel, onshore-directed mean flows were re-placed by offshore-directed undertow.

MATERIALS AND METHODS

Experimental Site and Procedures

The field experiment was conducted from August 23, 2002,to September 4, 2002, at Skallingen, which is located on theNorth Sea coast of Denmark. The shoreface has a gentleslope, b ø 0.007, the mean annual offshore significant waveheight is 1 m, and the mean tidal range is 1.5 m, increasing

to 1.8 m at spring tides. The shoreface exhibits 2–3 subtidalbars and additionally, one or two intertidal bars are common.

At the outset of the experiment, the upper shoreface hadsubtidal bar crests located at x 5 250 m and x 5 150 mrelative to the survey baseline, a rather large intertidal barcentered at x 5 90 m and an upper swash bar/berm at x 555 m (Figure 1). During the course of the experiment, theintertidal bar moved landward, closed the runnel, and weldedto the beach. The behavior was consistent with that typicallydisplayed by such bars at Skallingen: Intertidal bars tend tomigrate landward until they weld to the beach; after weldingand runnel infilling, the bar(s) may be eroded during high-energy situations and the sediment recycled to the lower in-tertidal zone (AAGAARD et al., 1998a).

Initially, the survey (and instrument) transect was locatedacross the intertidal bar approximately midway between tworip channels, which were spaced about 175 m apart. The dif-ference in elevation between the bar crest and the landwardrunnel was approximately 0.6 m and the bar form wasoblique to the beach, with the northern part of the bar locatedcloser to the shoreline, consistent with the dominant south-erly longshore sediment transport at the site (Figure 2).

The sediment on this bar was well sorted with a meangrain size of 200–240 mm. Wave-energy levels during the ex-periment were quite low (Figure 3). The significant offshorewave height (recorded 18 km offshore in a water depth of ø12m) remained below 0.5–0.6 m until August 29, when a galeoccurred and waves increased to 1.2 m and further to 2.1 mon August 31, and subsequently wave heights decreasedagain. Peak spectral wave periods increased from ø4–8 sec-onds during the event, which was also associated with a smallsurge of ø0.2 m (Figure 3) due to the onshore winds. Tideswere recorded at the ebb delta, about 3 km away from thefield site. Unfortunately, tidal records are missing prior tothe event, which was initiated 3 days after a spring tide; the

249Intertidal Bars


Figure 2. Three-dimensional topographic surfaces of the beach and in-tertidal zone at the experimental site before (August 24, 2002) and after(September 2, 2002) the event described in the article. The instrumenttransect was located at the longshore coordinate y 5 0 m.

Figure 3. Mean water level (top panel) and (bottom panel) significantoffshore wave height (solid line) and peak spectral wave period (dashedline) during the experiment period. Tidal records are missing prior toAugust 29. The dashed line in the upper panel indicates the mean annualwater level.

tidal stage was thus between spring and neap, with decreas-ing tidal ranges.

Four instrument stations were established in the surveyedtransect (Figure 1). These stations consisted of H-frames jet-ted approximately 1.5 m into the bed and all were equippedwith a Marsh-McBirney OEM 512 electromagnetic currentmeter (EM) at a nominal elevation of 0.20 m above the bedand an array of three OBS-1P optical backscatter sensors atnominal elevations of 0.05, 0.10, and 0.20 m above the bedfor sediment-transport measurements. At the uppermost sta-tion (S4), however, the current meter was deployed at a nom-inal elevation of 0.12 m and the lower OBS at 0.035 m. Wavetransformations and mean water levels were measured withpressure sensors (Viatran Model 2406A at S1 and S2 andDruck Model PTX1830 at S3 and S4). At the upper stations(S3 and S4), the pressure sensor elevation was kept at, orslightly below, bed level in order to measure water depths inthe swash zone. These upper stations were also equippedwith three-dimensional sideways-looking Sontek 10 MHzAcoustic Doppler Velocimeters (ADV) at nominal elevationsof 0.02–0.03 m above the bed and at S3, a vertical array offive D&A Instruments UFOBS-7 fiber-optical backscattersensors was installed. The UFOBS-7 uses an infrared laserto detect sediment concentration within a very small sam-pling volume (nominally ø10 mm3) that is centered 10–15

mm away from the sensor head. Due to the small size of thesensor head (8 mm outer diameter), the instrument is capableof recording sediment concentrations and, in combinationwith the ADV, suspended sediment transport very close tothe bed. In this experiment, sampling volumes were nomi-nally centered at z 5 0.01, 0.02, 0.03, 0.04, and 0.05 m. Allsensors were colocated in the cross-shore and readjustedwhen necessary to maintain a constant elevation relative tothe mobile bed. Sensors were hardwired to a mobile field sta-tion in the dunes where the signals were recorded on laptopcomputers. When instruments were covered by water, databursts of 45-minute duration were recorded almost continu-ously at a frequency of 10 Hz. Given the relatively close spac-ing of the instrument stations (ø15 m), convergences and di-vergences of suspended sediment transport could be evalu-ated and compared with morphological changes.

