Huff and Jarett 07
Vol. 345: 75–82, 2007 MARINE ECOLOGY PROGRESS SERIES
Published September 13
doi: 10.3354/meps06998 Mar Ecol Prog Ser
Sand addition alters the invertebrate community of
intertidal coralline turf
Tonya M. Huff1,*, Jessica K. Jarett2
1
Scripps Institution of Oceanography, La Jolla, California 92093, USA
2
Southampton College, Long Island University, Southampton, New York 11968, USA
ABSTRACT: Many rocky intertidal areas are subject to periodic sand inundations due to a natural
cycle of sand movement that is being altered and intensified by human activities. Though sand is
thought to be a major structuring force in intertidal communities, little experimental research has
been done to investigate its effects on intertidal organisms. Assemblages of meio- and macrofauna
that inhabit intertidal coralline algal turf have been especially neglected in regards to manipulative
research on the effects of sand. In this study sand was added daily to coralline turf plots to maintain
depths of either 3 or 6 cm for 1 mo. Within 1 h of sand addition, faunal community composition had
changed significantly due to a decrease in the abundances of highly mobile animals. Another shift
was seen 2 wk later when abundances of psammophilic gastropods increased. One month after sand
addition had ceased, communities in treatment plots again resembled those of the controls. This
experiment demonstrated that turf communities rapidly respond to and recover from local physical
disturbances due to sand inundation.
KEY WORDS: Algal turf · Sand inundation · Meiofauna · Coralline algae · Rocky intertidal ·
Disturbance
Resale or republication not permitted without written consent of the publisher
INTRODUCTION 1999) and personal observations have shown densities
as high as 1.6 × 106 animals m–2 (> 63 µm in size).
In many areas worldwide, significant portions of These invertebrates form an important component of
rocky shorelines are covered by a carpet-like mat of the food chain in the rocky intertidal zone and, hence,
algal thalli referred to as algal turf. In San Diego are significant players in the system (Coull & Wells
County, California, turf is composed of a few anchor 1983, Coull 1988). According to Coull (1988), more
species that attach directly to the substrate (usually than 50 papers have been published since the early
Corallina spp.) and many epiphytes that attach to the 1970s that document the presence of meiofaunal prey
anchor species (Stewart 1982). At times, large amounts in the stomach contents of marine fish and invertebrate
of sediment can be observed within the algal mat and predators. Gut analysis of the intertidal blenny Helco-
the presence or absence of this sand is an important gramma medium indicated that amphipods were their
factor to be considered when studying the turf commu- primary prey (Coull & Wells 1983) and Hicks (1984)
nity (Stewart 1983). The complex mat of algae and found that benthic copepods were the dominant prey
associated sediment provides habitat for diverse for young flatfish. Additionally, Gosselfin & Chia
assemblages of small invertebrates and larvae (Dom- (1994) found that juvenile Nucella emarginata com-
masnes 1969, Neumann et al. 1970, Edgar 1983, Hicks monly preyed upon small bivalves such as Lasaea spp.
1985, Gibbons & Griffiths 1986, Akioka et al. 1999, and juvenile Mytilus spp. Dierschke (1994) also deter-
Kelaher et al. 2001). Abundances in excess of 200 000 mined that the main prey species of the purple sand-
animals m–2 (> 500 µm in size) have been previously piper Calidris maritima included small snails such
observed in these turf communities (Brown & Taylor as Littorina saxatilis, small crustaceans, polychaetes,
*Email: tohuff@ucsd.edu © Inter-Research 2007 · www.int-res.com
76 Mar Ecol Prog Ser 345: 75–82, 2007
juvenile Mytilus spp. and the isopod Idotea granulosa. consists mainly of Corallina pinnatifolia Daws with
All of these examples include prey species that are occasional C. officinalis Kütz as anchor species and epi-
often part of the turf community. phytic Ulva californica Wille, Gelidium spp. Lam-
Evidence suggests that 6000 yr ago most of the ouroux, Centroceras clavulatum Montagne, Leathesia
southern California shoreline was rocky habitat. With difformis Aresch, and Laurencia pacifica Kylin.
a rise in sea level sand that previously may have fallen Much of the rocky area at this site is subject to peri-
into the deep sea began to accrete on the ocean shelf odic burial by sand, ranging from a depth of several
and bury much of this rocky environment (Graham et centimeters to more than a meter (T. M. Huff unpubl.
al. 2003). This has resulted in a fragmentation of rocky data). During the duration of this study, however, very
areas and has caused a shift to an intertidal community little natural sand was present.
that appears to be moderately tolerant of episodic sand Experimental design and sampling procedure.
burial (Littler et al. 1983, 1991). Currently, many rocky Three sand treatments were applied to a total of 15
shores experience sand levels that are variable in both experimental plots: 5 shallow sand addition plots, 5
space and time. Additionally, these sand cycles are deep sand addition plots and 5 control plots to
being altered and intensified by human activities such which no sand was added. To maintain the desired
as the building of seawalls and beach replenishment sand depths for the 24 h between applications, plots
(AMEC 2002). were haphazardly sited in naturally occurring semi-
Though the dynamics of sand movement are thought enclosed circles of boulders where they were protected
to be a major structuring agent on rocky intertidal from the full force of waves. Sand treatments were
shores (e.g. Daly & Mathieson 1977, Taylor & Littler then randomly assigned to plots.
1982, McQuaid & Dower 1990, Airoldi 2003), there is a Sand of a size typical to natural inundations (mean
need for experimental work to investigate the effects of particle size <1 mm and > 500 µm) was taken from a
sand on the organisms that inhabit these shores (but nearby beach and placed on plots daily to maintain a
see Kendrick 1991, Airoldi & Cinelli 1997, and Airoldi depth of 3 cm (shallow sand treatment) or 6 cm (deep
& Virgilio 1998 for subtidal work). Limited observa- sand treatment). The tips of the algal turf remained
tional studies of the effects of sand inundation on exposed at the shallow sand depth, while turf was
assemblages of meio- and macrofauna in coralline completely covered by the deep sand treatment. Care
algal turf have been published (Kelaher et al. 2001, was taken to cover each plot with sand well outside of
Prathep et al. 2003). Kelaher et al. (2001) showed that its boundaries to reduce edge effects.
of 4 environmental variables, sediment showed the Each 0.50 × 0.75 m plot was divided into six 0.25 ×
strongest relationship with macrofaunal assemblages 0.25 m quadrats. Each of these quadrats was sampled
in coralline turf. However, experimental work was still at 1 of 6 times after initial sand addition: 1 h, 12 h, 1 d,
needed to follow up on this observation. 2 d, 2 wk or 4 wk. Samples were also taken immedi-
The goal of our study was to use experimental tech- ately before sand addition began (‘pre-impact’) and
niques to investigate the role of sediment in intertidal 1 mo after sand addition had ceased (‘recovery’). Each
coralline turf habitat, particularly in relation to com- quadrat within a plot was randomly assigned a sam-
plete burial by sand. Our primary questions were: pling time and no quadrat was sampled more than
(1) How does the coralline turf macro- and meiofauna once until recovery samples were taken. Pre-impact
community change with sand burial? (2) Which or- samples were taken from the area immediately outside
ganisms appear to be sand-tolerant or sand-intolerant? the plot frame and recovery samples were taken ran-
(3) Does the turf faunal community respond differently domly from any quadrat inside each plot. During every
to burial by different depths of sand? (4) Does the sampling period, 3 samples were taken randomly from
response of the turf faunal community to sand burial within each plot. In addition to taking algal samples,
change with time? each plot was watched for 5 min after the initial sand
addition and animals that emerged were recorded and
counted.
