Forgot your password?
Document Actions
Schiel et al 04
Ten Years of Induced Ocean Warming Causes Comprehensive Changes in Marine Benthic
Communities
Author(s): David R. Schiel, John R. Steinbeck, Michael S. Foster
Source: Ecology, Vol. 85, No. 7 (Jul., 2004), pp. 1833-1839
Published by: Ecological Society of America
Stable URL: http://www.jstor.org/stable/3450356
Accessed: 25/01/2010 19:04
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=esa.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.
http://www.jstor.org
REPORTS
Ecology, 85(7), 2004, pp. 1833-1839
? 2004 by the Ecological Society of America
TEN YEARS OF INDUCED OCEAN WARMING CAUSES COMPREHENSIVE
CHANGES IN MARINE BENTHIC COMMUNITIES
DAVID R. SCHIEL,'4 JOHN R. STEINBECK,2 AND MICHAEL S. FOSTER3
'Zoology Department, University of Canterbury, Christchurch, New Zealand
2Tenera Environmental, 225 Prado Road, Suite D, San Luis Obispo, California 93401 USA
3Moss Landing Marine Laboratories, 8272 Moss Landing Road, Moss Landing, California 95039 USA
Abstract. One of the most commonly predicted effects of global ocean warming on
marine communities is a poleward shift in the distributional boundaries of species with an
associated replacement of cold-water species by warm-water species. However, these types
of predictions are imprecise and based largely on broad correlations in uncontrolled studies
that examine changes in the distribution or abundances of species in relation to seawater
temperature. Our study used an 18-year sampling program in intertidal and subtidal habitats
and before-after, control-impact analyses. We show that a 3.5?C rise in seawater temper-
ature, induced by the thermal outfall of a power-generating station, over 10 years along 2
km of rocky coastline in California resulted in significant community-wide changes in 150
species of algae and invertebrates relative to adjacent control areas experiencing natural
temperatures. Contrary to predictions based on current biogeographic models, there was
no trend toward warmer-water species with southern geographic affinities replacing colder-
water species with northern affinities. Instead, the communities were greatly altered in
apparently cascading responses to changes in abundance of several key taxa, particularly
habitat-forming subtidal kelps and intertidal foliose red algae. Many temperature-sensitive
algae decreased greatly in abundance, whereas many invertebrate grazers increased. The
responses of these benthic communities to ocean warming were mostly unpredicted and
strongly coupled to direct effects of temperature on key taxa and indirect effects operating
through ecological interactions.
Key words: BACI analysis; benthic community structure; biogeographic distribution information;
climate change, predicting effects; Diablo Cove, California, USA; ocean warming; rocky reef; tem-
perature change; thermal discharge of a power plant, effects.
INTRODUCTION between natural changes in seawater temperature and
Ocean warming through climate change is expected species' occurrences or abundances (Lubchenco et al.
to have significant effects on the composition and struc- 1993), or else on point counts separated by large time
ture of marine communities (Fields et al. 1993). These periods (e.g., Sagarin et al. 1999). Few empirical stud-
include changes in distributional boundaries of species ies incorporate significant changes in seawater tem-
and replacement of cold-water taxa by others with perature over periods long enough to span the lives of
warm-water affinities. Paleoclimatic records showing most species, while also having control areas and con-
shifts in the distribution or abundance of species as- tinuous time series of sampling that allow specific spa-
sociated with changes in temperature (Kennett and tial and temporal trends in community structure to be
Stott 1991), localized changes in species abundances detected and tested.
over decades (Southward et al. 1995, Holbrook et al. Here, we address the hypothesis that elevated sea-
1997, Sagarin et al. 1999), and changes associated with water temperature brought about by the thermal dis-
El Ninio-Southern Oscillation events (Pearcy and charge of a power generating station in California un-
Schoener 1987, Dayton et al. 1999) support such sce- derlay changes to rocky intertidal and subtidal com-
narios. However, evidence for changes due to ocean munities. Eighteen years of sampling before and after
warming is limited and based largely on correlations the thermal discharge began, and the simultaneous sam-
pling of nearby control areas that experienced only
Manuscript received28 July2003; revised1 February2004; natural
3
accepted February 2004;finalversion received10 March
2004. temperatures, allowed a robust assessment of
Corresponding Editor: G. Morgan.
S. changes in species' abundances through time. The sus-
4 E-mail: tained elevation of seawater temperature enabled an
david.schiel@canterbury.ac.nz
1833
July 2004
Volume 85 No. 7 EC UBLICATION O LOGICAL SOCIETY OF AMERICA
A PUBLICATION OF THE ECOLOGICAL SOCIETY OF AMERICA
The Robert H. MacArthurAward Lecture
Toward metabolictheoryof ecology
a
Forum
Themetabolictheoryof ecology
Reports
Tenyearsof inducedocean warmingcauses comprehensivechanges
in marinebenthiccommunities
Articles
withdifferent
Responses preyfromhabitats
of predator and
regimes:localadaptation heritability
1834 DAVID R. SCHIELET AL. Ecology,Vol. 85, No. 7
exceptionally strong and direct test of responses by 18- Seawater temperature
nearshore benthic marine communities to ocean warm- -
k , AAMAt.I\krWAMb
ing. Furthermore, the study area was close to a source 0 14 IJ i'
I
,,
' 4tX 1 ^
of potential invaders because it is only 138 km north I-
o
of Point Conception, a recognized boundary between I--,
0
a warm-temperate biota to the south and a cool-tem-
perate biota to the north (Ricketts et al. 1985). a) 6-
QC
E
2- Delta temperature
METHODS
Diablo Cove is on an exposed rocky coast in central An -
_ A- - A.- --_
_I
California, USA (35?12' N, 120?52' W), has a shoreline . , ; I I
, Ii I i I I , I I I I
cO o 0 C 't o
) 0 NC\
-2 km long, a surface area of 15 ha, an average depth I Ir oo
O oo
O oo
O oo o a
O o,) o
) oC C C C) C c) n C)
of 8 m, and a maximum depth of 18 m. The nearby
power plant draws seawater from outside the cove, uses FIG. 1. Monthly mean temperatures control (dashed
at
it in a cooling system to re-condense freshwater steam, line) and impact (solid line) sites measuredat a depth of 3
and discharges it into the cove at a maximum rate of m below MLLWfrom 1976 to 1995. Delta temperatures are
9.5 X 109 L/d, creating a thermal plume that spreads the averagedifferencesbetween controland impactstations.
The vertical line at 1985 shows when thermalchangebegan
throughout the cove. The plant has operated at mostly within Diablo Cove.
full power since May 1985 and very little is discharged
other than heated water. Required regular water mon-
itoring and bioassays showed that concentrations of and "impact" areas was used to detect significant ther-
other constituents were very low relative to the large mal effects (Stewart-Oaten et al. 1986). The "before"
volume of the discharge and did not present acute or period was January 1978 through December 1984 and
chronic toxicity to marine larval organisms used in the the "after" period was January 1987 through June
tests. Temperature was recorded continuously (every 1995. Surveys from January 1985 through December
20 min) from 1976 with digital recorders (+0.2?C ac- 1986 were omitted to allow time for thermal effects to
curacy) in metal canisters at the -3 m MLLW level (3 occur. The high power to detect small changes in abun-
m below mean lowest low water) of several stations. dance was due to the number of replicate surveys in
Biological surveys were done six times each year each period. The analytical variable was the difference
from 1978 through 1988 and four times annually from between the mean of each impact station and the grand
1989 through 1995. There were four intertidal and five mean of all the control stations in each survey. The
subtidal stations within Diablo Cove and two subtidal differences between these paired means ("deltas")
and three intertidal control stations 1-2 km outside the were then compared statistically before and after the
cove beyond the influence of the discharge. Each in- putative impact. Data for each species were tested for
tertidal station had two 30 m fixed transects parallel to additivity, linear trend, serial correlation, and homo-
the shore, one each at the +0.3 m and +0.9 m MLLW geneity of variance, transformed appropriately to meet
tidal levels. Ten 1-m2 quadrats (initially chosen ran- test assumptions, and then analyzed using a two-way.
domly) per transect were sampled during each survey. ANOVA with factors for "period" and "station." If
All visible algae and invertebrates were counted or their the interaction was not significant, "period" was used
percentage cover assessed. At each subtidal station, a to test for effects of thermal change. A significant in-
fixed 28-m2 circular plot divided into four quadrants at teraction indicated an impact of varying magnitude
a depth of 3 m to 4 m was sampled on each occasion. among stations or no impact due to variable results
In each quadrant, understory algal and attached inver- among stations. For taxa that tested significant, we cal-
tebrate cover was estimated from 50 random points, culated percentage change in abundance in Diablo
and invertebrates and kelp were counted. The great Cove relative to controls as: Relative percentage
- - - -
majority of taxa were identified to species level. Fol- change = ([(XB, XBi) (XA XAi)]/(XA XAi)) X
lowing Sagarin et al. (1999), individual species were =
100, where XBC mean abundance at the pooled control
classified as "cosmopolitan" (i.e., with an extensive sites before thermal increase; XBi = mean abundance
north-south distribution along the west coast), "north- at the pooled Diablo Cove sites before thermal increase;
ern" (with a southern limit north of Point Conception) XAC = mean abundance at the pooled control sites after
or "southern" (with a northern limit south of Cape thermal increase; XAi = mean abundance at the pooled
Mendocino) based on information in Morris et al. Diablo Cove sites after thermal increase.