Such changes were quantified from changes in bed eleva-tion along a line of 62 survey rods located about 5 m southof the instrument transect. The rods were 5 mm in diameterand were established with 2-m individual spacing and theline spanned the entire intertidal zone. The top of the rodswere surveyed relative to a benchmark in the dunes and thedistance from the top of the rods to the sand surface wasmeasured using a specially designed ruler at each low tidethroughout the field campaign. Elevations were determinedto the nearest millimeter and survey errors on the flat, well-packed bed at low tide are estimated as being less than 5

250 Aagaard et al.


mm. This survey method provides an inexpensive and rea-sonably reliable means of estimating the net sediment (bed-load and suspended load) transport across the profile. Final-ly, area surveys were conducted at the beginning, middle, andend of the experiment period using a total station along sevencross-shore transects, spaced 25 m apart, from the dune crestto the low-tide limit of wading (Figure 2).

Date Processing and Analysis

Electromagnetic current meter offsets were determined inbuckets prior to the experiment as well as at times of low tidewhen sensors became intermittently exposed but were stillwet; sensor gains were determined in a large tow tank priorto the experiment. The pressure sensors were calibrated in astilling well at the field site. In the case of the Viatran sen-sors, offsets were adjusted for atmospheric pressure fluctua-tions during the experiment. This was not necessary for theDruck sensors, however, because they were vented. OBS- andUFOBS-sensors were post-calibrated in a large recirculationtank using sand samples from the deployment locations.Field offsets caused by minute amounts of permanently sus-pended organics and/or fine-grained sediment particles orig-inating from the inlet were determined from breaks in thecumulative frequency distribution (AAGAARD and GREEN-WOOD, 1994). These offsets were generally close to the secondand fifth percentile frequency output voltages for the UFOBSand OBS sensors, respectively, and, to maintain consistency,these percentiles were applied to all records.

Prior to analysis, the sensor outputs were screened andchecked for data quality and noisy and/or erroneous datawere discarded from further analysis. Such errors could occurdue to bed accretion resulting in (UFOBS/OBS) signal satu-ration or instrument emergence. Also, OBS signals some-times become spiky in very shallow waters depths (probablydue to surface foam associated with surf/swash bores propa-gating past the instrument), which generally results in in-verted sediment concentration profiles. This problem did notappear to affect the output of the UFOBS sensors, which werelocated closer to the bed.

Velocity measurements from the ADV tended to becomenoisy in highly turbulent or aerated flows. At such times,signal correlation values recorded by the ADV were used toidentify potentially inaccurate data. When signal correlationfor a given acoustic beam was less than 55%, the raw velocitydata was replaced by the filtered signal obtained by applyinga 1-Hz filter (cf. RAUBENHEIMER, 2002). Finally, in the swashzone, the sensor sometimes became emerged; a signal-to-noise ratio of less than 20 was employed to identify such oc-casions in which the flow velocity is undefined (HUGHES andBALDOCK, 2004).

Pressure records were detrended prior to computing waveheights, but correction for depth attenuation was not appliedbecause of the small water depths in the intertidal zone.Mean water depths and water levels were determinedthrough repeated surveys of instrument positions and mea-surements of sensor elevations relative to the bed.

Instantaneous sediment flux at a particular elevation wascalculated as the product of instantaneous sediment concen-

tration and fluid velocity. For the UFOBS array, sedimentconcentrations were paired with velocities from the ADV,whereas velocities from the EMs were used with the OBSrecords. For surf-zone data, sediment fluxes were partitionedinto mean and oscillatory terms generated by mean currentsand oscillatory wave motions (at both incident and infragrav-ity frequencies), respectively (see AAGAARD and GREEN-WOOD, 1994).

RESULTS

Morphological Change

Prior to the increased wave energy associated with the galeoccurring on August 30 to August 31 (Figure 3), the intertidalbar was largely inactive. Only when mean water levels be-came sufficient to inundate the intertidal bar crest in theafternoon of August 30 did the onshore bar migration com-mence. Figure 4 illustrates the morphological change occur-ring over the four tidal cycles between the early morning ofAugust 30 and the early morning of September 1.

During cycle 1 (08300015–08301300), only very limitedmorphological change occurred, while cycle 2 resulted in a 5–10-m onshore migration of the bar crest and the 10-m widelandward runnel began to infill as sediment was scoured fromthe bar crest and deposited into the trough. Large clouds ofsuspended sand were driven landward with each wave strokeand deposited on both the landward bar slope and in the run-nel (Figure 5). As wave energy levels were very low in therunnel due to wave dissipation by the shallow water depthsacross the bar crest, and mean longshore (rip feeder) currentswere not sufficiently strong to remobilize the sand, infillingprogressed rapidly and was almost completed during tidalcycle 3 (08310130–08311315; Figure 4). Tidal cycle 4 resultedin a smoothing of the convexity marking the former intertidalbar crest. The morphology of the intertidal zone prior to andafter intertidal bar welding is illustrated in Figure 6; thewelding process resulted in a virtually planar intertidalbeach face.