MATERIALS AND METHODS Samples were obtained by cutting through the turf
mat with a 4.4 cm diameter (13.8 cm2) metal coring
Study site. This experiment was conducted in the device and carefully scraping the turf from the bedrock
Scripps Coastal Reserve at Dike Rock, La Jolla, Califor- with a metal spatula. Samples were placed in tightly
nia (32° 87’ N, 117° 25’ W). Dike Rock has many boul- sealed plastic containers, taken back to the lab and
ders as well as a flat shelf of mudstone covered with immediately preserved in ethanol. They were later
coralline algal turf and is bordered on either side by rinsed on a 63 µm sieve. Samples were sorted manu-
sandy beach. Experimental plots were located on the ally with forceps under a 12× dissecting microscope.
shelf in the mid- to low intertidal zones. Turf in this area All invertebrates were removed, identified to the low-
Huff & Jarett: Sand alters coralline turf community 77
est possible taxonomic level and counted. Although computes the contribution of each species to the total
sessile animals attached to algal fronds (e.g. bryo- average dissimilarity between all pairs of inter-group
zoans, serpulorbid snails, sponges) were commonly samples. These analyses were performed using Ply-
found in the turf, these animals were not included in mouth Routines in Marine Ecological Research
the study because the methods used were not appro- (PRIMER) software v.5.2.9 (Primer-E 2002).
priate to quantify them accurately (Kelaher 2002). Abundance (N ), taxonomic richness (S) and Pielou’s
Once defaunated, the algae and sand were separated, evenness index (J’) were also calculated with PRIMER.
dried in a 60°C drying oven until a constant weight These indices were then used as response variables
was obtained (at least 24 h), and weighed. with time as a factor in additional ANOVAs to look for
To avoid bias in our results due to the inadvertent changes in abundance and diversity through time
addition of organisms to the study plots directly with within each sand treatment.
the addition of sand, samples of the sand were taken
back to the lab and inspected. Invertebrates were
removed and counted and those found in large abun- RESULTS
dances were noted.
Data analysis. For all analyses data from the 3 sam- Overview
ples taken during each time period from each plot
were averaged to give dry weights of sand and algae A total of 44 090 invertebrates from 133 taxa were
and average animal abundances. Inspection of the counted (Table 1). The taxonomic resolution of the
invertebrate community found in the sand itself along fauna varied among groups because some species
with comparison of pre- and post-impact species have not been described, others require specialized
assemblages in the turf revealed one organism, a taxonomic knowledge to identify and some were juve-
platyhelminth, which appeared to be a direct artifact niles that could not be conclusively identified. The use
of sand addition. We believe that this was the only of differing (i.e. higher) levels of taxonomic discrimina-
abundant organism imported to the plots with the tion in these types of multivariate analyses has little
sand. Therefore, it was removed from all further effect on the outcome (Herman & Heip 1988, Warwick
analyses. 1988a,b, James et al. 1995).
Because samples contained varying amounts of sand Animals were observed immediately emerging from
and algae, analyses were performed to determine if experimental plots after sand addition. Counts made
standardization of sample size was necessary (e.g. ani- during the 5 min after initial sand addition showed that
mals per dry weight algae or sand rather than animals these animals mostly included amphipods, isopods,
per sample). A multiple regression was first completed pycnogonids, hermit crabs and larger gastropods
with dry weight of sand and algae as predictor vari- (Table 2).
ables and total number of invertebrates as the
response variable. Regression coefficients showed that
the number of animals was significantly correlated Time and sand depth
with amount of algae (R2 = 0.287, p < 0.001), but not
with sand (R2 = 0.017, p = 0.080). No significant differences existed among the inver-
Then, to determine if the average amount of algae in tebrate assemblages of control, shallow and deep
each sample was significantly different among the 3 treatment plots before sand was added (ANOSIM,
treatments, an ANOVA was performed using average Table 3). With sand addition significant differences
dry weight of algae as the dependent variable and were found between both shallow and deep sand
sand treatment (shallow, deep or control) as a factor. treatments and control plots during every sampling
No significant differences were found among the interval with 2 exceptions; no significant difference
weights of algae in the 3 treatments (F2,117 = 7.191, p = was found between deep treatment and control plots in
0.110) and consequently no standardization of sample the 1 or 2 d samples (Table 3). The nMDS plots also
size was done. revealed a distinct separation between the communi-
Analysis of similarities (ANOSIM), non-metric multi- ties of control plots and those of plots to which sand
dimensional scaling (nMDS) and second stage nMDS had been added (Fig. 1). There were no significant dif-
were used to investigate patterns and quantify ferences between the communities of shallow and
changes in the turf communities. Additionally, the sim- deep sand treatment plots during any sampling period
ilarity percentages method (SIMPER) was used to (Table 3). Samples taken 1 mo after the cessation of
determine which taxa were contributing to any per- sand addition to determine the recovery response
ceived differences between samples. This type of showed no significant differences among the fauna of
analysis uses a Bray-Curtis dissimilarity matrix and shallow, deep and control plots (Table 3).
78 Mar Ecol Prog Ser 345: 75–82, 2007
Table 1. Taxa found in coralline turf samples
Phylum Class or subclass No. of taxa Highest resolution
Annelida Polychaeta 10 Family – 8, Genus – 2
Oligochaeta 1 Family – 1
Arthropoda Ostracoda 8 Genus – 6, Species – 2
Copepoda 1 Order – 1
Cirripedia 3 Genus – 2, Species – 1
Malacostraca 10 Order – 3, Suborder – 3, Genus – 2, Species – 2
Cheliceriformes 1 Suborder – 1
Pycnogonida 1 Class – 1
Insecta 1 Family – 1
Cnidaria Anthozoa 1 Genus – 1
Echinodermata Ophiuroidea 1 Class – 1
Echinoidea 1 Genus – 1
Mollusca Polyplacophora 4 Species – 4
Gastropoda 65 Order – 1, Genus – 9, Species – 55
Bivalvia 18 Family – 2, Genus – 5, Species – 11
Nematoda – 1 Phylum – 1
Platyhelminthes – 1 Phylum – 1
Sarcomastigophora Granuloreticulosea 4 Family – 4
Sipuncula – 1 Phylum – 1
Table 2. Visual estimates of numbers of invertebrates that migrated out of treatment plots within 5 min of first sand addition.
S = shallow sand treatment (3 cm), D = deep sand treatment (6 cm)
Treat- Alia sp. Amphipods Conus Fish Hermit Isopods Pachygrapsus Pycnogonids Other Total
ment califonicus crabs crassipes
S 10 30 1 0 1 0 0 0 0 42
S 25 50 0 0 4 0 0 1 1 81
S 1 100 1 1 2 0 0 0 0 105
S 10 90 1 1 1 5 1 1 1 111
S 80 35 0 0 3 1 0 6 0 125
D 30 20 0 0 3 1 0 3 0 57
D 35 20 1 0 2 2 0 4 4 68
D 40 15 0 0 6 3 2 6 1 73
D 3 65 0 0 3 0 0 5 3 79
D 135 15 0 1 6 3 0 3 0 163
Neither the shallow nor deep sand addition treat- for the majority of differences between treatment and
ment showed a different trajectory of community control plots during the early time periods (1 and 12 h,
change through time from that seen in the control plots 1 and 2 d). Abundances of these taxa show a rapid and
(second stage ANOSIM, df = 14, Global R = –0.02, p = sustained decrease with sand addition and an increase
0.539). A second stage nMDS plot (Fig. 2) also supports to near-control levels in recovery samples (Fig. 3a). A
the result that neither sand addition treatment had a second shift in community composition was seen in the
different trajectory of community change from the con- 2 and 4 wk samples when abundances of the gas-
trols. Points from all 3 treatments were relatively tropods Barleeia spp. and Amphithalamus spp. began
evenly dispersed across the plot, and no distinct to increase. These snails also returned to near-control
separations were seen. levels in recovery samples (Fig. 3b).
