GEOPHYSICAL RESEARCH LETTERS, VOL. 25, NO.18, PAGES 3533-3536, SEPTEMBER 15, 1998

Sediment resuspensionin the wakes of Hurricanes Edouard and Hortense

T. D. Dickey,• G. C. Chang,•Y. C. Agrawal,2A. J. Williams,III,3 andP.S. Hill4

Abstract. A unique set of physicaland optical observationsof The presentsite is locatedin a region known as the "Mud sedimentresuspension was obtainedduring the passageof two Patch"on the Mid-Atlantic Bight shelf off the eastcoast of the hurricanes,Edouard and Hortense. The eyes of thesehurricanes U.S. (Fig. 1) where extensiveobservations have been made in the passedwithin approximately110 and 350 km, respectively,of past(e.g., Biscaye et al. [1988]). The Mud Patchis a depositof 2 our study site on the continentalshelf (100 km south of Cape to 14 m thicknesscovering an areaof roughly100 km by 200 km Cod, )within a two-week period in the fall of composedof relativelyuniform fine-grained material overlying 1996. Sedimentswere resuspendedto more than 30 m abovethe coarsersand-sized sediment [Twichelt et at., 1987]. The oceanbottom during both hurricanes. The sedimentresuspension sedimentconsists primarily of fine-texturedmaterial. processesassociated with the two hurricanesare shownto differ primarilybecause of the separationdistances between the eyesof Methods the hurricanesand the observationalsite (e.g., local versusremote forcing).Observed particle-size distributions were shiftedtoward The presentexperiment consisted of four deployments smaller scalesduring HurricaneEdouard because of flocculate (spanningsummer 1996 through summer 1997) of a mooringand disaggregationcaused by high levels of localized shear and two bottomtripods located at approximately40.5øN, 70.5øW in turbulence. watersof about70 m depth(Fig. 1). The time seriesdescribed herewere obtained during the first deployment period. Introduction Several physical and optical measurementsystems were deployedon the mooring(Fig. 1). Theseincluded bio-optical The present report describes physical and optical systems(BIOPS) (as described in Changet al. [1997];Dickey et measurementsrelevant to sedimentresuspension on a continental al. [1998]),which were placed at 14, 37, and 52 rn depthsto shelf during the passageof two hurricaneswithin a two-week quantifyoptical and physicalvariability. A subsetof BIOPS period. Our resultsare especiallyimportant as few observations sensorsused for thisstudy included 1) ac-9s,2) fluorometers,and of resuspensionhave been made under extreme atmospheric 3) temperaturesensors. The samplingrate for the ac-9 wasonce forcing (e.g., Madsen et al. [1993]). Such observationsare of per hour for 30 secondsand the samplingrate for all other greatinterest as they enablethe studyof episodicprocesses that sensorswas eight times per hour. Vector winds were measured modify the near-bottomenvironment and control the fate of from a nearbysurface buoy with an anemometerat 3.3 m above bottommaterials including sediments, organisms, and pollutants the surface.In addition,wave data were measured by National that are transported,deposited, and transformedon continental Data Buoy Center (NDBC) buoy 44008, locatedat 40.5øN, shelves(e.g., Churchill [1989]; Cacchioneand Drake [1990]; 69.43øW(90 km east of theCMO site). Nittrouer and Wright [1994]). Physical processesthat can A nearby optical bottom tripod consistedof a suite of contributeto sedimentresuspension and transportinclude wind- instrumentsdesigned to measureproperties of particles(Fig. 1). generatedsurface waves, internal gravity and solitary waves, The BIOPS, as describedabove, was mounted at 1.5 meters tides, mean interior and boundarylayer currents,eddies, and abovethe bottom (mab) (-68 m depth).A photographiccamera buoyancy-drivenplumes (e.g., Cacchioneand Drake [1990]; was placedon the tripod at 1 mab and was usedto measuresize Nittrouer and Wright [1994]). Further, there is a complex distributionsof particleslarger than 500 gm in the bottom interplaybetween physical, geological, and biologicalprocesses boundary layer. The final set of instrumentsmeasured true involving particlesand organismsof varying size, shape,and volumetric concentrationof fine particlesderived from size buoyancy(e.g., Nowell [1983]). These processesresult in a distributionusing the LISST-100 [Agrawal and Pottsmith, 1994]. changein particle-sizedistributions. The presentmeasurements The LISST-100, measuringthe size distributionat 1.6 mab, was provide insights into the dynamical processesthat result in sampledat 15-minuteintervals and on the hour. sedimentresuspension during intenseforcing and the relaxation The otherbottom tripod (also located at a depth'of70 m and to more normalnear-bottom conditions, as well as variability in about400 m from the mooring)was used to measurenear-bottom near-bottomparticle-size distributions. These observations allow currents.BASS currentmeters [Williams et at., 1987] recorded investigatorsto testmodel hypotheses and performance. vector velocities at seven depthsfrom 0.38 to 7 mab. These measurements were made on the hour and half-hour with •OceanPhysics Laboratory, University of Californiaat SantaBarbara, samplingat 2 Hz for 28 min and 49 sec. Mean currentbottom Santa Barbara, California. shearstress (xc), wave-orbital velocity (u,,), current-wavebottom 2SequoiaScientific, Inc., Mercer Island, Washington. shearstress (To+,,), and dissipation rate (•½+,,) were computed using 3WoodsHole Oceanographic Institution, Woods Hole, Massachusetts. time seriesvelocity data collected by the BASS currentmeters to 4Departmentof Oceanography,Dalhousie University, Halifax, Nova characterizethe near-bottomflow field responsiblefor sediment Scotia, Canada. resuspension.Wave-orbital velocity was estimated by takingthe Copyright1998 by the AmericanGeophysical Union. squareroot of the integralof energyunder the wavepeak in the velocityspectra (periods from 8 to 25 s). Mean currentbottom Papernumber 98GL02635. shearstress was computed from BASS currentmeter data and the 0094-8534/98/98GL-02635505.00 "law of thewall" formulation[Tennekes and Lumley, 1972].

