BENTHIC COMMUNITY DEVELOPMENT ON EDGE VS. INTERIOR …dl.uncw.edu › Etd › 2011-1 › r1 › novakc › coreynovak.pdf · Melody Ovard, Trey Sherard, Heather Stoker, Chris Swanson,

BENTHIC COMMUNITY DEVELOPMENT ON EDGE VS. INTERIOR OF CREATED SALT MARSHES

Corey S. Novak

A Thesis Submitted to the University of North Carolina Wilmington in Partial Fulfillment

of the Requirements for the Degree of Master of Science

Department of Biology and Marine Biology

University of North Carolina Wilmington

2011

Approved by

Advisory Committee

Troy D. Alphin Lynn A. Leonard Thomas E. Lankford Martin H. Posey Chair

Accepted by

______________________________ Dean, Graduate School

ii  

TABLE OF CONTENTS

ABSTRACT................................................................................................................................... iii

ACKNOWLEDGEMENTS........................................................................................................... iv

DEDICATION.................................................................................................................................v

LIST OF TABLES......................................................................................................................... vi

LIST OF FIGURES ...................................................................................................................... vii

INTRODUCTION ...........................................................................................................................1

METHODS ......................................................................................................................................5

Study Sites ...........................................................................................................................5

Sample Collection................................................................................................................8

Data Analysis .......................................................................................................................9

RESULTS ......................................................................................................................................11

Benthic Infauna..................................................................................................................11

Sediment Characteristics....................................................................................................23

Vegetation ..........................................................................................................................26

DISCUSSION................................................................................................................................31

LITERATURE CITED ..................................................................................................................39

iii  

ABSTRACT

Created salt marshes are increasingly used to mitigate for the functions lost when natural

marshes are impacted. However, the issue of which functions have been replaced and to what

extent remains controversial. Certain nekton utilize the edge of the marsh to a greater extent

than the marsh interior, suggesting that there also may be spatial components to marsh

development. Benthic invertebrates serve as an important intermediate trophic link, and

understanding their spatial and developmental patterns is important to understanding overall food

web dynamics. The temporal and spatial development of the benthic infaunal community was

examined among 5 created (3 to 18 yr in age) and 5 natural marshes in southeastern North

Carolina, USA, to test whether edge and interior habitats become similar to natural marshes at

different rates. All created marshes were generally similar to natural marshes with regard to

infauna, vegetation, and sediment. Although there were site differences, similarities among

created and paired reference marshes suggest these are related to factors other than age.

Oligochaetes, dominant infauna at most sites, were more abundant in the interior than the edge.

Polychaetes were generally evenly distributed between the edge and interior. Capitellids were

the most dominant infauna at the youngest created marsh. Although oligochaetes were positively

correlated with sediment organic content and macro-organic matter, these variables did not

consistently differ between edge and interior habitats. These data suggest that while the edge

and interior exhibit some important differences, both habitats in created marshes may resemble

those of natural marshes early in development, in as little as 3 years. These results contrast with

previous studies where created marshes exhibited lower oligochaete abundance and sediment

organic content after longer time periods, some as much as 25 years.

iv  

ACKNOWLEDGEMENTS

I would like to thank the members of the UNCW Benthic Ecology Lab including Steve

Artabane, Sherry Banner, Russ Barbour, Jeremy LaRosa, Jason Lautenschleger, Anne Markwith,

Melody Ovard, Trey Sherard, Heather Stoker, Chris Swanson, Sharon Tatem, and Ashley Whitt

for assistance with field work and laboratory sample processing. Special thanks go to my wife,

Lori Novak, for providing great company and assistance with some weekend field work and for

many hours in the laboratory sorting benthic infauna.

Site selection was greatly aided by Dave Meyer, Ted Wilgis, and Chuck Wilson. These

gentlemen went above and beyond by not only providing valuable information, but also assisting

in data collection. Frank Yelverton and Tracy Skrabal also provided greatly appreciated

information regarding existing created marshes.

Some financial support for this project was provided by North Carolina Sea Grant.

Finally, I would like to thank my committee members for their guidance and support.

Courtney Hackney contributed to study design and gave valuable comments on the prospectus. I

am grateful that Tom Lankford was able to join the committee upon Dr. Hackney’s retirement.

Lynn Leonard has contributed her expertise in many ways, especially concerning marsh

sediment characteristics and dynamics. This thesis would not have been possible without the

guidance of my advisors, Troy Alphin and Martin Posey.

v  

DEDICATION

I would like to dedicate this thesis to my wife, Lori Novak, who has made many

sacrifices in order to make this thesis possible.

vi  

LIST OF TABLES

Table Page 1. Study site general information .............................................................................................7 2. Percent abundance for dominant taxa for non-river sites averaged between years (A) and non-river sites (B) .....................................................................................12 3. Results from ANCOVAs comparing benthic infaunal abundances among locations (edge, interior) and created marsh age ...............................................................20 4. Results from ANOVAs comparing benthic infaunal abundances among locations (edge, interior) and sites .....................................................................................22 5. Pearson correlation coefficients for 2007 data...................................................................27

vii  

LIST OF FIGURES

Figure Page 1. Overview maps of southeastern North Carolina, USA showing locations of study sites .......................................................................................................................6 2. Mean total infaunal abundance per core (0.01 m2) ............................................................13 3. Mean oligochaete abundance per core (0.01 m2)...............................................................14 4. Mean spionid abundance per core (0.01 m2) .....................................................................15 5. Mean nereid abundance per core (0.01 m2) .......................................................................16 6. Mean capitellid abundance per core (0.01 m2) ..................................................................17 7. Mean total abundance (excluding oligochaetes) per core (0.01 m2)..................................18 8. MDS Plot for infaunal abundance at all sites.....................................................................24 9. Mean percent sediment organic content ............................................................................25 10. Mean macro-organic matter content ..................................................................................28 11. Spartina alterniflora height (mean of tallest stem of 4 m2 quadrats) ................................30

INTRODUCTION

As the ecological consequences (e.g. declining fisheries yields) of salt marsh impacts

have been realized, much effort has been made to mitigate for the loss of these areas and their

associated functions and values. One of the most important challenges associated with salt

marsh creation is establishing criteria for determining whether the lost functions of the original

marsh have been replaced by the created marsh (Moy and Levin 1991; Minello and Zimmerman

1992; Posey et al. 1997; Craft et al. 1999; Streever 2000). Numerous studies have shown that the

vegetation component of created salt marshes can resemble that of natural marshes within ~5

years of marsh construction (Broome et al. 1988; Craft et al. 1999; Streever 2000). However, the

development of sediment and benthic infaunal community characteristics in created marshes is

more complex and may lag behind that of vegetation.

Several studies of species composition in created marshes have revealed that

opportunistic species, such as the polychaetes Streblospio benedicti and Capitella capitata, are

the first to colonize, and then over time these opportunists may become mixed with poor-

dispersing and/or more specialized taxa that are more characteristic of natural marshes, such as

oligochaetes (Moy and Levin 1991; Levin et al. 1996; Alphin and Posey 2000; Craft and Sacco

2003). Oligochaetes may be more common in natural marshes than young created marshes due

to their life cycles and feeding strategies. In general, oligochaetes are subsurface-deposit feeders

that derive nutrition by ingesting sediment that is high in organics (McCann and Levin 1989).

Oligochaetes have also been noted to have a limited dispersal stage (Levin et al. 1996; Brusca

and Brusca 2003). They produce few, large young with direct development, lacking planktonic

2  

larvae. Relative to many polychaetes, oligochaetes have longer generation times with a slower

rate of population increase and shorter dispersal distance.

