Forgot your password?
Document Actions
Barth et al 07
Delayed upwelling alters nearshore coastal ocean
ecosystems in the northern California current
John A. Barth*†, Bruce A. Menge‡, Jane Lubchenco†‡, Francis Chan‡, John M. Bane§, Anthony R. Kirincich*,
Margaret A. McManus¶, Karina J. Nielsen , Stephen D. Pierce*, and Libe Washburn**
*College of Oceanic and Atmospheric Sciences and ‡Department of Zoology, Oregon State University, Corvallis, OR 97331; §Department of Marine Sciences,
University of North Carolina, Chapel Hill, NC 27599; ¶Department of Oceanography, University of Hawaii, Manoa, HI 96822; Department of Biology,
Sonoma State University, Rohnert Park, CA 94928; and **Department of Geography, University of California, Santa Barbara, CA 93106
Contributed by Jane Lubchenco, January 22, 2007 (sent for review December 9, 2006)
Wind-driven coastal ocean upwelling supplies nutrients to the of upwelling-favorable winds might be influenced by climate vari-
euphotic zone near the coast. Nutrients fuel the growth of phy- ability is an area of active research (5, 6).
toplankton, the base of a very productive coastal marine ecosys- Spatial and temporal changes in nutrient availability and phyto-
tem [Pauly D, Christensen V (1995) Nature 374:255–257]. Because plankton biomass propagate up marine food webs and may strongly
nutrient supply and phytoplankton biomass in shelf waters are mediate the structure and dynamics of nearshore ecological com-
highly sensitive to variation in upwelling-driven circulation, shifts munities (7). Decreased upwelling can have two effects: reduced
in the timing and strength of upwelling may alter basic nutrient nutrient supply to phytoplankton (8) and reduced offshore trans-
and carbon fluxes through marine food webs. We show how a port of phytoplankton and planktonic fish and invertebrate larvae
1-month delay in the 2005 spring transition to upwelling-favorable that are crucial for replenishing coastal populations (9). Nutrient
wind stress in the northern California Current Large Marine Eco- and phytoplankton reductions may decrease zooplankton (includ-
system resulted in numerous anomalies: warm water, low nutrient ing larvae) survival, and thus decrease recruitment rates of plank-
levels, low primary productivity, and an unprecedented low re- totrophic larvae. Slower offshore transport could have the opposite
cruitment of rocky intertidal organisms. The delay was associated effect, with greater retention and increased recruitment rates.
with 20- to 40-day wind oscillations accompanying a southward
When intraseasonal variations in upwelling forcing are substantial,
shift of the jet stream. Early in the upwelling season (May–July) off
changes in the timing of phytoplankton blooms can potentially
decouple food availability from consumer demands (10). Such
Oregon, the cumulative upwelling-favorable wind stress was the
temporal trophic mismatches may have particularly important
lowest in 20 years, nearshore surface waters averaged 2°C warmer
effects on recruitment class strength in fish and invertebrates (11).
than normal, surf-zone chlorophyll-a and nutrients were 50% and
Recruitment patterns of rocky shore barnacles and mussels can
30% less than normal, respectively, and densities of recruits of
ECOLOGY
provide insight into the dynamics of inner-shelf ecosystems (12).
mussels and barnacles were reduced by 83% and 66%, respec- Sessile adult barnacles and mussels release larvae or gametes into
tively. Delayed early-season upwelling and stronger late-season the water that spend 2–4 weeks in the plankton, feeding on small
upwelling are consistent with predictions of the influence of global plankton for their sustenance (13). Larvae develop until they are
warming on coastal upwelling regions. competent to settle, then they must be transported back to shore if
they are to recruit and begin life as tiny sessile barnacles or mussels.
climate variability coastal marine ecosystems coastal ocean upwelling Growth rates, settlement sizes, and survival of mussel larvae, and
marine ecology probably barnacle larvae, depend on phytoplankton concentration
GEOPHYSICS
that, in turn, depends on nutrients provided by upwelling events
E quatorward winds along the eastern boundaries of the world’s
oceans drive offshore surface Ekman transport and the up-
welling of cold, nutrient-rich water into the euphotic zone near the
(14). Previous work has documented that both barnacles and
mussels recruit heavily but intermittently throughout summer in
Oregon (9). Recruitment events of barnacles in California (15) and
coast. These nutrient pulses stimulate high phytoplankton produc- Oregon†† often follow upwelling relaxation, i.e., times when larvae
tion, which, in turn, supports a rich coastal marine ecosystem and are transported shoreward.
productive fisheries (1). Examples of such dynamics include the Barnacles and mussels create 3D spatial structure (habitat) in
California Current, the Humboldt Current, the Benguela Current, rocky intertidal communities and are the primary food for many
and the Canary Current (2). predators (16). Barnacle and mussel juveniles and adults are
The strength and extent of the seasonal cycle in upwelling- abundant and relatively easy to monitor and have similar life
favorable winds varies along the U.S. west coast. In the northern histories to numerous other ecologically or economically important
pelagic and benthic inner-shelf species. Consequently, they are good
California Current Large Marine Ecosystem (CCLME), there is a
surrogates for many other inner-shelf species with respect to the
strong seasonal cycle with upwelling-favorable winds, the appear-
ance of cold, saline, nutrient-rich water near the coast, and equa-
torward currents over the shelf occurring after a spring transition Author contributions: J.A.B., B.A.M., J.L., F.C., M.A.M., and L.W. designed research; J.A.B.,
(3). Alongshore winds in the northern CCLME are more variable B.A.M., J.L., F.C., J.M.B., A.R.K., M.A.M., K.J.N., S.D.P., and L.W. performed research; J.A.B.,
than those farther south because they are more frequently influ- B.A.M., J.L., F.C., J.M.B., A.R.K., M.A.M., K.J.N., S.D.P., and L.W. analyzed data; and J.A.B.,
B.A.M., J.L., F.C., J.M.B., A.R.K., M.A.M., K.J.N., S.D.P., and L.W. wrote the paper.
enced by eastward-traveling Gulf of Alaska low-pressure systems.
The authors declare no conflict of interest.
The intermittent cessation of upwelling-favorable winds is called
Freely available online through the PNAS open access option.
relaxation and plays an important role in coastal circulation and the
recruitment of marine organisms††. The timing of the spring Abbreviations: CCLME, California Current Large Marine Ecosystem; chl-a, chlorophyll-a;
ISO, intraseasonal oscillation; NDBC, National Data Buoy Center; C-MAN, Coastal-Marine
transition and the total amount of upwelling-favorable winds during Automated Network.
the spring–summer upwelling season have a considerable impact on †To whom correspondence may be addressed. E-mail: barth@coas.oregonstate.edu or
coastal ecosystem responses. Farther south in the CCLME, winds lubchenco@oregonstate.edu.
are more persistently upwelling-favorable and the transition to a ††Grantham, B. A., Menge, B. A., Lubchenco, J. (2002) Eos Trans Am Geophys Union
more productive spring–summer season is less pronounced. The 83(Suppl):OS12K-03 (abstr.).
extent to which the timing of the spring transition and the intensity © 2007 by The National Academy of Sciences of the USA
www.pnas.org cgi doi 10.1073 pnas.0700462104 PNAS March 6, 2007 vol. 104 no. 10 3719 –3724
A
A
B
B
C
Fig. 2. Maps of 200-hPa surface height (m, blue contours) and wind speed at
300 hPa (m s 1, color shading), from six hourly National Centers for Environmen-
tal Prediction reanalyses (25) when the jet stream is located anomalously south
(May 20, 2005, 1200 Coordinated Universal Time; A) and in its more typical
summer position (July 19, 2005, 0000 Coordinated Universal Time; B). Wind
speeds 35 m s 1 are shown, with color increments at 45, 55, and 65 m s 1.
change. We use a long-term set of physical, chemical, and biological
measurements to examine the unusual presence of warm water, low
nutrients and low chlorophyll-a (chl-a) near the coast during 2005
that led to unprecedented low recruitment of rocky intertidal
organisms in the northern CCLME during the early upwelling
season.
