Cross-Shelf SedimentTransport by Anticyclonic Eddy Off ...washburn/docs/washburn_et_al_1993_science.pdf · transported mainly along continental shelves. Cross-shelfflows are muchweaker,

"finger" happens to "find" the ion. Theexistence of these "fingers" was proved inmy earlier study of the neat surface (7), andFig. 3 shows an example of such a "finger"interacting with the ion.

Finally, I have considered (but do notshow) the reverse process, namely, the ionstarting in the middle of the water phase (Z- -12 A) and the polarity of the electricfield reversed so that the ion will climb thefree energy "hill." The interesting featurehere is that the interaction between thewater and the ion is still quite appreciableeven when the ion reaches the "bulk" ofthe DCE phase (as judged by the ion'slocation Z). This clearly shows that the ioncarries at least part of the hydration shell.An examination of the animated trajecto-ries shows that the structure of the interfacebecomes highly disordered, the effectivesurface region (the region where the densityof each liquid is within 10 to 90% of thebulk value) is broadened, and liquid capil-laries longer than the one characteristic ofthe neat surface are observed.

The molecular model described abovehas provided both a detailed molecularpicture of ion transport dynamics and an-swers to long-standing questions. In partic-ular, direct, molecularly based evidence forthe existence of a barrier for the transferprocess has been given, and the roles ofsurface roughness and capillary fluctuationsin the ion transfer have been stressed.However, there are several other key issuesthat this model is capable of addressing. Forexample, the role of ion pairs in facilitatingthe transport has been suggested in theliterature, and calculations on ion-pairtransport similar to the one reported abovewould be desirable. One important step isto perform comprehensive calculations onseveral ions with increasing size. This willallow the investigation of the experimen-tally relevant case of ions whose net freeenergy of transfer is close to zero. Theresults of these calculations will be reportedelsewhere (9).

REFERENCES AND NOTES

1. H. H. J. Girault and D. J. Schiffrin, in Electroana-lytical Chemistry, A. J. Bard, Ed. (Dekker, NewYork, 1989), p. 1.

2. Interfacial kinetics in solution" [Faraday Discuss.Chem. Soc. 77 (1984)].

3. B. H. Honig, W. L. Hubbell, R. F. Flewelling, Annu.Rev. Biophys. Biophys. Chem. 15, 163 (1986).

4. Y. Y. Gurevich and Y. I. Kharkats, J. Electroanal.Chem. 200, 3 (1986); Z. Samec, Y. I. Kharkats, Y.Y. Gurevich, ibid. 204, 257 (1986).

5. A. M. Kuznetsov and Y. I. Kharkats, in The InterfaceStructure and Electrochemical Processes at theBoundary Between Two Immiscible Liquids, V. E.Kazarinov, Ed. (Springer-Verlag, Berlin, 1987), p. 1 1.

6. G. J. Hanna and R. D. Noble, Chem. Rev. 85, 583(1985).

7. I. Benjamin, J. Chem. Phys. 97,1432 (1992).8. J.-P. Hansen and 1. R. McDonald, Theory of

Simple Liquids (Academic Press, London, ed. 2,1986), p. 179.

1560

9. K. J. Schweighofer and 1. Benjamin, in prepara-tion.

10. Y. Marcus, Ion Solvation (Wiley, New York, 1985),chap. 6.

11. For consistency, the dielectric constants used inthis model are calculated by molecular dynamicssimulation using the same Hamiltonian as the onethat was applied in this work.

12. H. Reiss, H. L. Frisch, J. L. Lebowitz, J. Chem.

Phys. 31, 369 (1959).13. W. C. Swope, H. C. Andersen, P. H. Berens, K. R.

Wilson, ibid. 76, 637 (1982).14. This work was supported by the National Science

Foundation (grants CHE-9015106 and CHE-9221580), the Petroleum Research Fund (ACS-PRF-22862), and the University of California, Santa Cruz.