The available literature comparing infauna in created marshes to natural marshes have

reached variable conclusions regarding whether the created marshes are considered successful.

LaSalle et al. (1991) reported that an older created marsh (eight years) was similar to a newly

created marsh (four years) in the same area in Winyah Bay, South Carolina. However, more

commonly the published literature indicates that differences between created and natural marshes

in infaunal species composition and/or densities may persist over a longer time period (Moy and

Levin 1991; Minello and Zimmerman 1992; Sacco et al. 1994; Levin et al. 1996; Posey et al.

1997; Craft et al. 1999; Alphin and Posey 2000; Streever 2000; Craft and Sacco 2003). Some of

these studies have found infaunal community differences between natural reference marshes and

created marshes as old as 25 years, with the most common difference being lower densities of

oligochaetes in the created marshes.

Most comparative studies of created marshes and natural marshes do not address smaller

scale spatial distribution (edge vs. interior) within the marsh. However, there is evidence that

different infaunal community types occupy the edge and interior of the marsh. Minello et al.

(1994), studying the effects of creating channels in a planted marsh (increasing the amount of

marsh edge), found significantly higher densities of polychaetes and decapod predators within 1

m of the channels and oligochaetes frequently more abundant 10 m from the edge. Whaley and

Minello (2002) also found more polychaetes within 1 m of the edge, while oligochaetes were

more evenly distributed throughout the marsh and often most abundant at 10 m from the edge.

However, Minello and Webb (1997) failed to find differences in infauna densities 5 m from the

edge versus 1 m from the edge. The spatial distribution of benthic infauna in marshes may affect

3  

fish use of the marsh or vice versa, as many fish species forage for infaunal prey along the marsh

edge but not the interior (Boesch and Turner 1984; McIvor and Odum 1988; Baltz et al. 1993;

Peterson and Turner 1994).

In addition to the nekton and infaunal community differences between edge and interior,

the vegetation may also differ as well. In S. alterniflora marshes, often the tall form will

dominate along the edge while the short form is more prevalent in the interior (Howes et al.

1986). Edge S. alterniflora may also contain a less extensive below-ground root matrix than

interior S. alterniflora. Whaley and Minello (2002) found that the amount of below-ground

living macroorganic matter increased with distance from the marsh edge along with oligochaete

abundance.

It is possible that the edge of created marshes may develop more quickly (relative to

natural marshes) due to a combination of factors. First, many infaunal species that are common

in the edge (i.e. most polychaetes) are good dispersers, exhibiting high vagility and many with

planktonic larvae. The marsh edge may be thought of as a disturbed community due to impacts

by waves and tidal flow, in which recolonization by opportunistic species may occur frequently.

The marsh edge also has greater connectivity to sources of planktonic recruitment. In contrast,

oligochaetes, which are more frequently associated with the interior, have low vagility with

direct development. Second, the edge may contain more planktonic food for infauna, which

should enhance colonization of this area. Third, edge infauna are more susceptible to predation,

meaning that infaunal abundance in the edge of natural marshes after predation has peaked may

be low, so that the edge of created marshes may resemble that of natural marshes early in

development with low infaunal abundance. Finally, the interior of natural marshes frequently

contains a higher percentage of organic sediment than the edge (Whaley and Minello 2002).

4  

This organic material can take several years to develop (Broome et al. 1988; Broome et al.

2000). Thus, the sediment in the interior of created marshes may not resemble that of natural

marshes until the organic material has had sufficient time to develop, since created marshes are

often constructed with sediment low in organics. In contrast, the edge of natural marshes is often

sandier with less organic material (Whaley and Minello 2002). These sediment differences

between the edge and interior should be reflected in the infaunal community, as oligochaetes

may take longer to colonize the interior of created marshes due to their preference for higher

organic sediment and their relatively poor dispersal abilities. Some representative polychaete

taxa should be able to colonize the edge of created salt marshes relatively quickly (assuming a

source habitat exists in close proximity), as many are superior dispersers and do not necessarily

require sediment high in organic content.

The principle goal of this study is to test the hypothesis that the infaunal community

inhabiting the edge of created marshes will differ from the interior and will approach that of

similar habitats in natural marshes at a faster rate than the interior. In order to answer the

question of whether the edge and interior mature at different rates, I examined both created

marshes of varying ages and spatial distribution patterns. I predicted that the infauna of created

marshes would become similar to that of natural marshes over time, but this process would take

several years (>3 yrs) with the rate varying among sites. I also predicted that the marsh edge

would contain more surface dwelling polychaetes than the interior, while the interior would

contain more deposit feeding oligochaetes than the edge. By understanding these dynamics, the

process of salt marsh creation can be refined to better achieve desired outcomes, i.e. constructing

marshes with edge:interior ratios that mimic natural marshes.

5  

MATERIALS AND METHODS

Study sites

To examine whether the benthic communities inhabiting the edge of Spartina alterniflora

marshes become similar to natural marshes more quickly than the interior areas, two groups of

created marshes of varying ages in southeastern North Carolina were sampled. All of the created

marshes in this study were excavated from dredge spoil material for the purpose of mitigation.

The first group of marshes consists of 3 euhaline coastal embayment marshes (mean salinity =

33.6 ppt), while the second group includes 2 mesohaline marshes in the lower Cape Fear River

(mean salinity = 13.7 ppt) (Fig. 1, Table 1). The euhaline non-river created marshes are Mason

(Fig. 1B), Army Reserve Center (ARC) and Port (Fig. 1C). The mesohaline river marshes are

Island 13 and Port 2 (Fig. 1A). All sites, except the Port 2 created and reference marshes,

contain interior tidal creeks. Within each group, sites were selected that appeared to have similar

geomorphology, salinity, soil characteristics and vegetation but different ages (following a

“chronosequence approach”, Craft and Sacco 2003). Sites were selected to be larger restoration

projects with similar construction methods and known histories. Several of the created marsh

sites were monitored at the time of establishment, and some were monitored annually for a few

years after establishment. Previous studies have described both the Port (Craft and Sacco 2003;

Currin et al. 1996; Levin et al. 1996; Piehler et al. 1998) and ARC (Piehler et al. 1998) created

marshes.

The effects of both marsh age and distance from the marsh edge on infaunal composition

were explored at each site by differences between created and paired natural reference marshes

6  

0 1 20.5Kilometers

0 0.50.25Kilometers

ATLANTIC OCEAN

NORTH CAROLINA

WILMINGTON

SWANSBORO

SNEADSFERRY

BEAUFORTINLETBOGUE INLET

NEW RIVER INLET

MOREHEAD CITY

MASON INLET SITE

PORT/ARMY MARSH SITES

PORT 2 SITE

ISLAND 13 SITE

CAPE FEAR RIVER

NE C

APE FEAR RIV

ER

Morehead CityState Port

PortCreated

ARCRef

ARCCreated

PortRef

MasonCreated

MasonRef

Figu

re 8

Islan

d

Wrig

htsv

ille B

each

0 1 20.5Kilometers

Port 2 Ref

Port 2Created

Island 13Created

Island 13Ref

Wilm

ingtonState Port

Rive r Rd .

Figure 1. Overview maps of southeastern North Carolina, USA showing locations of study sites. Each “site” consists of a created marsh and its nearby natural reference (ref) marsh. Map source: NAPP 2009 Aerial Photography, Carteret and New Hanover Counties, North Carolina.

A  B  C 

7  

Table 1.  Study site general information.  * Island 13 size represents the portion of the mitigation site sampled in this study.  The study area is part of an approximately 12‐hectare created salt marsh/tidal creek complex. 

