Sediment Organic Carbon Pools and Sources in a Recently Constructed Mangrove and Seagrass Ecosystem

MISSING IMAGE

Material Information

Title:
Sediment Organic Carbon Pools and Sources in a Recently Constructed Mangrove and Seagrass Ecosystem
Physical Description:
1 online resource (168 p.)
Language:
english
Creator:
Pries, Caitlin E. Hicks
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Soil and Water Science
Committee Chair:
Reddy, Konda R.
Committee Members:
Sickman, James
Osborne, Todd Z.
Schuur, Edward A.

Subjects

Subjects / Keywords:
accumulation, carbon, constructed, decomposition, ecosystem, isotopes, lability, mangrove, pools, seagrass, sediments, sink, stable
Soil and Water Science -- Dissertations, Academic -- UF
Genre:
Soil and Water Science thesis, M.S.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
Coastal ecosystems are significant natural carbon sinks. If constructed coastal ecosystems can obtain the same carbon sink capacity as their natural counterparts, then construction and restoration of these systems has the potential to become a tool for reducing atmospheric CO2. In this study, sediment organic carbon (OC) of a recently constructed mangrove and seagrass system in the Indian River Lagoon, Florida was compared with sediment OC of nearby mature, reference systems. Total OC, extractable OC, and microbial biomass C pools were measured to compare C storage. Organic C lability in the constructed and reference sites was also measured. The main sediment OC sources were determined using 13C isotopes and C:N ratios and were compared among systems. Organic C pools were generally larger in sediments of reference systems than in sediments of the constructed systems, but differences in pool sizes were much greater between the constructed and reference mangrove systems. Organic C lability was greater in the constructed systems indicating their sediments could not store OC for as long as the references. Seston was a major source of sediment OC in all systems. Other main sources of OC were higher plant-derived in constructed and reference mangrove and reference seagrass sediments, but were algal-derived in constructed seagrass sediments. After one year, the C sink capacity of the constructed systems is less than the capacity of the reference systems, but the constructed seagrass system is functioning more like its reference than the constructed mangrove system. In the long term, however, the potential C sink capacity of the constructed mangrove system is greater.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2007.
Local:
Adviser: Reddy, Konda R.
Statement of Responsibility:
by Caitlin E Hicks.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2007
System ID:
UFE0021541:00001


This item has the following downloads:


Full Text





situ acidification techniques and subsequent analysis with an elemental analyzer as the standard

method for OC measurements (Nieuwenhuize et al. 1994).

Lastly, research is needed that addresses the permanency of C storage in constructed and

restored coastal systems. Disturbances like changes in nutrient loading, invasive species, and

hurricanes can affect C storage. A nutrient loading experiment in a North Carolina salt marsh

increased microbial respiration and caused a subsequent net loss of SOC over 12 years (Morris

and Bradley 1999). Alternatively, the spread of a native, yet invasive, grass species in a natural

coastal marsh in France was found to increase SOC storage (Valery et al. 2004). These studies

occurred in natural marshes, and studies that examine disturbance effect on SOC in constructed

coastal systems are needed because constructed systems may be less resilient than natural

systems. In order for constructed coastal systems to become a viable option for C offsets, these

effects need to be understood and quantified.

More thorough studies are needed on the C sink capabilities of restored and constructed

coastal ecosystems. Thus far, the vast maj ority of studies have been carried out in salt marshes.

Studies are needed in coastal ecosystems like mangrove forests and seagrass beds whose

destruction is also routinely mitigated with restoration and construction. If researchers can prove

that constructed systems are effective C sinks by demonstrating that they not only follow

traj ectories of increasing SOC, but also have OC accumulation rates similar to natural

ecosystems, then constructing coastal ecosystems may become an accepted way to offset CO2

emissions. Institutions such as Climate Neutral (www.climateneutral .org) and the Chicago

Climate Exchange (www. chicagoclimatex.com) could use coastal ecosystem construction to

offset emissions like they now do with certain forestry and agricultural practices.









and that more mature forests, with their higher C:N ratios and ligno-cellulosic debris, were

dominated by mangrove-derived OM. They also found that the upper sediments of the younger

forests and the deeper sediments of the older forests had a lot of OLC, indicating a trapping of

allochthonous OM from the river. With this method, known proportions of various OM

components can be measured directly, in contrast to isotopic methods. However, OM

components such as TLC and GP cannot be directly attributed to any one species of primary

producer, but rather to broad classes of producers. This study did not take seagrasses into

consideration, which may have similar-looking partially decomposed ligno-cellulosic tissues as

mangroves, making it hard to differentiate between those two sources using this method.

Nuclear Magnetic Resonance Spectroscopy

Nuclear magnetic resonance spectroscopy (NMR), specifically 13C-NMR, is another

method that can be used to identify SOC sources. However, no known studies document this

method in seagrass, mangrove, or salt marsh sediments, and therefore an estuarine study is used

to illustrate this method. In NMR spectroscopy, a sediment sample is subjected to a magnetic

Hield which causes the nuclei in the 13C atoms to process as a gyroscope does (Stevenson 1994).

A second alternating magnetic Hield is then added, and when the frequency of that second

magnetic field matches the frequency of the nuclei's precession, the nuclei of the atom then

resonate causing a voltage change that is amplified and recorded. A spectrum is produced from

this resonance signal. The nuclei resonate at different frequencies depending on their chemical

environment. Each sample resonates at several frequencies, and from the different resonance

signals, spectra with several distinct peaks are produced. Spectra are plotted using the chemical

shift--the difference between resonance frequencies of samples and the resonance frequency of a

standard, tetramethylsilane (TMS) solution. This chemical shift calculation is analogous to the

calculation of 613C ValUeS based on how much the samples' values differ from the PDB standard.









allochthonous terrestrial plant matter and autochthonous marine plankton. Such simple systems

may be encountered in estuaries that lack submerged aquatic vegetation (Golding et al. 2004).

Sometimes researchers simply group the sources of OC into two end-member groups. For

example, in salt marshes the SOC inputs from Spartina can be distinguished from the inputs

from all other sources because they differ in their 613C ValUeS (Middelburg et al. 1997). These

two end-member models are often useful in estimating the maj or categories of OC sources (i.e.

whether allochthonous or autochthonous), but they cannot fully partition the individual SOC

sources.

The variety of methods used to determine OC inputs range widely in terms of time and

equipment involved. Methods can be as simple as comparing C:N ratios of possible sources with

sediment C:N ratios or as complicated as searching for a biomarker and then isolating and

concentrating that specific compound for isotopic analysis. The most widely used method

involves stable isotopes--either comparing bulk composition of 613C in pOssible sources and

sediments or comparing composition of 613C in lipids found in possible sources and sediments.

Lipids and other biomarkers can also be used singly to determine sources. Other methods, which

include petrographic analysis and nuclear magnetic resonance spectroscopy (NMR), involve

comparing relative amounts of different OC structures in the soil.

Stable Isotopes

In salt marshes, mangrove forests, and seagrass beds many researchers have tried to

determine SOC sources by matching the 613C iSotopic signatures of the bulk sediment to the 613C

isotopic signatures of the sources via strait comparison of the numbers, mixing models, or

diagrams (Table 2-5). In order for stable isotopes to elucidate OC sources, the sources need to

have consistently distinct isotopic signatures (Papadimitriou et al. 2005). The various primary

producers in coastal systems develop distinct isotopic signatures through their discrimination

















































































I I


1200


1000


-
-$--
-0-
-


Seagrass floc
Mangrove algal mat


800-


600-


400-


200-


0
30000


I I I I


25000-


20000-


15000-


10000-


5000-


I I I I


Feb06 May06 Jul06


Nov06


Figure 3-5. The changes in organic carbon parameters over the first year after construction in SL
15 seagrass and mangrove surface layers. The symbols are the mean values for each

sampling date (n=4) and error bars are + SE.



101









aided by water movement into interstitial spaces, once compaction-causing construction ceased

and tides could access the site. The seagrass section of SL 15 was completed a month before the

rest of SL 15. Seagrass sediments therefore decompressed earlier and may have experienced a

similar bulk density decrease before sampling began.

OC parameters in SL 15 sediments did not follow a trajectory, although OC pools in SL

15 seagrass sediments seemed to follow a pattern (Fig. 3-4). External, seasonal factors, not

ecosystem development, were likely the force driving these patterns. A review of physical and

chemical water column data in IRL from November 2005 to November 2006 revealed potential

correlations that could explain the pattern (SFWMD 2007; station IRL 36). Lows in OC

parameters corresponded with lows in salinity, highs in total Kj eldahl nitrogen, and the lowest

(February) and highest (July) water temperatures of the year. Nitrogen probably did not cause

these trends because N levels in the IRL are not high enough to be toxic to bacteria, but

temperature or salinity may have. If the overlying water affected OC parameters in seagrass

sediments, it explains why mangrove sediments, which are only in contact with water at high

tide, experienced the pattern to a much lesser extent.

In both SL 15 surface layers, TOC and MBC followed a traj ectory where they increased

over time (Fig. 3-5). As a surface layer, seagrass floc is more likely to respond to water column

changes than sediments. Floc TOC and MBC, however, followed a different pattern than

seagrass sediments and IRL salinity and temperature. Floc OC pool increases match increases in

IRL total suspended solids from February to November 2006 (SFWMD 2007; station IRL 36).

Since the floc is mostly water (95%), it is likely that its solids are correlated to water column

solids, which include OC substrate and microbes. Algal mat MBC increases are likely due to the

algal mat' s maturation as it became larger and denser throughout the year (personal observation).











2500


(mg kg ')
S0-5 cm
5-10 cm


Storage (g
a


2000


H 1500






3 000


2500



100

3 000


2500




0)


Reference SL 15 Reference SL 15


Figure 3-8.Comparisons between microbial biomass carbon (MBC) in reference and SL 15
mangrove (top) and seagrass (bottom) sediments. The bars are mean MBC averaged
over month (July and November 2006) for each depth of sediment (n=12 for SL 15
and n=9 for reference). Error bars are + SE. Depths in the seagrass systems are as
follows: 1= SL 15 accreted and reference 0-5, 2= SL 15 0-5 and reference 5-10, 3=
SL 15 5-10 and reference 10-15. An asterisk indicates a significant site effect (Table
3-5). Capital letters are results of a Tukey test performed after a significant site x
depth interaction, and lowercase letters are results of a Tukey performed after an
insignificant site x depth interaction, but a significant one way ANOVA. Bars that
share letters are not significantly different.










measuring functional trajectories, OC liability, and OC accumulation rates. If constructed

systems are similar to natural systems, then constructing coastal ecosystems may become an

accepted way to offset CO2 emiSsions, which would encourage more restoration.









5.8 to 9.3 for seston (Gonnocea et al. 2004; Zhou et al. 2006), and of 7 to 30 for macroalgae

(Kristensen 1994; Thimdee et al. 2003).

813C Of plants vary within different tissues (Vizzini et al. 2003; Papadimitriou et al. 2005;),

within a single species (Hemminga and Mateo 1996), across sites (Kennedy et al. 2004), seasons

(Vizzini et al. 2003), and years (Anderson and Fourqurean 2003; Fourqurean et al. 2005).

Variations are most pronounced in seagrasses (Thimdee et al. 2003) and macroalgae. In

submerged vegetation variation is due to the relative uses of dissolved CO2 and bicarbonate, the

source of inorganic C in the water, temperature, irradiance, and subsequent photosynthesis rates

(Lin et al. 1991; Hemminga and Mateo 1996). Seston 613C can also vary temporally, spatially,

and between ebb and flood tides (Hemminga et al. 1994). These variations in source 613C make

it necessary to measure all potential sources' 613C for each study area, instead of relying on

literature values, and ideally, measure significant sources across tissues, sites, and seasons. 613C

variations within individual sources and plant groups in this study were generally smaller than

differences among main sources, so the variations most likely do not affect our source

determinations. Furthermore, where 613C did overlap among main sources their C:N ratios set

them apart, as with seston and mangroves, or they were not both end members for the same

ternary diagram.

There is some concern about whether 613C Of plant tissues changes during diagenesis

because large changes in 613C COuld lead to misleading source determinations. Studies that

measured fresh and senescent mangrove leaves and seagrass found small (generally >1%o)

differences (Thimdee et al. 2003; Gonnocea et al. 2004). Decomposition studies found

significant but minor (0.55 to 2%o) changes in seagrass, mangrove, and macroalgae 613C during

diagenesis (Fenton and Ritz 1988; Fourqurean and Schrlau 2003), but others found no significant











Table 2-3. Continued.


C accumulation
(g: C m yr )
99

30-36

146

107

34

15

58.9

21.3

125

115

27

28-32

99

159

18

105-115

167

207


Data
Source"
15

15

17

17

15

15

17

17

18

18

15

15

18

18

15

15

3

3


Location

"DOT," North Carolina



Jacob's Creek,
North Carolina


"Marine Lab "
North Carolmna


Oregon Inlet,
North Carolina


Pine Knoll Shores,
North Carolina



"Port," North Carolina



Snow's cut,
North Carolmna


Swansboro
North Carolmna



Aransas NWR, Texas
San Bernard NWR,
Texas
United States Average


Site

1 yr-old constructed

Natural reference

Irregularly-flooded
streamside
Irregularly-flooded
backmarsh

26-vr-old constructed

Natural reference

Regularly-flooded
streamside
Regularly-flooded
backmarsh

21-vr-old constructed

Natural reference

8-vr-old constructed

Natural reference

25-vr-old constructed

Natural reference

11-vr-old constructed

Natural reference


Time Scale
1 Year

Decadal

Decadal

Decadal

26 Year

Decadal

Decadal

Decadal

11 Year

11 Year

8 Year

Decadal

11 Year

11 Year

11 Year

Decadal

Decadal

Decadal


Method
aoc /Time
'37Cs and
210p poie
'3bC profiles

'3Cs profiles


Cs0p profiles


13bC profiles

13Cs profiles


aoc /Time

aoc /Time

aoc /Time
137Cs and
210b profiles
aoc /Time

aoc /Time

aoc /Time
'37Cs and
210b profiles
13Cs profiles

'3Cs profiles


Compiled 19


Seagrass Beds

Aburtsub Bay1.2-1.55 NA Modeled 20
Japan
Celestun Lagoon, 40 Century "9''b profiles 7
Mexico
Terminos Lagoon, 53-65 Century "9rb profiles 7
Mexico

Cala Culip, Spain 19-191 600 Year Shipwreck 21

Fanals Point, Spain 182 Annual Sediment trap 22

"l, Brunskill et al. 2002; 2, Alongi et al. 2005; 3, Callaway et al. 1997: 4, Cahoon and Lynch 1997: 5, Ong 1993: 6,
Alongi et al. 2004; 7, Gonneea et al. 2004; 8, Alongi et al. 2001; 9, Cahoon and Turner 1989: 10, Cahoon 1994: 11,
Connor et al. 2001; 12, Choi and Wang 2004; 13, Hussein et al. 2004; 14, Middelburg et al. 1997: 15, Craft et al.









CHAPTER 2
LITERATURE REVIEW

Introduction

In this literature review, rates of organic carbon (OC) accumulation are compiled and

compared for three coastal ecosystems--salt marshes, mangrove forests, and seagrass beds.

Studies comparing sediment organic carbon (SOC) pools in restored or constructed salt marshes

to SOC pools in natural salt marshes are then examined. This section does not discuss mangrove

forests or seagrass beds because the literature on the functioning of restored or constructed

coastal ecosystems is currently limited to salt marshes. Third, methods for determination of SOC

sources are discussed for the three coastal ecosystems. These coastal ecosystems are dominated

by vascular, halophytic macrophytes, with mangroves dominated by trees and salt marshes and

seagrass beds dominated by grasses and other herbaceous species. Sediments in these systems

are C sinks due to their high net primary production, trapping of material from the water column,

and 02 limited conditions.

These systems are globally distributed. Salt marshes and mangroves occupy non-rocky,

sedimentary-driven intertidal zones of the world. Salt marshes predominate in temperate

climates, while mangroves predominate in subtropical and tropical climates. Salt marshes are

generally replaced by mangroves at a latitude of 250 N or S (Mitsch and Gosselink 2000).

Seagrass systems are subtidal and are found from tropical through temperate climates where

needs such as low light attenuation in the water column are met (Hemminga and Duarte 2000).

Seagrasses are often found adj acent to their intertidal counterparts--salt marshes or mangroves.

