PhytoplanktonprimaryproductivityintheSantaBarbaraChannel: …washburn/docs/brzezinski_washburn... · 2012-07-03 · 2.2. Water Sampling [11] Water samples were collected using a CTD/rosette

Phytoplankton primary productivity in the Santa BarbaraChannel:Effects of wind-driven upwelling and mesoscale eddies

Mark A. Brzezinski1 and Libe Washburn2

Received 20 June 2011; revised 26 September 2011; accepted 27 September 2011; published 13 December 2011.

[1] The patterns and drivers of phytoplankton primary productivity in the Santa BarbaraChannel (SBC) were examined on 16 cruises conducted 3 times each year from 2001to 2006. Empirical orthogonal function (EOF) analysis revealed 3 modes of variability thatexplained 89% of the variance in the productivity data set. The first mode, strongest duringspring, describes seasonal productivity changes driven by coastal upwelling. The secondmode, largest in spring and fall, describes productivity changes over the western SBCassociated with cyclonic eddies. Eddy-enhanced productivity changes were superimposedon variable levels of channel-wide productivity caused by upwelling within the SBC.The cyclonic eddies influenced productivity through enhanced nutrient supply associatedwith the uplift of isopycnal surfaces and through the occasional entrainment ofphytoplankton and nutrients from water upwelled north of Point Conception. The thirdEOF mode describes productivity gradients on the continental shelf along the mainlandcoast with enhanced productivity in the east during spring and fall. Overall, our analysisshows that coastal upwelling combined with the effects of cyclonic circulation onparticle retention and vertical nutrient supply combine to enhance phytoplanktonbiomass and productivity in the western SBC.

Citation: Brzezinski, M. A., and L. Washburn (2011), Phytoplankton primary productivity in the Santa Barbara Channel: Effectsof wind-driven upwelling and mesoscale eddies, J. Geophys. Res., 116, C12013, doi:10.1029/2011JC007397.

1. Introduction

[2] The Santa Barbara Channel (SBC) is a region ofenhanced phytoplankton biomass and primary productivitywithin the region from Point Conception to the borderbetween the United States and Mexico known as theSouthern California Bight [Mantyla et al., 1995]. The SBCis a 40 km wide by 100 km long area that is bounded to thenorth by the California mainland and to the south by theNorthern Channel Islands (Figure 1a). The western SBCencompasses a 600 m deep depression in the seafloor calledthe Santa Barbara Basin (Figure 1b). Although the distribu-tion of phytoplankton biomass in the SBC is complex, thelong-term climatology of chlorophyll concentration derivedfrom satellite ocean color (Figure 1a) reveals two distinctivefeatures: (1) a maximum north of Santa Rosa Island and overthe Santa Barbara Basin and (2) another maximum extend-ing along the mainland coast in the eastern SBC where thecontinental shelf widens (Figure 1b). In this paper weexamine how these spatial patterns in phytoplankton biomassobserved from space are related to underlying spatial andtemporal patterns in phytoplankton primary productivity.

[3] Phytoplankton primary productivity (hereafter calledproductivity) in the SBC supports a productive pelagicecosystem with large populations of fishes, seabirds andmarine mammals [Beers, 1986; Fiedler et al., 1998]. Thisproductivity also constitutes a subsidy to nearshore benthicand rocky reef communities [Miller et al., 2011]. In someareas of the SBC, near Coal Oil Point for example, the bot-tom slopes steeply (Figure 1b) such that processes occurringoffshore can readily transport materials to and from shallowwater ecosystems [Bassin et al., 2005; McPhee-Shaw et al.,2007; Fram et al., 2008]. In other areas, such as the easternSBC, the bottom slopes gradually and any transport pro-cesses providing nearshore subsidies are less clear. Under-standing the mechanisms driving productivity within theSBC may also help explain the increasing frequency ofharmful algal blooms in the region since 2000 [Andersonet al., 2006, 2008, 2009].[4] The overall high levels of phytoplankton biomass in

the SBC are consistent with upwelling-favorable regionalwind patterns. Wind patterns driving upwelling within theSBC differ from those causing upwelling along most of thewestern coast of the United States owing to the transition ofthe coastline from a north-south orientation north of PointConception to an east-west orientation east of Point Con-ception. Consequently, wind stress often has an offshorecomponent with respect to the mainland coast that con-tributes to upwelling [Cudaback et al., 2005; Fewings et al.,2008]. Upwelling-favorable winds occur during all seasons,but in spring they are more persistent and more uniformly

1Department of Ecology Evolution and Marine Biology, MarineScience Institute, University of California, Santa Barbara, California, USA.

2Department of Geography, Marine Science Institute, University ofCalifornia, Santa Barbara, California, USA.

Copyright 2011 by the American Geophysical Union.0148-0227/11/2011JC007397

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, C12013, doi:10.1029/2011JC007397, 2011

C12013 1 of 17

http://dx.doi.org/10.1029/2011JC007397

equatorward over the SBC [Oey et al., 2001; Dorman andWinant, 2000]. In fall there is a larger gradient in equator-ward winds with weaker winds in the eastern channel. Fol-lowing Harms and Winant [1998], we define equatorwardto be eastward for locations east of Point Conception orsouthward for locations north of Point Conception. Polewardis westward east of Point Conception and northward northof Point Conception. Upwelling-favorable winds are morepersistent from Point Conception northward where theydrive offshore Ekman transport from spring into fall [e.g.,Harms and Winant, 1998; Dorman and Winant, 2000].These temporal and spatial patterns of upwelling-favorablewinds lead to strongly seasonal patterns of phytoplanktonbiomass and productivity within the SBC [Shipe andBrzezinski, 2003] with highest levels in early spring andoccasional high levels in summer and fall [Shipe andBrzezinski, 2003].[5] The maximum in chlorophyll north of Santa Rosa

Island and over the Santa Barbara Basin (Figure 1a) coin-cides with the location of a recurring cyclonic flow; Harms

and Winant [1998, Figure 10c] depict this flow, and wediscuss and show it below (see Figure 9b). The cyclonicflow arises from a balance between poleward pressure gra-dient forces and equatorward wind stress, and is mostprevalent during summer and fall [Dever et al., 1998; Harmsand Winant, 1998; Winant et al., 2003]. The polewardpressure gradient forces result from lower sea level northof Point Conception that is driven by strong upwelling-favorable winds and higher sea level in the more shelteredSouthern California Bight. They drive poleward currentsalong the mainland coast and are part of a general pattern ofpoleward flow in the Southern California Bight [Lynn andSimpson, 1987]. The other component of the cyclonic flowis equatorward currents along the northern Channel Islandsthat are driven by the equatorward wind stress. The cyclonicflow often closes to form eddies that typically last a fewdays, but occasionally persist for weeks [Nishimoto andWashburn, 2002]. Beckenbach and Washburn [2004] foundthat these eddies interact strongly with the Santa BarbaraBasin. After forming, the eddies often propagate westward

Figure 1. (a) SeaWiFS chlorophyll a climatology during 1999–2008 for the Santa Barbara Channel.(b) Map of study area showing shoreline, bathymetry (first at 50 m, then at 100 m intervals from 100 m),track for towed vehicle surveys (solid line), location of conductivity-temperature-depth grid stations(numbered circles), and Plumes and Blooms stations (numbered pluses). Letters A–H designate cross-channel transects along the cruise track. Gray triangles indicate locations of high-frequency radars formeasuring surface currents. Boxes and highlighted segments of the cruise track denote regions used invorticity calculations as described in text.