(1980) and Abbott and Hollenberg (1976). Taxa were
RESULTS
excluded from this classification if they were identified
only to genus and may have contained species of dif- Seawater temperatures within and outside Diablo
ferent geographical affinities. Cove (California, USA) were similar from 1976 to
BACI (before-after, control-impact) analysis of var- 1985, when the heated discharge began (Fig. 1). How-
iance using "before" and "after" data from "control" ever, from 1986 through 1995 they were elevated by
July 2004 COMMUNITYCHANGESAFTER OCEAN WARMING 1835
TABLE 1. The numberof species and the percentageof all taxa combinedthat significantly
increased(Inc.), decreased(Dec.), or did not change (Nc) in abundance within Diablo Cove
(California,USA) relative to control stations outside the cove over the 10 years after the
rise in seawatertemperature.
Cosmopolitan Northern Southern
Species types Inc. Dec. Nc Inc. Dec. Nc Inc. Dec. Nc
Intertidal algae 1 15 1 0 1 0 0 1 0
Intertidal invertebrates 11 8 1 1 0 1 2 0 2
Subtidalalgae 6 6 2 1 1 0 1 0 0
Subtidalinvertebrates 11 16 9 0 1 0 6 3 1
All taxa (%) 26.6 41.3 11.9 1.8 2.8 0.9 8.3 3.7 2.8
Notes: Species werecategorizedas cosmopolitan,northern, southern theirbiogeographic
or in
affinities.Significancewas based on BACI analyses (see Methods).
an average of 3.5?C inside the cove, with monthly were sufficiently abundant to be analyzed statistically.
means ranging from 2.0 to 5.1?C above ambient. Max- Of these, 150 species (87%) changed significantly in
imum temperatures occurred during the El Ninio- abundance at stations with elevated temperatures rel-
Southern oscillation (ENSO) years of 1983, 1987, and ative to control stations, with 65 species (37.8%) in-
1992-1993. The coast in and around Diablo Cove ini- creasing and 85 species (49.4%) decreasing.
tially had marine communities typical of other semi- The biogeographic affinities of 109 of the 172 spe-
protected sites in the region (Ricketts et al. 1985). In- cies could be assigned with confidence (see Appendix).
tertidal areas were dominated by foliose (e.g., Maz- There was no consistent pattern of change in abun-
zaella flaccida) and branched (e.g., Chondracanthus dances related to geographic affinities (Table 1). Most
canaliculatus) red algae, with some grazing gastropods intertidal algae (89%) were cosmopolitan. Of these, 15
and sessile invertebrates in the understory. Subtidal
species decreased in abundance, as did the one northern
areas had a canopy of northern bull kelp, Nereocystis and one southern algal species. Twenty of the 26 in-
luetkeana, an understory dominated by the kelps Pter- tertidal invertebrate species (77%) were cosmopolitan,
ygophora californica and Laminaria setchellii, and pri- and 11 species increased, eight species decreased, and
mary cover of coralline and foliose red algae, with a one species remained unchanged relative to controls.
lush assemblage of smaller algae and invertebrates in- The few northern and southern species had equal num-
terspersed. bers increasing or showing no change. Most subtidal
During our 18-year study, 714 taxa were counted or algae (82%) were cosmopolitan, with six species in-
assessed for percentage cover, but only 172 species
creasing, six species decreasing, and two species re-
maining unchanged in abundance. One northern species
A) Algae increased and one decreased, while the only southern
20-
species increased. Thirty-six of the 47 subtidal inver-
15- tebrates (77%) were cosmopolitan, of which 11 species
increased, 16 species decreased, and nine species had
10 - no change. The one northern species decreased, while
six southern species increased, three species decreased,
and one species remained unchanged. Overall, of the
n
O
IC
a) AiLiL..
C, 5- . F]I
I I I I I I I I I I I I I I .... I I I I I I
71111n 11nnn
80% of species that were cosmopolitan, far more de-
creased (41.3%) than increased or remained un-
j 25- B) Invertebrates T changed. Of the 6% of species that were northern, more
aG 20- decreased (2.8%) than increased (1.8%). And of the
a r,
southern 15% of species, more increased (8.3%) than
15- decreased (3.7%).
'?:
10- Most species within Diablo Cove underwent large
'7,;
relative changes in abundance in the 10 years following
5-
n-
L
I .I.._.
I1 I I I
| I | *
7,
I
n
I I
flrr-
seawater warming. Fifty-four percent of algae (Fig. 2A)
and 27% of invertebrates (Fig. 2B) declined by at least
-100-80 -60 -40 -20 0 1 2 3 4 5 50%. Dominant algae (with an initial combined cover
Decrease (%) Log(% increase) >60%) declined dramatically, especially the cosmo-
FIG.2. Percentage politan red algae Mastocarpus papillatus (-65%) in
changesin abundance (A) algae and the
of
(B) invertebratespecies relative to controls after thermal upper zone, Endocladia muricata (-63%), Maz-
change, shown as decreases(solid bars),no change (mottled zaella flaccida (-98%), and Chondracanthus canali-
bars), and log,0 increases (open bars). culatus (-51%). Algal taxa richness declined by 40%.
1836 DAVID R. SCHIELET AL. Ecology, Vol. 85, No. 7
In contrast, non-calcified algal crusts increased by 62%.
100
The few other increases were in cosmopolitan taxa that,
taken together, comprised 1 % of total algal cover. 90
Subtidal understory kelps declined by 82% from an 80
average of 7.3 to 1.3 plants/m2. The initially abundant 70
cosmopolitan kelps P. californica and L. setchellii de- 60
clined by -90%, but a cosmopolitan understory red
50
alga, Cryptopleura violacea, increased by 3000% (i.e.,
cm
E
a log percentage increase of 3.47) to cover 30-50% of
the substratum. Some of the most important changes a)
N
involved the large kelps. N. luetkeana declined by 97% g d
in the cove and was replaced by the initially rare south- o 0)
ern giant kelp Macrocystis pyrifera, which dominated 6
c
the surface canopy with an average density of 1.5 ()
0
plants/7 m2 after 10 years. a) c
Intertidal grazers became very abundant, especially (0
C -0
the cosmopolitan Tegula funebralis, which increased
by 79% to a 10-year average of 175 individuals/m2 on a)
c
the upper shore, and the cosmopolitan limpet Macclin-
co
tockia scabra that increased by >400% to 86 individ- 0)
uals/m2. Some initially rare species became common,
including limpets of the cosmopolitan genus Lottia,
which increased from -3 to 9 limpets/m2, and the cos-
mopolitan keyhole limpet Fissurella volcano, which
increased by >700% to 13 limpets/m2 on the lower
shore. The cosmopolitan sea urchin Strongylocentrotus
purpuratus increased by 3000% to 8.2 individuals/m2
in the low intertidal zone. 0 o cmj 1 cO 0oo c'
N 't C
Almost all changes were in the abundances of species 0 0 0 o0 COcO 0 0 0
0)0)0 0)0)0)CD0)0)0)
already present. A few new invertebrates appeared in
impact sites after thermal change, but they were never FIG.3. Annualmeanpercentcover of foliose algae (open
of
squares;left-handaxis) and abundance grazersper square
abundant, such as brittle stars Ophiothrix spp. (prob- zone of
meter(solid circles; right-handaxis) in the intertidal
ably southern; from 0 to 1.1 individuals/0.25 m2), Aply- (A) control sites and (B) impactsites.