5 March 1993; accepted 26 July 1993

Cross-Shelf Sediment Transport by an AnticyclonicEddy Off Northern California

Libe Washburn, Mark S. Swenson, John L. Largier,P. Michael Kosro, Steven R. Ramp

A combination of satellite imagery, shipboard profiles, drifter tracks, and moored currentobservations reveals that an anticyclonic eddy off the coast of northern California trans-ported plumes of suspended sediments from the continental shelf into the deep ocean. Thehorizontal scale of the eddy was about 90 kilometers, and the eddy remained over thecontinental shelf and slope for about 2 months during the summer of 1988. The total massof sediments transported by the eddy was of order 105 metric tons. Mesoscale eddies arerecurrent features in this region and occur frequently in eastern boundary currents. Theseresults provide direct evidence that eddies export sediments from continental shelves.

Over continental margins, the rotation ofthe Earth and the presence of coastalboundaries cause ocean currents to bemainly parallel to isobaths, and, as a result,heat, mass, nutrients, and particles aretransported mainly along continentalshelves. Cross-shelf flows are much weaker,are intermittent, and consequently aremore poorly understood (1). However,property gradients across continentalshelves are much larger than gradients par-allel to continental shelves, and thus evenweak cross-shelf flows can produce largefluxes. Cross-shelf transport processes linkthe coastal environment and the deep seaand affect the global cycling of carbon andnitrogen (2). The movements of bottomsediments across continental shelves andthe export of sediments to the deep oceaninfluence the fate of primary productionand the transport of waste in the coastalocean, but specific transport mechanismsare difficult to identify (1). We report phys-ical and bio-optical oceanographic observa-tions (3, 4) that show an anticyclonicmesoscale eddy moving from the deepocean up onto the continental shelf offnorthern California during the summer of

L. Washburn, Center for Remote Sensing and Environ-mental Optics, Department of Geography, Universityof California, Santa Barbara, CA 93106.M. S. Swenson, National Oceanic and AtmosphericAdministration Atlantic Oceanographic and Meteoro-logical Laboratory, 4301 Rickenbacker Causeway,Miami, FL 33149.J. L. Largier, Scripps Institution of Oceanography, LaJolla, CA 92093.P. M. Kosro, College of Oceanography, Oregon StateUniversity, Corvallis, OR 97331.S. R. Ramp, Department of Oceanography, NavalPostgraduate School, Monterey, CA 93943.

SCIENCE * VOL. 261 * 17 SEPTEMBER 1993

1988. Layers of resuspended bottom sedi-ments extending as high as 50 m above thesea floor were advected by the eddy acrossthe shelf. Around the periphery of theeddy, sediment plumes extended tens ofkilometers beyond the continental shelfbreak.

The sediment plumes occurred within acoastal upwelling system off northern Cali-fomia between Point Reyes and Point Are-na (5). The distribution of sea surfacetemperature (SST) on 17 July 1988 [Julianday JD) 199] shows an extensive region ofcold water adjacent to the coast that result-ed from wind-driven upwelling (Fig. lA).The velocity field of the anticyclonic eddyin which the sediment plumes were ob-served is clearly evident as a rotary circula-tion pattern centered near 38039'N,123054'W (labeled A in Fig. 1A) (6). Theeddy was located inshore of a narrow, me-andering current jet that coincided with afilament of cold water and extended about300 km offshore (7). The anticyclonic ro-tation of the eddy is also evident in asatellite SST image from 11 July UD 193),which shows a plume of cold upwelledwater wrapping around the southern por-tion of the eddy in a clockwise sense (Fig.iB).

Moored temperature and current timeseries from the continental slope show that,as the eddy moved onshore, both watertemperature and equatorward (along shore)current speed at a depth of 10 m increased(Fig. 2, A and B). This general warming ofsurface waters by about 3.5VC occurreddespite strongly upwelling-favorable winds,which normally lead to cooling of surface

on

July

1, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

V. . .V.#V4.,'4 . . .

waters. On the basis of these temperatureand current time series (4) and a sequenceof current velocity surveys (7), we believethat the eddy was present over the conti-nental shelf for 2 months beginning onabout JD 128. The eddy produced episodesof strong offshore flow near the surface (Fig.2B, shaded regions), which generally oc-curred during periods of increased equator-ward flow that lasted up to a few weeks. Themore consistent offshore flow beginning onabout JD 170 may have resulted from pole-ward movement of the eddy that broughtthe offshore-flowing (southern) part of theclockwise circulation pattern to the moor-ing location. Deeper in the water column at150 m, episodes of equatorward and off-shore flow were also observed; these lastedfor up to a week (Fig. 2C). Peak offshorecurrent speeds exceeded 0.5 m/s at 10 mand 0.07 m/s at 150 m.

