JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. C4, PAGES 8555-8571, APRIL 15, 1997 Subtidal velocity correlation scales on the northern California shelf E. P. Dever • Woods Hole OceanographicInstitution,Woods Hole, Massachusetts Abstract. Along- and cross-shelfcorrelationscalesof subtidalcross-shelf(u) and alongshelf (v) velocitiesare estimatedusingmoored recordsfrom severalfield programsover the northern California shelf. Over record lengthsof 4-6 months, along-shelfcorrelation scalesof v are greater than maximummooringseparations(60 km). In the cross-shelf direction, v isgenerally correlated between the60 and130rnisobaths (10-15km separation).Along-shelfcorrelationscalesof u are much smallerthan thoseof v and are often not resolvedby minimum mooring separations.Time seriesbetweenNovember 1988 and May 1989 do resolvealong-shelfcorrelationscalesof near-surfaceu and indicate that they are 15-20 km. During this time the along-shelfcorrelationscaleof near-surfaceu showsvariabilityon a monthlyscale.It is generallylong (30 km or more) when correlation of u with wind stressis highand short(15 km or less)whencorrelationwith wind stressis low. On at least one occasion,short along-shelfcorrelation scalescoincidewith the intrusion of an offshore mesoscale feature onto the shelf. Cross-shelf correlation scales of u are resolvedfor typicalmooringseparations.In general,u is correlatedbetween the 90 and 130 rn isobaths(7-13 km separation)and betweenthe 60 and 90 rn isobaths(-5 km). 1. Introduction There existsa considerablebody of theory concerningthe dynamicsof subtidalvelocity on wind-drivenshelves[Allen, 1980].This includessurfaceand bottom boundarylayer theory [Brink, 1983; Lentz, 1992; Trowbridgeand Lentz, 1991], twodimensionalupwelling models [Csanady,1982; Janowitzand Pietrafesa, 1980; Mitchurn and Clarke, 1986], and coastaltrapped waves (CTW) theory [Chapman, 1987; Brink, 1991]. These theories often assume the circulation is two-dimensional or, at most, slowlyvarying in the along-shelfdirection. Observationalestimatesof along-shelfvelocity(v) correlation lengthsover wind-forcedshelves[KunduandAllen, 1976; Winant et al., 1987] generally agree with this theoretical assumption, and subtidal v can be correlated over along-shelf distancesgreater than maximum mooring separations(60 km and more). The long v correlationscalesobservedwhen wind stressforcing dominates do not preclude shorter correlation scaleswhen wind stressforcing is less important [Winant, 1983]. Understandingcross-shelfflow remainsa major challengein coastal oceanography[Smith, 1995]. However, along-shelf scalesof subtidalcross-shelf velocity(u) are not only shorter than those predicted by wind-forced theory but have usually not been resolved[e.g.,Kundu and Allen, 1976; 145'nant et al., 1987;145'nant, 1983].Successful estimatesshouldthereforeprovide insightinto the dynamicsof u and how observedvelocities differ from simplewind-forcedtheory.They may alsocontribute to a better understandingof the spatial scalesover which point measurementsapplyto u on wind-drivenshelveswhich should be useful to interpretation of existing measurements and to future experiment design. The objectivesof this study are to summarize the spatial scales of both subtidal v and u on the northern California shelf and to gain insight into the processeswhich affect u spatial scalesin particular.The spatial scaleswill be defined here in terms of correlation lengths, the length over which subtidal fluctuationsare correlatedat somesignificancelevel. They will be estimated from time serieslasting several months and encompassingall seasons. The remainder of the paper is divided into five sections.In section2 I give a brief overviewof the general conditionsover the northern California shelf and an introduction to the field programsused.In section3 I list the proceduresusedto estimate u and v correlation scalesand give resultsfor the various experiments. In section 4 the u correlation scalesbetween November 1988 and May 1989 are further examined with a view to identifyingthe processeswhich affect them. In section 5 I compare the observedcorrelation scaleswith those expected from wind-forced theory. In section6 the results are summarized. 2. 2.1. Background and Observations Background The region examinedin this studyextendsfrom Point Reyes to Point Arena (Figures 1 and 2). It is noted for its strong along-shelfwind forcing and relatively straight narrow shelf [Beardsley and Lentz, 1987],and in theserespectsit is perhaps an archetypeof wind-forcedcirculationpresentalong much of the U.S. Pacific coast and elsewhere. Wind stress here exhibits •Now at Centerfor CoastalStudies,ScrippsInstitutionof Ocean- strong seasonalvariability. In winter and spring it is distin- ography,La Jolla, California. guishedby weak monthlymeansand polewardand equator- Copyright1997 by the American GeophysicalUnion. wardfluctuations ontimescales ofdays. In summer i{ isdistin- Paper number 96JC03451. guishedby strongmonthly meanswith periods of upwelling favorablestresswhich persistfor severalweeks [Halliwell and 0148-0227/97/96 JC-03451 $09.00 8555 8556 DEVER: 124 ø VELOCITY SCALES ON THE NORTHERN 123 ø CALIFORNIA SHELF 124 ø 123 ø ß NDBCbuoys ß STRESS Moorings o ß 39 ø NCCCSMoorings SMILEMoorings 39 ø 39 o 39 ø 38 ø 38 ø 38 ø Pt. Arena 38 ø 124 ø 123 ø Figure 1. Map of northernCaliforniashelfshowingnominal Coastal Ocean DynamicsExperiment(CODE) (circles)and National Data Buoy Center (NDBC) (solidsquares)mooring locations.Exact locationsvaried slightlyfrom deploymentto deployment,and most locationsincludednearbysurfaceand subsurfacecomponents.Solid circlesdenote locationsoccupied duringCODE-2, and open circlesdenotelocationsoccupied duringCODE-1. The centralC line (with the exceptionof C1) was present during both CODE-1 and CODE-2. The southerlyR3 locationwas occupiedduringCODE-1 and fall and winter CODE deploymentsbut was later moved north during CODE-2. 124 ø 123 ø Figure 2. Map of northernCaliforniashelfshowingnominal Northern California Coastal Circulation Study (NCCCS) (opencircles),ShelfMixed LayerExperiment(SMILE) (solid circles),SedimentTransportEventsover the Shelfand Slope (STRESS)(solidtriangles),and NDBC (solidsquares)mooring locations.NCCCS mooringlocationsvaried slightlyfrom deploymentto deployment. 3. Velocity Correlation Length Scales 3.1. Procedures Correlationswere estimatedfrom low-passfiltered velocity records(Table 2). The PL64 filter usedhas a 38 hour halfpower point and is describedby Limeburner[1985]. Record Allen, 1987; Strubet al., 1987;Beardsleyet al., 1987]. Figure 3 showsthat the low-frequencyalong-shelfwind stresshas long spatialscalesin all seasons[see alsoBeardsleyet al., 1987]. 4:• + O•o+ +½ + Data used to estimate the correlation scalesof along-shelf wind stress are shown in Table 2.2. Field Programs o o) o 1. 0.5 In part becauseof the characteristics listedabove,the northern California shelf has been the site of several large field programs.Theseincludethe CoastalOceanDynamicsExperiment (CODE) [see,e.g.,Beardsley and Lentz, 1987;Winantet al., 1987; Lentz and Chapman, 1989], Northern California CoastalCirculationStudy(NCCCS) [Largieret al., 1993],Shelf Mixed Layer Experiment(SMILE) [Deverand Lentz, 1994], and Sediment Transport Events over the Shelf and Slope (STRESS) study.Nominal mooringlocationsused here are shownfor CODE in Figure 1 and for SMILE, STRESS, and NCCCS in Figure 2. Table 2 liststhe exactlocationsand start andstoptimesof time seriesusedin thisstudy.Comprehensive informationconcerningthesemooredmeasurementsare given by RosenfeM[1983], Limeburner[1985], EG&G, Inc. [1989, 1990a,b], Alessiet al. [1991],and Fredericks et al. [1993]. o -0.5 0 5'0 1•0 along-shelf separation in km Figure 3. Correlationsof low-passedalong-shelfwind stress, •-Y,asa functionof along-shelfmooringseparationfor summer (CODE-2) NDBC 14,N3, C3, R3, andNDBC 13 (opencircles) and winter and spring(SMILE) NDBC 14, G3, C3, M3, and NDBC 13 (pluses).The along-shelfdirectionis defined as 317øTfor all locationsexceptfor the NDBC 14 locationwhere it is defined as 341øT. Correlation scales of •-y are not sensitive to the precisedefinitionof along-shelfdirectionand exceed 100 km. DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF 8557 Table 1. Data Used to Estimate Along-Shelf Wind StressCorrelation Scales Position Mooring Latitude Longitude NDBC14 C3 R3 NDBC13 39011.98 38036.40 38025.33 38011.98 124000.00 123027.70 123016.36 123018.00 N3 38048.07 ' 123ø41.71 ' G3 38044.40 ' 123036.40 ' C3 M3 NDBC13 38ø38.71 ' 38032.67 ' 38011.98 ' 123029.56 ' 123022.97 ' 123018.00 ' NDBC14 39011.98 ' 124000.00 ' Principle Axis, øT •-7, dyncm-2 •'•, dyncm-2 O.x, dyncm-2 O-y, dyncm-2 rx, hours Ty, hours CODE-2:0400 UT, April 5, 1982, to 2100 UT, July 25, 1982 ' ' ' ' ' ' ' ' 333 313 313 312 -0.1 0.1 0.1 0.2 -0.9 - 1.1 -1.0 -1.0 0.2 0.1 0.1 0.2 1.3 1.3 1.2 1.2 47 58 53 53 44 58 55 57 1.1 46 46 0.8 38 38 0.9 0.8 0.9 38 38 38 38 38 43 0.9 38 38 CODE-2:0400 UT, April 11, 1982, to 2100 UT, July 25, 1982 335 -0.3 -1.0 0.1 SMILE: 0700 UT, Nov. 16, 1988, to 1100 UT, April 6, 1989 342 -0.3 -0.2 0.3 SMILE.' 