Total abundance (N ) significantly decreased through
time in both the shallow and deep sand addition treat-
Community response ments while evenness (J’) significantly increased
(ANOVA, Shallow N: F7,32 = 3.044, p = 0.014; Shallow
To determine which taxa were responsible for the J’: F7,32 = 2.820, p = 0.021; Deep N: F7,32 = 3.515, p =
dissimilarity between treatments a SIMPER analysis 0.007; Deep J’: F7,32 = 4.224, p = 0.002) (Fig. 4). How-
was performed. Highly mobile taxa including cope- ever, no significant differences were seen in abundance
pods, gammarid amphipods and ostracods accounted or evenness through time in the control plots (ANOVA,
Huff & Jarett: Sand alters coralline turf community 79
Table 3. Results of ANOSIM analyses to test for the effect of
sand addition. Bonferroni correction for multiple comparisons
has been applied (n = 5) and df = 14 for all tests. S = shallow
sand treatment (3 cm), D = deep sand treatment (6 cm),
C = control. *Significant value (p < 0.05)
Time Treatments R-statistic p-value
Pre-impact Global 0.094 0.132
S,D 0.056 1
S,C 0.292 0.160
D,C –0.068 1
1h Global 0.560 0.001*
S,D –0.040 1
S,C 0.768 0.040*
D,C 0.792 0.040*
12 h Global 0.608 0.001*
S,D –0.036 1
S,C 0.896 0.040*
D,C 0.828 0.040*
1d Global 0.300 0.008*
S,D 0.076 0.950
S,C 0.596 0.040*
D,C 0.096 1
2d Global 0.338 0.014*
S,D –0.04 1 Fig. 1. Representative nMDS plots. (a) nMDS plot showing
S,C 0.772 0.040* change in turf community structure 1 h after sand addition.
D,C 0.292 0.280 (b) nMDS plot showing change in turf community structure
4 wk after sand addition
2 wk Global 0.419 0.001*
S,D 0.018 1
S,C 0.925 0.040*
D,C 0.960 0.040*
4 wk Global 0.423 0.001*
S,D –0.068 1
S,C 0.588 0.040*
D,C 0.760 0.040*
Recovery Global –0.060 0.684
S,D –0.012 1
S,C –0.088 1
D,C –0.080 1
p > 0.05 in both cases). No significant differences were Fig. 2. Second stage nMDS plot indicating no difference in
seen through time in taxonomic richness (S) in any of the trajectory of community change among control and
sand-added plots
the treatments (ANOVA, p > 0.05 in all cases).
DISCUSSION mobile, sand-intolerant animals such as amphipods and
ostracods that rapidly dispersed from sand inundated
This study has established that the experimental plots. Amphipods and ostracods both tend to live in in-
addition of sand to intertidal coralline turf has almost terstitial spaces of the algal turf (Coull & Wells 1983,
immediate and sustained effects on the associated Gibbons 1988). It is possible that the addition of sedi-
meio- and macrofauna. Two distinct shifts in commu- ment clogged the coralline algae, thus eliminating their
nity composition were seen with sand inundation: a spatial niche and refuge from predators (Coull & Wells
rapid exodus of mobile sand-intolerant animals and a 1983, Dean & Connell 1987). While they have an ex-
more gradual increase in psammophilic (‘sand-loving’) oskeleton, these animals are not protected by a hard
gastropods. As early as 1 h after sand addition, signifi- shell and increased scour associated with sand addition
cant differences were seen between control and treat- could also be a cause of their decrease in abundance.
ment plots. Both observation and statistical analyses A second difference in community composition be-
suggest that these differences were caused by highly tween treatment and control plots was apparent begin-
80 Mar Ecol Prog Ser 345: 75–82, 2007
90 50
a a
Average number of taxa
80 45
Average number
70 40
per sample
60 35
50 30
40 25
30 20
15
20
10
10
5
0
0
14
Average number of animals
b 400
12 b
Average number
350
per sample
10 300
8 250
6 200
150
4
100
2
50
0
Pre 1h 12 h 24 h 2d 2 wk 4 wk Rec 0
Sampling time
0.90
Control Shallow Deep c
0.85
Average evenness
Fig. 3. (a) Average copepod abundance through time for each
sand treatment. (b) Average Amphithalamus inclusus abun- 0.80
dance through time for each sand treatment. Error bars
0.75
indicate SE and n = 5
0.70
ning in the 2 wk samples when abundances of the 0.65
snails Amphithalamus tenuis, A. inclusus and Barleeia
0.60
spp. showed significant increases in treatment plots. Pre 1h 12 h 24 h 2d 2 wk 4 wk Rec
Microgastropods are able to move about and disperse Sampling time
into new habitats as adults within a period of days or
Control Shallow Deep
weeks (Olabarria & Chapman 2001, Olabarria 2002),
so they may respond to habitat changes and move to Fig. 4. (a) Average taxonomic richness (S) through time
preferred sites. Amphithalamus spp. are commonly for each sand treatment. (b) Average total abundance (N )
reported to be positively correlated with the presence through time for each sand treatment. (c) Average Pielou’s
of sediment (Olabarria & Chapman 2001, Kelaher et al. evenness (J’) through time for each sand treatment. Error bars
indicate SE and n = 5
2003) and Barleeia spp. also tend to have higher abun-
dances when more sediment is present (T. M. Huff
pers. obs.). added and control plots. This may be because the
No significant differences were seen between the majority of community change in treatment plots
communities of the shallow and deep sand addition occurred in 2 pulses. Between these events control and
plots. The 2 depths were employed to determine if the treatment plots would have been subject to similar nat-
community responds differently to different levels of ural community fluctuations due to variables such as
sand burial. While organisms did not distinguish settlement events and disturbances and, therefore,
between the 2 sand depths employed in this study, would have had similar trajectories of change.
deeper sand depths might produce other changes in The anomalous non-significant data points seen in
the community. Given the immediate response of the deep treatment plots for 1 and 2 d samples deserve
sand-intolerant organisms to sand addition in this some consideration. A random number chart was
study it is also possible that even minimal levels of employed when plots were assigned a particular sand
sand may affect turf communities. treatment. In hindsight, we noticed that several of the
No significant differences were seen among the tra- deep treatment plots were located in more energetic
jectories of community change through time for sand- areas with more water flow than were the shallow
Huff & Jarett: Sand alters coralline turf community 81
plots. In the short term (i.e. 1 and 2 d samples), this may Vilchis and C. Catton provided indispensable statistical
have changed the impact of the sand addition. advice. Many wet, dark early mornings were spent in the
intertidal by our field assistants D. Taniguchi, J. Oswald, B.
It appears that although sand addition may not sig-
Pister and C. Gonzales. Finally, we acknowledge the impor-
nificantly alter the number of taxa living in an area, it tant support of A. Knight, I. Castillo, S. Rouse, K. Riser, J.
does alter the relative abundances of these taxa. Since Shaffer, S. Malagong, A. Bachter, J. Cattalano, L. Rouse, H.
numerically dominant mobile taxa like copepods dra- Huff, N. and D. Tortellini and D. Shaffer.
matically decreased in sand plots as compared with
control and pre-impact samples (Fig. 3a), we saw a sig- LITERATURE CITED
nificant decrease in total abundance (N ) of organisms
in our samples through time in the sand-added plots Airoldi L (2003) The effects of sedimentation on rocky coast
(Fig. 4). Additionally, since we saw an increase in assemblages. Oceanogr Mar Biol Annu Rev 41:161–236
Airoldi L, Cinelli F (1997) Effects of sedimentation on subtidal
abundance of scarcer, sand-tolerant taxa in sand plots
macroalgal assemblages: an experimental study from a
as the more dominant taxa decreased, an increase in Mediterranean rocky shore. J Exp Mar Biol Ecol 215:
Pielou’s evenness index (J’) in these plots makes sense. 269–288
The results of this experiment are significant not only Airoldi L, Virgilio M (1998) Responses of turf-forming algae to
because such dramatic changes were seen in the meio- spatial variations in the deposition of sediments. Mar Ecol
Prog Ser 165:271–282
and macrofaunal communities, but also because these Akioka H, Baba M, Masaki T, Johansen HW (1999) Rocky
shifts have the potential to be seen up through the shore turfs dominated by Corallina (Corallinales, Rhodo-
marine food web. As stated previously, meiofauna are phyta) in northern Japan. Phycol Res 47:199–206
important prey for many species of fish and inverte- AMEC Earth and Environment Inc. (2002) Regional beach
sand project post-construction monitoring report for inter-
brates. Small macrofauna like those found in turf are
tidal, shallow subtidal, and kelp forest resources. Pre-
also frequently reported as prey items for larger preda- pared for San Diego Association of Governments
tors (Gibbons & Griffiths 1986, Dugan et al. 2002). (SANDAG), San Diego, CA
Additionally, meiofauna may play an important role in Brown PJ, Taylor RB (1999) Effects of trampling by humans on
making detritus available to macroconsumers either animals inhabiting coralline algal turf in the rocky inter-
tidal. J Exp Mar Biol Ecol 235:45–53
through their enhancement of microbial activity or by Coull BC (1988) Ecology of the marine meiofauna. In: Higgins
ingestion of the meiofauna themselves (Coull 1988). RP, Thiel H (eds) Introduction to the study of meiofauna.