The horizontal scales of the eddy and itsanticyclonic rotation are apparent from thelooping trajectory of a surface layer drifter(8) that transited through the eddy from 6to 11 July UD 188 to 193) (Fig. 1B). Weinterpret the dimensions of the drifter track

to be a lower bound on the size of the eddy(at least 90 km in along shore extent and 50km in cross shore extent). The track of thedrifter extended well up onto the continen-tal shelf, reaching almost to the 100-misobath offshore of Point Arena. For 1.5days beginning on 8 July OD 190) when thedrifter was over the shelf, its average speedwas about 0.6 m/s (9), in good agreementwith the moored current observations froma depth of 10 m at that time (verticaldashed lines, Fig. 2B). A section of geo-strophic velocity (10) on 14 July UD 196)shows that the velocity field of the eddy wasdeep, penetrating to at least 400 m in deepwater, and that its center was over thecontinental slope near station 52 (Fig. 3).

At most stations on the continentalshelf and slope within the eddy, layers ofresuspended sediments were observed nearthe sea floor (black squares, Fig. iB). Atstation 204 (data obtained on JD 195) onthe shelf, a deep turbidity layer (11) ex-tended from 150 m down to the bottom at200 m (Fig. 4A). This finding indicatesthat sediment particles had been transport-ed upward from the sea floor (12, 13),

although the vertical transport may haveoccurred at another location and the layeradvected to this location by the eddy. Asecond turbidity layer occurred above 100m, but it exhibited high levels of chloro-phyll fluorescence, which indicates thatthis layer contained phytoplankton. Phyto-plankton are abundant in near-surface wa-ters in this productive upwelling system(14). The sediments near the bottom oc-curred in a relatively well mixed layer. Anabrupt decrease in the vertical density gra-dient (15) marks the top of this layer andindicates a transition from well-stratifiedwaters at mid-depth to the weakly stratifiedwaters in the bottom layer (Fig. 4B). Sim-ilar bottom mixed layers and turbidity layersare common features in shelf waters in thisregion (16, 17) and result from turbulenceand sediment resuspension processes in thebottom boundary layer. Near-bottom cur-rent velocities as a result of the eddy alonewere probably insufficient to produce theobserved turbidity layers. More likely, par-ticles in these layers were initially resus-pended by other processes such as oscillato-ry currents resulting from internal waves or

Fig. 1. Satellite SST images of coastal waters off northern California. (A)Image from 0306 UT on 17 July 1988 (JD 199) [adapted from (7)]. Lightareas indicate cooler water and dark areas, warmer water. The cold areaalong the left side of the image between 370 and 390 is due to clouds.Arrows show current direction and magnitude at a depth of 25 m from ashipboard survey (13 to 18 July; JD 195 to 200). The velocity scale isindicated by the upward arrow labeled 50 cm/s. The center of theanticyclonic eddy in which plumes of resuspended bottom sedimentswere observed is at A. Wind-driven upwelling produced the band ofcolder water adjacent to the coast. (B) False-color satellite SST imagefrom 1 1 July 1988 (JD 193) at 1557 UT showing an anticyclonic eddy nearPoint Arena. The area of the image is shown by the square in (A). A

surface drifter made a single loop around the eddy before it wasreentrained by a strong current jet and advected away toward thesouthwest. The drifter track is shown by the solid line and arrows indicatedirection. Tick marks on the drifter track are 1 day apart and the drifterlocation at the beginning of JD 187 (5 July) is shown. The dashed lineshows the location of part of the continental shelf break at the 200-misobath. Numbers identify specific stations mentioned in the text. Thetriangle near station 204 indicates the mooring position in a water depth of400 m. Squares mark station positions and black squares show stationswhere resuspended bottom sediments were observed. Sediments wereadvected well beyond the continental shelf by the eddy.

SCIENCE * VOL. 261 * 17 SEPTEMBER 1993 1561

on

July

1, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

surface waves (18). We hypothesize that,once the sediment layers were resuspended,the eddy provided a quasi-steady currentthat transported particles off the shelf (19).At some locations such as station 203 (dataobtained on JD 195), turbidity layers wereobserved at depths below 300 m within theeddy, an indication that sediment resuspen-sion and transport processes were not onlyconfined to the shelf but also occurred onthe continental slope.