BRZEZINSKI AND WASHBURN: SANTA BARBARA CHANNEL PRODUCTIVITY C12013C12013

2 of 17

out of the SBC [Beckenbach and Washburn, 2004; Harmsand Winant, 1998].[6] The role of eddies in driving phytoplankton rate pro-

cesses and carbon cycling in the open ocean has been stud-ied in mesoscale eddies off Hawaii [Benitez-Nelson et al.,2007] and in the Sargasso Sea [McGillicuddy et al., 2007,1998; Siegel et al., 1999]. Open-ocean cyclones and modewater eddies influence productivity through isopycnal upliftduring the spin-up phase of eddy formation that raisesnutrient-rich deep waters into the lower euphotic zonestimulating productivity. Similar isopycnal uplift has beenobserved in cyclonic eddies in the western SBC [e.g.,Nishimoto and Washburn, 2002], but links to productivityhave not been examined.[7] A number of processes may explain the maximum in

chlorophyll along the mainland coast in the eastern SBC(Figure 1a). For example, Otero and Siegel [2004] speculatethat it is caused by enhanced productivity arising from ter-restrial runoff and sediment resuspension events. Rainfalland runoff are restricted mainly to winter when growthconditions for phytoplankton in the offshore regions of theSBC are generally poor. When runoff occurs under morefavorable condition during late winter or early spring, ele-vated phytoplankton biomass has been observed offshorealong the periphery of freshwater runoff plumes [Otero andSiegel, 2004]. Freshwater runoff occurs all along the main-land coast where numerous streams drain into the SBC, butinputs are greatest in the eastern SBC where the shelfbroadens and where the two largest rivers in the region, theSanta Clara and Ventura Rivers, discharge into the SBC(Figure 1b). In addition to terrestrial influences, internaltides may bring nutrients from offshore to the inner conti-nental shelf. Lucas et al. [2011] suggest nutrients suppliedby the internal tide drive productivity on the inner shelf inthe Southern California Bight. These processes may beimportant along the mainland in the eastern SBC wherewind-driven upwelling is typically weak.[8] In this study we evaluate the patterns and drivers of

spatial and temporal variability in productivity across theSBC as part of the Santa Barbara Coastal Long-Term Eco-logical Research (SBC-LTER) project. The dominant spatialand temporal patterns of productivity were accessed relativeto physical processes in the SBC including wind-drivenupwelling, eddy circulation, and freshwater runoff.

2. Methods

2.1. Seasonal Cruises

[9] Productivity and environmental conditions wereassessed on 16 cruises from 2001 to 2006 in the SBC aboardthe R/V Point Sur. Three cruises per year were timed tocapture the extremes in the annual production cycle. Cruisesduring January through March (hereafter referred to aswinter cruises) were timed to capture the winter low inproductivity and the influence of terrestrial runoff fromwinter storms. Cruises during April and May (hereafterreferred to as spring cruises) sampled the spring bloomperiod when upwelling-favorable winds are strongest andmost persistent. Cruises during September and October(hereafter referred to as fall cruises) sampled the morestratified, oligotrophic fall period. Cruises were not con-ducted earlier in summer for both logistical reasons and the

finding of previous studies of similar productivity duringsummer and fall [Shipe et al., 2002; Shipe and Brzezinski,2001]. During each cruise, two surveys were conductedusing towed vehicles and shipboard acoustic Doppler currentprofilers (ADCPs). Data from the towed vehicles are notshown here, but ADCP data from those surveys are. Thecruise track for the towed vehicle surveys consisted of eightcross-channel transects denoted A through H beginning atPoint Conception and ending near Ventura, CA (Figure 1b).For most cruises, one survey was conducted at the beginningof the cruise and a second at the end of the cruise.[10] Between the towed surveys, conductivity-temperature-

depth (CTD) casts to collect water samples for chemical andbiological measurements were made at 25 stations (hereaftercalled the grid surveys) distributed along the survey trackand also along a 7 station cross-channel transect that ismonitored approximately monthly by the Plumes andBlooms (PnB) project (see www.icess.ucsb.edu/PnB/PnB.html and Figure 1b). In this analysis only data from the PnBline collected during the 16 cruises were used; monthly datafrom the PnB project were not used. ADCP data were alsocollected during the grid surveys.

2.2. Water Sampling

[11] Water samples were collected using a CTD/rosettesystem equipped with 12 L Niskin bottles fitted with vinylcoated springs, a Sea-Bird Electronics 911 CTD, a ChelseaAqua 3 fluorometer, a Wetlabs transmissometer, and a Bio-Spherical PAR sensor. Water samples for inorganic nutrientand chlorophyll a concentration analysis were collected atall 25 grid stations at depths of 0, 5, 10, 15, 25, 50 and 75 mand at the PnB stations at each of the seven light depths usedfor measuring profiles of primary productivity (see methodsfor productivity below). Samples for inorganic nutrientanalyses were frozen in plastic scintillation vials and ana-lyzed by the UCSB Marine Science Institute’s AnalyticalLaboratory using flow injection techniques [Johnson et al.,1985] on a QuickChem 8000 analyzer (Lachat InstrumentsDivision, Zellweger Analytics) as described in the work ofAnderson et al. [2006]. Detection limits for nitrate, ortho-phosphate, and silicic acid were 0.1, 0.05 and 0.2 mM,respectively. For chlorophyll a analysis, 125–250 mL ofseawater were filtered through Millipore HAWP 45 mmcellulose filters, which were immediately frozen at �20°C.Just prior to analysis the filters were extracted in 90% ace-tone for 24 h at�20°C. The fluorescence of each extract wasmeasured with and without acidification to determine chlo-rophyll a concentrations on a Turner Designs 10AU digitalfluorometer that had been calibrated with pure chlorophyll a(SIGMA Chemical). Analytical precision for the chlorophylla determinations is better than 5%.

2.3. Primary Productivity Measurements

[12] Productivity was measured at 5 m at all grid stationsand at 7 depths for all PnB stations on each cruise (exceptcruise 14 in fall 2005 when productivity was not measured)using 14C-bicarbonate tracer according to the methodsdescribed by Anderson et al. [2006]. Briefly, two acid-cleaned 250 mL polycarbonate bottles were filled with waterfrom 5 m at each station. One bottle from each pair that hadbeen wrapped in electrical tape served to measure dark 14Cuptake. Each bottle received �10 mCi of 14C bicarbonate (in


3 of 17

�3 mM NaHCO3). Sampling the grid took approximately36 h so each sample pair was incubated for 24 h to provideeach with a full diurnal light cycle. Light bottles were placedin a deck incubator screened to 50% of incident light withneutral density screens. Dark bottles placed in an opaqueplastic incubator on deck. Surface seawater was circulatedthrough all incubators to maintain temperature.[13] Sampling for productivity measurements along the

PnB transect began approximately an hour after dawn withall seven stations sampled before sunset so that light depthscould be determined at each station using a photosyntheti-cally active radiation (PAR) sensor mounted on the CTD/rosette. Sampling depths at each station were chosen tocorrespond to 100, 54, 35, 16, 7, 3.6 and 1.7% of theintensity of PAR just below the sea surface. Light depths oneach profile were determined using extinction coefficientscalculated from linear regressions of the log of light intensityversus depth. Changes in light intensity from clouds weredetected using a mast-mounted PAR sensor. SubsurfacePAR measurements influenced by clouds were not used inthe regression analyses for the determination of lightextinction coefficients. The samples were incubated andprocessed as described for the sampling grid except thatneutral density screens were used to simulate the light levelsat each collection depth. Incubations lasted 24 h to eliminatebias in the light exposure among incubated samples to day-light caused by sampling stations throughout the day.[14] At the end of incubations total radioactivity in each

sample was determined by adding 100 mL of incubatedseawater to 100 mL of b-phenethylamine in a 7 mL glassscintillation vial followed by 5 mL of Ultima Gold XRscintillation cocktail. Each vial was shaken vigorously for30 s and 14C activity assayed aboard ship after bubbles hadcleared and chemoluminescence had subsided (�2 h) using aBeckman 5801 liquid scintillation counter using an internalquench curve. The remaining volume of each sample wasfiltered through a 25 mm GF/F filter. Filters were placed inindividual 20 mL glass scintillation vials and 0.25 mL of0.5 N HCl pipette onto each filter in a fume hood to drive offexcess tracer. Vials were left uncapped for a minimum of7 h. Then 10 mL of Ultima Gold XR scintillation cocktailwas added to each vial, the vials capped and shaken andthe filters allowed to clear (2–3 weeks). 14C activity wasanalyzed ashore by scintillation counting on a Beckman6500 scintillation counter using an internal quench curve.Primary productivity rates, P, in both the light and darkincubations were calculated as:

P ¼ DPMP=DPMTot∗ DIC½ �∗1:05=t ð1Þ

where DPMP is the14C activity (disintegrations per minute)

per liter in particulate matter, DPMTot is the total14C activity

per liter in the sample, [DIC] is the concentration of inor-ganic carbon in seawater (25 mg C L�1), 1.05 is a correctionfactor to account for the preferential uptake of 12C over 14Cand t is the duration of the incubation (�24 h). Daily pri-mary production (g C m�3 d�1) was calculated as the dif-ference in productivity between light and dark bottles.Productivity-to-biomass ratio was calculated as the dailyprimary productivity (g C m�3 d�1) per unit chlorophyll a(g chl a m�3) with units of gC (g chl a)�1 d�1.

[15] Productivity measurements were not routinely repli-cated. Duplicate measurements made occasionally on eachcruise generally agreed to within 10%. Integrated rates ofproductivity from the PnB stations were calculated using thetrapezoidal method.