Detailed patterns of erosion and deposition across the in-tertidal bar during tidal cycles 2–4 are shown in Figure 7.Initially, deposition prevailed around station S3 and in therunnel, where up to 0.40 m of accretion occurred, while ero-sion occurred around station S4 at the bar crest and acrossthe lower seaward slope of the bar. During the two final tidalcycles, erosion prevailed across most of the lower and upperseaward slope of the bar and accretion was limited to therunnel, where accretion rates systematically declined withtime as accommodation space decreased. The shifting pat-terns of limited erosion/accretion around station S1 was prob-ably due to longshore migrating bedforms driven by the long-shore current; visual observations indicated a prevalence ofripples and megaripples seaward of station S2. The netbathymetric change over the three tidal cycles is also illus-trated in Figure 7. A maximum of 0.65 m of accretion oc-curred in the runnel, while erosion prevailed everywhere else,with a maximum of 0.22 m at station S4. The net sedimentdeposition landward of station S3 on the upper seaward slopeof the bar was 0.74 m3/m.

Alternating zones of erosion and deposition (or nonerosion)

251Intertidal Bars


Figure 4. Morphological change observed in the intertidal survey tran-sect during tidal cycles 1–4 (top to bottom panels) on August 30, 2002, toSeptember 1, 2002. Note the progressive landward migration of the in-tertidal bar and the associated closure of the runnel. The inner subtidalbar also exhibited a net onshore migration.

occurred across the seaward slope of the bar, and there issome evidence to suggest that there was a landward propa-gation of these zones, which could be analogous to the on-shore migrating bed oscillations described by GREENWOOD,

AAGAARD, and NIELSEN (2004). Moreover, the height andspacing of the present undulations (0.1–0.2 m and 20–30 m,respectively) are consistent. Alternatively, the spatially shift-ing zones of erosion/accretion might indicate temporallychanging positions of sediment-transport convergence/diver-gence across the bar.

Waves and Currents

Wave breaker patterns were quite persistent over the galeevent. Almost all waves broke through spilling across the in-ner subtidal bar and they reformed in the trough betweenthe subtidal and intertidal bars. Due to the filtering effectsof the subtidal bar, the secondary breakpoint on the intertidalbar was located close to station S2; at high tide, the mainbreakpoint was typically displaced some 5 m landward andat low tide some 5 m seaward of this instrument station. Thebreaker type at the intertidal bar was predominantly spilling.Thus, shoaling waves were almost always observed at stationS1, where the relative wave height (Hs /h) remained belowø0.4; S2 was in the shoaling zone with occasionally breakingwaves at high tide, or in the inner surf zone at low tide. S3and S4 were in the inner surf zone with spilling bores at hightide and in the swash zone (or dry) at low tide. The relativewave height was almost consistently .0.6 at these stations.

Typical surface elevation spectra from a high tide are il-lustrated in Figure 8 at a time when the significant waveheight at the outer edge of the instrument array was 0.6 m.The figure shows a spectral peak at the incident wave fre-quency ( f ø 0.13 Hz) at stations S1 and S2, with suggestionsof a harmonic peak at twice that frequency, which indicatesthe skewed form of these shoaling waves. As waves brokeacross the intertidal bar, incident wave energy was dissipatedand infragravity waves with a peak frequency of f 5 0.01 Hzincreased progressively in amplitude.

Two instrument records have been selected for illustrationof the general hydrodynamic characteristics across the inter-tidal bar (Figure 9). These two examples were collected athigh tide on August 30 and August 31, respectively, with ap-proximately similar water levels but with different bathym-etries. The mean water levels at the upper instrument stationwere 11.12 m DNN (Danish Ordnance Datum) and 11.03 mDNN, respectively. In both cases, significant wave-height at-tenuation occurred due to breaking landward of station S2;this dissipation caused a mean water level setup of 0.15 macross the seaward slope of the bar (Figure 9). The limitedwave dissipation and the relative set-down between stationsS1 and S2 confirm that waves were not (or only weakly)breaking seaward of station S2. Even though the basic hy-drodynamic process regimes were thus identical in the twosituations, the mean cross-shore current characteristics atthe bar crest (station S4) were different.

The mean current velocities shown in Figure 9 were mea-sured close to the bed by the ADVs at stations S3 and S4 andby electromagnetic current meters at stations S1 and S2. Atstations S1–S3, the cross-shore currents were directed off-shore with speeds of U ø 0.1–0.15 m/s. The smallest currentvelocities were recorded around the breakpoint at station S2.These currents were probably undertows, driven by the sea-

252 Aagaard et al.


Figure 5. Low surf bores propagating across the intertidal bar crest and generating a hydraulic jump at the seaward edge of the deep runnel. Wavesare reforming in the runnel. Note the large amounts of sediment trapped in the hydraulic jump; this sediment eventually settles on the landward slopeof the intertidal bar and contributes to the onshore form migration.

ward-directed setup gradient generated by waves breakingacross the intertidal bar. The relatively large mean cross-shore (and longshore) current velocities observed at stationS1 were probably due to horizontal mixing and onshore sur-face mass transport associated with the breaking bores acrossthe inner subtidal bar at x 5 150 m (cf. CHURCH and THORN-TON, 1993; GARCEZ-FARIA et al., 2000). At the uppermost sta-tion, S4, however, the mean cross-shore currents were di-rected onshore at the bar crest with a speed of 0.05 m/s inthe first example and offshore with a speed of 0.10 m/s in thesecond example.