The eddy carried sediments well be-yond the continental shelf into the deepocean. At station 218 (data obtained onJD 196), about 45 km from the shelfbreak, a turbidity layer 60 m thick wascentered at 225 m (Fig. 4C). The top ofthis layer coincided with an inversion intemperature where the water was 0.P1Cwarmer and saltier by 0.03 practical salin-ity unit (psu) than water immediatelyabove. Because of the increased salinity,the density is greater and the interface isstable. The differences in properties be-tween the sediment layer and ambientwaters at this station indicate that thelayer intruded into the surrounding watercolumn and was derived from a differentwater mass. Comparison of propertieswithin this sediment layer and the bottom

layer of station 204 (Fig. 4A) reveals thatthe temperature, salinity, and density ofthese two layers were virtually identicaland supports the hypothesis that this sed-iment layer originated on the shelf. Lowvertical density gradients within the layerat station 218 (Fig. 4D), particularly inthe upper 25 m, are consistent with strongmixing from contact with the bottom. Thelayer advected from the shelf to station218 in about 9 days (20). Several othersediment layers were observed within theeddy at stations offshore of the shelf break(Fig. 1B).

These observations indicate that plumesof sediments originating on the shelf andslope were advected into deep offshore wa-ters by the eddy. Turbidity layers were ob-served seaward of the continental shelf in allfour hydrographic surveys conducted duringJune and July 1988. This suggests that sedi-ment transport was frequent while the eddywas in the vicinity of the shelf and slope.Mesoscale eddies often occur in the vicinityof Point Arena (21), so eddy transport ofsediments off the continental shelf may be afrequent occurrence here. The observed tur-bidity level of about 0.45 m' in the 50-m-thick layer at station 218 corresponds rough-ly to a suspended mass concentration of 0.2

Fig. 2. Moored time series 15 Aof (A) temperature at a & 14-depth of 10 m, (B) current °...

013--velocity atadepth of 10 m, \ Aand (C) current velocity at 12--adepth of 150 m from JD 11-120toJD 200, 1988. Moor- E 10Aing was in water 400 m ' v .| ,,,. .Ideep at the position indi-cated by the triangle in Fig. 120 130 140 150 160 170 180 190

1 B. Along shore currentspeed is positive toward -Cross shore3170 true (poleward), and B -Along shorecross shore current speed 0.2is positive toward 470 true 00(onshore). All data have to 0 0

been smoothed with a 1.5- "2-0.2day running mean. Vertical /

0dashed lines between JD ° -0.4190 and JD 192 show >when the surface layer -0.6-drifter was over the conti-nental shelf. Warming of 120 130 140 150 160 170 180 190surface water beginningon JD 128 indicates thewarm core of the eddy 0.2 -Cross shorelmoving shoreward and co- 0Along shoreincides with increasing 0equatorward flow at 10 m. E 01The eddy produced 1/strong, persistent offshore o 0.0 vV vflow at 10 m (shaded re- :gions) between JD 170 -0.1

1

VJ

and JD 196. Episodes of . .,ay. ...1..1Julyoffshore flow lasting sever- 120 130 140 150 160 170 180 190al days occurred at 150 m JD (1988)(shaded regions) and mayhave been partly responsible for the transport of sediment plumes off the continental shelf.

mg/liter. Over the shelf at station 204,concentrations in the bottom layer were 0.3to 0.5 mg/liter (22). The decrease in turbid-ity within the plumes around the eddy mayresult from mixing and dilution of the sedi-ments with clearer ambient waters and sug-gests that the eddy produced a net cross-shelftransport of sediments. Sediment layers werenot observed at the few stations off PointArena where the eddy flow was directedback toward shore.