D
Results and Discussion
Perturbations to Upwelling Regime. Year-to-year differences in the
timing of the spring transition and the total amount of seasonal
Fig. 1. Alongshore wind stress off the U.S. west coast. (A–C) Alongshore wind upwelling-favorable winds may have considerable impact on coastal
stress from three west-coast buoys: Newport, OR (A), Monterey Bay, CA (B), and
Point Conception, CA (C). Values for 2005 (blue) are plotted on top of the
ecosystem responses. Off central Oregon (44.62°N) where winds
climatological mean (black) 1 SD (gray shading) for 1985–2005. In A, the dashed have a pronounced seasonal cycle, anomalous winds in 2005 were
curve is the north–south position of the jet stream, and the two arrows indicate dominated by five strong, northward wind events, separated by
the times of maps shown in Fig. 2. (D) Cumulative alongshore (north–south) wind 20–40 days, during March–July (Fig. 1A). The onset of unusually
stress from National Oceanic and Atmospheric Administration NDBC Buoy 46050 strong, upwelling-favorable winds off central Oregon in early to
offshore of Newport, OR, starting from the spring transition. The black curve and mid-July coincided with a northward shift of 1,000 km in the
shading represent the mean 1 SD for 1985–2005, and the blue curve is for 2005. position of the jet stream (Fig. 1 A). A map of 200-hPa surface
At zero cumulative wind stress, the black curve and shading represent the mean
height during the mid-May northward wind event that occurred
and 1 SD of the date of the spring transition, and the blue line represents the
date of the 2005 spring transition. The 2005 curves are dark blue when absolute
before this northward shift shows a jet-stream position, and hence
values exceed any observed values during the previous 20 years. Data locations storm track, more typical of wintertime conditions (17) compared
are indicated at lower left. with a typical summertime jet-stream location in mid-July after the
shift (Fig. 2).
Off central California (37.36°N), 2005 winds were not unusual
impact of environmental changes in the inner shelf. There are few except for perhaps some stronger-than-normal southward winds
long-term, ecological time series available for the detection and from April to June (Fig. 1B). Off of southern California (34.72°N),
quantification of ocean ecosystem responses to environmental but outside the southern California Bight lee region, there were two
3720 www.pnas.org cgi doi 10.1073 pnas.0700462104 Barth et al.
A
B
C
Fig. 3. Surface (0 –2 m) temperatures during 2005 (solid lines) compared with climatological means (dashed lines) with 1 SD (shaded) from inner-shelf
moorings in 15 m of water off of central Oregon (44.25°N, 124.13°W) (1998 –2004 mean) (A), 21 m of water off of Monterey Bay, CA (36.97°N, 122.16°W)
(1999 –2004 mean) (B), and 15 m of water off of Santa Barbara, CA (34.46°N, 120.29°W) (1999 –2004) (C). Temperatures during 2005 that are warmer (colder) than
the climatological mean 1 SD are shaded in red (blue).
unusually strong northward wind events in February-March fol- transition off of Oregon contrasts strongly with the long-term
lowed by normal or slightly stronger upwelling-favorable winds the average and reveals the unusual timing of the upwelling-favorable
remainder of the year (Fig. 1C). winds (Fig. 1D). By examining wind and sea-level records, we
In 2005, the cumulative alongshore wind stress since the spring determined that the 2005 spring transition was on May 24, over a
ECOLOGY
A B C D 46
45
44
43
GEOPHYSICS
42
41
40
E F G H
Fig. 4. Surf-zone chlorophyll and nutrients measured along the Oregon and north/central California coasts. (A–D) Chl-a measured along the coast during
May-August 2005 (F), the long-term climatological mean (E) with 95% C.I. (bars), and 2005 anomalies from the mean (colored bars). (E–H) As in A–D, but for
nitrate plus nitrite (N N). The arrows at 44.25°N indicate the location of time series shown in Fig. 5.
Barth et al. PNAS March 6, 2007 vol. 104 no. 10 3721
month later than average and outside of the 1-SD range of the last A
20 years. The first substantial upwelling did not occur until late
June, 2 months later than average, but was terminated by the last
of the northward wind events in July 2005. Stronger-than-average
upwelling-favorable winds commenced in early to mid-July and
persisted until the seasonal accumulation of upwelling-favorable
winds reached the long-term average in September.
Effects on Temperature. Near-surface temperatures measured close
to shore illustrate the coastal ocean response to the unusual winds
of 2005 (Fig. 3). Off of central Oregon from May to mid-July 2005,
nearshore surface waters averaged 2°C warmer than normal, with
a maximum warming of 6.4°C (Fig. 3A). Upwelling-favorable winds
in mid-June (weak) and late in June (stronger) cooled upper-ocean
temperatures to nearly normal. However, after each of these
cooling periods, nearshore temperatures returned to above-average
levels, as warmer offshore waters were advected onshore by surface B
Ekman transport driven by northward wind events. In mid-July,
nearshore temperatures cooled to below normal as a result of
persistent and vigorous coastal upwelling. Off of central California,
nearshore surface temperatures were 2–3°C higher than normal
until early to mid-April, in part because of the northward wind event
in mid-March, after which temperatures returned to normal (Fig.
3B). Off of southern California, nearshore surface temperatures
were near normal, except for perhaps slight above-average tem-
perature in late February–early March in response to two north-
ward wind events (Fig. 3C).
Bottom-Up Effects. In 2005, coastwide negative chl-a anomalies were
evident during May, with the largest anomalies off of the central
Oregon coast (Fig. 4A). Depression of surf-zone chl-a continued
through July before reverting to neutral or increased levels in Fig. 5. Time series of surf-zone chlorophyll and nutrients off the central Oregon
August (Figs. 4A and 5). The May 2005 negative chl-a anomalies coast. (A) Time series of chl-a (circles) measured at the coast off of central Oregon
were accompanied by strong, coastwide decrease in nitrate con- (44.25°N): 2005 (filled symbols); 1993–2004 climatological mean (open symbols)
centration (Fig. 4B). Negative nitrate anomalies did not accompany with 95% C.I. (bars). (B) As in A, but for nitrate plus nitrite (diamonds).