sia californica (cosmopolitan; from 0 to 0.2 individ-
uals/7 m2), Cryptochiton stelleri (cosmopolitan; from
0 to 0.1 individual/7 m2), and Norrisia norrisi (south- trol areas (7400%, 350%, and 49000%, respectively),
ern; from 0 to 0.1 individual/7 m2). their overall abundances remained low. The total cover
Intertidal habitats were greatly modified in Diablo of all sessile invertebrates increased from -2.0-4.7%
Cove, especially in the vertical layering of algae. All in the intertidal zone of Diablo Cove sites. There were
sites initially had a dense cover (70-80%) of foliose also slight increases of predatory gastropods such as
algae up to 10 cm thick, with grazing invertebrates Acanthina spp. and Ocenebra spp. on the upper shore,
interspersed below. At control sites these algae de- but these were not statistically significant.
clined around 1985 and grazers increased, most likely Community changes in subtidal habitats were more
as delayed responses to the severe 1983 ENSO (Fig. complex. Control sites (Fig. 4A) had a primary sub-
3A). However, dense algae returned within two years stratum cover of foliose algae from 1978 to 1995 but
and by 1995 both algal and grazer abundances were understory kelps underwent large changes in abun-
virtually the same as in 1978. At the thermally affected dance, ranging from 20 to 225 kelps/m2. Kelps had high
sites, foliose algae also declined in 1985 but did not recruitment after the 1983 and 1987 ENSO events but
recover (Fig. 3B). Instead, grazers dominated for 10 then gradually declined. Grazers fluctuated in abun-
years and, as might be predicted given their high den- dance. By the end of the study, however, the control
sities (cf. Lubchenco and Cubit 1980), this was accom- sites looked remarkably similar to their initial condi-
panied by large increases in bare substratum, low-lying tion. The thermally affected sites had far greater chang-
ephemeral turf, and thin algal crusts. Other changes, es (Fig. 4B). Bull kelp was replaced by giant kelp,
especially at higher tidal levels, were increases in subcanopy kelps declined and had no major episodes
space-occupying species such as barnacles, Chthama- of recruitment, and grazers fluctuated and then declined
lus fissus and Tetraclita rubescens, and the sand tube after 1988 as dense foliose algae dominated the sub-
worm Phragmatopoma californica. Although these stratum. In contrast to control sites, the grazing assem-
species had large percentage increases relative to con- blage changed after 1985 when sea urchins began in-
July 2004 COMMUNITY CHANGES AFTER OCEAN WARMING 1837
creasing in abundance while gastropods declined (Fig.
4C). Furthermore, only within Diablo Cove did the
dominant benthic algae change (Fig. 4D). The cos-
a)
E
mopolitan red alga Cryptopleura ruprechtiana was re-
0 6
O placed by its cosmopolitan congener C. violacea.
a)
a) 0 DISCUSSION
Ca)
D. c"- Key results were that after a sustained rise in sea-
water temperature the great majority of nearshore spe-
cies changed in abundance and whole communities
changed in character. Mechanisms of change are usu-
o ally complex, involve the interaction of many factors,
and, as abundances change, can be highly sensitive to
140 the strengths of interactions among species (Menge
2000). The level of experimentation (i.e., bay-wide) in
120
this study limits the resolution of the processes re-
a)
100 sponsible for change and, therefore, the sequence of
0
a)
O (I-
a) observed changes and their consistency give the best
80
0 clues for evaluating the processes responsible.
60 The deterioration of algae clearly preceded increases
0
a)
EL ~-. in grazers. Dominant algae, especially Mazzaellaflac-
40 t:
c. cida, Nereocystis, Pterygophora, and Laminaria, had
20 -0 acute responses to the increased temperature and de-
clined rapidly, becoming bleached with deteriorating
0
thalli within a few months. This provided new settle-
ment space and an altered light environment to primary
120 substrata (Reed and Foster 1984) and may have trig-
CM
gered a series of responses such as altered mortality,
E 100
growth, and reproductive and recruitment rates of other
d 80 species (Ives and Gilchrist 1993), behavioral changes
c relating to temperature preferences and avoidance
c
60 -
C\ (Schroeter et al. 1993, Sanford 1999), and modified
(I-
food webs and species interactions (Wootton 1993).
40
a Initially dominant species had little or no subsequent
20 recruitment, even during the periods when grazer num-
(D bers declined (e.g., 1988).
0 In contrast, grazing gastropods showed a positive
response to temperature, which most likely involved
80 physiological tolerances, trophic responses, availabil-
ity of space and recruitment dynamics. Many intertidal
70
molluscs are highly resistant to the extreme tempera-
60 tures experienced during low tide and can repair ther-
50 mal damage to proteins (Tomanek and Somero 1999).
40 Their strong recruitment intertidally was most likely a
a) direct consequence of having more free space to settle
0 30
a) after the decline of algae. Once they reached the very
20
high densities seen in Diablo Cove (up to 370 molluscs/
10 m2), they would have precluded effective recruitment
of macroalgae for much of the post-impact period (Fos-
CO
N- CM C <D D CD C\ C C ter 1982). Grazers often recruit into areas cleared of
algae along the west coast (Foster et al. 2003), as was
FIG. 4. Annual mean percent cover of foliose algae (left- seen in all sites after the 1983 ENSO. Given the quick
hand axes) and numbers of subcanopy kelps and grazers recovery of control sites, however, it is likely that the
(right-hand axes) at (A) subtidal control sites and (B) subtidal
trophic response of grazing mollusks in the impacted
impact sites, with (C) a separation of grazers into gastropods
and sea urchins, and (D) the percent cover of the two most areas was secondary to the direct effects of temperature
abundant benthic foliose algae (Cryptopleura spp.). Percent on algae.
cover could total >100% because of layering of algae; the Subtidal changes were more complex. The switch in
sample area was 7 m2. canopy kelp was predictable because Nereocystis was
1838 DAVID R. SCHIELET AL. Ecology,Vol. 85, No. 7
almost at its southern limit and is intolerant of warm lighted, for example, by Fissurella volcano, which is
temperatures, whereas Macrocystis extends into Mex- considered to be a warm-water limpet but was classified
ico. Both species recruited annually but Nereocystis as "cosmopolitan" because its range extended just
juveniles did not survive. The switch in dominance of north of Cape Mendocino, while species such as Lottia
the two benthic Cryptopleura species soon after the pelta extend to Alaska, so were also cosmopolitan.
temperature change and then their decline were prob- The temperatures within Diablo Cove were within
ably responses both to temperature change and to the the range of those experienced during ENSO events
altered light environment as understory kelps declined (Glynn 1988) but they were sustained over 10 years.
and the Macrocystis canopy developed. Changes in However, they were up to 3 times greater than the nat-
grazers were complicated. Urchins were initially rare ural increase in temperature recorded along the Cali-
(<0.1 individual/m2) in Diablo Cove, but after 1987 fornian coast over the past 50 years (Roemmich 1992).
aggregated at up to 300 individuals/m2. These changes If ocean warming continues, therefore, our study im-
occurred in tandem with overall benthic changes. The plies that large and mostly unpredicted changes may
abrupt decline of gastropods after 1988 was probably occur in nearshore communities. Furthermore, there are
due to predation by fishes known to feed on gastropods potentially greater long-term changes through new in-
such as the sheephead Semicossyphus pulcher (Feder vaders and exotic species (Groshotz 2002), changes in
et al. 1974) and bat rays, Myliobatis californica (Gray nutrient concentrations and productivity that drive
et al. 1997). These species were scarce before 1985 but "bottom-up" processes (Roemmich and McGowan
then increased to average abundances of 0.4 and 1.4 1995, Menge 2000), a rise in sea level, and an increase
individuals per 50 X 5 m transect, respectively, within in extreme storms (Easterling et al. 2000).