In both cases, the ADV at station S4 was permanently sub-merged throughout the instrument record. Unfortunately, nomean water level measurements were obtained in the runnel,but it is likely that the onshore current at station S4 was dueto either (a) a landward-directed pressure gradient generatedby a relative set-down in the runnel where incident waveswere reforming (Figure 5) and/or (b) the presence of the run-

nel reduced the offshore component of the oscillatory flow bychanneling this flow alongshore. Whatever the origin, the on-shore-directed mean current at the bar crest (S4) representedthe onshore-directed limb of a cell circulation pattern withthe mass transport of water across the bar crest drainingalong the runnel and subsequently seaward through thedowndrift rip channel (Figure 2). When the intertidal bar hadwelded to the beach and the runnel had closed (hour 187.7;Figure 9), the mean cross-shore current at station S4 clearlybecame part of the undertow circulation.

These mean current characteristics were consistentthroughout the four tidal cycles for the two bathymetric con-figurations (Figure 10). Prior to bar welding (tidal cycle 2),mean currents were persistently directed onshore at the barcrest (station S4) with speeds of 0.05–0.10 m/s, at instrumentelevations of 0.02–0.07 m above the bed. When the runnelinfilled at the beginning of the tidal cycle 3 (Figure 4), thecross-shore currents at the bar crest reversed and became

253Intertidal Bars


Figure 6. The morphology of the intertidal zone prior to and after bar welding. The upper photo shows the .10-m-wide runnel existing prior to theevent; and in the lower photo, taken during the final phase of the experiment, the beach is near planar and the former runnel position is indicated by aslightly darker color due to increased surface moisture. The instrument stations are seen in the center of the photos.

254 Aagaard et al.


Figure 7. Detailed intertidal bathymetric change measured at the sur-vey rods over the three tidal cycles when significant morphological chang-es occurred. The accumulated net change is shown by the thick line. Thecross-shore profile and instrument positions are shown in the lower panelfor reference.

Figure 8. Water surface elevation spectra recorded at stations S1–S4 athigh tide, hour 187.7. The spectra have 50 degrees of freedom.

Figure 9. Cross-shore hydrodynamics recorded during two high tideruns at hours 162.3 (tidal cycle 2) and 187.7 (tidal cycle 4). From the topdown, the panels illustrate cross-shore (U, solid lines) and longshore (V,dashed lines) mean currents, mean water level setup relative to stationS1, and significant wave height (Hs, solid lines) and relative significantwave height (Hs /h, dashed lines). Onshore- (U) and northward- (V) di-rected mean currents are positive. The beach profiles are shown in thebottom panels for reference.

offshore directed with speeds of 0.10–0.15 m/s, similar tomean currents at the other three instrument stations. Hence,a morphodynamic feedback existed between the morphologyand the mean current circulation across the bar crest.

Cross-Shore Suspended Sediment Transport

Visual observations and the calculated sediment fluxes in-dicate that considerable amounts of sediment were movedlandward across the upper seaward slope and crest of theintertidal bar during the three tidal cycles when the bar wasactive (Figure 11). Sediment-transport rates were estimatedby summing the sediment fluxes calculated for each opticalsensor bin. At S1 and S2, velocity measurements determinedby the current meters at z ø 0.2 m were paired with sedimentconcentrations determined by the OBS at 0.05, 0.1, and 0.2m; each OBS sensor output was assumed representative fora 0.05 m (0.10 m) vertical bin. At S4, velocity measurementsfrom the ADV at z ø 0.03 m were paired with the OBS sen-sors at z 5 0.035, 0.085, and 0.135 m, and finally, at S3, ADVvelocity measurements at z ø 0.03 m were paired with sed-iment concentrations measured at z 5 0.01, 0.02, 0.03, 0.04,and 0.05 m, and EM velocity measurements at z 5 0.2 mwere paired with concentrations at z 5 0.1 and 0.2 m. Thecomputed estimates at S3 are considered to approximate thetotal suspended sediment transports occurring at this sta-tion, while transport estimates at the other stations may only

255Intertidal Bars


Figure 10. Mean cross-shore current velocities at the four instrumentstations. Positive values represent onshore currents. Times of low tideare indicated by the vertical dashed lines and the tidal cycle number isshown at the top of the figure.

Figure 11. Net cross-shore suspended sediment-transport rates acrossthe instrument array. At station S3, transports are illustrated for theupper instrument array (EM-OBS, dashed line), the lower array (ADV/FOBS, thin solid line), and sum of the two (heavy line). Positive transportrates are onshore directed. The numbers of the tidal cycles are shown atthe top of the figure, and low tide occurred at hours 156, 168, 181, and193.

be indicative, as sediment concentrations were not measuredvery close to the bed.