0700 UT, Nov. 16, 1988, to 1900 UT, May 13, 1989 311 312 303 0.0 -0.0 0.1 -0.4 -0.3 -0.4 0.1 0.1 0.3 SMILE: 1200 UT, Jan. 15, 1989, to 0100 UT, May 4, 1989 342 -0.0 -0.0 0.1 Start and stoptimesgivenare thoseusedfor correlationcalculations.The along-shelfdirectionis taken as 317øTexceptat NDBC 14 where it is taken as 34 løT. start and stop times were chosensuchthat completerecords existedat most mooringsin any given field program deployment. To compareNCCCS and SMILE, the NCCCS records were dividedinto three periods,the secondof whichcoincided with SMILE. Record lengthsrangedfrom 90 to 200 days(Table 2). The (one-sided)autocorrelationtimescalesprovide an estimateof the number of degreesof freedom for crosscorrelations between two time series. Autocorrelation timescales were generally3-4 daysfor v and 2-3 daysfor u. Longer time seriesalso tend to have longer autocorrelationtimes as they start to pick up seasonalvariability.The autocorrelationtimescalesgenerally suggestat least 30 degreesof freedom. This estimate is conservative in that cross-correlation timescales are these recordswere in the surfaceboundarylayer, though several are probablybelow it duringweak wind stressconditions, especiallyin summer[Lentz, 1992]. Middepth recordswere as near as possibleto half the water depth at each mooring. Near-bottom velocityrecordswere 10 to 20 m abovethe bottom. Though 20 m is near the upper limit of bottom boundary layer [Lentzand Trowbridge,1991], it was felt the larger number of records gained was preferable to the limited records within 3.2. 10 m of the bottom. Along-Shelf Correlation Scales Correlations as a function of along-shelf separation were estimated primarily along the 90 m isobath.Additional estimates along the 60 and 130 m isobathswere made using the CODE-2 data. Minimum along-shelfseparationsranged from 4 km during the common SMILE/NCCCS time period to 26 alwayslessthan autocorrelationtimescales.For 30 degreesof freedom,correlationsof 0.30 (0.35) or greaterare significantat the 90% (95%) level. Velocity recordswere rotated into local isobath reference km in CODE-2. Along-shelfvelocities(Figure 4 and Table 3) were signififramesat eachmooringlocation(Table 2). The CODE-2 local isobathdirections[Winant et al., 1987] were also used for the cantly correlated at all along-shelfmooring separationsfor analogousCODE-i, NCCCS, SMILE, and STRESS locations. near-surface, middepth, and near-bottom depths; hence v The CODE-1 C1 isobathorientationis givenby Lentz [1994], along-shelfcorrelation scalesare greater than 60 km in this while the CODE-1 R3 and M3 and the SMILE M3 and G3 area. Correlationsappearto showlittle variationbetweenfield isobathorientationswere estimatedfrom charts.As noted by programswhich occurred during different seasons.Though Smith [1981],u is sensitiveto the referenceframe. To test this, correlated at all depths, correlations of interior and nearcorrelations were calculated in which all records were rotated bottom v were slightlylarger than near-surfacev (Table 3 and ñ2 ø and ñ5 ø from the local isobaths. The v correlation coefFigure 4). Along-shelfcorrelationsof v are not a strongfuncficients varied by only ñ0.02. Correlation coefficientsfor u tion of total water depth thoughCODE-2 observationssuggest showedmore scatter(about ñ0.10). Sensitivityof u correla- some decrease in near-surface v correlation at the 130 m isotions to coordinate rotation was greatestfor near-bottom in- bath (N4, C4, and R4) relativeto the 60 (N2, C2, and R2) and struments and those located at the 30 or 60 m isobath. How90 m (N3, C3, and R3) isobaths(Table 3). ever, qualitativeresultswere unaffectedin that u correlations Along-shelfcorrelationscalesof u (Figure 5 and Table 3) which were significantin the local isobath reference frame were resolved only for near-surfaceu during SMILE and remained significantin the rotated referenceframes. NCCCS (and possiblyby CODE-1 C3 and M3) alongthe 90 m Correlations were calculated for near-surface, middepth, isobath. SMILE and NCCCS correlations indicate that alongand near-bottomrecords.These depthswere chosenbasedon shelf correlation scales of near-surface u are from 15 to 20 km. the divisionof flow into a surfaceboundary layer, an inviscid Little information is available about subsurfaceu along-shelf interior (exceptat the 30 m C1 site), and a bottom boundary correlationscales.There is the suggestionthat middepth and layer. Near-surfacerecordswere at 10 m, or failing that, at the near-bottom u are correlated at separationsof 4-10 km but shallowestavailable depth (always20 m or less). Generally, that correlation lengths are lessthan 25 km. 8558 DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF Table 2. VelocityData Used to EstimateCorrelationScales Position Instrument Water Mooring Latitude Longitude Depth, m Depth, m Isobath, øT •, 10-2 m s-• •, 10-2 m s-• o-., 10-2 m s-• o-•,, 10-2 m s-• T., T,,, Principle hours hours Axis, øT CODE-I: 0400 UT, April 17, 1981, to 1500 UT, July10, 1981 ' ' ' ' ' ' ' ' ' 38ø36.11 38ø36.19 38ø36.19 38ø31.30 38021.80 38021.70 38021.70 ' ' ' ' ' ' ' 123ø27.18 123ø27.18 123ø27.18 123ø40.10 123013.00 123ø13.10 123ø13.10 ' ' ' ' ' ' ' 38ø36.11 38ø36.10 38ø36.10 38031.30 38ø31.30 38021.50 38ø21.95 38ø21.95 ' ' ' ' ' ' ' ' 123ø27.18 123027.40 123027.40 123040.60 123ø40.10 123013.00 123ø13.15 123ø13.15 ' ' ' ' ' ' ' ' 9 55 75 9 150 9 55 75 123ø25.32 123025.32 123025.32 123ø27.71 123032.70 123032.70 ' ' ' ' ' ' R4 38ø38.16 ' 38ø38.16 38ø38.16 38ø38.38 38ø34.30 38ø34.30 38ø33.26 38ø33.26 38ø33.26 38ø30.80 38ø30.88 38ø49.50 38ø49.50 38ø48.07 38ø48.09 38ø48.09 38ø45.79 38ø45.71 38ø45.71 38ø27.17 38ø27.14 38ø27.14 38ø25.38 38ø25.33 38ø25.38 38ø20.36 38ø20.84 38ø20.84 ' ' ' ' ' ' ' ' ' ' ' 10 35 53 10 53 83 10 70 121 20 150 10 35 10 53 83 10 70 121 20 35 53 20 53 70 10 70 110 C2 38038.33 ' 123024.60 ' 10 C2 C3 C3 C3 C4 C4 C4 C5 C5 M3 M3 M3 R3 R3 R3 ' 30 30 63 90 90 90 133 133 133 402 402 90 90 90 90 90 90 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' C1 38039.80 38ø39.80 38ø39.20 38ø36.38 38ø37.20 38ø37.20 38ø34.50 38ø34.50 38034.50 38031.27 38031.27 38031.60 38031.60 38031.60 38ø21.60 38ø21.65 38ø21.65 4 123ø25.10 123ø25.10 123025.60 123ø27.71 123028.30 123028.30 123032.60 123032.60 123032.60 123ø40.41 123ø40.41 123023.20 123023.30 123023.30 123ø13.00 123ø13.00 123013.00 C1 27 4 9 39 83 19 65 123 9 152 9 55 74 9 55 75 335 335 325 317 317 317 319 319 319 330 330 317 317 317 336 336 336 -1.6 0.4 -4.4 -9.8 -2.7 -0.5 -6.3 -1.2 -1.1 -6.5 1.9 -4.0 2.7 1.3 -3.8 2.6 2.2 -3.4 1.0 -2.4 -12.7 -2.3 0.4 -22.8 -9.1 -1.6 -23.5 -0.6 -1.4 2.8 4.3 -5.6 1.3 2.5 1.2 1.0 3.8 4.5 4.5 2.7 9.1 5.7 4.5 14.4 5.0 4.8 4.1 3.6 4.2 3.2 2.7 6.1 4.1 12.1 24.0 19.5 13.7 21.2 14.4 10.3 21.1 14.5 22.9 18.5 17.2 15.2 12.1 10.9 38 38 38 38 40 42 45 46 39 54 53 44 50 43 38 43 51 38 40 54 46 51 45 58 56 46 52 91 61 56 54 66 58 51 344 344 340 330 342 315 330 322 312 16 340 331 330 317 347 324 316 18.9 13.1 11.1 9.6 12.8 9.0 8.2 76 75 38 52 38 38 38 121 87 76 88 80 78 66 337 333 319 337 356 332 327 21.3 11.6 8.8 15.2 12.2 14.3 9.9 9.1 42 38 38 77 63 38 38 38 58 51 39 104 105 51 40 38 332 324 320 12 337 345 328 323 19.3 13.3 8.2 24.9 15.7 11.6 21.9 12.3 10.7 19.6 11.7 25.9 15.7 28.9 17.1 13.9 24.4 15.0 14.0 13.7 10.6 6.3 17.8 12.4 11.9 18.4 9.9 7.9 38 