Multiple estimates of global area covered by each system differ, but in each system the estimates

are within the same order of magnitude. According to the average of the estimates, mangrove

forests cover 220,000 km2, Salt marshes cover 350,000 km2, and seagrass beds cover 450,000










Table 3-1. Mean (+ SE) bulk density, % shell pieces, pH, and Eh redoxx potential) of the sediments according to depth and site. The
bulk density and % shell data were averaged over the July and November 2006 sampling periods (n=24 for SL 15 and n=18
for references). The pH and Eh data were measured in September 2006 (n=3).


Bulk density
(g cnf3)


Shells >1mm pH
(%)
Reference SL 15 Reference SL 15


Eh
(mV)
Reference SL 15 Reference
NA NA NA
04) 7.6 (0.07) -71 (90) -130 (40)
03) 7.5 (0.03) -5.7 (100) -160 (8)


System Depth


SL 15
Mangrove Algal mat/ Litter 1.03 (0.2)
0-5 cm 1.62 (0.03)
5-10 cm 1.48 (0.03)


0.41 (0.1)
0.95 (0.04)
0.93 (0.05)


0
24 (2)
30 (2)


0
21 (5)
14 (2)


Seagrass Floc
Accreted
0-5 cm
5-10 cm
10-15 cm


0.32 (0.04) 0.54 (0.09) 0
0.91 (0.05) 6.0 (1)
1.51 (0.04) 0.89 (0.04) 20 (3)
1.48 (0.04) 1.03 (0.03) 19 (5)
1.20 (0.02)


0

0.33 (0.1)
0.49 (0.2)
0.48 (0.1)


NA NA
8.2 (0.04)
8.3 (0.04) 8.0 (0.18)
8.2 (0.03) 8.3 (0.07)
8.3 (0.14)


NA
-98 (50)
-230 (4)
-180 (60)


NA

-150 (30)
-240 (8)
-320 (40)













~trrltil fl~,~


Figure 3-2. Core from SL 15 seagrass system illustrating the surface layer (floc) and different
sediment depths accretedd layer, 0-5 cm, 5-10 cm). Note the difference in color
between the accreted layer and 0-5 cm depth.










Table 2-4. Continued
Age Constructed OC Reference OC Depth sampled
Location Site earsr) (units) (units) (cm) Method" Sourceb
Georgia Sappelo Island 42 1264 g C m 1372 g C m 0-10 1 5
North Carolina Pamlico River Estuary 5 886 kmol C ha 10270 kmol C ha' 0-30 3 6
15 1866

Virginia Gloucester Point 5 95 g C m 129 163 g C m 0-2 1 7
12 120
5 50 146 174 14-16
12 53

North Carolina2 "DOT" 1 400 g C m 3800 g C m 0-30 1 8
Consultant 3 600 4600
Port 8 900 2000
Swansboro 11 1000 4600
Dill's Creek 13 1800 4900
Pine Knoll 24 1200 1000
Marine Lab 26 2900 5100
Snow's Cut 28 2900 10000
al, loss-on-ignition: 2, Walkley-Black oxidation: 3, CHN analyzer. bl, Simenstad and Thom 1996: 2, Morgan and Short 2002; 3, Cammen 1975; 4, Zedler and
Calloway 1999; 5, Craft 2001; 6, Craft et al. 2002; 7, Havens et al. 2002; 8, Craft et al. 2003
1Signifies study measured organic matter (OM) only, not organic carbon (OC). 2Signifies study did not measure the same wetland overtime but instead used a
space-for-time substitution.









accretion rates are less straightforward measurements than SOC pool measurements.

Radioisotope dating of cores using either 210Pb and 14C activity or 6137CS and 14C peaks from

nuclear bomb fallout were the most commonly used methods to date sediments (e.g. Callaway et

al. 1997; Connor et al. 2001; Choi and Wang 2004; Hussein et al. 2004; Alongi et al. 2005).

Other methods of dating sediments were Romero' s (1994) use of a shipwreck whose date was

known and that had been buried by seagrasses over hundreds of years and Chmura et al.'s (2001)

use of pollen stratigraphy. Short term (1-3 years) accretion rates were measured with feldspar

markers ( Cahoon and Turner 1989; Cahoon 1994; Cahoon and Lynch 1997) or sediment traps

(Gacia et al. 2002). The third step is to calculate C accumulation rates. The amount of OC in a

unit of sediment is divided by the age of that sediment unit, or OC in a unit of sediment is

multiplied by the rate at which that sediment unit accreted. Sediment ages in restored or

constructed systems do not have to be determined because the site ages are known. C

accumulation rates can be calculated by the difference between SOC content at the beginning of

the restoration process and SOC content at subsequent points after the initial restoration, divided

by site's age (Cammen 1975; Craft et al. 2003).

Calculated rates of C accumulation may be dependent on time scale, which is dependent on

the method used. With a half-life of 5730 years, 14C methods are suitable for measuring rates

over many millennia, while with a half-life of 22.3 years, 210Pb methods are suitable for

measuring rates over a century (Bierman et al. 1998). Bomb fallout methods using 8137CS and

14C peaks can only measure rates over the last 40 years as the peaks generally occur in 1963.

Many of the highest rates of C accumulation were measured using the feldspar marker technique,

which measures C accumulation rates over a year or two. These rates may be high because

surface pools of SOC are relatively labile compared to deeper pools of SOC. Much of the










FISCHLER, K. C. 2006. Observations and Characterizations of Subaqueous Soils and
Seagrasses in a Recently Constructed Habitat in the Indian River Lagoon, Florida.
University of Florida.

FOURQUREAN, J. W., AND J. E. SCHRLAU. 2003. Changes in nutrient content and stable
isotope ratios of C and N during decomposition of seagrasses and mangrove leaves along a
nutrient availability gradient in Florida Bay, USA. Chemistry and Ecology 19: 373-390.

------, S. P. ESCORCIA, W. T. ANDERSON, AND J. C. ZIEMAN. 2005. Spatial and seasonal
variability in elemental content, delta C-13, and delta N-15 of Thala~ssia testudinum from
South Florida and its implications for ecosystem studies. Estuaries 28: 447-461.

GACIA, E., C. M. DUARTE, AND J. J. MIDDELBURG. 2002. Carbon and nutrient deposition
in a Mediterranean seagrass (Posidonia oceanica) meadow. Limnol. Oceanogr. 47: 23-32.

GOLDING, C. J., R. J. SMERNIK, AND G. F. BIRCH. 2004. Characterisation of sedimentary
organic matter from three south-eastern Australian estuaries using solid-state C-13-NMR
techniques. Marine and Freshwater Research 55: 285-293.

GONI, M. A., M. J. TEIXEIRA, AND D. W. PERKEY. 2003. Sources and distribution of
organic matter in a river-dominated estuary (Winyah Bay, SC, USA). Estuar. Coast. Shelf
Sci. 57: 1023-1048.

GONNEEA, M. E., A. PAYTAN, AND J. A. HERRERA-SILVEIRA. 2004. Tracing organic
matter sources and carbon burial in mangrove sediments over the past 160 years. Estuar.
Coast. Shelf Sci. 61: 211-227.

HAINES, E. B. 1976. Stable carbon isotope ratios in biota, soils and tidal water of a Georgia
salt-marsh. Estuarine And Coastal Marine Science 4: 609-616.

HARRIS, D., W. R. HORWATH, AND C. VAN KESSEL. 2001. Acid fumigation of soils to
remove carbonates prior to total organic carbon or carbon-13 isotopic analysis. Soil Sci.
Soc. Am. J. 65: 1853-1856.

HAVENS, K. J., L. M. VARNELL, AND B. D. WATTS. 2002. Maturation of a constructed tidal
marsh relative to two natural reference tidal marshes over 12 years. Ecol. Eng. 18:
305-315.

HEDGES, J. I. 1992. Global biogeochemical cycles Progress Aa problems. Mar. Chem. 39:
67-93.

------, AND J. H. STERN. 1984. Carbon And Nitrogen Determinations Of Carbonate-Containing
Solids. Limnol. Oceanogr. 29: 657-663.

HEMMINGA, M. A., AND C. M. DUARTE. 2000. Seagrass Ecology. Cambridge.

------, AND M. A. MATEO. 1996. Stable carbon isotopes in seagrasses: Variability in ratios and
use in ecological studies. Mar. Ecol.-Prog. Ser. 140: 285-298.










VIZZINI, S., G. SARA, M. A. MATEO, AND A. MAZZOLA. 2003. delta C-13 and delta N-1 5
variability in Posidonia oceanica associated with seasonality and plant fraction. Aquatic
Botany 76: 195-202.

WANG, X. C., R. F. CHEN, AND A. BERRY. 2003. Sources and preservation of organic matter
in Plum Island salt marsh sediments (MA, USA): Long-chain n-alkanes and stable carbon
isotope compositions. Estuar. Coast. Shelf Sci. 58: 917-928.

WU, J., R. G. JOERGENSEN, B. POMMERENING, R. CHAUSSOD, AND P. C. BROOKS.
1990. Measurement of soil microbial biomass c by fumigation extraction an automated
procedure. Soil Biology & Biochemistry 22: 1167-1169.

ZEDLER, J. B. 2004. Compensating for wetland losses in the United States. Ibis 146: 92-100.

------, AND J. C. CALLAWAY. 1999. Tracking wetland restoration: Do mitigation sites follow
desired traj ectories? Restor. Ecol. 7: 69-73.

------, AND S. KERCHER. 2005. Wetland resources: Status, trends, ecosystem services, and
restorability. Annual Review Of Environment And Resources 30: 39-74.

ZHOU, J., Y. WU, J. ZHANG, Q. KANG, AND Z. LIU. 2006. Carbon and nitrogen composition
and stable isotope as potential indicators of source and fate of organic matter in the salt
marsh of the Changjiang Estuary, China. Chemosphere 65: 310-3 17.

ZHU, Z. B., R. C. ALLER, AND J. MAK. 2002. Stable carbon isotope cycling in mobile coastal
muds of Amapa, Brazil. Cont. Shelf Res. 22: 2065-2079.

ZIEMAN, J. C., J. W. FOURQUREAN, AND T. A. FRANKOVICH. 1999. Seagrass die-off in
Florida Bay: Long-term trends in abundance and growth of turtle grass, Jhala~ssia
testudinum. Estuaries 22: 460-470.









(Connor et al. 2001). It is therefore important to know if restoration and construction of coastal

systems returns the C accumulation and storage capacity of these C sinks. Such research can

indicate whether mitigation is effective and if coastal wetland restoration can become a policy

tool for reducing CO2 emiSsions as was suggested by Connor et al. (2001). Studies that focus on

functional traj ectories of OC in restored/constructed systems and compare OC between

restored/constructed and natural systems help answer these questions.

Functional traj ectories are used to track the progress of constructed systems over time

and to compare constructed and reference systems (Simenstad and Thom 1996; Zedler and

Callaway 1999; Morgan and Short 2002). Functional trajectory studies examine many

"ecological attributes" that act as indicators of more complex ecosystem functions (Simenstad

and Thom 1996; Craft et al. 2003). Attributes reach functional equivalence when they have a

value similar to the reference. Functions can follow linear, asymptotic, and sigmoidal

traj ectories (Kentula et al. 1993) or no traj ectory at all (Zedler and Calloway 1999). Craft et al.

(2003) proposed that different attributes follow one of three traj ectories depending on whether

they are part of hydrologic, biological, or "soil" development processes. OC pool formation is a

soil development process, and soil development processes generally follow the longest traj ectory

before reaching functional equivalence (Craft et al. 2003). There have been many studies

documenting functional trajectories of sediment OC (SOC) or organic matter (OM) in restored

and constructed tidal marshes (Simenstad and Thom 1996; Craft 2001; Havens et al. 2002;

Morgan and Short 2002; Craft et al. 2003) but, to our knowledge, only one in seagrass beds

(Evans and Short 2005) and only a comparison study in mangrove forests (McKee and Faulkner

2000).









The North Carolina study shows how the molecular distribution (diversity of types and

dominant types) of lipids and isotopic signatures of lipids can be used to determine SOC sources.

The method was not without problems, however. Use of compound classes that may not be the

"most" diagnostic for vascular plants may have skewed results. Furthermore, this technique is

biased toward extractable, not bound, lipids in sediments (Canuel et al. 1997). It is important to

note that only the top 0.5 cm of sediment was analyzed in this study. Thus, depths where

macrophyte roots may contribute to SOC through exudates or senescent tissue were ignored.

Not all studies examine a wide range of lipids. It is common to examine only one lipid

class such as n-alkanols (Bull et al. 1999) or n-alkanes (Wang et al. 2003). Studying the 613C

values of one type of lipid biomarker can help solve the intermediate-value problem that muddles

analysis of SOC sources in bulk isotope studies. The use of compounds that are specific to

vascular plants (n-alkanes) or to plankton and algae (HBI alkenes; Canuel et al. 1997) for isotope

studies can help clarify their contributions to SOC (i.e.: whether a sediment bulk 813C Value Of -

18 %o is due to an even mix of C4 and C3 plants, only plankton, or a mixture of all three). The n-

alkanol homologue, C32, WAS chosen for a study addressing contributions ofS. alterniflora and

Puccinellia maritima to salt marsh SOC in the United Kingdom (Bull et al. 1999). By only

examining 613C ValUeS of an n-alkanol, which plankton cannot produce, plankton's confounding

intermediate 813C ValUe WaS removed as a factor from the isotopic mixing equation. Using a

two-member mixing model based on 813C ValUeS of the C32 n-homologue, contributions of S.

alterniflora to SOC were calculated. S. alterniflora contributed about 100% of primary biomass

to sediments directly beneath S. alterniflora stands and about 50% to sediments beneath P.

maritima stands. To fully understand all sources to SOC, this method should be expanded to

include group-specific biomarkers for both vascular plants and plankton. Otherwise when









all three OC pools--TOC, Extractable OC (ExOC), and MBC. The 0-10 cm depths, also

reached equivalence for ExOC pools. On a storage (areal) basis, equivalence was also reached

by mangrove 0-5 cm depths for ExOC and MBC and seagrass 0-10 cm depths for MBC. This

equivalence was only on a storage basis because it was driven by greater bulk densities in the

constructed sediments. Seagrass sediments reached SOC pool equivalence more than mangrove

sediments due to their constant inundation, parent material, and lower equivalence goal

(reference seagrass sediments had less TOC and ExOC than reference mangrove sediments).

SOC pool sizes were not the only factors that indicated if constructed systems had attained

functionally equivalent OC storage--information about OC accumulation rates and OC liability

was also needed. OC accumulation rates in the constructed mangrove and seagrass systems were

similar to literature values if accumulation in surface layers was considered. It was unknown if

the constructed systems could sustain these accumulation rates over the long term or if the rates

reflected an immediate response in SOC after construction. Larger proportions of the TOC pool

were MBC in constructed 0-10 cm sediments indicating greater SOC liability and therefore less

SOC storage in constructed systems.

Objective Three: Comparisons of Sediment Organic Carbon Lability

Generally, constructed systems SOC was more labile than reference system SOC for both

mangrove and seagrass sediments (Fig. 5-1 and 5-2). Lability was only similar between

constructed and reference systems in the upper portion of seagrass sediments and in seagrass

floc. These results confirm hypothesis two. Greater liability of OC in the constructed system

indicates that the constructed system does not function as well as reference systems in terms of

OC storage. Even when SOC pool sizes are similar to references', as in the 5-10 cm depth of

constructed seagrass sediments, greater OC liability indicates that OC storage is not functionally

equivalent. The more labile OC is, the more likely it will be mineralized by microbes and










].2 Reference Seagrass Sediments


0.20 -

0.18 -


06-
0.14 -

0.12 -


t ji .