During all three tidal cycles, the cross-shore sediment-transport rate at S3 was large and directed onshore in smallwater depths, with a tendency for a transport reversal at hightide (Figure 11). There was a trend toward more seaward-directed sediment fluxes in the lower part of the water col-umn. On balance, however, the net estimated transport wasclearly onshore directed even though mean currents wereconsistently directed offshore (Figure 10). Sediment transportat the upper station S4 is most likely underestimated becausethe sediment concentrations were not measured closer than0.03–0.04 m above the bed and visual observations indicatedthat a significant fraction of the sediment transport occurredas a thin carpet very close to the bed. Given this uncertainty,the estimated transport at S4 was directed onshore duringcycle 2 and the beginning of cycle 3. Close to high tide duringcycle 3, however, a transport reversal occurred at this stationand offshore-directed sediment fluxes became very large.During tidal cycle 4, the transport again became onshore di-rected.

Given that the direction of the cross-shore sediment trans-port at station S3 depended on water depth, the total trans-port rates at S3 were correlated against local water depth, h,and relative wave height for surf-zone conditions (Hs /h .0.4); see Figure 12. Apparently, there is some form of rela-tionship between transport rate and water depth, or relativewave height, and both regressions are significant at a 5 0.05.The functional dependencies are not convincing, however, be-

256 Aagaard et al.


Figure 12. Net sediment-transport rates obtained under surf-zone con-ditions at S3 plotted against local water depth (upper panel) and relativewave height (lower panel). The lines of best fit are indicated by the dashedlines. Coefficients of determination for the linear fits are r2 5 0.127 and0.256, respectively.

Figure 13. Absolute values of net suspended sediment transport (thesum of mean, incident, and infragravity fluxes) estimated at stations S1–S3, plotted as a function of local relative wave height.

Figure 14. Normalized cross-shore suspended sediment-transport ratesdue to mean currents, incident waves, and infragravity waves recordedduring hours 162.3 and 187.7. Positive transports are directed onshore.The beach profiles are shown in the lower panels for reference.

cause the linear fits only explain 13 and 26% of the variancein sediment transport, respectively.

At the two lower stations (S1 and S2) further down theseaward slope of the bar, the suspended sediment transportwas consistently directed offshore. The only exception oc-curred when water levels became very low at station S2, suchthat this station was located in the inner surf zone, and rel-atively large onshore-directed transport rates were recordedbriefly on the ebbing tide. Overall, the data (Figure 11) in-dicate that sediment-transport rates were about a factor offive larger in the inner surf and swash zones (stations S3,S4) than in the shoaling/outer surf zones (stations S1, S2).The average estimated suspended sediment-transport rates(absolute values) were: S1: 0.077 kgm22 s21; S2: 0.061 kgm22

s21; S3 (upper instrument array only to provide a compari-son): 0.254 kgm22 s21; S4: 0.464 kgm22 s21.

The impact of relative wave height/intensity of wave break-ing on cross-shore suspended sediment-transport rate is fur-ther illustrated in Figure 13, which plots absolute values ofsuspended sediment transports estimated at S1–S3 as a func-tion of relative wave height. For nonbreaking wave conditions(Hs /h less than 0.35), absolute net transports remain small

whereas maximum recorded net transports increase abruptlyat the onset of wave breaking (Hs /h . ø 0.35). Even thoughnet transports can still remain small under intensely break-ing wave conditions due to the balancing effects of mean andoscillatory fluxes (OSBORNE and ROOKER, 1999; see also Fig-ure 14), there is a generally increasing trend in net sedimenttransport with increasing relative wave height.

Integrated over time, there was a suspended sediment-transport divergence between stations S2 and S3, with theformer characterized by (small) seaward-directed transports

257Intertidal Bars


Figure 15. Cospectra of sediment concentration and cross-shore oscil-latory velocity at stations S3 and S4 recorded during the rising tide (hour173, dashed lines) and high tide (hour 176, solid lines) of tidal cycle 3.Optical sensor elevations above the bed are noted in the figure. The co-spectra have 50 degrees of freedom.

and the upper stations exhibiting a landward-directed trans-port (Figure 11). This sediment-transport pattern is consis-tent with the bar form migration and the runnel infilling. Thenet calculated sediment transport (summed over the threetidal cycles) at station S3 was 11130 kg/m (corresponding to0.71 m3 m21), which closely corresponds to the amount ofsand deposited landward of that station.

At station S3, the largest cross-shore transport rates oc-curred when swash conditions prevailed (Figure 11), and atthose times, the transport was directed landward. Landward-directed transport also persisted for a significant part of thetime when the station was subjected to surf-zone conditionseven though mean currents were directed offshore; the land-ward sediment transport was driven by waves at both inci-dent and infragravity frequencies. Only at high tide did themean currents become sufficiently important to cause a netseaward-directed transport. Figure 14 illustrates the relativesignificance of mean currents, incident and infragravitywaves to the net cross-shore sediment transport for two ex-ample instrument records close to high tide. The normalizedtransport rate due to incident waves during an instrumentrecord was computed as

qincQ 5 (1)inc zq z 1 zq z 1 zq zinc ig mean

where qinc, qig, and qmean are the sediment-transport rates ac-complished by incident waves, infragravity waves, and meancurrents, respectively. Normalized transport rates due to in-fragravity waves and mean currents were computed accord-ingly.