90 44 53 65 38 38 38 38 54 56 46 43 73 43 55 81 92 49 42 38 38 45 48 50 40 38 45 62 61 60 76 74 65 98 69 49 160 133 61 60 60 60 53 110 80 55 49 50 50 61 59 57 85 70 59 323 325 321 327 324 314 332 320 314 337 332 316 308 325 315 302 338 322 305 316 310 310 333 323 317 347 333 322 101 103 317 CODEW2: 0400 UT, Aug. 7, 1981, to 2100 UT, Dec. 7, 1981 C3 C3 C3 C5 R3 R3 R3 90 90 90 400 90 90 90 9 55 75 150 9 55 75 317 317 317 330 336 336 336 -1.9 4.6 0.3 0.5 -1.4 1.3 -0.7 7.6 9.4 8.1 4.4 6.4 7.2 5.1 4.2 2.1 1.5 4.1 5.2 2.1 1.7 CODEW3: 0400 UT, Dec. 16, 1981, to 2100 UT, March 19, 1982 C3 C3 C3 C5 C5 R3 R3 R3 90 90 90 400 400 90 90 90 317 317 317 330 330 336 336 336 -1.6 2.0 0.6 -0.8 1.7 -0.3 1.3 0.4 3.5 5.6 5.4 -7.3 3.6 0.2 3.4 3.5 6.3 2.9 2.4 8.8 5.0 6.5 2.4 2.3 CODE-2:0400 UT, April 5, 1982,to 2100 UT, July25, 1982 C2 C2 C2 C3 C3 C3 C4 C4 C4 C5 C5 N2 N2 N3 N3 N3 N4 N4 N4 R2 R2 R2 R3 R3 R3 R4 R4 123ø31.68 ' 123ø31.56 ' 123ø31.56 ' 123ø40.25 123ø40.41 123ø40.11 123ø40.11 123ø41.71 123ø41.77 123ø41.77 123ø45.60 123045.55 123045.55 123ø13.97 123ø13.94 123ø13.94 123ø16.40 123ø16.36 123ø16.40 123022.94 123022.95 123022.95 ' 60 60 60 93 90 90 130 130 130 400 400 60 60 90 90 90 129 130 130 60 60 60 90 90 90 130 130 130 325 325 325 317 317 317 319 319 319 330 330 316 316 319 319 319 319 319 319 319 319 319 329 329 329 339 339 339 -1.8 0.4 -0.4 -3.8 0.7 0.0 -7.3 0.7 0.1 0.7 1.9 -2.1 1.1 -3.8 1.9 -0.8 -5.6 1.6 1.0 1.4 1.4 0.2 0.2 0.6 0.5 -0.7 1.8 -0.1 3.6 3.5 1.9 -6.4 0.5 0.9 -20.7 -4.7 1.4 -16.9 4.8 -2.7 1.7 -8.3 2.1 3.5 -17.3 -2.3 4.3 6.5 5.8 3.1 -0.0 1.2 1.6 - 18.3 -4.9 -0.1 2.5 2.7 1.1 6.6 2.7 1.8 9.2 4.6 3.4 9.5 3.8 4.0 1.7 6.0 2.6 3.3 9.0 4.6 3.3 3.0 1.9 1.7 4.1 2.3 1.9 6.8 5.2 4.0 NCCCS: 0100 UT, March 23, 1988, to 0000 UT, Aug. 16, 1988 60 325 -0.0 -1.3 2.1 17.5 DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF 8559 Table 2. (continued) Position Instrument Mooring Latitude Longitude Depth, m Water Depth, m Isobath, øT •, 10-2 m s-• •, 10-2 m s-• rru, 10-2 m s-• rr•,, Tu, T•,, Principle 10-2 m s-• hours hours Axis, øT NCCCS: 0100 UT, March 23, 1988, to 0600 UT, Nov. 16, 1988 C3 C3 C3 C4 C5 38ø36.77 ' 38ø36.77 ' 38ø36.77 ' 38ø33.60 ' 38ø30.00 ' 123ø26.93 ' 123ø27.48 ' 123ø26.93 ' 123ø30.78 ' 123ø38.55 ' 10 45 75 10 150 90 90 90 130 400 C5 38ø30.35 ' 123ø39.73 ' 10 C5 38ø30.00 ' 123ø38.55 ' 150 C3 C3 C3 C4 38ø36.77 ' 38ø36.77 ' 38ø36.77 ' 38ø33.60 ' 123ø26.93 123ø27.48 123ø26.93 123ø30.78 ' ' ' ' 10 45 75 10 C2 38ø38.33 ' 123ø24.60 ' 10 C2 C3 C3 C3 C4 38ø38.33 ' 38ø36.77 ' 38ø36.77 ' 38ø36.77 ' 38ø33.60 ' 123ø24.60 ' 123ø26.93 ' 123ø27.48 ' 123ø26.93 ' 123ø30.78 ' 10 10 45 75 10 C3 38ø38.71 ' 123029.56 ' 11.5 C3 C4 C4 G3 M3 38ø38.71 ' 38036.78 ' 38036.78 ' 38044.40 ' 38032.67 ' 123029.56 123031.87 123031.87 123ø36.40 123ø22.97 44.5 10 52 10 10 C2 38039.80 ' 123027.82 ' 10 C3 C3 38038.44 ' 38038.44 ' 123029.64 ' 123029.64 ' 79 91 C3 38ø38.14 ' 123029.97 ' 77 C3' 38036.89 ' 123ø27.18 ' 86 C2 C3 C3 C4 C4 38039.50 ' 38ø38.14 ' 38ø38.14 ' 38ø35.64 ' 38035.64 ' 123ø25.64 ' 123ø28.31 ' 123ø28.31 ' 123032.52 ' 123032.52 ' 39 59 80 59 120 317 317 317 319 330 -3.0 2.1 1.3 -8.0 0.8 -6.1 2.7 3.4 -17.9 7.0 5.5 2.7 3.1 7.1 3.3 23.4 14.4 7.2 21.5 6.6 65 156 55 215 92 434 194 76 268 102 321 324 300 348 330 22.0 105 101 360 8.8 47 92 318 17.4 11.4 8.1 17.5 47 85 55 110 108 92 56 162 325 319 310 329 17.1 47 62 339 13.2 27.0 19.3 11.5 21.6 68 66 38 54 49 75 103 60 72 213 339 323 320 307 327 14.4 38 104 325 10.9 17.3 13.3 16.1 17.4 52 95 109 95 38 92 149 134 118 115 320 337 327 322 326 18.8 38 59 326 7.6 6.8 38 38 43 38 313 309 9.9 38 50 314 6.7 38 38 313 7.1 12.1 8.9 15.8 9.6 38 38 38 46 38 47 64 51 105 49 323 317 310 328 312 NCCCS: 0900 UT, May 4, 1988, to 0700 UT, Aug. 17, 1988 400 330 -1.3 -19.6 13.0 NCCCS: 0700 UT, Nov. 16, 1988, to 0100 UT, Feb. 12, 1989 400 330 -1.2 6.9 3.3 NCCCS: 0700 UT, Nov. 16, 1988, to 1900 UT, May 13, 1989 90 90 90 130 317 317 317 319 -1.5 2.8 0.2 -3.2 -3.9 3.7 3.7 -13.5 5.2 2.7 2.3 7.3 NCCCS: 0200 UT, Feb. 21, 1989, to 1900 UT, May 13, 1989 60 325 -1.9 -2.9 2.1 NCCCS*: 2000 UT, May 13, 1989, to 2300 UT, Oct. 17, 1989 60 90 90 90 130 325 317 317 317 319 1.2 -4.3 0.6 1.0 -5.6 4.6 -6.8 -5.2 4.1 - 19.1 2.0 4.4 3.1 2.8 6.4 SMILE: 0700 UT, Nov. 16, 1988, to 1400 UT, April 21, 1989 93 317 -2.2 -3.3 4.3 SMILE: 0700 UT, Nov. 16, 1988, to 1900 UT, May 13, 1989 ' ' ' ' ' 93 117 117 93 93 317 319 319 317 317 1.7 -5.5 0.4 -3.6 -4.2 1.8 -12.0 -4.1 -5.9 -4.0 2.3 7.8 3.8 5.7 4.6 SMILE: 0400 UT, Feb. 28, 1989, to 1900 UT, May 13, 1989 80 325 -2.6 -8.5 4.9 STRESS-I: 1800 UT, Dec. 8, 1988, to 0700 UT, Feb. 25, 1989 97 97 317 317 -0.1 -1.3 5.7 3.5 1.8 1.9 STRESS-l: 1200 UT, March 6, 1989, to 0100 UT, May 5, 1989 95 317 -0.3 1.5 1.9 STRESS-I: 1000 UT, Dec. 9, 1988, to 0300 UT, Feb. 24, 1989 92 317 -0.6 4.2 1.9 STRESS-2:1900 UT, Nov. 23, 1990, to 0600 UT, March 7, 1991 49 90 90 130 130 325 317 317 319 319 0.7 1.9 -1.0 1.3 -0.3 4.8 9.8 5.9 3.1 3.8 1.4 2.7 1.9 4.0 2.4 Start and stop times given are thoseusedfor correlationcalculations. *From overlappingmultipledeployments, positiongivenis recordedpositionin ScrippsData Zoo files(deployment1). 3.3. Cross-Shelf Correlation Scales large-scalewind forcing,interpretationof along-shelfcorrelaCorrelations as a function of cross-shelfseparationwere tion scalesis fairly straightforwardin that we expect alongestimatedprimarilyalongthe C line. Mooringsalongthe C line shelfcorrelationsto be primarilya functionof separationand rangedfrom the inner shelfto the outer shelf,and separations not along-shelfposition. In contrast,there is no reason to were from i to 12 km becomingprogressively larger offshore. expect cross-shelfcorrelation scalesto be independentof Additional estimates of cross-shelf correlations were made cross-shelfposition. Coastal oceanographersoften divide shelvesinto three regions[see, e.g., Lentz, 1995;Allen et al., alongthe CODE-2 N and R lines. Cross-shelfscalesin general are more difficult to interpret 1983]:an inner shelfwhere surfaceand bottom boundarylaythan along-shelfcorrelationscales.For a straight shelf with ers interact, a midshelf where surface and bottom boundary 8560 DEVER: VELOCITY (a) '- ON THE near-surface v 1 •) SCALES (b) + •o o0.5 CALIFORNIA 8' SHELF middepthv 1 -•++ • NORTHERN 0 ß • (D + o x +o e o 0.5 o • 0 ............... 0 o o • -0'50 2'0 40 60 along-shelfseparationin km (c) •q) 1 o o 0.5 o • 6O o ß o 4O near-bottom v --k •- o 2O along-shelfseparation in km e ß x o + CODE-1 CODEW2 CODEW3 CODE-2 SMILE/NCCCS-2 O ............... o o -0.5 20 40 60 along-shelfseparationin km Figure 4. Correlationsof along-shelfvelocity v as a function of along-shelfmooring separationfor (a) near-surface,(b) middepth,and (c) near-bottominstruments.Lagsat maximumcorrelationare usuallyvery closeto thoseat 0 lag, and for consistency, 0 laggedcorrelationsare plotted throughout.Correlationsare also presentedin Table 3. Along-shelfcorrelationscalesof v exceedthe maximummooringseparation(--•60km). layersare thin comparedto the total water depthbut wherethe shelf is generally isolated from off-shelf variability, and an outer shelfwhere open oceanvariability is important.Because the processeswhich govern the velocity field on the shelf are themselvesoften a strong function of cross-shelfposition, cross-shelfcorrelationsestimatedhere will be discussed in light of cross-shelfposition as well as separation. Cross-shelf correlationscalesof v (Figure6 andTable 4) are 10-15 km. Despite showingmore scatter than v along-shelf correlations,they are well resolvedin the sensethat observations are significantlycorrelated to their nearest neighbors. Scatterbetweenexperimentsagainshowsno seasonalvariability. Most observationsare over the 60, 90, and 130 m isobaths. Within this region, v correlationsare not a strongfunctionof position; that is, correlationsbetween v at 60 and 90 m (a distanceof about 5 km) appear to be about the sameas correlationsbetweenv at 90 and 130m (a distanceof about8 km). Similarly, cross-shelfcorrelationsof v exhibit little systematic variabilitywith instrumentdepth. Like the along-shelfu correlations,the cross-shelfu correlations where were best resolved in the near-surface the cross-shelf correlation observations scale is about 10 km. In con- trast to along-shelfcorrelationsof u, cross-shelfcorrelationsof u (Figure 7 and Table 4) were resolvedby typical mooring separations.Comparisonof u correlationsbetweenthe 60 and 90 isobathswith thosebetween90 and 130 m (mostlyfrom summerobservations)givessomesuggestion that near-surface u is more highlycorrelatedbetweenthe outer-shelfpair despite the greater mooringseparation.Though the CODE-2 observationsgivesomeindicationthat u correlationsare higherfor nearsurface and near-bottom instruments than for interior instru- ments,the large scatterand limited numberof time seriesmake definitivestatementsaboutdepth dependenceimpossible. 