Created Marsh Year 

Created  Type Size (ha) 

Center Coordinates (N,W) 

Reference Marsh  Center Coordinates (N,W) 

Size** (ha) 

Distance to Created Marsh (km) 

Mason  2003  Non‐river  3.5  34.262336, ‐77.768022  34.230301, ‐77.789825  1.1  4.04 Army Reserve Center(ARC)  1995  Non‐river  2.6  34.732687, ‐76.695652  34.734620, ‐76.696343  2.6  0.18 

Port  1990  Non‐river  1.2  34.726653, ‐76.697366  34.728982, ‐76.698056  0.8  0.29 

Island 13  2001  River  1.1*  34.164638, ‐77.955733  34.149902, ‐77.959567  0.5  1.68 

Port2  1989  River  2.1  34.187737, ‐77.959454  34.200981, ‐77.960301  0.5  1.59  

**Relevant area of reference marshes estimated by relevant boundary creeks and upland (likely underestimated relative to total created marsh area).

8  

that appeared to be similar with regard to location (geographic area and physical setting),

salinity, substrate, elevation, and vegetation. Each natural reference marsh is located within 4

km of its analogous created marsh, and several of them are much closer.

Sample collection

Sampling of benthic infauna at the 3 created and 3 reference non-river sites was

conducted in summer (June-August) 2006 and again in summer 2007. The 2 created and 2

reference river sites were sampled in summer 2007 only. Each marsh site consisted of 2 habitats

(edge, interior) sampled along 4 transects, for a total of 128 infaunal cores. The marsh edge was

defined as the vegetated area within 1.5 m of open water (Minello et al. 1994; Minello and Webb

1997; Whaley and Minello 2002). The marsh interior was defined as the vegetated marsh

between 10 and 15 m from open water (Minello et al. 1994; Whaley and Minello 2002). To

reduce potential confounding effects of elevation, edge and interior locations were selected that

had similar elevations based on water depth after submergence.

All infaunal sediment cores were collected after marsh exposure. Cores were 12 cm in

diameter and 15 cm deep. The entire core samples were placed in plastic bags in the field and

then returned to the laboratory. In the laboratory, cores were preserved in 10% buffered formalin

with rose Bengal dye (buffered to saturation with sodium tetraborate), allowed to fix for at least

48 hours, sieved through a 0.5 mm screen, then transferred to 70% isopropanol until sorted.

Infauna were sorted from the sediment and root material using a dissecting microscope. All

infauna were identified to the family taxonomic level where practical, dried in an oven at 70° C

until all water was removed, and weighed.

9  

During the 2007 sampling, two additional sediment cores were taken at each habitat at

each site to compare sediment grain size and sediment organic content (SOC) among locations.

Sediment grain size was determined in the laboratory using a particle analyzer (Beckman

Coulter, LS200). SOC was determined using the combustion method. The SOC cores were

sieved through a 3 mm screen to remove stems and roots, placed in an oven at 70° C until all

water was removed, ground with mortar and pestle, weighed, fired at 500° C in a muffle furnace

for 4 hours, then weighed again to determine the SOC through subtraction.

Sediment macroorganic matter content (MOM) was determined from the infaunal cores

collected during the second year of sampling (2007). After all invertebrates were sorted and

removed from the sample, roots were separated from the sample, dried in an oven at 70° C until

all water was removed, then weighed.

During each of the two annual sampling events, four 0.25m X 0.25m PVC vegetation

quadrats were placed randomly within 2m of the locations where the infaunal cores were taken.

Within the quadrat, all stems from each species were counted. In addition, the height of the

tallest stem within each quadrat was measured.

Data analysis

Analysis of Variance (ANOVA) was used to test for differences among sites and between

locations (edge vs. interior), with years analyzed separately because of interannual variation in

community structure. All data was log10-transformed where variances were heterogeneous. To

examine patterns in benthic infaunal abundance, dependent variables tested included total

benthic infaunal abundance (excluding nematodes) and abundance of common taxa including

oligochaetes, capitellids, spionids, and nereids. Vegetation and sediment characteristics analyzed

10  

using ANOVA included Spartina height, Spartina density, SOC, sediment grain size, and MOM.

The two years of vegetation data were analyzed separately. In all ANOVA tests, data were

analyzed separately by site and by location where significant interactions occurred.

Analysis of Covariance (ANCOVA) was used to examine relations with created marsh

age, using the dependent variables indicated above with marsh age as the covariate and

edge/interior as the independent variable. Reference marshes were not included in these analyses

since their ages are unknown. For the second year of sampling (2007), data from the river and

non-river sites were combined since only 2 river sites were sampled. An independent variable

for marsh type (river vs. non-river) was added. Where significant interactions existed between

marsh type and site and/or location, non-river sites were tested separately. Where significant

interactions existed between site and location, linear regression was employed to test for

relations with marsh age within locations separately. All ANOVAs and ANCOVAs were

performed using SAS (v 9.1, Cary, North Carolina, USA).

Community similarity and dominance patterns were explored through Analysis of

Similarity (ANOSIM) and Multi-Dimensional Scaling (MDS) using PRIMER (v6, Roborough,

Plymouth, United Kingdom). Since fauna may covary with soil and vegetation characteristics

among sites, between locations, and with marsh age, a correlation analysis was performed to

assess relationships between vegetation and sediment characteristics and the dependent variables

used in the infaunal abundance ANOVAs. In order to assess whether young created marshes are

dominated by opportunistic taxa while older created marshes and natural marshes are dominated

by more specialized taxa such as oligochaetes, dominance patterns as indicated by rank

abundance were examined. Dominant taxa are defined as those taxa comprising at least 3

percent of the total abundance within a site during each annual sampling event.

11  

RESULTS

Benthic Infauna

Dominant taxa at most non-river sites (in decreasing order) were capitellids, oligochaetes,

spionids, and nereids (Table 2). With the exception of the ARC site, the non-river created

marshes contained a higher percentage of capitellids than oligochaetes while the reference

marshes were more dominated by oligochaetes than by capitellids. The river sites were generally

dominated by oligochaetes, spionids, and insects, with capitellids being much less common.

With regard to infaunal abundance patterns, the younger created marshes (Mason and

ARC) were more similar to their reference marshes than the oldest created marsh (Port). During

2006, there were no significant differences between the Mason created marsh and its reference

marsh pair for total infaunal abundance (Fig. 2), oligochaetes (Fig. 3), spionids (Fig. 4), or

nereids (Fig. 5). The Mason created marsh contained more capitellids than its analogous

reference marsh in 2006 (F=4.901,12, p=0.0470; Fig. 6). The ARC created marsh exhibited

significantly greater total abundance than its reference marsh in 2006 (F=9.391,12, p=0.0098)

while other taxa did not differ. When oligochaetes were removed from total abundance, it was

still significantly greater at the ARC created marsh than at its reference marsh (F=13.391,12,

p=0.0033; Fig. 7). The oldest non-river created marsh (Port) exhibited significantly lower total

abundance (F=5.931,12, p=0.0314) and oligochaetes (F=10.451,12, p=0.0072) than its reference

marsh in 2006. When oligochaetes were removed from total abundance, the Port created marsh

no longer differed from its reference marsh (Fig. 7). Other taxa differed in the marsh interior but

not in the edge. Within the interior, the Port created marsh contained significantly more spionids

12  

Table 2.  Percent abundance for dominant taxa for non‐river sites averaged between years (A) and  non‐river sites (B).   Taxa comprising at least 3% of total abundance at at least one site are shown in rank order for all sites combined. 