chl-a reductions in June and July at the coastwide scale, although
persistent negative nitrate anomalies were evident from a number
of sites along the central and southern Oregon coasts (Figs. 4 and 5). The influence of a delayed transition to upwelling was most
dramatic in the northern CCLME (18) as might be expected
Effects on Recruitment. The 2-month delay in the onset of sub- because the seasonal cycle in wind stress is most pronounced in this
stantial upwelling in 2005 had serious consequences for recruitment region (Fig. 1 A). Because upwelling was abnormally low in the early
of mussels and barnacles. Mussel recruitment during May-August season, and because weak upwelling should lead to retention of
in 2005 was the lowest ever recorded for these 4 months at every site larvae close to shore, the unprecedented low early-season recruit-
(Fig. 6). During the next 2 months, September–October, after ment of mussels and barnacles was probably caused by low food
unusually intense and continuous upwelling, recruitment re- supply (i.e., low phytoplankton, a consequence of low nutrients)
bounded to values higher than average except for the two stations rather than by offshore loss of larvae caused by advection. By
in the extreme south and north of the measurement region. midsummer, however, upwelling-favorable winds were stronger
Barnacles and mussels often show different recruitment patterns than normal, resulting in higher phytoplankton biomass (Figs. 4 and
(8). Early in the season (May–July), barnacle (Balanus glandula) 5). During the late summer season, recruitment of both mussels and
recruitment was lower than normal at a number of, but not all, sites barnacles rebounded. Because spawning to settlement of these
(Fig. 6; see Methods for details of analysis). During August– organisms takes 2 weeks to 2 months (13), it is likely that these
October, after strong and persistent upwelling, B. glandula recruit- late-season recruits (August–November) were the result of repro-
ment was higher, indicating a rebound, at many, but not all, sites. duction that occurred shortly after the resumption of upwelling in
For both barnacles and mussels, recruitment in 2005 differed mid-July. Spatial differences in recruitment patterns may reflect
greatly between early (June–August for mussels, May–July for among-site differences in nearshore circulation.
barnacles) versus late (September–November for mussels, August–
October for barnacles) recruitment at most sites (Fig. 6). Consequences for the Marine Food Web. The ecological conse-
Summing results across all 10 sites indicates that in 2005 during quences of these changes are potentially severe. The low recruit-
June–August, mussel recruitment was reduced by 83%. During the ment of the intertidal species reported here is consistent with
next 3 months (September–November), mussel recruitment re- lower-than-normal concentrations of zooplankton (17) in the
bounded for an increase of 53% compared with normal years. northern CCLME. Reproductive failure of a planktivorous seabird,
Overall, however, total mussel recruitment for the entire 6-month Cassin’s auklet (Ptychoramphus aleuticus) in the CCLME during
recruitment season (June–November) in 2005 was normal. Barna- 2005 (17) is also a likely result of the chain of reductions at lower
cle recruitment in 2005 was depressed 66% during May–July and trophic levels in coastal food webs initiated by the reduced nutrients
was nearly normal (2.5% increase) during August–October, al- and phytoplankton at a critical time in the life history of coastal
though this late-season response was variable among sites. Total seabirds. Reduced phytoplankton is likely to slow growth and
barnacle recruitment for the 6-month recruitment season (May– suppress reproductive output of sessile invertebrates (19). Reduced
October) in 2005 was 38% less than normal. recruitment of sessile invertebrates could diminish the food supply
3722 www.pnas.org cgi doi 10.1073 pnas.0700462104 Barth et al.
ECOLOGY
Fig. 6. Mussel (Mytilus spp.) (Left) and barnacle (B. glandula) (Right) recruitment at sites along the Oregon coast: 2005 (F) compared with climatological
monthly means (E) computed for 8 –17 years (including 2005) depending on site. Means and SEM are shown for all data. Dotted vertical lines mark the ‘‘early’’
GEOPHYSICS
and ‘‘late’’ recruitment seasons used in data analysis. The absence of dotted lines indicates sites where no differences between years or season occurred.
for predators such as whelks, sea stars, crabs, and shorebirds, with In experiments using a regional climate model, albeit in the
potentially far-reaching consequences elsewhere in the food absence of an active ocean submodel, increased greenhouse gas
web (20). forcing results in a 1-month delay in the onset of the coastal
upwelling season in the northern CCLME and an increase in
Intraseasonal Oscillations (ISOs). Recent studies have shed light on upwelling intensity later in the season (June–September) (6). The
the importance of ISOs of the wind (20- to 40-day periods) on mechanism by which this happens, changes in radiative forcing, is
coastal upwelling (21, 22) and coastal ecosystem dynamics (J.M.B., different from the dynamics described above involving wind ISOs
Y. Spitz, R. Letelier, and W. T. Peterson, unpublished work). The and an anomalously southward jet-stream position that led to
five large wind ISOs played a key role in delaying the spring delayed upwelling. Sorting out the relative roles of these atmo-
transition in the northern CCLME in 2005. Wind ISOs at midlati- spheric processes is key to our ability to predict future changes in
tude are formed through interactions of the midlatitude jet stream coastal ocean ecosystems caused by climate variability and climate
with large-scale topography (23). During the summer of 2001, ISOs change.
in alongshore wind stress off of Oregon (44.6°N) correlated well
with the north–south position of the jet stream (22). The cessation Methods
of strong wind ISOs and the onset of unusually strong upwelling- Meteorological Data. Wind stress was calculated by using measure-
favorable wind in the northern CCLME in early to mid-July 2005 ments from the National Oceanic and Atmospheric Administration
were consistent with a shift of the jet stream from the south to the National Data Buoy Center (NDBC) buoys and Coastal-Marine
north of this region (Figs. 1 A and 2). Automated Network (C-MAN) stations (24) and then low-pass-
Delayed early-season upwelling and stronger late-season up- filtered with a filter with a 40-h width at half-amplitude to remove
welling documented here for the northern CCLME during 2005 are short-period (e.g., diurnal) fluctuations. The alongshore wind stress
consistent with predictions of the influence of global warming on was computed by rotating into a coordinate system aligned with the
coastal upwelling regions (5, 6). The global-warming scenario relies local coastline (rotation angle indicated for each site below). Winds
on increased land–sea temperature contrasts, created by preferen- were measured at NDBC buoy 46050 off of central Oregon
tial heating of the land caused by elevated greenhouse gases, driving (44.62°N, 124.53°W; 3°), buoy 46012 off of Monterey, CA (37.36°N,
stronger equatorward winds, and hence stronger upwelling. 122.88°W; 327°), and buoy 46023 off of Point Arguello, CA
Barth et al. PNAS March 6, 2007 vol. 104 no. 10 3723
(34.72°N, 120.97°W; 329°). Gaps in the NDBC buoy records were (Tuffys; The Clorox Company, Oakland, CA), and barnacle re-
filled through regression with nearby buoys and/or C-MAN sta- cruitment was quantified by using settlement plates [0.10 0.10
tions: 46050 with Newport, OR, C-MAN (NWPO3) and occasion- 0.004-m poly(vinyl chloride) plates with Saf-T-Walk (3M Company,
ally the Cape Arago, OR, C-MAN (CARO3); 46012 with buoys St. Paul, MN), a textured plastic tape, on the top side] (4, 28).
46026 and 46042; 46023 with buoy 46011 and Point Conception, Collectors were deployed/recovered monthly and processed in the
CA, C-MAN (PTGC1). A 5-day running mean was then applied to laboratory where counts of mussels (number per collector) and
all time series. To assess the cumulative alongshore wind stress off barnacles (number per 10 2 m2) were made under dissecting
of central Oregon during the upwelling season, north–south stress microscopes. Barnacle recruits were separated by species with B.
was summed starting from the spring transition, defined as when glandula and Chthamalus dalli as the most abundant barnacles for
winds turn to predominantly upwelling-favorable (southward) usu- which we have long-term recruitment data. Mussel species cannot
ally during March to April (3). be reliably identified visually. Judging from the species composition
The latitudinal position of the jet stream in the northern CCLME of mussels growing to identifiable sizes on the shore after recruit-
was taken as the location, along 125°W longitude, of the strongest ment, most mussel recruits are Mytilus trossulus, but very small
horizontal gradient in the height of the 200-hPa surface (22), numbers of Mytilus californianus also settle jointly with Mytilus
determined from six hourly National Centers for Environmental trossulus.