Diablo Cove while staying at low numbers outside the Despite uncertainty about the exact mechanisms, the
cove. changes themselves were real, lasting, and compre-
Our study provides little support for the hypothesis hensive, providing both clues and a caution about pre-
of predictable directional changes in northern and dicting effects of climate change. Knowledge of key
southern species based on biogeographic models and ecological interactions is necessary but not sufficient
the metrics used by Sagarin et al. (1999). Although five to predict changes to communities. Identification of key
more southern species increased than decreased, and species and knowledge about their responses to tem-
one more northern species decreased than increased perature are also necessary. Because of the large num-
(Table 1), one might have expected that if predictions ber of species present, the resolution of predictions may
were robust, the great majority of species within these remain coarse. Our findings highlight that complex and
groups would have exhibited the predicted changes. unforeseen impacts can occur in nearshore benthic
The utility of the model is also weakened because so communities and also caution against using simplified
few species fell into these categories. Moreover, except models of species' responses to ocean warming based
for the substitution of giant kelp for bull kelp, the great on their present biogeographic distributions.
majority of changes were in widely distributed species, ACKNOWLEDGMENTS
the mortality of which should have been relatively in-
sensitive to variation in temperature. For example, two We thank J. Blecha, J. Carroll, C. Ehrler,S. Kimura,J.
Tupen,and many others for collecting the data, and Pacific
of the most-affected species (Mazzaella flaccida and Gas and Electric Companyand the CentralCaliforniaRe-
Endocladia muricata) were cosmopolitan and consid- gional WaterQuality ControlBoard,especially M. Thomas,
ered to be warm tolerant (Abbott and North 1972), and for full access and use of the data set. Thanksto Prof. Mike
Winterbourn and two anonymousreferees for useful com-
Pterygophora is common in kelp forests of southern ments on the
California (Dayton et al. 1999), yet all of these virtually manuscript.
disappeared from Diablo Cove. Overall, there were LITERATURE CITED
large and unforeseen changes to the benthos that re- Abbott, I. A., and G. J. Hollenberg. 1976. Marinealgae of
sulted in communities neither typical of sites further California.StanfordUniversityPress,Stanford, California,
south, where warmer seawater temperatures prevail, USA.
I. and W. J. North. 1972. Temperature influences
nor of sites in central California. However, once large Abbott, A.,
on floral composition in Californiacoastal waters. Pages
changes occurred in dominant algae, some of the com- 72-79 in K. Nisizawa. Proceedings of the International
munity-wide changes became more likely, such as the Seaweed Symposium. Volume 7. University of Tokyo
influx of grazers to the intertidal zone after space be- Press, Tokyo, Japan.
came available. It is apparent that biogeographic dis- Dayton, P. K., M. J. Tegner,P. B. Edwards,and K. L. Riser.
1999. Temporal spatialscales of kelp demography:
and the
tribution as defined by current models was a poor pre- role of oceanographic climate. Ecological Monographs 69:
dictor of overall changes resulting from warming. Per- 219-250.
haps other measures of species' positions along lati- Easterling,D. R., G. A. Meehl,C. Parmesan, A. Changnon,
S.
tudinal or temperature gradients relative to the range T. R. Karl, and L. O. Mearns. 2000. Climate extremes:
limits for each species may result in more accurate observations,modelling, and impacts. Science 289:2068-
2074.
predictions, but such information is currently limited. Feder,H. M., C. H. Turner, C. Limbaugh.1974. Obser-
and
The problem of characterizing distributions is high- vationson associationsof fishes with kelp beds in southern
July 2004 COMMUNITY CHANGES AFTER OCEAN WARMING 1839
California. California Department of Fish and Game Fish- Menge, B. A. 2000. Top-down and bottom-up community
ery Bulletin 160:1-44. regulation in marine rocky intertidal habitats. Journal of
Fields, P. A., J. B. Graham, R. J. Rosenblatt, and G. N. So- Experimental Marine Biology and Ecology 250:257-289.
mero. 1993. Effects of expected global climate change on Morris, R. H., D. P. Abbott, and E. C. Haderlie. 1980. In-
marine faunas. Trends in Ecology and Evolution 8:361- tertidal invertebrates of California. Stanford University
367. Press, Stanford, California, USA.
Foster, M. S. 1982. Factors controlling the intertidal zonation Pearcy, W. G., and A. Schoener. 1987. Changes in the marine
of Iridaeaflaccida (Rhodophyta). Journal of Phycology 18: biota coincident with the 1982-1983 El Nifio in the north-
285-295. eastern subarctic Pacific Ocean. Journal of Geophysical
Foster, M. S., E. W. Nigg, L. M. Kiguchi, D. R. Hardin, and Research 92:14417-14428.
J. S. Pearse. 2003. Temporal variation in an algal-domi- Reed, D. A., and M. S. Foster. 1984. The effects of canopy
nated high intertidal assemblage. Journal of Experimental shading on algal recruitment and growth in a giant kelp
Marine Biology and Ecology 289:15-39. (Macrocystis pyrifera) forest. Ecology 65:937-948.
Glynn, P. W. 1988. El Nifo-Southern oscillation 1982-1983: Ricketts, E. E, J. Calvin, and J. W. Hedgpeth. 1985. Between
nearshore population, community and ecosystem respons- Pacific tides. Stanford University Press, Stanford, Califor-
es. Annual Reviews of Ecology and Systematics 19:309- nia, USA.
345. Roemmich, D. 1992. Ocean warming and sea level rise along
the southwest U.S. coast. Science 257:373-375.
Gray, A. E., T. J. Mulligan, and R. W. Hannah. 1997. Food
habits, occurrence, and population structure of the bat ray, Roemmich, D., and J. McGowan. 1995. Climatic warming
and the decline of zooplankton in the California current.
Myliobatis californica, in Humboldt Bay, California. En-
vironmental Biology of Fishes 49:227-238. Science 267:1324-1326.
Groshotz, E. 2002. Ecological and evolutionary consequenc- Sagarin, R. D., J. P. Barry, S. E. Gilman, and C. H. Baxter.
es of coastal invasions. Trends in Ecology and Evolution 1999. Climate-related change in an intertidal community
over short and long time scales. Ecological Monographs
17:22-27.
69:465-490.
Holbrook, S. J., R. J. Schmitt, and J. S. Stephens. 1997.
Sanford, E. 1999. Regulation of keystone predation by small
Changes in an assemblage of temperate reef fishes asso-
ciated with climate change. Ecological Applications 7: changes in ocean temperature. Science 283:2095-2097.
1299-1310. Schroeter, S. C., J. D. Dixon, J. K. Kastendiek, R. O. Smith,
and J. R. Bence. 1993. Detecting ecological effects of en-
Ives, A. R., and G. Gilchrist. 1993. Climate change and eco- vironmental impacts: a case study of kelp forest inverte-
logical interactions. Pages 120-146 in P. M. Kareiva, J. G. brates. Ecological Applications 3:331-350.
Kingsolver, and R. H. Huey, editors. Biotic interactions and Southward, A. J., S. J. Hawkins, and M. T. Burrows. 1995.
global change. Sinaeur Associates, Sunderland, Massachu- Seventy years' observations of changes in distribution and
setts, USA. abundance of zooplankton and intertidal organisms in the
Kennett, J. P., and L. D. Stott. 1991. Abrupt deep-sea warm- western English Channel in relation to rising sea temper-
ing, palaeoceanographic changes and benthic extinctions ature. Journal of Thermal Biology 20:127-155.
at the end of the Palaeocene. Nature 353:225-229. Stewart-Oaten, A., W. M. Murdoch, and K. R. Parker. 1986.
Lubchenco, J., and J. D. Cubit. 1980. Heteromorphic life Environmental impact assessment: "pseudoreplication" in
histories of certain marine algae as adaptations to variations time? Ecology 67:929-940.
in herbivory. Ecology 61:676-687. Tomanek, L., and G. N. Somero. 1999. Evolutionary and
Lubchenco, J., S. A. Navarrete, B. N. Tissot, and J. C. Cas- acclimation-induced variation in the heat-shock responses
tilla. 1993. Possible ecological responses to global climate of congeneric marine snails (genus Tegula) from different
change: nearshore benthic biota of northeastern Pacific thermal habitats: implications for limits of thermotolerance
coastal ecosystems. Pages 147-166 in H. A. Mooney, E. and biogeography. Journal of Experimental Biology 202:
R. Fuentes, and B. I. Kronberg, editors. Earth system re- 2925-2936.
sponses to global change-contrasts between North and Wootton, J. T. 1993. Indirect effects and habitat use in an
South America. Academic Press, San Diego, California, intertidal community: interaction chains and interaction
USA. modifications. American Naturalist 141:71-89.