3533 3534 DICKEY ET AL.: SEDIMENT RESUSPENSION IN THE WAKE OF TWO HURRICANES

ß m s-] (Fig. 2a). The water columnwas highly stratifiedbefore the 10m

I passageof Edouard as indicatedby the temperaturetime series shownin Fig. 2b. The mixed-layerdepth (MLD; computedusing 14m! BIOPb a 1ø temperaturedifference criterion) deepened from 12 m to a depthof 67 m shortlyafter the passageof Edouard(Fig. 2b). The ! I generalmixing effect of the hurricaneled to entrainmentof cool, I deeperwaters into the upperlaye? and to mixing of upperlayer warm waterswith cooler,deeper waters. The warming of deeper 37m[ BIOP$ waters (see 68 m record) continued for about 10 days when ! coolingbegan again. i The physicaleffects of HurricaneHortense at the mooring site I are less obvious than those associated with Edouard because of the much greaterdistance between the eye of Hortense and the site, and possiblybecause of the greater translationalspeed of Hortense. Winds associated with were more I I sporadicat the mooring site, occasionallyexceeding 10-15 m s4 from September 14 through September 19, 1996. Hortense 68mBIOPS on• resulted in further mixing of the water column with the MLD increasingfrom 18 m to near 50 m. The passageof Hortenseis more evidentin the sedimentresuspension event (seenin optical from time series)than in the physicalrecord. The time series of beam attenuation coefficient at 676 nm Figure 1. Geographicmap indicatingsite of mooringand tripods used for the presentstudy. Map showingtracks of Hurricanes (hereafterbeam c) collectedwith the ac-9 on BIOPS packagesat Edouard and Hortense. Mooring and optical tripod schematic 37, 52, and 68 m depthsare illustratedin Figs. 2c, d, and e (note diagrams. differences in ordinate scales for the three panels). This wavelengthwas selectedfor beam c becauseof its covariance with suspendedparticulate matter and its minimal absorptiondue A number of models have been formulated to characterize to dissolved humic acids. Sediment resuspensioncaused by near-bottomcombined wave and current flow (e.g., Grant and Madsen [1979], and Christoffersenand Jonsson[1985]). More advancedversions of bottom boundarylayer modelsthat include effectsof armoring,moveable beds, and/or stratifiedflows have beendescribed by Trowbridgeand Madsen [1984] and Glenn and Grant [1987]. These advancedversions were not used in the presentstudy due to the absenceof bottomsediment condition data and the presenceof a well-mixed bottom boundarylayer. The Christoffersenand Jonsson [1985] model was chosen to computebottom shearstress from combinedcurrent and wave motionat 0.38 mab. The modelis quitesimilar to the commonly

