Full text of "What happens after mussels die? Biogenic legacy effects on community structure and ecosystem processes"

See other formats

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

ELSEVIER

Contents lists available at ScienceDirect

Journal of Experimental Marine Biology and Ecology

journal homepage: www.elsevier.com/locate/jembe

What happens after mussels die? Biogenic legacy effects on community
structure and ecosystem processes^ c ^ r

John A. Commito a,b ’*, Brittany R. Jones a<1 , Mitchell A. Jones 3 , Sondra E. Winders 3 , Serena Como c

a Environmental Studies Department, Gettysburg College, Gettysburg, PA 17325, USA

b Unita di Biologia Marina ed Ecologia, Dipartimento di Biologia, Universita di Pisa, Via Dema 1, 56126 Pisa, Italy

c Consiglio Nazionale delle Ricerche, Istituto per VAmbiente Marino Costiero (CNR-IAMC), Localitd Sa Mardini, Torregrande 09170, Oristano, Italy

ARTICLE INFO

ABSTRACT

Keywords:

Ecosystem engineer
Legacy effects
Mussel bed
Mytilus edulis
Shell hash
Soft-bottom

Mussels are well-known ecosystem engineers in soft-bottom systems. Mytilus edulis beds have myriad effects on
sediment, benthic organisms, and ecosystem processes such as hydrodynamic transport of sediment and animals.
When mussels die, they may leave behind massive amounts of whole (empty) and fragmented shells. The legacy
effects of this long-lasting biogenic material (i.e., shell hash) on benthic systems are poorly understood. We
measured percent cover values of 4 bottom cover types, i.e., live mussels, whole shells, fragmented shells, and
bare sediment, at the mussel bed in Carrying Place Cove, Harrington, Maine, USA, and examined their effects on
sediment characteristics, community structure of macrofauna and meiofauna, and ecosystem processes of se¬
diment flux and dispersal of postlarval macrofauna and meiofauna. We predicted that live mussels are the cover
type with the greatest effects compared to bare sediment, followed by fragmented shells and then whole shells.
We discovered mostly bare sediment, substantial cover of whole and fragmented shells, and almost no live
mussels in what had in past years been a robust bed. We found significant univariate and multivariate differences
in sediment and animals across cover types, especially for meiofauna. Fragmented shell material in particular
may be an important driver in this system. Our results are the first to quantify the 4 mussel bed cover types and
demonstrate their effects. Mussel beds in the Gulf of Maine have experienced severe declines in the past two
decades, attributed primarily to climate change and the invasive green crab, Carcinus maenas. Our results may be
useful in predicting the responses of soft-bottom systems as intact mussel beds die off, leaving large areas of bare
sediment and shell hash.

1. Introduction

Mussels are well-known ecosystem engineers in soft-bottom sys¬
tems. Blue mussel (Mytilus edulis ) beds have myriad effects on sediment,
infauna, epifauna, and ecosystem processes like wind-generated bed¬
load transport and animal dispersal (Bouma et al., 2009; Buschbaum
et al., 2009; Commito et al., 2005, 2008; Gutierrez et al., 2011). When
soft-bodied ecosystem engineers such as polychaetes die, their impact
may soon begin to wane (Gutierrez et al., 2011; Reise, 2002). But
mussels and other hard-bodied ecosystem engineers may leave behind
massive amounts of whole (empty) and fragmented shells (Fig. 1A;
Commito et al., 2008, 2014). The legacy effects of this long-lasting
biogenic material (i.e., shell hash) on community structure and eco¬
system processes are poorly understood.

Mussel beds consist of intermingled patches of live mussels, bare
sediment, and whole and fragmented shells (Fig. IB). Because beds in
Maine have a hierarchical, fractal spatial structure down to the milli¬
meter scale (Snover and Commito, 1998; Commito and Rusignuolo,
2000; Commito et al., 2016), even a small mussel bed patch may consist
of smaller patches of all 4 components. Moreover, live mussels attach to
each other and to whole and fragmented shells, bound together by
byssal threads. Thus, ecosystem engineering effects of mussel beds
cannot be attributed solely to the live mussel component. Yet to our
knowledge, no study of the effects of mussel shell hash has ever been
conducted at an intertidal, soft-bottom mussel bed anywhere in the
world.

Mollusk shell material can have important impacts on habitat pro¬
vision, water flow, recruitment, food supply, predation, and other

^ Author contribution: JAC, BRJ, MAJ, and SC conceived and designed the field work. JAC, BRJ, and MAJ carried out the field work. JAC, BRJ, MAJ, and SEW carried out the
laboratory work. JAC, BRJ, MAJ, SEW, and SC analyzed the data and wrote the manuscript.

* Corresponding author at: Environmental Studies Department, Gettysburg College, Gettysburg, PA 17325, USA.

E-mail address: jcommito@gettysburg.edu (J.A. Commito).

1 Present address: College of Fisheries and Ocean Sciences, University of Alaska Fairbanks, Fairbanks, AK 99775, USA.

https://doi.org/10.1016/jjembe.2018.05.004

Received 22 August 2017; Received in revised form 11 May 2018; Accepted 12 May 2018

Available online 25 May 2018

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Live mussels Whole shells

Fragmented shells Bare sediment

Fig. 1. Mussel bed bottom cover. (A) Shell hash at our mussel bed study site:
Carrying Place Cove, Harrington, Maine, USA. (B) Types of bottom cover found
in Maine mussel beds.

Patch attribute Strength of effects on benthic systems

Weak Strong

Size

Age

Vertical dimension

Surface topography

Mobility of components

Porosity (gap widths)

Production of materials
(e.g., feces, pseudofeces)

Percent cover

Areal extent

Fig. 2. Mussel bed patch characteristics. Attributes presented across the range
from weak to strong effects on soft-bottom sediment, community structure, and
ecosystem processes of sediment flux and animal transport in bedload.

factors (Gutierrez et al., 2003, 2011), often causing an increase in in¬
fauna, epifauna, and species diversity in benthic systems (Guay and
Himmelman, 2004; Gutierrez et al., 2003; Hily, 1991; Hubbard, 2016;
Kraeuter et al., 2003; Ribeiro et al., 2005; Rodney and Paynter, 2006;
Summerhayes et al., 2009; Tomatsuri and Ron, 2017; Wilding and
Nickell, 2013). However, some studies have found mixed, weak, or no
significant effects (Bomkamp et al., 2004; Gutierrez and Iribarne, 1999;
Hewitt et al., 2005; Mann et al., 2016; Nicastro et al., 2009; Turner
et al., 1997). The wide-ranging responses to living bivalves and non¬
living biogenic structure make it difficult to predict the expected
magnitude and even the direction of differences among live mussels,
whole shells, and fragmented shells relative to bare sediment. The im¬
pacts most likely stem from variations in bed patch attributes such as
those presented in Fig. 2.

It seems plausible that effects on ambient sediment and organisms,
as well as on flow-related sediment flux and animal transport, are

JM*

Large

Thick, deep

w Smooth

Rough

Easily transported Stationary

Open (wide)

Closed (narrow)

J—

Shell hash

Live mussels

— ■ —,-^ ^

Low

High

w v

yf- It) .#!■*

Small

Large

greatest for live mussels, followed by fragmented shells and then whole
shells. We make this prediction because live mussels produce copious
amounts of feces and pseudofeces that increase the silt-clay fraction and
create low-oxygen, high sulfide conditions that are detrimental to most
species but favorable to a few others, e.g., the oligochaete Tubificoides
benedeni and opportunistic polychaetes like Capitella capitata (Albrecht
and Reise, 1994, Albrecht, 1998; Commito et al., 2005, 2008; Kent
et al., 2017; Ragnarsson and Raffaelli, 1999). In addition, live mussels
project up into the water column above the bottom, and this roughness
profile has strong effects on flow dynamics that increase the capture of
sediment, postlarval macrofauna, and meiofauna moving across the
bottom (Commito and Rusignuolo, 2000; Commito et al., 2005).