_


0.10-

0.08-

0.06-

0.04-

0.02-
-30


-25


813C


Figure 4-6. Continued









Total OC (TOC) and total nitrogen (TN) were measured on freeze-dried sediment and

surface layer samples. Freeze-dried samples were composite by plot and sieved through a 1

mm mesh screen to remove large shell pieces and carbonate rock, which were weighed so their

mass could be accounted for in calculations. Samples were then ball-milled to a fine powder in

stainless steel canisters. Inorganic C (IC) in was removed from samples via vapor acidification

(Hedges and Stern 1983; Harris et al. 2001). Samples were weighed out into 9 x 5 mm or 10 x

10 mm silver capsules (Thermo Scientific, Waltham MA and CE Elantech, Lakewood, NJ),

moistened with deionized water, and placed in an airtight container with a beaker of concentrated

HCI (12 M) for 24 hours before being dried at 60oC for 24 hours. Samples were then rolled and

analyzed for OC on an elemental analyzer (ECS 4010, Costech Analytical Technologies,

Valencia, CA). Peach leaves (NIST 1547) were used for calibration standards, and sucrose and

an internal soil standard were used for quality control. Tests were run on sand samples with

various carbonate percentages and total weights to assess the efficacy of vapor acidification and

to determine the maximum sample mass that still ensured complete removal of IC. Furthermore,

concurrent measures of 13C were used to confirm complete removal oflIC, and if incomplete IC

removal was suspected, samples were rerun at a lower total mass. Unacidified samples were run

separately in tin capsules (Costech) on a Flash EA 1112 series elemental analyzer (Thermo

Scientific, Waltham, MA) for TN. Acetinilide was used for calibration standards, and peach

leaves (NIST 1547) and an internal soil standard were used for quality control.

Extractable organic C (ExOC) and microbial biomass C (MBC) were measured using a

modified fumigation-extraction procedure (Vance et al. 1987; Joergensen and Mueller 1995).

Approximately 5 g of moist sample was weighed out in duplicate for sediment, algal mat, and

litter samples and 10 g was weighed out for floc samples. One set of samples was immediately








SL 15 Seagrass


Water Column
Seagrass (513C
Seston 1513C tMC:N
Macroalgae t)613C
tC:N


RefSeagrass


Water Column
Seston l513C Seagrass (513C
JC:NMacroalgae M5130 tC:N


Figure 4-9. Continued.









Water column


Mang rove/Trerrestrial
Imports CO,


CO2


Figure 1-1. The carbon cycle in seagrass beds.






























































3 -vr-old constructed

Natural reference

Bare spoil

Spoil planted with
Spartina
Fertilized spoil
with Spartina
13-vr-old constructed


Table 2-3. Continued
C accumulation Data
Location Site (g C m yr ) Time scale Method sources


Salt marshes

Upper Bay of Fundy'
Canada


Outer Bay of Fundy,
Canada



St Marks NWR, Florida












Lafourche Parish,
Louisiana


13Cs profiles

13Cs profiles

'37 poie
13Cs profiles

'C bomb
uptake
'C bomb
uptake
'C bomb
uptake
14C rfl
1C profiles2

1C profiles2



Feldspar
marker
Feldspar
marker


210b profiles2

1C profiles2

210b profiles2

1C profiles2

Modeled

13Cs profiles

Modeled

aoc /Time2
'37Cs and
210b profiles

aoc /Time3

aoc /Time3

aoc / Time3


Low marsh

High marsh

Low marsh

High marsh

Low marsh

Mid marsh

High marsh

Low marsh

Mid marsh

High marsh

Continuous canal
Discontinuous
canal

Natural waterway


39

194

76

188

117

101

65

25

22

20

300'

200'

650'

89

18.5

78

39.8

96

180

105

39

35-51

80

87

96.8

62


30 Year

30 Year

30 Year

30 Year

12 Year

12 Year

12 Year
400-600
Year
400-600
Year
400-600
Year
Annual

Annual

Annual

150 Year

Millennia

150 Year

Millennia

NA

Decadal

NA

3 Year

Decadal

16 Month

16 Month

16 Month

13 Year


Cedar Creek, Maryland



Hell Hook, Maryland



Barnstable,
Massachusetts
Biloxi Bay, Mississippi
Waarde Marsh,
Netherlands
Consultant,
North Carolmna


Drum Inlet
North Carolmna




Dill's Creek
North Carolmna











Table 3-4. Mean (+ SE) organic carbon concentrations (%) and storage (g m-2), nitrogen concentrations, and carbon to nitrogen molar
ratios of SL 15 (n=4) and reference (n=3) mangrove and seagrass sediments according to depth and month. TOC=total
organic carbon and TN=total nitrogen.
TOC TOC TN C:N
Month and system Depth (%/) (g m-2) (%) (mOlar ratio)


SL 15
2.49 (0.7)
0.13 (0.04)
0.11 (0.03)

3.27 (0.7)
0.17 (0.02)
0.14 (0.03)

1.8 (0.4)
0.9 (0.01)
2260 (0.07)
0.27 (0.1)


5.0 (1)

0.25 (0.05)
0.41 (0.1))


Reference SL 15 Reference SL 15


Reference SL 15 Reference


July mangrove



Nov. mangrove



July seagrass





Nov. seagrass


Algal mat/ litter
0-5 cm
5-10 cm

Algal mat/ litter
0-5 cm
5-10 cm


11.9 (3)

1.4 (0.4)

18.0 (3)
1.3 (0.3)
1.7 (0.5)


170 (20)
110 (30)
77 (20)

310 (60)
140 (10)
110 (30)


670 (50)
610 (50)
620 (90)

1300 (400)
600 (90)
760 (200)


0.26 (0.07)
0.018 (0.005)
0.010 (0.003)

0.37 (0.1)
0.024 (0.004)
0.013 (0.001)

0.20 (0.06)
0.096 (0.01)
0.023 (0.008)
0.027 (0.01)


0.62 (0.1)
0.13 (0.02)
0.027 (0.006)
0.041 (0.01)


0.096 (0.02)
0.11 (0.03)

0.79 (0.1)
0.12 (0.02)
0.14 (0.03)


8.2 (.22)
5.0 (0.99)
8.0 (1.7)

8.1 (0.92)
5.5 (0.33)
6.2 (0.77)


9.3 (0.29)
9.6 (2.5)
9.2 (1.3)

21 (5.4)

8.8 (0.9)

5.7 (0.55)

6.7 (0.16)
6.6 (0.19)
7.0 (0.21)

10.8 (0.65)

5.4 (0.22)
5.6 (0.35)
6.1 (0.36)


Floc
Accreted
0-5 cm
5-10 cm
10-15 cm

Floc
Accreted
0-5 cm
5-10 cm
10-15 cm


2.7 (1) 60 (20)
260 (50)
0.91 (0.6) 170 (40)
0.65 (0.8) 200 (80)
0.63 (1.0)

3.7 (0.3) 130 (30)
340 (40)
0.97 (0.1) 180 (30)
0.65 (0.6) 280 (60)
0.62 (0.3)


84 (40)

440 (20)
330 (20)
370 (40)

96 (10)

390 (30)
340 (10)
370 (1)


0.39 (0.1) 8.2 (0.95)

0.12 (0.01) 6.9 (0.27)
0.084 (0.01) 7.0 (0.98)
0.077 (0.01)

0.30 (0.04) 6.8 (0.08)
6.1 (0.38)
0.15 (0.02) 6.7 (0.76)
0.10 (0.01) 73(.)
0.089 (0.009)









measured in Terminos Lagoon, Mexico (Gonneea et al. 2004) and the highest rate measured at

the low marsh in Jiulonglijang Estuary in China (Alongi et al. 2005). Rates in salt marshes

ranged from 2 to 300 g C m-2 -1l with the lowest rate measured at a natural site in Dell's Creek,

North Carolina (Craft et al. 2003) and the highest rate measured behind a continuous canal in

Lafourche Parish, Louisiana (Cahoon and Turner 1989). Rates in seagrass beds ranged from 19

to 191 g C m-2 -1l, a range measured offshore of Cala Culip, Spain (Romero et al. 1994).

Comparing the compiled rates for these systems, there were no trends of one system

having consistently higher C accumulation rates than the other systems (Table 2-3). The system

accumulating C at the highest rates even varied within the same region. For example, in

Celestun Lagoon, Mexico, mangroves accumulated more carbon in their sediments than

seagrasses, but in Terminos Lagoon, Mexico, the reverse was true (Gonneea et al. 2004). This

lack of a trend is supported by Chmura' s (2003) review that found no significant differences

between C accumulation rates in salt marshes and mangroves. It should be noted, however, that

contributions of mangrove forests to C storage on an ecosystem scale may be greater than salt

marshes and seagrass beds because large amounts of C are stored for decades in woody biomass

of mangrove trees (Twilley et al. 1992).

Measuring Rates of Carbon Accumulation

Calculating rates of C accumulation typically involves three steps. The first step is to

measure SOC pools. SOC pools for local rate studies were directly measured using either an

elemental analyzer after acidification of the sample to get rid of carbonates (Gacia et al. 2002), a

TOC analyzer (Brunskill et al. 2002; Alongi et al. 2004), or a mass spectrometer (Choi and

Wang 2004). SOC pools were indirectly measured by using loss-on-ignition (LOI) values in

regression equations describing the relationship between SOM and SOC (Connor et al. 2001).

The second step is to age the sediment or measure rates of sediment accretion. Sediment age and









indicates that 50% of a biomarker is from S. alterniflora and the other 50% is from P. maritima

does not mean each species contributes to 50% of the bulk SOC. Furthermore, when examining

isotopes of group-specific lipids, one must make sure that the lipids being examined have similar

abundances in each species. Otherwise, what seems like a greater abundance of one species in

SOC might actually be due to a greater abundance of that biomarker in tissues of that species

relative to other species. In cases where biomarkers are not likely to be at the same

concentration among species, relative abundances of biomarkers within each plant should be

included in isotopic mixing equations (Bull et al. 1999).

Petrographic Analysis

Petrographic analysis of sediments is a lesser-used method to determine SOC sources.

Petrographic analysis microscopically examines organic matter for recognizable organic

components such as macrophytic tissues, differentiated based on their level of decomposition,

and algae. Marchland et al. (2003) examined SOC sources in mangrove forests of various ages

using six different categories of plant tissues: Translucent ligno-cellulosic debris (TLC), which

exhibited preserved cell wall structures, degraded ligno-cellulosic debris (DLC), which exhibited

decaying cell walls, gelified particles (GP), which were orange brown gel-like particles produced

by cellulose degradation, reddish amorphous organic matter (RAOM), in which the cellulose is

completely degraded, oxidized opaceous ligno-cellulosic debris (OLC), which were dark and

structureless refractory land-derived OM, and grayish amorphous organic matter (GAOM),

which were the remains of algae and phytoplankton. This study looked at relative proportions of

these various components to understand whether SOC sources to mangrove forests were

autochthonous algae, mangroves, or allocthonous riverine detritus. Combining proportions of

these OM components with C:N ratios, they found that sediments of younger mangrove forests,

with their low C:N ratios and high proportion of GAOM, were dominated by algal-derived OM








0-5

5-10
5-101 0-5

"
..
I
: :
-
I I
-2 -


l itter


-


613(
r
-30


-28


-26


-"18


epiphytes


I
-16


algal mat




seag rass


-14 -12 -


I
10


fI


iacroalgae



2. 20



loc = loc
: :

5-10

0-8*
accreted


1


0-5


~Potential Source
Ref Mangrove
*******----- SL 15 Mangrove
Ref Seagrass
*********** SL 15 Seagrass


5-15


Figure 4-3. 613C averaged over July and November 2006 for SL 15 and reference sediments and surface layers compared to mean 613C
of potential sources.


seston

mang roves
te rrestri al











Table 2-3. Rates of carbon accumulation in coastal ecosystem sediments and the methods used
to calculate the time component of the rates.
C accumulation Data
Location Site (g C nt yr ) Time scale Method sources


Mangrove forests
Herbert River
Estuan Australia
Jiulonglijang Estuart
China





Florida Keys, Florida



Rookery Bay, Florida


210b profiles

20b profiles

20b profiles

20b profiles

'3Cs profiles

'3Cs profiles
Feldspar
marker
Feldspar
marker
Feldspar
marker
Feldspar
marker
Estimate

210b profiles

210b profiles

210b profiles

210b profiles

210b profiles

210b profiles
'37Cs and
210b profiles


Feldspar
marker
Feldspar
marker
Feldspar
marker
Feldspar
marker
Feldspar
marker
Feldspar
marker


180

168

204

841

159

105

228

328

291

191

150

101

110

127

55-70

67-104

33

184-281



700

35

30

75

10

335


Century

Century

Century

Century

30 Year

30 Year

Annual

Annual

Annual

Annual

8,000
Year
Century

Century

Century

Century

Century

Century

Decadal



Annual

Annual

Annual

Annual

Annual

Annual


High intertidal

Mid intertidal

Low intertidal

Rhizophora mangle

4vicennia germinans

Fringe

Basin


Exposed island

Sheltered island


Matang Forest Presen e,
Malaysia


5-yr-old stand

18-yr-old stand

85-yr-old stand


Celestun Lagoon, Mexico

Chelem Lagoon, Mexico
Terminos Lagoon'
Mexico

Sawi Bay, Thailand

Brackish Marshes
Cameron Parish,
Louisiana




Fina La Terre,
Louisiana


Rockefeller Refuge,
Louisiana


Natural waterway

Restricted canal
Restricted natural
waterway
Unmanaged

Managed

Unmanaged










Table 3-6. Mean (+ SE) organic carbon liability of organic carbon in SL 15 (n=4) and reference
(n=3) sites according to depth and month.


Lability
(mg 02 g-10C lu- )


System and month


Julymnagrove



Noy.nuangrove



July seagrass





Nov. seagrass


Depth


SL 15
1120 (300)
1480 (580)
360 (120)

2230 (540)
1210 (250)
572 (350)

631(73)
387 (45)
1170 (120)
856 (120)


333 (79)
555 (52)
1280 (110)
800 (69)


Reference
760 (290)
520 (170)
332 (130)

320 (100)
355 (21)
198 (48)

782 (320)

527 (34)
677 (48)
469 (29)

421 (120)

492 (25)
625 (38)
626 (40)


Algalnxit/ litter
0-5 cm
5-10 cn1

Algalnxit/litter
0-5 cm
5-10 cn1

Floc
Accreted
0-5 cm
5-10 cn1
10-15 cm

Floc
Accreted
0-5 cm
5-10 cm
10-15 cm










extracted with 25 mL of 0.5 M K2SO4 for an hour and then filtered through a Whatman 42 filter.

The second set was fumigated in an ethanol free-chloroform atmosphere for 24 hours before

being extracted as above. Extracts were diluted, acidified, and run for OC on a Shimadzu TOC-

5050A (Shimadzu North America, Columbia, MD). OC in non-fumigated samples was ExOC.

The difference between OC in fumigated and non-fumigated samples, multiplied by a correction

factor of 2.22 (Wu et al. 1990; Joergensen 1996; Jenkinson et al. 2004), was MBC.

Sediment oxygen demand (SOD; APHA 1992), normalized to TOC, was used as a measure

of OC liability. SOD was measured by mixing 10 mg of wet sample with about 300 mL of

oxidized, salt water in dark biological oxygen demand (BOD) bottles. The salt water was

created by dissolving Instant Ocean Sea Salt (Marineland Labs) into deionized water until the

solution reached 25 ppt. Dissolved oxygen (DO) content of the water was measured initially and

after 24 hours by a Fisher Accumet AR40 DO meter. Measurements were taken after the water

and sample in each BOD bottle were thoroughly mixed on stir plates for 30 minutes. While

abiotic and chemotrophic reactions can cause decreases in DO, these reactions most likely did

not cause a significant Oz reduction during this experiment because samples were already

exposed to 02 during processing. Furthermore, NH4' leVOIS in the samples were low

(unpublished data) and pH did not change during incubation, which would have indicated

oxidation of sulfide in the samples,. The maj ority of 02 depletion was therefore assumed to be

due to biological, heterotrophic oxidation of OC.

OC accumulation rates (g OC m-2 -1l) in SL 15 were calculated using equation 3-1

(Cammen 1975; Craft et al. 1999).

OC, -OC, + OC
OC a(3-1)
GCCHImill/onl/ A
system










OTERO, X. L., T. O. FERREIRA, P. VIDAL-TORRADO, AND F. MACIAS. 2006. Spatial
variation in pore water geochemistry in a mangrove system (Pai Matos island, Cananeia-
Brazil). Applied Geochemistry 21: 2171-2186.

PAPADIMITRIOU, S., H. KENNEDY, D. P. KENNEDY, C. M. DUARTE, AND N. MARBA.
2005. Sources of organic matter in seagrass-colonized sediments: A stable isotope study of
the silt and clay fraction from Posidonia oceanica meadows in the western Mediterranean.
Org. Geochem. 36: 949-961.

RABENHORST, M. C. 1995. Carbon storage in tidal marsh soils, p. 93-104. hz R. Lal, J.
Kimble, E. Levine and B. A. Stewart [eds.], Soils and Global Change. Advances in Soil
Science. CRC Lewis.