Figure 14 indicates that sediment transport at the lowerstations in the shoaling wave and outer surf zones was dom-inated by the mean currents, which contributed about 80%of the total transport rate. Note, however, that, because sus-pended sediment concentrations at stations S1 and S2 weresmall, the net sediment-transport rates were also small. Atstation S3, onshore sediment fluxes due to incident and in-fragravity wave action balanced, or exceeded, the offshoresediment flux due to the undertow. At the uppermost station,S4, all transport components were onshore directed at thetime when mean currents were due to the cell circulation.When the undertow occurred at the upper station, the off-shore sediment flux caused by this current was balanced byan oscillatory onshore-directed flux. Interestingly, incidentwaves contributed increasingly large proportions of the totaltransport as the shoreline was approached, possibly becauseof offshore wave-stroke attenuation due to flow diversionalong the runnel, while the infragravity contribution waslargest around station S3. In summary, the net onshore-di-rected sediment transport at S3 and S4 appears to have beendriven by swash processes at low tide and mainly by oscil-latory wave motions at high tide.

Figure 14 provides a general impression of the relative im-portance of the different sediment-transport mechanismsacross the intertidal bar, but exceptions to that pattern didexist. As mentioned earlier, the net transport at station S4momentarily reversed from onshore to offshore and increaseddramatically around hour 176 (Figure 11). This was due to asudden reversal in the direction of the sediment flux due to

infragravity waves (Figure 15). The infragravity transportrate also increased significantly and, during hour 176, it con-tributed about 60% of the total sediment transport at stationS4. A simultaneous switch in infragravity transport directionalso occurred at station S3 (Figure 15).

BUTT and RUSSELL (1999) suggested that infragravitytransport direction could depend on the higher order mo-ments of the oscillatory infragravity velocity field, such asvelocity or acceleration skewness. This may not have beenthe case here, however. Normalized velocity skewness can becomputed as S 5 u3/(u2)1.5 and acceleration skewness as A 5a3/(a2)1.5, where a 5 du/dt (BUTT and RUSSELL, 1999). Infra-gravity velocity and acceleration skewnesses were computedfrom low-passed velocity records with a high frequency cut-off of 0.067 Hz (Table 1). At both stations, velocity skewness-es were consistently negative and no convincing relationshipwas apparent between the skewness magnitude and the in-fragravity fluxes. With respect to the acceleration skewness,this was at least an order of magnitude smaller than the ve-

258 Aagaard et al.


Table 1. Normalized infragravity velocity skewness (S) and accelerationskewness (A) for low-tide records (hour 173) and high-tide records (hour176).

Station S A

S3, hour 173hour 176

S4, hour 173hour 176

21.41921.43620.45021.461

20.00220.03020.00220.136

locity skewness. Negative acceleration skewness did increasesignificantly when infragravity transport reversed offshore,but because acceleration skewness was consistently negative,it is difficult to convincingly attribute the observed transportreversal to increased negative accelerations.

DISCUSSION

This field experiment demonstrated an example of onshoremigration of an intertidal bar with subsequent bar weldingto the beach and the development of a planar intertidal beachprofile (Figure 6). The intertidal bar at Skallingen was of theslip-face ridge type (cf. WIJNBERG and KROON, 2002) and theoutcome of the bar evolution was a significant onshore sedi-ment supply from the nearshore zone to the beach.

Previously, AAGAARD et al. (1998a) reported observationsof sediment transport and hydrodynamics across a landward-migrating intertidal bar at Skallingen. Measurements werethen obtained at a single location on the seaward slope of theintertidal bar under more energetic conditions than encoun-tered in the present experiment. It was concluded that thelandward migration of that intertidal bar was mainly due toan onshore-directed sediment transport driven by the meancurrent, the direction of which depended on the presence orabsence of a runnel landward of the bar. In the study re-ported here, a much denser array of sensors was used, veloc-ity and sediment transport were measured very close to thebed, and similar conclusions on the mean current circulationwere reached: Onshore-directed currents persisted on the barcrest until the runnel closed, subsequent to which the under-tow extended landward of the bar crest. Similar onshore-di-rected mean currents have been observed in three-dimen-sional bar settings by DRøNEN et al. (1999) and KROON andDE BOER (2001).

In this experiment, however, the onshore directed meancurrents at the bar crest did not appear critically importantto the onshore migration of the bar and the current speedwas smaller than in the example reported by AAGAARD et al.(1998a), the reason probably being the lower wave-energylevels. Here, onshore sediment transport did prevail acrossthe upper seaward slope and crest of the bar, but it was main-ly caused by oscillatory wave motions under swash and innersurf-zone conditions (Figures 11 and 13). At the upper sea-ward slope of the bar, onshore sediment-transport rates oc-curred when mean water depths were less than approxi-mately 0.5 m or relative wave heights . ø0.7 (Figure 12).This trend was not entirely consistent at all stations; for ex-ample, large offshore transport rates developed at station S4

when h 5 0.1–0.15 m and Hs /h . 1. The reason was thatinfragravity transport momentarily became large and off-shore directed. The sudden and dramatic switch in infra-gravity sediment-transport direction and magnitude aroundhour 176 could not be confidently related to changes in infra-gravity velocity or acceleration skewness. Examination of thetime-series records suggests that the reversal may have hadless to do with the hydrodynamic forcing than with the pro-cesses of sediment resuspension. This is a topic of ongoingresearch but falls outside the scope of the present article.