We believe that the investigation of the potential Smithsonian Institute Press, Washington, DC, p 18–38
effects of changes in the turf community on the larger Coull BC, Wells JBJ (1983) Refuges from fish predation —
experiments with phytal meiofauna from the New
coastal food web is an important area for future
Zealand rocky intertidal. Ecology 64:1599–1609
research. Daly MA, Mathieson AC (1977) Effects of sand movement on
Many questions still remain unanswered about turf intertidal seaweeds and selected invertebrates at Bound
communities in relation to periodic sand inundation. Rock, New Hampshire, USA. Mar Biol 43:45–55
This experiment showed a rapid response of turf fauna Dean RL, Connell JH (1987) Marine invertebrates in algal
succession. III. Mechanisms linking habitat complexity
followed by a relatively rapid recovery as well. How- with diversity. J Exp Mar Biol Ecol 109:249–273
ever, our experimental plots were small in scale Dierschke V (1994) Food and feeding ecology of purple sand-
compared with the broad areas of habitat that can be pipers Calidris maritima on rocky intertidal habitats (Hel-
covered with sand during a natural inundation event. goland, German Bight). Neth J Sea Res 31:309–317
Dommasnes A (1969) On the fauna of Corallina officinalis L.
Recovery of such a large area might take significantly
in western Norway. Sarsia 38:71–86
longer since animals may crawl in from the edges of a Dugan JE, Hubbard DM, McCrary MD, Pierson MO (2002)
small experimental plot more easily than they could if The response of macrofauna communities and shorebirds
an entire stretch of habitat was inundated. Addition- to macrophyte wrack subsidies on exposed sandy
ally, this study did not examine how the community is beaches of southern California. Estuar Coast Shelf Sci
58S:133–148
affected by anoxia caused by sand burial or by very Edgar GJ (1983) The ecology of south-east Tasmanian phytal
long-term inundation events. These are also likely to animal communities. I. Spatial organization on a local
be fruitful areas for future research. scale. J Exp Mar Biol Ecol 70:129–157
Gibbons MJ (1988) The impact of sediment accumulations,
relative habitat complexity, and elevation on rocky shore
Acknowledgements. This work was funded by California Sea meiofauna. J Exp Mar Biol Ecol 122:225–241
Grant, the Edna Bailey Sussman Fund and a San Diego Foun- Gibbons MJ, Griffiths CL (1986) A comparison of macrofaunal
dation Blasker Environment Grant. The work was done while and meiofaunal distribution and standing stock across a
J.K.J. performed an internship at Scripps made possible by rocky shore, with an estimate of their productivities. Mar
Dr. L. B. Liddle and P. Jackson. We are grateful for permission Biol 93:181–188
from reserve manager I. Kay to complete the work in the Gosselfin LA, Chia FS (1994) Feeding habits of newly hatched
Scripps Coastal Reserve within the University of California juveniles of an intertidal predatory gastropod, Nucella
Natural Reserve System. We thank P. Dayton and J. Leichter emarginata (Deshayes). J Exp Mar Biol Ecol 176:1–13
for valuable guidance, advice and support. P. E. Parnell, L. Graham MH, Dayton PK, Erlandson JM (2003) Ice ages and
82 Mar Ecol Prog Ser 345: 75–82, 2007
ecological transitions on temperate coasts. Trends Ecol Littler MM, Littler DS, Murray SN, Seapy RR (1991) Southern
Evol 18:33–40 California rocky intertidal ecosystems. In: Mathieson AC,
Herman PMJ, Heip C (1988) On the use of meiofauna in eco- Nienhuis PH (eds) Ecosystems of the world. Vol 24. Else-
logical monitoring: Who needs taxonomy? Mar Pollut Bull vier, New York, p 273–296
19:665–668 McQuaid CD, Dower KM (1990) Enhancement of habitat het-
Hicks GRF (1984) Spatio-temporal dynamics of a meiobenthic erogeneity and species richness on rocky shores inun-
copepod and the impact of predation disturbance. J Exp dated by sand. Oecologia 84:142–144
Mar Biol Ecol 81:47–72 Neumann AC, Gebelein CD, Scoffin TP (1970) The composi-
Hicks GRF (1985) Meiofauna associated with rocky shore tion, structure, and erodability of subtidal mats, Abaco,
algae. In: Moore PG, Seed R (eds) Ecology of rocky coasts. Bahamas. J Sediment Petrol 40:274–297
Hodder and Stoughton, London, p 36–56 Olabarria C (2002) Role of colonization in spatio-temporal
James RJ, Lincoln Smith MP, Fairweather PG (1995) Sieve patchiness of microgastropods in coralline turf habitat.
mesh-size and taxonomic resolution needed to describe J Exp Mar Biol Ecol 274:121–140
natural spatial variation of marine macrofauna. Mar Ecol Olabarria C, Chapman MG (2001) Comparisons of patterns of
Prog Ser 118:187–198 spatial variation of microgastropods between two con-
Kelaher BP (2002) Influence of physical characteristics of trasting intertidal habitats. Mar Ecol Prog Ser 220:201–211
coralline turf on associated macrofaunal assemblages. Prathep AR, Marrs H, Norton TA (2003) Spatial and temporal
Mar Ecol Prog Ser 232:141–148 variations in sediment accumulation in an algal turf and
Kelaher BP, Chapman MG, Underwood AJ (2001) Spatial pat- their impact on associated fauna. Mar Biol 142:381–390
terns of diverse macrofaunal assemblages in coralline turf Stewart JG (1982) Anchor species and epiphytes in intertidal
and their associations with environmental variables. J Mar algal turf. Pac Sci 36:45–59
Biol Assoc UK 81:917–930 Stewart JG (1983) Fluctuations in the quantity of sediments
Kelaher BP, Underwood AJ, Chapman MG (2003) Experimen- trapped among algal thalli on intertidal rock platforms in
tal transplantations of coralline algal turf to demonstrate southern California. J Exp Mar Biol Ecol 73:205–211
causes of differences in macrofauna at different tidal Taylor PR, Littler MM (1982) The roles of compensatory mor-
heights. J Exp Mar Biol Ecol 282:23–41 tality, physical disturbance, and substrate retention in the
Kendrick GA (1991) Recruitment of coralline crusts and fila- development and organization of a sand-influenced, rocky
mentous turf algae in the Galapagos Archipelago — effect intertidal community. Ecology 63:135–146
of simulated scour, erosion, and accretion. J Exp Mar Biol Warwick RM (1988a) The level of taxonomic discrimination
Ecol 147:47–63 required to detect pollution effects on marine benthic
Littler MM, Martz DR, Littler DS (1983) Effects of recurrent communities. Mar Pollut Bull 19:259–268
sand deposition on rocky intertidal organisms: importance Warwick RM (1988b) Analysis of community attributes of the
of substrate heterogeneity in a fluctuating environment. macrobenthos of Frierfjord/Langesundfjord at taxonomic
Mar Ecol Prog Ser 11:129–139 levels higher than species. Mar Ecol Prog Ser 46:167–170
Editorial responsibility: Howard Browman (Associate Editor- Submitted: June 8, 2006; Accepted: March 27, 2007
in-Chief), Storebø, Norway Proofs received from author(s): August 30, 2007
Published September 13
doi: 10.3354/meps06998 Mar Ecol Prog Ser
Sand addition alters the invertebrate community of
intertidal coralline turf
Tonya M. Huff1,*, Jessica K. Jarett2
1
Scripps Institution of Oceanography, La Jolla, California 92093, USA
2
Southampton College, Long Island University, Southampton, New York 11968, USA
ABSTRACT: Many rocky intertidal areas are subject to periodic sand inundations due to a natural
cycle of sand movement that is being altered and intensified by human activities. Though sand is
thought to be a major structuring force in intertidal communities, little experimental research has
been done to investigate its effects on intertidal organisms. Assemblages of meio- and macrofauna
that inhabit intertidal coralline algal turf have been especially neglected in regards to manipulative
research on the effects of sand. In this study sand was added daily to coralline turf plots to maintain
depths of either 3 or 6 cm for 1 mo. Within 1 h of sand addition, faunal community composition had
changed significantly due to a decrease in the abundances of highly mobile animals. Another shift
was seen 2 wk later when abundances of psammophilic gastropods increased. One month after sand
addition had ceased, communities in treatment plots again resembled those of the controls. This
experiment demonstrated that turf communities rapidly respond to and recover from local physical
disturbances due to sand inundation.