A (Non‐River Sites) 

               Mason ARC Port Created Reference Created Reference Created Reference Total Abundance 350 316 761 583 822 1441 Capitellidae  41.4%  15.9%  17.3%  32.2%  25.2%  28.5% Oligochaeta  26.4%  38.1%  8.6%  20.4%  14.8%  41.9% Spionidae  0.0%  0.2%  14.4%  18.2%  27.6%  15.7% Nereidae  0.0%  2.9%  10.3%  13.7%  13.0%  5.3% Insecta  13.4%  9.6%  3.2%  2.9%  1.4%  2.6% Gastropoda  2.5%  0.6%  17.4%  3.4%  11.6%  0.3% Isopoda  6.9%  10.2%  0.0%  0.0%  0.0%  0.0% Cirratulidae  0.0%  0.0%  12.9%  0.1%  0.6%  0.2% Bivalvia  0.4%  0.2%  7.5%  0.2%  1.5%  0.0% Ostracoda  1.1%  0.0%  4.4%  0.1%  0.8%  0.0%  

B (River Sites) 

Island 13 Port 2 Created Reference Created ReferenceTotal Abundance 1561 1380 3007 387 Oligochaeta 58.6% 57.8% 87.2% 59.9% Spionidae 25.6%  29.1%  9.2%  33.3% Insecta 4.5%  7.8%  1.2%  9.6% Capitellidae 2.5%  0.1%  0.4%  3.4% Ampharetidae 3.5%  0.1%  0.5%  0.0% 

13  

A (2006 Non‐River) 

B (2007 Non‐River) 

C (2007 River) 

Figure 2. Mean total infaunal abundance per core (0.01 m2). Each mean is based on 4 cores. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Asterisk indicates created marsh is significantly different from its reference marsh (p-value between 0.001 and 0.05).

*

*

*

*

14  

A (2006 Non‐River) 

B (2007 Non‐River) 

C (2007 River) 

Figure 3. Mean oligochaete abundance per core (0.01 m2). Each mean is based on 4 cores. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Asterisks indicate created marsh is significantly different from its reference marsh. One asterisk indicates p<0.05, and two asterisks indicate p<0.001.

**

*

*

15  

A (2006 Non‐River) 

B (2007 Non‐River) 

C (2007 River) 

Figure 4. Mean spionid abundance per core (0.01 m2). Each mean is based on 4 cores. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Port sites are significantly different in Interior only for 2006 (p=0.0402) and Edge only for 2007 (p=0.0089). Other created marshes do not differ significantly from reference marshes.

16  

A (2006 Non‐River) 

B (2007 Non‐River) 

C (2007 River) 

Figure 5. Mean nereid abundance per core (0.01 m2). Each mean is based on 4 cores. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Port sites are significantly different in Interior only for 2006 (p=0.0280). Other created marshes do not differ significantly from reference marshes.

17  

A (2006 Non‐River)

B (2007 Non‐River) 

C (2007 River) 

Figure 6. Mean capitellid abundance per core (0.01 m2). Each mean is based on 4 cores. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Asterisk indicates created marsh is significantly different from its reference marsh (p-value between 0.001 and 0.05). Port sites are significantly different in Interior only for 2006 (p=0.0072).

* 

*

18  

A (2006 Non‐River) 

B (2007 Non‐River) 

C (2007 River) 

Figure 7. Mean total abundance (excluding oligochaetes) per core (0.01 m2). Each mean is based on 4 cores. Error bars represent one standard error. Created marshes are in order of increasing age. Site abbreviations same as previous figures. Asterisk indicates created marsh is significantly different from its reference marsh (p-value between 0.001 and 0.05). Port sites are significantly different in Interior only for 2007 (p=0.0433).

* 

19  

(F=5.291,12, p=0.0402) and nereids (F=6.241,12, p=0.0280) but less capitellids (F=10.441,12,

p=0.0072).

Infaunal abundance patterns in the non-river marshes were generally similar between

2006 and 2007. For the younger created marshes (Mason and ARC) in 2007, there were no

significant differences between created and reference marsh pairs for total infaunal abundance

(Fig. 2), oligochaetes (Fig. 3), spionids (Fig. 4), nereids (Fig. 5), or capitellids (Fig. 6). In 2007,

the oldest non-river created marsh (Port) exhibited lower total abundance (F=9.521,12, p=0.0094)

and oligochaete abundance (F=21.491,12, p=0.0004) than its reference marsh. When oligochaetes

were removed from total abundance, the Port created marsh still contained significantly lower

total abundance than its reference marsh in the interior (F=5.101,12, p=0.0433) but not in the edge

(Fig. 7). Spionid abundance was significantly lower at the Port created marsh than at its

reference marsh in the edge (F=9.721,12, p=0.0089) but not in the interior. Capitellid abundance

at the Port created marsh was significantly higher in the edge (F=5.081,12, p=0.0437) but lower in

the interior (F=18.381,12, p=0.0001) compared to its reference marsh. Nereid abundance did not

differ between the Port created and reference marshes in 2007 (Fig. 5). The river sites exhibited

few differences between created and reference marsh pairs. The oldest created marsh (Port 2)

had greater total abundance (F=7.971,12, p=0.0154; Fig. 2) and greater oligochaete density

(F=6.831,12, p=0.0227; Fig. 3) than its reference marsh. When oligochaetes were removed from

total abundance, there were no significant differences within either created/reference marsh pair

for the river sites (Fig. 7). There were no significant differences within either created/reference

marsh pair for spionids (Fig. 4), nereids (Fig. 5), or capitellids (Fig. 6) among the river sites.

Total infaunal abundance did not consistently exhibit clear patterns in relation to created

marsh age. In 2006, total abundance was not significantly related to marsh age (Fig. 2, Table 3).

20  

Table 3.  Results from ANCOVAs comparing benthic infaunal abundances among locations (edge, interior) and created marsh age.   For non‐river sites, N=24 and degrees of freedom (Model, error)=3,20.  For river and non‐river sites combined, N=40 and degrees of freedom (Model, error)=7,32.  Where age is significant, r2 values are given in parentheses.  Nonsignificant (ns) p‐values greater than 0.05 are not listed.  L=Location, A=Age, and T=Type (river, non‐river).         

A 

  2006 Non‐River Parameter  F  Model  Location  Age  Interaction 

Total Abundance  3.38  0.0385  0.0353  ns  ns Oligochaetes  8.28  0.0009  <0.0001  ns  ns Total Excluding Oligochaetes  3.44  0.0364  0.0405  ns  ns Spionidae  8.95  0.0060  ns  <0.0001 (0.57) ns Capitellidae  0.27  ns  ns  ns  ns Nereidae  15.21  <0.0001 ns  <0.0001 (0.70) 0.0441 

B 

  2007 Non‐River Parameter  F  Model  Location  Age  Interaction 

Total Abundance  4.13  0.0197  ns  0.0027 (0.38)  ns Oligochaetes  2.89  ns  0.0280  ns  ns Total Excluding Oligochaetes  3.95  0.0230  ns  0.0032 (0.37)  ns Spionidae  9.48  0.0004  ns  <0.0001 (0.59)  ns Capitellidae  2.49  ns  ns  0.0429 (0.52)  ns Nereidae  6.33  0.0034  ns  0.0003 (0.50)  ns 

C 

2007 River & Non‐River Combined Parameter  F  Model  Location  Age  Type  Interactions 

Total Abundance  7.62  <0.0001  ns  0.0274 (0.63)  0.0002  L*A 0.0154               A*T 0.0256             L*A*T  0.0007Oligochaetes  12.68  <0.0001  0.0114  ns  <0.0001  L*A 0.0008              L*A*T 0.0419 Total Excluding Oligochaetes  3.22  0.0106  ns  ns  ns  A*T 0.0034              L*A*T 0.0142 Spionidae  11.10  <0.0001  ns  0.0092 (0.71)  <0.0001  L*A 0.0235               A*T <0.0001             L*A*T 0.1407 Capitellidae  6.23  <0.0001  0.0425  ns  <0.0001  A*T 0.0076 Nereidae  11.66  <0.0001  ns  <0.0001 (0.72) <0.0001  A*T <0.0001 

21  

In 2007, total abundance significantly increased with created marsh age in the non-river sites.