Offshore transport across the upper seaward bar slopemainly occurred at high tide when h . 0.4 m and Hs /h # 0.7(Figure 12). To some extent, this supports observations byHOUSER and GREENWOOD (2003), who found landward- andseaward-directed sediment transports being separated forrelative wave heights ø0.54. However, the functional rela-tionship found here between transport rate and relative waveheight is certainly not convincing (r2 5 0.26) and other mech-anisms were clearly important to the transport rate and di-rection.

Under weakly breaking or shoaling waves (stations S1 andS2), the recorded suspended sediment transport was consis-tently directed offshore. The offshore transport under shoal-ing waves (station S1) is somewhat surprising but was dueto offshore-directed mean currents probably generated bybreaking across the subtidal bar located further seaward (e.g.,Figure 4). The main point is that estimated sediment-trans-port rates under shoaling and weakly breaking waves weregenerally about an order of magnitude smaller than transportrates in the inner surf and swash zones (Figure 13), mainlybecause suspended sediment concentrations in the water col-umn are small under such conditions (AAGAARD, BLACK, andGREENWOOD, 2002). This observation is of importance to thequestion whether there is any fundamental difference be-tween the mobile intertidal bars (slip-face ridge type) studiedhere and the low-amplitude quasi-static ridge-and-runneltype of bars, which mainly occur in meso-macrotidal settingsand with low-energy wave conditions (KING and WILLIAMS,1949; MULRENNAN, 1992; ORFORD and WRIGHT, 1978). Thepresent measurements suggest that inner surf/swash zoneconditions are required in order to generate large suspendedsediment concentrations and transport rates. In settings witha large tidal range and/or low waves, such conditions willgenerally last only a small fraction of each tidal cycle at aspecific bar. This could be the reason why ridge and runnelmorphology is not very mobile and do not develop a formasymmetry through landward migration.

If this interpretation is correct, then it is likely that manyintertidal bars or intertidal bar sequences may oscillate be-tween one type and the other, for example, through spring-neap tidal cycles or as incident wave energy varies tempo-rally on a seasonal cycle. It is therefore questionable whethera distinction should be made between slip-face ridges andlow-amplitude ridges (ridge-and-runnels). It would seemmore prudent to use the term intertidal bar for both bartypes. The term swash bar would also appear inappropriateas both swash and surf-zone processes are critical to the be-havior of the mobile intertidal bars.

259Intertidal Bars


CONCLUSIONS

One of the main mechanisms for shoreline progradation isthe migration of bars across the intertidal zone and their de-position on the beach face. This study has provided hydro-dynamic and suspended sediment-transport measurementsduring a beach accretion event of this type. Onshore bar mi-gration was achieved mainly by swash processes and by theoscillatory flows of both short and long waves under surf-zoneconditions. The transport competency of the onshore strokeof the waves was considerably larger than the competency ofthe offshore stroke because water transported over the barcrest was subsequently channeled alongshore in the runnel.This transport asymmetry and the large sediment fluxes nat-urally associated with surf and swash in very shallow water(, ø0.5 m), resulted in a relatively rapid bar-migration ratecorresponding to 10–20 m/d. Both the existence of the inter-tidal bar and its disappearance once it infilled the runnelproduced a strong feedback effect on the hydrodynamics andsediment dynamics. The presence of the runnel was associ-ated with horizontal cell circulation in the intertidal zone,characterized by onshore-directed mean flows across the barcrest, which augmented the sediment transport due to wavemotions. When the runnel was infilled, an offshore-directedundertow developed which opposed the wave-induced sedi-ment transport.

ACKNOWLEDGMENTS

We are grateful to Per Sørensen (the Danish Coastal Au-thority) and Erik Brenneche (Esbjerg Port Authorities) forgiving us access to offshore wave and tidal data, respectively.Ulf Thomas and Niels Vinther helped out in the field—undersunny conditions this time! This research was funded by theDanish Technical Sciences Research Council (grant99012287) and by the European Union through the Coast-View Project (contract EVK3-CT-2001-0054).

LITERATURE CITED

AAGAARD, T.; BLACK, K.P., and GREENWOOD, B., 2002. Cross-shoresuspended sediment transport in the surf zone: a field-based pa-rameterization. Marine Geology, 185, 283–302.

AAGAARD, T. and GREENWOOD, B., 1994. Suspended sediment trans-port and the role of infragravity waves in a barred surf zone. Ma-rine Geology, 118, 23–48.

AAGAARD, T.; NIELSEN, J., and GREENWOOD, B., 1998b. Suspendedsediment transport and nearshore bar formation on a shallow in-termediate-state beach. Marine Geology, 148, 203–225.