APPENDIX
Tables summarizing data used for biogeographic analysis are available in ESA's Electronic Data Archive: EcologicalArchives
E085-051-A1.
Communities
Author(s): David R. Schiel, John R. Steinbeck, Michael S. Foster
Source: Ecology, Vol. 85, No. 7 (Jul., 2004), pp. 1833-1839
Published by: Ecological Society of America
Stable URL: http://www.jstor.org/stable/3450356
Accessed: 25/01/2010 19:04
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=esa.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.
http://www.jstor.org
REPORTS
Ecology, 85(7), 2004, pp. 1833-1839
? 2004 by the Ecological Society of America
TEN YEARS OF INDUCED OCEAN WARMING CAUSES COMPREHENSIVE
CHANGES IN MARINE BENTHIC COMMUNITIES
DAVID R. SCHIEL,'4 JOHN R. STEINBECK,2 AND MICHAEL S. FOSTER3
'Zoology Department, University of Canterbury, Christchurch, New Zealand
2Tenera Environmental, 225 Prado Road, Suite D, San Luis Obispo, California 93401 USA
3Moss Landing Marine Laboratories, 8272 Moss Landing Road, Moss Landing, California 95039 USA
Abstract. One of the most commonly predicted effects of global ocean warming on
marine communities is a poleward shift in the distributional boundaries of species with an
associated replacement of cold-water species by warm-water species. However, these types
of predictions are imprecise and based largely on broad correlations in uncontrolled studies
that examine changes in the distribution or abundances of species in relation to seawater
temperature. Our study used an 18-year sampling program in intertidal and subtidal habitats
and before-after, control-impact analyses. We show that a 3.5?C rise in seawater temper-
ature, induced by the thermal outfall of a power-generating station, over 10 years along 2
km of rocky coastline in California resulted in significant community-wide changes in 150
species of algae and invertebrates relative to adjacent control areas experiencing natural
temperatures. Contrary to predictions based on current biogeographic models, there was
no trend toward warmer-water species with southern geographic affinities replacing colder-
water species with northern affinities. Instead, the communities were greatly altered in
apparently cascading responses to changes in abundance of several key taxa, particularly
habitat-forming subtidal kelps and intertidal foliose red algae. Many temperature-sensitive
algae decreased greatly in abundance, whereas many invertebrate grazers increased. The
responses of these benthic communities to ocean warming were mostly unpredicted and
strongly coupled to direct effects of temperature on key taxa and indirect effects operating
through ecological interactions.
Key words: BACI analysis; benthic community structure; biogeographic distribution information;
climate change, predicting effects; Diablo Cove, California, USA; ocean warming; rocky reef; tem-
perature change; thermal discharge of a power plant, effects.
INTRODUCTION between natural changes in seawater temperature and
Ocean warming through climate change is expected species' occurrences or abundances (Lubchenco et al.
to have significant effects on the composition and struc- 1993), or else on point counts separated by large time
ture of marine communities (Fields et al. 1993). These periods (e.g., Sagarin et al. 1999). Few empirical stud-
include changes in distributional boundaries of species ies incorporate significant changes in seawater tem-
and replacement of cold-water taxa by others with perature over periods long enough to span the lives of
warm-water affinities. Paleoclimatic records showing most species, while also having control areas and con-
shifts in the distribution or abundance of species as- tinuous time series of sampling that allow specific spa-
sociated with changes in temperature (Kennett and tial and temporal trends in community structure to be
Stott 1991), localized changes in species abundances detected and tested.
over decades (Southward et al. 1995, Holbrook et al. Here, we address the hypothesis that elevated sea-
1997, Sagarin et al. 1999), and changes associated with water temperature brought about by the thermal dis-
El Ninio-Southern Oscillation events (Pearcy and charge of a power generating station in California un-
Schoener 1987, Dayton et al. 1999) support such sce- derlay changes to rocky intertidal and subtidal com-
narios. However, evidence for changes due to ocean munities. Eighteen years of sampling before and after
warming is limited and based largely on correlations the thermal discharge began, and the simultaneous sam-
pling of nearby control areas that experienced only
Manuscript received28 July2003; revised1 February2004; natural
3
accepted February 2004;finalversion received10 March
2004. temperatures, allowed a robust assessment of
Corresponding Editor: G. Morgan.
S. changes in species' abundances through time. The sus-
4 E-mail: tained elevation of seawater temperature enabled an
david.schiel@canterbury.ac.nz
1833
July 2004
Volume 85 No. 7 EC UBLICATION O LOGICAL SOCIETY OF AMERICA
A PUBLICATION OF THE ECOLOGICAL SOCIETY OF AMERICA
The Robert H. MacArthurAward Lecture
Toward metabolictheoryof ecology
a
Forum
Themetabolictheoryof ecology
Reports
Tenyearsof inducedocean warmingcauses comprehensivechanges
in marinebenthiccommunities
Articles
withdifferent
Responses preyfromhabitats
of predator and
regimes:localadaptation heritability
1834 DAVID R. SCHIELET AL. Ecology,Vol. 85, No. 7
exceptionally strong and direct test of responses by 18- Seawater temperature
nearshore benthic marine communities to ocean warm- -
k , AAMAt.I\krWAMb
ing. Furthermore, the study area was close to a source 0 14 IJ i'
I
,,
' 4tX 1 ^
of potential invaders because it is only 138 km north I-
o
of Point Conception, a recognized boundary between I--,
0
a warm-temperate biota to the south and a cool-tem-
perate biota to the north (Ricketts et al. 1985). a) 6-
QC
E
2- Delta temperature
METHODS
Diablo Cove is on an exposed rocky coast in central An -
_ A- - A.- --_
_I
California, USA (35?12' N, 120?52' W), has a shoreline . , ; I I
, Ii I i I I , I I I I
cO o 0 C 't o
) 0 NC\
-2 km long, a surface area of 15 ha, an average depth I Ir oo
O oo
O oo
O oo o a
O o,) o
) oC C C C) C c) n C)
of 8 m, and a maximum depth of 18 m. The nearby
power plant draws seawater from outside the cove, uses FIG. 1. Monthly mean temperatures control (dashed
at
it in a cooling system to re-condense freshwater steam, line) and impact (solid line) sites measuredat a depth of 3
and discharges it into the cove at a maximum rate of m below MLLWfrom 1976 to 1995. Delta temperatures are
9.5 X 109 L/d, creating a thermal plume that spreads the averagedifferencesbetween controland impactstations.
The vertical line at 1985 shows when thermalchangebegan
throughout the cove. The plant has operated at mostly within Diablo Cove.
full power since May 1985 and very little is discharged
other than heated water. Required regular water mon-
itoring and bioassays showed that concentrations of and "impact" areas was used to detect significant ther-
other constituents were very low relative to the large mal effects (Stewart-Oaten et al. 1986). The "before"
volume of the discharge and did not present acute or period was January 1978 through December 1984 and
chronic toxicity to marine larval organisms used in the the "after" period was January 1987 through June
tests. Temperature was recorded continuously (every 1995. Surveys from January 1985 through December
20 min) from 1976 with digital recorders (+0.2?C ac- 1986 were omitted to allow time for thermal effects to
curacy) in metal canisters at the -3 m MLLW level (3 occur. The high power to detect small changes in abun-
m below mean lowest low water) of several stations. dance was due to the number of replicate surveys in
Biological surveys were done six times each year each period. The analytical variable was the difference
from 1978 through 1988 and four times annually from between the mean of each impact station and the grand
1989 through 1995. There were four intertidal and five mean of all the control stations in each survey. The
subtidal stations within Diablo Cove and two subtidal differences between these paired means ("deltas")
and three intertidal control stations 1-2 km outside the were then compared statistically before and after the
cove beyond the influence of the discharge. Each in- putative impact. Data for each species were tested for
tertidal station had two 30 m fixed transects parallel to additivity, linear trend, serial correlation, and homo-
the shore, one each at the +0.3 m and +0.9 m MLLW geneity of variance, transformed appropriately to meet
tidal levels. Ten 1-m2 quadrats (initially chosen ran- test assumptions, and then analyzed using a two-way.
domly) per transect were sampled during each survey. ANOVA with factors for "period" and "station." If
All visible algae and invertebrates were counted or their the interaction was not significant, "period" was used
percentage cover assessed. At each subtidal station, a to test for effects of thermal change. A significant in-
fixed 28-m2 circular plot divided into four quadrants at teraction indicated an impact of varying magnitude
a depth of 3 m to 4 m was sampled on each occasion. among stations or no impact due to variable results
In each quadrant, understory algal and attached inver- among stations. For taxa that tested significant, we cal-
tebrate cover was estimated from 50 random points, culated percentage change in abundance in Diablo
and invertebrates and kelp were counted. The great Cove relative to controls as: Relative percentage
- - - -
majority of taxa were identified to species level. Fol- change = ([(XB, XBi) (XA XAi)]/(XA XAi)) X
lowing Sagarin et al. (1999), individual species were =
100, where XBC mean abundance at the pooled control
classified as "cosmopolitan" (i.e., with an extensive sites before thermal increase; XBi = mean abundance
north-south distribution along the west coast), "north- at the pooled Diablo Cove sites before thermal increase;
ern" (with a southern limit north of Point Conception) XAC = mean abundance at the pooled control sites after
or "southern" (with a northern limit south of Cape thermal increase; XAi = mean abundance at the pooled
Mendocino) based on information in Morris et al. Diablo Cove sites after thermal increase.
