Download Abrupt community change on a rocky shore – biological

Ecology Letters, (2004) 7: 441–445
doi: 10.1111/j.1461-0248.2004.00601.x
LETTER
Abrupt community change on a rocky shore –
biological mechanisms contributing to the potential
formation of an alternative state
Robert T. Paine* and Alan
C. Trimble
Department of Biology, Box
351800, University of
Washington, Seattle,
WA 98195-1800, USA
*Correspondence: E-mail:
[email protected]
Abstract
The 1997/1998 El Niño initiated a major shift in the intertidal assemblage on the
Washington State outer coast. A 25 year time series (1978–2003) shows stands of
dominant canopy algae replaced by mussel beds. A prior experiment had indicated that
mussels can become too large to be eaten by starfish; newly initiated starfish removals
predict mussel attainment of a size refuge. Such escapes inhibit recovery towards prior
community composition and enhance development of alternative community states
which may persist long after the originating forcing has lessened or disappeared.
Keywords
Alternative states, Hedophyllum, intertidal, mussels, Pisaster, size escapes, stability.
Ecology Letters (2004) 7: 441–445
INTRODUCTION
Lewontin (1969) proposed that shifts in community
composition be cast in the language of mathematics
governing dynamical space. Lewontin’s framework suggested that alternative states might be identified from long-term
studies, and assemblage stability tested by deliberate
perturbations of species abundance. Observational (Brooks
& Dodson 1965) and experimental (Paine 1966) field studies
conducted earlier had identified that dramatic compositional
shifts could be forced through alterations of consumer
abundance. Slightly later, Sutherland (1974) explicitly cast
his fouling community experiments in the framework
suggested by Lewontin. Recently, increasing concerns about
the biological impacts of climate change (Alley et al. 2003)
and regime shifts leading to alternative assemblages
(Scheffer et al. 2001, Scheffer & Carpenter 2003, Beisner
et al. 2003, Carpenter 2003) have heightened awareness of
their sweeping consequences.
This paper provides preliminary details on a physicallyforced but abrupt natural change in assemblage composition
on an intensely studied marine rocky intertidal shore. The
relationships we describe focus on three species: a predatory
starfish Pisaster ochraceus (Brandt, 1835) preferentially consumes the mussel Mytilus californianus Conrad, 1837 which in
turn can outcompete a brown alga (kelp) Hedophyllum sessile
(Setchell, 1901). Although the local species richness is two
to three orders of magnitude greater, and spatial structure is
known to be important, we believe the essence of the
described dynamic is captured by the interplay of just these
three dominant species.
Our data meet Connell & Sousa’s (1983) criteria for
alternative (multiple) states: (1) Two or more different
biological populations or assemblages must be able to
occupy the same environment. (2) The novel state should be
able to persist indefinitely without maintained biological
forcing. (3) The data must meet reasonable temporal and
spatial constraints. We argue that for systems fundamentally
limited by space, the critical characteristic of persistence or
stability (as in alternative ÔstableÕ states) is the demonstration
of hysteresis. Furthermore, our study may, despite a spatial
scale measured as only 1–2 m along a 5 m vertical tidal
gradient, meet the regime shift criteria of Carpenter (2003):
a Ôrapid modification of ecosystem organizationÕ with
Ôprolonged consequencesÕ.
It is impossible to discuss alternative states without
employing certain terms, all with vague or contentious
definitional histories. In general, we follow Beisner et al.
(2003) and Scheffer & Carpenter (2003). To us, stability
implies a dynamic response to a perturbation, one that
promotes recovery to the pre-perturbation state. Persistence is simply an acknowledgment that the species of
interest did not go extinct during the observational
interval and, following Frank (1968), is especially vexing
when individuals are long-lived. Hysteresis is revealed
when a system can exist in multiple states, and in which
the shift or collapse trajectory differs from the recovery
one.
2004 Blackwell Publishing Ltd/CNRS
442 R. T. Paine and A. C. Trimble
METHODS
The research took place on the Washington State (USA)
outer coast, on rocky intertidal shores frequently exposed to
heavy wave action. The experimental studies at Mukkaw Bay
(4819¢ N, 12440¢ W) were conducted from 1963 to 1968;
the site has been monitored irregularly since. Research at
Tatoosh Island (4824¢ N, 12444¢ W) was initiated in 1968;
data acquisition and manipulative experiments began in
1970 and have continued to the present.