2.4. EOF Analysis of Productivity

[16] We used empirical orthogonal function (EOF) analy-sis as described by Emery and Thompson [2001] to identifydominant spatial and temporal patterns of 5 m productivityamong cruises. In the procedure, measured 5 m productivityvalues pj(n) at each station j (out of M = 25 stations) oncruise n were normalized as,

pj′ nð Þ ¼ pj nð Þ � Pjh i

=sj ð2Þ

where Pj and sj are the mean and standard deviation at stationj over N = 15 cruises for which productivity was measured.Changes in productivity were separated into modes as,

pj′ nð Þ ¼XMi

ai nð Þ8ij ð3Þ

where ai(n) is the amplitude function of mode i for cruise n;8ij is spatial mode i at station j; summation is over all Mmodes. Spatial modes are orthogonal such that ∑M

i 8ij 8kj =1 if i = k and it equals 0 if i ≠ k; summation is over allM stations. Amplitude functions ai(n) are uncorrelated suchthat ∑N

i ai(n) ak(n) = 0 if i ≠ k and it equals the variance ofmode i, si

2, if i = k; summation is over all N cruises.

2.5. Current Measurements

[17] Currents at depth were measured with shipboardADCPs operating at 150 and 300 kHz (ADCPs were man-ufactured by Teledyne RD Instruments, San Diego, CA).The depth range of current measurements extended from thefirst ADCP bin at 15–20 m depth to the deepest bin between110 and 370 m depth depending on the ADCP frequencyand configuration. Currents were consistently measured inthe upper 120 m except during cruise 10 in May 2004 whenno ADCP data were available. The ADCPs used bottomtracking in shallow water depths.[18] Currents at the surface (�1 m depth) were measured

using high-frequency (HF) radars located at Point Argello,Point Conception, Refugio, Coal Oil Point, Summerland,and the Mandalay Generating Station (Figure 1b). Not allHF radars were operating during all cruises. Surface currentvectors were derived from hourly radial currents measuredby the HF radars and then interpolated onto a 2 km grid.Current vectors were averages over circles 3 km in radiusfrom each grid point. Additional details on the HF radarnetwork and its performance are given by Emery et al.[2004].

2.6. Estimating Flow Rotation

[19] Flow rotation was quantified by relative vorticity as,

z ¼ ∂V=∂x� ∂U=∂y ≈ DV=Dx�DU=Dy ð4Þ

where x was positive eastward, y was positive northward,


4 of 17

Tab

le1.

MeanPropertiesObservedon

EachCruisea

Cruise

Dates

0–10

m5m

Nitrate

(mmol

L�1)

5m

chla

(ugL�1)

R chla

(mgm

�2)

5m

Produ

ctivity

(mgCm

�3d�

1)

5m

PB

(gC(g

Chl

a)�1d�

1)

R prod

(gCm

�2d�

1)

Start

End

Tem

perature

(°C)

Salinity

(psu)

119

Mar

2001

26Mar

2001

13.71�

0.45

33.46�

0.05

1.9�

1.5

1.7�

1.0

67�

3595

�51

60�

202.0�

1.1

22Sep

2001

14Sep

2001

17.31�

1.12

33.62�

0.04

0.6�

1.1

0.8�

0.5

73�

3254

�34

69�

211.1�

0.20

319

Feb

2002

26Feb

2002

13.55�

0.38

33.60�

0.02

0.4�

0.8

2.0�

1.0

124�

4375

�39

39�

121.3�

0.6

425

Apr

2002

2May

2002

12.65�

0.82

33.79�

0.04

4.7�

5.2

6.9�

5.2

177�

9926

1�

190

51�

262.2�

1.1

54Sep

2002

11Sep

2002

14.66�

1.19

33.56�

0.02

3.5�

2.4

1.8�

0.8

65�

2616

5�

132

86�

301.7�

0.5

626

Feb

2003

4Mar

2003

13.93�

0.35

33.32�

0.02

1.3�

1.1

2.0�

1.0

84�

4610

8�

5056

�11

1.6�

0.8

715

May

2003

22May

2003

12.03�

0.86

33.81�

0.06

7.9�

5.4

14.1

�8.4

302�

115

516�

254

46�

365.8�

2.4

89Oct

2003

16Oct

2003

16.80�

1.18

33.31�

0.02

0.6�

0.8

2.4�

2.2

54�

1611

3�

8954

�13

1.6�

0.7

925

Feb

2004

3Mar

2004

12.67�

0.40

33.19�

0.03

2.2�

0.9

4.7�

2.4

194�

9015

6�

7233

�7

1.9�

0.9

106May

2004

12May

2004

13.31�

1.16

33.56�

0.08

6.4�

4.2

3.6�

2.0

124�

6622

3�

108

74�

452.2�

0.8

119Sep

2004

15Sep

2004

20.11�

1.04

33.42�

0.03

0.3�

0.2

0.7�

0.5

71�

3760

�31

96�

421.1�

0.4

1220

Jan20

0527

Jan20

0514

.85�

0.41

32.98�

0.11

0.4�

0.4

1.2�

0.6

60�

2247

�29

38�

100.4�

0.2

1323

Apr

2005

30Apr

2005

12.85�

0.51

33.45�

0.09

3.8�

4.2

6.6�

4.4

328�

209

330�

280

49�

213.9�

1.9

147Oct

2005

18Oct

2005

14.06�

0.98

33.39�

0.02

3.7�

2.9

3.4�

1.4

87�

34–b

––

154Feb

2006

6Feb

2006

12.79�

0.76

33.50�

0.06

5.9�

3.5

3.6�

1.5

90�

3813

8�

5140

�8

1.2�

0.2

1629

Apr

2006

3May

2006

12.81�

0.81

33.53�

0.10

4.5�

3.8

6.1�

2.7

152�

4233

8�

131

65�

142.7�

0.5

Springseason

almean

12.89�

0.56

33.60�

0.16

4.9�

2.1

6.5�

4.2

192�

103

294�

140

56�

113.1�

1.5

Fallseason

almean

16.58�

2.40

33.46�

0.13

1.7�

1.7

1.8�

1.1

70�

1298

�52

76�

181.4�

0.3

Winterseason

almean

13.56�

0.89

33.32�

0.25

2.4�

2.3

2.7�

1.4

110�

5210

5�

4541

�9

1.3�

0.6

a Tem

perature,salinity

,5mnitrate,5mchloroph

ylla

,integratedchloroph

ylla

,and

5m

prod

uctiv

itymeasurementsarefrom

the25

gridstations.Integratedprod

uctiv

itydataareaverages

from

theseven-stationPnB

transect.

bPrimaryprod

uctiv

itywas

notmeasuredon

cruise

14.


5 of 17

and U, V were velocity components positive eastward andnorthward, respectively. Estimates of z, DV/Dx and�DU/Dy were normalized for display by the Coriolisparameter f ( = 8.196 � 10�5 s�1 at 34.3°N) for comparisonwith Earth’s rotation rate; positive values were counter-clockwise (cyclonic). Patterns of surface currents from theHF radars showed that cyclonic eddies were present in thewestern SBC when DV/Dx and �DU/Dy were each bothlarge and positive.[20] Components of relative vorticity at depth zd,DUd/Dy

and DVd/Dx, were estimated from the ADCP data. Ud/Dywas computed as,

DUd=Dy ¼ UnC � UsCð Þ=Dy1 þ UnD � UsDð Þ=Dy1½ �=2 ð5Þ

where UnC and UnD were the spatially averaged eastward

components of velocity over the northern sections of lines Cand D between 34.3 and 34.4°N (thick black bars onnorthern parts of lines C and D; see Figure 1b) at 20–120 mdepth. UsC and UsD were similarly averaged over thesouthern sections of lines C and D, but between 34.1 and34.2°N (thick black bars on southern parts of lines C and D;see Figure 1b). Dy1 was the distance between the centers ofthe sections along the C and D lines (22.2 km).DVd/Dx wascomputed as,

DVd=Dx ¼ VD � VBð Þ=Dx1 þ VE � VBð Þ=Dx2½ �=2 ð6Þ

where VB, VD, and VE were the spatially averaged north-ward components of velocity over the midsections of linesB, D, and E between 34.2 and 34.3°N (thick gray bars incentral parts of lines B, D, and E; see Figure 1b) and

Figure 2. Mean and seasonal distribution of (a–d) primary productivity, (e–h) chlorophyll a concentra-tion, and (i–l) the productivity to biomass ratio. The magnitude of each parameter is proportional to thediameter of the circles as scaled by the labeled circle at bottom left.