KEY WORDS: Algal turf · Sand inundation · Meiofauna · Coralline algae · Rocky intertidal ·
Disturbance
Resale or republication not permitted without written consent of the publisher
INTRODUCTION 1999) and personal observations have shown densities
as high as 1.6 × 106 animals m–2 (> 63 µm in size).
In many areas worldwide, significant portions of These invertebrates form an important component of
rocky shorelines are covered by a carpet-like mat of the food chain in the rocky intertidal zone and, hence,
algal thalli referred to as algal turf. In San Diego are significant players in the system (Coull & Wells
County, California, turf is composed of a few anchor 1983, Coull 1988). According to Coull (1988), more
species that attach directly to the substrate (usually than 50 papers have been published since the early
Corallina spp.) and many epiphytes that attach to the 1970s that document the presence of meiofaunal prey
anchor species (Stewart 1982). At times, large amounts in the stomach contents of marine fish and invertebrate
of sediment can be observed within the algal mat and predators. Gut analysis of the intertidal blenny Helco-
the presence or absence of this sand is an important gramma medium indicated that amphipods were their
factor to be considered when studying the turf commu- primary prey (Coull & Wells 1983) and Hicks (1984)
nity (Stewart 1983). The complex mat of algae and found that benthic copepods were the dominant prey
associated sediment provides habitat for diverse for young flatfish. Additionally, Gosselfin & Chia
assemblages of small invertebrates and larvae (Dom- (1994) found that juvenile Nucella emarginata com-
masnes 1969, Neumann et al. 1970, Edgar 1983, Hicks monly preyed upon small bivalves such as Lasaea spp.
1985, Gibbons & Griffiths 1986, Akioka et al. 1999, and juvenile Mytilus spp. Dierschke (1994) also deter-
Kelaher et al. 2001). Abundances in excess of 200 000 mined that the main prey species of the purple sand-
animals m–2 (> 500 µm in size) have been previously piper Calidris maritima included small snails such
observed in these turf communities (Brown & Taylor as Littorina saxatilis, small crustaceans, polychaetes,
*Email: tohuff@ucsd.edu © Inter-Research 2007 · www.int-res.com
76 Mar Ecol Prog Ser 345: 75–82, 2007
juvenile Mytilus spp. and the isopod Idotea granulosa. consists mainly of Corallina pinnatifolia Daws with
All of these examples include prey species that are occasional C. officinalis Kütz as anchor species and epi-
often part of the turf community. phytic Ulva californica Wille, Gelidium spp. Lam-
Evidence suggests that 6000 yr ago most of the ouroux, Centroceras clavulatum Montagne, Leathesia
southern California shoreline was rocky habitat. With difformis Aresch, and Laurencia pacifica Kylin.
a rise in sea level sand that previously may have fallen Much of the rocky area at this site is subject to peri-
into the deep sea began to accrete on the ocean shelf odic burial by sand, ranging from a depth of several
and bury much of this rocky environment (Graham et centimeters to more than a meter (T. M. Huff unpubl.
al. 2003). This has resulted in a fragmentation of rocky data). During the duration of this study, however, very
areas and has caused a shift to an intertidal community little natural sand was present.
that appears to be moderately tolerant of episodic sand Experimental design and sampling procedure.
burial (Littler et al. 1983, 1991). Currently, many rocky Three sand treatments were applied to a total of 15
shores experience sand levels that are variable in both experimental plots: 5 shallow sand addition plots, 5
space and time. Additionally, these sand cycles are deep sand addition plots and 5 control plots to
being altered and intensified by human activities such which no sand was added. To maintain the desired
as the building of seawalls and beach replenishment sand depths for the 24 h between applications, plots
(AMEC 2002). were haphazardly sited in naturally occurring semi-
Though the dynamics of sand movement are thought enclosed circles of boulders where they were protected
to be a major structuring agent on rocky intertidal from the full force of waves. Sand treatments were
shores (e.g. Daly & Mathieson 1977, Taylor & Littler then randomly assigned to plots.
1982, McQuaid & Dower 1990, Airoldi 2003), there is a Sand of a size typical to natural inundations (mean
need for experimental work to investigate the effects of particle size <1 mm and > 500 µm) was taken from a
sand on the organisms that inhabit these shores (but nearby beach and placed on plots daily to maintain a
see Kendrick 1991, Airoldi & Cinelli 1997, and Airoldi depth of 3 cm (shallow sand treatment) or 6 cm (deep
& Virgilio 1998 for subtidal work). Limited observa- sand treatment). The tips of the algal turf remained
tional studies of the effects of sand inundation on exposed at the shallow sand depth, while turf was
assemblages of meio- and macrofauna in coralline completely covered by the deep sand treatment. Care
algal turf have been published (Kelaher et al. 2001, was taken to cover each plot with sand well outside of
Prathep et al. 2003). Kelaher et al. (2001) showed that its boundaries to reduce edge effects.
of 4 environmental variables, sediment showed the Each 0.50 × 0.75 m plot was divided into six 0.25 ×
strongest relationship with macrofaunal assemblages 0.25 m quadrats. Each of these quadrats was sampled
in coralline turf. However, experimental work was still at 1 of 6 times after initial sand addition: 1 h, 12 h, 1 d,
needed to follow up on this observation. 2 d, 2 wk or 4 wk. Samples were also taken immedi-
The goal of our study was to use experimental tech- ately before sand addition began (‘pre-impact’) and
niques to investigate the role of sediment in intertidal 1 mo after sand addition had ceased (‘recovery’). Each
coralline turf habitat, particularly in relation to com- quadrat within a plot was randomly assigned a sam-
plete burial by sand. Our primary questions were: pling time and no quadrat was sampled more than
(1) How does the coralline turf macro- and meiofauna once until recovery samples were taken. Pre-impact
community change with sand burial? (2) Which or- samples were taken from the area immediately outside
ganisms appear to be sand-tolerant or sand-intolerant? the plot frame and recovery samples were taken ran-
(3) Does the turf faunal community respond differently domly from any quadrat inside each plot. During every
to burial by different depths of sand? (4) Does the sampling period, 3 samples were taken randomly from
response of the turf faunal community to sand burial within each plot. In addition to taking algal samples,
change with time? each plot was watched for 5 min after the initial sand
addition and animals that emerged were recorded and
counted.