RAMOS E SILVA, C. A., A. P. DA SILVA, AND S. R. DE OLIVEIRA. 2006. Concentration,
stock and transport rate of heavy metals in a tropical red mangrove, Natal, Brazil. Mar.
Chem. 99: 2-11.

ROCHETTE, P., AND E. G. GREGORICH. 1998. Dynamics of soil microbial biomass C,
soluble organic C and CO2 evolution after three years of manure application. Canadian
Journal of Soil Science 78: 283-290.

ROMERO, J., M. PEREZ, M. A. MATEO, AND E. SALA. 1994. The belowground organs of
the Mediterranean seagrass Posidonia-oceanica as a biogeochemical sink. Aquatic Botany
47: 13-19.

SCHLESINGER, W. H. 1990. Evidence from chronosequence studies for a low carbon-storage
potential of soils. Nature 348: 232-234.

------. 1997. Biogeochemistry: An Analysis of Global Change. Academic Press.

SENECA, E. D. S. W. B., W. W. WOODHOUSE, L. M. CAMMEN, J. T. LYON. 1976.
Establishing Spartina alterniflora marsh in North Carolina. Environmental Conservation
3: 185-188.

SFWMD. 2005-2006. Environmental Database (DBHYDRO). South Florida Water Management
Di stri ct. URL: www. sfwm d.gov/org/ema/db hy dro/. D ate acce s sed (July 200 O7).

SHORT, F. T., E. W. KOCH, J. C. CREED, K. M. MAGALHAES, E. FERNANDEZ, AND J.
L. GAECKLE. 2006. SeagrassNet monitoring across the Americas: case studies of
seagrass decline. Marine Ecology-an Evolutionary Perspective 27: 277-289.

SIGUA, G. C., AND W. A. TWEEDALE. 2003. Watershed scale assessment of nitrogen and
phosphorus loadings in the Indian River Lagoon basin, Florida. Journal of Environmental
Management 67: 363-372.

SIMENSTAD, C. A., AND R. M. THOM. 1996. Functional equivalency trajectories of the
restored Gog-Le-Hi-Te estuarine wetland. Ecol. Appl. 6: 38-56.












SL 15 Mangrove


-~5 cm
S5-10 cm


2.2 -


2.0-


1.8-


1.6-


1.4-


Nov05 Feb06 May06 Jul06 Nov06


Figure 3-3. The functional traj ectory the bulk density of SL 15 mangrove sediments followed
over the first year after construction. The symbols are the mean values for each
sampling date (n=12) and error bars are d- SE.



































To my parents, Michelle and Richard Hicks, for their wonderful support of all my academic
endeavors from preschool onwards











Storage (g m )


S 0-5 cm
lc 0-10 cm


120~ EA


O
r 80

o
S60

40

20

0
100


80

O





2 0


20


Reference SL 15 Reference SL 15


Figure 3-7. Comparisons between extractable organic carbon (ExOC) in reference and SL 15
mangrove (top) and seagrass (bottom) sediments. The bars are mean ExOC averaged
over month (July and November 2006) for each depth of sediment (n=12 for SL 15
and n=9 for reference). Error bars are + SE. Depths in the seagrass systems are as
follows: 1= SL 15 accreted and reference 0-5, 2= SL 15 0-5 and reference 5-10, 3=
SL 15 5-10 and reference 10-15. An asterisk indicates a significant site effect (Table
3-5). Capital letters are results of a Tukey test performed after a significant site x
depth interaction, and lowercase letters are results of a Tukey performed after an
insignificant site x depth interaction, but a significant one way ANOVA. Bars that
share letters are not significantly different.


(mg kg')


M1
@2
m 3










generally had less than 10% of their MBC as TOC (Boschker et al. 2000; Bouillon et al. 2004;

Cordova-Kreylos et al. 2006). OC limitation is a possible reason for high microbial biomass.

The low C:N ratios of constructed and reference sediments suggest a C limitation (Sterner and

Elser 2002). When microbes are C limited they tend to sequester C in their cells instead of

respiring C for energy (Anderson 2003). This mechanism is supported by another study with

high MBC percentages (23 to 50% of TOC), as its sediments also had low TOC (<1.0%)

(Joergensen and Mueller 1995).

Constructed sediments do not store OC as well as reference sediments because the liability

of SL 15 OC was greater than references at all depths except for seagrass floc and accreted

layers. Lability is a proxy for the decomposability of OC-the greater the liability, the faster OC

is decomposed releasing C back to the atmosphere. It is therefore unlikely that labile OC would

be stored in sediments for long periods of time. One study of macro organic matter (MOM), a

precursor of sediment OM, in constructed marshes showed that younger marshes had more labile

MOM than older marshes indicating they were less likely to sequester OC in the long term (Craft

et al. 2003)

Organic carbon accumulation

Rates of OC accumulation are another factor that determines how well constructed

systems function as OC stores. Pool sizes measure how much C systems are keeping from the

atmosphere, liability indicates how long C is likely sequestered, and accumulation rates measure

how much C is being actively taken from the atmosphere (via plant production). Salt marsh

studies found equal and even greater OC accumulation rates in constructed marshes (Cammen

1975; Craft et al. 1999; Craft et al. 2003). In this study, OC accumulation rates in constructed

seagrass sediments were similar to those in other studies, but rates of constructed mangrove

sediments were much lower than other studies unless the algal mat was included (Table 7, Fig. 3-









material had accumulated on top of the seagrass section, which was collected and analyzed

separately from the original sediment depths as an accreted layer. Surface layers--floc from

seagrass systems, algal mats from the SL 15 mangrove system, and litter layers from the

reference mangrove system--were collected from each core and were composite by plot.

Differences in color and texture were used to separate accreted and surface layers from original

depths except for floc, which was the fraction of the accreted layer that poured off (Fig. 4-2).

Laboratory Analyses

Rocks, roots, and detritus were removed from each sample prior to homogenization.

Samples were then freeze-dried for 48 hours. Freeze-dried sediment samples were composite

by plot and sieved through a 1 mm mesh to remove large shell pieces and carbonate rock, which

were weighed so their mass could be accounted for in calculations. Sediment and surface layer

samples were then ball-milled to a fine powder in stainless steel canisters.

TOC, TN, and 813C were measured in sediment, surface layers, seston filters, and plant

samples. TOC and TN were used to calculate C:N ratios on a molar basis. Inorganic carbon (IC)

was removed from sediment, surface layer, and seston samples via vapor acidification (Hedges

and Stern 1983; Harris et al. 2001; Gonneea et al. 2004). Sediment and surface layer samples

were weighed out into 9 x 5 mm or 10 x 10 mm silver capsules (Thermo Scientific, Waltham

MA and CE Elantech, Lakewood, NJ), which were arranged in plastic well plates and moistened

with deionized water before acidification. Three holes (7 mm in diameter) were cut from each

seston filter with a hole punch and arranged in plastic well plates. The filled well plates were

then placed in a glass desiccator with a beaker of concentrated HCI (12 M) for 24 hours before

being dried at 60oC for another 24 hours. Seston filter samples were then put into 10 x 10 mm

silver capsules. Plant samples were weighed into 9 x 5 mm tin capsules (Costech Analytical

Technologies, Valencia, CA). All samples were combusted on an elemental analyzer (ECS











Table 4-5. Mean 613 01 Of 3C ranges of means for sources in this study and in the literature.
Means are averaged across and ranges are across plant parts, species, and sites for this
study and where applicable in the literature. The sources listed here are the main
SOC sources that were used in this study's ternary diagrams.

8 C, (% o) 8 3C (% o
Source This study Other studies Data sources
Seagrass Mean: -11.54 -10.0 1
Range: -20.07 to -9.23 -19.7 to -10.7 2
-13.3 to -5.8 3
-12.4 4
-12.2 5
-16.1 to -11.9 6
-10.5 7
-23 to -3, -10 8
(mode)
-10.4 to -7.2 9
-14.6 to -8.8 10
-12.7 to -11.4 11
Mangrove Mean: -26.95 -27.0 12
Range: -27.77 to -25.26 -28.3 to -24.1 2
-29.0 to -27.0 13
-28.4 to -27.9 3
-28.8 7
-28.2 14
-27.9 15
-30.1 to -28.3 16
-29.7 to -25.9 17
Macroalgae Mean: -21.00 -31.7 to -16.6 11
Range: -32.06 to -17.11 -21.5 to -15.0 18
-26.0 to -20.9 15
-15.61 7
Seston Mean: -26.29 -22.0 to -21.0 12
Range: -30.10 to -24.23 -23.0 to -20.5 13
-18.4 1
-23.3 to -13.7b 2
-27.6 to -12.1 3
-22.1 4
-24.7 5
-25.32 to -22.06b 6
-20.6 7
-22.6b 19
-28.1 to -20.8b 20
-26.4b 21
Terrestrial (C3) Mean: -27.54 -28 to -25 22
Range: -28.30 to -26.33 -30 to -25 23
-26 24
"l, Canuel et al. 1997: 2, Hemminga et al. 1994: 3, Kennedy et al. 2004; 4, Papadimitriou et al. 2005; 5, Gacia et al.
2002; 6, Gonnocea et al. 2004; 7, Thimdee et al. 2003; 8, Hemminga and Mateo 1996: 9, Anderson and Fourqurean
2003; 10, Vizzini et al. 2003; 11, Smit et al. 2005; 12, Jennerjahn and Ittekkot 2002; 13, Bouillon et al. 2003; 14,
Bouillon et al. 2004; 15, Bouillon et al. 2002; 16, Lallier-Verges et al. 1998: 17, Muzuka and Shunula 2006; 18,
Fenton and Ritz 1988: 19, Zhou et al. 2006; 20, Dittmar et al. 2001; 21, Cifuentes et al. 1996: 22, Miserocchi et al.
2007; 23, Kang et al. 2007; 24, Ogrinc et al. 2005
bCalled particulate organic matter (POM) or suspended particulate matter (SPM) by the authors













SL 15 Mangrove


SL 15 Seagrass


- -


0-5 cm
5-10 cm
algal mat
Ref litter
Ref0-10


S5-10 cm


- -- accreted
-0- floc
-- Ref 0-15
-- -- Ref floce (July)


-19E


-19 F - -


-21 E


-21 t


Nov05 JanO6 May06 Jul06 Nov06


Nov05 Feb06 May06 Jul06 Nov06


Nov05 Feb06 May06 Jul06 Nov06


Figure 4-4. Mean 613C Of SL 15 sediments and surface layers over the first year after construction. Error bars are d-SE. Reference
lines are 613C averaged over depth (for sediments) and month (except for reference floc) for the respective reference
systems.


ir









against heavy isotopes during carbon uptake and fixation. Discrimination against the heavy

isotope is highest when the inorganic C exceeds supply. Generally, C3 plants are lighter (613C

-35 to -20 %o) than C4 plants (813C = -15 to -9 %o) in their 613C Signatures due to the strong

isotopic discrimination of the carboxylase Rubisco, an enzyme that is not found in the C fixation

pathway of C4 plants (Hemminga and Mateo 1996; Hemminga and Duarte 2000). Luckily for C

source determination in coastal systems, the principal primary producers of seagrass beds,

mangrove forests, and salt marshes all have isotopic signatures distinct from the signatures of the

less abundant primary producers within these systems. Seagrasses are relatively heavy

isotopically with average 613C ValUeS of -10 to -11 %o (Hemminga and Mateo 1996).

Mangroves, a C3 plant, have isotopic signatures close to that of many terrestrial primary

producers with 813C ValUeS around -28 %o (Jennerj ahn and Ittekkot 2002; Kennedy et al. 2004).

Spartina species that dominate salt marshes are C4 plants with 813C ValUeS around -12 to -13 %o

(Haines 1976; Middelburg et al. 1997). The isotopic signatures of other primary producers such

as plankton and epiphytes generally fall below that of seagrasses and Spartina and above that of

mangroves (Kennedy et al. 2004; Papadimitriou et al. 2005), but this is not always the case.

In order for the bulk stable isotope method to be accurate, the 6 813C Signature of the

sources must not change during decomposition, or if they do change, the magnitude of the

change needs to be small when compared to inter-source differences (Papadimitriou et al. 2005).

Changes during decomposition are often small, like the 0.7 %o difference found between fresh

and senescent mangrove leaves in Brazil (Jennerj ahn and Ittekot 2002), but are variable in

direction and magnitude depending on the plant (Dai et al. 2005).

A study of OC inputs into seagrass (Posidonia oceanica) sediments of 22 sites in the

northwestern Mediterranean by Papadimitriou et al. (2005) is a good example of the potential of









813C and a relatively low C:N were found in seagrass fronds. S. alterniflora had a low 613C and

high C:N. Seston had low 813C and low C:N. Seston samples had greater 613C in fall than in

winter (ANOVA, df=4, p=0.0002) but did not differ between ebb and flood tides (ANOVA,

df-1, p=0.54). Compared to variation among plant groups, variation of 613C within plant groups

was usually low with mangrove tissues of all species varying by less than 2.5%o and seagrass

tissues (except H. johnsonii) by less than 3.4%o. The exception was the macroalgae group,

whose 613C varied by 15%o. Macroalgae had high variability with C:N ratios as well. Plant

tissue type influenced C:N ratios with greater C:N in roots than in leaves for both mangroves and

seagrasses.

Plants also differed in their decay rates, even within groups (Table 4-2). The greatest

decay constants, and fastest rates of decay, were for a macroalgae (A. spicifera) and a seagrass

(S. filiforme). The slowest decay rates were for a seagrass (H. beaudettei) and a mangrove (R.

mangle).

Sediments and Surface Layers

613C Of SL 15 sediments and surface layers, with the exception of the 0-10 cm seagrass

sediments, changed significantly over time (month effect, p<0.034, Table 4-3, Fig. 4-4).

Mangrove sediments and seagrass accreted layers had 813C that increased towards the mean 613C

of their respective references over the first year after construction (Fig. 4-4). The mangrove

algal mat' s 613C also increased but moved away from reference values (Fig. 4-4). Most of the

layers did not have changing %OC or C:N ratios throughout the year. C:N ratios changed

significantly without direction in mangrove sediments and seagrass floc (Chapter 3). TOC

significantly changed in seagrass 0-10 cm sediments and floc, but only with direction in floc,

where it increased over time (Chapter 3). The 613C, TOC, or C:N values did not differ among

sediment depths (Table 4-3, Chapter 3).











Table 2-5. Continued
Source Sediment Main How main sources Data
Location Potential sources 8' C Site description 61 C sources determined source
Chelem, Mexico Seston -22.1 Fringe mangrove -23.1 to 1,2 Ternary mixing diagram 15
Seagrass -15.42 -26.16 4,2 of 61 C and N:C
Mangrove -27.12 Seagrass bed -17.2 to


-22.46
-26
-16


-22.7





-15.7


Terminos, Mexico




Santa Barbara,
Philipines




Buenavista, Philipines


Seston
Seagrass
Mangrove

Seston
Seagrass
Epiplwte
Mangrove

Seston
Seagrass
Epiplwte
Mangrove

Seston
Seagrass
Epiplwte
Mangrove

Seston
Sea grass
Macroalgae
Mangrove
Shrimp feed

Seston
Seagrass
Mangrove


-25.3
-11.9
-28.62

-19.0
-10.9
-12.9
-28.6

-17.7
-11.7
-13.1
-28.1

-27.6
-12.3
-22.9
-28.4

-20.6
-10.5
-15.6
-28.8 ,
-22.5


Fringe mangrove
Seagrass bed


Seagrass bed





Seagrass bed


Ternary mixing diagram
of 61 C and N:C


Percent contribution ranges
from mixing equation




Percent contribution ranges
from mixing equation


Umalagan, Philipines


Seagrass bed


2 or 4 Percent contribution ranges
from mixing equation


-26.6


Khung Krabaen Bay,
Thailand


Canals
Mangroves
Inner bay
Mouth of bay
Offshore


-26.5
-26.3
-15.1
-19.2
-17.5

-24.6


Comparison


Ghia Luan, Vietnam


-21.6
-13.3
-27.9


Seagrass


Percent contribution ranges
from mixing equation


"The numbers in the main sources column signify the following: 1, seagrass: 2, seston: 3, epiplwtes: 4, mangroves: 5, Spartina; 6, other; 7, macroalgae.
bl, Jennerjahn and Ittekkot 2002; 2, Hemminga et al. 1994: 3, Bouillon et al. 2003; 4, Soto-Jimenez et al. 2003; 5, Johnson and Calder 1973; 6, Haines 1976; 7,
Chmura et al. 1987 : 8, Wang et al. 2003; 9, Middelburg et al. 1997: 10, Canuel et al. 1997: 11, Bull et al. 1999; 12, Kennedy et al. 2004; 13, Papadimitriou et al.
2005; 14, Gacia et al. 2002; 15, Gonneea et al. 2004; 16, Thimdee et al. 2003.