used Grant and Madsen [1979] model, but utilizes an iterative 68m approachfor computingfriction factors. Dissipationrate was estimatedusing the relationshipdescribed in Grosset al. [ 1994]. •..2[ c. 37m E [ H Observations

o.5 The present observations were obtained during the deploymentperiod from July 8 throughSeptember 26, 1996. On o August 29, 1996, Hurricane Edouard'swinds reachedtheir maximumspeed of 140 mph (category4 on the Saffir-Simpson HurricaneScale), and the translationalspeed peaked at 1000 km

d4 approximately700 km south of the mooringsite (Fig. 1). 30 • , Hurricane Edouard maintained category 4 status for 2O approximatelyseven days. HurricaneHortense reached category 4 statuson September13, 1996, roughly 800 km south of the mooringsite, but decreasedin intensityto category3 on the same 235 240 245 250 255 260 265 day. At that point, winds were 120 mph, and the translational Aug22 Sept1 Sept11 Sept21 Year Day (1996) speedwas 1500 km d'l. The eye of HurricaneEdouard passed within about 110 km of the mooring on September2, 1996 and Figure 2. a. Time seriesof wind speed. Passagesof Hurricane the eye of HurricaneHortense passed within 350 km of the on Edouard (E) and Hurricane Hortense(H) are indicated. b. Time September14, 1996. The physicaland optical effects of the series of temperature(sampling interval is 7 5 min) collected at passagesof Hurricane Edouard and Hurricane Hortense are 14, 37, 52, and 68 m depths. c. Time seriesof hourly average illustratedin the time seriesshown in Fig. 2. beam attenuationcoefficient (676 nm) at 37 m depth.d. Sameas Wind speed(10 m abovethe surface)did not exceed10 m s'l c. for 52 m depth. e. Same as c. for 68 m depth. Note until the passageof HurricaneEdouard when winds peaked at 28 differences in ordinate scalesbetween c, d, and e. DICKEY ET AL.: SEDIMENT RESUSPENSION IN THE WAKE OF TWO HURRICANES 3535