The exact mechanism by which theplumes were entrained into the eddy fromthe continental shelf and slope is not clear,but the cross-shelf transport could occur ina number of ways, including (i) during theepisodes of deep offshore flow while the

Geostrophic velocity (mis)47 48 49 50 51 52 53 54 55

,:.2000.010

0. Ann

40 1Distance (km)

t Fig. 3. Vertical section of geostrophic velocity(500-dbar reference level) across an anticy-clonic eddy for stations 47 to 53 (Fig. 1 B)obtained on 14 July 1988 (JD 196). Southward

20flow is indicated with stippling; the continental

200 shelf and slope are shown with cross-hatch-ing. Strong northward flow between stations49 and 52 and southward flow between sta-tions 52 and 53 resulted from the anticycloniceddy. Southward flow between stations 47and 49 was part of the flow field of a strongcurrent jet lying offshore of the eddy (Fig. 1A).A southward current speed of 0.6 m/s at adepth of 15 m near station 54 was estimatedfrom the path of a surface layer drifter (Fig.1 B) that crossed over the continental shelf 2days before this velocity section was ob-

200 tained. The velocity field of the eddy extendedto a depth of at least 400 m and transportedplumes of sediments from the continental shelfand slope to the deep ocean.

SCIENCE * VOL. 261 * 17 SEPTEMBER 1993

Ew

1562

on

July

1, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

eddy was over the shelf and slope, such asthose observed at the mooring location at adepth of 150 m (Fig. 2C) and (ii) duringthe more persistent episodes when the near-bottom flow over the outer shelfand slope ispoleward and the Ekman transport in thebottom boundary layer is offshore. In thislatter case, the sediment layers may bemoved toward the shelf break until they areentrained in the eddy and swept into thedeep ocean. More observations are requiredto clarify the interaction between eddies inthis region and the deeper flow fields of theshelf and slope.

To obtain an estimate of the mass flux ofsediments across the shelf due to the eddy,we assume that the layer at station 218 ispart of a continuous plume extendingaround the eddy from the shelf break, about80 km. Taking the layer width to be 10 km(23) and the thickness to be 50 m and usingthe advection time scale of 9 days, weestimate that the cross-shelf sediment massflux was of order 106 kg/day. Assuming thatthis eddy remained in the vicinity of theshelf and slope for about 2 months, we

Fig. 4. Vertical profilesat station 204 over the 25.0 25.5continental shelf (ob-tained on JD 195) of: 0.4 0.6 0.8(A) turbidity (c), poten-tial temperature (9), sa- 31.5 32.0linity (S), potential den- 0

sity anomaly (a0), chlo- Erophyll fluorescence - 100 Statia(F), and (B) buoyancy A150frequency (N), a mea- 200sure of the vertical den- 250sity gradient [in cycles 0 2per hour (cph)]. The 0 1bottom is shown bycross-hatching. A layerof resuspended sedi-ments corresponding 25.0 25.to high c and low Foc-curred between 150 mand the bottom. Within 0,5this bottom turbidity 31.5 32.0layer, profiles of 0, S, 0 3.and cr were relativelywell mixed and vertical 100density gradients (N)were reduced as a re- F c

suit of bottom turbu- E 200lence processes. High , ...

levels of c and F above 300100 m indicate phyto- Station 2plankton. (C) As in (A) 400but for station 218 (dataobtained on JD 196) lo- 500cated 45 km seaward 0 2of the continental shelfbreak and along the 0 2track of the surface

estimate that the total mass of sedimentstransported to the deep ocean may be 105metric tons. This amounts to about 4% ofthe total mass of sediments, 2 x 106 to 3 x106 metric tons, discharged onto the shelfannually from the Russian River (19). Al-though this is a small fraction of the totalinput, it probably represents a large fractionof the annual flux of sediments reaching theregion of the shelf around Point Arena.A flux of this order is likely to be a

dominant factor in the cross-shelf transportof sediments during summer when othertransport processes are generally weak; mostsediment flux on this shelf occurs duringwinter storms (13). The sediments on thisshelf contain organic carbon (24). Thus thesuspended particles within the turbidity lay-ers are also likely to contain carbon, and asignificant cross-shelf carbon flux to thedeep ocean may result from eddy transportof sediments here. Mesoscale eddies arecommon features of the California Currentsystem and other eastern boundary cur-rents, and these observations suggest thatthese eddies may play an important role in

a (ktmi3)26.0 26.5 27.0

Beam c(ar1)1.0 1.2 1.4 1.6 1.8 2.0

32.5 33.D 33.5 34.0 34.5A

'n 2040 'e S

4 6 8 10 12O(OC)

2 3 4 5FMV)

a.(kglm3)5 26.0 26.5 27.0

Beam c (m'1)1.0 1.5

6 8O(oC)

N(cph)

2.0

4 6 8 10F(V)

drifter (Fig. 1 B). (D) As in (B) but for station 218. A 60-m-thick turbidity layer (bracketed by the dottedlines) occurred below 200 m and resulted from the advection of sediments from the continental shelfinto the deep ocean. Low values of N and relatively well mixed 0, S, and cr profiles within the layerare consistent with previous mixing at the sea floor.