River sites were excluded from analysis of total abundance and marsh age due to interaction

between age and marsh type (river vs. non-river). Oligochaete abundance was not significantly

related to marsh age in 2006 (Fig. 3, Table 3). However, when all sites were considered in 2007,

interaction existed between age and location (Table 3). In 2007, oligochaete abundance

significantly increased with marsh age within the marsh interior but not within the edge,

suggesting some differential development between habitats (Fig. 3, Table 3). River sites

contained more oligochaetes than non-river sites (Fig. 3).

Capitellid abundance was not significantly related to created marsh age in 2006 (Table 3).

For nereids, interaction existed between age and location. As with oligochaete abundance,

nereid abundance was significantly related to marsh age in the marsh interior (r2=0.82,

F=46.271,10 , p<0.0001) but not in the edge (r2=0.20, F=2.441,10 , p=0.1490). Spionid abundance

increased with marsh age in 2006 (Fig. 4, Table 3). In 2007, interaction existed between marsh

type and created marsh age for abundances of spionids, capitellids and nereids. Within the non-

river sites, abundances of all three polychaete families were significantly related to marsh age

(Figs. 4-6, Table 3).

Within the non-river sites, total abundance was significantly greater in the marsh interior

than in the edge in 2006 (Fig. 2, Table 4). When oligochaetes were excluded from the 2006

analysis, location was still significant (Fig. 7, Table 4). In 2007, total abundance was

consistently greater in the interior, although this was not significant. The Port2 created marsh

was the only river site that exhibited significantly greater total abundances in the interior

(F=17.191,24, p=0.0004; Fig. 2). Oligochaetes were more abundant in the marsh interior than in

the edge during both years for all sites except the Island 13 created marsh (Fig. 3, Table 4).

22  

Table 4.  Results from ANOVAs comparing benthic infaunal abundances among locations (edge, interior) and sites.  For non‐river sites, N=48 and degrees of freedom (Model, error)=11,36.  For river sites, N=32 and degrees of freedom (Model, error)=7,24.  Nonsignificant (ns) p‐values greater than 0.05 are not listed.         

A 

2006 Non‐River Parameter  F  Model  Location  Site  Location*Site 

Total Abundance  4.06  0.0007  0.0089  0.0004  ns Oligochaetes  5.03  0.0001  <0.0001  0.0146  0.0227 Total Excluding Oligochaetes  5.06  <0.0001 0.0244  <0.0001 ns Spionidae  7.35  <0.0001 ns  <0.0001 0.0023 Capitellidae  2.93  0.0074  ns  0.0180  ns Nereidae  14.32  <0.0001 0.0035  <0.0001 0.0099  

B 

2007 Non‐River Parameter  F  Model  Location  Site  Location*Site 

Total Abundance  6.04  <0.0001  ns  <0.0001  ns Oligochaetes  4.95  0.0001  0.0002  0.0007  ns Total Excluding Oligochaetes  3.47  0.0023  ns  0.0004  ns Spionidae  8.42  <0.0001  ns  <0.0001  0.0314 Capitellidae  3.56  0.0019  ns  0.0008  ns Nereidae  3.21  0.0040  ns  0.0020  ns 

C 

2007 River Parameter  F  Model  Location  Site  Location*Site 

Total Abundance  5.47  0.0008  0.0400  0.0049  0.0045 Oligochaetes  7.67  <0.0001  0.0008  0.0014  0.0036 Total Excluding Oligochaetes  3.90  0.0056  ns  0.0057  0.0287 Spionidae  3.61  0.0085  ns  0.0171  0.0175 Capitellidae  1.01  ns  ns  ns  ns Nereidae  0.55  ns  ns  ns  ns 

23  

However, in 2006, the difference was only significant at half of the sites (ARC created

F=5.051,36, p=0.0309 , Port created F=11.421,36, p=0.0018, Port reference F=19.471,36, p<0.0001).

No significant differences existed for spionids (Fig. 4, Table 4) or capitellids (Fig. 6, Table 4)

during either year. Nereid abundance did not differ between the edge and interior in 2007.

However, in 2006 there were significantly more nereids in the interior than in the edge at the

ARC reference (F=7.501,36, p=0.0095) and Port created marshes (F=19.251,36, p<0.0001).

Based on Analysis of Similarity, benthic infaunal communities were most similar within

the same geographic area (Fig. 1, Fig. 8). As expected from frequent interaction between marsh

type (river vs. non-river) and other variables in the ANCOVAs, river sites clearly differed from

non-river sites in the MDS plot. Within the non-river marshes, the sites near Morehead City

(ARC and Port) were most similar while the Mason sites comprised a different grouping.

Patterns involving created vs. natural marshes and edge vs. interior habitats were not clear.

Infaunal biomass patterns generally paralleled abundance patterns among sites. There

were no significant differences between created and reference marshes or between edge and

interior habitats for annelids, crustaceans, mollusks, or total infaunal biomass.

Sediment characteristics

Percent sand and SOC were generally inversely related. The ARC created marsh was the

only site that contained significantly lower SOC (Fig. 9) and higher percent sand (F=27.351,12,

p=0.0002) than its analogous reference marsh. The Island 13 created marsh was the only site

that exhibited less sand than its reference marsh (F=6.981,12, p<0.0215). Within non-river sites,

SOC and percent sand were significantly related to created marsh age (SOC F=38.331,20,

p<0.0001, sand F=13.141,20, p=0.0017). River sites were excluded from these analyses since

24  

Transform: Square rootResemblance: S17 Bray Curtis similarity

SITEARCI13MASPORTPORT2

Arc C E

Arc C IArc R E

Arc R I

Isle13 C E

Isle13 C I

Isle13 R E

Isle13 R I

Mas C EMas C I

Mas R E

Mas R I

Port C EPort C I

Port R E

Port R I

Port2 C E

Port2 C I

Port2 R E

Port2 R I

2D Stress: 0.13

Figure 8. MDS Plot for infaunal abundance at all sites. Abundance is averaged between years for non-river sites. Circles indicate geographic regions (left=river sites, top right=Beaufort sites, bottom right=Wilmington non-river Mason sites). I13=Island 13, MAS=Mason, C=Created, R=Reference, E=Edge, and I=Interior.

25  

A (Non‐River Sites)  

B (River Sites) 

Figure 9. Mean percent sediment organic content. Each mean is based on 4 samples. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Three asterisks indicates created marsh is significantly different from its reference marsh (p<0.00001).

26  

interaction existed between marsh type and age. SOC and percent sand did not significantly

differ between edge and interior locations within non-river sites. However, within river sites,

SOC was significantly higher (F=39.891,24, p<0.0001) while percent sand was lower

(F=41.521,24, p<0.0001) in the interior. SOC was positively correlated with oligochaetes,

spionids, and nereids (Table 5). Percent sand was negatively correlated with capitellids and

nereids.

MOM was significantly higher in the Port 2 created marsh than in its reference marsh

(Fig. 10). There were no significant differences within any other created/reference marsh pair.

Within non-river sites, MOM was positively related to marsh age (F=24.991,20, p<0.0001). River

sites were not included in this analysis since marsh type interacted with both age and location.