AAGAARD, T.; NIELSEN, J.; NIELSEN, N., and GREENWOOD, B.,1998a. Suspended sediment transport and morphological evolu-tion on an intertidal beach. Proceedings Coastal Dynamics ’97,Plymouth, ASCE, pp. 824–833.

BUTT, T. and RUSSELL, P., 1999. Suspended sediment transportmechanisms in high-energy swash. Marine Geology, 161, 361–375.

CHURCH, J.C. and THORNTON, E.B., 1993. Effects of breaking waveinduced turbulence within a longshore current model. Coastal En-gineering, 20, 1–28.

DAWSON, J.C.; DAVIDSON-ARNOTT, R.G.D., and OLLERHEAD, J.,2002. Low-energy morphodynamics of a ridge and runnel system.Journal of Coastal Research, Special Issue No. 36, pp. 198–215.

DRøNEN, N.; KARUNARATHNA, H.; FREDSøE, J.; SUMER, B.M., andDEIGAARD, R., 1999. The circulation over a longshore bar with ripchannels. Proceedings Coastal Sediments ’99, Long Island, ASCE,pp. 576–587.

GARCEZ-FARIA, A.F.; THORNTON, E.B.; LIPPMANN, T.C., and STAN-TON, T.P., 2000. Undertow over a barred beach, Journal of Geo-physical Research, 105, 16999–17010.

GREENWOOD, B.; AAGAARD, T., and NIELSEN, J., 2004. Swash barmorphodynamics in the Danish Wadden Sea: sand bed oscillationsand suspended sediment flux during an accretionary phase of theforeshore cycle. Danish Journal of Geography, 104(1), 15–30.

GREENWOOD, B. and DAVIDSON-ARNOTT, R.G.D., 1979. Sedimenta-tion and equilibrium in wave-formed bars: a review and case-study. Canadian Journal of Earth Sciences, 16, 312–332.

HAAS, K.A.; SVENDSEN, I.A.; HALLER, M.C., and ZHAO, Q., 2003.Quasi-three-dimensional modeling of rip current systems. Journalof Geophysical Research, 108, C7, 10.1–10.21.

HOUSER, C. and GREENWOOD, B., 2003. Response of a swash bar toa sequence of storm events. Proceedings Coastal Sediments ’03,Clearwater, Florida, USA. CD-ROM published by East MeetsWest Productions, 13 p.

HUGHES, M.G and BALDOCK, T.E., 2004. Eulerian flow velocities inthe swash zone: field data and model predictions. Journal of Geo-physical Research, 109, C08009, DOI:10.1029/2003JC002213.

KING, C.A.M. and WILLIAMS, W.W., 1949. The formation and move-ment of sand bars by wave action. Geographical Journal, 113, 68–85.

KROON, A. and DE BOER, A., 2001. Horizontal flow circulation on amixed energy beach. Proceedings Coastal Dynamics ’01, Lund,ASCE, pp. 548–557.

KROON, A. and MASSELINK, G., 2002. Morphodynamics of intertidalbar morphology on a macrotidal beach under low-energy wave con-ditions, North Lincolnshire, England. Marine Geology, 190, 591–608.

MASSELINK, G. and SHORT, A.D., 1993. The effect of tide range onbeach morphodynamics: a conceptual model. Journal of CoastalResearch, 9, 785–800.

MULRENNAN, M.E., 1992. Ridge and runnel beach morphodynamics:an example from the central east coast of Ireland. Journal ofCoastal Research, 8, 906–918.

ORFORD, J.D. and WRIGHT, P., 1978. What’s in a name—descriptiveor genetic implications of ‘‘ridge and runnel’’ topography. MarineGeology, 28, M1-M8.

OSBORNE, P.D. and ROOKER, G.A., 1999. Sand re-suspension eventsin a high energy infragravity swash zone. Journal of Coastal Re-search, 15, 74–86.

RAUBENHEIMER, B., 2002. Observations and predictions of fluid ve-locities in the surf and swash zones. Journal of Geophysical Re-search, 107, C11, 11.1–11.7

SOULSBY, R.L., 1999. Coastal sediment transport: the COAST3Dproject. Proceedings ICCE98, Copenhagen, ASCE, pp. 2548–2558.

STEPANIAN, A.; VLASWINKEL, B.; LEVOY, F., and LARSONNEUR, C.,2001. Sediment transport on a macrotidal ridge and runnel beachduring accretionary conditions. Proceedings Coastal Dynamics ’01,Lund, ASCE, pp. 1017–1027.

VOULGARIS, G.; SIMMONDS, D.; MICHEL, D.; HOWA, H.; COLLINS,M.B., and HUNTLEY, D.A., 1998. Measuring and modelling sedi-ment transport on a macrotidal ridge and runnel beach: an inter-comparison. Journal of Coastal Research, 14, 315–330.

WIJNBERG, K.M. and KROON, A., 2002. Barred beaches. Geomor-phology, 48, 103–120.

Hydrodynamics and Sediment Fluxes across an Onshore Migrating Intertidal Bar

Documents