4. Interpreting u Along-Shelf Correlation Length Scales During the Winter and Spring 1988-1989 Both along-shelfand cross-shelfcorrelation scalesof v and cross-shelf correlationscalesof u were resolvedadequatelyby most field programsexamined in section3. However, alongshelf u correlationswere not resolvedby typical along-shelf mooringseparationsof 30 km, and short mooring separations (5-15 km) are required.This is muchshorterthan the scaleof the along-shelfwind stress(see Figure 3) which is generally acknowledgedto be a dominant driving force of near-surface cross-shelf circulation[e.g.,Dever, 1995;Lentz, 1992] and warrants a closer examination. To gain insightinto the processeswhich reduce along-shelf correlation scales of u, observations from the combined SMILE, STRESS,and NCCCS moorings(Figure2) are examined here in further detail. These mooringscome closestto meetingthe requirementsfor resolvingalong-shelfscalesof u, though spatial coverageis only extensivenear the surface. They provide correlation scale estimatesfor near-surfaceu from 4 to 30 km along the 90 m isobathbetween November 1988 and May 1989. Time seriesof low-passfiltered along-shelfwind stress(r y) and u (Figure 8) showvisuallythat numerouswind forcing events(e.g., December 6-7, December 22-24, March 1, and April 10) are reflected to some extent at all mooring sites. However, there is additionalu variabilitywhich is not evident DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF 8561 Table 3. Correlations of u and v asa Functionof Along-ShelfSeparation Max Lag Distance, Moorings Hours km Max Lag 0 Lag u Correlations , 0 Lag u Correlations Hours v Correlations v Correlations Hours Near-SurfaceCorrelations CODE-1 C3(009):M3(009) M3(009):R3(009) C3(009):R3(009) 2028 2028 2028 11.61 25.00 36.45 0.47 0.34 0.22 0.52 0.34 0.23 - 11 -2 -5 0.67 0.82 0.56 0.73 0.85 0.62 -13 -10 -14 2946 33.56 0.52 0.52 -4 0.78 0.83 -14 2250 34.00 0.28 0.28 -2 0.63 0.64 -6 2682 2682 2682 2682 2682 2682 2682 2682 2682 26.18 29.98 56.07 26.29 29.64 55.89 26.42 30.76 56.83 0.03 0.24 0.06 0.03 0.07 -0.01 -0.03 0.08 -0.03 -0.17 0.26 0.13 -0.08 -0.15 0.19 0.26 0.21 -0.20 -89 + 100 + 100 -43 - 100 +49 + 56 +55 +49 0.87 0.78 0.75 0.87 0.72 0.71 0.68 0.50 0.55 0.87 0.78 0.75 0.87 0.74 0.72 0.68 0.50 0.57 +1 -5 +0 -3 -9 -5 - 14 +0 -16 3752 4285 3752 3752 4285 4285 4285 4.67 10.04 14.46 14.71 19.12 29.16 5.02 0.73 0.54 0.35 0.48 0.39 0.11 0.89 0.73 0.54 0.37 0.48 0.39 -0.27 0.89 +0 +2 -4 +0 -2 -53 - 1 0.93 0.90 0.86 0.77 0.70 0.55 0.97 0.93 0.90 0.86 0.77 0.71 0.55 0.97 +0 +0 -2 -2 -4 -4 +0 -9 -4 -9 CODEW2 C3(009):R3(009) CODEW3 C3(009):R3(009) CODE-2 C2(010):R2(020) N2(010):C2(010) N2(010):R2(020) C3(010):R3(020) N3(010):C3(010) N3(010):R3(020) C4(010):R4(010) N4(010):C4(010) N4(010):R4(010) SMILE C3(010):NC(010) NC(010):M3(010) G3(010):C3(010) C3(010):M3(010) G3(010):NC(010) G3(010):M3(010) C4(010):NC(010) Middepth Correlations CODE-1 C3(039):M3(055) M3(055):R3(055) C3(039):R3(055) 2028 2028 2028 10.42 25.15 35.36 0.66 -0.12 -0.23 0.67 --0.28 -0.30 -2 -43 - 17 0.81 0.85 0.64 0.85 0.86 0.67 2946 33.74 -0.06 -0.20 -22 0.92 0.93 2250 33.12 -0.06 -0.22 - 19 0.67 0.68 2682 2682 2682 2682 2682 2682 2682 2682 2682 26.25 29.98 56.14 26.29 29.72 55.97 26.28 30.60 56.54 0.22 0.28 0.64 -0.12 -0.35 0.28 0.01 -0.08 0.03 0.24 0.29 0.64 -0.12 -0.37 0.34 0.30 -0.28 0.23 + 15 -9 - 1 +7 + 14 - 16 + 52 + 88 -57 0.79 0.67 0.72 0.92 0.79 0.79 0.82 0.77 0.73 0.80 0.92 O.80 +0 0.80 0.82 0.77 0.73 -3 + 1 -3 -3 4285 4.67 0.56 0.56 +0 0.96 0.96 +0 -14 + 17 + 10 0.84 0.85 0.65 0.86 0.86 0.68 -6 -5 -8 -40 0.92 0.93 -5 CODEW2 C3(055):R3(055) CODEW3 C3(055):R3(055) -3 CODE-2 C2(035):R2(035) N2(035):C2(035) N2(035):R2(035) C3(053):R3(053) N3(053):C3(053) N3(053):R3(053) C4(070):R4(070) N4(070):C4(070) N4(070):R4(070) 0.69 0.72 +4 -10 - 1 SMILE C3(045):NC(045) Near-Bottom Correlations CODE-1 C3(083):M3(074) M3(074):R3(075) C3(083):R3(075) 2028 2028 2028 10.42 25.15 35.36 0.35 0.25 0.44 0.45 0.36 0.49 2946 33.74 0.16 2250 33.12 0.10 0.26 + 15 0.78 0.78 +0 2682 2682 2682 2682 2682 2682 2682 26.25 26.29 29.72 55.97 26.28 30.60 56.54 0.41 0.44 0.31 0.73 0.03 0.34 0.26 0.41 0.44 0.32 0.75 0.16 0.38 0.39 -2 - 1 -6 -8 +56 -14 -24 0.80 0.95 0.86 0.84 0.85 0.88 0.75 0.80 0.95 0.86 0.84 0.85 0.88 0.75 +2 +1 -3 -1 -2 +0 -3 3268 1842 4.38 4.58 0.51 0.66 0.51 0.66 +2 -2 0.94 0.96 0.94 0.96 +1 - 1 CODEW2 C3(075):R3(075) -0.31 CODEW3 C3(075):R3(075) CODE-2 C2(053):R2(053) C3(083):R3(070) N3(083):C3(083) N3(083):R3(070) C4(121):R4(110) N4(121):C4(121) N4(121):R4(110) SMILE C3(079):NC(075) C3(091):C3'(086) Positivelagsdenotethe first listedseriesleadingthe second. 8562 DEVER: VELOCITY SCALESON THE NORTHERN CALIFORNIA SHELF (a) c o o c (D+ 0.5 middepth u 1 øo0.5 ' + + (9 o ß- (b) near-surface u 1 , + o O X(9 o e 0 0 o O- (•DOX o o . o 20 40 20 60 (c) c near-bottom o u 0.5 o ß- 60 1 ...-, o 40 along-shelfseparationin km along-shelfseparationin km e CODE-1 ß x CODEW2 CODEW3 o + CODE-2 SMILE/NCCCS-2 0- o • -0.5• 20 40 60 along-shelfseparationin km Figure5. Correlations of cross-shelf velocity u asa function of along-shelf mooring separation for (a) near-surface, (b) middepth, and(c) near-bottom instruments. Correlations arealsopresented in Table3. Along-shelf correlation scales ofnear-surface u are15-20km.Along-shelf scales ofsubsurface u arenotwell resolved but are less than 25 km. (a) c (b) near-surface v middepth v 1 1 o• .m o o o 0.5 0.5 o o o o+ © © ,- 0 0 o 10 2o 3o cross-shelf separation in km (c) c near-bottom 10 2o 3o cross-shelfseparation in km v 1 '• ß o o 0.5 © ß x o ß + [] ß o o e o © '.,• • -0-50 o o o '0.50 10 20 CODE-1 CODEW2 CODEW3 CODE-2 NCCCS-1 SMILE/NCCCS-2 NCCCS-3 STRESS-2 30 cross-shelf separation in km Figure6. Correlations of along-shelf velocity v asa function of cross-shelf mooring separation for (a) near-surface, (b) middepth, and(c) near-bottom instruments. Correlations arealsopresented in Table4. Cross-shelf correlation scales of v are 10-15 km. DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF 8563 Table 4. Correlations of u and v asa Functionof Cross-Shelf Separation Maximum Lag Moorings Hours Distance, km 0 Lag u Correlations u Correlations Maximum Lag 0 Lag Hours v Correlations +1 + 10 + 12 +8 -76 -29 +35 -38 -74 +3 0.82 0.70 0.50 0.61 -0.02 0.22 0.13 -0.06 0.02 0.07 0.86 0.79 0.71 0.68 0.36 0.39 0.41 0.15 -0.14 0.14 +8 +13 +22 +20 + 100 +44 +53 +94 +43 -81 -3 0.13 0.22 + 89 +2 +2 -2 +2 +2 + 100 + 12 +86 -78 +95 +83 +3 0.88 0.86 0.87 0.45 0.53 0.57 0.21 0.36 0.46 0.13 0.01 0.47 0.89 0.87 0.87 0.54 0.58 0.57 0.32 0.42 0.53 0.18 -0.21 0.47 v Correlations Hours Near-Surface Correlations CODE-1 C1(004):C2(004) C2(004):C3(009) C1(004):C3(009) C3(009):C4(019) C4(019):C5(009) C2(004):C4(019) C1(004):C4(019) C3(009):C5(009) C2(004):C5(009) C1(004):C5(009) 2028 2028 2028 2028 2028 2028 2028 2028 2028 2028 1.18 5.52 6.68 7.99 11.46 13.49 14.64 18.53 23.56 24.60 0.73 0.42 0.36 0.38 0.15 -0.08 0.09 -0.03 0.13 0.33 0.73 0.50 0.46 0.39 -0.42 -0.21 0.17 -0.21 0.17 0.33 2250 21.39 0.49 0.49 2682 2682 2682 2682 2682 2682 2682 2682 2682 2682 2682 2682 3.52 4.74 4.86 7.03 8.06 12.77 10.50 12.80 13.37 20.94 25.57 17.63 0.57 0.30 0.51 0.75 0.66 0.02 0.31 0.04 0.11 0.00 0.07 -0.20 0.57 0.30 0.51 0.75 0.66 0.17 0.35 0.14 -0.23 0.18 0.23 -0.20 3504 5718 3504 2519 2519 2488 4.52 8.06 12.57 13.21 21.11 25.63 0.38 0.55 0.05 0.44 0.48 0.35 0.39 0.55 0.11 0.45 0.49 0.35 -4 +0 -53 +6 +5 -3 0.76 0.75 0.40 0.47 0.34 0.14 1962 4285 1962 4.36 7.37 11.73 0.44 0.69 0.00 0.48 0.69 -0.30 -6 -1 +34 0.83 0.58 0.20 0.83 +2 0.58 0.21 + 3 +9 3772 3772 3772 4.57 7.47 