Hurricane Edouard first became evident in the beam c record 41a 0 38 mab whenthe stormwas greater than 900 km from the CMO site. The greatestbeam c values(up to about30 m'•) are seenin the near bottom (68 m) time seriesduring the time of closestpassage, howeverthe signalis clearlyseen in the 37 m depthrecord (up to 235 240 245 250 255 260 265 2 m'•) as well. The beamc valuesare reducedby abouta factor 25 of five between 68 and 52 m and by about a factor of two between 52 and 37 m. The NDBC buoy shows that the A20tb038mab j• significantwave heightexceeded 9 m at this time. It is worth notingthat the onsetof the increasedbeam c valuesoccurred first at 68 m, then about a half-day later at 52 m, and finally an 235 240 245 250 •55- additionalhalf-day later at 37 m (Fig. 2). The return of beam c 051c 038mab valuesto pre-hurricanelevels was slowerat the shallowerdepths than at the deeper depths. This is consistentwith size distributionsbiased toward smaller particles (with lower settling velocities)at the shallowerdepths. 0 ...... ^ ..,,,._•A,...J Figure 3 showscurves for size distributionstaken six hours 235 240 245 250 255 260 265 apart,beginning 12 hoursbefore the peak in HurricaneEdouard- 20 d 2 mab induced bottom pressure. The volume distributionsclearly •. indicatethe presence,disappearance, and reestablishment of large particlesin the water columnas the intensityof the hurricane- induced bottom currents and oscillations increased to their o 235 240 245 250 255 260 265 Aug22 Sept 1 Sept 11 Sept21 maximumand then weakenedas the hurricanemoved away from Year Day (1996) the bottomtripod. Thesedata supportthe photographicresults Figure 4. a. Time seriesof current-wavebottom shear stressat that show high levels of localizedshear and turbulence(e.g., 0.38 mab computedusing near-bottomcurrent measurements turbulentdissipation rate) in the water column, due to currents (half-hour averages),wave orbital velocity, and the model and waves, broke up flocculates. These flocs then rapidly presentedin Christoffersenand Jonsson (1985). b. Time seriesof reappearedas the intensityof turbulenceweakened. wave-orbitalvelocity at 0.38 mab. c. Time seriesof dissipation The greatestincreases in beamc did not coincidewith the time rate at 0.38 mab calculatedusing friction velocity derivedfrom of closestpassage of Hortense. At the time of closestpassage, the Christoffersenand Jonsson(1985) model. d. Beam windspeaked at 11 m s4 and significantwave heightwas 4 m at attenuationcoefficient (676 nm) at 68 m (2 mab). the NDBC buoy site. The cause(s)for the abrupt dip and then rise in the beam c reeordat 68 m (Fig. 2e), is not obvious.One hypothesisis that the sedimentscould have been resuspended some distanceaway from the site and then advectedpast the The current-wave interaction has been shown to be a non-linear mooring. This possibilityseems unlikely in that the increasein phenomenon(e.g., Grantand Madsen[1979]; Christoffersenand beamc did not occurat all depthsat the sametime. The signal Jonsson[1985]; Glenn and Grant [1987]). The frequency was first observed at the bottom, then about three-fourths of a differences between waves and mean currents result in an day later at 52 m, and then an additionalthree-fourths of a day oscillatorywave boundary layer within a relativelysteady current later at 37 m. It is possiblethat two trains of wind-generated boundarylayer. The smallscale of the wave boundarylayer surfacewaves could have caused the successivepeaks. causes the wave boundary shear stress to be an order of The principlephysical process that determinesthe structureof magnitudegreater than that of the shearstress a•sociated with a the near-bottom flow field on a continental shelf is the interaction currentof comparablemagnitude. The suspensionof sediment of surfacewaves with mean currents[Glenn and Grant, 1987]. by waves in the absenceof strong,subcritical currents is not unexpected(e.g., Grant and Madsen [1979]; Glenn and Grant [ 1987]);it is associatedwith thetime-varying part of shearstress, which has oscillatoryamplitudes clearly in excessof critical

a 0745, September2 45 values. To characterizethe near-bottomflow field, combined

b 1345, September2 current-wave shear stresses must be considered. 4 c 1945, September2 (Peak Edouard) Sedimentis expectedto be resuspendedwhen bottom shear d 0145, September3 d stressesexceed a "critical"shear stress [Twichell et al, 1987]. '• 3 e0745, September 3 / The critical combined current-wave shear stress was first exceededat the time of increasedsediment resuspension during the passageof HurricaneEdouard (Fig. 4a). Current-waveshear