Prediction reanalysis pressure maps (25). The jet-stream position Because we were interested in comparing among-year and with-
time series was then filtered with a 35-d cosine low-pass filter. in-season differences, recruitment data were analyzed by using
two-way ANOVA. Factors tested were year (2005 vs. the full data
Mooring-Based Measurements. Temperature was measured at 1- to set) and recruitment season (early and late). Both factors were
2-min intervals just below the surface (0–3 m) by using StowAway fixed. To improve the fit of the data to a normal distribution, we
XTI Temperature Loggers (Onset Computer Corp.) on moorings transformed the response variable to ln (recruits per day 1); we
deployed in water depths of 15 m (central Oregon, 44.25°N, used P 0.05 as our level of significance. Sample size (n) in these
124.13°W; Santa Barbara, CA, 34.46°N, 120.29°W) or 21 m analyses varied among sites from 181 to 517 (mussels) and 184 to
(Monterey, CA, 36.97°N, 122.16°W). 571 (barnacles) and is based on five to eight recruit collector
samples for each species per month per site for the time periods
Coastal Transects. Surf-zone chl-a and inorganic nutrient measure- listed above. Sample size varied because of varying numbers of
ments (nitrate plus nitrite) were made between 2001 and 2005 years in which sampling was done and occasional losses of collec-
during May and June and between 1997 and 2005 during July and tors. We report P values for interactions (year season), which
August, at 30 sites along the Oregon and north/central California subsumes main effects when significant, or main effects (year or
coasts (46–38°N). Once a month from May through August, season) when interactions were not significant (Fig. 6). For barna-
synoptic measurements were all made on the same day along the cles, analysis of data using May–July and August–October provided
coast between 3 h before and 3 h after low tide. In central Oregon the clearest contrast between early and late recruitment seasons,
(44.25°N), additional data were collected more frequently (daily to respectively (except for Cape Arago and Cape Blanco). For mus-
biweekly) between 1993 and 2005. Anomalies were computed by sels, analysis of data using June–August and September–November
differencing values from 2005 with a long-term mean for each provided the clearest contrast between early and late recruitment
month: May–June (2001–2005) and July–August (1997–2005). seasons (except for Rocky Point) The variation in which season was
Chl-a and nitrate plus nitrite measurements were made according analyzed is shown by the vertical dotted lines in Fig. 6.
to standard methods (26, 27).
We thank the many scientists, technicians, and students of the Partner-
ship for Interdisciplinary Studies of Coastal Oceans who were involved
Mussel and Barnacle Recruitment. Monthly recruitment (the appear- with the collection of the long-term data sets and who made this study
ance of new young on rocky shores) of mussels and barnacles has possible. The Partnership for Interdisciplinary Studies of Coastal Oceans
been measured monthly at 10 sites spanning two-thirds of the is supported by the David and Lucile Packard Foundation and the
Oregon coast: for 17 years at Boiler Bay (44.83°N, 124.06°W) and Gordon and Betty Moore Foundation. Additional support for J.A.B. was
Strawberry Hill (44.25°N, 124.13°W) (since 1989); for 12 years at provided by National Science Foundation Grants OCE-9907854, OCE-
Fogarty Creek (44.84°N, 124.06°W) and Seal Rock (44.50°N, 0435619, OCE-0453071, and OCE-0527168. Additional support for
124.10°W) (since 1994); for 11 years at Cape Meares (45.47°N, B.A.M. and J.L. was provided by the A. W. Mellon Foundation, the
Wayne and Gladys Valley Foundation, and the Robert and Betty
123.97°W) and Cape Arago (43.31°N, 124.40°W) (since 1995); for Lundeen Marine Biology Fund. This paper is Partnership for Interdis-
9 years at Yachats Beach (44.32°N, 124.11°W) (since 1997); and for ciplinary Studies of Coastal Oceans contribution 250 and contribution
8 years at Tokatee Klootchman (44.20°N, 124.12°W), Cape Blanco number 516 of the U.S. Global Ocean Ecosystem Dynamics program,
(42.84°N, 124.56°W), and Rocky Point (42.72°N, 124.47°W) (since jointly funded by the National Science Foundation and National Oceanic
1998). Mussel recruitment was quantified by using plastic mesh balls and Atmospheric Administration.
1. Pauly D, Christensen V (1995) Nature 374:255–257. 16. Menge BA, Branch GM (2001) in Marine Community Ecology, eds Bertness MD,
2. Wooster WS, Reid JL (1963) in The Sea, ed Hill MN (Wiley, New York), Vol 2, pp 253–280. Gaines SD, Hay ME (Sinauer, Sunderland, MA), pp 221–251.
3. Huyer A, Sobey EJC, Smith RL (1979) J Geophys Res 84:6995–7011. 17. Sydeman WJ, Bradley RW, Warzybok P, Abraham CL, Jahncke J, Hyrenbach KD,
4. Menge BA, Daley BA, Lubchenco J, Sanford E, Dahlhoff E, Halpin PM, Hudson G, Kousky V, Hipfner JM, Ohman MD (2006) Geophys Res Lett 33:L22S09.
Burnaford J (1999) Ecol Monogr 69:297–330. 18. Thomas AC, Brickley P (2006) Geophys Res Lett 33:L22S05.
5. Bakun A (1990) Science 247:198–201. 19. Leslie HM, Breck E, Chan F, Lubchenco J, Menge BA (2005) Proc Natl Acad Sci USA
6. Snyder MA, Sloan LC, Diffenbaugh NS, Bell JL (2003) Geophys Res Lett 30:1823. 102:10534–10539.
7. Menge BA, Daley BA, Wheeler PA, Dahlhoff E, Sanford E, Strub PT (1997) Proc 20. Menge BA (1995) Ecol Mon 65:21–74.
Natl Acad Sci USA 94:14530–14535. 21. Breaker LC, Liu PC, Torrence C (2001) Cont Shelf Res 21:727–750.
8. MacIsaac JJ, Dugdale RC, Barber RT, Blasco D, Packard TT (1985) Deep-Sea Res 22. Bane JM, Levine MD, Samelson RM, Haines SM, Meaux MF, Perlin N, Kosro PM,
32:503–529. Boyd T (2005) J Geophys Res 110:C10S02. .
9. Connolly SR, Menge BA, Roughgarden J (2001) Ecol 82:1799–1813. 23. Lott F, Robertson AW, Ghil M (2001) Geophys Res Lett 28:1207–1210.
10. Edwards M, Richardson AJ (2004) Nature 430:881–884. 24. Large WG, Pond S (1981) J Phys Oceanogr 11:324–336.
11. Beaugrand G, Brander KM, Lindley JA, Souissi S, Reid PC (2003) Nature 426:661–664. 25. Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha
12. Menge BA, Lubchenco J, Bracken MES, Chan F, Foley MM, Freidenburg TL, Gaines S, White G, Woollen J, et al. (1996) Bull Am Meteorol Soc 77:437–471.
SD, Hudson G, Krenz C, Leslie H, et al. (2003) Proc Natl Acad Sci USA 100:12229–12234. 26. Grasshoff K, Kremling K, Ehrhardt M (1999) Methods of Seawater Analysis (Wiley,
13. Strathmann MF (1987) Reproduction and Development of Marine Invertebrates of the Weinheim, Germany).
Northern Pacific Coast (Univ Wash Press, Seattle). 27. Welshmeyer NA (1994) Limnol Oceanogr 39:1985–1992.