MATERIALS AND METHODS Samples were obtained by cutting through the turf
mat with a 4.4 cm diameter (13.8 cm2) metal coring
Study site. This experiment was conducted in the device and carefully scraping the turf from the bedrock
Scripps Coastal Reserve at Dike Rock, La Jolla, Califor- with a metal spatula. Samples were placed in tightly
nia (32° 87’ N, 117° 25’ W). Dike Rock has many boul- sealed plastic containers, taken back to the lab and
ders as well as a flat shelf of mudstone covered with immediately preserved in ethanol. They were later
coralline algal turf and is bordered on either side by rinsed on a 63 µm sieve. Samples were sorted manu-
sandy beach. Experimental plots were located on the ally with forceps under a 12× dissecting microscope.
shelf in the mid- to low intertidal zones. Turf in this area All invertebrates were removed, identified to the low-
Huff & Jarett: Sand alters coralline turf community 77
est possible taxonomic level and counted. Although computes the contribution of each species to the total
sessile animals attached to algal fronds (e.g. bryo- average dissimilarity between all pairs of inter-group
zoans, serpulorbid snails, sponges) were commonly samples. These analyses were performed using Ply-
found in the turf, these animals were not included in mouth Routines in Marine Ecological Research
the study because the methods used were not appro- (PRIMER) software v.5.2.9 (Primer-E 2002).
priate to quantify them accurately (Kelaher 2002). Abundance (N ), taxonomic richness (S) and Pielou’s
Once defaunated, the algae and sand were separated, evenness index (J’) were also calculated with PRIMER.
dried in a 60°C drying oven until a constant weight These indices were then used as response variables
was obtained (at least 24 h), and weighed. with time as a factor in additional ANOVAs to look for
To avoid bias in our results due to the inadvertent changes in abundance and diversity through time
addition of organisms to the study plots directly with within each sand treatment.
the addition of sand, samples of the sand were taken
back to the lab and inspected. Invertebrates were
removed and counted and those found in large abun- RESULTS
dances were noted.
Data analysis. For all analyses data from the 3 sam- Overview
ples taken during each time period from each plot
were averaged to give dry weights of sand and algae A total of 44 090 invertebrates from 133 taxa were
and average animal abundances. Inspection of the counted (Table 1). The taxonomic resolution of the
invertebrate community found in the sand itself along fauna varied among groups because some species
with comparison of pre- and post-impact species have not been described, others require specialized
assemblages in the turf revealed one organism, a taxonomic knowledge to identify and some were juve-
platyhelminth, which appeared to be a direct artifact niles that could not be conclusively identified. The use
of sand addition. We believe that this was the only of differing (i.e. higher) levels of taxonomic discrimina-
abundant organism imported to the plots with the tion in these types of multivariate analyses has little
sand. Therefore, it was removed from all further effect on the outcome (Herman & Heip 1988, Warwick
analyses. 1988a,b, James et al. 1995).
Because samples contained varying amounts of sand Animals were observed immediately emerging from
and algae, analyses were performed to determine if experimental plots after sand addition. Counts made
standardization of sample size was necessary (e.g. ani- during the 5 min after initial sand addition showed that
mals per dry weight algae or sand rather than animals these animals mostly included amphipods, isopods,
per sample). A multiple regression was first completed pycnogonids, hermit crabs and larger gastropods
with dry weight of sand and algae as predictor vari- (Table 2).
ables and total number of invertebrates as the
response variable. Regression coefficients showed that
the number of animals was significantly correlated Time and sand depth
with amount of algae (R2 = 0.287, p < 0.001), but not
with sand (R2 = 0.017, p = 0.080). No significant differences existed among the inver-
Then, to determine if the average amount of algae in tebrate assemblages of control, shallow and deep
each sample was significantly different among the 3 treatment plots before sand was added (ANOSIM,
treatments, an ANOVA was performed using average Table 3). With sand addition significant differences
dry weight of algae as the dependent variable and were found between both shallow and deep sand
sand treatment (shallow, deep or control) as a factor. treatments and control plots during every sampling
No significant differences were found among the interval with 2 exceptions; no significant difference
weights of algae in the 3 treatments (F2,117 = 7.191, p = was found between deep treatment and control plots in
0.110) and consequently no standardization of sample the 1 or 2 d samples (Table 3). The nMDS plots also
size was done. revealed a distinct separation between the communi-
Analysis of similarities (ANOSIM), non-metric multi- ties of control plots and those of plots to which sand
dimensional scaling (nMDS) and second stage nMDS had been added (Fig. 1). There were no significant dif-
were used to investigate patterns and quantify ferences between the communities of shallow and
changes in the turf communities. Additionally, the sim- deep sand treatment plots during any sampling period
ilarity percentages method (SIMPER) was used to (Table 3). Samples taken 1 mo after the cessation of
determine which taxa were contributing to any per- sand addition to determine the recovery response
ceived differences between samples. This type of showed no significant differences among the fauna of
analysis uses a Bray-Curtis dissimilarity matrix and shallow, deep and control plots (Table 3).
78 Mar Ecol Prog Ser 345: 75–82, 2007
Table 1. Taxa found in coralline turf samples
Phylum Class or subclass No. of taxa Highest resolution
Annelida Polychaeta 10 Family – 8, Genus – 2
Oligochaeta 1 Family – 1
Arthropoda Ostracoda 8 Genus – 6, Species – 2
Copepoda 1 Order – 1
Cirripedia 3 Genus – 2, Species – 1
Malacostraca 10 Order – 3, Suborder – 3, Genus – 2, Species – 2
Cheliceriformes 1 Suborder – 1
Pycnogonida 1 Class – 1
Insecta 1 Family – 1
Cnidaria Anthozoa 1 Genus – 1
Echinodermata Ophiuroidea 1 Class – 1
Echinoidea 1 Genus – 1
Mollusca Polyplacophora 4 Species – 4
Gastropoda 65 Order – 1, Genus – 9, Species – 55
Bivalvia 18 Family – 2, Genus – 5, Species – 11
Nematoda – 1 Phylum – 1
Platyhelminthes – 1 Phylum – 1
Sarcomastigophora Granuloreticulosea 4 Family – 4
Sipuncula – 1 Phylum – 1
Table 2. Visual estimates of numbers of invertebrates that migrated out of treatment plots within 5 min of first sand addition.
S = shallow sand treatment (3 cm), D = deep sand treatment (6 cm)
Treat- Alia sp. Amphipods Conus Fish Hermit Isopods Pachygrapsus Pycnogonids Other Total
ment califonicus crabs crassipes
S 10 30 1 0 1 0 0 0 0 42
S 25 50 0 0 4 0 0 1 1 81
S 1 100 1 1 2 0 0 0 0 105
S 10 90 1 1 1 5 1 1 1 111
S 80 35 0 0 3 1 0 6 0 125
D 30 20 0 0 3 1 0 3 0 57
D 35 20 1 0 2 2 0 4 4 68
D 40 15 0 0 6 3 2 6 1 73
D 3 65 0 0 3 0 0 5 3 79
D 135 15 0 1 6 3 0 3 0 163
Neither the shallow nor deep sand addition treat- for the majority of differences between treatment and
ment showed a different trajectory of community control plots during the early time periods (1 and 12 h,
change through time from that seen in the control plots 1 and 2 d). Abundances of these taxa show a rapid and
(second stage ANOSIM, df = 14, Global R = –0.02, p = sustained decrease with sand addition and an increase
0.539). A second stage nMDS plot (Fig. 2) also supports to near-control levels in recovery samples (Fig. 3a). A
the result that neither sand addition treatment had a second shift in community composition was seen in the
different trajectory of community change from the con- 2 and 4 wk samples when abundances of the gas-
trols. Points from all 3 treatments were relatively tropods Barleeia spp. and Amphithalamus spp. began
evenly dispersed across the plot, and no distinct to increase. These snails also returned to near-control
separations were seen. levels in recovery samples (Fig. 3b).