215 stressvalues periodically exceeded the criticalvalue (up to a f maximumof about3.5 dynescm '2 duringthe time of greatest 1 sedimentresuspension) for aboutfive daysfollowing the peakof O5 the Hurricane Edouard winds. Mean current shear stress

lO1 lO• lO3 exceededthe criticalvalue during the passage of Edouardas well, ParticleD•ameter (lam) with xc exceeding2.5 dynescm '2 (data not shown). Sediment Figure 3. Six curvesof volumesize distributionof particles(gl resuspensionassociated with Edouardwas possibly due to intense 1-•) at six hour intervals,beginning 12 hoursprior to the peak of bottom mean currents with little interaction with waves. In Hurricane Edouard-inducedbottom rms pressure (1945 on contrast, the sediment resuspensioncaused by Hurricane September2). Hortenseoccurred when boundaryshear stresses associated with 3536 DICKEY ET AL.: SEDIMENT RESUSPENSION IN THE WAKE OF TWO HURRICANES currents were below critical values and current-wave shear References stresseswere high. The current-wavebottom shearstress reached 2 dynescm '2. Agrawal,Y. C., andH. C. Pottsmith,Laser diffraction particle sizing in STRESS, Cont. Shelf Res., 14, 1101-1121, 1994. Hurricane Edouard resulted in a ten-fold increase in wave- Biscaye,P. E., R. F. Anderson,and B. L. Deck, Fluxesof particlesand orbital velocity (u,,) at a depthof 0.38 mab (Fig. 4b); u,, increased constituentsto the easternUnited Statescontinental slope and rise: from approximately2 to 15 cm s'• with the passageof Hurricane SEEP I, Cont. Shelf Res., 8, 855-904, 1988. Hortense. Hurricane Edouard resultedin values of dissipation Cacchione,D. A., andD. E. Drake,Shelf sediment transport, in TheSea, rate (• ..... at 0.38 mab) increasing20-fold (Fig. 4c). Dissipation 9, editedby B. LeMehauteand D.M. Hanes,pp. 729-774,John Wiley, ratesfollowing the passageof HurricaneHortense also increased , NY, 1990. by a factor of 20. For both hurricanes,peaks in •,+,, and u,, Chang,G. C., T. D. Dickey, D. V. Manov, D. E. Sigurdson,and J. D. coincided with the highest sedimentresuspension seen in the McNeil, Data report:Coastal Mixing andOptics deployment I, OPL- 4-97, Ocean Physics Laboratory, University of California, Santa beam c records(Fig. 4). Barbara, CA 1997. Regressionanalysis between mean currentbottom shear stress Christoffersen,J. B., andI. G. Jonsson,Bed frictionand dissipation in a (0.38 mab) and beam c (68 m depth) resultedin a correlationof combinedcurrent and wave motion, Ocean Eng., 12, 387-423, 1985. 0.68 over the periodof passageof HurricaneEdouard (August 22 Churchill, J. H., The effect of commercialtrawling on sediment through September11, 1996). The correlationbetween x, (0.38 resuspensionand transportover the Middle Atlantic Bight continental mab) and beam c (68 m depth) for the period of Hurricane shelf,Cont. Shelf Res., 9, 841-864, 1989. Hortense(September 11 throughSeptember 21, 1996) was 0.007. Dickey,T., D. Frye, H. Jannasch,E. Boyle,D. Manov,D. Sigurdson,J. Regression analysis performed for combined current-wave McNeil,M. Stramska,A. Michaels,N. Nelson,D. Siegel,G. Chang, J. Wu, and A. Knap, Initial results from the Testbed bottomshear stress (0.38 mab) and beam c (68 m) over the period Mooringprogram, Deep-Sea Res., 45, 771-794, 1998. of passageof Edouardgave a correlationof 0.80. A correlation Glenn, S. M., and W. D. Grant, A suspendedsediment stratification of 0.75 was found betweenx,+,, (0.38 mab) and beam c (68 m) correctionfor combinedwave and current flows,d. Geophys. Res., 92, duringthe passageof HurricaneHortense. These analyses suggest 8244-8264, 1987. that the sediment resuspensionassociated with Hurricane Grant, W. D., and O. S. Madsen, Combined wave and current interaction Edouardwas primarily forced by currentswith lesserinteraction