Fragmented shells may be next in importance because they can
often be observed in dense, tightly packed patches that may act as a
barrier between the sediment and the water column. Packing theory
demonstrates clearly that packing is tighter, with lower porosity, when
the objects are of many sizes, allowing small ones to fill in the gaps
between large ones (Chen et al., 2003). Thus, at our sites in Maine we
observe that fragmented shell pieces, which exist in angular shapes of
all sizes, are generally more tightly packed than whole shells. Frag¬
mented shells also alter sediment structure by contributing directly to
the coarse sediment fraction. Species abundances in Maine mussel beds
can be positively or negatively correlated with coarse, terrestrially-de¬
rived gravel (Commito et al., 2008), so they may respond similarly to
coarse, fragmented shell material as well. In particular, fragmented
shells might be expected to depress animal abundance by blocking the
movement of oxygen into the sediment below. Their influence on flow
dynamics, hence the movement of sediment and animals, is probably
less than that of live mussels because fragments do not project as high
up into the water column.

Whole shells are often loosely packed. They are generally larger
than fragments, and their large-radius curves leave sizable gaps be¬
tween neighboring shells even when touching because they have no
straight, parallel sides. This porosity due to gaps within a patch may
create less of a sediment-water column barrier compared to fragmented
shell cover. We also observe that whole shells tend to lie flat on the
bottom, often concave-side down, presenting a relatively smooth bed
surface that may not induce as much turbulent flow as does a bed with
the rough topography of live mussels or fragmented shell pieces. Thus
we expect whole shells to have less of an impact on sediment, animals,
and hydrodynamics than do live mussels and fragmented shells.

In this study we investigated biogenic legacy effects by comparing
sediment characteristics, community structure of macrofauna and
meiofauna, and the ecosystem processes of sediment flux and faunal
transport in isolated patches of live mussels, bare sediment, whole
shells, and fragmented shells in a Maine soft-bottom mussel bed. Mussel
beds in the Gulf of Maine have recently experienced severe declines,
with reduced larval settlement and decimated abundances of juveniles
and adults in the past two decades (Petraitis and Dudgeon, 2015; Sorte
et al., 2011, 2016). The bottom cover proportions of whole shells,
fragmented shells, and bare sediment may be increasing relative to that
of live mussels. If so, our results could be useful in understanding how
soft-bottom systems respond to the apparent mussel bed decline.

2. Methods

2.1. Study site

The research was conducted at the intertidal, soft-bottom Mytilus
edulis bed in Carrying Place Cove, Harrington, Maine, USA (44.5451°N,
— 67.7844°W), a relatively protected embayment with a bottom of
muddy sand (Fig. 3). As is typical in this region, the bed extends across
the mouth of the cove near the low tide line. The ecology of eastern
Maine soft-bottom mussel beds like this one has been well studied,
including their spatial abundance patterns (Commito et al., 2006, 2014;
Crawford et al., 2006), sediment and macrofauna (Commito et al..

J.A. Commito et al.

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Fig. 3. (A) Location of study site in eastern Maine, USA. (B) Aerial view of Carrying Place Cove, Harrington, Maine, with mussel bed and transects indicated. Cores
were taken and traps deployed immediately south of Transect C. Map data copyright 2015 Google.

2005, 2008), and impacts on sediment flux and postlarval dispersal
dynamics (Commito et al., 2005). The intertidal, soft-bottom mussel
beds in eastern Maine that we have studied for decades typically had =
50% cover of live mussels in complex fractal power-law spatial patterns
(Commito et al., 2006, 2014; Crawford et al., 2006). The remaining bed
surface consisted primarily of bare sediment with small percent cover
values of whole (empty) mussel valves and mussel shell fragments.
Today, some of the largest, densest beds where we have conducted
research since the mid-1970s contain virtually no live mussels (John A.
Commito, personal observation; Brian F. Beal, personal communica¬
tion). Our original intent was to conduct this investigation at the nearby
Guard Point mussel bed because we had results from past studies there
for possible comparison (Commito et al., 2005, 2006, 2014). However,
we could not locate any live mussels except for some scattered in¬
dividual on this once thriving bed, requiring us to move = 1 km to
Carrying Place Cove, which has similar sediment and other character¬
istics. No long-term time series of aerial photographs suitable for mussel
cover analysis exists for Carrying Place Cove. However, our field ob¬
servations at the site over the last 30 years suggest that mussel abun¬
dance at Carrying Place Cove has dropped precipitously, resulting in a
scattered array of small, isolated patches of live mussels in a complex
pattern with bare sediment, whole mussel shells, and mussel shell
fragments over the bed surface (Fig. 1A).

In this project we used photo transects to quantify the percent cover
of each of the 4 cover types: live mussels, whole mussel shells, frag¬
mented mussel shells, and bare sediment. In patches of each cover type,
we used: (1) core samples to compare univariate and multivariate
measures of sediment characteristics, macrofauna, and meiofauna; and
(2) bottom traps to compare the ecosystem processes of sediment flux
and dispersal of postlarval macrofauna and meiofauna.

2.2. Cover types

Digital photographs were taken every meter along 3 transects
(Transect C on 4 August 2011, Transects A and B on 31 July 2012) that
ran from the upper to the lower margin of the mussel bed (Fig. 3). No
transects were established in the southern portion of the bed because of
possible effects on the bottom in that area from occasional light activity
by small boats. Images were cropped to an area of 0.25 m 2 . They were
uploaded into ArcMap and hand digitized into the 4 cover types (live
mussels, whole shells, fragmented shells, and bare sediment). Percent
cover of each cover type was determined using pixel counts. Transect

results were plotted with a 3-cell running average for smoothing.

2.3. Ambient sediment, macrofauna, and meiofauna

We followed Hurlburt's (1984) recommendations for treatment in-
terspersion when spatial heterogeneity is expected, particularly in field
projects of this size. Our sample design utilized his systematic model to
achieve treatment interspersion, reducing the likelihood of treatment
segregation, type 1 error, and spurious treatment effects. Our goal had
been to sample large patches similar in size to those from a previous
study at Guard Point, where we were able to select live mussel patches
> 6.0 m in the smallest dimension (Commito et al., 2005). Although we
commonly observed such large patches in past years at Carrying Place
Cove, none approaching that size could be found at the time of this
study. Live mussel patches were smaller and far less common than the
other 3 cover types, so we searched for live mussels and chose the 10
mussel patches >0.5m x 0.5 m in size nearest to Transect C within
a = 10 m wide band from the upper to the lower mussel bed margin (~
100 m). We then selected the patches >0.5m x 0.5 m of each of the
other 3 cover types closest to each live mussel patch. Thus, N = 10 for
each of the 4 cover types.

At low tide on 15 May 2011, 1.3 cm diameter cores (cross-sectional
area = 1.33 cm 2 ) were taken to a depth of 5 cm at the 40 sampling
locations. Cores of this size were used successfully to sample sediment
characteristics, macrofauna, and meiofauna at similar sites nearby,
collecting approximately the same numbers of macrofaunal species as
larger cores and allowing us to utilize a larger sample size with the
same sampling and laboratory processing effort (Commito and Tita,
2002; Commito et al., 2005, 2008). Core contents were stained with
rose bengal and fixed in situ with buffered formalin. In the laboratory,
samples were wet sieved on 0.5 mm (to obtain macrofauna and the
coarse sediment fraction) and 0.063 mm (meiofauna and fine sediment
fraction) mesh, and the material that passed through the sieves (silt-
clay sediment fraction) was retained. Macrofauna (to species level when
possible, using Pollock, 1998) and meiofauna (to family or higher le¬
vels, e.g., Copepoda, Nematoda, Foraminifera) were identified and
counted in gridded Petri dishes. The 3 sediment size-classes were dried
at 85 °C for 24 h and weighed. Organic matter was calculated using loss-
on-ignition by burning at 500 °C for 4h (Dean, 1974).

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

0 50

Distance from upper to lower mussel bed margin (m)

Fig. 4. Percent cover values for cover types along transects. Note different
transect lengths on horizontal axes. See Table 1 for summary values.

2.4. Sediment flux and animal dispersal as ecosystem processes

On 15 May 2011, 15.0 cm tall bottom traps of the same diameter as
the coring device (aspect ratio = 11.5:1) were filled with seawater and
inserted flush with the sediment surface into the holes created during
the core sampling described above. Traps of this design were used
successfully to measure sediment flux and dispersal of macrofauna and
meiofauna at similar sites nearby (Commito and Tita, 2002; Commito
et al., 2005). To assure that the traps did not lose their capture effi¬
ciency by filling with sediment and effectively lowering their aspect
ratio, they were removed and replaced with new traps inserted into the
same holes at low tide on each of the following 3 days. Trap samples
were processed in the same way as the cores. Data from the daily trap
contents were combined to create a cumulative sample for the 3 day
deployment period. Following Commito et al. (1995a, 2005) and
Commito and Tita (2002), the sediment flux rate = g sediment trap -1
deployment -1 and the animal dispersal rate = number of individuals
trap -1 deployment -1 were calculated. Despite being marked with
flags, some traps were impossible to locate due to burial by shell hash
moved by water currents. Because our multivariate statistical approach
(see below) required a balanced design, trap samples from each cover
type were eliminated at random to achieve an equal number of re¬
plicates (IV = 8) for each cover type.