(1980) and Abbott and Hollenberg (1976). Taxa were
RESULTS
excluded from this classification if they were identified
only to genus and may have contained species of dif- Seawater temperatures within and outside Diablo
ferent geographical affinities. Cove (California, USA) were similar from 1976 to
BACI (before-after, control-impact) analysis of var- 1985, when the heated discharge began (Fig. 1). How-
iance using "before" and "after" data from "control" ever, from 1986 through 1995 they were elevated by
July 2004 COMMUNITYCHANGESAFTER OCEAN WARMING 1835
TABLE 1. The numberof species and the percentageof all taxa combinedthat significantly
increased(Inc.), decreased(Dec.), or did not change (Nc) in abundance within Diablo Cove
(California,USA) relative to control stations outside the cove over the 10 years after the
rise in seawatertemperature.
Cosmopolitan Northern Southern
Species types Inc. Dec. Nc Inc. Dec. Nc Inc. Dec. Nc
Intertidal algae 1 15 1 0 1 0 0 1 0
Intertidal invertebrates 11 8 1 1 0 1 2 0 2
Subtidalalgae 6 6 2 1 1 0 1 0 0
Subtidalinvertebrates 11 16 9 0 1 0 6 3 1
All taxa (%) 26.6 41.3 11.9 1.8 2.8 0.9 8.3 3.7 2.8
Notes: Species werecategorizedas cosmopolitan,northern, southern theirbiogeographic
or in
affinities.Significancewas based on BACI analyses (see Methods).
an average of 3.5?C inside the cove, with monthly were sufficiently abundant to be analyzed statistically.
means ranging from 2.0 to 5.1?C above ambient. Max- Of these, 150 species (87%) changed significantly in
imum temperatures occurred during the El Ninio- abundance at stations with elevated temperatures rel-
Southern oscillation (ENSO) years of 1983, 1987, and ative to control stations, with 65 species (37.8%) in-
1992-1993. The coast in and around Diablo Cove ini- creasing and 85 species (49.4%) decreasing.
tially had marine communities typical of other semi- The biogeographic affinities of 109 of the 172 spe-
protected sites in the region (Ricketts et al. 1985). In- cies could be assigned with confidence (see Appendix).
tertidal areas were dominated by foliose (e.g., Maz- There was no consistent pattern of change in abun-
zaella flaccida) and branched (e.g., Chondracanthus dances related to geographic affinities (Table 1). Most
canaliculatus) red algae, with some grazing gastropods intertidal algae (89%) were cosmopolitan. Of these, 15
and sessile invertebrates in the understory. Subtidal
species decreased in abundance, as did the one northern
areas had a canopy of northern bull kelp, Nereocystis and one southern algal species. Twenty of the 26 in-
luetkeana, an understory dominated by the kelps Pter- tertidal invertebrate species (77%) were cosmopolitan,
ygophora californica and Laminaria setchellii, and pri- and 11 species increased, eight species decreased, and
mary cover of coralline and foliose red algae, with a one species remained unchanged relative to controls.
lush assemblage of smaller algae and invertebrates in- The few northern and southern species had equal num-
terspersed. bers increasing or showing no change. Most subtidal
During our 18-year study, 714 taxa were counted or algae (82%) were cosmopolitan, with six species in-
assessed for percentage cover, but only 172 species
creasing, six species decreasing, and two species re-
maining unchanged in abundance. One northern species
A) Algae increased and one decreased, while the only southern
20-
species increased. Thirty-six of the 47 subtidal inver-
15- tebrates (77%) were cosmopolitan, of which 11 species
increased, 16 species decreased, and nine species had
10 - no change. The one northern species decreased, while
six southern species increased, three species decreased,
and one species remained unchanged. Overall, of the
n
O
IC
a) AiLiL..
C, 5- . F]I
I I I I I I I I I I I I I I .... I I I I I I
71111n 11nnn
80% of species that were cosmopolitan, far more de-
creased (41.3%) than increased or remained un-
j 25- B) Invertebrates T changed. Of the 6% of species that were northern, more
aG 20- decreased (2.8%) than increased (1.8%). And of the
a r,
southern 15% of species, more increased (8.3%) than
15- decreased (3.7%).
'?:
10- Most species within Diablo Cove underwent large
'7,;
relative changes in abundance in the 10 years following
5-
n-
L
I .I.._.
I1 I I I
| I | *
7,
I
n
I I
flrr-
seawater warming. Fifty-four percent of algae (Fig. 2A)
and 27% of invertebrates (Fig. 2B) declined by at least
-100-80 -60 -40 -20 0 1 2 3 4 5 50%. Dominant algae (with an initial combined cover
Decrease (%) Log(% increase) >60%) declined dramatically, especially the cosmo-
FIG.2. Percentage politan red algae Mastocarpus papillatus (-65%) in
changesin abundance (A) algae and the
of
(B) invertebratespecies relative to controls after thermal upper zone, Endocladia muricata (-63%), Maz-
change, shown as decreases(solid bars),no change (mottled zaella flaccida (-98%), and Chondracanthus canali-
bars), and log,0 increases (open bars). culatus (-51%). Algal taxa richness declined by 40%.
1836 DAVID R. SCHIELET AL. Ecology, Vol. 85, No. 7
In contrast, non-calcified algal crusts increased by 62%.
100
The few other increases were in cosmopolitan taxa that,
taken together, comprised 1 % of total algal cover. 90
Subtidal understory kelps declined by 82% from an 80
average of 7.3 to 1.3 plants/m2. The initially abundant 70
cosmopolitan kelps P. californica and L. setchellii de- 60
clined by -90%, but a cosmopolitan understory red
50
alga, Cryptopleura violacea, increased by 3000% (i.e.,
cm
E
a log percentage increase of 3.47) to cover 30-50% of
the substratum. Some of the most important changes a)
N
involved the large kelps. N. luetkeana declined by 97% g d
in the cove and was replaced by the initially rare south- o 0)
ern giant kelp Macrocystis pyrifera, which dominated 6
c
the surface canopy with an average density of 1.5 ()
0
plants/7 m2 after 10 years. a) c
Intertidal grazers became very abundant, especially (0
C -0
the cosmopolitan Tegula funebralis, which increased
by 79% to a 10-year average of 175 individuals/m2 on a)
c
the upper shore, and the cosmopolitan limpet Macclin-
co
tockia scabra that increased by >400% to 86 individ- 0)
uals/m2. Some initially rare species became common,
including limpets of the cosmopolitan genus Lottia,
which increased from -3 to 9 limpets/m2, and the cos-
mopolitan keyhole limpet Fissurella volcano, which
increased by >700% to 13 limpets/m2 on the lower
shore. The cosmopolitan sea urchin Strongylocentrotus
purpuratus increased by 3000% to 8.2 individuals/m2
in the low intertidal zone. 0 o cmj 1 cO 0oo c'
N 't C
Almost all changes were in the abundances of species 0 0 0 o0 COcO 0 0 0
0)0)0 0)0)0)CD0)0)0)
already present. A few new invertebrates appeared in
impact sites after thermal change, but they were never FIG.3. Annualmeanpercentcover of foliose algae (open
of
squares;left-handaxis) and abundance grazersper square
abundant, such as brittle stars Ophiothrix spp. (prob- zone of
meter(solid circles; right-handaxis) in the intertidal
ably southern; from 0 to 1.1 individuals/0.25 m2), Aply- (A) control sites and (B) impactsites.