12.03 0.47 0.64 0.24 0.48 0.65 0.29 -5 - 1 - 16 0.75 0.80 0.54 0.76 +8 0.81 + 12 0.57 + 27 1259 3752 1792 3.23 4.90 8.11 0.81 0.85 0.75 0.81 0.85 0.75 +0 +1 +0 0.90 0.82 0.39 0.90 0.82 0.39 +0 +3 +6 -72 -56 +31 0.74 0.32 0.05 0.76 0.45 0.27 +8 +86 - 100 CODEW3 C3(009):C5(009) CODE-2 N2(010):N3(010) C2(010):C3(010) R2(020):R3(020) N3(010):N4(010) C3(010):C4(010) R3(020):R4(010) N2(010):N4(010) C2(010):C4(010) C4(010):C5(020) C3(010):C5(020) C2(010):C5(020) R2(020):R4(010) +6 +8 +4 +30 +23 +4 +33 +24 +94 +68 -99 +5 NCCCS-1 C2(010):C3(010) C3(010):C4(010) C2(010):C4(010) C4(010):C5(010) C3(010):C5(010) C2(010):C5(010) 0.77 +8 0.76 +9 0.46 0.57 0.36 0.21 +60 + 100 - 100 - 100 NCCCS-2 C2(010):C3(010) C3(010):C4(010) C2(010):C4(010) NCCCS-3 C2(010):C3(010) C3(010):C4(010) C2(010):C4(010) SMILE C2(010):C3(010) C3(010):C4(010) C2(010):C4(010) Middepth Correlations CODE-I C3(039):C4(065) 2028 C4(065):C5(152) 2028 C3(039):C5(152) 2028 8.13 12.64 20.75 0.02 0.06 -0.36 -0.30 -0.41 -0.40 2946 20.80 0.13 0.14 -78 0.17 0.17 -3 2250 20.83 0.03 0.16 + 75 0.25 0.25 +3 2682 2682 2682 2682 2682 2682 2682 2682 2682 2682 2682 2682 3.55 4.74 4.86 7.02 8.06 12.68 10.54 12.80 13.37 20.94 25.57 17.54 0.63 0.02 0.67 0.65 0.29 0.37 0.41 0.11 0.06 -0.19 0.03 0.23 0.63 -0.04 0.68 0.65 -0.29 0.44 0.41 0.14 -0.24 -0.22 -0.12 0.24 +0 -62 -3 +1 - 60 -13 +0 -22 -97 - 14 + 100 -4 0.91 0.86 0.88 0.73 0.66 0.65 0.51 0.38 0.33 0.24 0.15 0.54 0.92 0.87 0.88 0.74 0.67 0.65 0.53 0.44 0.34 0.24 0.27 0.54 +5 +6 +3 +6 +7 +2 + 10 +84 -8 +1 +94 +4 5718 21.11 -0.25 -0.38 -36 0.23 0.24 + 100 CODEW2 C3(055):C5(150) CODEW3 C3(055):C5(150) CODE-2 N2(035):N3(053) C2(035):C3(053) R2(035):R3(053) N3(053):N4(070) C3(053):C4(070) R3(053):R4(070) N2(035):N4(070) C2(035):C4(070) C4(070):C5(150) C3(053):C5(150) C2(035):C5(150) R2(035):R4(070) NCCCS-1 C3(045):C5(150) 8564 DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF Table 4. (continued) Maximum Lag Maximum Lag Distance, Moorings 0 Lag km Hours 0 Lag u Correlations u Correlations Hours v Correlations v Correlations Hours -0.02 -0.14 +26 MiddepthCorrelations(continued) NCCCS-2 C3(045):C5(150) 2107 21.44 0.19 0.19 +2 4285 4.90 0.50 0.52 -4 0.83 0.83 +0 2484 7.66 0.47 0.48 +3 0.78 0.78 +0 SMILE C3(045):C4(052) STRESS-2 C3(059):C4(059) Near-Bottom Correlations CODE-1 C1(027):C3(083) 2028 C3(083):C4(123) 2028 C1(027):C4(123) 2028 7.12 8.13 15.00 0.14 0.67 0.17 0.23 0.68 0.18 -12 +3 -6 0.49 0.81 0.30 0.66 0.81 0.44 + 19 + 1 + 18 4.74 0.36 0.37 -3 0.77 4.86 7.02 8.06 12.68 12.80 17.54 0.70 0.76 0.70 0.50 0.35 0.31 0.70 0.76 0.70 0.53 0.35 0.34 +1 +1 -1 -7 -4 -7 0.84 0.85 0.76 0.66 0.44 0.54 0.79 0.85 0.86 0.76 0.66 0.46 0.54 +6 +6 + 1 +2 -3 +9 +3 4.61 7.66 12.26 0.47 0.58 0.38 0.47 0.58 0.38 +2 +0 - 1 0.66 0.86 0.48 0.66 0.86 0.48 +3 - 1 +0 CODE-2 C2(053):C3(083) R2(053):R3(070) N3(083):N4(121) C3(083):C4(121) R3(070):R4(110) C2(053):C4(121) R2(053):R4(110) 2682 2682 2682 2682 2682 2682 2682 STRESS-2 C2(039):C3(080 ) 2484 C3(080):C4(120) C2(039):C4(120) 2484 2484 Positivelagsdenotethe first listedseriesleadingthe second. whichreducecorrelationlengthsof neareverywhere. For example,from earlyto mid-Februaryoffshore standthe processes surfaceu, 28 day subsetsof the 6 monthtime period are next absentat G3. Similarly,in late April, offshoreflow is observed considered.Twenty-eightdayswas chosenin order to give 10 degreesof freedomper subset.Correlations at M3, while onshoreflow is observedat G3. To better under- approximately flow at 10 m is observed at C3, NCCCS C3, and M3 but is (a) near-surface (b) u c 1 middepth u 1 ++ oo o 0.5 0.5 + o • o o o o ..• 0(9 0 (90 ,o 0 10 0 20 30 cross-shelfseparation in km (c) near-bottom ---0-• 10 20 30 cross-shelfseparation in km u 1 e ß x o ß + [] ß o% 0.5 • o o • o 0 • -0.% 10 20 CODE-1 CODEW2 CODEW3 CODE-2 NCCCS-1 SMILE/NCCCS-2 NCCCS-3 STRESS-2 30 cross-shelfseparation in km Figure7. Correlations of cross-shelf velocityu as a functionof cross-shelf mooringseparation for (a) near-surface, (b) middepth,and (c) near-bottom instruments. Correlations are alsopresented in Table4. Cross-shelfcorrelationscalesof u are approximately10 km. DEVER: VELOCITY NOV SCALES DEC ON THE JAN NORTHERN CALIFORNIA SHELF APR 8565 FEB MAR 20 30 10 20 30 10 20 30 10 20 10 20 30 10 20 30 MAY 10 20 30 10 20 30 10 20 30 10 20 10 20 30 10 20 30 10 3 -4 2O 10 0 -10 -20 -30 20 10 0 -20 -30 20 10 0 -10 -20 -30 20 10 0 -10 -20 -,30 NOV DEC JAN FEB MAR APR MAY Figure8. Timeseriesof low-pass filteredr y (dyncm-2) andu (cms-1) at 10m. Mooringlocations areas denotedin Figure2. Correlationsof u with r y rangefrom 0.5 to 0.6. Periodsof strongalong-shelfvariability in early February and late April are bracketedby solid lines. of 0.50 (0.58) are significantat the 90% (95%) level for 10 degreesof freedom. 4.1. Monthly Variation in Correlation Scales of Near-Surface Cross-Shelf Velocity Monthly plots of near-surfaceu correlationas a functionof along-shelfseparation(Figure9) showa greatdeal of variability. From mid-Novemberto mid-January,and againfrom midFebruary to early April, along-shelfcorrelationsshow little drop-off as a function of separation.However, from midJanuaryto early February and againfrom early April to early May, shorter along-shelfcorrelationscalesexist. Correlation scalesare always_>4km as the SMILE C3 and NCCCS C3 mooringsremain well correlatedover all months. Monthly variabilityin •'"g •h• ..... ,..,;..... ,• • .... surfaceu appearsto be related to the importanceof alongshelf wind stressforcing relative to other processes.During monthsin which correlation length scalesare long, all mooringstend to be highlycorrelatedwith localwind stress(Table 5). Conversely, when correlationlengthscalesare short,some or mostmooringsare lesswell correlatedwith along-shelfwind stress.Wind stressvariability, as indicated by its standard deviation, remains similar through most of the experiment. Therefore lower correlationsof near-surfaceu with along-shelf wind stressare primarilyattributednot to a weakeningof wind stressbut to the greater importanceof other processes.However,there is a slightweakeningof wind stressvariabilityfrom mid-Januarythroughearly February,and this playsa role in reducingcorrelationwith wind stressand hence along-shelf correlation scales. Betweenmid-Januaryand early February,near-surfaceu is we!!corr,q•t,•dbetwe...... i-gs NCCCS C3 and SMILE C3 and M3 but not between G3 and the other moorings. This suggests a break in cross-shelfcirculationbetweenthe northern and southernpartsof the studyarea rather than a general increasein short-scaleu fluctuations.This along-shelfvariability is alsomarkedby a three-dimensional heatbalancebetween January30 and February12 [Deverand Lentz, 1994,Figure 9] and by a stronglythree-dimensionalmassbalanceas indicated by depth-averaged cross-shelf flow at C3 betweenJanuary30 and February20 [Dever,1995].This depth-averaged cross-shelf flow, which is uncorrelated with the wind, also probably re- ducesalong-shelfcorrelationsof near-surface u betweenearly Februaryand early March (Figure 9) as shiftingthe start and stop times of the 28 day periodsshowsreducedalong-shelf 8566 DEVER: VELOCITY SCALES ON THE NORTHERN Nov 16, 1988 to Dec 14, 1988 o o CALIFORNIA Dec 14, 1988 to Jan 11, 1989 o o SHELF o o o 0.5 O.5 o o -0.5 -0.5 -1 o¸ o o -1 10 0 20 30 10 0 20 30 along-shelfseparation in km along-shelfseparation in km Jan 11, 1989 to Feb 8, 1989 Feb 8, 1989 to Mar 8, 1989 o o o 0.5 o 0.5 o o o o 1'0 2'0 3'0 o o o -0.5 -0.5 -1 1'0 0 2'0 -1 3'0 0 along-shelfseparation in km along-shelfseparation in km Mar 8, 1989 to Apr 5, 1989 Apr 5, 1989 to May 3, 1989 , o o o o 0.5 0.5 o o o 0 -0.5 -o.5 -10 1'0 2'0 3'0 -10 along-shelfseparationin km 1'0 2'0 3'0 along-shelfseparationin km Figure 9. Correlationsof near-surfacecross-shelf velocityu as a functionof along-shelfmooringseparation for six 1-month periods between November 1988 and May 1989. Correlations with wind stressare also presentedin Table 5. Table 5. Correlationsof Near-Surfaceu With Along-ShelfWind StressOver 1-Month Periods Period Nov. 16, 1988, to Dec. 14, 1988 Wind o-,dyncm-2 u scale, km 0.74 30 Dec. 14, 1988, to Jan. 11, 1989 0.72 30 Jan. 11, 1989, to Feb. 8, 1989 0.65 15 Feb. 8, 1989, to March 8, 1989 1.02 30 March 8, 1989, to April 5, 1989 0.95 30 April 5, 1989, to May 3, 1989 0.71 15 Mooring G3 C3 NCCCS M3 C3 0.77 0.78 0.68 0.71 0.82 0.63 0.67 0.73 0.44 0.41 0.38 0.69 0.61 0.60 0.79 0.75 0.81 0.63 0.79 0.73 Rough estimatesof u correlation scalesand the standarddeviations,o-,of r y for each month are also shown. *Short record. 