"rate of OC sequestration," "rate of POC burial," "rate of refractory accumulation," and "organic

accumulation rate." Nuances of these terms could be gleaned from methodology. Some like

"rate of C accumulation" referred to additions of both labile and refractory OC to the SOC pool

(Craft et al. 2003). Others like "rate of refractory accumulation" referred to long-term burial of

OC that is unlikely to decompose on a human time scale (Cebrian 2002), and others like "POC

burial" were vague (Alongi et al. 2005). Some studies reported OM accumulation, not OC

accumulation. Those rates were divided by two to obtain OC accumulation rates. Also it was

assumed that "C accumulation rates" referred to accumulation rates of OC, not total C, because

studies that used the term reported measuring OC. Lastly, for indirectly measured rates

(modeled or based on mass balance equations) it was not always clear whether rates included

amounts of OC from both autochthonous and allocthonous sources. This review reports all

values as "C accumulation rates," which refers to the build up of OC in sediments though there

may be discrepancies in the liability of accumulating OC. Generally, the longer the timescale of a

study, the more likely rates represent long-term burial. Only the term "C burial rates"

definitively refers to long-term storage of refractory OC. There is a need for future studies to

clearly define rate terminology and to be consistent in its use.

Global Rates

Many scientists have estimated global rates of C accumulation for coastal ecosystems due

to their important role in the global C cycle (Table 2-2). Global rates of C accumulation for

these systems are calculated in several ways. The most common way was averaging published

accumulation rates for many sites (Duarte and Cebrian 1996; Chmura et al. 2003). Other

methods included graphing frequency distributions of published accumulation rates (Cebrian

2002), scaling up from model-derived rates (Suzuki et al. 2003), or using mass balance equations

derived from production and burial estimates (Jennerjahn and Ittekkot 2002). Despite the









Low C:N ratios of many sediment samples are concerning because it may have lead us to

overstate the importance of seston as a source because it is the only source with equally low C:N

ratios. Just as microbial activity likely lowered C:N ratios in seston (Cifuentes et al. 1996), it

could lower C:N ratios in sediments. Decreasing sediment C:N ratios during diagenesis has also

occurred in other source determination studies (Thimdee et al. 2003; Gonnocea et al. 2004;

Kennedy et al. 2004). Changes due to bacteria are likely because a high percentage of TOC in

these sediments is microbial biomass (11 to 63%; Chapter 3). Decreases in source C:N ratios

during decomposition may also explain the relatively low C:N ratios of the sediments compared

with living source material. Unfortunately, C:N ratios during decomposition were not measured

in this study. Results of studies that measured decomposition in similar systems were equivocal

(see source characteristics section). Another reason for low C:N ratios is the eutrophication of

the IRL, which has greatly increased the availability of inorganic N sources (Sigua and Tweedale

2003). Due to the influence of factors other than source identity in determining sediment C:N

ratios, caution is emphasized in interpreting ternary diagram results.

Conclusion

In all sediments, seston was a dominant source and diagenesis of organic matter within

sediments lowered sediment C:N ratios (Fig. 4-9). Because the other main sources differed

between SL 15 and reference sediments (Fig. 4-9), their abilities to sequester SOC probably

differ too. The litter bag decomposition study suggests which SOC sources are likely to be

sequestered in sediments the longest. This information allows us to predict how OC storage will

differ in sediments of SL 15 and reference sites. Since seston is an OC source for all sediments,

the fact that its decomposition was not measured should not greatly affect these predictions.

Because fast-decaying macroalgae OC dominates in SL 15 seagrass sediments, they are unlikely

to store OC for as long as reference seagrass sediments. A year after construction, SL 15's









CHAPTER 3
SEDIMENT ORGANIC CARBON STORAGE INT A CONSTRUCTED MANGROVE AND
SEAGRASS SYSTEM

Introduction

Coastal ecosystems such as salt marshes, mangrove forests, and seagrass beds are being

degraded and lost worldwide as a result of the eutrophication, sedimentation, and destruction that

accompany coastal development for human habitation, agriculture, and aquaculture (Valiela et al.

2001; Kennish 2002; Zedler 2004). In the United States development and infilling are the main

causes of coastal ecosystem loss (Dahl 2000). In the last two decades, humans have caused the

loss of 18% of the known worldwide area of seagrass beds (Green and Short 2003), and in the

last Hyve decades, have caused the loss of about 3 5% of the world' s mangrove forests (Valiela et

al. 2001; Alongi 2002). In the United States, about 50% of salt marshes have been lost

historically (Kennish 2001) and 25% of mangrove forests have been lost since the 1950's

(Bridgham et al. 2006). United States seagrass beds had a relatively constant area between 1986

and 1997 in, what is to our knowledge, the only nationwide seagrass inventory (Dahl 2000).

Smaller scale studies, however, have demonstrated local declines in the extent of seagrasses

(Zieman et al. 1999; Short et al. 2006). When coastal systems are lost, we lose not only wildlife

habitat, storm surge protection, and economically-important fish and shellfish nurseries, but also

biogeochemical functions like phosphorus retention, denitrification, and carbon (C) sequestration

(Alongi 2002; Duarte 2002; Zedler and Kercher 2005).

The United States has a policy of no net wetland loss that includes coastal wetlands as part

of the Clean Water Act (Zedler 2004; Zedler and Kercher 2005). Florida policy applies this no-

net-loss principle to seagrass beds as well (Florida Administrative Code, Chapter 18-21).

Destruction of mangrove and seagrass ecosystems in Florida requires compensatory mitigation

via restoration of an existing ecosystem or construction of a new ecosystem. Mitigation can













~trrltil fl~,~


Figure 4-2. Core from SL 15 seagrass system illustrating the surface layer (floc) and different
sediment depths accretedd layer, 0-5 cm, and 5-10 cm). Note the difference in color
between the accreted layer and 0-5 cm depth.



























I I I I I I I I I I I


i I


NovO5 Feb06 May06 Ju1O6 Nov06


- SL 15 M~angrove +osa
+ 5-10 em






-it


200 E SL 15 Seagrass


+ 0-sam
-0- 5-10 cm
-F- accreted


175 -

150 -

125 -

100 -

75


50 ~


oPb~g~~


0
3500


3000-

2500

2000-

1500-


l l l


2.0-


1.5-


1.0-


0.5-


Nov05 Feb06 May06 Jul06 Nov06


Figure 3-4. The changes in organic carbon parameters over the first year after construction in SL
15 seagrass and mangrove sediments. The symbols are the mean values for each
sampling date (n=12 for ExOC and MBC and n=4 for TOC) and error bars are + SE.









ACKNOWLEDGMENTS

I thank the entire Reddy Lab group, all of who helped me in some way throughout this

process. Within the lab group, I especially thank Cory Catts, who endured long, arduous days in

the field with me, Angelique Keppler and Melissa Martin, who answered my countless questions

about lab procedures and graduate school, and Ms. Yu, whose running of the lab is nothing short

of miraculous and who was always there to clarify procedures and offer encouragement during

long days in the lab. I also thank my advisor, Dr. Reddy, for allowing me to work on a project of

my own design.

Thanks to my committee, Todd Osborne, Jim Sickman, and Ted Schuur, who each helped

me understand certain concepts or lab procedures. Todd Osborne was an integral part of my

committee as he mentored me through much of this process. Thanks to Kelly Fischler, Rex Ellis,

and Hanna Lee for help in the field and Meghan Brennan for help with statistics.

Last but not least, I thank Alex Pries, who has assisted me in the field, lab, with computer

glitches, and in editing portions of this thesis, and who has supported me through both tough

times and good times.





For Pier


N


O


350 Meters


87.5 175
I I I I I


Figure 4-1. The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset). SL 15
is the large island in the center. Circles are mangrove system plots and squares are
seagrass system plots. Symbols outside of SL 15 are the reference sites, which have
one plot each.


SL 15


Indian
RFiver
Lagoon









different ways of calculating global rates, accumulation rates for intertidal systems are basically

in agreement (Table 2-2). Estimated global accumulation rates for mangrove forests range from

92 to 200 g C m-2 -1l and rates for salt marshes range from 50 to 175 g C m-2 -1l (ignOring the

high estimate of Rabenhorst (1995). Rates for seagrass beds are more variable and range from

16.5 to 270 g C m-2 -1l. Higher variability for seagrass ecosystems is likely because C

accumulation rates in seagrass sediments have been studied less than in mangroves and salt

marshes. Suzuki et al. (2003) estimated that seagrasses caused an accumulation of 1.2 g C m-2

yr- in deep ocean sediments due to export of their primary production to the open ocean and its

subsequent burial. While this review concentrates on in situ accumulation, it is important to

recognize that there are other ways these systems contribute to the global C sink.

Coastal ecosystems accumulate C at a rates several orders of magnitude greater than rates

in terrestrial systems and the open ocean (Table 2-2). C cycling in terrestrial systems should

reach a steady-state condition, making them neither a C source nor C sink (Hussein et al. 2004).

Disturbances such as fire, however, occur before climax stages causing terrestrial systems to

become C sources to the atmosphere. Coastal ecosystem C accumulation rates are greater than

open ocean rates because their primary producers differ. Open ocean phytoplankton have a

much lower net primary production (NPP) per unit area, have a greater percentage of their NPP

consumed by herbivores, and contain more easily decomposed OM than coastal macrophytes (

Duarte and Cebrian 1996; Cebrian 2002).

Local Rates

Mean global rates of C accumulation (2-2) were calculated using rates of C accumulation

from numerous local studies (Table 2-3). The majority of C accumulation rates were measured

in salt marshes while accumulation rates in seagrass beds were the least measured.

Accumulation rates in mangrove forests ranged from 33 to 841 g C m-2 -1l with the lowest rate









CHAPTER 5
SYNTHESIS

Coastal ecosystems including salt marshes, seagrass beds, and mangrove forests are more

effective carbon (C) sinks than terrestrial systems and freshwater wetlands (Chapter 2). These

ecosystems store large amounts of OC and actively accumulate OC at high rates; about 44.6 Pg

C is stored and about 120 Tg C yl accumulates in salt marsh and mangrove sediments globally

(Jennerjahn and Ittekkot 2002; Chmura et al. 2003). Many coastal ecosystems have been

degraded or lost due to anthropogenic disturbances (Valiela et al. 2001; Kennish 2002; Zedler

2004). Destruction of coastal ecosystems increases atmospheric CO2 COncentrations because

their organic C (OC) stores are often mineralized as a result and their future OC sequestration

capacity is lost (Duarte et al. 2005; Bridgham et al. 2006). Generally in the United States, the

destruction of coastal ecosystems must be mitigated by restoring or creating coastal ecosystems

elsewhere. Whether mitigation of seagrass and mangrove systems restores the C sink capacity is

currently not well-studied.

In the present study, functional traj ectories of sediment OC (SOC) parameters in a

constructed mangrove and seagrass system in the Indian River Lagoon, Florida were measured.

Sediment OC (SOC) parameters in constructed systems were also compared to mature reference

systems to indicate if constructed sediments had reached functional equivalence in terms of OC

storage. The objectives of this study were: 1) to determine short term trajectories of SOC pools

in a constructed mangrove forest and seagrass bed; 2) to compare SOC pools in the constructed

system with those in reference systems; 3) to compare the liability of SOC in the constructed and

reference systems; 4) to determine and compare significant sources to the total SOC pool in the

constructed and reference systems. The hypotheses were: 1) in the short term, storage in the

three OC pools studied would increase in the constructed systems, but would not reach the level









diagram (Fig. 4-7b). Macroalgae was also a dominant source for November reference floc

samples.

Ternary diagrams indicated that seston is a dominant source to almost all sediments and

surface layers regardless of site. Seston is not the only source, however, because all sediments

and surface layers are more enriched in 13C than seston (Fig. 4-3). A review of source

determination studies in mangrove and seagrass sediments found that seston was a dominant

source at 47% of sites (Chapter 2). OC in sediments of young mangrove forests were dominated

by algal and seston sources, just as the constructed sediments were in this study (Marchland et al.

2003; Alongi et al. 2004). In sediments with low %OC, as in this study (Chapter 3), the

dominant macrophytes such as mangroves or seagrass seemed less likely to be significant OC

sources (Gonnocea et al. 2004; Kennedy et al. 2004). Middelburg et al. (1997) showed a

significant relationship of decreasing 613C (mOre depleted than in situ macrophyte 613C) with

decreasing %SOC. These trends may be because when seston settles onto sediments, it does so

with inorganic particles, which dilute SOC, lowering the %OC.

Seston comes from a variety of sources as it is made up of phytoplankton, zooplankton,

bacteria, and detritus (Fig. 4-8). Its high 813C in this study is indicative of a mangrove or

terrestrial origin. Its high N:C (low C:N), however, indicates a mixture of phytoplanktom, which

have C:N ratios from 7.7 to 10.1 (Holligan et al. 1984), and bacterioplankton, which have C:N

ratios from 2.6 to 4.3 (Lee and Fuhrman 1987). Other estuarine studies similarly had seston with

low 613C and low C:N ratios (Hemminga et al. 1994; Cifuentes et al. 1996; Zhou et al. 2006).

Cifuentes et al. (1996) demonstrated that bacteria in the water column were likely immobilizing

N in the process of decomposing terrestrial-derived organic matter, which could lead to

incorporation of that nitrogen into organic matter during humification and a lower C:N ratio.









(McKee and Faulkner 2000; Alongi et al. 2001; Jennerj ohn and Ittekkot 2002; Alongi et al.

2004; Bouillon et al. 2004; Otero et al. 2006). The functional equivalence "bar" is therefore

lowest for seagrass sediments, which was true in this study where reference sediments' mean

%OC was 1.4 in mangroves and only 0.74 in seagrass. Lower than reported %OC values in this

study's reference mangrove sediments are likely due to their position around spoil islands--

mangrove reference sites, just as SL 15, began development in dredge spoil.

No known functional traj ectory studies have measured OC pools with short turnover

times like ExOC and MBC. These OC pools were the only pools to approach equivalence in

mangrove sediments. Because these pools are more active (Buyanovsky et al 1994; Rochette

and Gregorich 1998), they are likely to develop faster in sediments. Constructed and reference

sediments in this study had MBC that was about equal to greater than MBC in a North Sea tidal

flat, a Brazilian mangrove forest, and an arctic salt marsh (Joergensen and Mueller 1995; Otero

et al. 2006; Buckeridge and Jefferies 2007). Those other studies are the only known to measure

MBC via fumigation extraction in a marine environment. MBC measured by fumigation-

extraction has been found to correlate well with MBC measured by phospholipids fatty acid

(PLFA) analysis but not by DNA analysis or substrate-induced respiration (Bailey et al. 2002;

Leckie et al. 2004). A relationship between fumigated and extracted C and total PLFA

concentrations has been developed by Bailey et al. (2002).

CFEr,us, = 2.4(PLFAors) + 46.2 (3 -2)

In equation 3-2, CFE is the uncorrected flush of OC (ug C g-l soil) resulting from fumigation and

PLFAtotal (nmol g-l soil) is the total amount PLFA extracted from the soil. Multiplying the

results by the 2.22 CFE to MBC correction factor, MBC from this study was compared to MBC

in studies that used the PLFA method. Converted measurements of PLFA yielded MBC values











Table 3-2. Results of factorial ANOVAs comparing SL 15 and references. Sediments and surface layers of the mangrove and
seagrass systems were each run individually. BD=Bulk Density, ExOC=Extractable organic carbon, MBC=Microbial
biomass carbon, TOC= Total organic carbon, and TN= Total nitrogen. Concentration (conc) parameters are reported in mg
kg-1 dry soil, and storage parameters are reported in g m-2.
ANOVA Effect BD TOC TN C:N ExOC MBC Lability
(conc) (storage) (conc) (conc) (storage) (conc) (storage)
Sediment


Mangrove







Seagrass


Site *** *** ***
Month NS NS NS
Depth NS NS
Site*Month NS NS NS
Site*Depth NS NS NS
Month*Depth NS NS NS

Site *** ** **
Month NS NS NS
Depth *** *** ***
Site*Month NS NS NS
Site*Depth *** NS
Month*Depth NS NS NS


*** ** *** ***
NS NS NS NS
NS NS ** **
NS NS NS NS
* NS NS
NS NS NS NS

*** NS *** ***
** NS *** ***
*** NS *** ***
NS NS NS NS
* NS *** **
NS NS NS NS


** ** NS NS
NS NS NS NS
NS NS NS


** ***
*** ***
** ***
* *
NS NS
NS NS

*** ***
*** ***
*** ***
NS **
NS ***
NS NS


NS NS
NS NS
NS NS


Surface layers
Mangrove algae/litter Site *** ***
Month NS NS *
Site*Month NS NS NS


Seagrass floc Site NS NS NS NS NS NS NS NS
Month NS NS NS NS NS NS ** NS NS
Site*Month NS NS NS ** NS NS NS NS NS
For significance NS=not significant, p = or <0.05, **p < 0.01, ***p < 0.0001. Please see Appendix A for a table listing how these data were transformed prior
to running the factorial ANOVA.