Total abundance (N ) significantly decreased through
time in both the shallow and deep sand addition treat-
Community response ments while evenness (J’) significantly increased
(ANOVA, Shallow N: F7,32 = 3.044, p = 0.014; Shallow
To determine which taxa were responsible for the J’: F7,32 = 2.820, p = 0.021; Deep N: F7,32 = 3.515, p =
dissimilarity between treatments a SIMPER analysis 0.007; Deep J’: F7,32 = 4.224, p = 0.002) (Fig. 4). How-
was performed. Highly mobile taxa including cope- ever, no significant differences were seen in abundance
pods, gammarid amphipods and ostracods accounted or evenness through time in the control plots (ANOVA,
Huff & Jarett: Sand alters coralline turf community 79
Table 3. Results of ANOSIM analyses to test for the effect of
sand addition. Bonferroni correction for multiple comparisons
has been applied (n = 5) and df = 14 for all tests. S = shallow
sand treatment (3 cm), D = deep sand treatment (6 cm),
C = control. *Significant value (p < 0.05)
Time Treatments R-statistic p-value
Pre-impact Global 0.094 0.132
S,D 0.056 1
S,C 0.292 0.160
D,C –0.068 1
1h Global 0.560 0.001*
S,D –0.040 1
S,C 0.768 0.040*
D,C 0.792 0.040*
12 h Global 0.608 0.001*
S,D –0.036 1
S,C 0.896 0.040*
D,C 0.828 0.040*
1d Global 0.300 0.008*
S,D 0.076 0.950
S,C 0.596 0.040*
D,C 0.096 1
2d Global 0.338 0.014*
S,D –0.04 1 Fig. 1. Representative nMDS plots. (a) nMDS plot showing
S,C 0.772 0.040* change in turf community structure 1 h after sand addition.
D,C 0.292 0.280 (b) nMDS plot showing change in turf community structure
4 wk after sand addition
2 wk Global 0.419 0.001*
S,D 0.018 1
S,C 0.925 0.040*
D,C 0.960 0.040*
4 wk Global 0.423 0.001*
S,D –0.068 1
S,C 0.588 0.040*
D,C 0.760 0.040*
Recovery Global –0.060 0.684
S,D –0.012 1
S,C –0.088 1
D,C –0.080 1
p > 0.05 in both cases). No significant differences were Fig. 2. Second stage nMDS plot indicating no difference in
seen through time in taxonomic richness (S) in any of the trajectory of community change among control and
sand-added plots
the treatments (ANOVA, p > 0.05 in all cases).
DISCUSSION mobile, sand-intolerant animals such as amphipods and
ostracods that rapidly dispersed from sand inundated
This study has established that the experimental plots. Amphipods and ostracods both tend to live in in-
addition of sand to intertidal coralline turf has almost terstitial spaces of the algal turf (Coull & Wells 1983,
immediate and sustained effects on the associated Gibbons 1988). It is possible that the addition of sedi-
meio- and macrofauna. Two distinct shifts in commu- ment clogged the coralline algae, thus eliminating their
nity composition were seen with sand inundation: a spatial niche and refuge from predators (Coull & Wells
rapid exodus of mobile sand-intolerant animals and a 1983, Dean & Connell 1987). While they have an ex-
more gradual increase in psammophilic (‘sand-loving’) oskeleton, these animals are not protected by a hard
gastropods. As early as 1 h after sand addition, signifi- shell and increased scour associated with sand addition
cant differences were seen between control and treat- could also be a cause of their decrease in abundance.
ment plots. Both observation and statistical analyses A second difference in community composition be-
suggest that these differences were caused by highly tween treatment and control plots was apparent begin-
80 Mar Ecol Prog Ser 345: 75–82, 2007
90 50
a a
Average number of taxa
80 45
Average number
70 40
per sample
60 35
50 30
40 25
30 20
15
20
10
10
5
0
0
14
Average number of animals
b 400
12 b
Average number
350
per sample
10 300
8 250
6 200
150
4
100
2
50
0
Pre 1h 12 h 24 h 2d 2 wk 4 wk Rec 0
Sampling time
0.90
Control Shallow Deep c
0.85
Average evenness
Fig. 3. (a) Average copepod abundance through time for each
sand treatment. (b) Average Amphithalamus inclusus abun- 0.80
dance through time for each sand treatment. Error bars
0.75
indicate SE and n = 5
0.70
ning in the 2 wk samples when abundances of the 0.65
snails Amphithalamus tenuis, A. inclusus and Barleeia
0.60
spp. showed significant increases in treatment plots. Pre 1h 12 h 24 h 2d 2 wk 4 wk Rec
Microgastropods are able to move about and disperse Sampling time
into new habitats as adults within a period of days or
Control Shallow Deep
weeks (Olabarria & Chapman 2001, Olabarria 2002),
so they may respond to habitat changes and move to Fig. 4. (a) Average taxonomic richness (S) through time
preferred sites. Amphithalamus spp. are commonly for each sand treatment. (b) Average total abundance (N )
reported to be positively correlated with the presence through time for each sand treatment. (c) Average Pielou’s
of sediment (Olabarria & Chapman 2001, Kelaher et al. evenness (J’) through time for each sand treatment. Error bars
indicate SE and n = 5
2003) and Barleeia spp. also tend to have higher abun-
dances when more sediment is present (T. M. Huff
pers. obs.). added and control plots. This may be because the
No significant differences were seen between the majority of community change in treatment plots
communities of the shallow and deep sand addition occurred in 2 pulses. Between these events control and
plots. The 2 depths were employed to determine if the treatment plots would have been subject to similar nat-
community responds differently to different levels of ural community fluctuations due to variables such as
sand burial. While organisms did not distinguish settlement events and disturbances and, therefore,
between the 2 sand depths employed in this study, would have had similar trajectories of change.
deeper sand depths might produce other changes in The anomalous non-significant data points seen in
the community. Given the immediate response of the deep treatment plots for 1 and 2 d samples deserve
sand-intolerant organisms to sand addition in this some consideration. A random number chart was
study it is also possible that even minimal levels of employed when plots were assigned a particular sand
sand may affect turf communities. treatment. In hindsight, we noticed that several of the
No significant differences were seen among the tra- deep treatment plots were located in more energetic
jectories of community change through time for sand- areas with more water flow than were the shallow
Huff & Jarett: Sand alters coralline turf community 81
plots. In the short term (i.e. 1 and 2 d samples), this may Vilchis and C. Catton provided indispensable statistical
have changed the impact of the sand addition. advice. Many wet, dark early mornings were spent in the
intertidal by our field assistants D. Taniguchi, J. Oswald, B.
It appears that although sand addition may not sig-
Pister and C. Gonzales. Finally, we acknowledge the impor-
nificantly alter the number of taxa living in an area, it tant support of A. Knight, I. Castillo, S. Rouse, K. Riser, J.
does alter the relative abundances of these taxa. Since Shaffer, S. Malagong, A. Bachter, J. Cattalano, L. Rouse, H.
numerically dominant mobile taxa like copepods dra- Huff, N. and D. Tortellini and D. Shaffer.
matically decreased in sand plots as compared with
control and pre-impact samples (Fig. 3a), we saw a sig- LITERATURE CITED
nificant decrease in total abundance (N ) of organisms
in our samples through time in the sand-added plots Airoldi L (2003) The effects of sedimentation on rocky coast
(Fig. 4). Additionally, since we saw an increase in assemblages. Oceanogr Mar Biol Annu Rev 41:161–236
Airoldi L, Cinelli F (1997) Effects of sedimentation on subtidal
abundance of scarcer, sand-tolerant taxa in sand plots
macroalgal assemblages: an experimental study from a
as the more dominant taxa decreased, an increase in Mediterranean rocky shore. J Exp Mar Biol Ecol 215:
Pielou’s evenness index (J’) in these plots makes sense. 269–288
The results of this experiment are significant not only Airoldi L, Virgilio M (1998) Responses of turf-forming algae to
because such dramatic changes were seen in the meio- spatial variations in the deposition of sediments. Mar Ecol
Prog Ser 165:271–282
and macrofaunal communities, but also because these Akioka H, Baba M, Masaki T, Johansen HW (1999) Rocky
shifts have the potential to be seen up through the shore turfs dominated by Corallina (Corallinales, Rhodo-
marine food web. As stated previously, meiofauna are phyta) in northern Japan. Phycol Res 47:199–206
important prey for many species of fish and inverte- AMEC Earth and Environment Inc. (2002) Regional beach
sand project post-construction monitoring report for inter-
brates. Small macrofauna like those found in turf are
tidal, shallow subtidal, and kelp forest resources. Pre-
also frequently reported as prey items for larger preda- pared for San Diego Association of Governments
tors (Gibbons & Griffiths 1986, Dugan et al. 2002). (SANDAG), San Diego, CA
Additionally, meiofauna may play an important role in Brown PJ, Taylor RB (1999) Effects of trampling by humans on
making detritus available to macroconsumers either animals inhabiting coralline algal turf in the rocky inter-
tidal. J Exp Mar Biol Ecol 235:45–53
through their enhancement of microbial activity or by Coull BC (1988) Ecology of the marine meiofauna. In: Higgins
ingestion of the meiofauna themselves (Coull 1988). RP, Thiel H (eds) Introduction to the study of meiofauna.