0.56 0.76* 0.70 0.18 DEVER: (a)1 VELOCITY SCALES i i ON THE NORTHERN i CALIFORNIA SHELF 8567 i '.--0.,5 o I 0 0 I 0 3 O.4 O' .6 0.5 frequency,cpd (b) 15 10 E C3 • 1 ½. • .-- -5 -NCCCS C3 0.4 _1o[ 0.2 M3 I 15 .1 0 02 O.3 O.4 O.5 ] O.6 frequency, cpd Figure 10. (a) Coherenceof near-surfaceu at C3 with •-Y. Coherenceof other near-surfaceu recordswith •-Yalso declinewith decreasingfrequency.(b) Coherenceof near-surfaceu with distancefrom C3. The 90% confidence level is 0.45, and the 95% confidence level is 0.50. correlation scales (and correlations with along-shelf wind stress)which are most evident from late January to midFebruary and increaseagain in late February. During April, correlationscalesare reducedprobablyby the presenceof a mesoscalefeature over much of the shelf. This feature, describedby Largieret al. [1993], originatesfrom offshore as indicated by its high temperature and low salinity [Alessiet al., 1991]. It is responsiblefor strongequatorward flow over the studyarea despitethe absenceof persistentequatorwardwind forcing(Figure 8), and it affectsboth the nearsurfacecirculationand the heat and salt balances[Dever and Lentz,1994].Similarfeatureshavealsobeenobservedat different timesin previousyears[Lentz,1987;Washburnet al., 1993]. Largier et al. [1993] postulate that oceanic mesoscalefeatures commonlyoccur over the northern California shelf and that they are a significantsourceof forcingfor shelfcirculation at periodslongerthan 10 daysfor whichwind stressvarianceis decline for longer periods where oceanic mesoscaleforcing may become more important. The associationof the longestspatial scalesof near-surface u with wind stressforcing is also supportedby empirical orthogonalfunctions(EOFs) of near-surfaceu during SMILE and NCCCS (Figures 11 and 12). The lowest mode, which accountsfor 57% of the variance,is highly correlated(0.76) with along-shelfwind stress,is important throughoutthe November to May period, and exhibitslittle along-shelfstructure. Higher modes have more complex spatial structuresand are uncorrelatedwith the wind stress.They becomeimportant on timescales of weeks when correlation scales are reduced. Modes 2 and 3 become important during January and early February. Together they represent u fluctuations at G3 (in January)and C3 and NCCCS C3 (in February).Mode 2 again becomesimportant in late April when it representsoffshore flow 1-2_1 ....... at M3 _1__ and onshore ,,_1 ß flow at G3. The timescales of the mgnm muuc•, tnmr intermittent nature, and the lack of correancepeaksat periodsbetween2.5 and 5 days[Dever,1995]. If lation with wind stressall suggestthat they are not associated wind stressforcing dominates at short periods and oceanic with wind forcing. mesoscalevariabilityis the primary sourceof forcingfor longer periods,then spatial coherenceof near-surfaceu should be 4.2. Monthly Variation in Cross-Shelf Velocity Correlation Between SMILE C3 and NCCCS C3 highestin the wind band asthere is no reasonto expectoceanic Velocity measurementat most SMILE and NCCCS moormesoscaleforcingto be two-dimensional.Figure 10 lendssome supportto this idea. Both coherencewith the wind stressand ingswas limited to the near surface(10 m). However, C3 was along-shelfcoherenceare largestat the samefrequenciesand instrumentedthroughout the water column. The C3 sites of -- WlllU 8568 DEVER: 15 VELOCITY SCALES E NORTHERN CALIFORNIA SHELF Therefore, in the absenceof other processes,theoretical u correlation scaleswithin the surface and bottom boundary layersshouldreflect thoseof along-shelfwind stressand interior v, respectively.Over the northern California shelf these along-shelfcorrelationscalesare 100 km or more (Figure 3) and over 60 km (Figure 4). The theoretical correlation scales above are larger than thoseobservedon the northern California shelf. This is especially true for u. Observedv correlation scalesare next dis- G3% e• E 10 o ON THE 5 o cussed,and a more extensive discussionof u correlation scales c 0 C3 follows. o....,` Correlations of v estimated in this study and previously [Winant et al., 1987] over the northern California shelf and '1,,i elsewhere[KunduandAllen, 1976]are in manywaysconsistent ..c -5 NCCCS C3 with those expectedfrom CTWs. The correlation scale of v estimatedhere, though almost certainly less than the limits o impliedby CTWs, is greaterthan 60 km (Figure 4). Maximum < -lO lagged v along-shelfcorrelation coefficientstend to exhibit behavior consistentwith CTWs in that southern velocity recordsgenerallylead northern velocityrecordsby about 3 to -15 12 hours (Table 3). Application of CTWs is limited to the -5 0 5 midshelfand outer shelfwhere surfaceand bottom boundary Mode Amplitudein cm/s layerscan be consideredthin. The increasingimportanceof Figure 11. Along-shelfstructureof near-surfaceu empirical mesoscaleprocessesoffshoreand the cross-shelfstructureof orthogonalfunctions(EOFs). The solidline indicatesmode 1 CTW modeson the northernCaliforniashelf[Chapman,1987] which accounts for 57% of the total variance. Modes 2, 3, and would suggestthat along-shelfcorrelationscalesdecreaseoff4, indicated by the dashed,dash-dotted,and dotted lines, ac- shore. In this study, examination of along-shelfcorrelation countfor 24%, 13%, and 5% of the total variance,respectively. scales of v was limited to CODE-2 observations between 60 and 130 m. Within this region, little cross-shelfvariation in along-shelfv correlationsis evident except for near-surface NCCCS, SMILE, and STRESS were locatedapproximately4 observationswhere along-shelf v correlations at the 130 m km apart in the along-shelfdirection.They providean oppor- isobathare reducedslightlyrelative to thoseat the 60 and 90 m tunity to examinecorrelation in the along-shelfdirection for isobaths.This is a possibleindicationthat mesoscaleprocesses near-surface,middepth,and near-bottomdepths.Correlations are more important near the surface. In the cross-shelfdirection, v is generallycorrelatedfrom 60 betweenthesetwo sites(Figure 13) showedthat near-surface correlationswere the highestand that monthly trends in cor- to 130m. Despiteits smallermagnitudeand possibleeffectsof relationswere the same at all depths.This indicatesthat re- localbathymetry,inner-shelfv (30 m) is alsosignificantlycormarksconcerningcorrelationscalesof near-surfaceu may also relatedout to the midshelf(90 m), arguingfor a similarresponsein v to wind forcingover muchof the shelf.Maximum be applicableto interior and near-bottomu. laggedv cross-shelfcorrelationstend to exhibitbehaviorconsistentwith two-dimensionalupwellingmodelsin that inshore 5. Discussion velocityrecordsgenerallylead offshorevelocityrecords(Table On wind-forced shelves,CTWs and Ekman theory have 4), suggesting a morerapidresponse to windforcingin shallow been compared quantitatively with observed interior [e.g., water. Along-shelf correlation scalesof u are much shorter than Chapman, 1987] and boundarylayer flows [e.g.,Dever, 1995; Lentz, 1992]. CTWs and Ekman theory both imply correlation those of v and were best resolvedby near-surfacemeasurescalesfor u and v. For v, velocities associatedwith CTWs are ments during SMILE and NCCCS (Figure 5). Near-surface much larger than surfaceand bottom Ekman velocitiesso that and near-bottomalong-shelfscalesof u are substantiallyless the wind-forcedv is dominatedby the CTW responsethrough- than those of along-shelfwind stressand interior along-shelf out the water column.Brink et al. [1987, 1994] have examined velocitywhich set the scalesof u in models of surfaceand the scalesof motionimpliedby full (i.e., no long-waveassump- bottomboundarylayers.Figures8, 9, 10, and 12 and Table 5 all tion) CTWs and found that v is resonantfor low wavenumber suggestthat wind forcingis associatedwith the longestscalesof along-shelfwind stressvariabilityimplyingcorrelationscalesof near-surfaceu in winter and springand that other processes 