sia californica (cosmopolitan; from 0 to 0.2 individ-
uals/7 m2), Cryptochiton stelleri (cosmopolitan; from
0 to 0.1 individual/7 m2), and Norrisia norrisi (south- trol areas (7400%, 350%, and 49000%, respectively),
ern; from 0 to 0.1 individual/7 m2). their overall abundances remained low. The total cover
Intertidal habitats were greatly modified in Diablo of all sessile invertebrates increased from -2.0-4.7%
Cove, especially in the vertical layering of algae. All in the intertidal zone of Diablo Cove sites. There were
sites initially had a dense cover (70-80%) of foliose also slight increases of predatory gastropods such as
algae up to 10 cm thick, with grazing invertebrates Acanthina spp. and Ocenebra spp. on the upper shore,
interspersed below. At control sites these algae de- but these were not statistically significant.
clined around 1985 and grazers increased, most likely Community changes in subtidal habitats were more
as delayed responses to the severe 1983 ENSO (Fig. complex. Control sites (Fig. 4A) had a primary sub-
3A). However, dense algae returned within two years stratum cover of foliose algae from 1978 to 1995 but
and by 1995 both algal and grazer abundances were understory kelps underwent large changes in abun-
virtually the same as in 1978. At the thermally affected dance, ranging from 20 to 225 kelps/m2. Kelps had high
sites, foliose algae also declined in 1985 but did not recruitment after the 1983 and 1987 ENSO events but
recover (Fig. 3B). Instead, grazers dominated for 10 then gradually declined. Grazers fluctuated in abun-
years and, as might be predicted given their high den- dance. By the end of the study, however, the control
sities (cf. Lubchenco and Cubit 1980), this was accom- sites looked remarkably similar to their initial condi-
panied by large increases in bare substratum, low-lying tion. The thermally affected sites had far greater chang-
ephemeral turf, and thin algal crusts. Other changes, es (Fig. 4B). Bull kelp was replaced by giant kelp,
especially at higher tidal levels, were increases in subcanopy kelps declined and had no major episodes
space-occupying species such as barnacles, Chthama- of recruitment, and grazers fluctuated and then declined
lus fissus and Tetraclita rubescens, and the sand tube after 1988 as dense foliose algae dominated the sub-
worm Phragmatopoma californica. Although these stratum. In contrast to control sites, the grazing assem-
species had large percentage increases relative to con- blage changed after 1985 when sea urchins began in-
July 2004 COMMUNITY CHANGES AFTER OCEAN WARMING 1837
creasing in abundance while gastropods declined (Fig.
4C). Furthermore, only within Diablo Cove did the
dominant benthic algae change (Fig. 4D). The cos-
a)
E
mopolitan red alga Cryptopleura ruprechtiana was re-
0 6
O placed by its cosmopolitan congener C. violacea.
a)
a) 0 DISCUSSION
Ca)
D. c"- Key results were that after a sustained rise in sea-
water temperature the great majority of nearshore spe-
cies changed in abundance and whole communities
changed in character. Mechanisms of change are usu-
o ally complex, involve the interaction of many factors,
and, as abundances change, can be highly sensitive to
140 the strengths of interactions among species (Menge
2000). The level of experimentation (i.e., bay-wide) in
120
this study limits the resolution of the processes re-
a)
100 sponsible for change and, therefore, the sequence of
0
a)
O (I-
a) observed changes and their consistency give the best
80
0 clues for evaluating the processes responsible.
60 The deterioration of algae clearly preceded increases
0
a)
EL ~-. in grazers. Dominant algae, especially Mazzaellaflac-
40 t:
c. cida, Nereocystis, Pterygophora, and Laminaria, had
20 -0 acute responses to the increased temperature and de-
clined rapidly, becoming bleached with deteriorating
0
thalli within a few months. This provided new settle-
ment space and an altered light environment to primary
120 substrata (Reed and Foster 1984) and may have trig-
CM
gered a series of responses such as altered mortality,
E 100
growth, and reproductive and recruitment rates of other
d 80 species (Ives and Gilchrist 1993), behavioral changes
c relating to temperature preferences and avoidance
c
60 -
C\ (Schroeter et al. 1993, Sanford 1999), and modified
(I-
food webs and species interactions (Wootton 1993).
40
a Initially dominant species had little or no subsequent
20 recruitment, even during the periods when grazer num-
(D bers declined (e.g., 1988).
0 In contrast, grazing gastropods showed a positive
response to temperature, which most likely involved
80 physiological tolerances, trophic responses, availabil-
ity of space and recruitment dynamics. Many intertidal
70
molluscs are highly resistant to the extreme tempera-
60 tures experienced during low tide and can repair ther-
50 mal damage to proteins (Tomanek and Somero 1999).
40 Their strong recruitment intertidally was most likely a
a) direct consequence of having more free space to settle
0 30
a) after the decline of algae. Once they reached the very
20
high densities seen in Diablo Cove (up to 370 molluscs/
10 m2), they would have precluded effective recruitment
of macroalgae for much of the post-impact period (Fos-
CO
N- CM C <D D CD C\ C C ter 1982). Grazers often recruit into areas cleared of
algae along the west coast (Foster et al. 2003), as was
FIG. 4. Annual mean percent cover of foliose algae (left- seen in all sites after the 1983 ENSO. Given the quick
hand axes) and numbers of subcanopy kelps and grazers recovery of control sites, however, it is likely that the
(right-hand axes) at (A) subtidal control sites and (B) subtidal
trophic response of grazing mollusks in the impacted
impact sites, with (C) a separation of grazers into gastropods
and sea urchins, and (D) the percent cover of the two most areas was secondary to the direct effects of temperature
abundant benthic foliose algae (Cryptopleura spp.). Percent on algae.
cover could total >100% because of layering of algae; the Subtidal changes were more complex. The switch in
sample area was 7 m2. canopy kelp was predictable because Nereocystis was
1838 DAVID R. SCHIELET AL. Ecology,Vol. 85, No. 7
almost at its southern limit and is intolerant of warm lighted, for example, by Fissurella volcano, which is
temperatures, whereas Macrocystis extends into Mex- considered to be a warm-water limpet but was classified
ico. Both species recruited annually but Nereocystis as "cosmopolitan" because its range extended just
juveniles did not survive. The switch in dominance of north of Cape Mendocino, while species such as Lottia
the two benthic Cryptopleura species soon after the pelta extend to Alaska, so were also cosmopolitan.
temperature change and then their decline were prob- The temperatures within Diablo Cove were within
ably responses both to temperature change and to the the range of those experienced during ENSO events
altered light environment as understory kelps declined (Glynn 1988) but they were sustained over 10 years.
and the Macrocystis canopy developed. Changes in However, they were up to 3 times greater than the nat-
grazers were complicated. Urchins were initially rare ural increase in temperature recorded along the Cali-
(<0.1 individual/m2) in Diablo Cove, but after 1987 fornian coast over the past 50 years (Roemmich 1992).
aggregated at up to 300 individuals/m2. These changes If ocean warming continues, therefore, our study im-
occurred in tandem with overall benthic changes. The plies that large and mostly unpredicted changes may
abrupt decline of gastropods after 1988 was probably occur in nearshore communities. Furthermore, there are
due to predation by fishes known to feed on gastropods potentially greater long-term changes through new in-
such as the sheephead Semicossyphus pulcher (Feder vaders and exotic species (Groshotz 2002), changes in
et al. 1974) and bat rays, Myliobatis californica (Gray nutrient concentrations and productivity that drive
et al. 1997). These species were scarce before 1985 but "bottom-up" processes (Roemmich and McGowan
then increased to average abundances of 0.4 and 1.4 1995, Menge 2000), a rise in sea level, and an increase
individuals per 50 X 5 m transect, respectively, within in extreme storms (Easterling et al. 2000).