14. Phillips NE (2002) Ecology 83:2562–2574. 28. Menge BA, Berlow EL, Blanchette CA, Navarrete SA, Yamada SB (1994) Ecol
15. Farrell TM, Bracher D, Roughgarden J (1991) Limnol Oceanogr 36:279–288. Monogr 64:249–286..
3724 www.pnas.org cgi doi 10.1073 pnas.0700462104 Barth et al.
ecosystems in the northern California current
John A. Barth*†, Bruce A. Menge‡, Jane Lubchenco†‡, Francis Chan‡, John M. Bane§, Anthony R. Kirincich*,
Margaret A. McManus¶, Karina J. Nielsen , Stephen D. Pierce*, and Libe Washburn**
*College of Oceanic and Atmospheric Sciences and ‡Department of Zoology, Oregon State University, Corvallis, OR 97331; §Department of Marine Sciences,
University of North Carolina, Chapel Hill, NC 27599; ¶Department of Oceanography, University of Hawaii, Manoa, HI 96822; Department of Biology,
Sonoma State University, Rohnert Park, CA 94928; and **Department of Geography, University of California, Santa Barbara, CA 93106
Contributed by Jane Lubchenco, January 22, 2007 (sent for review December 9, 2006)
Wind-driven coastal ocean upwelling supplies nutrients to the of upwelling-favorable winds might be influenced by climate vari-
euphotic zone near the coast. Nutrients fuel the growth of phy- ability is an area of active research (5, 6).
toplankton, the base of a very productive coastal marine ecosys- Spatial and temporal changes in nutrient availability and phyto-
tem [Pauly D, Christensen V (1995) Nature 374:255–257]. Because plankton biomass propagate up marine food webs and may strongly
nutrient supply and phytoplankton biomass in shelf waters are mediate the structure and dynamics of nearshore ecological com-
highly sensitive to variation in upwelling-driven circulation, shifts munities (7). Decreased upwelling can have two effects: reduced
in the timing and strength of upwelling may alter basic nutrient nutrient supply to phytoplankton (8) and reduced offshore trans-
and carbon fluxes through marine food webs. We show how a port of phytoplankton and planktonic fish and invertebrate larvae
1-month delay in the 2005 spring transition to upwelling-favorable that are crucial for replenishing coastal populations (9). Nutrient
wind stress in the northern California Current Large Marine Eco- and phytoplankton reductions may decrease zooplankton (includ-
system resulted in numerous anomalies: warm water, low nutrient ing larvae) survival, and thus decrease recruitment rates of plank-
levels, low primary productivity, and an unprecedented low re- totrophic larvae. Slower offshore transport could have the opposite
cruitment of rocky intertidal organisms. The delay was associated effect, with greater retention and increased recruitment rates.
with 20- to 40-day wind oscillations accompanying a southward
When intraseasonal variations in upwelling forcing are substantial,
shift of the jet stream. Early in the upwelling season (May–July) off
changes in the timing of phytoplankton blooms can potentially
decouple food availability from consumer demands (10). Such
Oregon, the cumulative upwelling-favorable wind stress was the
temporal trophic mismatches may have particularly important
lowest in 20 years, nearshore surface waters averaged 2°C warmer
effects on recruitment class strength in fish and invertebrates (11).
than normal, surf-zone chlorophyll-a and nutrients were 50% and
Recruitment patterns of rocky shore barnacles and mussels can
30% less than normal, respectively, and densities of recruits of
ECOLOGY
provide insight into the dynamics of inner-shelf ecosystems (12).
mussels and barnacles were reduced by 83% and 66%, respec- Sessile adult barnacles and mussels release larvae or gametes into
tively. Delayed early-season upwelling and stronger late-season the water that spend 2–4 weeks in the plankton, feeding on small
upwelling are consistent with predictions of the influence of global plankton for their sustenance (13). Larvae develop until they are
warming on coastal upwelling regions. competent to settle, then they must be transported back to shore if
they are to recruit and begin life as tiny sessile barnacles or mussels.
climate variability coastal marine ecosystems coastal ocean upwelling Growth rates, settlement sizes, and survival of mussel larvae, and
marine ecology probably barnacle larvae, depend on phytoplankton concentration
GEOPHYSICS
that, in turn, depends on nutrients provided by upwelling events
E quatorward winds along the eastern boundaries of the world’s
oceans drive offshore surface Ekman transport and the up-
welling of cold, nutrient-rich water into the euphotic zone near the
(14). Previous work has documented that both barnacles and
mussels recruit heavily but intermittently throughout summer in
Oregon (9). Recruitment events of barnacles in California (15) and
coast. These nutrient pulses stimulate high phytoplankton produc- Oregon†† often follow upwelling relaxation, i.e., times when larvae
tion, which, in turn, supports a rich coastal marine ecosystem and are transported shoreward.
productive fisheries (1). Examples of such dynamics include the Barnacles and mussels create 3D spatial structure (habitat) in
California Current, the Humboldt Current, the Benguela Current, rocky intertidal communities and are the primary food for many
and the Canary Current (2). predators (16). Barnacle and mussel juveniles and adults are
The strength and extent of the seasonal cycle in upwelling- abundant and relatively easy to monitor and have similar life
favorable winds varies along the U.S. west coast. In the northern histories to numerous other ecologically or economically important
pelagic and benthic inner-shelf species. Consequently, they are good
California Current Large Marine Ecosystem (CCLME), there is a
surrogates for many other inner-shelf species with respect to the
strong seasonal cycle with upwelling-favorable winds, the appear-
ance of cold, saline, nutrient-rich water near the coast, and equa-
torward currents over the shelf occurring after a spring transition Author contributions: J.A.B., B.A.M., J.L., F.C., M.A.M., and L.W. designed research; J.A.B.,
(3). Alongshore winds in the northern CCLME are more variable B.A.M., J.L., F.C., J.M.B., A.R.K., M.A.M., K.J.N., S.D.P., and L.W. performed research; J.A.B.,
than those farther south because they are more frequently influ- B.A.M., J.L., F.C., J.M.B., A.R.K., M.A.M., K.J.N., S.D.P., and L.W. analyzed data; and J.A.B.,
B.A.M., J.L., F.C., J.M.B., A.R.K., M.A.M., K.J.N., S.D.P., and L.W. wrote the paper.
enced by eastward-traveling Gulf of Alaska low-pressure systems.
The authors declare no conflict of interest.
The intermittent cessation of upwelling-favorable winds is called
Freely available online through the PNAS open access option.
relaxation and plays an important role in coastal circulation and the
recruitment of marine organisms††. The timing of the spring Abbreviations: CCLME, California Current Large Marine Ecosystem; chl-a, chlorophyll-a;
ISO, intraseasonal oscillation; NDBC, National Data Buoy Center; C-MAN, Coastal-Marine
transition and the total amount of upwelling-favorable winds during Automated Network.
the spring–summer upwelling season have a considerable impact on †To whom correspondence may be addressed. E-mail: barth@coas.oregonstate.edu or
coastal ecosystem responses. Farther south in the CCLME, winds lubchenco@oregonstate.edu.
are more persistently upwelling-favorable and the transition to a ††Grantham, B. A., Menge, B. A., Lubchenco, J. (2002) Eos Trans Am Geophys Union
more productive spring–summer season is less pronounced. The 83(Suppl):OS12K-03 (abstr.).
extent to which the timing of the spring transition and the intensity © 2007 by The National Academy of Sciences of the USA
www.pnas.org cgi doi 10.1073 pnas.0700462104 PNAS March 6, 2007 vol. 104 no. 10 3719 –3724
A
A
B
B
C
Fig. 2. Maps of 200-hPa surface height (m, blue contours) and wind speed at
300 hPa (m s 1, color shading), from six hourly National Centers for Environmen-
tal Prediction reanalyses (25) when the jet stream is located anomalously south
(May 20, 2005, 1200 Coordinated Universal Time; A) and in its more typical
summer position (July 19, 2005, 0000 Coordinated Universal Time; B). Wind
speeds 35 m s 1 are shown, with color increments at 45, 55, and 65 m s 1.
change. We use a long-term set of physical, chemical, and biological
measurements to examine the unusual presence of warm water, low
nutrients and low chlorophyll-a (chl-a) near the coast during 2005
that led to unprecedented low recruitment of rocky intertidal
organisms in the northern CCLME during the early upwelling
season.