MOM was greater in the marsh interior than the edge at all non-river sites except Mason Created,

although this was significant at only two of the sites (ARC created F=6.211,36, p=0.0175 , ARC

reference F=28.011,36, p<0.0001). All of the river sites exhibited greater MOM in the interior

than the edge (F=28.011,24, p<0.0001). MOM was positively correlated with oligochaetes and

spionids (Table 5).

Vegetation

All sites were primarily dominated by Spartina alterniflora, and this was the only species

present at most sites. Vegetation at the non-river sites was similar among years. The Mason

created marsh contained 85.9% Spartina alterniflora, 8.5% Salicornia virginica, 3.7% Limonium

carolinianum, and 1.9% Sueda maritima (averaged for both years). The Mason reference marsh

contained 50.9% Spartina alterniflora, 39.4% Salicornia virginica, 5.9% Sueda maritima , and

27  

Table 5.  Pearson correlation coefficients for 2007 data, N=80.  MOM=macro‐organic matter, SOC=sediment organic content.    ***p<0.0001, **p<0.001, *p<0.01, ns=nonsignificant p>0.05.     

   Total Abundance  Oligochaeta  Capitellidae  Spionidae  NereidaeSOC  ns  0.38**  ns  0.28     0.23 % Sand  ns  ns  ‐0.28  ns   ‐0.38** MOM  ns  0.38**  ns   0.34*  ns Spartina Height   0.3*            0.32*  ns        0.44***     ‐0.22 Spartina Density   0.27            0.27  ns   0.35*  ns      

28  

A (Non‐River Sites)  

B (River Sites) 

Figure 10. Mean macro-organic matter content. Each mean is based on 4 samples. Error bars represent one standard error. Created marshes are in order of increasing age. MC=Mason Created, MR=Mason Reference, AC=ARC Created, AR=ARC Reference, PC=Port Created, PR=Port Reference, I13C=Island 13 Created, I13R=Island 13 Reference, P2C=Port 2 Created, and P2R=Port 2 Reference. Asterisk indicates created marsh is significantly different from its reference marsh (p-value between 0.001 and 0.05).

*

29  

3.8% Limonium carolinianum (averaged for both years). Phragmites australis was present (21

stems total) at only one sampling location at the Island 13 reference marsh.

Spartina height was significantly lower at the youngest created marsh (Mason) than it

was at its analogous reference marsh in both years (Fig. 11). Spartina height was also

significantly lower at the oldest non-river created marsh (Port) in 2006 (Fig. 11). There were no

significant differences in Spartina height within any other created/reference site pairs during

either year. Spartina height was significantly related to created marsh age in both years (2006

F=28.211,20, p<0.0001; 2007 F=31.431,32, p<0.0001). Within the non-river sites, Spartina height

was significantly greater on the marsh edge than in the interior for both years (2006 F=24.191,36,

p<0.0001; 2007 F=18.291,36, p<0.0001). River sites showed no significant difference between

locations. Spartina height was positively correlated with total infaunal abundance, oligochaetes,

and spionids and negatively correlated with nereids (Table 5).

Spartina density was significantly higher at two of the created sites (Mason 2006

F=60.581,12 , p=<0.0001; Mason 2007 F=18.161,12 , p=0.0011; Port 2 F=21.561,12 , p=0.0006 )

than at their reference marshes. In 2006, interaction existed between created marsh age and

location. Age was significant (r2=0.58, F=13.531,10 , p=0.0043) in the edge but not in the interior.

In 2007, Spartina density was significantly related to marsh age (F=4.131,32, p=0.0432). The

marsh interior exhibited significantly greater Spartina density within the river sites (F=5.061,24,

p=0.0339). Within the non-river sites, location was not significant during either year. Spartina

density was positively correlated with total infaunal abundance, oligochaetes, and spionids

(Table 5).

30  

A (2006 Non‐River)

B (2007 Non‐River) 

C (2007 River) 

Figure 11. Spartina alterniflora height (mean of tallest stem of 4 m2 quadrats). Error bars represent one standard error. Created marshes are in order of increasing age. Site abbreviations same as previous figures. Asterisks indicate created marsh is significantly different from its reference marsh. One asterisk indicates p<0.05, and two asterisks indicate p<0.001.

*

31  

DISCUSSION

Some of the most common taxa found in salt marshes by this and other studies are

spionids, oligochaetes, and capitellids. Streblospio benedicti and Capitella capitata are often the

first to colonize created marshes while oligochaetes recruit later in marsh development (Moy and

Levin 1991; Levin et al. 1996; Alphin and Posey 2000; Craft and Sacco 2003). This idea is

supported by my results for capitellids in the non-river marshes but not for spionids. Although

infauna in this study were identified to family rather than species, Streblospio benedicti and

Capitella capitata were observed to be the most abundant species within the Spionidae and

Capitellidae, respectively, and my results for these families largely represent these common

species.

The major difference reported for infauna among natural and created Spartina

alterniflora marshes after several years is low numbers of oligochaetes in conjunction with low

sediment organic matter content in the created marshes (Moy and Levin 1991; Levin et al. 1996;

Posey et al. 1997; Broome et al. 2000; Craft 2000; Craft and Sacco 2003). My findings suggest

that this may be a variable pattern, since most of the created marshes in this study (including the

3-4 year old site) exhibited oligochaete abundances and SOC that were comparable to their

reference marshes. Although both oligochaete abundance and SOC increased with created marsh

age, the reference sites followed the same trajectories, indicating that other factors may be

responsible for these patterns. These apparent age patterns for oligochaete abundance and SOC

could be due to the near ocean, barrier island proximity of the non-river sites, resulting in higher

base sand contents in these marshes as compared to the river sites. Even within the non-river

marshes, infaunal communities were most similar within the same geographic area, potentially

32  

obscuring effects of created marsh age. This emphasizes the need for paired reference sites in

assessing patterns. All of the marshes in this study were located adjacent to potential infaunal

source habitats; this is particularly important for species with direct development such as

oligochaetes. The accessibility of the created marshes used in this study to infaunal source

populations could contribute to the accelerated development rates observed here as compared to

previous studies (Moy and Levin 1991; Minello and Zimmerman 1992; Sacco et al. 1994; Levin

et al. 1996; Posey et al. 1997; Craft et al. 1999; Alphin and Posey 2000; Streever 2000; Craft and

Sacco 2003).

The time required for a created marsh to achieve functional equivalence with a natural

marsh depends in large part on the characteristics of the reference marsh as well as the landscape

position of the created marsh. In order to ascertain effects of created marsh age amongst the

many other factors affecting benthic community development, researchers may be tempted to

group several created marshes of varying ages with different geomorphologies. However, the

dissimilarities in several attributes between the river and non-river sites in this study suggest that

these are very different habitats and that most comparisons between marshes with different

geomorphic positions should not be made even when the sites are in close proximity. Reference

marshes must be closely paired with created marshes with regard to geomorphology and

geographic location to account for the wide variability among natural marshes. The interaction

between marsh type and created marsh age for spionids, capitellids, and nereids observed in this

study suggests that marshes with different geomorphologies develop at different rates.

Compared to euhaline coastal marshes, mesohaline river marshes have more variable disturbance

regimes, lower faunal diversity, and differing salinity and flow characteristics (Mallin et al.

1998; Mallin et al. 2008). Even when created marshes are of similar geomorphological types,

33  

subtle differences between them could affect infaunal community development rates. For

example, the Port 2 created marsh in this study was the only river site where total abundance was

greater in the marsh interior than on the edge. A greater amount of fringing marsh edge that is

more exposed at this site compared to the other river sites may influence this observed habitat

difference.