2.5. Data analysis

Analysis of the cores and traps included the sediment mass in each
size-class (silt-clay, fine sand, coarse material) and percent organic
matter, as well as the total abundance, species richness, Shannon-
Wiener Index (H') using log base e, and abundance of each dominant
taxon in the macrofauna and meiofauna, respectively. Because the

oligochaete Tubiflcoides benedeni often dominates the macrofauna in
mussel beds and exhibits abundance patterns different from other
macrofaunal species (Commito et al., 2005, 2008), the non-oligochaetes
were also analyzed together as one group.

For each core and trap variable, differences among cover types (4
levels: fixed) were analyzed with a 1-factor ANOVA. Homogeneity of
variances was checked using Cochran's C-test, and, when necessary,
data were transformed to reduce heterogeneity to acceptable levels, as
indicated in the ANOVA tables. In these cases, back-transformed means
and 95% confidence intervals were calculated. When significant dif¬
ferences among cover types were found (alpha = 0.05), a posteriori
comparisons were made using SNK tests. Cases occurred where SNK
could not discriminate among alternative hypotheses despite the sig¬
nificant ANOVA differences. Because a posteriori multiple comparisons
have less power than the original ANOVA F-test, such cases indicated
that the 2 cover types with the largest and smallest means were sig¬
nificantly different, and no further resolution among cover types could
be determined (Underwood, 1997). All analyses were done with the
GMAV 5 computer program (A. J. Underwood and M. G. Chapman,
unpublished).

For macrofauna and meiofauna in the ambient community and
dispersing assemblages, differences among cover types (4 levels: fixed)
were analyzed with a distance-based permutational multivariable
analysis of variance (PERMANOVA) based on Bray-Curtis dissimilarity
measures (Anderson, 2001; McArdle and Anderson, 2001). Raw, square
root transformed, and presence/absence data were used in the analysis.
Results were the same for all 3, so only results from raw data are pre¬
sented. PERMDISP was utilized to determine if significant PERMAN¬
OVA differences were due to cover type data dispersion or location in
multivariate space (Anderson, 2006; Anderson et al., 2006). PERMA¬
NOVA and PERMDISP were executed using the functions “adonis” and
“betadisper” in the “vegan” package for R v3.4.2 (R Core Development
Team, 2015). The non-metric multidimensional scaling (nMDS) ordi¬
nation model based on the Bray-Curtis dissimilarity matrix was con¬
ducted with the PRIMER v5.2 package (Clarke and Warwick, 2001) to
visualize cover type separation when PERMANOVA differences were
significant.

3. Results

3.1. Cover types

The 3 transects revealed consistent trends from the upper margin to
the lower margin of the mussel bed (Fig. 4, Table 1). Live mussel cover
was only 1-4% of the bed surface. A few individuals were found at the
upper margin of the bed, but most were in the lower part of the bed
near the low tide line. Whole shell cover (7-21%) and fragmented shell
cover (4-15%) were much more extensive, particularly in the upper
portion of the bed. Bare sediment (62-88%) was by far the largest cover
type, especially in the middle and low portions of the bed.

Table 1

Cover values for 3 transects at Carrying Place Cove. Values may not add up to
100% due to rounding. Transects extended from the upper margin to the lower
margin of the mussel bed. See Fig. 3 for aerial view of the cove and Fig. 4 for
transect details.

Transect

Length (m)

Cover (%)

Live

Whole

Fragmented

Bare

mussels

shells

sediment

1.5

21.3

15.2

61.8

107

0.7

6.6

4.0

88.7

112

3.7

10.1

10.4

75.3

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Cores: Sediment

Fig. 5. Sediment results for cores. (A) Mass values of silt-clay (light gray), fine
sand (medium gray), and coarse material (dark gray). (B) Percent total organic
matter (TOM). Values are mean ± 1 SE, except back-transformed mean and
95% confidence interval for coarse material, which required transformation to
meet ANOVA assumptions. Live = live mussels, Whole = whole shells.
Frag = fragmented shells. Bare = bare sediment. See Results and Table 2 for
statistical analysis.

3.2. Ambient sediment, macrofauna, and meiofauna

The silt-clay and coarse sediment size-classes were significantly
different among cover types (Fig. 5, Table 2). Silt-clay mass was

Table 2

ANOVA results for ambient community cores.

Parameter

Transform

$ 3,36

SNK

Sediment

Silt-clay (g)

4.78

0.005

Bare > Whole = Live = Frag

Fine sand (g)

2.1

0.117

Coarse material (g)

Ln(X + 1)

5.51

0.002

Frag > Whole = Live = Bare

Total organic matter

1.47

0.24

(%)

Macrofauna

Number of species

2.09

0.119

0.92

0.439

Total abundance

Ln(X + 1)

1.22

0.318

Oligochaetes

1.89

0.149

N on-oligochaetes

Ln(X + 1)

1.73

0.179

Capitella capitata

Ln(X + 1)

3.34

0.03

NAH (Live > Frag)

Meiofauna

Number of taxa

0.51

0.678

1.53

0.223

Total abundance

4.21

0.012

NAH (Bare > Whole)

Nematodes

4.4

0.01

NAH (Bare > Whole)

Copepods

2.66

0.065

Foraminiferans

2.35

0.089

Significant differences (P < 0.05) are in bold. For parameters with significant
differences among cover types, the SNK results are presented. NAH = no al¬
ternative hypothesis; the 2 cover types with the largest and smallest means were
significantly different, and no further resolution among cover types was pos¬
sible. Live = live mussels, Whole = whole shells, Frag = fragmented shells,
Bare = bare sediment.

significantly greater in cores from the bare sediment cover type than
from the other cover types, which were not significantly different from
each other. Coarse material mass was significantly greater in cores from
the fragmented shell cover type than from other cover types, which
were not significantly different from each other (Table 2). The per¬
centage of organic material did not vary significantly among cover
types (Fig. 5, Table 2).

Sixteen macrofaunal species were found (Appendix A), of which the
most abundant was the oligochaete Tubificoides benedeni (74% of the
total), followed by the polychaete Capitella capitata and much smaller
numbers of the other species.No significant differences were found
among cover types except for C. capitata abundance (Fig. 6,
Table 2).The a posteriori comparison test could not discriminate among
alternative hypotheses, indicating that C. capitata was significantly
more abundant in live mussel cover than in fragmented shell cover, and
no further resolution among cover types was possible (Table 2).At the
multivariate level, PERMANOVA revealed no significant differences in
macrofaunal community structure among cover types with or without
oligochaetes (Table 3).

Six meiofaunal taxa were found (Appendix A), of which the most
abundant were nematodes (81% of the total), followed by copepods,
foraminiferans, and much smaller numbers of the other taxa.Total
abundance and nematode abundance were significantly different
among cover types (Fig. 7, Table 2).In both cases the a posteriori
comparison test could not discriminate among alternative hypotheses,
indicating that total meiofauna and nematodes were significantly more
abundant in bare sediment cover than in whole shell cover, and no
further resolution among cover types was possible (Table 2).At the
multivariate level, PERMANOVA revealed significant differences in
meiofaunal community structure among cover types (Table 3), and
nMDS plots showed separation among cover types (Fig. 8), consistent
with the PERMDISP result of no significant data dispersion
0*3,36 = 0.47, P = 0.74).Dissimilarity values were large for fragmented
and whole shells compared to bare sediment and, to a lesser degree, live
mussels, highlighting the impact of shell hash on meiofaunal commu¬
nity structure.

3.3. Sediment flux and animal dispersal as ecosystem processes

The silt-clay and coarse sediment size-classes collected by traps
were significantly different among cover types (Fig. 9, Table 4). In both
cases the a posteriori comparison test could not discriminate among
alternative hypotheses, indicating that trap values were significantly
higher for silt-clay in live mussel cover than fragmented shell cover and
for coarse sediment in fragmented shell cover than live mussel cover,
and no further resolution among cover types was possible (Table 4).
Total sediment mass collected by traps was not significantly different
among cover types (Table 4), nor was the percentage of organic ma¬
terial (Fig. 9, Table 4).