6 of 17

between 20 and 120 m depth. Dx1 was the distance betweenlines B and D along 34.26° (27.4 km) and Dx2 was thedistance between lines B and E along 34.26° (36.4 km).When data from multiple surveys were available for a cruise,multiple values of UnC, UsC, UnD, UsD, VE, and VB wereaveraged together and used in equations (5) and (6).[21] Components of relative vorticity at the surface zs,

DUs/Dy and DVs/Dx, were computed from the HF radarsusing grid points near the survey lines to be consistent withestimates from equations (5) and (6). DUs/Dy was com-puted as,

DUs=Dy ¼ UnCD � UsCDð Þ=Dy2 ð7Þ

where UnCD was the eastward surface velocity averaged overgrid points in the black rectangle at the north end of lines Cand D in Figure 1b and UsCD was similar, but averaged overthe black rectangle at the south end of lines C and D. Dy2was the distance between the centers of the black rectangles(22.7 km). DVs/Dx was computed as,

DVs=Dx ¼ VeB � VwEð Þ=Dx3 ð8Þ

where VeB was the northward surface velocity averaged overthe gray rectangle just east of the center of line B and VwE

was similar, but averaged over the gray rectangle just west ofline E. Dx3 was the distance between the centers of the grayrectangles (25.7 km). Grid points in the gray rectangles were

the closest to lines B and E with consistent HF radar datacoverage during all cruises.[22] Tidal currents were aliased into the velocities

measured by the ADCPs but, as shown below, estimatesDVd/Dx, DUd/Dy and DVs/Dx, DUs/Dy were consistentindicating tidal aliasing effects were not severe. Tidal cur-rent speeds measured by the HF radars in the study areawere 0.1 m s�1 or less and were typically smaller than theADCP velocity estimates used in equations (5) and (6).

2.7. Meteorological and Satellite Observations

[23] Wind velocity data were obtained from NOAA DataBuoy Center (NDBC) for buoys 46011, 46023, 46053, and46054; Figures 1b and 5b show the locations of these buoys.Wind stress t was estimated using algorithms described bySmith [1988]. Following previous studies [e.g., Harms andWinant, 1998; Dorman and Winant, 2000; Cudaback et al.,2005; Melton et al., 2009], winds were rotated into prin-cipal axis coordinates and we use the term “upwelling-favorable” to denote winds that blow equatorward alongthe principal axis directions at the buoys.[24] Satellite ocean color images from the Sea-viewing

Wide Field-of-view Sensor (SeaWiFS) were processed forchlorophyll using the algorithms of McClain et al. [2004].Satellite sea surface temperature images were obtained fromthe Advanced Very High Resolution Radiometer (AVHRR)using the NOAA Comprehensive Large Array-Stewardship

Figure 3. (top) Contour plots of integrated chlorophyll a concentrations and (bottom) integrated primaryproductivity at the seven Plumes and Blooms stations between 2001 and 2006. Each year begins at theyear label along the x axis.


7 of 17

System (CLASS) and the Nonlinear Sea Surface Tempera-ture (NLSST) algorithm [McClain et al., 1985, 2004].

3. Results

3.1. Mean and Seasonal Productivity Patterns

[25] A summary of the physical, chemical and biologicalconditions averaged over the 25 grid surveys is presented inTable 1. Seasonal changes in the average temperature andsalinity in the upper 10 m reflected the annual pattern ofstratification and upwelling in the SBC with relatively cooland fresh waters present during winter, colder more salinewaters associated with upwelling present during spring, andwaters of intermediate salinity and warmer, but more vari-able, temperature present during fall reflecting varying levelsof stratification and surface warming (Table 1). Seasonalaverages of productivity, chlorophyll a, and nitrate followedthe known annual cycle of productivity, phytoplanktonbiomass, and nutrient concentrations. Maxima in 5 m chlo-rophyll a concentration, integrated chlorophyll a concentra-tion (to 75 m), and productivity occurred during spring when

nitrate was most abundant, and lower biomass and produc-tivity occurred during the fall and winter when nitrate wasless abundant (see Table 1 and Figure 2).[26] Chlorophyll a concentrations and productivity mea-

sured at 5 m showed local maxima along line D duringspring and fall (Figures 2b–2d and 2f–2h). Mean produc-tivity along line D during spring was 403 mg C m�3 d�1

compared with a mean of 166 mg C m�3 d�1 during fall asexpected on the basis of the annual pattern of upwelling-favorable winds. Productivity-to-biomass (PB) ratios did notshow an enhancement over the Santa Barbara Basin onaverage or seasonally (Figures 2i–2l). PB values werehighest during fall (see Table 1 and Figures 2j–2l) consistentwith the known tendency of PB to increase in warmer watersand when nanoflagellates are more dominant, both of whichare characteristic of the SBC during fall [Anderson et al.,2008].[27] Vertical profiles of productivity along the PnB tran-

sect allowed assessment of integrated productivity whichwas not possible for the grid stations where productivity wasmeasured at 5 m only. Integrated productivity generally

Figure 4. First three spatial modes of primary productivity at 5 m: (a) 81j, (b) 82j, and (c) 83j where j = 1–25are the station numbers as indicated in Figure 1a. Diameters of the circles give the relative magnitude ofthe spatial modes. Positive values are indicated by gray shading, and negative values are indicated byblack shading. The percent of variance explained by each mode is given at top. (d) Corresponding ampli-tude functions for each cruise a1(n), a2(n), and a3(n), as indicated by the legend and where n = 1–16 arethe cruise numbers; S, F, and W denote spring, fall, and winter. No productivity data were available forcruise 14.


8 of 17

Figure 5. (a) Primary productivity at 5 m depth during the conductivity-temperature-depth (CTD) gridsurvey of cruise 7. The magnitude of productivity is proportional to the diameters of the circles as scaledby the labeled circle at bottom left. (b) Wind velocities averaged during the grid survey (solid arrows) andaveraged over the 2 weeks before the grid survey (dashed arrows). Scales indicate 5 m s�1 wind speed.NDBC buoys 46011, 46023, 46054, and 46053 are indicated by open circles. Winds at land-based mete-orological stations are also indicated. (c) Time series of principal axis wind stress at NDBC buoys 46053,46054, and 46023 as indicated by the legend. Principal axis directions at each buoy are approximately par-allel to the wind vectors in Figure 5b. Negative wind stress is upwelling favorable and toward the east orsoutheast. Vertical lines indicate the time of the grid survey.

Figure 6. As in Figure 5 but for cruise 12.


9 of 17

followed the spatial and seasonal pattern of integratedchlorophyll a (Figure 3); consistent with this observation,5 m productivity correlated strongly with 5 m chlorophyllduring the grid surveys (r2 = 0.87, p < 0.02, N = 15). Inte-grated chlorophyll a concentrations (to the 1.7% light depth)showed blooms of varying intensity and spatial extentoccurring each spring (Figure 3, top). The largest bloomswere nearly contiguous across the SBC with integratedchlorophyll a concentrations between 200 and 300 mg chla m�2 consistent in magnitude with the integrated chloro-phyll values observed at the more widely distributed gridstations (Table 1). An exceptionally strong bloom wasapparent during spring 2003 (cruise 7) when integratedproductivity values up to 8.5 gC m�2 d�1 were measured atindividual stations with an overall mean for the seven PnBstations of 5.8 gC m�2 d�1 (see Table 1 and Figure 3,bottom). Smaller high-productivity events confined tothe mainland (2002) or islands (2001, 2004) were also

evident (Figure 3, bottom). Mean integrated productivity foreach season, averaged across the PnB stations, was nearlyequal during winter and fall with a winter mean of 1.3 �0.6 gC m�2 d�1 and a fall mean of 1.4 � 0.3 gC m�2 d�1

(Table 1). The spring mean was higher 3.1 � 1.5 gC m�2

d�1 with a larger standard deviation reflecting the variableintensity of the blooms encountered in that season thatranged from a maximum of 5.8 gC m�2 d�1 during springof 2003 (cruise 7) to a minimum of 2.0 gC m�2 d�1 duringspring 2002 (cruise 4; see Table 1).