We believe that the investigation of the potential Smithsonian Institute Press, Washington, DC, p 18–38
effects of changes in the turf community on the larger Coull BC, Wells JBJ (1983) Refuges from fish predation —
experiments with phytal meiofauna from the New
coastal food web is an important area for future
Zealand rocky intertidal. Ecology 64:1599–1609
research. Daly MA, Mathieson AC (1977) Effects of sand movement on
Many questions still remain unanswered about turf intertidal seaweeds and selected invertebrates at Bound
communities in relation to periodic sand inundation. Rock, New Hampshire, USA. Mar Biol 43:45–55
This experiment showed a rapid response of turf fauna Dean RL, Connell JH (1987) Marine invertebrates in algal
succession. III. Mechanisms linking habitat complexity
followed by a relatively rapid recovery as well. How- with diversity. J Exp Mar Biol Ecol 109:249–273
ever, our experimental plots were small in scale Dierschke V (1994) Food and feeding ecology of purple sand-
compared with the broad areas of habitat that can be pipers Calidris maritima on rocky intertidal habitats (Hel-
covered with sand during a natural inundation event. goland, German Bight). Neth J Sea Res 31:309–317
Dommasnes A (1969) On the fauna of Corallina officinalis L.
Recovery of such a large area might take significantly
in western Norway. Sarsia 38:71–86
longer since animals may crawl in from the edges of a Dugan JE, Hubbard DM, McCrary MD, Pierson MO (2002)
small experimental plot more easily than they could if The response of macrofauna communities and shorebirds
an entire stretch of habitat was inundated. Addition- to macrophyte wrack subsidies on exposed sandy
ally, this study did not examine how the community is beaches of southern California. Estuar Coast Shelf Sci
58S:133–148
affected by anoxia caused by sand burial or by very Edgar GJ (1983) The ecology of south-east Tasmanian phytal
long-term inundation events. These are also likely to animal communities. I. Spatial organization on a local
be fruitful areas for future research. scale. J Exp Mar Biol Ecol 70:129–157
Gibbons MJ (1988) The impact of sediment accumulations,
relative habitat complexity, and elevation on rocky shore
Acknowledgements. This work was funded by California Sea meiofauna. J Exp Mar Biol Ecol 122:225–241
Grant, the Edna Bailey Sussman Fund and a San Diego Foun- Gibbons MJ, Griffiths CL (1986) A comparison of macrofaunal
dation Blasker Environment Grant. The work was done while and meiofaunal distribution and standing stock across a
J.K.J. performed an internship at Scripps made possible by rocky shore, with an estimate of their productivities. Mar
Dr. L. B. Liddle and P. Jackson. We are grateful for permission Biol 93:181–188
from reserve manager I. Kay to complete the work in the Gosselfin LA, Chia FS (1994) Feeding habits of newly hatched
Scripps Coastal Reserve within the University of California juveniles of an intertidal predatory gastropod, Nucella
Natural Reserve System. We thank P. Dayton and J. Leichter emarginata (Deshayes). J Exp Mar Biol Ecol 176:1–13
for valuable guidance, advice and support. P. E. Parnell, L. Graham MH, Dayton PK, Erlandson JM (2003) Ice ages and
82 Mar Ecol Prog Ser 345: 75–82, 2007
ecological transitions on temperate coasts. Trends Ecol Littler MM, Littler DS, Murray SN, Seapy RR (1991) Southern
Evol 18:33–40 California rocky intertidal ecosystems. In: Mathieson AC,
Herman PMJ, Heip C (1988) On the use of meiofauna in eco- Nienhuis PH (eds) Ecosystems of the world. Vol 24. Else-
logical monitoring: Who needs taxonomy? Mar Pollut Bull vier, New York, p 273–296
19:665–668 McQuaid CD, Dower KM (1990) Enhancement of habitat het-
Hicks GRF (1984) Spatio-temporal dynamics of a meiobenthic erogeneity and species richness on rocky shores inun-
copepod and the impact of predation disturbance. J Exp dated by sand. Oecologia 84:142–144
Mar Biol Ecol 81:47–72 Neumann AC, Gebelein CD, Scoffin TP (1970) The composi-
Hicks GRF (1985) Meiofauna associated with rocky shore tion, structure, and erodability of subtidal mats, Abaco,
algae. In: Moore PG, Seed R (eds) Ecology of rocky coasts. Bahamas. J Sediment Petrol 40:274–297
Hodder and Stoughton, London, p 36–56 Olabarria C (2002) Role of colonization in spatio-temporal
James RJ, Lincoln Smith MP, Fairweather PG (1995) Sieve patchiness of microgastropods in coralline turf habitat.
mesh-size and taxonomic resolution needed to describe J Exp Mar Biol Ecol 274:121–140
natural spatial variation of marine macrofauna. Mar Ecol Olabarria C, Chapman MG (2001) Comparisons of patterns of
Prog Ser 118:187–198 spatial variation of microgastropods between two con-
Kelaher BP (2002) Influence of physical characteristics of trasting intertidal habitats. Mar Ecol Prog Ser 220:201–211
coralline turf on associated macrofaunal assemblages. Prathep AR, Marrs H, Norton TA (2003) Spatial and temporal
Mar Ecol Prog Ser 232:141–148 variations in sediment accumulation in an algal turf and
Kelaher BP, Chapman MG, Underwood AJ (2001) Spatial pat- their impact on associated fauna. Mar Biol 142:381–390
terns of diverse macrofaunal assemblages in coralline turf Stewart JG (1982) Anchor species and epiphytes in intertidal
and their associations with environmental variables. J Mar algal turf. Pac Sci 36:45–59
Biol Assoc UK 81:917–930 Stewart JG (1983) Fluctuations in the quantity of sediments
Kelaher BP, Underwood AJ, Chapman MG (2003) Experimen- trapped among algal thalli on intertidal rock platforms in
tal transplantations of coralline algal turf to demonstrate southern California. J Exp Mar Biol Ecol 73:205–211
causes of differences in macrofauna at different tidal Taylor PR, Littler MM (1982) The roles of compensatory mor-
heights. J Exp Mar Biol Ecol 282:23–41 tality, physical disturbance, and substrate retention in the
Kendrick GA (1991) Recruitment of coralline crusts and fila- development and organization of a sand-influenced, rocky
mentous turf algae in the Galapagos Archipelago — effect intertidal community. Ecology 63:135–146
of simulated scour, erosion, and accretion. J Exp Mar Biol Warwick RM (1988a) The level of taxonomic discrimination
Ecol 147:47–63 required to detect pollution effects on marine benthic
Littler MM, Martz DR, Littler DS (1983) Effects of recurrent communities. Mar Pollut Bull 19:259–268
sand deposition on rocky intertidal organisms: importance Warwick RM (1988b) Analysis of community attributes of the
of substrate heterogeneity in a fluctuating environment. macrobenthos of Frierfjord/Langesundfjord at taxonomic
Mar Ecol Prog Ser 11:129–139 levels higher than species. Mar Ecol Prog Ser 46:167–170
Editorial responsibility: Howard Browman (Associate Editor- Submitted: June 8, 2006; Accepted: March 27, 2007
in-Chief), Storebø, Norway Proofs received from author(s): August 30, 2007