Twenty-four macrofaunal species were found in traps, and the
species ranks were different from those in the ambient community cores
(Appendix A). The most abundant species was still Tubificoides benedeni,
but dropping to 43% of the total, followed by smaller numbers of other
species, with no species a clear co-dominant, including Capitella capi¬
tata. No significant differences were found for any macrofaunal species,
total abundance, species richness, or H' (Fig. 10, Table 4). At the
multivariate level, PERMANOVA revealed no significant differences in
macrofaunal dispersal assemblage structure among cover types
(Table 5).

Eight meiofaunal taxa were found in traps, and the taxon ranks were
different from those in cores, with a dramatic shift to copepods as the
dominant taxon (54% of the total), followed by nematodes, for¬
aminiferans, and much smaller numbers of the other taxa (Appendix A).
Total abundance was significantly lower in traps from fragmented shell
cover than from other the cover types, which were not significantly
different from each other (Fig. 11, Table 4). H' and the abundances of

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Cores: Macrofauna

Number of species H'

Oligochaetes

Non-oligochaetes

Total abundance

Live Whole Frag Bare Live Whole Frag Bare

Live Whole Frag Bare

Fig. 6. Macrofauna results for cores. Values are mean ± 1 SE for number of species, Shannon-Wiener diversity index (HO, and abundance of the dominant species
(the oligochaete Tubificoides benedeni ), and back-transformed mean and 95% confidence interval for total abundance, the non-oligochaetes as a group, and the second
most abundant species (the polychaete Capitella capitatd), which required transformation to meet ANOVA assumptions. Live = live mussels, Whole = whole shells,
Frag = fragmented shells, Bare = bare sediment. See Results and Table 2 for statistical analysis.

Table 3

PERMANOVA results for ambient community cores.

Parameter MS F 3.36 P

Macrofauna

All species

3118.75

0.24

0.275

Without oligochaetes

5283.76

1.44

0.275

Meiofauna

3364.62

2.48

0.031

Within cover type:

Live = 43.53
Whole = 49.23
Frag = 49.61
Bare = 39.91
Among cover types:
Live vs Whole = 48.88
Live vs Frag = 48.24
Live vs Bare = 42.67
Whole vs Frag = 47.69
Whole vs Bare — 54.22
Frag vs Bare = 55.53

Significant differences (P < 0.05) are in bold. For parameters with significant
differences among cover types, the mean Bray-Curtis dissimilarity (%) values
within and among cover types are presented. Live = live mussels,
Whole = whole shells. Frag = fragmented shells, Bare = bare sediment.

nematodes and foraminiferans also differed significantly among cover
types (Fig. 11, Table 4). In all 3 cases the a posteriori comparison test
could not discriminate among alternative hypotheses, indicating that
trap values were significantly higher for H' in bare sediment than live
mussels, nematodes in bare sediment than fragmented shells, and for¬
aminiferans in bare sediment than live mussels, and no further resolu¬
tion among cover types was possible (Table 4). At the multivariate
level, PERMANOVA revealed significant differences in meiofaunal
community structure among cover types (Table 5), and the nMDS plot
showed separation among cover types (Fig. 8), consistent with the

PERMDISP result of no significant data dispersion (F 3:28 = 0.97,
P = 0.45). Dissimilarity values were large for fragmented shell cover
compared to all the other cover types (Table 5), demonstrating its im¬
portant impact on the meiofaunal dispersal assemblage.

4. Discussion

4.1. Mussel bed cover

The results presented here are the first to quantify live mussel,
whole shell, fragmented shell, and bare sediment percent cover values
and demonstrate effects of the 4 cover types. The mussel bed at
Carrying Place Cove was primarily bare sediment, with substantial
cover of whole shells and fragmented shells and almost no live mussels.
The combined cover value of whole and fragmented shells was an order
of magnitude larger than live mussel cover, revealing the importance of
non-living biogenic material when considering the physical structure of
this and probably other mussel beds.

4.2. Does biogenic structure matter?

The effects of mussel beds on sediment, ambient community, and
ecosystem processes have been well studied in recent years (Bouma
et al., 2009; Buschbaum et al., 2009; Commito et al., 2005, 2008, 2014;
Gutierrez et al., 2011; Smith and Shackley, 2004; Thiel and Ullrich,
2002). However, investigations have generally treated mussel beds as a
single biotic unit, with little attempt at teasing apart the effects of their
different structural components - live mussels, whole shells, and frag¬
mented shells - compared to bare sediment.

We predicted that effects on ambient sediment and organisms, as
well as on flow-related sediment flux and animal transport, are greatest
for live mussels, followed by fragmented shells and then whole shells.
We found significant differences across cover types at the univariate

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Cores: Meiofauna

Number of species

Copepods

Total abundance

Foraminiferans

Live Whole Frag Bare Live Whole Frag Bare Live Whole Frag Bare

Fig. 7. Meiofauna results for cores. Values are mean ± 1 SE for number of taxa, Shannon-Wiener diversity index (HO, total abundance, and abundances of the
dominant taxa (nematodes, foraminiferans), and back-transformed mean and 95% confidence interval for the dominant taxon of copepods, which required trans¬
formation to meet ANOVA assumptions. Live = live mussels, Whole = whole shells, Frag = fragmented shells. Bare = bare sediment. See Results and Table 2 for
statistical analysis.

level in sediment structure and animal abundances for both the ambient
community and the transported sediment and animals. The fragmented
shell cover type was often most different from the others. We also found
significant differences at the multivariate level for meiofauna in both
ambient community and dispersing assemblage structure. The frag¬
mented shell cover type contributed most to the dissimilarity. Thus it
seems that cover type is an important driver in mussel beds and that
fragmented shell material may be particularly important.

We studied patches >0.5 m x 0.5 m in size for all 4 cover types
because we wanted all the patches to be similar in size, and live mussels
were never observed in patches larger than this size. Most of our pat¬
ches were right at the lower size limit. Had we been able to use larger
patches with lower perimeter: area ratios, it is possible that the influ¬
ence of cover type may have been stronger. Thus, our results are likely
to be conservative estimates of cover type effects.

4.2.1. Sediment

Organic material in soft-bottom systems is typically linked posi¬
tively to the silt-clay fraction and negatively to the coarse fraction (Gray
and Eliot, 2009). However, despite significant differences in the sedi¬
ment size-classes among cover types, with bare sediment cover having
the most silt-clay and fragmented shell cover having the most coarse
material, no significant differences occurred in the percentage of total
organic material among cover types in cores or traps. This result was
surprising because mussels produce feces and pseudofeces known to
enhance local sedimentation and alter sediment structure (Albrecht and
Reise, 1994, Albrecht, 1998; Commito et al., 2005; Kent et al., 2017;
Ragnarsson and Raffaelli, 1999). However, it is possible that organi¬
cally rich feces and pseudofeces were exported from patches of live
mussels to nearby patches of the other cover types, reducing differences

in organic content (Donadi et al., 2013).

Surface topography is generally roughest for live mussels, less rough
for fragmented and then whole shells, and smoothest for bare sediment
(Commito and Rusignuolo, 2000). These topographic differences shape
hydrodynamic profiles because turbulence increases as a function of the
roughness, number, and spacing of patches, as well as the amount of
patch edge, with more turbulence at patch edges than at patch centers
(Folkard and Gascoigne, 2009; Widdows et al., 2009). The result is
generally more sediment erosion and resuspension due to higher tur¬
bulent kinetic energy and bed shear stress in patches of physical
structure compared to bare sediment (wa Kangeri et al., 2016; Widdows
et al., 2009), with increased sediment deposition in M. edulis beds
(Albrecht and Reise, 1994; Albrecht, 1998; Commito et al., 2008;
Ragnarsson and Raffaelli, 1999). Yet in this study we found no sig¬
nificant differences across cover types in total sediment collected by
traps. One possible explanation is that the cover type patches were too
small to produce an effect. Another possibility is that the differences in
physical structure among the cover types, or even the presence of
physical structure per se, were not enough to enhance sediment capture
dramatically. In fact, the opposite may be true. A surface cover of gravel
acts as “bed armor” that reduces bedload transport (Singer and
Anderson, 1984). It is possible that live mussels, whole shells, and
fragmented shells can also armor the surface, thus countering their
effects as sedimentation-enhancing roughness elements.