Diablo Cove while staying at low numbers outside the Despite uncertainty about the exact mechanisms, the
cove. changes themselves were real, lasting, and compre-
Our study provides little support for the hypothesis hensive, providing both clues and a caution about pre-
of predictable directional changes in northern and dicting effects of climate change. Knowledge of key
southern species based on biogeographic models and ecological interactions is necessary but not sufficient
the metrics used by Sagarin et al. (1999). Although five to predict changes to communities. Identification of key
more southern species increased than decreased, and species and knowledge about their responses to tem-
one more northern species decreased than increased perature are also necessary. Because of the large num-
(Table 1), one might have expected that if predictions ber of species present, the resolution of predictions may
were robust, the great majority of species within these remain coarse. Our findings highlight that complex and
groups would have exhibited the predicted changes. unforeseen impacts can occur in nearshore benthic
The utility of the model is also weakened because so communities and also caution against using simplified
few species fell into these categories. Moreover, except models of species' responses to ocean warming based
for the substitution of giant kelp for bull kelp, the great on their present biogeographic distributions.
majority of changes were in widely distributed species, ACKNOWLEDGMENTS
the mortality of which should have been relatively in-
sensitive to variation in temperature. For example, two We thank J. Blecha, J. Carroll, C. Ehrler,S. Kimura,J.
Tupen,and many others for collecting the data, and Pacific
of the most-affected species (Mazzaella flaccida and Gas and Electric Companyand the CentralCaliforniaRe-
Endocladia muricata) were cosmopolitan and consid- gional WaterQuality ControlBoard,especially M. Thomas,
ered to be warm tolerant (Abbott and North 1972), and for full access and use of the data set. Thanksto Prof. Mike
Winterbourn and two anonymousreferees for useful com-
Pterygophora is common in kelp forests of southern ments on the
California (Dayton et al. 1999), yet all of these virtually manuscript.
disappeared from Diablo Cove. Overall, there were LITERATURE CITED
large and unforeseen changes to the benthos that re- Abbott, I. A., and G. J. Hollenberg. 1976. Marinealgae of
sulted in communities neither typical of sites further California.StanfordUniversityPress,Stanford, California,
south, where warmer seawater temperatures prevail, USA.
I. and W. J. North. 1972. Temperature influences
nor of sites in central California. However, once large Abbott, A.,
on floral composition in Californiacoastal waters. Pages
changes occurred in dominant algae, some of the com- 72-79 in K. Nisizawa. Proceedings of the International
munity-wide changes became more likely, such as the Seaweed Symposium. Volume 7. University of Tokyo
influx of grazers to the intertidal zone after space be- Press, Tokyo, Japan.
came available. It is apparent that biogeographic dis- Dayton, P. K., M. J. Tegner,P. B. Edwards,and K. L. Riser.
1999. Temporal spatialscales of kelp demography:
and the
tribution as defined by current models was a poor pre- role of oceanographic climate. Ecological Monographs 69:
dictor of overall changes resulting from warming. Per- 219-250.
haps other measures of species' positions along lati- Easterling,D. R., G. A. Meehl,C. Parmesan, A. Changnon,
S.
tudinal or temperature gradients relative to the range T. R. Karl, and L. O. Mearns. 2000. Climate extremes:
limits for each species may result in more accurate observations,modelling, and impacts. Science 289:2068-
2074.
predictions, but such information is currently limited. Feder,H. M., C. H. Turner, C. Limbaugh.1974. Obser-
and
The problem of characterizing distributions is high- vationson associationsof fishes with kelp beds in southern
July 2004 COMMUNITY CHANGES AFTER OCEAN WARMING 1839
California. California Department of Fish and Game Fish- Menge, B. A. 2000. Top-down and bottom-up community
ery Bulletin 160:1-44. regulation in marine rocky intertidal habitats. Journal of
Fields, P. A., J. B. Graham, R. J. Rosenblatt, and G. N. So- Experimental Marine Biology and Ecology 250:257-289.
mero. 1993. Effects of expected global climate change on Morris, R. H., D. P. Abbott, and E. C. Haderlie. 1980. In-
marine faunas. Trends in Ecology and Evolution 8:361- tertidal invertebrates of California. Stanford University
367. Press, Stanford, California, USA.
Foster, M. S. 1982. Factors controlling the intertidal zonation Pearcy, W. G., and A. Schoener. 1987. Changes in the marine
of Iridaeaflaccida (Rhodophyta). Journal of Phycology 18: biota coincident with the 1982-1983 El Nifio in the north-
285-295. eastern subarctic Pacific Ocean. Journal of Geophysical
Foster, M. S., E. W. Nigg, L. M. Kiguchi, D. R. Hardin, and Research 92:14417-14428.
J. S. Pearse. 2003. Temporal variation in an algal-domi- Reed, D. A., and M. S. Foster. 1984. The effects of canopy
nated high intertidal assemblage. Journal of Experimental shading on algal recruitment and growth in a giant kelp
Marine Biology and Ecology 289:15-39. (Macrocystis pyrifera) forest. Ecology 65:937-948.
Glynn, P. W. 1988. El Nifo-Southern oscillation 1982-1983: Ricketts, E. E, J. Calvin, and J. W. Hedgpeth. 1985. Between
nearshore population, community and ecosystem respons- Pacific tides. Stanford University Press, Stanford, Califor-
es. Annual Reviews of Ecology and Systematics 19:309- nia, USA.
345. Roemmich, D. 1992. Ocean warming and sea level rise along
the southwest U.S. coast. Science 257:373-375.
Gray, A. E., T. J. Mulligan, and R. W. Hannah. 1997. Food
habits, occurrence, and population structure of the bat ray, Roemmich, D., and J. McGowan. 1995. Climatic warming
and the decline of zooplankton in the California current.
Myliobatis californica, in Humboldt Bay, California. En-
vironmental Biology of Fishes 49:227-238. Science 267:1324-1326.
Groshotz, E. 2002. Ecological and evolutionary consequenc- Sagarin, R. D., J. P. Barry, S. E. Gilman, and C. H. Baxter.
es of coastal invasions. Trends in Ecology and Evolution 1999. Climate-related change in an intertidal community
over short and long time scales. Ecological Monographs
17:22-27.
69:465-490.
Holbrook, S. J., R. J. Schmitt, and J. S. Stephens. 1997.
Sanford, E. 1999. Regulation of keystone predation by small
Changes in an assemblage of temperate reef fishes asso-
ciated with climate change. Ecological Applications 7: changes in ocean temperature. Science 283:2095-2097.
1299-1310. Schroeter, S. C., J. D. Dixon, J. K. Kastendiek, R. O. Smith,
and J. R. Bence. 1993. Detecting ecological effects of en-
Ives, A. R., and G. Gilchrist. 1993. Climate change and eco- vironmental impacts: a case study of kelp forest inverte-
logical interactions. Pages 120-146 in P. M. Kareiva, J. G. brates. Ecological Applications 3:331-350.
Kingsolver, and R. H. Huey, editors. Biotic interactions and Southward, A. J., S. J. Hawkins, and M. T. Burrows. 1995.
global change. Sinaeur Associates, Sunderland, Massachu- Seventy years' observations of changes in distribution and
setts, USA. abundance of zooplankton and intertidal organisms in the
Kennett, J. P., and L. D. Stott. 1991. Abrupt deep-sea warm- western English Channel in relation to rising sea temper-
ing, palaeoceanographic changes and benthic extinctions ature. Journal of Thermal Biology 20:127-155.
at the end of the Palaeocene. Nature 353:225-229. Stewart-Oaten, A., W. M. Murdoch, and K. R. Parker. 1986.
Lubchenco, J., and J. D. Cubit. 1980. Heteromorphic life Environmental impact assessment: "pseudoreplication" in
histories of certain marine algae as adaptations to variations time? Ecology 67:929-940.
in herbivory. Ecology 61:676-687. Tomanek, L., and G. N. Somero. 1999. Evolutionary and
Lubchenco, J., S. A. Navarrete, B. N. Tissot, and J. C. Cas- acclimation-induced variation in the heat-shock responses
tilla. 1993. Possible ecological responses to global climate of congeneric marine snails (genus Tegula) from different
change: nearshore benthic biota of northeastern Pacific thermal habitats: implications for limits of thermotolerance
coastal ecosystems. Pages 147-166 in H. A. Mooney, E. and biogeography. Journal of Experimental Biology 202:
R. Fuentes, and B. I. Kronberg, editors. Earth system re- 2925-2936.
sponses to global change-contrasts between North and Wootton, J. T. 1993. Indirect effects and habitat use in an
South America. Academic Press, San Diego, California, intertidal community: interaction chains and interaction
USA. modifications. American Naturalist 141:71-89.
APPENDIX
Tables summarizing data used for biogeographic analysis are available in ESA's Electronic Data Archive: EcologicalArchives
E085-051-A1.