SL 15 seagrass sediments had lower 613C than reference seagrass sediments (site effect,

p<0.0001, Tables 4-3 and 4-4), but SL 15 mangrove sediments had 813C Similar to reference

mangrove sediments (p=0.40, Tables 4-3 and 4-4). SL 15 floc was more depleted than reference

floc in July but more enriched than reference floc in November (month x site interaction,

p<0.0001, Table 4-3 and 4-4). SL 15 algal mat was more enriched than reference litter in both

months, but the difference was greater in November (month x site interaction, p<0.0001, Table

4-3 and 4-4). TOC (%) was generally lower in SL 15 sediments than references with the

exception of the SL 15 seagrass accreted layer and floc, which had similar TOC to the

reference's 0-5 cm depth and floc, respectively (Chapter 3). C:N ratios were similar in seagrass

sediments and floc but were lower in SL 15 mangrove sediments and surface layers than in

respective mangrove references (Chapter 3).

Source Determination

Putting source (plants and seston) and sediment 813C data together indicates potential

sources to the various sediments and surface layers (Fig. 4-3). Using observations from the field

and Fig. 4-3, the three ternary diagram end members for SL 15 mangrove sediment were seston,

algal mat, and terrestrial plants (Fig. 4-5a). Seston, litter, and mangroves were the end members

for reference mangrove sediments (Fig. 4-5b). Seston, seagrass, and macroalgae were the end

members for SL 15 and reference seagrass sediments and floc (Fig. 4-6 and 4-7b). Seston,

seagrass, and mangroves were the end members for reference mangrove litter (Fig. 4-7a). SL 15

algal mats did not need a diagram because they are their own source as primary producers.

Ternary diagrams explained 74% of SL 15 mangrove sediment samples, 92% of reference

mangrove sediment samples, 71% of SL 15 seagrass sediment samples, 33% of reference

seagrass sediment samples, 100% of reference mangrove litter samples, and 89% of seagrass floc












































Figure 4-5. N:C vs. 513C in ternary mixing diagrams of three potential OC sources and
mangrove sediments. Circles are the mean end member values and boxes are +
standard deviation ofN:C and 613C. Triangles are mangrove sediment values for SL
15 (A) and the reference (B).


I I I I I I I I I I I


0.30-



0.25-



0.20-



0.15-



0.10-



0.05-


15 Mangrove Sediments


~"",~~~~~................~~~~


terrestrial


0.00


-30 -28 -26 -24 -22 -20 -18 -16 -14 -12 -10









increase over 7 years (Havens et al. 2002). Sedimentation of mineral particles dilutes SOC

concentrations. Creek banks often have lower OC concentrations than the interior of marshes

(Craft et al. 2002) because they experience greater sedimentation of mineral particles (e.g.

Temmerman et al. 2003). Simenstad and Thom (1996) cited sedimentation as a reason why

SOM in a restored marsh did not increase with time.

Storing Carbon Versus Sinking Carbon

Even though most constructed marshes do not yet store the same amount of C as their

natural counterparts, they may still be acting as C sinks. A few studies examined OC

accumulation rates as well as SOC pools and found that OC accumulation rates in constructed

marshes are as high as or higher than rates in constructed marshes (Cammen 1975; Craft et al.

2002; Craft 2001). The mean OC accumulation rate of 8 different-aged constructed wetlands in

North Carolina was 42 g C m-2 -1l COmpared to 43 g C m-2 -1l in the reference wetlands, even

though the OC pools (g C m-2) in the constructed wetlands were significantly lower (Craft et al.

2003). Additionally, some young marshes have high sedimentation rates (Morgan and Short

2002). Sedimentation may encourage OC accumulation while reducing SOC concentrations

resulting in a reciprocal relationship as was demonstrated in Bay of Fundy marshes (Connor et

al. 2001). High sedimentation rates may have prevented SOC pools from increasing in the

Tacoma and San Diego constructed marshes while encouraging OC accumulation, which

unfortunately was not measured in those studies.

New Directions

The extensive studies on coastal marsh functional traj ectories have been broad in scope

and therefore unable to examine OC dynamics in constructed marshes with sufficient detail.

SOC is a conglomeration of pools that include a labile pool, a slowly oxidized pool, a very

slowly oxidized pool, and a recalcitrant pool (Eswaran et al. 1995). The pool matters, as the one










1.00


0.50


r Mangrove
*Seagrass


o.oo r


-0.50


-1.00


TOC (S)


MBC (S) ExOC (S)


Figure 5-3. Recovery indices of three organic carbon (OC) pools and OC liability parameters for
constructed mangrove and seagrass systems. For each system, OC pools are summed
for all sediment depths and surface layers and liability is the average of all depths and
surface layers. TOC=total organic carbon, MBC=microbial biomass carbon,
ExOC=extractable organic carbon, and (S) indicates that these OC pool parameters
were calculated on a storage basis.


Lability









Given the limited scope of these studies, many questions remain unexplored. First, the

maj ority of studies on restored coastal systems have been performed in temperate salt and

brackish marshes. Second, these studies only measured SOC or sediment OM as one of a suite

of variables and did not deeply examine various SOC pools or characteristics. Third, these

studies only examined long term trends and not short term changes that may occur immediately

following construction of an ecosystem.

Whether constructed mangrove and seagrass ecosystems provide the same ecological

services as their natural counterparts with respect to the C sink, and if the restoration of this

service follows a functional trajectory is currently unknown. In this study, OC storage in a

constructed seagrass and mangrove system in the Indian River Lagoon, FL was examined and its

OC storage functioning was compared with the functioning of adj acent mature systems. Specific

objectives were to: 1) determine whether extractable OC, microbial biomass C, total OC pools,

and OC liability follow a short term traj ectory in sediments of a constructed mangrove forest and

seagrass bed and 2) evaluate whether the constructed system has reached functional equivalence

by comparing SOC between constructed and natural systems. We hypothesized that, in the short

term, SOC storage would increase in the constructed system but would not reach the level of

SOC storage in natural systems.

Methods

Study Site

SL 15 (Fig. 3-1) is a mitigation site located in the Indian River Lagoon (IRL) adj acent to

Fort Pierce, Florida. It is one of many spoil islands created in the IRL during the construction of

the Atlantic Intracoastal Waterway that sit several meters above sea level. These islands are

populated by many exotics, such as Australian Pine (Casuvina ca~suvina) and Brazilian Pepper

(.\hullnt terebinthifolius), in their interiors and by native red, black, and white mangroves






























mangj ve 1 ......... --


0.22-

0.20-

0.18-

0.16-

0.14-

0.12-

0.10-

0.08-

0.06-

0.04 -

0.02
-30


referencee Mangrove Sediments


-28 -26 -24 -22 -20 -18


-16


813C


Figure 4-5. Continued











Table 4-4. Mean 613C and C:N (+ SE) for sediments and surface layers of SL 15 and reference
mangrove and seagrass systems.


System and month Depth


C:N (molar)
Reference SL 15 Reference


SL 15
-15.72 (0.46)
-23.57 (1.03)
-23.70 (0.50)

-11.96 (0.47)
-21.96 (0.18)
-24.11 (0.46)

-21.05 (0.14)
-21.00 (0.44)
-21.03 (0.35)
-20.93 (0.60)


-19.50 (0.15)
-20.06 (0.20)
-21.55 (0.41)
-20.80 (0.48)


July mangrove


Algal Mat/ Litter
0-5 cm
5-10 cm

Algal Mat/ Litter
0-5 cm
5-10 cm


-18.21 (0.63)
-24.12 (0.91)
-22.17 (1.40)

-24.43 (1.56)
-23.42 (0.70)
-24.10 (0.31)

-18.93 (0.27)

-18.97 (0.29)
-19.11 (0.20)
-18.79 (0.33)

-24.46 (0.56)

-17.87 (0.21)
-18.80 (0.46)
-19.10 (0.48)


8.2 (0.2)
5.0 (1)
8.0 (2)

8.1 (0.9)
5.5 (0.3)
6.2 (0.8)

8.2 (1)
7.6 (1)
6.9 (0.3)
7.0 (1)


6.8 (0.1)
6.1 (0.4)
6.7 (0.8)
7.3 (1)


9.3 (0.3)
9.6 (3)
9.2 (1)

21 (5)

8.8 (0.9)

5.7 (0.6)

6.7 (0.2)
6.6 (0.2)
1.0 (0.2)

10.8 (0.6)

5.4 (0.2)
5.6 (0.4)
6.1 (0.4)


Nov. mangrove


July seagrass





Nov. seagrass


Floc
Accreted
0-5 cm
5-10 cm
10-15 cm

Floc
Accreted
0-5 cm
5-10 cm
10-15 cm











Table 2-5. Continued


Source
8' C
-13.3
-25.3


-12.7
-25.5

-18.4
-10.0
-12.6
-26.0

-12.1
-26.9


-12.5'
-25.5'


Main
sources
both
both
Both

6


2
2




5
5 (50%)
5 (40%)

5
6


How main sources
determined
Comparison and
distributions of long
chain n-Alkanes

Compared to curve of
2 source mixing model

Comparison with lipid
distributions and lipid
8 3C


Mixing model using
compound specific
61 C

Compared to curve of
2 source mixing model


Data
source
8


Location
Plum Island,
Massachusetts


Potential sources
S. 41terniflora
T. latifolia


Spartina
Allochthonous OM

Seston
Seagrass
Spartina
J. roeinerianus

Spartina anglica
Pucinella inaritina



Spartina
Allochthonous OM


Site description
Mid marsh
Upper marsh
Mudflat


Waarde Marsh,
Netherlands

Cape Lookout Bight,
North Carolina




Dorset,
United Kingdom


B\ arnstable,
Massachusetts


Marsh


Sediment
61 C
-18.9
-22.81
-19.39

-22 to
-24.6

-17.8
-20.3




-17.6
-21.4
-20.4

-13.4 to
-14.5
-21 to
-19.5


Fall
Spring



S. anglica
P. inaritina
Mudflat

High marsh

Low marsh


Seagrass beds

Gazi Bay, Kenya


Seagrass
Mangroves
Sediment Traps


-19.7
-26.75
-23.3

-18.3
-26.75
-22.5

-15.8
-26.75
-19.2

-10.70
-26.75
-13.7


Closest to mangroves


-22.9


4,1 Comparison


Gazi Bay, Kenya


Seagrass
Mangroves
POM

Sea grass
Mangroves
POM

Seagrass
Mangroves
POM


Closer to mangroves


-20.6


4,1 Comparison


Gazi Bay, Kenya




Gazi Bay, Kenya


Farther from mangroves




Farthest from mangroves


-18.5




-15.14


Comparison




Comparison










APPENDIX A
STATISTICAL TRANSFORMATIONS

Table A-1. How data were transformed to meet the normality assumption prior to running
ANOVAs. For parameters, TOC=total organic carbon, TN= total nitrogen, C:N= carbon to
nitrogen ratio, ExOC=extractable organic carbon, and MBC=microbial biomass carbon. For
transformations, NT= not transformed, Sqrt=square root, and a C (as in X-C) indicates that a
constant was subtracted or added to a parameter before it was transformed via square root, log,
arcsmne, etc.

Chapter ANOVA Depths Parameter Transformation


3 Mangrove factorial Sediments 0-10


Bulk density
TOC (conc)
TOC (storage)
TN (conc)
C:N
ExOC (conc)
ExOC (storage)
MBC (conc)
MBC (storage)
Lability
Bulk density
TOC (conc)
TOC (storage)
TN (conc)
C:N
ExOC (conc)
ExOC (storage)
MBC (conc)
MBC (storage)
Lability
Bulk density
TOC (conc)
TOC (storage)
TN (conc)
C:N
ExOC (conc)
ExOC (storage)
MBC (conc)
MBC (storage)
Lability
Bulk density
TOC (conc)
TOC (storage)
TN (conc)
C:N
ExOC (conc)
ExOC (storage)
MBC (conc)
MBC (storage)
Lability
Bulk density
TOC (conc)
ExOC (conc)


NT
Log 10 (Arcsin)
Log 10
Log 10 (Arcsin)
Sqrt
Log e
Log e
Sqrt (MIBC-C)
Log e
Log 10
NT
NT
NT
NT
Log e
Log e(ExOC-C)
Log 10
Sqrt (MIBC-C)
NT
Log 10
Sqrt
Log 10 (Arcsin)
Log 10
Log 10 (Arcsin)
Log e(C:N-C)
Log e(ExOC-C)
Log 10
Sqrt
Sqrt
Log e
NT
Arcsin (sqrt)
NT
Sqrt
Sqrt
Sqrt
Sqrt
Sqrt
Log e
NT
Log e
Sqrt
NT


Seagrass factorial


Sediments 1-3


Mangrove factorial Surface layers









Seagrass factorial Surface layers









Mangrove repeated Sediments 0-10
Measures









Overall, during the first year following construction, with the exception of the mangrove

algal mat, OC changes in SL 15 are due to seasonality and water quality. These seasonality-

caused changes are large and may obscure any changes due to increasing functions. High

interannual variability that mask directional changes has been observed in a restored California

salt marsh (Zedler and Callaway 1999). SL 15 changes were greatest in ExOC and MBC, pools

with fast turnover rates. One year may not be ample time to observe changes in more stable OC

pools like TOC.

Constructed and Reference Equivalence

Organic carbon pools

A lack of traj ectories does not preclude OC on SL 15 from being functionally equivalent

to reference OC. Examining depths separately, 0-5 cm SL 15 mangrove sediments approached

functional equivalence on a storage basis for ExOC and MBC (Fig. 3-7 and 3-8). Most depths of

SL 15 seagrass sediments were at or exceeded functional equivalence for all OC pools on a

storage basis (Fig. 3-7 and 3-8). The reason for this equivalence was bulk density. Because bulk

density of SL 15 0-10 cm sediments is greater than reference sediments, when OC concentrations

are multiplied by bulk density in order to be reported on a storage basis, the resulting parameters

in SL 15 are often greater than or equal to the resulting parameters in reference sediments.

Accreted layers were an exception because their bulk densities were the same as the references'

and their heights were usually less than the references' 5 cm.

TOC equivalence did not occur on a concentration or a storage basis in the mangrove

sediments but occurred for accreted and 0-5 cm depths in seagrass sediments. Accreted layers

reached equivalence because the material accumulating from the water column is likely the same

material being trapped by seagrasses in reference sediments. It is odd, at first, that 5-10 cm

depths reached equivalence before 0-5 cm depths because inputs of OC to SL 15 sediments were









A study of different-aged New England salt marshes found that SOM increased steadily from 2%

at a 1-year-old site to 15% at a 15-year-old site (Morgan and Short 2002). Studies from the

western coast did not find strong directional trends of SOM over time. In Tacoma, Washington

SOM stayed between 2-4% over 5 years (Simenstad and Thom 1996) and in San Diego,

California only a slight increase of 3% was found over 11 years (Zedler and Calloway 1999).

These differences in traj ectories may be more a case of land use than geography. Both of the

west coast studies took place in large urban areas, whereas the east coast studies took place in a

variety of locales, none as developed as Tacoma and San Diego.

Only two studies documented tidal marshes that reached functional equivalence with their

natural references in terms of SOC. The tidal marshes were 25 (Craft et al. 1999; Craft et al.

2003) and 42 (Craft 2001) years old. Both these marshes are located in the southeastern U. S.