4.2.2. Macrofauna

Our results revealed little difference in macrofauna among cover
types. The only significant ANOVA difference was for the opportunistic
polychaete Capitella capitata, the numerically dominant species. The
ambiguous SNK result allowed the conclusion that its abundance was

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Meiofauna

■ = Live mussels O = Fragmented shells

O = Whole shells A = Bare sediment

Fig. 8. Non-metric multi-dimensional scaling (nMDS) results for cases where
PERMANOVA revealed significant differences among cover types. Meiofauana:
(A) Cores, (B) Traps. No significant differences were observed for macrofauna.

higher in live mussel cover than in fragmented shell cover. At the
multivariate level, no significant differences were found among cover
types. Numerous studies that have found higher total macrofaunal
abundance in mussel beds compared to bare intertidal flats, probably
due to the production of organically enriched, low oxygen, high sulfide
sediment favorable to oligochaetes and opportunistic species like C.
capitata, especially those without free-swimming larvae (Buschbaum
et al., 2009; Commito et al., 2005, 2008; Gutierrez et al., 2011;
Ysebaert et al., 2009). The lack of major differences in macrofauna
among cover types in our study most likely reflects the similarities in
sediment structure and organic material observed across cover types.

With respect to the dispersing assemblage, macrofaunal abundances
in traps are often positively correlated with sediment mass in traps
because animals may be transported like passive sediment particles
(Commito et al., 1995a, 1995b; 2008; Turner et al., 1997; Valanko
et al., 2010a, 2010b). Thus, the lack of clear patterns among cover types
for macrofaunal abundance in traps in our study is consistent with the
lack of significant differences in total sediment captured by traps.

Our trap results demonstrated that the mussel bed system is highly
dynamic, suggesting a potential for change over time. Traps on average
collected approximately the same number of macrofaunal individuals in
the 3-day deployment as were found in ambient assemblage cores, in¬
dicating a turnover time of = 3 days, where turnover time = (no. in¬
dividuals per core) 4- (no. individuals per trap per deployment period).
This value is within the range of 1.4-4.3 days at a nearby mussel bed in

Traps: Sediment

Live Whole Frag Bare

Fig. 9. Sediment results for traps. (A) Mass values of silt-clay (light gray), fine
sand (medium gray), and coarse material (dark gray). (B) Percent total organic
matter (TOM). Values are mean ± 1 SE. Live = live mussels, Whole = whole
shells, Frag = fragmented shells, Bare = bare sediment. See Results and Table 4
for statistical analysis.

Table 4

ANOVA results for dispersal assemblage traps.

Parameter

Transform

F 3,28

SNK

Sediment

Total (g)

1.11

0.361

Silt-clay (g)

5.12

0.006

NAH (Live > Frag)

Fine sand (g)

0.51

0.676

Coarse material (g)

4.56

0.01

NAH (Frag > Live)

Total organic

0.25

0.861

matter (%)

Macrofauna

Number of species

0.5

0.688

0.65

0.591

Total abundance

Sqrt(X + 1)

0.64

0.595

Oligochaetes

1.15

0.349

Non-oligochaetes

Ln(X + 1)

0.16

0.92

Meiofauna

Number of taxa

0.64

0.594

Ln(X + 1)

3.44

0.03

NAH (Bare > Live)

Total abundance

Ln(X + 1)

3.3

0.035

Bare = Whole = Live < F rag

Nematodes

Ln(X + 1)

3.47

0.029

NAH (Bare > Frag)

Copepods

1.86

0.16

Foraminiferans

3.71

0.023

NAH (Bare > Live)

Significant differences (P < 0.05) are in bold. For parameters with significant
differences among cover types, the SNK results are presented. NAH = no al¬
ternative hypothesis; the 2 cover types with the largest and smallest means were
significantly different, and no further resolution among cover types was pos¬
sible. Live = live mussels. Whole = whole shells. Frag = fragmented shells,
Bare = bare sediment.

Maine (Commito et al., 2005) and 2.5-33 days calculated from data in
Valanko et al. (2010a, 2010b) for a shallow, non-tidal, subtidal site in
the Baltic Sea. It is longer than the range of 0.5-1.1 days for a small
bivalve (Gemma gemma ) at a sandflat in Virginia (Commito et al.,
1995a), most likely because the sandflat had high wind velocities, and

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Traps: Macrofauna

Number of species

2.5

0 J——-l_l-l_l

Oligochaetes

H' Total abundance

Non-oligochaetes

Live Whole Frag Bare

Fig. 10. Macrofauna results for traps. Values are mean ± 1 SE for number of species, Shannon-Wiener diversity index (HO, and abundance of the dominant species
(the oligochaete Tubificoides benedeni), and back-transformed mean and 95% confidence interval for total number of individuals and the non-oligochaetes as a group,
which required transformation to meet ANOVA assumptions. Live = live mussels, Whole = whole shells, Frag = fragmented shells, Bare = bare sediment. See
Results and Table 4 for statistical analysis.

Table 5

PERMANOVA results for dispersal assemblage traps.

Parameter MS F 3i2 s P

Macrofauna

All species

2854.10

0.97

0.522

Without oligochaetes

3466.50

0.86

0.68

Meiofauna

2273.29

2.19

0.043

Within cover type:

Live = 35.99
Whole = 49.50
Frag = 42.71
Bare — 32.31
Among cover types:

Live vs Whole = 41.83
Live vs Frag = 47.53
Live vs Bare = 34.07
Whole vs Frag = 48.02
Whole vs Bare = 41.05
Frag vs Bare = 48.12

bedload transport and benthic animal transport are positively related to
wind-generated hydrodynamic forces (Commito et al., 1995a, 1995b;
Turner et al., 1997; Valanko et al., 2010a, 2010b).

Traps collected macrofauna in different proportions than were
found in the ambient community cores. Tubificoides benedeni, the
dominant macrofaunal species in the ambient community, was the most
abundant macrofaunal species in traps but dropped substantially as a
proportion of the total. Yet for macrofauna, ambient community
structure seems to persist in soft-bottom systems, despite the daily

transport across the bottom of organisms with a different species mix
than in cores (Commito et al., 2005; Turner et al., 1997; Valanko et al.,
2010a, 2010b). In the face of potential forcing from dispersing organ¬
isms, the strength of this resistance to change is a ripe topic for re¬
search.

4.2.3. Meiofauna

Cover type had stronger effects on meiofauna than on macrofauna.
Significant ANOVA differences were found for total meiofauna and the
numerically dominant nematodes. The ambiguous SNK results allowed
the conclusion that abundances for both parameters were higher in bare
sediment cover than in whole shell cover. At the multivariate level,
significant differences in meiofaunal community structure occurred
among cover types. Dissimilarity values indicated that fragmented and
whole shells had a large impact on the meiofauna community. Some
live bivalves (e.g., Cerastoderma glaucum, Macoma balthica, Mya are-
naria\ Urban-Malinga et al., 2016) are known to enhance nematode,
copepod, and total abundance values. We did not find this effect for live
M. edulis.

With respect to the dispersing assemblage, cover type again had
stronger effects on meiofauna than on macrofauna. ANOVA revealed
significant differences in several parameters, and total abundance was
significantly lower in fragmented shell cover than in the other cover
types. One possible explanation is that fragmented shell cover acted as
bed armor to reduce the erosion and capture of local, within-patch
meiofauna, especially in comparison with bare sediment.

Traps collected 2.3-fold more meiofaunal individuals than were in
cores, indicating a turnover time of = 1.3 days, which is shorter than
the 4-day turnover time for meiofauna at a nearby Maine mudflat
(Commito and Tita, 2002), most likely because wind velocities were
higher during this study. As with macrofauna, traps collected meio¬
faunal taxa in different rank order than were found in the ambient

J.A. Commito et al

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Traps: Meiofauna

Number of species h'

Total abundance

Fig. 11. Meiofauna results for traps. Values are mean ± 1 SE for number of taxa and abundances of the dominant taxa (copepods, foraminiferans), and back-
transformed mean and 95% confidence interval for Shannon-Wiener diversity index (FT'), total number of individuals, and the dominant taxon of nematodes, which
required transformation to meet ANOVA assumptions. Live = live mussels, Whole = whole shells, Frag = fragmented shells, Bare = bare sediment. See Results and
Table 4 for statistical analysis.

community. The rank orders of the 2 dominant taxa in cores, nematodes
and copepods, were reversed in traps, with copepods the most abundant
taxon by a wide margin, as was also observed in traps at a nearby
mudflat (Commito and Tita, 2002). This reversal indicates clearly that
copepods and nematodes do not disperse in the same way, most likely
because copepods can live close to the sediment-water interface and
engage in active emergence from the bottom in response to flow
(Commito and Tita, 2002; Pacheco et al., 2013).