D
Results and Discussion
Perturbations to Upwelling Regime. Year-to-year differences in the
timing of the spring transition and the total amount of seasonal
Fig. 1. Alongshore wind stress off the U.S. west coast. (A–C) Alongshore wind upwelling-favorable winds may have considerable impact on coastal
stress from three west-coast buoys: Newport, OR (A), Monterey Bay, CA (B), and
Point Conception, CA (C). Values for 2005 (blue) are plotted on top of the
ecosystem responses. Off central Oregon (44.62°N) where winds
climatological mean (black) 1 SD (gray shading) for 1985–2005. In A, the dashed have a pronounced seasonal cycle, anomalous winds in 2005 were
curve is the north–south position of the jet stream, and the two arrows indicate dominated by five strong, northward wind events, separated by
the times of maps shown in Fig. 2. (D) Cumulative alongshore (north–south) wind 20–40 days, during March–July (Fig. 1A). The onset of unusually
stress from National Oceanic and Atmospheric Administration NDBC Buoy 46050 strong, upwelling-favorable winds off central Oregon in early to
offshore of Newport, OR, starting from the spring transition. The black curve and mid-July coincided with a northward shift of 1,000 km in the
shading represent the mean 1 SD for 1985–2005, and the blue curve is for 2005. position of the jet stream (Fig. 1 A). A map of 200-hPa surface
At zero cumulative wind stress, the black curve and shading represent the mean
height during the mid-May northward wind event that occurred
and 1 SD of the date of the spring transition, and the blue line represents the
date of the 2005 spring transition. The 2005 curves are dark blue when absolute
before this northward shift shows a jet-stream position, and hence
values exceed any observed values during the previous 20 years. Data locations storm track, more typical of wintertime conditions (17) compared
are indicated at lower left. with a typical summertime jet-stream location in mid-July after the
shift (Fig. 2).
Off central California (37.36°N), 2005 winds were not unusual
impact of environmental changes in the inner shelf. There are few except for perhaps some stronger-than-normal southward winds
long-term, ecological time series available for the detection and from April to June (Fig. 1B). Off of southern California (34.72°N),
quantification of ocean ecosystem responses to environmental but outside the southern California Bight lee region, there were two
3720 www.pnas.org cgi doi 10.1073 pnas.0700462104 Barth et al.
A
B
C
Fig. 3. Surface (0 –2 m) temperatures during 2005 (solid lines) compared with climatological means (dashed lines) with 1 SD (shaded) from inner-shelf
moorings in 15 m of water off of central Oregon (44.25°N, 124.13°W) (1998 –2004 mean) (A), 21 m of water off of Monterey Bay, CA (36.97°N, 122.16°W)
(1999 –2004 mean) (B), and 15 m of water off of Santa Barbara, CA (34.46°N, 120.29°W) (1999 –2004) (C). Temperatures during 2005 that are warmer (colder) than
the climatological mean 1 SD are shaded in red (blue).
unusually strong northward wind events in February-March fol- transition off of Oregon contrasts strongly with the long-term
lowed by normal or slightly stronger upwelling-favorable winds the average and reveals the unusual timing of the upwelling-favorable
remainder of the year (Fig. 1C). winds (Fig. 1D). By examining wind and sea-level records, we
In 2005, the cumulative alongshore wind stress since the spring determined that the 2005 spring transition was on May 24, over a
ECOLOGY
A B C D 46
45
44
43
GEOPHYSICS
42
41
40
E F G H
Fig. 4. Surf-zone chlorophyll and nutrients measured along the Oregon and north/central California coasts. (A–D) Chl-a measured along the coast during
May-August 2005 (F), the long-term climatological mean (E) with 95% C.I. (bars), and 2005 anomalies from the mean (colored bars). (E–H) As in A–D, but for
nitrate plus nitrite (N N). The arrows at 44.25°N indicate the location of time series shown in Fig. 5.
Barth et al. PNAS March 6, 2007 vol. 104 no. 10 3721
month later than average and outside of the 1-SD range of the last A
20 years. The first substantial upwelling did not occur until late
June, 2 months later than average, but was terminated by the last
of the northward wind events in July 2005. Stronger-than-average
upwelling-favorable winds commenced in early to mid-July and
persisted until the seasonal accumulation of upwelling-favorable
winds reached the long-term average in September.
Effects on Temperature. Near-surface temperatures measured close
to shore illustrate the coastal ocean response to the unusual winds
of 2005 (Fig. 3). Off of central Oregon from May to mid-July 2005,
nearshore surface waters averaged 2°C warmer than normal, with
a maximum warming of 6.4°C (Fig. 3A). Upwelling-favorable winds
in mid-June (weak) and late in June (stronger) cooled upper-ocean
temperatures to nearly normal. However, after each of these
cooling periods, nearshore temperatures returned to above-average
levels, as warmer offshore waters were advected onshore by surface B
Ekman transport driven by northward wind events. In mid-July,
nearshore temperatures cooled to below normal as a result of
persistent and vigorous coastal upwelling. Off of central California,
nearshore surface temperatures were 2–3°C higher than normal
until early to mid-April, in part because of the northward wind event
in mid-March, after which temperatures returned to normal (Fig.
3B). Off of southern California, nearshore surface temperatures
were near normal, except for perhaps slight above-average tem-
perature in late February–early March in response to two north-
ward wind events (Fig. 3C).
Bottom-Up Effects. In 2005, coastwide negative chl-a anomalies were
evident during May, with the largest anomalies off of the central
Oregon coast (Fig. 4A). Depression of surf-zone chl-a continued
through July before reverting to neutral or increased levels in Fig. 5. Time series of surf-zone chlorophyll and nutrients off the central Oregon
August (Figs. 4A and 5). The May 2005 negative chl-a anomalies coast. (A) Time series of chl-a (circles) measured at the coast off of central Oregon
were accompanied by strong, coastwide decrease in nitrate con- (44.25°N): 2005 (filled symbols); 1993–2004 climatological mean (open symbols)
centration (Fig. 4B). Negative nitrate anomalies did not accompany with 95% C.I. (bars). (B) As in A, but for nitrate plus nitrite (diamonds).
chl-a reductions in June and July at the coastwide scale, although
persistent negative nitrate anomalies were evident from a number
of sites along the central and southern Oregon coasts (Figs. 4 and 5). The influence of a delayed transition to upwelling was most
dramatic in the northern CCLME (18) as might be expected
Effects on Recruitment. The 2-month delay in the onset of sub- because the seasonal cycle in wind stress is most pronounced in this
stantial upwelling in 2005 had serious consequences for recruitment region (Fig. 1 A). Because upwelling was abnormally low in the early
of mussels and barnacles. Mussel recruitment during May-August season, and because weak upwelling should lead to retention of
in 2005 was the lowest ever recorded for these 4 months at every site larvae close to shore, the unprecedented low early-season recruit-
(Fig. 6). During the next 2 months, September–October, after ment of mussels and barnacles was probably caused by low food
unusually intense and continuous upwelling, recruitment re- supply (i.e., low phytoplankton, a consequence of low nutrients)
bounded to values higher than average except for the two stations rather than by offshore loss of larvae caused by advection. By
in the extreme south and north of the measurement region. midsummer, however, upwelling-favorable winds were stronger
Barnacles and mussels often show different recruitment patterns than normal, resulting in higher phytoplankton biomass (Figs. 4 and
(8). Early in the season (May–July), barnacle (Balanus glandula) 5). During the late summer season, recruitment of both mussels and
recruitment was lower than normal at a number of, but not all, sites barnacles rebounded. Because spawning to settlement of these
(Fig. 6; see Methods for details of analysis). During August– organisms takes 2 weeks to 2 months (13), it is likely that these
October, after strong and persistent upwelling, B. glandula recruit- late-season recruits (August–November) were the result of repro-
ment was higher, indicating a rebound, at many, but not all, sites. duction that occurred shortly after the resumption of upwelling in
For both barnacles and mussels, recruitment in 2005 differed mid-July. Spatial differences in recruitment patterns may reflect
greatly between early (June–August for mussels, May–July for among-site differences in nearshore circulation.