3.2. Upwelling and Productivity

[28] The first three EOF modes explained 89% of the var-iance in 5 m productivity (Figures 4a–4c). Mode 1 accoun-ted for 64% of the variance in 5 m productivity. Values of81j had the same sign at all stations (Figure 4a) meaning thatmode 1 described variations in productivity that changed inunison at all stations. The pattern of 81j (Figure 4a) resem-bled the pattern of mean productivity from all cruises(Figure 2a) in that except for cruise 1, a1 was positive duringspring and negative during fall and winter (Figure 4d) somode 1 approximately reproduced the seasonal cycle inproductivity shown in Figures 2b–2d. The correlationbetween the average productivity on each cruise and a1 wasvery strong (r2 = 0.99, data not shown).[29] The maximum in a1 corresponded to the intense phy-

toplankton bloom during cruise 7 (grid survey: 16–18 May2003) with 5 m productivity exceeding 1000 mg C m�3 d�1

at some stations (Figure 5a); 5 m productivity averaged overall stations during the cruise was 516 mg C m�3 d�1

(Table 1) The highest values of integrated primary produc-tion on the PnB line were also measured on this cruise.Wind stresses were strongly upwelling favorable, withmagnitudes exceeding 0.3 Pa at times, before and duringcruise 7 (Figures 5b and 5c). In contrast, the largest nega-tive value of a1 corresponded to cruise 12 (grid survey: 24–25 January 2005) when productivity was very low at allstations on the survey grid (Figure 6a); 5 m productivityaveraged only 47 mg C m�3 d�1 for the cruise. Windstresses at NDBC buoys 46023 and 46053 were weak andvariable before cruise 12 and were generally downwellingfavorable during the cruise (Figures 6b and 6c); Wind stressdata at 46054 data were not available during cruise 12.[30] Two possible drivers for variability in a1 were

examined, wind stress and nutrient concentration. Despitethe apparent connection between wind stress t and produc-tivity suggested in Figures 5 and 6, no consistent relation-ship was found between t and a1. This was examined byseparately averaging t at NDBC buoys 46023, 46053, and

Figure 7. First EOF amplitude function a1 versus nitrateconcentration at 5 m for 15 cruises. Symbols denote cruisesconducted during spring (open circles), fall (squares), andwinter (triangles). Numerals denote cruise numbers for adja-cent symbols. Line shows model II regression (reducedmajor axis) of the relationship between nitrate concentrationat 5 m and the time series amplitude function for mode 1, a1.

Table 2. Wind Forcing and Near-Surface Oceanographic Conditions and Water Properties for Selected Cruise Pairs

Cruise Season t46054 (Pa) Na (cph)ML

Depthb (m) zs /fc

[NO3]5 m

(mmol L�1)T0–10 m

(°C)a1

(mg C m�3 d�1)

7 spring �0.26 � 0.07 6.1 � 2.5 8.9 � 8.6 0.33 � 0.16 7.9 � 5.4 12.03 � 0.86 193111 fall �0.25 � 0.11 11.8 � 1.8 7.2 � 5.0 0.25 � 0.11 0.3 � 0.2 20.11 � 1.04 �6451 spring �0.044 � 0.026 6.0 � 2.2 10.4 � 8.2 0.17 � 0.01 1.9 � 1.5 13.71 � 0.45 �44116 spring �0.029 � 0.036 7.3 � 1.7 3.9 � 3.0 0.16 � 0.04 4.5 � 3.8 12.81 � 0.81 9426 winter �0.15 � 0.05 4.4 � 2.2 18.7 � 8.9 0.19 � 0.08 1.3 � 1.1 13.93 � 0.35 �40815 winter �0.18 � 0.13 5.9 � 2.5 10.9 � 11.6 0.14 � 0.06 5.9 � 3.5 12.79 � 0.76 �250

aBuoyancy frequency N in upper 30 m.bMixed layer (ML) depth where temperature drops by 0.10°C from its surface value.cSurface relative vorticity zs from HF radars as described in text.


10 of 17

46054 over a wide range of time intervals before and duringthe grid surveys and computing the correlation r2 betweenaverage t and a1 over the 15 cruises when a1 was available.No significant correlation (p < 0.05) with a1 was found withany averaging interval for wind stress from any of the threebuoys. The mean nitrate concentration at 5 m [NO3]5 m oneach cruise was significantly correlated with a1 (Figure 7,r2 = 0.63, N = 15, p = 0.0004). Except for cruise 1, highvalues of a1 and [NO3]5 m occurred in spring and, except forcruise 5, low values occurred in fall. Winter concentrationswere generally intermediate between the spring and fallobservations.[31] We hypothesized that the lack of correlation between

wind stress and a1 resulted from upwelling of source waterswith variable nutrient concentrations. We examined thishypothesis by comparing [NO3]5 m, near-surface tempera-ture T0–10 m, and a1 for pairs of cruises with similar,upwelling-favorable wind stress. In doing so, we assumedthat the variations in near-surface temperature and nitrate

concentrations reflected variations in upwelled sourcewaters. During spring cruise 7 and fall cruise 11, wind stressat buoy 46054 t46054 was similar, but [NO3]5 m and a1during cruise 7 were much higher while T0–10 m was muchlower (Table 2). Higher [NO3]5 m and a1 and lower T0–10 m

occurred during spring cruise 16 compared with springcruise 1 even though t46054 was higher (but weak) duringcruise 1. During winter cruises 6 and 15, t46054 was similar,but [NO3]5 m and a1 were higher and T0–10 m was lowerduring cruise 15. Cruise pairs with similar t46023 and t46053(data not shown) also exhibited a wide range of [NO3]5 m,T0–10 m, and a1.[32] Other factors controlling the response of near-surface

waters to upwelling-favorable wind stress, including mixedlayer depth, buoyancy frequency N, and z /f, were compa-rable between cruise pairs (Table 2). Exceptions were higherN in cruise 11 compared with cruise 7 and deeper mixedlayers during cruise 1 compared with cruise 16 and duringcruise 6 compared with cruise 15.

Figure 8. Spatial patterns of 5 m productivity for (a) cruise 5, (b) cruise 13, and (c) cruise 7. The mag-nitude of productivity is proportional to the diameters of the circles as scaled by the labeled circle at bottomleft. Satellite-derived chlorophyll distributions during (d) cruise 5 at 20:38 UTC on 7 September 2002,(e) cruise 13 at 21:00 UTC on 25 April 2005, and (f) cruise 7 at 20:02 UTC on 18 May 2003 are shown.


11 of 17

3.3. Cyclonic Flow and Productivity

[33] Mode 2 explained 17% of the variance in productiv-ity. The spatial pattern of 82j (Figure 4b) shows that as a2varied, productivity on line D, at station 8 on line C, and atstation 17 on line F, changed oppositely relative to the otherstations. a2 was near zero during fall and negative duringspring (Figure 4d) except for the largest a2 values duringfall cruise 5 (grid survey: 7–9 September 2002) and springcruise 13 (grid survey: 25–27 April 2005).[34] The large value of a2 during cruise 5 with enhanced

productivity on line D (Figure 8a) was observed after(DVs/Dx)/f and (�DUs/Dy)/f had increased from aroundzero on 23 August to more than 0.2 on 7 September whenmeasurements on line D were made (Figure 9a). zs /fincreased from around zero to more than 0.5 over thisperiod, consistent with strengthening cyclonic flow. Thepattern of surface currents averaged over 5–6 Septemberillustrates the cyclonic flow (Figure 9b). The simultaneousoccurrence of high, localized productivity on line D andstrong cyclonic flow suggests a relationship between thetwo.[35] To examine this relationship, flow rotation was

quantified at the surface by components DVs/Dx and�DUs/Dy for each cruise (Figure 10a) and at depth by

components DVd/Dx and �DUd/Dy (Figure 10b). Com-ponents were generally lower for winter cruises comparedwith summer and fall cruises with points scattering widelyabout the regression lines. Components were positive for allcruises except during winter cruises 3 and 12 when DVs/Dxand DVd/Dx were negative. Over all cruises, �DU/Dy andDV/Dx were significantly correlated at the surface and atdepth. Cyclonic flow was asymmetric in the sense that�DU/Dy > DV/Dx with all but a few points falling belowthe one-to-one lines in Figures 10a and 10b. This asymmetryis also indicated by the observation that alongshore flowspeeds near the islands and mainland comprising �DU/Dyexceeded cross-channel flow speeds comprising DV/Dx(data not shown). Cyclonic flow was strongest during cruises5, 7, 8, and 13 in the sense that these had the largest compo-nents simultaneously with (�DU/Dy)/f > 0.2 and (DV/Dx)/f >0.05; these criteria were met at depth for cruise 14, but not atthe surface.[36] For cruises 5, 13, and 7 when cyclonic flow was

strong, productivity was high on line D. Productivity at theother stations in the channel exhibited a progression fromlow levels during fall cruise 5 (Figure 8a), to higher levelsduring spring cruise 13 (Figure 8b), and much higher levelsduring spring cruise 7 (Figure 8c). This progression isquantified by the increasing values of a1 over cruises 5, 13,

Figure 9. (a) Time series of relative vorticity zs at the surface and its componentsDVs/Dx and�DUs/Dynormalized by the Coriolis parameter f during 23 August through 16 September 2002. Details of thevorticity calculations are explained in text. Horizontal gray bar on the zero line indicates duration ofcruise 5, with the black square corresponding to the timing of productivity measurements along line D.(b) Surface currents averaged over 24 h beginning at 18:00 UTC on 5 September 2001. Scale indicates0.6 m s�1 surface current speed.