100-200 km. act to reducecorrelationscaleson periodsof severalweeksto Brink et al. [1987, 1994] also found u velocitiesassociated i month (Figures8 and 9 and Table 5). The timescalesand with CTWs to be more sensitiveto high wavenumberalong- spatialstructuresof variabilityin short-scaleu are consistent shelfwind stresswith scalesof 15 km or less,implyingsimilarly with those of mesoscale features. The similar time variation of shortcorrelationscalesfor u forcedby CTWs. However,unlike correlationsbetween two nearby moorings at near-surface, v the character of the total wind-forced u is not determined middepth,and near-bottomlocations(Figure 13) suggests the mainly by the CTW response. Cross-shelfboundary layer processes which act to reduce near-surfaceu correlationare transportsare the samemagnitudeas CTW cross-shelftrans- also important throughoutthe water column. Examinationof u along-shelfcorrelationscaleswas limited ports, and becausethey are confinedto relativelythin layers, the velocitiesassociatedwith them are greater [Dever,1995]. to the period between November 1988 and May 1989. This M3 i .,"½1" ;/ i DEVER: VELOCITY NOV 20 30 SCALES ON THE NORTHERN DEC 10 20 JAN 30 10 20 30 CALIFORNIA FEB MAR 10 20 10 20 SHELF APR 30 10 20 8569 MAY 30 10 O -3 -4 15 i i -lO - - -15 15- - 10- - "'-'-'V E- 1 -15 v 10 E- 1 -15 10 -5 -lO -- I 5 ••j••••••••••••• 20 ,30 10 NOV 20 DEC / ,30 10 20 JAN 30 10 20 FEB 10 20 MAR 30 10 20 APR ,30 10 MAY Figure12. Timeseries of •-Y(dyncm-2) andnear-surface u (cms-l) EOF modes. To lengthen theEOF timeseriesintolateApril, theSMILE timeseriesat C3 hasbeenextended bypatchingthe8.5 and11.5m time seriestogether.The correlationcoeflScient of mode1 with •-Yat C3 is 0.76.Higher modesare uncorrelated with •-Y. period coveredthe winter and springstormseasonand ended with this front, and with later fronts causedby wind stress daysafter the springtransitionto upwelling;hencethe appli- relaxationand subsequentupwelling,are a ready sourceof cability of the above observationsto u correlation scalesin the three-dimensional variabilityin summer.In thisway,upwelling summer(upwelling)seasonis uncertain.Section4.1 showed favorablewindswhichpersistlong enoughto causethe develthat for someperiodsof 1 month or more, along-shelfcorre- opment of an upwellingfront may be an indirect sourceof lation scaleswere at least 30 km, the maximumalong-shelf short correlationlength scales. mooringseparationin SMILE. Though30 km is the minimum The importanceof short-scalewind stressin reducingcoralong-shelf mooringseparationduringCODE-2, makingit dif- relation scalesof interior u remainsuncertain.Thoughalongficult to comparespatialscales,a similartreatment of summer shelfwindstressis highlycorrelatedoverdistances of up to 130 CODE-2 data revealed no time when correlation scales were as km for both SMILE and CODE-2 (Figure 3), aircraftobsergreat as 30 km. This suggests the possibilityof seasonal(or vationsmake it clear that short-scalewind stressis present interannual)variation in along-shelfscalesof near-surfaceu. southof PointArena [Winantet al., 1988;Dotmanet al., 1995]. Seasonalvariabilitycouldbe due to greateroffshoremesoscale Neither the aircraftobservations nor the availablebuoyobseractivityin the summer[Kosroet al., 1991]whichwould tend to vations allow the estimationof the completewavenumber shorten correlation length scalesin summer. Offshore me- spectrum.Thoughinclusionof an estimatedshort-scale (down soscalevariability in summermay or may not be related to to 10 km) wind stressin a stochastic model[Brinket al., 1994] instabilityof an upwellingfront. Barth [1994] examinedthe reduces interior u scales, observed u fluctuations are more developmentof frontal instabilitiesand found that the time- energeticthan predictionsof the linear model. Cross-shelf correlation scales of u are about 10 km. These scalefor development of a short-wavelength (20 km) baroclinic instabilitywas about 1.5 days.In winter, wind forcingevents correlationscalesextend over a substantialportion of the tendto be relativelybrief anddo not lead to the development northernCaliforniashelfand are only about5 km lessthan the of an upwellingfront. After the springtransitionto upwelling, cross-shelf correlation scales of v. There is some indication a front developsand moves offshore.Instabilities associated that u is more weakly correlatedfrom the inner shelf to the 8570 DEVER: VELOCITY SCALES ON THE NORTHERN CALIFORNIA SHELF i 0.8- 00.6o •0.4o ! ! 0.2- 0 -do i o i do do i do i 1989 yearday Figure 13. Correlationsof cross-shelf velocityu betweenSMILE and NCCCS C3 moorings.Near-surface (solid),middepth(dashed),and near-bottom(dotted) instrumentsare consideredfor sixone-monthperiods betweenNovember 1988 and May 1989. Though correlationsare highestbetweennear-surfaceinstruments, monthlytrends are similar for all instrumentdepths. midshelf(60 to 90 m, a distanceof approximately5 km) than from the midshelfto the outer shelf(90 to 130 m, a distanceof ---8 km). Becausewe expect offshoremesoscalefeaturesto but are less than 25 km. In the cross-shelf direction, u corre- reduce lation correlation scales over the outer shelf rather than the and indicate that they are 15-20 km. Along-shelf correlation scalesof subsurfaceu are not well resolvedby availabledata scales are ---10 km. There is some indication that u is inner shelf, it is likely some other processis responsiblefor morehighlycorrelatedbetweenthe 90 and 130m isobaths than this. As most of the observations between used in Table 4 come from summer, one reasonableexplanationis near-shorevariability causedby relaxationfrom upwelling[Sendet al., 1987].During the 60 and 90 rn isobaths. To investigatefurther the processes whichdeterminealong- shelf correlation scalesof u, SMILE and NCCCS recordsfrom November 1988 to May 1989 were examinedin greater detail. from Point Reyesoccursin a wedgeconfinedprimarilyto the Monthly variation of correlationscaleswas comparedto corinner-shelf and midshelf locations. This could cause a break in relation with along-shelfwind stress,the heat balance, and the character of the flow field between the 60 and 90 m isoother descriptiveinformation about shelf circulation.During baths. This type of three-dimensionalvariability could over- several months the along-shelf correlation scale of nearwhelm the wind-drivensignalespeciallyover the inner shelf surfaceu was at least 30 km, the maximummooring separawhere models suggestwind-driven u is reduced relative to tion, and it was alwaysgreater than 4 km, the minimum moordeeper water by the overlap of surfaceand bottom stresses ing separation.Along-shelf correlation scaleswere generally [Mitchurnand Clarke, 1986;Lentz, 1994]. longwhencorrelationof velocitywith wind stresswashighand relaxation events, northward advection of warm saline water short when correlation 6. Summary Correlationsof subtidalcross-shelf(u) and along-shelf(v) velocitiesare estimatedas a functionof along-shelfand crossshelfdistanceusingmooredtime seriesfrom severalfield programsover the northernCalifornia shelf.Distancesoverwhich velocities are significantlycorrelated are termed correlation with wind stress was low. Short corre- lation scalescoincided with three-dimensional heat, salt, and volumebalances.During April 1989, short correlationscales coincided with the intrusion of an offshore mesoscale feature onto the shelf. Along-shelf correlationsof near-surface,middepth, and near-bottomu between two nearby (4 km) sites showed near-surface correlations were highest but that monthlytrendsin correlationswere similar at all depths. scales. Over periods of 4-6 months, correlation scalesof v are resolvedin both the along-shelfand cross-shelfdirections.In the along-shelfdirection, v is significantlycorrelatedfor distancesgreaterthan 60 km, the maximummooringseparation. In the cross-shelfdirection, v is generallycorrelatedbetween the 60 and 130 m isobaths(10-15 km). Correlationsof v show little dependenceon instrumentdepth;subsurface v is perhaps more highly correlatedthan surfacev. Correlationscalesof u overperiodsof 4-6 monthsare much smallerthan thoseof v and are often not resolvedby minimum mooringseparations.Mooringsdeployedin SMILE and NCCCS