4.3. Implications for mussel bed decline

A serious decline in the abundance of intertidal mussels over the
past several decades has been documented in the Gulf of Maine, in¬
cluding larvae, juveniles, and adults (Petraitis and Dudgeon, 2015;
Sorte et al., 2011, 2016). We have observed this decline in eastern
Maine, where we have conducted mussel bed research since the mid-
1970s. The intertidal, soft-bottom mussel beds in eastern Maine that we
have studied over the years typically had ~ 50% cover of live mussels
in complex fractal power-law spatial patterns (Crawford et al., 2006).
Many formerly robust beds contain virtually no live mussels today
(John A. Commito, personal observation; Brian F. Beal, personal com¬
munication). They now consist largely of bare sediment with patches of
shell hash.

Individual soft-bottom mussel beds can experience short-term po¬
pulation swings (Folmer et al., 2014; Khaitov and Lentsman, 2016;
Nehls and Thiel, 1993). But the Gulf of Maine decline encompasses
much broader spatial and temporal scales and is generally attributed to
climate change (Helmuth et al., 2006; Jones et al., 2010; Lesser, 2016;
Sorte et al., 2011, 2016) and predation by the invasive green crab,
Carcinus maenas (Grosholz and Ruiz, 1996; Tan and Beal, 2015;
Whitlow and Grabowski, 2012). The decline may also be related to
long-term oscillating oceanographic conditions in the northwest
Atlantic region (Beaugrand et al., 2008; Edwards et al., 2013; Greene

et al., 2013). Thus it may be premature to state with certainty that a
permanent Gulf of Maine regime shift is occurring. Regardless of the
reasons behind the reduction in mussels, their decline means that
bottom cover values of whole shells, fragmented shells, and bare sedi¬
ment are likely to be increasing relative to live mussels. To the degree
that Carrying Place Cove may serve as a general model for a mussel bed
that has suffered declines, our results can be useful in helping to predict
the responses to these changes.

As mussels die from causes that do not affect the shell directly, they
produce an increase in whole shell cover, followed by breakdown into
coarse fragmented shell pieces. Crushing and chipping predators like
Carcinus maenas contribute directly to the production of coarse shell
fragments (Tan and Beal, 2015). In addition, whole and fragmented
shells held by mussel byssal threads (Commito et al., 2014) are released
as mussels die. Moreover, in our study we found that coarse material
was captured at higher rates in fragmented shell cover than in bare
sediment, a form of positive feedback that may contribute to the
maintenance of the fragmented shell cover type. Despite this increase in
shell hash, our transect results showed that bare sediment by far com¬
prised the largest bottom cover area in the Carrying Place Cove mussel
bed. This result suggests that shell material is probably being exported,
buried, or broken down into fine particles. Interestingly, although we
have observed changing relative values of each cover type at Carrying
Place Cove over the past several decades, the boundaries of the mussel
bed have remained stationary. Live mussels are still present at the upper
(landward) and lower (seaward) margins of the bed. We observe some
internal shifting of whole and fragmented material on a daily basis,
apparently resulting in net transport to the upper part of the bed, as
seen in Fig. 4. We can find windrows of whole and fragmented mussel
shells along the shoreline above the mussel bed. But unlike storm-in¬
duced movement of empty snail shells described by Nicastro et al.
(2009), relatively little mussel shell material at our site seems so far to
have been exported outside of the original bed boundary, perhaps due

J.A. Commito et al.

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

to its position in a sheltered embayment.

Small patch size may be at least partly responsible for the relatively
small differences we found across cover types for many parameters.
Small patch size may be the norm as intact mussel beds die off and large
patches of live mussels give way to small patches of live mussels, whole
shells, and fragmented shells scattered within large areas of bare sedi¬
ment. Intact mussel beds and other bivalve reefs are important eco¬
system engineers and provide valuable ecosystem services (Borsje et al.,
2011; Bouma et al., 2009; Buschbaum et al., 2009; Commito et al.,
2005, 2008; Gutierrez et al., 2011; Koch et al., 2009; Palumbi et al.,
2009; Stone et al., 2005). If patches reach a small enough size, then a
tipping point may be reached when the system is no longer a functional
mussel bed. Then its ecosystem engineering effects will be gone, as will
the ecosystem services it once provided.

5. Conclusion

M. edulis beds comprise a spatial mosaic of live mussels, whole
shells, fragmented shells, and bare sediment. Sediment and benthic
fauna differ among these cover types, as do their rates of movement in
bedload. Particularly in regions such as the Gulf of Maine, where mussel
beds may be declining in response to factors including climate change
and invasive predators, consideration of these differences will lead to
better understanding of the ecology of mussel beds and soft-bottom
systems overall.

Supplementary data to this article can be found online at https://
doi.org/10.1016/j.jembe.2018.05.004.

Acknowledgments

We thank Angela Commito and Ann Commito for assistance in the
field and the University of Pisa and Gettysburg College for their sup¬
port. Seth Barker, Brian Beal, Tom Crawford, William DeVoe, Heidi
Leighton, Denis Nault, and Ann Thayer provided useful information on
the availability of aerial mussel bed images. Two anonymous reviewers
made comments that significantly improved this paper. Our work was
funded by grants from the Gettysburg College Research and
Professional Development Program (JAC), Gettysburg College Senior
Honors Research Program (BRJ, MAJ, SEW), and Howard Hughes
Medical Institute undergraduate research program to Gettysburg
College (52007540) (SEW).

References

Albrecht, S.H., 1998. Soft bottom versus hard rock: community ecology of macroalgae on
intertidal mussel beds in the Wadden Sea. J. Exp. Mar. Biol. Ecol. 229, 85-109.
Albrecht, A., Reise, K., 1994. Effects of Fucus vesiculosus covering intertidal mussel beds in
the Wadden Sea. Helgol. Meeresunters. 48, 243-256.

Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of var¬
iance. Aust. J. Ecol. 26, 32-46.

Anderson, M.J., 2006. Distance-based tests for homogeneity of multivariate dispersions.
Biometrics 62, 245-253.

Anderson, M.J., Ellingsen, K.E., McArdle, B.H., 2006. Multivariate dispersion as a mea¬
sure of beta diversity. Ecol. Lett. 9, 683-693.

Beaugrand, G., Edwards, M., Brander, K., Luczak, C., Ibanez, F., 2008. Causes and pro¬
jections of abrupt climate-driven ecosystem shifts in the North Atlantic. Ecol. Lett.
11, 157-1168.

Bomkamp, R.E., Page, H.M., Dugan, J.E., 2004. Role of food subsidies and habitat

structure in influencing benthic communities of shell mounds at sites of existing and
former offshore oil platforms. Mar. Biol. 146, 201-211.

Borsje, B.W., van Wesenbeeck, B.K., Dekker, F., Paalvast, P., Bouma, T.J., van Katwijk,

M. M., Vries, M.B., 2011. How ecological engineering can serve in coastal protection.
Ecol. Eng. 37, 113-122.

Bouma, T.J., Olenin, S., Reise, K., Ysebaert, T., 2009. Ecosystem engineering and biodi¬
versity in coastal sediments: posing hypotheses. Helgol. Mar. Res. 63, 95-106.
Buschbaum, C., Dittmann, S., Hong, J.-H., Hwang, I.-S., Strasser, M., Thiel, M., Valdiva,

N. , Yoon, S.-P., Reise, K., 2009. Mytilid mussels: global habitat engineers in coastal
sediments. Helgol. Mar. Res. 63, 47-58.

Chen, P., Fu, Z., Lim, A., Rodrigues, B., 2003. Two-dimensional packing for irregular
shaped objects. In: Proceedings of the 36th Hawaii International Conference on
System Sciences. Institute of Electrical and Electronics Engineers, http://dx.doi.org/
10.1109/HICSS.2003.1174211.