12 of 17

and 7 (Figure 4d). Satellite chlorophyll images around thetimes of measurements on line D during cruises 5 and 13(Figures 8d and 8e) showed circular maxima centeredaround line D. High chlorophyll levels throughout the

channel obscured any local maximum along line D duringcruise 7 (Figure 8f).[37] During cruise 8 (grid survey: 11–13 October 2003),

productivity was enhanced on the southern half of line D(Figure 11a) and at stations 4, 5 and 9 when a cyclonic flowpattern was evident (Figure 11b). Satellite surface chloro-phyll and surface temperature distributions (Figures 11cand 11d) during ship-based productivity measurements online D were consistent with advection of surface watersfrom south and west of Point Conception into the SBC bythe cyclonic flow. Entrained waters with high chlorophylland lower temperatures were clearly evident along the northsides of the San Miguel and Santa Rosa Islands and thewestern tip of Santa Cruz Island (Figures 11c and 11d)consistent with the higher productivity on the southernportion of line D.

3.4. Alongshore Productivity Gradients

[38] Mode 3 accounted for 8% of the variance in produc-tivity and described alongshore productivity gradients nearthe mainland coast (Figure 4c). 83j at stations 9 and 14 nearthe mainland coast in the western channel had opposite signscompared with stations 20, 24, and 25 in the eastern channel.a3 attained its highest positive value during cruise 4 (gridsurvey: 28–30 April 2002) when a large east-to-west pro-ductivity gradient occurred across stations 20, 24 and 25 andstations 9 and 14 (Figure 12a). This gradient coincidedwith a developing bloom in the eastern channel (Figures 12band 12c). The opposite trend with a strong west-to-eastgradient occurred during cruise 7 (Figure 8c) when a3attained its maximum negative value. a3 was also negativeduring cruise 8 (Figure 11a) corresponding to a gradientacross stations 9, 14, 20, and 25. a3 exceeded a1 and a2 inmagnitude only during cruise 4 (Figure 4d) so its con-tributions are difficult to see for other cruises.[39] Although mode 3 described only a small fraction of

the total productivity variance, it accounted for more vari-ance at stations near the mainland. Six stations had >8%of their variance explained by mode 3: stations 9 (9%),14 (13%), 20 (13%), 21(13%), 24 (46%) and 25 (25%).Except for station 21 all of these were located in <50 mwater depth along the mainland coast (Figure 1b) with sta-tions 24 and 25 also being near the mouths of the SantaClara and Ventura Rivers. Shelf width may also affectmode 3 since the along-channel gradient it describes spansthe transition between the wide and narrow shelves along theSBC (Figure 1b).

4. Discussion

[40] Our measurements of productivity confirm pastobservations that the SBC is a highly productive region inthe Southern California Bight. Shipe and Brzezinski [2003]measured integrated primary production in the SBC duringthe onset of El Niño in 1997 at station 4 on the PnB transectto range from 0.72 to 1.9 g C m�2 d�1, similar to the averagevalue of 1.6 g C m�2 d�1 reported by Hardy [1993]. Ourseasonal averages across the seven PnB stations are withinthe range of these past observations with values of 1.3, 3.3and 1.4 g C m�2 d�1 during winter, spring and fall, respec-tively (Table 1). Our spring maximum equals or exceedsvalues reported from other regions in the Southern California

Figure 10. (a) Cruise-averaged components of relative vor-ticity zs at the surface DVs/Dx and �DUs/Dy normalizedby the Coriolis parameter f. DVs/Dx and �DUs/Dy werecomputed from surface currents measured by the HF radars.Crossed lines through symbols indicate �1 standard devia-tion during each cruise. (b) Similar to Figure 10a but fordepth-averaged components DVd/Dx and �DUd/Dy asdetermined from the shipboard ADCPs. Crossed lines indicaterange of two measurements of DVd/Dx and �DUd/Dy.Details of the vorticity calculations are explained in text.In Figures 10a and 10b, triangles indicate winter cruises,circles indicate spring cruises, and squares indicate fallcruises. Cruise numbers are given near each symbol.Dashed lines have slopes of 1 and y intercepts of zero. Solidlines are linear least square fits, and regression statistics aregiven at bottom right. Vertical and horizontal lines spanningx and y axes correspond to thresholds (DV/Dx)/f = 0.05 and(�DU/Dy)/f = 0.2, respectively.


13 of 17

Bight [Eppley et al., 1979]. Maximum productivity occurredduring spring 2003 when integrated productivity reached8.5 g C m�2 d�1 at some stations on the PnB transect and 5 mproductivity exceeded 1000 mg Cm�3 d�1 at several stationson the grid survey.[41] Previous studies in the SBC [e.g., Shipe and

Brzezinski, 2003; Mantyla et al., 1995; McPhee-Shaw et al.,2007; Fram et al., 2008] have revealed a seasonal cycle inproductivity driven by a seasonal cycle in nutrient con-centrations in the upper ocean. These cycles were evident inour study by the contrast between channel-wide, high pro-ductivity and high near-surface nutrient concentrations inspring, such as during cruise 7 (see Figure 5 and Table 1),and much lower productivity and lower near-surface nutri-ent concentrations in fall and winter, such as during wintercruise 12 (Figure 6). We conclude that near-surface nutri-ents resulted from upward transport of deeper source watersby wind-driven upwelling. The following observationssupport this conclusion:[42] 1. Within each season, higher near-surface nitrate

concentrations generally corresponded with lower near-surface temperatures. Only during one cruise (cruise 9),when it rained heavily, were nutrient inputs from runoff alikely source.[43] 2. During all but one cruise (cruise 12) wind stress at

the buoys shown in Figures 5 and 6 was upwelling-favorable.

[44] Our observations indicate that when stronglyupwelling-favorable wind stress occurs in fall and winter,the resulting upwelling usually delivers low nutrient con-centrations to surface waters with consequent low produc-tivity. In contrast, when strongly upwelling-favorable stressoccurs in spring, upwelled source waters usually containhigh nutrient concentrations and drive higher productivity.Therefore, wind stress alone is insufficient for predictingproductivity across seasons in the SBC. It is the combinationof upwelling-favorable wind stress and source waters withhigh nutrient concentrations that leads to high productivity.We interpret EOF mode 1 as representing nutrient deliveryto near-surface waters by wind-driven coastal upwelling.These inputs drive the seasonal productivity cycle includingthe spring productivity maximum.[45] Spatial patterns in productivity in the SBC resembled

spatial patterns in the long-term chlorophyll climatologyfrom satellite ocean color imagery [e.g., Halpern et al.,2006; Otero and Siegel, 2004] (see Figure 1a). This is con-sistent with our observations showing that 5 m productivityand 5 m chlorophyll a concentrations averaged over the gridsurveys were strongly correlated (Table 1; r2 = 0.87, p <0.02, N = 15). The maximum in satellite-derived chlorophyllconcentration centered over the Santa Barbara Basin(Figure 1a) was consistent with our measurements of meanproductivity (Figure 2a) and mean chlorophyll a concentra-tion (Figure 2e). This maximum appeared in our grid

Figure 11. (a) Five meter productivity for cruise 8. The magnitude of productivity is proportional tothe diameters of the circles as scaled by the labeled circle at bottom left. (b) Surface currents during01:00–20:00 UTC on 12 October 2003. (c) Satellite-derived chlorophyll concentration at 20:06 UTCon 12 October 2003. (d) Satellite sea surface temperature image at 18:25 UTC on 12 October 2003.