barnacles) versus late (September–November for mussels, August–
October for barnacles) recruitment at most sites (Fig. 6). Consequences for the Marine Food Web. The ecological conse-
Summing results across all 10 sites indicates that in 2005 during quences of these changes are potentially severe. The low recruit-
June–August, mussel recruitment was reduced by 83%. During the ment of the intertidal species reported here is consistent with
next 3 months (September–November), mussel recruitment re- lower-than-normal concentrations of zooplankton (17) in the
bounded for an increase of 53% compared with normal years. northern CCLME. Reproductive failure of a planktivorous seabird,
Overall, however, total mussel recruitment for the entire 6-month Cassin’s auklet (Ptychoramphus aleuticus) in the CCLME during
recruitment season (June–November) in 2005 was normal. Barna- 2005 (17) is also a likely result of the chain of reductions at lower
cle recruitment in 2005 was depressed 66% during May–July and trophic levels in coastal food webs initiated by the reduced nutrients
was nearly normal (2.5% increase) during August–October, al- and phytoplankton at a critical time in the life history of coastal
though this late-season response was variable among sites. Total seabirds. Reduced phytoplankton is likely to slow growth and
barnacle recruitment for the 6-month recruitment season (May– suppress reproductive output of sessile invertebrates (19). Reduced
October) in 2005 was 38% less than normal. recruitment of sessile invertebrates could diminish the food supply
3722 www.pnas.org cgi doi 10.1073 pnas.0700462104 Barth et al.
ECOLOGY
Fig. 6. Mussel (Mytilus spp.) (Left) and barnacle (B. glandula) (Right) recruitment at sites along the Oregon coast: 2005 (F) compared with climatological
monthly means (E) computed for 8 –17 years (including 2005) depending on site. Means and SEM are shown for all data. Dotted vertical lines mark the ‘‘early’’
GEOPHYSICS
and ‘‘late’’ recruitment seasons used in data analysis. The absence of dotted lines indicates sites where no differences between years or season occurred.
for predators such as whelks, sea stars, crabs, and shorebirds, with In experiments using a regional climate model, albeit in the
potentially far-reaching consequences elsewhere in the food absence of an active ocean submodel, increased greenhouse gas
web (20). forcing results in a 1-month delay in the onset of the coastal
upwelling season in the northern CCLME and an increase in
Intraseasonal Oscillations (ISOs). Recent studies have shed light on upwelling intensity later in the season (June–September) (6). The
the importance of ISOs of the wind (20- to 40-day periods) on mechanism by which this happens, changes in radiative forcing, is
coastal upwelling (21, 22) and coastal ecosystem dynamics (J.M.B., different from the dynamics described above involving wind ISOs
Y. Spitz, R. Letelier, and W. T. Peterson, unpublished work). The and an anomalously southward jet-stream position that led to
five large wind ISOs played a key role in delaying the spring delayed upwelling. Sorting out the relative roles of these atmo-
transition in the northern CCLME in 2005. Wind ISOs at midlati- spheric processes is key to our ability to predict future changes in
tude are formed through interactions of the midlatitude jet stream coastal ocean ecosystems caused by climate variability and climate
with large-scale topography (23). During the summer of 2001, ISOs change.
in alongshore wind stress off of Oregon (44.6°N) correlated well
with the north–south position of the jet stream (22). The cessation Methods
of strong wind ISOs and the onset of unusually strong upwelling- Meteorological Data. Wind stress was calculated by using measure-
favorable wind in the northern CCLME in early to mid-July 2005 ments from the National Oceanic and Atmospheric Administration
were consistent with a shift of the jet stream from the south to the National Data Buoy Center (NDBC) buoys and Coastal-Marine
north of this region (Figs. 1 A and 2). Automated Network (C-MAN) stations (24) and then low-pass-
Delayed early-season upwelling and stronger late-season up- filtered with a filter with a 40-h width at half-amplitude to remove
welling documented here for the northern CCLME during 2005 are short-period (e.g., diurnal) fluctuations. The alongshore wind stress
consistent with predictions of the influence of global warming on was computed by rotating into a coordinate system aligned with the
coastal upwelling regions (5, 6). The global-warming scenario relies local coastline (rotation angle indicated for each site below). Winds
on increased land–sea temperature contrasts, created by preferen- were measured at NDBC buoy 46050 off of central Oregon
tial heating of the land caused by elevated greenhouse gases, driving (44.62°N, 124.53°W; 3°), buoy 46012 off of Monterey, CA (37.36°N,
stronger equatorward winds, and hence stronger upwelling. 122.88°W; 327°), and buoy 46023 off of Point Arguello, CA
Barth et al. PNAS March 6, 2007 vol. 104 no. 10 3723
(34.72°N, 120.97°W; 329°). Gaps in the NDBC buoy records were (Tuffys; The Clorox Company, Oakland, CA), and barnacle re-
filled through regression with nearby buoys and/or C-MAN sta- cruitment was quantified by using settlement plates [0.10 0.10
tions: 46050 with Newport, OR, C-MAN (NWPO3) and occasion- 0.004-m poly(vinyl chloride) plates with Saf-T-Walk (3M Company,
ally the Cape Arago, OR, C-MAN (CARO3); 46012 with buoys St. Paul, MN), a textured plastic tape, on the top side] (4, 28).
46026 and 46042; 46023 with buoy 46011 and Point Conception, Collectors were deployed/recovered monthly and processed in the
CA, C-MAN (PTGC1). A 5-day running mean was then applied to laboratory where counts of mussels (number per collector) and
all time series. To assess the cumulative alongshore wind stress off barnacles (number per 10 2 m2) were made under dissecting
of central Oregon during the upwelling season, north–south stress microscopes. Barnacle recruits were separated by species with B.
was summed starting from the spring transition, defined as when glandula and Chthamalus dalli as the most abundant barnacles for
winds turn to predominantly upwelling-favorable (southward) usu- which we have long-term recruitment data. Mussel species cannot
ally during March to April (3). be reliably identified visually. Judging from the species composition
The latitudinal position of the jet stream in the northern CCLME of mussels growing to identifiable sizes on the shore after recruit-
was taken as the location, along 125°W longitude, of the strongest ment, most mussel recruits are Mytilus trossulus, but very small
horizontal gradient in the height of the 200-hPa surface (22), numbers of Mytilus californianus also settle jointly with Mytilus
determined from six hourly National Centers for Environmental trossulus.