Density estimates of starfish were obtained from quadrats
or line transects. Pisaster removals were manual, with the
individuals moved 10–40 m away. Per cent cover estimates
of benthic algae and mussels were visual; the Dethier et al.
(1993) study suggests that such estimates may provide the
best representation of relative cover.
Many details of the Mukkaw Bay site have been published
(Paine 1966, Paine et al. 1985). At Tatoosh, study began in
1978 of six original algal census areas on the heavily exposed
north side, each with extensive Hedophyllum and few if any
Pisaster or mussels. These were successfully invaded by
mussels by 1999. In April 2000, we initiated a predator
removal experiment. To quantify effects of Pisaster removal
or retention, numbered plastic markers were epoxied as
close to the mussel bed edge as possible at the removal and
three (of four) Pisaster present sites. Because of PisasterÕs
mobility, we designated only areas separated by significant
surge channels as removal (n ¼ 1) or sites with starfish
(n ¼ 4). The procedure required combining three original
algal census sites into the single removal site, and combining
two others into one of the sites with Pisaster. The remaining
three in that treatment involved one of the original algal
sites and two new ones. The Pisaster removal site had an
initial perimeter of 45 m; sites with Pisaster had perimeters of
16, 18, 20 and 20 m. The distance of the markers to the
nearest mussel was measured whenever conditions permitted. However, marker loss, burial under encroaching
mussels or hidden under a developing Hedophyllum canopy
has produced a dataset with numerous gaps and variable
durations. Advances or retreats of the bed’s perimeter were
calculated as the difference in distance (cm) from initial
placement to terminal observation, normalized by duration
(months).
Figure 1 Field estimates at Tatoosh Island of the per cent cover by
Hedophyllum sessile. From July 1978 to March 1983, one to four sites
were censused, since then, six sites have been. Pisaster removal was
initiated in April 2000 with three algal census sites in each
treatment. Angular transformation of the per cent cover estimates
was followed by calculation of mean values. These have then been
back transformed and plotted. In comparison, Hedophyllum
recovery, following manual removal, was illustrated for three of
the current sites (Paine 1984).
summer, declines to an annual low by early spring and then
recovers. Figure 1 illustrates this characteristic pattern for
the upper half of the Hedophyllum band (+1.0 to +2.0 m
above tidal datum). This typical pattern showed signs of
change in summer 1997 associated with early stages of an El
Niño event, and was fully expressed by early 1998 when per
cent cover in spring reached a 20 year low. The spike in
Hedophyllum per cent cover in 1999–2000 is an artefact of
the cover estimation procedure. Mussels outcompete all
adjacent species (Paine 1974, 1984), but the process takes
time and when the mussels are small, the process is
inefficient and some Hedophyllum persist. However, as
mussels grow larger, they outcompete the algae for space.
Figure 1 illustrates the recovery, gradual in comparison with
all prior years, because it also takes time for Pisaster to
eliminate the competitively superior mussels. Where starfish
have been removed, mussels dominate and Hedophyllum is
minimally present.
The role of Pisaster
RESULTS
The shift in assemblage structure
Photographs dated 1968 and 1970 of the mid- and lower
intertidal zone identify the canopy forming alga Hedophyllum
as the most conspicuous occupant. Quantitative estimates
of per cent cover initiated in 1978 substantiate the
impression: Hedophyllum attains 80–100% cover by late
2004 Blackwell Publishing Ltd/CNRS
We quantified the cumulative effects of PisasterÕs presence or
absence by changes in the mussel bedsÕ perimeters relative
to the numbered markers. At the removal site (n ¼ 1) there
has been no net change (0.0 cm month)1, n ¼ 10 markers,
maximum months, 28). Where Pisaster has been allowed to
remain (n ¼ 4), net change is significantly different
()1.43 cm month)1, n ¼ 27, maximum months, 27;
Mann–Whitney U-test, P < 0.01). At the last observation,
Alternative states on rocky intertidal shores 443
September 2003, mussels had been eliminated from one site
with Pisaster present, were nearly gone from another, and
were dwindling at the remaining two.
Mussel bed size structure
Aggregations of mature M. californianus have collective
properties not dissimilar to those of terrestrial forests.