These marshes achieved functional equivalence possibly because they, or their references,

differed from most of the marshes studied. The 42-year-old marsh differed because it was a

restored marsh and not a marsh constructed from dredge spoil. Instead, it had been disturbed by

a dike that prevented tidal inundation but was removed after only 8 years (Craft 2001). The 25-

year-old marsh differed because its reference marsh was a relatively new natural salt marsh,

which contrasted to other reference marshes that are greater than 2,500 years old (Craft et al.

1999). Because the reference was relatively young, its soils resembled spoil (90% sand) more

than a histosol (>10% OM). They were mineral entisols with a high bulk density and low OC

content (<1.4%). Reference marshes can determine whether or not a constructed marsh reaches

functional equivalence because their mean attribute values represent functional equivalence

"finish lines."











Table 3-7. Organic carbon accumulation rates in mangrove and seagrass systems in this and
other studies.
Rate
System (g OC no y-') Location and remarks Sources
Seagrass 195 Florida, USA This study
40-65 Mexico 1
19-191 Spain 2
182 Spain 3

Mangrove -189 Florida, USA: sediment of 1-year This study
old planted system
120 Florida, USA: above system with This study
algal mat included
180 Australia 4
168-841 China 5
105-159 Florida, USA 6
191-3281 Florida, USA 7
101-127 Malaysia 8
33-104 Mexico 1
184-281 Thailand 9
a 1, Gonneea et al. 2004; 2, Romero et al. 1994: 3, Gacia et al 2002; 4, Brunskill et al. 2002; 5, Alongi et al. 2005; 6,
Callaway et al. 1997: 7, Cahoon and Lynch 1997: 8, Alongi et al. 2004; 9, Alongi et al. 2001
'This author reported organic matter accumulation rates, so rates were divided by 2 to obtain these numbers.










CANUEL, E. A., K. H. FREEMAN, AND S. G. WAKEHAM. 1997. Isotopic compositions of
lipid biomarker compounds in estuarine plants and surface sediments. Limnol. Oceanogr.
42: 1570-1583.

CAPONE, D. G., AND R. P. KIENE. 1988. Comparison of microbial dynamics in marine and
fresh-water sediments contrasts in anaerobic carbon catabolism. Limnol. Oceanogr. 33:
725-749.

CEBRIAN, J. 2002. Variability and control of carbon consumption, export, and accumulation in
marine communities. Limnol. Oceanogr. 47: 11-22.

------, M. F. PEDERSEN, K. D. KROEGER, AND I. VALIELA. 2000. Fate of production of the
seagrass Cymodocea nodosa in different stages of meadow formation. Mar. Ecol.-Prog.
Ser. 204: 119-130.

CHAPIN, F. S., P. A. MATSON, AND H. A. MOONEY. 2002. Principles of Terrestrial
Ecosystem Ecology. Springer Science.

CHMURA, G. L., P. AHARON, R. A. SOCKI, AND R. ABERNETHY. 1987. An inventory of
C-13 abundances in coastal wetlands of Louisiana, USA Vegetation And Sediments.
Oecologia 74: 264-271.

------, S. C. ANISFELD, D. R. CAHOON, AND J. C. LYNCH. 2003. Global carbon
sequestration in tidal, saline wetland soils. Glob. Biogeochem. Cycle 17:1111.

------, L. L. HELMER, C. B. BEECHER, AND E. M. SUNDERLAND. 2001. Historical rates of
salt marsh accretion on the outer Bay of Fundy. Canadian Journal Of Earth Sciences 38:
1081-1092.

CHOI, Y. H., AND Y. WANG. 2004. Dynamics of carbon sequestration in a coastal wetland
using radiocarbon measurements. Glob. Biogeochem. Cycle 18.

CIFUENTES, L. A., R. B. COFFIN, L. SOLORZANO, W. CARDENAS, J. ESPINOZA, AND
R. R. TWILLEY. 1996. Isotopic and elemental variations of carbon and nitrogen in a
mangrove estuary. Estuar. Coast. Shelf Sci. 43: 781-800.

CONNOR, R. F., G. L. CHMUJRA, AND C. B. BEECHER. 2001. Carbon accumulation in Bay
of Fundy salt marshes: Implications for restoration of reclaimed marshes. Glob.
Biogeochem. Cycle 15: 943-954.

CORDOVA-KREYLOS, A. L. AND OTHERS. 2006. Diversity, composition, and geographical
distribution of microbial communities in California salt marsh sediments. Applied And
Environmental Microbiology 72: 3357-3366.

CRAFT, C. 2001. Soil organic carbon, nitrogen, and phosphorus as indicators of recovery in
restored Spartina marshes. Ecological Restoration 19: 87-91.










------, S. BROOME, AND C. CAMPBELL. 2002. Fifteen years of vegetation and soil
development after brackish-water marsh creation. Restor. Ecol. 10: 248-258.

------, AND OTHERS. 2003. The pace of ecosystem development of constructed Spartina
alterniflora marshes. Ecol. Appl. 13: 1417-1432.

------, READER, J. N. SACCO, AND S. W. BROOME. 1999. Twenty-five years of ecosystem
development of constructed Spartina alterniflora (Loisel) marshes. Ecol. Appl. 9:
1405-1419.

------, E. D. SENECA, AND S. W. BROOME. 1993. Vertical accretion in microtidal regularly
and irregularly flooded estuarine marshes. Estuar. Coast. Shelf Sci. 37: 371-386.

DABY, D. 2003. Some quantitative aspects of seagrass ecology in a coastal lagoon of Mauritius.
Marine Biology 142: 193-203.

DAHL, T. E. 1990. Wetland losses in the United States 1780's to 1980's. F. a. W. S. U. S.
Department of the Interioir, Washington, D.C. [ed.].

DAI, J. H., M. Y. SUN, R. A. CULP, AND J. E. NOAKES. 2005. Changes in chemical and
isotopic signatures of plant materials during degradation: Implication for assessing various
organic inputs in estuarine systems. Geophysical Research Letters 32.

DAUBY, P. 1989. The stable carbon isotope ratios in benthic food webs of the Gulf of Calvi,
Corsica. Cont. Shelf Res. 9: 181-195.

DITTMAR, T., R. J. LARA, AND G. KATTNER. 2001. River or mangrove? Tracing major
organic matter sources in tropical Brazilian coastal waters. Mar. Chem. 73: 253-271.

DUARTE, C. M. 2002. The future of seagrass meadows. Environmental Conservation 29:
192-206.

------, AND J. CEBRIAN. 1996. The fate of marine autotrophic production. Limnol. Oceanogr.
41: 1758-1766.

------, AND C. L. CHISCANO. 1999. Seagrass biomass and production: a reassessment. Aquat.
Bot. 65: 159-174.

------, J. J. MIDDELBURG, AND N. CARACO. 2005. Major role of marine vegetation on the
oceanic carbon cycle. Biogeosciences 2: 1-8.

EVANS, N. T., AND F. T. SHORT. 2005. Functional trajectory models for assessment of
transplanted eelgrass, Zostera marina L., in the Great Bay Estuary, New Hampshire.
Estuaries 28: 936-947.

FENTON, G. E., AND D. A. RITZ. 1988. Changes in carbon and hydrogen stable isotope ratios
of macroalgae and seagrass during decomposition. Estuar. Coast. Shelf Sci. 26: 429-436.










LIST OF FIGURES


Figure page

1-1 The carbon cycle in seagrass beds. ............. ...............18.....

1-2 The carbon cycle in mangrove forests. ............. ...............19.....

3-1 The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset) ...............97

3-2 Core from SL 15 seagrass system illustrating the surface layer (floc) and different
sediment depths accretedd layer, 0-5 cm, 5-10 cm) .............. ...............98....

3-3 The functional traj ectory the bulk density of SL 15 mangrove sediments followed
over the first year after construction ..........._ .....___ ...............99.

3-4 The changes in organic carbon parameters over the first year after construction in SL
15 seagrass and mangrove sediments .............. ...............100....

3-5 The changes in organic carbon parameters over the first year after construction in SL
15 seagrass and mangrove surface layers. ............. ...............101....

3-6 Comparisons between total organic carbon (TOC) in reference and SL 15 mangrove
(top) and seagrass (bottom) sediments ................. ...............102........... ...

3-7 Comparisons between extractable organic carbon (ExOC) in reference and SL 15
mangrove (top) and seagrass (bottom) sediments ................. ............... ......... ...103

3-8 Comparisons between microbial biomass carbon (MBC) in reference and SL 15
mangrove (top) and seagrass (bottom) sediments ................. ............... ......... ...104

3-9 Organic carbon (OC) pools in SL 15 and reference mangrove and seagrass
sedim ents ........... __... ......... ...............105....

4-1 The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset) .............133

4-2 Core from SL 15 seagrass system illustrating the surface layer (floc) and different
sediment depths accretedd layer, 0-5 cm, and 5-10 cm) .............. .....................3

4-3 613C averaged over July and November 2006 for SL 15 and reference sediments and
surface layers compared to mean 613C Of potential sources. ................ ................. .. 135

4-4 Mean 613C Of SL 15 sediments and surface layers over the first year after
con strcti on........... ..... ..___ ...............136....

4-5 N:C vs. 813C in ternary mixing diagrams of three potential OC sources and
mangrove sediments............... ...............13










MEAD, R., Y. P. XU, J. CHONG, AND R. JAFFE. 2005. Sediment and soil organic matter
source assessment as revealed by the molecular distribution and carbon isotopic
composition of n-alkanes. Org. Geochem. 36: 363-370.

MEWS, M., M. ZIMMER, AND D. E. JELINSKI. 2006. Species-specific decomposition rates of
beach-cast wrack in Barkley Sound, British Columbia, Canada. Mar. Ecol.-Prog. Ser. 328:
155-160.

MIDDELBURG, J. J., J. NIEUWENHUIZE, R. K. LUBBERTS, AND O. VAN DE
PLASSCHE. 1997. Organic carbon isotope systematics of coastal marshes. Estuar. Coast.
Shelf Sci. 45: 681-687.

MISEROCCHI, S., L. LANGONE, AND T. TESI. 2007. Content and isotopic composition of
organic carbon within a flood layer in the Po River prodelta (Adriatic Sea). Cont. Shelf
Res. 27: 338-358.

MITSCH, W. J., AND J. G. GOSSELINK. 2000. Wetlands, 3rd ed. John Wiley and Sons.

MOORE, T. N., AND P. G. FAIRWEATHER. 2006. Decay of multiple species of seagrass
detritus is dominated by species identity, with an important influence of mixing litters.
Oikos 114: 329-337.

MORGAN, P. A., AND F. T. SHORT. 2002. Using functional trajectories to track constructed
salt marsh development in the Great Bay Estuary, Maine/New Hampshire, USA. Restor.
Ecol. 10: 461-473.

MORIARTY, D. J. W. AND OTHERS. 1985. Microbial biomass and productivity in seagrass
beds. Geomicrobiology Journal 4: 21-51.

MORRIS, J. T., AND P. M. BRADLEY. 1999. Effects of nutrient loading on the carbon balance
of coastal wetland sediments. Limnol. Oceanogr. 44: 699-702.

MOSER, M., C. PRENTICE, AND S. FRAZIER. 1996. A global overview of wetland loss and
degradation. RAMSAR, 6th Meeting of the Conference of Contracting Parties.

MOY, L. D., AND L. A. LEVIN. 1991. Are spartina marshes a replaceable resource a
functional-approach to evaluation of marsh creation efforts. Estuaries 14: 1-16.

MUZUKA, A. N. N., AND J. P. SHUNULA. 2006. Stable isotope compositions of organic
carbon and nitrogen of two mangrove stands along the Tanzanian coastal zone. Estuar.
Coast. Shelf Sci. 66: 447-458.

NIEUWENHUIZE, J., Y. E. M. MAAS, AND J. J. MIDDELBURG. 1994. Rapid analysis of
organic-carbon and nitrogen in particulate materials. Mar. Chem. 45: 217-224.

ONG, J. E. 1993. Mangroves--a carbon source and sink. Chemosphere 27: 1097-1107.









long amount of time as a well-established mangrove system whose main OC sources are the

more recalcitrant leaves and roots of mangroves.

There are a myriad of methods researchers utilize to determine OC sources. The most

widely used method measures bulk stable isotopes (usually 13C and 15N) in possible sources and

sediments. Bulk analyses measure isotopic signatures of entire OC pools in sediments or of

whole plant parts. Sources are then determined by a simple comparison of source and sediment

isotopic signatures (Haines 1976; Hemminga et al. 1994; Jennerjahn and Ittekkot 2002; Thimdee

et al. 2003) or by mixing models (Dauby 1989; Kennedy et al. 2004; Papadimitriou et al. 2005;

Zhou et al. 2006;). Other parameters are used with isotopic signatures to determine sources

using ternary diagrams of N:C ratios plotted against 813C (Gonnocea et al. 2004; Miserocchi et

al. 2007) or more complex mixing models using 613C and biomass or %OC as parameters

(Chmura et al. 1987; Middelburg et al. 1997; Bouillon et al. 2003;). Sources must have

consistently distinct stable isotopic signatures for this method to be useful (Papadimitriou et al.

2005). Lipids are also used as biomarkers to determine OC sources (Wang et al. 2003). The

lipids, generally sterols, fatty acids, or hydrocarbons, vary in specificity as some can identify

groups of organisms such as vascular plants or algae while others may be specific to one genera

or species (Canuel et al. 1997). Finer resolution of sources is possible when the isotopic

signatures of lipids are measured in compound specific stable isotope analyses (Canuel et al.

1997; Bull et al. 1999; Hernandez et al. 2001; Mead et al. 2005). Some lesser-used methods

involve comparing relative amounts of certain OC structures in the soil, either visually as in

petrographic analysis (Lallier-Verges et al. 1998; Marchland et al. 2003) or chemically as in

nuclear magnetic resonance spectroscopy (Golding et al. 2004).









sediments, SL 15 0-5 cm depths had similar MBC to reference 5-10 cm depths on a storage basis

(Fig. 3-8a; one-way ANOVA, df=3, p<0.0001). On a concentration basis, depths two and three

of SL 15 seagrass sediments had significantly lower MBC than those depths in reference

sediments, while depth one MBC was similar across sites (Fig. 3-8b; one-way ANOVA, df=5,

p<0.0001). On a storage basis, depths two and three had similar MBC across sites, but depth one

had significantly lower MBC in SL 15 (site x depth interaction, p<0.0001, Table 2; Fig. 3-8b).

MBC was significantly greater in November than in July for both mangrove and seagrass

sediments (month effect, p<0.0009 Table 2; Table 5). MBC was significantly greater in upper

depths of both mangrove and seagrass sediments (depth effect, p<0.0066, Table 2; Table 5 and

6). Surface layers had similar MBC to respective references (Table 2; Table 5).

SL 15 systems had significantly greater OC liability than reference systems in all sediments

and surface layers except for floc (site effect, p<0.013, Table 2; Table 6). Only depth one in

seagrass sediments was similar across sites (site x depth interaction, p<0.0001, Table 2; Table 6).

In mangrove sediments, the 0-5 cm depth had significantly greater liability than the 5-10 cm

depth while in seagrass sediments, depth two had the greatest liability (depth effect, p<0.0027,

Table 2; Table 6). In mangrove surface layers, liability of the SL 15 algal mat increased while

liability of reference litter decreased from July to November (site x month interaction, p<0.0001,

Table 2; Table 6).

Organic Carbon Accumulation Rates

OC accumulation rates in SL 15 sediments were between 168 to 231 g OC m-2 -1l in the

seagrass sediments, but were between -119 to -148 g OC m-2 -1l in the mangrove sediments.

When algal mat accumulations were added to mangrove sediments accumulations, rates ranged

from 29 to 236 g OC m-2 -1l. Floc OC accumulations were not added to seagrass sediments due

to the transient nature of floc, which is easily swept away by currents.