Prediction reanalysis pressure maps (25). The jet-stream position Because we were interested in comparing among-year and with-
time series was then filtered with a 35-d cosine low-pass filter. in-season differences, recruitment data were analyzed by using
two-way ANOVA. Factors tested were year (2005 vs. the full data
Mooring-Based Measurements. Temperature was measured at 1- to set) and recruitment season (early and late). Both factors were
2-min intervals just below the surface (0–3 m) by using StowAway fixed. To improve the fit of the data to a normal distribution, we
XTI Temperature Loggers (Onset Computer Corp.) on moorings transformed the response variable to ln (recruits per day 1); we
deployed in water depths of 15 m (central Oregon, 44.25°N, used P 0.05 as our level of significance. Sample size (n) in these
124.13°W; Santa Barbara, CA, 34.46°N, 120.29°W) or 21 m analyses varied among sites from 181 to 517 (mussels) and 184 to
(Monterey, CA, 36.97°N, 122.16°W). 571 (barnacles) and is based on five to eight recruit collector
samples for each species per month per site for the time periods
Coastal Transects. Surf-zone chl-a and inorganic nutrient measure- listed above. Sample size varied because of varying numbers of
ments (nitrate plus nitrite) were made between 2001 and 2005 years in which sampling was done and occasional losses of collec-
during May and June and between 1997 and 2005 during July and tors. We report P values for interactions (year season), which
August, at 30 sites along the Oregon and north/central California subsumes main effects when significant, or main effects (year or
coasts (46–38°N). Once a month from May through August, season) when interactions were not significant (Fig. 6). For barna-
synoptic measurements were all made on the same day along the cles, analysis of data using May–July and August–October provided
coast between 3 h before and 3 h after low tide. In central Oregon the clearest contrast between early and late recruitment seasons,
(44.25°N), additional data were collected more frequently (daily to respectively (except for Cape Arago and Cape Blanco). For mus-
biweekly) between 1993 and 2005. Anomalies were computed by sels, analysis of data using June–August and September–November
differencing values from 2005 with a long-term mean for each provided the clearest contrast between early and late recruitment
month: May–June (2001–2005) and July–August (1997–2005). seasons (except for Rocky Point) The variation in which season was
Chl-a and nitrate plus nitrite measurements were made according analyzed is shown by the vertical dotted lines in Fig. 6.
to standard methods (26, 27).
We thank the many scientists, technicians, and students of the Partner-
ship for Interdisciplinary Studies of Coastal Oceans who were involved
Mussel and Barnacle Recruitment. Monthly recruitment (the appear- with the collection of the long-term data sets and who made this study
ance of new young on rocky shores) of mussels and barnacles has possible. The Partnership for Interdisciplinary Studies of Coastal Oceans
been measured monthly at 10 sites spanning two-thirds of the is supported by the David and Lucile Packard Foundation and the
Oregon coast: for 17 years at Boiler Bay (44.83°N, 124.06°W) and Gordon and Betty Moore Foundation. Additional support for J.A.B. was
Strawberry Hill (44.25°N, 124.13°W) (since 1989); for 12 years at provided by National Science Foundation Grants OCE-9907854, OCE-
Fogarty Creek (44.84°N, 124.06°W) and Seal Rock (44.50°N, 0435619, OCE-0453071, and OCE-0527168. Additional support for
124.10°W) (since 1994); for 11 years at Cape Meares (45.47°N, B.A.M. and J.L. was provided by the A. W. Mellon Foundation, the
Wayne and Gladys Valley Foundation, and the Robert and Betty
123.97°W) and Cape Arago (43.31°N, 124.40°W) (since 1995); for Lundeen Marine Biology Fund. This paper is Partnership for Interdis-
9 years at Yachats Beach (44.32°N, 124.11°W) (since 1997); and for ciplinary Studies of Coastal Oceans contribution 250 and contribution
8 years at Tokatee Klootchman (44.20°N, 124.12°W), Cape Blanco number 516 of the U.S. Global Ocean Ecosystem Dynamics program,
(42.84°N, 124.56°W), and Rocky Point (42.72°N, 124.47°W) (since jointly funded by the National Science Foundation and National Oceanic
1998). Mussel recruitment was quantified by using plastic mesh balls and Atmospheric Administration.
1. Pauly D, Christensen V (1995) Nature 374:255–257. 16. Menge BA, Branch GM (2001) in Marine Community Ecology, eds Bertness MD,
2. Wooster WS, Reid JL (1963) in The Sea, ed Hill MN (Wiley, New York), Vol 2, pp 253–280. Gaines SD, Hay ME (Sinauer, Sunderland, MA), pp 221–251.
3. Huyer A, Sobey EJC, Smith RL (1979) J Geophys Res 84:6995–7011. 17. Sydeman WJ, Bradley RW, Warzybok P, Abraham CL, Jahncke J, Hyrenbach KD,
4. Menge BA, Daley BA, Lubchenco J, Sanford E, Dahlhoff E, Halpin PM, Hudson G, Kousky V, Hipfner JM, Ohman MD (2006) Geophys Res Lett 33:L22S09.
Burnaford J (1999) Ecol Monogr 69:297–330. 18. Thomas AC, Brickley P (2006) Geophys Res Lett 33:L22S05.
5. Bakun A (1990) Science 247:198–201. 19. Leslie HM, Breck E, Chan F, Lubchenco J, Menge BA (2005) Proc Natl Acad Sci USA
6. Snyder MA, Sloan LC, Diffenbaugh NS, Bell JL (2003) Geophys Res Lett 30:1823. 102:10534–10539.
7. Menge BA, Daley BA, Wheeler PA, Dahlhoff E, Sanford E, Strub PT (1997) Proc 20. Menge BA (1995) Ecol Mon 65:21–74.
Natl Acad Sci USA 94:14530–14535. 21. Breaker LC, Liu PC, Torrence C (2001) Cont Shelf Res 21:727–750.
8. MacIsaac JJ, Dugdale RC, Barber RT, Blasco D, Packard TT (1985) Deep-Sea Res 22. Bane JM, Levine MD, Samelson RM, Haines SM, Meaux MF, Perlin N, Kosro PM,
32:503–529. Boyd T (2005) J Geophys Res 110:C10S02. .
9. Connolly SR, Menge BA, Roughgarden J (2001) Ecol 82:1799–1813. 23. Lott F, Robertson AW, Ghil M (2001) Geophys Res Lett 28:1207–1210.
10. Edwards M, Richardson AJ (2004) Nature 430:881–884. 24. Large WG, Pond S (1981) J Phys Oceanogr 11:324–336.
11. Beaugrand G, Brander KM, Lindley JA, Souissi S, Reid PC (2003) Nature 426:661–664. 25. Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha
12. Menge BA, Lubchenco J, Bracken MES, Chan F, Foley MM, Freidenburg TL, Gaines S, White G, Woollen J, et al. (1996) Bull Am Meteorol Soc 77:437–471.
SD, Hudson G, Krenz C, Leslie H, et al. (2003) Proc Natl Acad Sci USA 100:12229–12234. 26. Grasshoff K, Kremling K, Ehrhardt M (1999) Methods of Seawater Analysis (Wiley,
13. Strathmann MF (1987) Reproduction and Development of Marine Invertebrates of the Weinheim, Germany).
Northern Pacific Coast (Univ Wash Press, Seattle). 27. Welshmeyer NA (1994) Limnol Oceanogr 39:1985–1992.
14. Phillips NE (2002) Ecology 83:2562–2574. 28. Menge BA, Berlow EL, Blanchette CA, Navarrete SA, Yamada SB (1994) Ecol
15. Farrell TM, Bracher D, Roughgarden J (1991) Limnol Oceanogr 36:279–288. Monogr 64:249–286..
3724 www.pnas.org cgi doi 10.1073 pnas.0700462104 Barth et al.