SCIENCE * VOL. 261 * 17 SEPTEMBER 1993

the transport of shelf sediments to the deepocean.

REFERENCES AND NOTES

1. K. H. Brink et al., Coastal Ocean Processes(CoOP): A Science Prospectus (Publ. WHO1-92-18, Woods Hole Oceanographic Institution,Woods Hole, MA, April 1992).

2. J. J. Walsh, Nature 350, 53 (1991).3. Spatial surveys were conducted as part of the

Coastal Transition Zone (CTZ) Experiment duringJune, July, and August 1988. An overview of theexperiment is given by K. H. Brink and T. J.Cowles [J. Geophys. Res. 96, 14637 (1991)1.

4. Moored current observations were obtained aspart of the Northem California Coastal CirculationStudy in which a large array of current meters wasdeployed over 40 latitude off northern California. Adescription of the moored array and a discussionof the flow induced by the eddy on the continentalshelf are given by J. L. Largier, B. A. Magnel, andC. D. Winant (J. Geophys. Res., in press).

5. For a comprehensive summary of the seasonalstructure and dynamics of this upwelling system,see P. T. Strub, P. M. Kosro, and A. Huyer [J.Geophys. Res. 96, 14743 (1991)] and P. M. Kosroet al. (ibid., p. 14707).

6. We measured the currents shown in Fig. 1A at25-m depth using an acoustic Doppler currentprofiling (ADCP) system mounted on the vesselR.V. Point Sur. This system provides continuoussections of relative current to a depth of about 200m; we obtained absolute currents by combiningthe ADCP relative velocities with shipboard navi-gation. The survey was conducted along sevenparallel sampling lines oriented along shore from13 to 18 July (JD 195 to 200); the inner two linescovering the eddy were sampled on 13 and 14July. The track followed by the vessel is indicatedby the midpoints of the current vectors in Fig. 1A.Five such surveys were conducted from 20 Juneto 4 August 1988.

7. A. Huyer et al., J. Geophys. Res. 96, 14809(1991).

8. A total of 56 Tristar II mixed layer drifters werereleased during the experiment to track the near-surface flow field. Drogues on the drifters wereplaced at 15 m, and the positions were obtainedsix to eight times per day by the ARGOS trackingnetwork. M. S. Swenson, P. P. Niiler, K. H. Brink,M. R. Abbot, J. Geophys. Res. 97, 3593 (1992); K.H. Brink et al., ibid. 96, 14693 (1991).

9. Time series of current speed were formed fromsuccessive observations of drifter position. Drifterlatitude and longitude were interpolated to regularintervals of 0.125 day, and speed was obtainedfrom moving centered differences over 0.25 day.Final velocity estimates were smoothed with a21-hour moving average.

10. Geostrophic velocity was computed with respectto a reference level of 500 dbar along the line ofstations 47 through 53 (Fig. 1B). Data were ob-tained from the R.V. Thomas Washington on 14July (JD 196). At each station in all shipboardsurveys, vertical profiles of temperature, conduc-tivity, pressure, beam transmission, and chloro-phyll fluorescence were measured to a depth of500 m or to the bottom in shallower waters. Atstation 53 where the water depth was 450 m, weestimated dynamic height by first extrapolatingthe specific volume anomaly to 500 dbar and thenintegrating upward.

11. The beam attenuation coefficient or beam c is aquantitative indicator of water turbidity and iscomputed from the transmissometer signal by

c=-e- 'in(100)where T is the percent transmission over the pathlength e = 0.25 m. The transmissometers used inall surveys are manufactured by SeaTech, Inc.,Corvallis, OR, and operate at a wavelength of 660nm.