My findings suggest that fauna of created marshes may resemble natural marshes early in

development, in as little as 3 years. The youngest created marsh in this study significantly

differed from its reference marsh only with regard to capitellid abundance and Spartina height

and density. Two of the most common capitellids (Capitella capitata and Mediomastus

ambiseta) are often referred to as opportunistic species (Grassle and Grassle 1974; Tsutsumi et

al. 1990). Although the presence and/or dominance of opportunistic species may be expected at

young created marshes, capitellids were also common at the older created marshes in this study

as well as the natural marshes. The high numbers of capitellids observed at both the created and

natural marshes in this study may reflect overall high regional abundance. Lower Spartina

height compared to the reference marsh could reflect lower primary productivity at the youngest

created marsh, but Spartina density was higher and no direct measurement of primary

productivity was performed.

Although, to my knowledge, no previous comparative studies of created marshes that

focus on fine-scale spatial distribution of benthic infauna have been conducted, there is evidence

that polychaetes may be more abundant at the marsh edge (Minello et al. 1994; Whaley and

Minello 2002) while oligochaetes are more common in the interior (Whaley and Minello 2002)

in the Gulf of Mexico. In this study, oligochaetes were consistently more abundant in the marsh

interior. However, the dominant polychaetes (Capitellidae, Spionidae, and Nereidae) were

34  

generally evenly distributed between the edge and interior, with some exceptions. Spionid

abundance appeared to be greater in the edge than in the interior at several sites, and this

difference was significant at a few sites. Nereids were more common in the interior in some

cases. The persistent abundance of polychaetes in the marsh interior may contribute to

accelerated rates of created marsh similarity to adjacent reference areas. Comparisons between

salt marshes in the Gulf of Mexico and those on the U.S. Atlantic coast are limited by geographic

differences such as lunar tidal range and spatial extent of marsh habitat (Minello and Webb

1997). Whaley and Minello (2002) proposed that predation by nekton controls infaunal densities

at the marsh edge. Although my sampling was conducted during only one season (summer)

when recruitment had already occurred and predation was likely high, predation did not appear

to relate to differential patterns among habitats in this study. Although predation may have

affected the abundance of surface-dwelling fauna, it should not have strongly affected

oligochaete abundance. Predation was potentially responsible for lower nereid abundance at the

marsh edge, but this occurred at two sites during one year only.

Greater abundance of oligochaetes in the marsh interior is an important finding because

although other studies have found more oligochaetes in natural marshes than created marshes

(Moy and Levin 1991; Levin et al. 1996; Posey et al. 1997; Broome et al. 2000; Craft and Sacco

2003), they did not distinguish edge samples from interior samples. Since oligochaetes appear to

be more common in natural marshes than created marshes, and they are often more abundant in

the interior (Whaley and Minello 2002), then the spatial distribution of samples may greatly

affect the results of created marsh comparisons. Because subsurface infauna such as

oligochaetes are usually not strongly affected by epibenthic predation, other factors must be

involved. Several studies have suggested that high SOC is linked to oligochaete abundance

35  

(Moy and Levin 1991; Levin et al. 1996; Posey et al. 1997; Broome et al. 2000; Whaley and

Minello 2002; Craft and Sacco 2003). Previous studies have also proposed that oligochaete

abundance is linked to MOM (Craft 2000; Craft & Sacco 2003). In this study, both SOC and

MOM were positively correlated with oligochaete abundance. Although these physical variables

may explain greater oligochaete abundance in the interior within the river marshes, they did not

significantly differ between the edge and interior in most of the non-river marshes. However,

MOM appeared to be higher in the interior at most non-river sites, and this difference was

significant at two sites. SOC and MOM may also contribute to higher oligochaete abundance in

the river marshes than in the non-river sites, although salinity also differs between the river and

non-river sites. Greater oligochaete abundance in the river marshes along with greater SOC and

MOM may contribute to the correlation between oligochaete abundance and these sediment

characteristics. SOC and MOM were both significantly higher in the interior at the river sites.

Although SOC and MOM were expected to be positively correlated with oligochaete abundance,

the positive correlation observed between these sediment characteristics and spionid and nereid

abundances is surprising. These correlations were likely influenced by the low abundances of

these polychaetes at the sandy Mason created and reference marshes, which were low in SOC

and MOM. Also, spionids were common in the river marshes, which exhibited high SOC and

MOM.

Spartina height was the variable other than oligochaete abundance measured in this study

that clearly varied between the edge and interior in the non-river marshes. Spartina height and

density were both positively correlated with oligochaete abundance, even though Spartina height

was greater on the edge while oligochaetes were more abundant in the interior. The correlation

between Spartina height and oligochaete abundance in this study was likely influenced by

36  

greater Spartina height and oligochaete abundance in the river marshes compared to the non-

river sites. When river sites were excluded from the correlation analysis, there were no

significant correlations between oligochaete abundance and Spartina height or density. Spartina

density could indicate availability of detrital food resources for oligochaetes, although above-

ground biomass was not measured in this study. MOM, which represents food for oligochaetes,

was consistently higher in the interior at most of the sites. However, this difference was not

significant at many of the non-river sites.

Elevation generally increases from the marsh edge to the interior in most coastal salt

marshes, and this pattern was observed in this study. However, differences between relative

elevations for edge and interior sampling locations in this study (mean=17.2 cm) were small and

within the range that is considered similar for many benthic infauna (Minello and Zimmerman

1992; Minello et al. 1994; Posey et al. 2003), and this was not likely to significantly impact

infaunal distribution patterns. All of the interior locations in this study were dominated by

Spartina alterniflora and located below the normal high water line. Furthermore, since elevation

varied within and among sites, the consistent patterns of more oligochaetes in the interior and

even distribution of polychaetes observed across sites indicate that elevation likely was not a

major factor influencing my results.

My overall hypothesis that the edge of created marshes matures more quickly than the

interior is not supported by the data. Although oligochaetes were generally more abundant in the

interior, the most common surface-feeding polychaetes (spionids and nereids) were generally

evenly distributed between locations. Spionids appeared to be more abundant on the edge in

many cases, but the difference was not significant. Nereids appeared to be more abundant in the

interior in several cases, and this difference was significant at some sites. The youngest created

37  

marsh contained oligochaete abundances that were comparable to its natural reference marsh

after only 3 years. SOC and MOM, which should be related to oligochaete abundance, only

differed between edge and interior locations in the river sites generally. These results indicate

that created marshes may become similar to natural marshes early in development, in the interior

as well as on the edge. Oligochaete recruitment may be more dependent on other factors such as

accesibility by source populations rather than high levels of organic matter.

Relationships between created marsh age and the biological and physical variables

measured in this study were overwhelmed by effects of geographic location. Whenever age was

significant, the reference marshes showed the same pattern, casting doubt on whether the

observed differences were actually due to created marsh age. As evident in the MDS plots,

infaunal communities were most similar between sites that were located in the same area. These

groupings are likely due in part to the differences between marsh types (river vs. non-river).

However, sites still clustered by region within non-river sites. These patterns could be

influenced by the quality of adjacent source populations.

Although some differences were observed, all of the created marshes in this study were

generally similar to their natural reference marshes with regard to infauna, vegetation, and

sediment characteristics. Differences among the created sites appeared to be controlled by

multiple factors overriding the effect of marsh age. While the vegetation and sediment

characteristics sampled affected infaunal community abundance and composition, uncontrolled

factors such as geographic location and geomorphology also appeared to influence invertebrate

distribution among sites. Oligochaetes were more abundant in the marsh interior as compared to

the edge, while polychaetes were generally evenly distributed. Nereids appeared to be more

abundant in the interior at several sites, and this difference was significant at two sites. These

38  

spatial distribution patterns were not readily explained by the physical parameters sampled in

this study.