14 of 17

surveys as enhanced productivity and chlorophyll on line Dduring spring (Figures 2c and 2g) and fall (Figures 2dand 2h).[46] The enhanced productivity and chlorophyll a con-

centration over the Santa Barbara Basin suggest stimulationof productivity by the cyclonic eddies frequently occurringthere. Evidence for this is the simultaneous occurrence ofenhanced productivity on line D during cruises 5 and 13(Figures 8a and 8b) and high, positive values of DV/Dx and�DU/Dy (Figure 10). We interpret Figure 8 as showingeddy-enhanced productivity superimposed on progressivelyhigher levels of channel-wide productivity caused by wind-driven nutrient upwelling. The eddy enhancement asdescribed by a2 (Figure 4b) was apparent when channel-wideproductivity was weak (Figure 8a) or moderate (Figure 8b),but was obscured when channel-wide productivity wasstrong (Figure 8c) even though DV/Dx and �DU/Dy werehigh (Figure 10).[47] One mechanism for eddy enhancement of productiv-

ity is uplift of isopycnal surfaces with elevated nutrientconcentrations into the euphotic zone during spin-up ofcyclonic eddies [e.g., McGillicuddy et al., 2007, 1998;Benitez-Nelson et al., 2007]. Uplift of isopycnal surfaces inthe center cyclonic eddies in the western SBC has beenreported by Nishimoto and Washburn [2002]. It was evidentin our grid and towed surveys when cyclonic rotation wasstrong. For example, towed surveys during cruise 5 showeduplift of isopycnals of �30 m in the eddy along lines B, C,and D (data not shown). The increasing vorticity beforeand during the cruise (Figure 9a) implies that isopycnaluplift was recent which would account for the high [NO3

- ]of 4–12 mM observed in the euphotic zone during cruise 5

along line D (data not shown) compared with much lowervalues over the rest of the channel (Table 1).[48] A second mechanism for eddy enhancement of pro-

ductivity is convergence which would concentrate and retainphytoplankton cells within the flow. Beckenbach andWashburn [2004] observed convergence and inward radialflow in cyclonic eddies in the western SBC on the basis ofHF radar-derived surface currents. More recently, Kim et al.[2011] reported surface convergence within cyclonic sub-mesoscale eddies off San Diego, California also derivedfrom HF radar observations. In both studies, however, scat-ter in convergence estimates was large. Numerical modelingstudies [e.g., Capet et al., 2008a, 2008b] indicate that con-vergence and strong vertical velocities are associated withfrontal regions with short cross-frontal scales. These scalesmay be poorly resolved by the HF radars and the role ofconvergence in regulating productivity in cyclonic eddiesremains uncertain.[49] A third mechanism for eddy enhancement of pro-

ductivity is advection of phytoplankton and nutrients intothe western SBC from the upwelling center between PointConception and Point Arguello. This apparently occurredduring cruise 8 when a plume of cooler water with highchlorophyll extended southward from Point Conception anda band of high chlorophyll curved around the cyclonic flownorth of Santa Cruz Island (Figure 11). Both DV/Dx and�DU/Dy were large and positive then (Figure 10) andproductivity was high at the southern stations on lines Aand D (Figure 11a). Average nitrate concentrations in theupper 15 m at these stations exceeded 2.8 mM comparedwith values less than 0.7 mM at other stations in the channel(data not shown).

Figure 12. (a) Five meter productivity for cruise 4. The magnitude of productivity is proportional to thediameters of the circles as scaled by the labeled circle at bottom left. Satellite-derived chlorophyll concen-trations (b) at 20:36 UTC on 27 April 2002 and (c) at 19:40 UTC on 28 April 2002.


15 of 17

[50] The high productivity in the SBC resulting fromwind-driven upwelling and eddy processes can stimulatespring blooms of toxic diatoms of the genus Pseudo-nitzschia[Anderson et al., 2006, 2008, 2009]. Bloom events duringspring cruises in 2001, 2003, 2004 and 2005 containedPseudo-nitzschia (ibid). The largest of these occurred duringcruise 7 in 2003 and was dominated by Pseudo-nitzschiaaustralis with high levels of domoic acid [Anderson et al.,2006]. Thus the same mechanisms that drive spatial pat-terns in productivity may also influence the intensity anddistribution of harmful algal blooms in the SBC.[51] Our results only partially resolve productivity patterns

associated with the gradient in chlorophyll biomass alongthe eastern mainland coast (Figure 1a). EOF mode 3 cap-tured the alongshore gradient in productivity consistent withthis enhanced biomass. This mode accounted for only asmall fraction of total variance suggesting that mechanismsdriving productivity along the eastern portion of the main-land coast are ephemeral and poorly sampled by our thriceyearly cruises. However, mode 3 explained up to 46% ofproductivity variance at the two stations nearest the mouthsof the Santa Clara and Ventura Rivers. This may indicatesome relationship of high productivity at these stations to theriver mouths, although river discharge was negligible duringeach cruise except cruise 9. Other mechanisms that coulddrive this productivity include wave-driven suspension ofpreviously deposited, nutrient-laden sediments and enhancednutrient supply forced by internal tidal activity [Lucas et al.,2011]. Enhanced productivity in this area may also berelated to the broader continental shelf in the eastern SBC orthe proximity to the eastern entrance of the SBC wherepoleward flow from the greater Southern California Bightoccurs frequently [Lynn and Simpson, 1987; Harms andWinant, 1998].

5. Conclusions

[52] Spatial distributions of productivity, currents, andwater properties were measured during 16 oceanographiccruises in the SBC from 2001 to 2006. Our analysis reveals astrong coupling of patterns of productivity with wind-drivennutrient upwelling and cyclonic eddies. These processescause productivity in the SBC to exceed values reportedelsewhere in the Southern California Bight. The highestlevels of productivity in the SBC occur in spring owing tonutrient input by coastal upwelling. Cyclonic eddies enhanceproductivity in the western SBC through retention ofupwelled waters in the SBC or through eddy-induced iso-pycnal uplift or combinations of both mechanisms. Eddyprocesses are most apparent in fall when nutrient supply bywind-driven upwelling is low or absent. Enhanced produc-tivity on the continental shelf in the eastern SBC wasobserved, but was not well resolved.

[53] Acknowledgments. We thank the captains and crewmembers ofthe R/V Point Sur for excellent service on all of the cruises in this study. Wethank Dave Nelson, Doug Conlin, and Stewart Lamerdin for operating theCTDs, rosettes, and towed vehicles. David Salazar, Janice Jones, and ChrisGotschalk provided critical technical assistance. We thank Peter Franks andan anonymous reviewer for helpful comments on an earlier version of themanuscript. This work was supported by the U.S. National Science Founda-tion’s Long-Term Ecological Research Program under the Division ofOcean Sciences grants OCE9982105 and OCE 0620276. Funding for oper-ation of the HF radars was provided by the Partnership for Interdisciplinary

Studies of Coastal Oceans and by the Southern California Coastal OceanObserving System.

ReferencesAnderson, C. R., M. A. Brzezinski, L. Washburn, and R. Kudela (2006),Circulation and environmental conditions during a toxigenic Pseudo-nitzschia australis bloom in the Santa Barbara Channel, California,Mar. Ecol. Prog. Ser., 327, 119–133, doi:10.3354/meps327119.

Anderson, C. R., D. A. Siegel, M. A. Brzezinski, and N. Guillocheau(2008), Controls on temporal patterns in phytoplankton community struc-ture in the Santa Barbara Channel, California, J. Geophys. Res., 113,C04038, doi:10.1029/2007JC004321.

Anderson, C. R., D. A. Siegel, R. M. Kudela, and M. A. Brzezinski (2009),Empirical models of toxigenic Pseudo-nitzschia blooms: Potential use asa remote detection tool in the Santa Barbara Channel, Harmful Algae, 8,478–492, doi:10.1016/.hal.2008.10.005.

Bassin, C. J., L. Washburn, M. Brzezinski, and E. McPhee-Shaw (2005),Sub-mesoscale coastal eddies observed by high frequency radar: A newmechanism for delivering nutrients to kelp forests in the Southern Califor-nia Bight, Geophys. Res. Lett., 32, L12604, doi:10.1029/2005GL023017.

Beckenbach, E., and L. Washburn (2004), Low-frequency waves in theSanta Barbara Channel observed by high-frequency radar, J. Geophys.Res., 109, C02010, doi:10.1029/2003JC001999.

Beers, J. R. (1986), Organisms and the food web, in Plankton Dynamics ofthe Southern California Bight, edited by R. W. Eppley, pp. 84–175,Springer, New York.

Benitez-Nelson, C. R., et al. (2007), Mesoscale eddies drive increased silicaexport in the subtropical Pacific Ocean, Science, 316, 1017–1021,doi:10.1126/science.1136221.

Capet, X., J. C. McWilliams, M. J. Molemaker, and A. F. Shchepetkin(2008a), Mesoscale to submesoscale transition in the California CurrentSystem. Part I: Flow structure, eddy flux, and observational tests, J. Phys.Oceanogr., 38, 29–43, doi:10.1175/2007JPO3671.1.

Capet, X., J. C. McWilliams, M. J. Molemaker, and A. F. Shchepetkin(2008b), Mesoscale to submesoscale transition in the California CurrentSystem. Part II: Frontal processes, J. Phys. Oceanogr., 38, 44–64,doi:10.1175/2007JPO3672.1.

Cudaback, C. N., L. Washburn, and E. P. Dever (2005), Subtidal inner-shelfcirculation near Point Conception, California, J. Geophys. Res., 110,C10007, doi:10.1029/2004JC002608.