12. Surface sediments on the middle and outer shelf

1563

ll

----------------------

on

July

1, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

between Point Reyes and Point Arena are com-posed mainly of fine silt derived from the RussianRiver outflow and have mean grain sizes in therange 30 to 125 pm.

13. D. E. Drake and D. A. Cacchione, Cont. Shelf Res.4-5, 495 (1985).

14. B. H. Jones, C. N. K. Mooers, M. Reinecker, T. P.Stanton, L. Washbum, J. Geophys. Res. 96,22207 (1991); R. R. Hood, M. R. Abbot, A. Huyer,ibid., p. 14769.

15. The vertical density gradient is expressed as thebuoyancy frequency N and is a measure of thestatic stability of the water column. It is defined by

N= (-9 PA

where g is the acceleration of gravity, p is seawa-ter density, and z is the vertical coordinate, pos-itive upward. Vertical density gradients are esti-mated from moving centered differences of p andz over vertical scales of about 4 m followed bysmoothing with a running mean over a 12-mvertical scale.

16. S. J. Lentz and J. H. Trowbridge, J. Phys. Ocean-ogr. 21, 1186 (1991); S. R. Ramp, R. W. Garwood,C. 0. Davis, R. L. Snow, J. Geophys. Res. 96,14947 (1991).

17. D. E. Drake and D. A. Cacchione, J. Geophys.Res. 92, 1699 (1987).

18. D. A. Cacchione and D. E. Drake, Geo-Mar. Lett.6, 147 (1986).

19. A recent comprehensive review of sedimenttransport on continental shelves is given by D. A.Cacchione and D. E. Drake, in The Sea (Wiley,New York, 1990), vol. 9, chap. 21.

20. The length L scale over which the layer advectedis estimated from the distance along the driftertrack from the point at which it crossed the conti-nental shelf break to station 218 where the sedi-ment layer was observed, about 80 km. An ad-vective speed U of 0.1 m/s is estimated by as-

suming that the speed within the layer is similar tothe geostrophic current speed in the velocitysection of Fig. 3 at the layer depth. The advectiontime scale L/ is about 9 days.

21. G. S. E. Lagerloef, J. Geophys. Res. 97, 12557(1992).

22. Suspended sediment concentration is derivedfrom transmissometer and bottle sample data forthe Russian River shelf reported in figures 3 and 5and table 2 of (17). An approximately linearrelation is found between the beam attenuationcoefficient c in reciprocal meters and suspendedsediment mass concentration C in milligrams perliter of the form c = 0.475C + 0.37 (SE 0.14 m-1).

23. The scale for the width of the region containingresuspended sediments is estimated from themaximum distance the drifter moves inshore ofthe continental shelf break (200-m isobath) and isabout 10 km. Turbidity layers of this scale havebeen observed on this shelf (17).

24. The carbon content of sediments is indicated bytheir combustible fraction, which is also an indi-cator of phytoplankton abundance. Suspendedsediment samples obtained over the Russian Riv-er shelf have combustible fractions in the range of8 to 53% (17). In the CTZ experiment a fewbottom turbidity layers exhibited weak but mea-surable levels of fluorescence, which suggeststhat they contained phytoplankton that had previ-ously settled to the sea floor.

25. The CTZ Experiment was funded by the CoastalSciences Program of the Office of Naval Re-search. The Northem Califomia Coastal Circula-tion Study was funded by the Minerals Manage-ment Service (U.S. Department of the Interior) andwas conducted in collaboration with EG&G Wash-ington Analytical Services Center, Inc. D. Lawsonof the University of Califomia, Santa Barbara,produced the final satellite images.

7 April 1993; accepted 12 July 1993

Chicxulub Multiring Impact Basin: Size and OtherCharacteristics Derived from Gravity Analysis

Virgil L. Sharpton, Kevin Burke, Antonio Camargo-Zanoguera,Stuart A. Hall, D. Scott Lee, Luis E. Marin,

Gerardo Suarez-Reynoso, Juan Manuel Quezada-Mufleton,Paul D. Spudis, Jaime Urrutia-Fucugauchi

The buried Chicxulub impact structure in Mexico, which is linked to the Cretaceous-Tertiary (K-T) boundary layer, may be significantly larger than previously suspected.Reprocessed gravity data over Northern Yucatan reveal three major rings and parts ofa fourth ring, spaced similarly to those observed at multiring impact basins on otherplanets. The outer ring, probably corresponding to the basin's topographic rim, is almost300 kilometers in diameter, indicating that Chicxulub may be one of the largest impactstructures produced in the inner solar system since the period of early bombardmentended nearly 4 billion years ago.