with a roughbottom, d. Geophys,.Res., 84, 1797-1808,1979. with waves. In contrast, Hurricane Hortense sediment Gross,T. F., A. J. Williams,and E. A. Terray,Bottom boundary layer resuspensionwas likely dominatedby wave interactionwith little spectraldissipation estimates in the presenceof wave motions,Cont. Shelf Res., 14, 1239-1256, 1994. influence from mean currents. Madsen,O. S., L. D. Wright, J. D. Boon, and T. A. Chisholm,Wind stress,bed roughness,and sedimentsuspension on the inner shelf Summary duringan extremestorm event, Cont. ShelfRes., 13, 1303-1324,1993. Nittrouer, C. A., and L. D. Wright, Transportof particles across The presentcomprehensive physical and opticalmeasurements continentalshelves, Rev. of Geophys.,32, 85-113, 1994. have capturedthe impactsof two closelyspaced hurricanes. The Nowell, A. R. M., Effects of biologicalactivity on theentrainment of marine sediments,Mar. Geol., 42, 133-153, 1983. sedimentresuspension associated with Hurricane Edouard was Tennekes,H.. andJ. L. Lumley,A First Coursein Turbulence,300 pp., primarily locally driven through near-bottomcurrent processes The MIT Press,Cambridge, MA, 1972. with secondaryinfluence from waves. In contrast,Hurricane Trowbridge,J. H., andO. S. Madsen,Turbulent wave boundary layers, 1, Hortensepassed much further from the site, and causedsediment Model formulationand first-ordersolution, d. Geophys.Res., 89, resuspensionthrough remote forcing via near-bottom wave 7989-7997, 1984. processes(i.e., surfacegravity waves generatedsome distance Twichell, D.C., B. Butman,and R. S. Lewis, Shallowstructure, surficial away) with comparativelylittle interactionwith mean currents. geology,and the processescurrently shaping the bank,in Georges Evidence for this can be found in the correlation between mean Bank,edited by R.H. Backus,pp. 31-38,MIT Press,Cambridge, MA, 1987. current bottom shear stress and beam c, and between combined Wiiliams,A. J.,3rd, J. S.Tochko, R. L. Koehler,W. D. Grant,T. F. current-wave bottom shear stress and beam c. Our measurements Gross, and C. V. R. Dunn, Measurementof turbulence in the oceanic show that particleswere suspendedmore than 30 rn above the bottomboundary layer with an acousticcurrent meter array,d. Atmos. oceanbottom during both hurricanes.Volume size distributions and Ocean. Tech.,4, 312-327, 1987. were biased towards small particles due to fiocculate disaggregation during the peak of Hurricane Edouard. Throughoutthe periods of HurricanesEdouard and Hortense, Y. C. Agrawal,Sequoia Scientific, Inc., P.O. Box 592, MercerIsland, sedimentresuspension occurred in accordancewith the critical WA 98040 (e-mail:yøgi•sequøiasci'cøm) shear stress criterion for combined wave and current shear stress. G. C. Changand T. D. Dickey,Ocean Physics Laboratory, University of California at Santa Barbara,6487 Calle Real Unit A, Goleta, CA 93117(e-mail: grace•icess.ucsb.edu; tommy•icess.ucsb.edu) Acknowledgments. This work was supportedby the Office of Naval P.S. Hill, Departmentof Oceanography,Dalhousie University, Researchas part of the CoastalMixing and Optics program.We would Halifax,Nova Scotia, Canada B3H 4Jl (e-mail:phill•phys.ocean.dal.ca) alsolike to thankJohn Trowbridge, Steve Lentz, JoeMcNeil andZenitha A. J. Williams,3 rd, Bigelow 110- ms 12, AOP&E, WoodsHole Chroneerfor their assistancewith data analysisand developmentof the OceanographicInstitution, 98 WaterStreet, Woods Hole, MA 02543 (e- manuscript;David Porterand DonaldThompson for their satelliteimages mail: awilliams•whoi.edu) and analyses;Murray Levine who sharedthe mooringplatform with us; and Derek Manov, David Sigurdson,Erin Lutrick, and ChuckPottsmith (ReceivedApril 6, 1998;revised July 30, 1998; for their engineeringand technicalsupport. acceptedAugust 7, 1998)