Although the data do not support the hypothesis that the edge of created marshes matures

more quickly than the interior, future research on marsh creation/restoration should still address

spatial distribution in the marsh. Even if the edge and interior develop at similar rates, there are

differences in faunal characteristics. The edge and interior of most marshes are likely to differ

with regard to physical characteristics and potential forcing factors on infaunal populations (e.g.

predation, submergence time, and sediment inputs). The degree to which the edge and interior

differ biologically and physically varies by site. Many created marsh studies do not specify

whether samples were collected near open water, near creek channels, or deep in the marsh

interior. If one study sampled near the edge while another sampled in the interior, then the

results of those studies would likely not be comparable. Studies which concentrate their efforts

on one microhabitat may give a biased view of the overall marsh community. The complexity of

infaunal community composition and distribution patterns requires that future research consider

many variables, including location within the marsh. In order to gain a more complete

understanding of marsh succession amidst natural variability, multiple sites will need to be

compared over time. Comparison between studies would be greatly aided by positioning

sampling locations in transects that cover different zones of the marsh so that the same habitat

type could be compared across multiple studies at several sites during different years.

LITERATURE CITED

Alphin, T.D. and Posey, M.H. 2000. Long-term trends in vegetation dominance and

infaunal community composition in created marshes. Wetlands Ecology and

Management 8:317-325.

Baltz, D.M., Rakocinski, C., and Fleeger, J.W. 1993. Microhabitat use by marsh-edge

fishes in a Louisiana estuary. Experimental Biology of Fishes 36:109-126.

Boesch, D.F. and Turner, R.E. 1984. Dependence of fishery species on salt marshes:

the role of food and refuge. Estuaries 7(4A):460-468.

Broome, S.W., Seneca, E.D., and Woodhouse, W.W. 1988. Tidal salt marsh restoration.

Aquatic Botany 32:1-22.

Broome, S.W., Craft, C.B., and Toomey, W.A. Jr. 2000. Soil organic matter effects on

infaunal community structure in restored and created tidal marshes. In Concepts

and Controversies in Tidal Marsh Ecology, eds. M.P. Weinstein and D.A. Kreeger, 737-

747. The Netherlands: Kluwer Academic Publishers.

Brusca, R.C. and Brusca, G.J. 2003. Invertebrates, 2nd Ed. Sunderland, MA: Sinauer Associates,

Inc.

Craft, C., Reader, J., Sacco, J., and Broome, S.W. 1999. Twenty-five years of ecosystem

development of constructed Spartina alterniflora (Loisel) marshes. Ecological

Applications 9(4):1405-1419.

Craft, C. 2000. Co-development of wetland soils and benthic invertebrate communities

following salt marsh creation. Wetlands Ecology and Management 8:197-207.

Craft, C. and Sacco, J. 2003. Long-term succession of benthic infauna communities on

constructed Spartina alterniflora marshes. Marine Ecology Progress Series 257:

45-58.

40  

Currin, C.A., Joye, S.B., and Paerl, H.W. 1996. Diel rates of N2-fixation and denitrification in a

transplanted Spartina alterniflora marsh: implications for N-flux dynamics. Estuarine,

Coastal, and Shelf Science 42:597-616.

Grassle, J.F. and Grassle, J.P. 1974. Opportunistic life histories and genetic systems in

marine benthic polychaetes. Journal of Marine Research 32:253-284.

Howes, B.L., Dacey, J.W.H., and Goehringer, D.D. 1986. Factors controlling the growth

form of Spartina alterniflora: feedbacks between above-ground production,

sediment oxidation, nitrogen and salinity. Journal of Ecology 74:881-898.

LaSalle, M.W., Landin, M.C., and Sims, J.G. 1991. Evaluation of the flora and fauna of

a Spartina alterniflora marsh established on dredged material in Winyah Bay,

South Carolina. Wetlands 11(2):191-208.

Levin, L.A., Talley, D. and Thayer, G. 1996. Succession of macrobenthos in a created

salt marsh. Marine Ecology Progress Series 141:67-82.

Mallin, M.A., L.B. Cahoon, J.J. Manock, J.F. Merritt, M.H. Posey, R.K. Sizemore, W.D. Webster and T.D. Alphin. 1998. A Four-Year Environmental Analysis of New Hanover County Tidal Creeks, 1993-1997. CMSR Report No. 98-01. Center for Marine Science Research, University of North Carolina at Wilmington, Wilmington, N.C. Mallin, M.A., M.R. McIver and J.F. Merritt. 2008. Environmental Assessment of the Lower Cape Fear River System, 2007. CMS Report No. 08-03, Center for Marine Science, University of North Carolina at Wilmington, Wilmington, N.C. McCann, L.D. and Levin, L.A. 1989. Oligochaete influence on settlement, growth and

reproduction in a surface-deposit-feeding polychaete. Journal of Experimental

Marine Biology and Ecology 131:233-253.

McIvor, C.C. and Odum, W.E. 1988. Food, predation risk, and microhabitat selection in a marsh

fish assemblage. Ecology 69(5):1341-1351.

41  

Minello, T.J. and Webb, J.W. 1997. Use of natural and created Spartina alterniflora salt

marshes by fishery species and other aquatic fauna in Galveston Bay, Texas,

USA. Marine Ecology Progress Series 151:165-179.

Minello, T.J. and Zimmerman, R.J. 1992. Utilization of natural and transplanted Texas

salt marshes by fish and decapod crustaceans. Marine Ecology Progress Series

90:273-285.

Minello, T.J., Zimmerman, R.J., and Medina, R. 1994. The importance of edge for

natant macrofauna in a created salt marsh. Wetlands 14(3):184-198.

Moy, L.D. and Levin, L.A. 1991. Are Spartina marshes a replaceable resource? A

functional approach to evaluation of marsh creation efforts. Estuaries 14:1-16.

Peterson, G.W. and Turner, R.E. 1994. The value of salt marsh edge vs. interior as a

habitat for fish and decapod crustaceans in a Louisiana tidal marsh. Estuaries

17:235-262.

Piehler, M.F., Currin, C.A., Cassanova, R., and Paerl, H.W. 1998. Development and N2-fixing

activity of the benthic microbial community in transplanted Spartina alterniflora marshes

in North Carolina. Restoration Ecology 6(3):290-296.

Posey, M. H., Alphin, T.D., and Powell, C.M. 1997. Plant and infaunal communities

associated with a created marsh. Estuaries 20:42-47.

Posey, M.H., Alphin, T.D., Meyer, D.L., and Johnson, J.M. 2003. Benthic communities of

common reed Phragmites australis and marsh cordgrass Spartina alterniflora marshes in

Chesapeake Bay. Marine Ecology Progress Series 261:51-61.

Sacco, J.N., Seneca, E.D., and Wentworth, T.R. 1994. nfaunal community development

of artificially established salt marshes in North Carolina. Estuaries 17:489-500.

Streever, W.J. 2000. Spartina alterniflora marshes on dredged material: a critical review

of the ongoing debate over success. Wetlands Ecology and Management 8:295-316.

42  

Tsutsumi, H., Fukunaga, S., Fujita, N., and Sumida, M. 1990. Relationship between

growth of Capitella sp. and organic enrichment of the sediment. Marine Ecology

Progress Series 53:157-162.

Whaley, S.D. and Minello, T.J. 2002. The distribution of benthic infauna of a Texas salt

marsh in relation to the marsh edge. Wetlands 22(4):753-766.