Clarke, K.R., Warwick, R.M., 2001. Changes in Marine Communities: An Approach to
Statistical Analysis and Interpretation, 2 nd ed. PRIMER-E, Plymouth.

Commito, J.A., Rusignuolo, B.R., 2000. Structural complexity in mussel beds: the fractal
geometry of surface topography. J. Exp. Mar. Biol. Ecol. 225, 133-152.

Commito, J.A., Tita, G., 2002. Differential dispersal rates in an intertidal meiofauna as¬
semblage. J. Exp. Mar. Biol. Ecol. 268, 237-256.

Commito, J.A., Currier, C.A., Kane, L.R., Reinsel, K.A., Ulm, I.M., 1995a. Dispersal dy¬
namics of the bivalve Gemma gemma in a patchy environment. Ecol. Monogr. 65,
1 - 20 .

Commito, J.A., Thrush, S.A., Pridmore, R.D., Hewitt, J.E., Cummings, V.J., 1995b.
Dispersal dynamics in a wind-driven benthic system. Limnol. Oceanogr. 40,
1513-1518.

Commito, J.A., Celano, E.A., Celico, H.J., Como, S., Johnson, C.P., 2005. Mussels matter:
postlarval dispersal dynamics altered by a spatially complex ecosystem engineer. J.
Exp. Mar. Biol. Ecol. 316, 133-147.

Commito, J.A., Dow, W.E., Grupe, B.M., 2006. Hierarchical spatial structure in soft-
bottom mussel beds. J. Exp. Mar. Biol. Ecol. 330, 27-37.

Commito, J.A., Como, S., Grupe, B.M., Dow, W.E., 2008. Species diversity in the soft-
bottom intertidal zone: biogenic structure, sediment, and macrofauna across mussel
bed spatial scales. J. Exp. Mar. Biol. Ecol. 366, 70-81.

Commito, J.A., Commito, A.E., Platt, R.V., Grupe, B.M., Piniak, W.E.D., Gownaris, N.J.,
Reeves, K.A., Vissichelli, A.M., 2014. Recruitment facilitation and spatial pattern
formation in soft-bottom mussel beds. Ecosphere 5, artl60. http://dx.doi.org/10.
1890/ESI 4-00200.1 .

Commito, J.A., Gownaris, N.J., Haulsee, D.E., Coleman, S.E., Beal, B.F., 2016. Separation
anxiety: mussels self-organize into similar power-law clusters regardless of predation
threat cues. Mar. Ecol. Prog. Ser. 547, 107-119.

Crawford, T.W., Commito, J.A., Borowik, A.M., 2006. Fractal characterization of Mytilus
edulis L. spatial structure in intertidal landscapes using GIS methods. Landsc. Ecol. 21,
1033-1044.

Dean, W.E., 1974. Determination of carbonate and organic matter in calcareous sedi¬
ments and sedimentary rocks by loss on ignition: a comparison with other methods. J.
Sediment. Petrol. 44, 242-248.

Donadi, S., van der Heide, T., van der Zee, E.M., Eklof, J.S., van de Koppel, J., Weerman,
E.J., Piersma, T., Olff, H., Eriksson, B.K., 2013. Cross-habitat interactions among
bivalve species control community structure on intertidal flats. Ecology 94, 489-498.

Edwards, M., Beaugrand, G., Helaouet, P., Alheit, J., Coombs, S., 2013. Marine ecosystem
response to the Atlantic multidecadal oscillation. PLoS ONE 8 (2), e57212. http://dx.
doi.org/10.1371/journal.pone.0057212.

Folkard, A.M., Gascoigne, J., 2009. Hydrodynamics of discontinuous mussel beds: la¬
boratory flume simulations. J. Sea Res. 62, 250-257.

Folmer, E.O., Drent, J., Troost, K., Biittger, H., Dankers, N., Jansen, J., van Stralen, M.,
Millat, G., Herlyn, M., Philippart, C.J., 2014. Large-scale spatial dynamics of inter¬
tidal mussel ( Mytilus edulis L.) bed coverage in the German and Dutch Wadden Sea.
Ecosystems 17, 550-566.

Gray, J.S., Eliot, M., 2009. Ecology of Marine Sediments, 2 nd ed. Oxford University,
Oxford.

Greene, C.H., Meyer-Gutbrod, E., Monger, B.C., McGarry, L.P., Pershing, A.J., Belkin,

I. M., Fratantoni, P.S., Mountain, D.G., Pickart, R.S., Proshutinsky, A., Ji, R., Bisagni,

J. J., Hakkinen, S.M.A., Haidvogel, D.B., Wang, S.J., Head, E., Smith, P., Reid, P.C.,
Conversi, A., 2013. Remote climate forcing of decadal-scale regime shifts in
Northwest Atlantic shelf ecosystems. Limnol. Oceanogr. 58, 803-816.

Grosholz, E.D., Ruiz, G.M., 1996. Predicting the impact of introduced marine species:
lessons from the multiple invasions of the European green crab Carcinus maenas. Biol.
Conserv. 78, 59-66.

Guay, M., Himmelman, J.H., 2004. Would adding scallop shells ( Chlamys islandica ) to the
sea bottom enhance recruitment of commercial species? J. Exp. Mar. Biol. Ecol. 312,
299-317.

Gutierrez, J.L., Iribarne, O.O., 1999. Role of Holocene beds of the stout razor clam Tagelus
plebeius in structuring present benthic communities. Mar. Ecol. Prog. Ser. 185,
213-228.

Gutierrez, J.L., Jones, C.G., Strayer, D.L., Iribarne, O.O., 2003. Mollusks as ecosystem
engineers: the role of shell production in aquatic habitats. Oikos 101, 79-90.

Gutierrez, J.L., Jones, C.G., Byers, J.E., Arkema, K.K., Berkenbusch, K., Commito, J.A.,
Duarte, C.M., Hacker, S.D., Lambrinos, J.G., Hendriks, I.E., Hogarth, P.J., Palomo,
M.G., Wild, C., 2011. Physical ecosystem engineers and the functioning of estuaries
and coasts. In: Wolanski, E., McLusky, D.S. (Eds.), Treatise on Estuarine and Coastal
Science. Functioning of Ecosystems at the Land-Ocean Interface Vol. 7. Elsevier,
Academic Press, Waltham, MA, pp. 53-81.

Helmuth, B., Mieszkowska, N., Moore, P., Hawkins, S.J., 2006. Living on the edge of two
changing worlds: forecasting the responses of rocky intertidal ecosystems to climate
change. Annu. Rev. Ecol. Evol. Syst. 37, 373-404.

Hewitt, J.E., Thrush, S.F., Halliday, J., Duffy, C., 2005. The importance of small-scale
habitat structure for maintaining beta diversity. Ecology 86, 1619-1626.

Hily, C., 1991. Is the activity of benthic suspension feeders a factor controlling water
quality in the Bay of Brest? Mar. Ecol. Prog. Ser. 69 (179-18).

Hubbard, W.A., 2016. Benthic studies in upper Buzzards Bay, Massachusetts: 2011-12 as
compared to 1955. Mar. Ecol. 37, 532-542.

Hurlburt, S.H., 1984. Pseudoreplication and the design of ecological field experiments.
Ecol. Monogr. 54, 187-211.

Jones, S.J., Lima, F.P., Wethey, D.S., 2010. Rising environmental temperatures and bio¬
geography: poleward range contraction of the blue mussel, Mytilus edulis L., in the
western Atlantic. J. Biogeogr. 37, 2243-2259.

Kent, F.E.A., Last, K.S., Harries, D.B., Sanderson, W.G., 2017. In situ biodeposition mea¬
surements on a Modiolus modiolus (horse mussel) reef provide insights into ecosystem
services. Estuar. Coast. Shelf Sci. 184, 151-157.

J.A. Commito et ah

Journal of Experimental Marine Biology and Ecology 506 (2018) 30-41

Khaitov, V.M., Lentsman, N.V., 2016. The cycle of mussels: long-term dynamics of mussel
beds on intertidal soft bottoms at the White Sea. Hydrobiologia 781, 161-180.

Koch, E.W., Barbier, E.B., Silliman, B.R., Reed, D., Perillo, G.M.E., Hacker, S.D., Granek,
E.F., Primavera, J.H., Muthiga, N., Polaskey, S., Halpern, B.S., Kennedy, C.J., Kappel,
C.V., Wolanski, E., 2009. Non-linearity in ecosystem services: temporal and spatial
variability in coastal protection. Front. Ecol. Environ. 7, 29-37.