The Chicxulub structure is widely consid-ered to be an impact crater related to the-65-Ma (million years ago) K-T boundarylayer (1-3). Because this feature is buriedunder 300 to 1100 m of Tertiary carbonaterocks in the Northern Yucatain Platform,geophysical exploration is essential to un-

derstand its morphology and structure. In-deed, concentric patterns in gravity andmagnetic field data over this feature led toits discovery (1). Hildebrand et al. (2) later

1564

recognized two concentric rings in griddedgravity data then available and interpretedtheir 180-km-diameter outer ring as the rimcrest of Chicxulub. They could not resolveconcentric structure in the northern one-

third of the feature, however, and proposedthat an east-northeast trending fault hadremoved the crater's signature in this region(2, 4). To better understand the nature ofthe Chicxulub impact structure and its re-

gional setting, we compiled and analyzed a

SCIENCE * VOL. 261 * 17 SEPTEMBER 1993

new gravity anomaly map (5) of northernYucatin (Figs. 1 and 2).

The gravity data set comprises 3134offshore measurements and 3675 land sta-tions (Fig. 2A) between 19.50N to 22.5%Nand 880W to 90.50W (6). After removingobviously spurious points, we gridded thedata by a bivariate interpolation schemedesigned for irregularly distributed points(7). Gravity anomalies in the mapped re-gion range from -16.4 mgal (10-5 m s-2)to +53.6 mgal.

The Chicxulub basin is a broad, nearlycircular region in which gravity values are20 to 30 mgal lower than regional values. Adistinct 15- to 20-mgal high occupies thegeometric center that we place at 21.30Nand 89.6CW (Figs. 1 and 2). Analysis ofradial profiles compiled for each 100 incre-ment of azimuth clearly reveals multiplerings expressed as local maxima in thegravity anomaly data. Besides the centralgravity high, we recognize three major ringsand evidence of a fragmentary fourth ring(Figs. 1, 2, and 3). Basins with three ormore concentric rings are the largest impactlandforms observed on planetary surfaces.Analyses of multiring impact basins on allthe silicate planets of the inner solar systemhave shown that the radial positions ofthese topographic rings follow a "squareroot of 2" spacing rule (8). The concentricgravity highs within the Chicxulub basin(Figs. 1 and 2) also follow this spacing rule(Table 1), indicating that they correspondto topographic rings of this now-buriedimpact basin.

The central gravity high most likelyreflects the mass concentration associatedwith the dense impact melt sheet and theuplift of silicate basement rocks in themiddle of the structure. The concentricgravity trough separating the central domefrom the inner ring (Fig. 1, ring 1) couldmark the position where the dense meltrock sequence is thin enough for thelow-density breccias filling the crater todominate the gravity expression. The di-ameter of the inner ring is 105 + 10 km.Ring 1 probably corresponds to the topo-graphic central-peak ring associated withthe structural uplift of large complex im-pact craters (9).

Gravity values increase substantially andabruptly between -70 and -100 km fromthe basin center. Near the inside edge of

V. L. Sharpton, D. S. Lee, P. D. Spudis, Lunar andPlanetary Institute, 3600 Bay Area Boulevard, Hous-ton, TX 77058.K. Burke and S. A. Hall, Department of Geosciences,University of Houston, Houston, TX 77004.A. Camargo-Zanoguera and J. M. Quezada-Murieton,Gerencia Divisional de Programacion y Evaluaci6n,Petr6leos Mexicanos, Marina Nacional 329, MexicoCity, Mexico 01131.L. E. Marin, G. SuArez-Reynoso, J. Urrutia-Fucugau-chi, Instituto de Geofisica, Universidad Nacional Au-tonoma de M6xico, Mexico City, Mexico 04510.

:1 p ..Mwi....IM-Mam.~--, 9?~ ,~

on

July

1, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om