Individual body size renders them immune to consumption by the great majority of consumers whose attacks at
earlier life history stages prove fatal. For M. californianus to
achieve this long-term spatial dominance, three quasiindependent events must occur. First, recruitment must
be sufficient to overwhelm a host of consumers, for
instance, small carnivorous gastropods (Dayton 1971), and
at larger sizes, starfish (Paine 1976). Second, a refuge
from predation in large body size must be attainable. And
last, the numerical response of natural enemies must be
slower than the attainment of the size refuge. Figure 2
illustrates such an escape for the 5 year Pisaster removal
experiment (1963–1968) at Mukkaw Bay (Paine 1966,
1976, Paine et al. 1985). By 1968 a sufficient fraction of
the protected mussels were too large to be eaten and
therefore coexisted with high densities of Pisaster for
30 additional years. Gradual attrition of the matrix
individuals eventually lead to their local extinction. Once
eliminated, recovery has not occurred, probably due to
resident predators maintained by alternative prey. These
data and a comparable set from Tatoosh in which a
1970–1995 removal experiment (Paine 1984; R.T. Paine,
unpublished data) resulted in mussel matrices composed
of 16–24 cm individuals, document both the potential for
and reality of size escapes.
Figure 2 Long-term results at Mukkaw Bay of a Pisaster removal,
1963–1968. Field measures of mussel bed area suggest that it
persisted, although declining, for an additional 30 years despite the
presence in high density of Pisaster. 1968–1984 data from Paine
et al. (1985) have been averaged within calendar years. Data from
1984 to 2003 are new.
DISCUSSION
The 1997/1998 El Niño was exceptionally severe
(McPhaden 1999) even at Tatoosh’s latitude (49N), and
is the prime candidate for having forced the observed
changes in intertidal community structure. The mechanism(s) remain conjectural, but summer 1997 and early winter
1998 water temperatures appear to have been the highest
ever recorded. Hedophyllum is a northern kelp (Lüning &
Freshwater 1988), and thus seems a likely candidate for
stress at exceptionally high ambient temperatures. As a
presumed result, both the summer 1997 and early spring
1998 per cent cover estimates were the lowest observed in
the prior 20 years, presumably resulting from mortality of
adult plants and recruitment failures. Was this physically
forced event by itself sufficient? We believe the answer is
no. 1982/1983 was another powerful El Niño yet Hedophyllum recovered to 80–100% cover within a year (Fig. 1).
Rapid recovery also characterized the minor El Niños of
1987, 1992 and 1995/1996. What made the 1997/1998
event different was the associated recruitment events of two
other species. The barnacle Balanus glandula Darwin, 1854
recruited at exceptional densities (426 per 100 cm2, C.D.G.
Harley, personal communication), and was followed within a
year by mixed mussels, M. trossulus Gould, 1860 and
M. californianus. By summer 1999, the upper Hedophyllum
zone was being replaced by mussels, thereby meeting
successfully the first of the Connell–Sousa requirements.
A mussel bed and a stand of canopy-forming algae are
very different biological assemblages. They embody alternative states similar in basic character to those in Maine,
USA (Petraitis & Dudgeon 1999, Bertness et al. 2002). But
are any of these stable? All are, based on the criterion of
spatial persistence. On Washington’s outer coast Hedophyllum populations can persist for decades as the mid-intertidal
canopy dominant (Fig. 1). If it is a biannual species
(Widdowson 1965), individuals have turned over many
times during our lengthy observational period, suggesting
that its populations are stable as well. M. californianus
characteristically inhabits a belt immediately above this, and
although routinely stressed by winter storms, inevitably
recovers (Paine 1984). In a dynamical sense, because it
always recovers, and no alternative, non-successional states
capable of replacing it are known, mussels beds are stable.
However, the intertidal distribution of these mussels is
limited by Pisaster (see also Menge et al. 1994, Robles et al.
1995) suggesting that Hedophyllum dominance requires
PisasterÕs presence to constrain the development of mussel
beds. In addition, Hedophyllum itself is dependent on the
identity and density of its consumers, with alternative states
of coralline algae (Paine & Vadas 1969) or another kelp
(Paine 2002). In this sense, intertidal community structure
below the mussel bed is reminiscent of many other
2004 Blackwell Publishing Ltd/CNRS
444 R. T. Paine and A. C. Trimble
situations (Power 1985, Estes & Duggins 1995) in which a
clearly alternative state is maintained by the presence of
consumers. All states are persistent under these conditions.