Dever, E. P., M. C. Hendershott, and C. D. Winant (1998), Statisticalaspects of surface drifter observations of circulation in the Santa BarbaraChannel, J. Geophys. Res., 103, 24,781–24,797.

Dorman, C. E., and C. D. Winant (2000), The structure and variability ofthe marine atmosphere around the Santa Barbara Channel, Mon. WeatherRev., 128, 261–282, doi:10.1175/1520-0493(2000)128<0261:TSAVOT>2.0.CO;2.

Emery, B. M., L. Washburn, and J. A. Harlan (2004), Evaluating radial cur-rent measurements from CODAR high-frequency radars and mooredin situ current meters, J. Atmos. Oceanic Technol., 21, 1259–1271,doi:10.1175/1520-0426(2004)021<1259:ERCMFC>2.0.CO;2.

Emery, W. J., and R. E. Thompson (2001), Data Analysis Methods in Phys-ical Oceanography, Elsevier, New York.

Eppley, R. W., E. H. Renger, and W. G. Harrison (1979), Nitrate and phy-toplankton production in Southern California coastal waters, Limnol.Oceanogr., 24, 483–494, doi:10.4319/lo.1979.24.3.0483.

Fewings, M. R., S. J. Lentz, and J. Fredericks (2008), Observations ofcross-shelf flow driven by cross-shelf winds on the inner continentalshelf, J. Phys. Oceanogr., 38, 2358–2378, doi:10.1175/2008JPO3990.1.

Fiedler, P. C., et al. (1998), Blue whale habitat and prey in the CaliforniaChannel Islands, Deep Sea Res., Part II, 45, 1781–1801, doi:10.1016/S0967-0645(98)80017-9.

Fram, J. P., H. L. Stewart, M. A. Brzezinski, D. C. Reed, S. L. Williams,and S. MacIntyre (2008), Physical pathways and utilization of nitratesupply to the giant kelp, Macrocystis pyrifera, Limnol. Oceanogr.,53, 1589–1603.

Halpern, B. S., K. Cottenie, and B. R. Broitman (2006), Strong top-downcontrol in Southern California kelp forest ecosystems, Science, 312,1230–1232, doi:10.1126/science.1128613.

Hardy, J. T. (1993), Phytoplankton, in Ecology of the Southern CaliforniaBight: A Synthesis and Interpretation, edited by M. D. Dailey, D. J.Reish, and J. W. Anderson, pp. 233–258, Univ. of Calif. Press, Berkeley,Calif.

Harms, S., and C. D. Winant (1998), Characteristic patterns of the circula-tion in the Santa Barbara Channel, J. Geophys. Res., 103, 3041–3065,doi:10.1029/97JC02393.

Johnson, K. S., R. L. Petty, and J. Thompsen (1985), Flow-injection analy-sis for seawater micronutrients, in Mapping Strategies in Chemical


16 of 17

Oceanography, edited by A. Zirino, pp. 7–30, Am. Chem. Soc.,Washington, D. C.

Kim, S. Y., et al. (2011), Mapping the U.S. West Coast surface circulation:A multiyear analysis of high-frequency radar observations, J. Geophys.Res., 116, C03011, doi:10.1029/2010JC006669.

Lucas, A. J., C. L. Dupont, V. Tai, J. L. Largier, B. Palenik, and P. J. S.Franks (2011), The green ribbon: Multiscale physical control of phyto-plankton productivity and community structure over a narrow continentalshelf, Limnol. Oceanogr., 56, 611–626, doi:10.4319/lo.2011.56.2.0611.

Lynn, R. J., and J. J. Simpson (1987), The California Current System: Theseasonal variability of its physical characteristics, J. Geophys. Res., 92,12,947–12,966.

Mantyla, A. W., E. L. Venrick, and T. L. Hayward (1995), Primary produc-tion and chlorophyll relationships, derived from ten years of CALCOFImeasurements, Rep. 36, pp. 159–166, Calif. Coop. Oceanic Fish. Invest.,La Jolla, Calif.

McClain, E. P., W. G. Pichel, and C. C. Walton (1985), Comparativeperformance of AVHRR-based multichannel sea surface temperatures,J. Geophys. Res., 90(C6), 11,587–11,601, doi:10.1029/JC090iC06p11587.

McClain, C. R., G. C. Feldman, and S. B. Hooker (2004), An overview ofthe SeaWiFS Project and strategies for producing a climate research qualityglobal ocean bio-optical time series, Deep Sea Res. Part II, 51(1–3), 5–42,doi:10.1016/j.dsr2.2003.11.001.

McGillicuddy, D. J., et al. (1998), Influence of mesoscale eddies on newproduction in the Sargasso Sea, Nature, 394, 263–266, doi:10.1038/28367.

McGillicuddy, D. J., et al. (2007), Eddy/wind interactions stimulateextraordinary mid-ocean plankton blooms, Science, 316, 1021–1026,doi:10.1126/science.1136256.

McPhee-Shaw, E., et al. (2007), Mechanisms for nutrient delivery tothe inner shelf: Observations from the Santa Barbara Channel, Limnol.Oceanogr., 52, 1748–1766, doi:10.4319/lo.2007.52.5.1748.

Melton, C., L. Washburn, and C. Gotschalk (2009), Wind relaxations andpoleward flow events in a coastal upwelling system on the central Californiacoast, J. Geophys. Res., 114, C11016, doi:10.1029/2009JC005397.

Miller, R. J., D. C. Reed, and M. A. Brzezinski (2011), Partitioning ofprimary production among giant kelp (Macrocystis pyrifera), understorymacroalgae and phytoplankton on a temperate reef, Limnol. Oceanogr.,56, 119–132, doi:10.4319/lo.2011.56.1.0119.

Nishimoto, M. M., and L. Washburn (2002), Patterns of coastal eddy cir-culation and abundance of pelagic juvenile fish in the Santa BarbaraChannel, California, USA, Mar. Ecol. Prog. Ser., 241, 183–199,doi:10.3354/meps241183.

Oey, L.-Y., D.-P. Wang, T. Hayward, C. Winant, and M. Hendershott(2001), “Upwelling” and “cyclonic” regimes of the near-surface circula-tion in the Santa Barbara Channel, J. Geophys. Res., 106, 9213–9222,doi:10.1029/1999JC000129.

Otero, M. P., and D. A. Siegel (2004), Spatial and temporal characteristicsof sediment plumes and phytoplankton blooms in the Santa BarbaraChannel, Deep Sea Res., Part II, 51, 1129–1149.

Shipe, R. F., and M. A. Brzezinski (2001), A time-series study of silicaproduction and flux in an eastern boundary region: Santa BarbaraBasin, California, Global Biogeochem. Cycles, 15, 517–531, doi:10.1029/2000GB001297.

Shipe, R. F., and M. A. Brzezinski (2003), Siliceous plankton dominate pri-mary and new productivity during the onset of El Niño conditions in theSanta Barbara Basin, California, J. Mar. Syst., 42, 127–143, doi:10.1016/S0924-7963(03)00071-X.

Shipe, R. F., U. Passow, M. A. Brzezinski, W. M. Graham, D. K. Pak, D. A.Siegel, and A. L. Alldredge (2002), Effects of the 1997–1998 El Nino onseasonal variations in suspended and sinking particles in the SantaBarbara Basin, Prog. Oceanogr., 54, 105–127, doi:10.1016/S0079-6611(02)00045-9.

Siegel, D. A., D. J. McGillicuddy Jr., and E. A. Fields (1999), Mesoscaleeddies, satellite altimetry, and new production in the Sargasso Sea,J. Geophys. Res., 104, 13,359–13,379, doi:10.1029/1999JC900051.

Smith, S. D. (1988), Coefficients for sea surface wind stress, heat flux, andwind profiles as a function of wind speed and temperature. J. Geophys.Res., 93, 15,467–15,472.

Winant, C. D., E. P. Dever, and M. C. Hendershott (2003), Characteristicpatterns of shelf circulation at the boundary between central and southernCalifornia, J. Geophys. Res., 108(C2), 3021, doi:10.1029/2001JC001302.

M. A. Brzezinski, Department of Ecology Evolution and Marine Biology,Marine Science Institute, University of California, Santa Barbara, CA93106, USA. ([email protected])L. Washburn, Department of Geography, Marine Science Institute,

University of California, Santa Barbara, CA 93106, USA.


17 of 17

Documents

PhytoplanktonprimaryproductivityintheSantaBarbaraChannel: …washburn/docs/brzezinski_washburn... · 2012-07-03 · 2.2. Water Sampling [11] Water samples were collected using a CTD/rosette