Kraeuter, J.N., Kennish, M.J., Dobarro, J., Fegley, S.R., Flimlin, G.E., 2003. Rehabilitation
of the northern quahog (hard clam) (Mercenaria mercnaria) habitats by shelling - 11
years in Bamegat Bay, New Jersey. J. Shellfish Res. 22, 61-67.

Lesser, M.P., 2016. Climate change stressors cause metabolic depression in the blue
mussel, Mytilus edulis, from the Gulf of Maine. Limnol. Oceanogr. 61, 1705-1717.

Mann, R., Southworth, M., Fisher, R.J., Wesson, J.A., Erskine, A.J., Leggett, T., 2016.
Oyster planting protocols to deter losses to cownose ray predation. J. Shellfish Res.
35, 127-136.

McArdle, B.H., Anderson, M.J., 2001. Fitting multivariate models to community data: a
comment on distance-based redundancy analysis. Ecology 82, 290-297.

Nehls, G., Thiel, M., 1993. Large-scale distribution patterns of the mussel Mytilus edulis in
the Wadden Sea of Schleswig-Holstein: do storms structure the ecosystem? Neth. J.
Sea Res. 31, 181-187.

Nicastro, A., Bishop, M., Kelaher, B.P., Benedetti-Cecchi, L., 2009. Export of non-native
gastropod shells to a coastal lagoon: alteration of habitat structure has negligible
effects on infauna. J. Exp. Mar. Biol. Ecol. 374, 31-36.

Pacheco, A.S., Uribe, R.A., Thiel, M., Oliva, M.E., Riascos, J.M., 2013. Dispersal of post-
larval macrobenthos in subtidal sedimentary habitats: roles of vertical diel migration,
water column, bedload transport and biological traits' expression. J. Sea Res. 77,
79-92.

Palumbi, S.R., Sandifer, P.A., Allan, J.D., Beck, M.W., Fautin, D.G., Fogarty, M.J.,

Halpern, B.S., Incze, L.S., Leong, J.-A., Norse, E., Stachowicz, J.J., Wall, D.H., 2009.
Managing for ocean biodiversity to sustain marine ecosystem services. Front. Ecol.
Environ. 7, 204-211.

Petraitis, P.S., Dudgeon, S.R., 2015. Variation in recruitment and the establishment of
alternative community states. Ecology 96, 3186-3196.

Pollock, L.W., 1998. A Practical Guide to the Marine Animals of Northeastern North
America. Rutgers University Press, New Brunswick, New Jersey.

R Core Development Team, 2015. R: A Language and Environment for Statistical

Computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-
project.org/.

Ragnarsson, S.A., Raffaelli, D., 1999. Effects of Mytilus edulis L. on the invertebrate fauna
of sediments. J. Exp. Mar. Biol. Ecol. 241, 31-43.

Reise, K., 2002. Sediment mediated species interactions in coastal waters. J. Sea Res. 48,
127-141.

Ribeiro, P.D., Iribame, O.O., Daleo, P., 2005. The relative importance of substratum
characteristics and recruitment in determining the spatial distribution of the fiddler
crab Uca Uruguay ensis Nobili. J. Exp. Mar. Biol. Ecol. 314, 99-111.

Rodney, W.S., Paynter, K.T., 2006. Comparisons of macrofaunal assemblages on restored
and non-restored oyster reefs in mesohaline regions of Chesapeake Bay in Maryland.
J. Exp. Mar. Biol. Ecol. 35, 39-51.

Singer, J.K., Anderson, J.B., 1984. Use of total grain-size distributions to define bed
erosion and transport for poorly sorted sediment undergoing simulated bioturbation.
Mar. Geol. 57, 335-359.

Smith, J., Shackley, S.E., 2004. Effects of a commercial mussel Mytilus edulis lay on a

sublittoral, soft sediment benthic community. Mar. Ecol. Prog. Ser. 282, 185-191.

Snover, M.L., Commito, J.A., 1998. The fractal geometry of Mytilus edulis spatial dis¬
tribution in a soft-bottom system. J. Exp. Mar. Biol. Ecol. 223, 53-64.

Sorte, C.J.B., Jones, S.J., Miller, L.P., 2011. Geographic variation in temperature toler¬
ance as an indicator of potential population responses to climate change. J. Exp. Mar.
Biol. Ecol. 400, 209-217.

Sorte, C.J.B., Davidson, V.E., Franklin, M.C., Benes, K.M., Doellman, R.J., Etter, R.J.,
Hannigan, R.E., Lubchenco, J., Menge, B.A., 2016. Long-term declines in an intertidal
foundation species parallel shifts in community composition. Glob. Chang. Biol.

http://dx.d 0 i. 0 rg/l 0.1111/gcb. 13425.

Stone, G.W., Zhang, X., Sheremet, A., 2005. The role of barrier islands, muddy shelf and
reefs in mitigating the wave field along coastal Louisiana. J. Coast. Res. 44, 40-55.

Summerhayes, S.A., Bishop, M.J., Leigh, A., Kelaher, B.P., 2009. Effects of oyster death
and shell disarticulation on associated communities of epibiota. J. Exp. Mar. Biol.
Ecol. 379, 60-67.

Tan, E.B.P., Beal, B.F., 2015. Interactions between the invasive European green crab,
Carcinus maenas (L.), and juveniles of the soft-shell clam. My a arenaria L., in eastern
Maine, USA. J. Exp. Mar. Biol. Ecol. 462, 62-73.

Thiel, M., Ullrich, N., 2002. Hard rock versus soft bottom: the fauna associated with
intertidal mussel beds on hard bottoms along the coast of Chile, and considerations
on the functional role of mussel beds. Helgol. Mar. Res. 56, 21-30.

Tomatsuri, M., Kon, K., 2017. Effects of dead oyster shells as a habitat for the benthic
faunal community along rocky shore regions. Hydrobiologia 790, 25-232.

Turner, S.J., Grant, J., Pridmore, R.D., Hewitt, J.E., Wilkinson, M.R., Hume, T.M.,

Morrisey, D. J., 1997. Bedload and water-column transport and colonization processes
by post-settlement benthic macrofauna: does infaunal density matter? J. Exp. Mar.
Biol. Ecol. 216, 51-75.

Underwood, A.J., 1997. Experiments in Ecology: Their Logical Design and Interpretation
Using Analysis of Variance. Cambridge University, Cambridge.

Urban-Malinga, B., Zalewski, M., Barnes, N., 2016. Effects of various bivalves on meio-
benthic and nematode assemblages in shallow sandy sediments. Hydrobiologia 772,
131-144.

Valanko, S., Norkko, A., Norkko, J., 2010a. Rates of post-larval bedload dispersal in a
non-tidal soft-sediment system. Mar. Ecol. Prog. Ser. 416, 145-163.

Valanko, S., Norkko, A., Norkko, J., 2010b. Strategies of post-larval dispersal in non-tidal
soft-sediment communities. J. Exp. Mar. Biol. Ecol. 384, 51-60.

wa Kangeri, A.K., Jansen, J.M., Joppe, D.J., Dankers, N.M.J.A., 2016. In situ investigation
of the effects of current velocity on sedimentary mussel bed stability. J. Exp. Mar.
Biol. Ecol. 485, 65-72.

Whitlow, W.L., Grabowski, J.H., 2012. Examining how landscapes influence benthic
community assemblages in seagrass and mudflat habitats in southern Maine. J. Exp.
Mar. Biol. Ecol. 411, 1-6.

Widdows, J., Pope, N.D., Brinsley, M.D., Gascoigne, J., Kaiser, M.J., 2009. Influence of
self-organised structures on near-bed hydrodynamics and sediment dynamics within
a mussel ( Mytilus edulis ) bed in the Menai Strait. J. Exp. Mar. Biol. Ecol. 379, 92-100.

Wilding, T.A., Nickell, T.D., 2013. Changes in benthos associated with mussel ( Mytilus
edulis L.) farms on the West-Coast of Scotland. PLoS ONE 8 (7), e68313. http://dx.
doi.org/10.1371/journal.pone.0068313.

Ysebaert, T., Hart, M., Herman, P.M. J., 2009. Impacts of bottom and suspended cultures
of mussels Mytilus spp. on the surrounding sedimentary environment and macro-
benthic diversity. Helgol. Mar. Res. 63, 59-74.