However, none are stable because all revert to the
alternative assemblage if the consumer guild is substantially
altered. In this fashion they violate the second Connell &
Sousa (1983) criterion.
What then makes M. californianus domination different?
The answer is that, granted sufficient time, their growth in
size renders them unavailable to most consumers (Fig. 2).
The mussel bed, although perhaps not the individual
constituent mussels, survived 35 years at Mukkaw Bay
(1963–1998) and to date a minimum of 33 years (1970–
2003) at Tatoosh. However, maximum longevity and hence
turnover of all individuals remains unknown. How does this
persistence relate to stability rather than just persistence? If
M. californianus have mature gonads in their second year
(Suchanek 1981), and if we accept Slobodkin’s (1961) clock
for ecological time of 10 generations, these aggregations of
large and ÔescapedÕ mussels are stable. They have been
present for 16–18 mussel generations. However, because
longevity is unknown they fail the criterion of one turnover
of every species in the alternative assemblage (Connell &
Sousa 1983). This definition of stability is overly restrictive
because longevity of individuals in the initially dominant
population is not relevant to organisms comprising the
dominant population of the alternative state. Replacement is
independent of the lifespan of the current space occupiers; it
is dependent on whether or not the space is currently
occupied and the related ecological consequences. The
appropriate metric is whether the space remains occupied
for a sufficient period to prevent or significantly delay
recovery (the latter would demonstrate hysteresis), or
whether the temporarily novel state simply reverts to the
original condition. A biologically sensible, invader-centric,
metric would be the number of possible invasion periods,
including the time it would take for the alternative state to
fully establish, that the existing state is able to persist through.
We believe that M. californianus have persisted long
enough, using a criterion of generation time, to be
considered stable. The minimum spatial scale necessary
for persistence is a different issue. A mussel bed <5 m2
(Fig. 2) seems doomed to local extinction. At Tatoosh, the
two beds produced by the 1970–1995 removal and the
experiment currently underway (Fig. 1) both have or will
produce beds at least a factor of 10–20 greater in area. These
should persist (be stable) barring catastrophic disturbance.
Comparable examples are commonplace in anthropogenically modified terrestrial communities. Most grasslands
above a minimal precipitation of 250–1000 mm per year
would be forested without human intervention (Milchunas
et al. 1988). The lesson is that trees escape by becoming too
large; burning and apex predator suppression maintains the
2004 Blackwell Publishing Ltd/CNRS
grassland alternative state. Once escaped, the tree state is
capable of persisting many grass generations, especially if
the latter are annuals. Hysteresis aptly describes the situation
since the transition from grassland to forest, once a size
refuge has been attained, is demonstrably different than the
reverse. That is, grazer suppression could well be beneficial
to the forest whereas in grasslands it leads to rapid change.
A similar inequality characterizes the Tatoosh mussel beds
where a combination of stature and longevity permits the
establishment of a persistent and therefore stable assemblage, one which coexists without being dependent upon its
major consumer.
ACKNOWLEDGMENTS
RTP thanks the US National Science Foundation for
decades of research support. Both authors acknowledge
with gratitude the Andrew W. Mellon Foundation for
support and the Makah Tribal Council for permission to
conduct research on Tatoosh Island. We acknowledge with
thanks the clarifications and improvements made by three
anonymous reviewers.
REFERENCES
Alley, R.B., Marotzke, J., Nordhaus, W.D., Overpeck, J.T., Peteet,
D.M., Pielke, R.A., Jr et al. (2003) Abrupt climate change. Science,
299, 2005–2010.
Beisner, B.E., Haydon, D.T. & Cuddington, K. (2003) Alternative
stable states in ecology. Front. Ecol. Environ., 1, 376–382.
Bertness, M.D., Trussell, P.J., Ewancheck, P.J. & Silliman, B.R.
(2002) Do alternate stable community states exist in the Gulf of
Maine rocky intertidal zone? Ecology, 83, 3434–3448.
Brooks, J.L. & Dodson, S.I. (1965) Predation, body size, and
composition of plankton. Science, 150, 28–35.
Carpenter, S.R. (2003) Regime Shifts in Lake Ecosystems: Pattern and
Variation. Ecology Institute, Germany.
Connell, J.P. & Sousa, W.P. (1983) On the evidence needed to
judge ecological stability or persistence. Am. Nat., 121, 789–824.