do resolvealong-shelfcorrelationscalesof near-surfaceu Acknowledgments. This work was done as part of my thesisin the MIT/WHOI Joint Program in Oceanography.Financial supportwas providedby NSF grantOCE 91-15713.I thankespeciallymy advisorS. Lentz and alsocommitteemembersJ. Price, K. Brink, R. Beardsley,P. Malanotte-Rizzoli, and R. Weller for helpful discussions and comments.The careful commentsof two reviewersare also appreciated. Specialthanksare alsoextendedto B. Butman at USGS WoodsHole andJ. Trowbridge(WHOI) for makingavailableSTRESSdata and to J. L. Largier (SIO, Universityof Cape Town, South Africa), B. A. Magnell (EG&G), and C. D. Winant (SIO), who generouslymade availableNCCCS data. Supportfor preparationof the manuscriptat SIO was providedby The Mellon Foundationand by the Minerals ManagementServiceunder cooperativeagreement14-35-0001-30571. Woods Hole OceanographicInstitution contribution9187. DEVER: VELOCITY SCALES ON THE References NORTHERN CALIFORNIA SHELF 8571 Kundu, P. K., and J. S. Allen, Some three-dimensionalcharacteristics of low-frequencycurrentfluctuationsnear the Oregoncoast,J. Phys. Al½ssi,C. A., S. J. Lcntz, and R. C. Beardslay,The shelf mixed layer Oceanogr.,6, 181-199, 1976. experiment(SMILE): Programoverviewand moored and coastal Largier, J. L., B. A. Magnell, and C. D. Winant, Subtidalcirculation array data report, WHOI Tech.Rep. 91-39, 211 pp., Woods Hole over the northern California shelf, J. Geephys.Res., 98, 18,147Occanogr.Inst., WoodsHole, Mass.,1991. 18,179, 1993. Allen, J. S., Models of wind-driven currentson the continental shelf, Lentz, S. J., A heat budget for the northern California shelf during Annu. Rev. Fluid Mech., 12, 389-433, 1980. CODE 2, J. Geephys.Res., 92, 14,491-14,509, 1987. Allen, J. S., ½tel., Physicaloceanography of continentalshelves,Rev. Lentz, S. J., The surfaceboundarylayer in coastalupwellingregions,J. Geophys.,21, 1149-1181, 1983. Phys.Oceanogr.,22, 1517-1539, 1992. Berth, J. A., Short-wavelength instabilitieson coastaljets and fronts,J. Lentz, S. J., Current dynamicsoverthe northernCaliforniainner shelf, Geophys.Res., 99, 16,095-16,115, 1994. J. Phys. Oceanogr.,25, 2461-2478, 1994. Beardslay,R. C., and S. J. Lcntz, The CoastalOcean DynamicsEx- Lentz, S. J., U.S. contributionsto the physicaloceanographyof contiperimentcollection:An introduction,J. Geophys. Res.,92, 1455nental shelvesin the early 1990s,Rev. Geephys.,33, 1225-1236, 1995. 1464, 1987. Lentz, S. J., and D.C. Chapman,Seasonaldifferencesin the current Beardsley,R. C., C. E. Derman, C. A. Friehe, L. K. Rosenfeld,and and temperaturevariabilityover the northernCaliforniashelf durC. D. Winant, Local atmosphericforcingduringthe CoastalOcean ing the CoastalOcean DynamicsExperiment,J. Geephys.Res., 94, DynamicsExperiment, 1, A descriptionof the marine boundary 12,571-12,592, 1989. layer and atmosphericconditionsover a northern California up- Lentz, S. J., andJ. H. Trowbridge,The bottomboundarylayeroverthe welling region,J. Geephys.Res.,92, 1467-1488, 1987. northern California shelf,J. Phys. Oceanogr.,21, 1186-1201, 1991. Brink, K. H., The near-surfacedynamicsof coastal upwelling,Prog. Limeburner,R. (Ed.), CODE-2: Moored array and large-scaledata Oceanogr.,12, 223-257, 1983. report, WHOI Tech.Rep. 85-35, 234 pp., Woods Hole Oceanogr. Brink, K. H., Coastal-trapped wavesandwind-drivencurrentsover the Inst., Woods Hole, Mass., 1985. continental shelf,/lnnu. Rev. Fluid Mech., 23, 389-412, 1991. Mitchum, G. T., and A. J. Clarke, The frictionalnearshereresponseto Brink, K. H., D.C. Chapman,and G. R. Halliwell, A stochasticmodel forcing by synopticscale winds, J. Phys. Oceanogr.,16, 934-946, for wind-drivencurrentsover the continentalshelf,J. Geephys.Res., 1986. 92, 1783-1797, 1987. Rosenfeld,L. (Ed.), CODE-l: Moored array and large-scaledata Brink, K. H., J. H. LaCasce, and J. D. Irish, The effect of short-scale report, WHOI Tech.Rep. 83-23, 186 pp., Woods Hole Oceanogr. wind variations on shelf currents,J. Geephys.Res., 99, 3305-3314, Inst., Woods Hole, Mass., 1983. 1994. Send, U., R. C. Beardsley,and C. D. Winant, Relaxation from upChapman,D.C., Application of wind-forced,long, coastal-trapped welling in the Coastal Ocean DynamicsExperiment,J. Geephys. wave theory alongthe California coast,J. Geephys.Res., 92, 1798Res., 92, 1683-1698, 1987. 1816, 1987. Smith, R. L., A comparisonof the structureand variabilityof the flow Csanady,G. T., Circulationin the CoastalOcean,279 pp., D. Reidel, field in three coastalupwelling regions:Oregon, northwestAfrica, Norwell, Mass., 1982. and Peru, in Coastal Upwelling,CoastalEstuarineSci., vol. 1, edited Dever, E. P., Subtidal cross-shelfcirculation on the northern Califorby F. A. Richards,pp. 107-118, AGU, Washington,D.C., 1981. nia shelf, Ph.D. thesis, Mass. Inst. of Technol./Woods Hole OceanSmith, R. L., The physicalprocessesof coastaloceanupwellingsysogr. Inst., Woods Hole, 1995. tems, in Upwellingin the Ocean: Modern Processesand Ancient Dever, E. P., and S. J. Lentz, Heat and salt balancesover the northern Records,edited by C. P. Summerhayeset al., pp. 39-64, John Wiley, California shelf in winter and spring,J. Geephys.Res., 99, 16,001New York, 1995. 16,017, 1994. Strub, P. T., J. S. Allen, A. Huyer, R. L. Smith, and R. C. BeardsIcy, Derman, C. E., A. G. Enriquez, and C. A. Friehe, Structureof the Seasonalcyclesof currents,temperatures,winds, and sea level over loweratmosphereover the northernCaliforniacoastduringwinter, the northeast Pacific continental shelf: 35øN to 48øN, J. Geephys. Men. Weather Rev., 123, 2384-2404, 1995. EG&G, Inc., Northern California Coastal Circulation Study data re- port no. 2: Main measurementprogram, March-August 1988, Oceanogr.Serv. Rep. NCCCS-89-17, 2 vols., 442 pp., Washington Anal. Serv. Cent., Waltham, Mass., Sept. 1989. EG&G, Inc., Northern California Coastal Circulation Study data report no. 3: Main measurement program,August1988-March 1989, Oceanogr.Serv. Rep. NCCCS-90-4, 2 vols., 344 pp., Washington Res., 92, 1507-1526, 1987. Trowbridge,J. H., and S. J. Lentz, Asymmetricbehaviorof an oceanic bottom boundarylayer above a slopingbottom,J. Phys.Oceanogr., 21, 1171-1185, 1991. Washburn, L., M. S. Swenson,J. L. Largier, P.M. Kosro, and S. R. Ramp, Cross-shelfsedimenttransportby an anticycloniceddy off northern California, Science,261, 1560-1564, 1993. Winant, C. D., Longshorecoherenceof currents on the southern California shelf during the summer,J. Phys.Oceanogr.,13, 54-64, Anal. Serv. Cent., Waltham, Mass., March 1990a. EG&G, Inc., Northern California Coastal Circulation Study data re1983. port no. 4: Main measurementprogram, March-October 1989, Winant, C. D., R. C. Beardsley,and R. E. Davis, Moored wind, temOceanogr.Serv. Rep. NCCCS-90-5, 2 vols., 326 pp., Washington perature, and current observationsmade duringCoastalOcean DyAnal. Serv. Cent., Waltham, Mass., Aug. 1990b. namicsExperiments1 and 2 over the northernCaliforniashelfand Fredericks,J. J., J. H. Trowbridge,A. J. Williams III, S. J. Lentz, B. upper slope,J. Geephys.Res., 92, 1569-1604, 1987. Butman, and T. F. Gross, Fluid mechanical measurementswithin Winant, C. D., C. E. Derman, C. A. Friehe, and R. C. Beardsley,The the bottomboundarylayerover the northernCaliforniacontinental marine layer off northern California: An exampleof supercritical shelfduringSTRESS, WHOI Tech.Rep.93-32, 116pp., WoodsHole channel flow, J. Atmos. Sci., 45, 3588-3605, 1988. Oceanogr.Inst., WoodsHole, Mass.,1993. Halliwell, G. R., Jr., and J. S. Allen, The large-scalecoastalwind field E. P. Dever, Center for Coastal Studies0209, ScrippsInstitution of alongthe westcoastof North America,1981-1982,J. Geephys.Res., Oceanography,University of California, San Diego, 9500 Gilman 92, 1861-1884, 1987. Drive, La Jolla, CA 92093-0209.(e-mail: dever@coast.ucsd.edu) Janowitz, G. S., and L. J. Pietrafesa, A model and observationsof time-dependentupwellingover the mid-shelf and slope,J. Phys. Oceanogr.,10, 1574-1583, 1980. Kosro, P.M., et al., The structure of the transition zone between coastalwaters and the open ocean off northern California, winter (ReceivedApril 10, 1996; revisedSeptember19, 1996; acceptedOctober 3, 1996.) and spring1987,J. Geephys.Res.,96, 14,707-14,730,1991.