Dayton, P.K. (1971) Competition, disturbance and community
organization: the provision and subsequent utilization of space
in a rocky intertidal community. Ecol. Monogr., 41, 351–389.
Dethier, M.N., Graham, E.S., Cohen, S. & Tear, L.M. (1993) Visual
versus random-point percent cover estimations: ÔobjectiveÕ is not
always better. Mar. Ecol. Prog. Ser., 96, 93–100.
Estes, J.A. & Duggins, D.O. (1995) Sea otters and kelp forests in
Alaska: generality and variation in a community ecological
paradigm. Ecol. Monogr., 65, 75–100.
Frank, P.W. (1968) Life histories and community stability. Ecology,
49, 355–356.
Lewontin, R.C. (1969) The meaning of stability. Brookhaven Symp.
Quant. Biol., 22, 13–24.
Lüning, K. & Freshwater, W. (1988) Temperature tolerance of
northeast Pacific marine algae. J. Phycol., 24, 310–315.
McPhaden, M.J. (1999) Genesis and evolution of the 1997–98
El Niño. Science, 283, 950–954.
Alternative states on rocky intertidal shores 445
Menge, B.A., Berlow, E.L., Blanchette, C.A., Navarrete, S.A. &
Yamada, S.B. (1994) The keystone species concept: variation in
interaction strength in a rocky intertidal habitat. Ecol. Monogr., 64,
249–286.
Milchunas, D.G., Sala, O.E. & Lauenroth, W.K. (1988) A generalized model of the effects of grazing by large herbivores or
grassland community structure. Am. Nat., 132, 87–106.
Paine, R.T. (1966) Food web complexity and species diversity.
Am. Nat., 100, 65–75.
Paine, R.T. (1974) Intertidal community structure. Experimental
studies on the relationship between a dominant competitor and
its principal predator. Oceologia, 15, 93–120.
Paine, R.T. (1976) Size-limited predation: an observational and
experimental approach with the Mytilus-Pisaster interaction.
Ecology, 57, 858–873.
Paine, R.T. (1984) Ecological determinism in the competition for
space. Ecology, 65, 1339–1348.
Paine, R.T. (2002) Trophic control of production in a rocky
intertidal community. Science, 296, 736–739.
Paine, R.T., Castilla, J.C. & Cancino, J. (1985) Perturbation and
recovery patterns of starfish-dominated intertidal assemblages in
Chile, New Zealand, and Washington State. Am. Nat., 125, 679–
691.
Paine, R.T. & Vadas, R.L. (1969) The effects of grazing by sea
urchins, Strongylocentrotus spp., on benthic algal populations. Limn.
Ocean., 14, 710–719.
Petraitis, P.S. & Dudgeon, S.R. (1999) Experimental evidence for
the origin of alternative communities on rocky intertidal shores.
Oikos, 84, 239–245.
Power, M.E. (1985) Grazing minnows, piscivorous bass, and
stream algae: dynamics of a strong interaction. Ecology, 66, 1448–
1456.
Robles, C., Sherwood-Stephens, R. & Alvarado, M. (1995)
Responses of a key intertidal predator to varying recruitment of
its prey. Ecology, 76, 565–579.
Scheffer, M., Carpenter, S., Foley, J.A., Folkes, C. & Walker, B.
(2001) Catastrophic shifts in ecosystems. Nature, 413, 591–596.
Scheffer, M. & Carpenter, S.R. (2003) Catatrophic regime shifts in
ecosystems: linking theory to observation. Trends Ecol. Evol., 18,
648–656.
Slobodkin, L.B. (1961) Growth and Regulation of Animal Populations
Holt. Rinehart and Winston, New York.
Suchanek, T.H. (1981) The role of disturbance in the evolution of
life history strategies in the intertidal mussels Mytilus edulus and
Mytilus californianus. Oecologia., 50, 143–152.
Sutherland, J.P. (1974) Multiple stable points in natural communities. Am. Nat., 108, 859–873.
Widdowson, T.B. (1965) A taxonomic study of the genus Hedophyllum Setchell. Can. J. Bot., 64, 1499–1506.
Editor, Thomas Miller
Manuscript received 5 February 2004
First decision made 1 March 2004
Manuscript accepted 23 March 2004
2004 Blackwell Publishing Ltd/CNRS