].0 SL 15 Seagrass Sediments


0.25 -



0.20 -


a
r
r


0.15


0.10-



0.05-



0.00-
-30


-25


-20


-15


-10


813C



Figure 4-6. N:C vs. 813C in ternary mixing diagrams of three potential OC sources and seagrass
sediments. Circles are the mean end member values and boxes are + standard
deviation of N:C and 613C. Filled triangles are 0-10 cm sediment values for SL 15
(A) and 0-15 cm sediment values for reference (B). Open triangles are accreted layer
values for SL 15 (A).

















































Sediments 0-10
Sediments 1-3
Surface layers
Surface layers
Sediments 0-10
Algal mat
Sediments 0-10
Sediments accreted


Parameter

TN (conc)
C:N
Lability
Bulk density
TOC (conc)
ExOC (conc)
MBC (conc)
TN (conc)
C:N
Lability
Bulk density
TOC (conc)
ExOC (conc)
MBC (conc)
TN (conc)
C:N
Lability
Bulk density
TOC (conc)
ExOC (conc)
MBC (conc)
TN (conc)
C:N
Lability
Bulk density
TOC (conc)
ExOC (conc)
MBC (conc)
TN (conc)
C:N
Lability
Ac e
Ac e
Ac e
Ac e
Ac e
Ac e
Ac e
Ac e
Ac e


Transformation


Table A-1. Continued

Chapter ANOVA Depths
3


Sqrt
Log e
Sqrt
NT
NT
NT
Sqrt (MIBC-C)
Log e (TN-C)
Log e
Log e
Log e
Sqrt
NT
NT
Sqrt
Sqrt


Algal mat


Seagrass repeated
Measures


Sediments 0-10


Sediments accreted







Floc


NT
Log e
Sqrt (ExOC-C)
Sqrt
Log e
Log e(C:N-C)
NT
Log 10
NT
Sqrt (ExOC-C)
NT
NT
Log e(C:N-C)
NT
NT
NT
NT
Log 10 (8'C*-1)
NT
NT
NT
NT
NT


4 Mangrove factorial
Seagrass factorial
Mangrove factorial
Seagrass factorial
Mangrove repeated
measures
Seagrass repeated
measures









SEDIMENT ORGANIC CARBON POOLS AND SOURCES INT A RECENTLY
CONSTRUCTED MANGROVE AND SEAGRASS ECOSYSTEM




















By

CAITLINT E HICKS


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2007









0.047 g C g-l OC hr-l (C)
0.036 g C g-l OC hr-l (R)


Water Column


Litter Decomposition
k=0.0049 0.046 d-l (Seagrass)
k=0.019 0.070 d-l (Algae)


Figure 5-1. A modified seagrass bed carbon cycle showing values from this study in constructed
(C) and reference (R) systems. Organic carbon (OC) pools are the sum of sediment
and surface layer means (July and November data). Rates of microbial carbon
respiration are the mean of all depths (sediment and surface layers) in July and
November, adjusted from an Oz uptake rate to a carbon release rate by an assumed 1
02 to 6 C molar ratio. Bolded words are the main contributors to sediment OC pools.









such as Australian Pine (Casuvina ca~suvina) and Brazilian Pepper (Malrln terebinthifolius), in

their interiors and by native red, black, and white mangroves (Rhizophora mangle, Avicennia

germinans, and Laguncularia racemosa) around their edges. To mitigate destruction of a nearby

mangrove forest and seagrass bed, seagrass and mangrove systems were created on SL 15.

These systems were created by burning and removing interior vegetation and removing dredge

spoil down to several different elevations. The seagrass bed, which remains submerged during

low tide, is at the lowest elevation, the mangrove forest, which is exposed at low tide, is at the

middle elevation, and at the highest elevation, above sea level, is a maritime forest. The

mangrove fringe of SL 15 was left intact except for a few flushing channels. In between the

constructed seagrass and mangrove systems a thin Spartina alterniflora buffer was planted. The

mangrove forest was planted with R. mangle, and maritime forests were planted with Coccoloba

uvifera, Borrichia frutescens, Rapanea guinensis, Conocarpus erectus, and Distichlis spicata,

but seagrasses were left to colonize naturally. Natural systems near SL 15 include its original

mangrove forest fringe, surrounding seagrass beds, and mangrove fringes of adj acent spoil

islands, which are at least 40 years old.

Litter Bags

Plant material from Syringodeum fihforme, Thala~ssia testudinum, Halodule beaudettei,

Acanthophora spicifera, Salrga~ssum spp, A. germinans, and R. mangle were collected in July

2006. Living seagrass fronds were taken from the beds around SL 15, which is similar to the

material ripped off by wind and wave events (Moore and Fairweather 2006). Clumps of live

macroalgae were taken from the subtidal areas in and around SL 15. Yellow mangrove leaves,

the kind about to fall, were taken from trees on the edge of SL 15 and surrounding islands. Plant

material was transported back to the laboratory and rinsed. Epiphytes were removed from

seagrass fronds and macroalgae. Plant material was then air dried for several weeks before being









bulk isotopic studies and their inherent weaknesses. They measured 813C and 15N isotopic

signatures of the top 2cm of fine fraction (>63um) sediments and of potential sources-seston

(assumed to represent phytoplankton), above- and below-ground seagrass tissues, and epiphytes.

613C ValUeS of the sediments ranged from -15.8 %o to -21.5 %o and average 613C ValUeS of seston,

epiphytes, below-ground seagrass tissues, and above-ground tissues were -22.1 %o, -17.8 %o, -

12. 1 %o, and -12.6 %o, respectively. No systematic differences in the 1N values of the potential

sources were found; most likely because discrimination against different N isotopes is not due to

physiology of primary producers and because N is often a limiting nutrient. At all sites, SOC

was more depleted isotopically than seagrass tissues but less depleted than seston. Using a

mixing equation based on one developed by Dauby (1989), they were able to find a range of

fractional contribution values of each source.

13Csedimen t s feston 13 season + teiperphrtes1 3 ep~liphrte s s f~ egra~/ss 13 seagra/ss (2- 1)

In equation 2-1, J is the unknown proportion of the OC from source i in the SOC pool and 813C

is the isotopic signature of source i. This equation is used to find the range of values for each

source needed to satisfy the equation and equal the sediment 813C value. With this model, they

were able to determine which sites had seston as the maj or contributor to SOC and which sites

had seagrass as the major contributor to SOC. If this were simply a two end member system

involving seagrasses and seston, the elucidation of sources to the sediments using this model

would have been straightforward. But these sites also included epiphytes, and their intermediate

813C Signature made it impossible for the model to assign them reasonable contribution ranges

(often the ranges included a 0% contribution). Thus the relative contribution of epiphytes to

SOC could not be determined by bulk isotopic methods alone.









assumed to be a mixture of all three sources, samples that fell along a line connecting two end-

members were considered a mixture of those two sources, samples that fell near one end member

were assumed to have OC predominantly from that source, and samples that fell outside the

triangle were assumed to have OC contributions from additional sources. This method is limited

to systems with three main sources. Conclusions of research, like Papadimitriou et al.'s (2005)

study of seagrass, seston, and epiphytes, could have benefited from this method if had they

measured C:N ratios.

Comparisons of actual values to values from a predicted model can sometimes help

elucidate sources better than a mixing equation based on the actual data. These models are based

on biomass of potential sources (Chmura et al. 1987), primary productivity (Bull et al. 1999), or

%SOC (Middelburg et al. 1997). The problem with these models is that they assume sources

contribute to SOM in the same relative proportions as their biomass/productivity Thi s

assumption may not be correct because sources differ in their degrees of liability, in their

litterfall, and in the amount of their biomass that is exported out of the system. However, models

are good approximations, especially in more peaty coastal wetlands where sediments have high

OC and sedimentation of allocthonous OC inputs is minimal.

For the above methods of calculating SOM sources from isotope values, parameters other

than the 613C ValUeS, such as biomass or C:N ratios, are needed. These other parameters also

support conclusions based on isotopic values alone. Many studies combine C:N ratios with

isotopic measurements ( Middelburg et al. 1997; Bouillon et al. 2003; Soto-Jimenez et al. 2003;

Thimdee et al. 2003; Gonneea et al. 2004). Correlations between C:N ratios or %SOC and 813C

values are used to assist in determining SOC sources. A mild relationship (R2 = 0.26) was found

to exist between 613C ValUeS and C:N ratios in a Mexican salt marsh where less negative 613C












TABLE OF CONTENTS


page

ACKNOWLEDGMENT S .............. ...............4.....


LI ST OF T ABLE S ........._..... ...............7..._.........


LIST OF FIGURES .............. ...............9.....


AB S TRAC T ............._. .......... ..............._ 1 1..


CHAPTER


1 INTRODUCTION ................. ...............13.......... ......


2 LITERATURE REVIEW ................. ...............20................


Introducti on .................. .. .... ........ ...............20......
Rates of Organic Carbon Sequestration ................. ...............21........... ...
Global Rates .............. ...............22....
Local Rates ..................... ... ..................2
Measuring Rates of Carbon Accumulation ................ ...... ..... .. ...........2
Comparing Organic Carbon in Restored and Reference Coastal Marshes.............................26
Monitoring Constructed Coastal Marshes Using Functional Traj ectories ................... ...28
Functional Trajectory Case Studies............... ...............29
Factors Affecting Functional Equivalence ................ ...............31................
Storing Carbon Versus Sinking Carbon .............. ...............32....
New Directions ................. .......... ... .......... .............3
Sediment Organic Carbon Source Determination .............. ...............35....
Stable Isotopes ................. ........... ...............36.......
Lipid Biomarker Compounds ................ ...............43........... ....
Petrographic Analysis............... .. ...............4
Nuclear Magnetic Resonance Spectroscopy .............. ...............49....
Conclusion ........._.___..... .__ ...............51....


3 SEDIMENT ORGANIC CARBON STORAGE INT A CONSTRUCTED MANGROVE
AND SEAGRAS S SY STEM .............. ...............65....


Introducti on ............ _. .... ...............65....
M ethod s .............. ...............69....

Study Site............... ...............69..
Sediment Sampling............... ...............70
Laboratory Analyses ............ ..... ._ ...............71....
Statistical Analyses............... ...............74
Re sults............... ..._ ...............75...
Sediment Characteristics .............. ...............75....

Traj ectory of Constructed System .........._.... ...............76..._........











Table 4-1. 613C (%o) and C:N ratios for all potential sources of organic carbon to mangrove and
seagrass sediments in SL 15 and reference sites averaged over various collection
times and plant parts (unless otherwise noted). Values in parentheses are + SE; where
no standard error is listed, the value is for a single composite sample.
Location Species 813C (o ) C.8
Subtidal Seagrass -11.54 (0.84)
Leaves -10.95 (1.3) 14 (0.52)
Roots -12.42 (0.78) 31 (2.5)
Syringodium filiforme -9.23 (0.86) 14 (0.5)
Thalassia testudinum -9.83 (1.2) 13 (0.9)
Halodule beaudettei -12.62 11
Halophila johnsonii -20.07 16
Epiphytes -16.20 (1.9) 11 (0.6)
Macroalgae -21.00 (0.98) 18 (1.6)
Acanthophora spicifera -17.11 (0.78) 12 (0.3)
Caulerpa sertulariodes -18.36 14
Sargassum spp. -17.48 (0.29) 28 (0.9)
Daysa baillouviana -32.06 15
Ulva spp. -20.64 14
Chaetomorpha linum -25.29 25
SL subtidal macroalgae -21.75 (0.86) 16 (2)
Rosenviga intricate -20.94 (1.1) 16 (0.8)
Hypnea cervicornis -19.69 (1.2) 21 (0.1)
Gracilaria tikvahiae -22.50 (1.4) 11 (0.4)
Enteromorpha spp. -25.24 19
Seston -26.29 (0.47) 6.5 (0.2)
May 2006 -27.30 (0.44) 7 (0.5)
September 2006 -25.78 (2.2) 6 (0.5)
October 2006 -24.23 (0.21) 6 (0.2)
November 2006 -24.58 (1.3) 6.5 (0.7)
February 2007 -30.10 (0.84) 7.5 (1)
Intertidal Spartina alterniflora -12.94 (0.30)
Leaves -13.37 (0.27) 27 (2)
Roots -12.30 (0.04) 42 (11)
Sued'a linearis -29.22 14
Mangrove -26.95 (0.24)
Leaves -27.27 (0.32) 27 (1.4)
Roots -26.18 (0.20) 58 (3.7)
Avicennia germinans -27.46 (0.53)
Leaves -27.77 (0.49) 22 (1.7)
Roots -25.26 55
Rhizophora mangle -27.17 (0.36)
Leaves -27.32 (0.47) 31 (1.7)
Roots -26.34 (0.30) 55 (9)
Laguncularia racemosa -26.31 (0.84) 26 (3.0)
Terrestrial C3 terrestrial -27.54 (0.44) 33 (4)
Schinus terebenthifolius (leaves) -28.30 31
Casuarina equisetifolia (needles) -26.33 34
Coccoloba uvifera (leaves) -27.77 (0.32) 32 (11)
Borrichia fr~utescens -27.86 (0.60) 21 (0.9)
Distichlis spicata -13.67 (0.86) 27 (2.6)









analyzing only 613C ValUeS of a biomarker for one plant group, contributions of plants outside

that group remain unknown.

Mead et al. (2005) took the examination of group-specific series of homologues a step

farther when they used an n-alkane-based proxy, Paq, along with compound-specific stable

isotopes to elucidate sources along a gradient of freshwater marsh to estuarine mangrove forests

to marine seagrass beds in the Florida Everglades. Paq is calculated from abundances of

different n-alkane homologues.

Paq = (C23- 25 2)(3i 25 2z+ 9 3 1 ) (2-2)

In equation 2-2, Cx is the amount of the Cx n-alkane. Submerged and floating macrophytes like

seagrasses contained more abundant mid chain n-alkanes and therefore had a higher Paq than

emergent macrophytes and terrestrial plants like mangroves. This method was able to resolve

sources to a greater extent than studies using only isotopes or only biomarkers in estuaries

because sources with similar Paq values were differentiated using n-alkane 613C ValUeS and vice

versa. Generally, as the gradient went from freshwater marsh to seagrass beds there was a trend

of increasing sediment 813C ValUeS and increasing Paq values. These trends were further

connected to contributions of individual sources through a PCA based on compound specific

813C and Paq.

An example using a species-specifie biomarker is the study of different homologues of the

n-alkane-2-ones lipid series to elucidate SOC sources in the Harney River estuary and the

adj acent Florida shelf (Hernandez et al. 2001; Mead et al. 2005). Lipids in this series generally

have odd-numbered C chains ranging from 19 to 33 C's in length. There has been some debate

about whether n-alkane-2-ones arise is sediments directly from plant detritus or whether they

arise from microbial oxidation of alkanes. However, Hernandez et al. (2001) were able to Eind










LIST OF TABLES


Table page

2-1 Global area of mangrove forests, salt marshes, and seagrass beds ................. ................52

2-2 Global rates of carbon accumulation in coastal ecosystem sediments .............. ................53

2-3 Rates of carbon accumulation in coastal ecosystem sediments ................ ................ ...54

2-4 Studies comparing organic carbon in restored and constructed coastal marshes to OC
in natural reference marshes .............. ...............58....

2-5 Stable isotope values and dominant source conclusions from carbon source
determination studies in coastal ecosystems ................. ...............60........... ...

3-1 Mean (a SE) bulk density, % shell pieces, pH, and Eh redoxx potential) of the
sediments according to depth and site .............. ...............90....

3-2 Results of factorial ANOVAs comparing SL 15 and references. ................ ................. 91

3-3 Results of the repeated measures ANOVAs for SL 15 mangrove and seagrass
sediments ................. ...............92.................

3-4 Mean organic carbon concentrations (%) and storage (g m-2), nitrogen
concentrations, and carbon to nitrogen molar ratios of SL 15 and reference
mangrove and seagrass sediments according to depth and month............... .................9

3-5 Mean concentration (mg kg- ) and storage (g m-2) Of two relatively labile types of
organic carbon in SL 15 and reference mangrove and seagrass sediments .......................94

3-6 Mean organic carbon liability of organic carbon in SL 15 and reference sites
according to depth and month ................. ...............95........... ...

3-7 Organic carbon accumulation rates in mangrove and seagrass systems in this and
other studies .............. ...............96....

4-1 613C (%o) and C:N ratios for all potential sources of organic carbon to mangrove and
seagrass sediments in SL 15 and reference sites............... ...............128.

4-2 Decay constants (a SE) and turnover times calculated from a nonlinear regression
(exponential decay) of litter bag experiment data ................. .............................129

4-3 Results of ANOVAs comparing 613C ValUeS in SL 15 and reference mangrove and
seagrass sediments and surface layers and of repeated measures ANOVAs of 613C
values in SL 15 sediments and surface layers............... ...............130

4-4 Mean 613C and C:N (A SE) for sediments and surface layers of SL 15 and reference
mangrove and seagrass systems ................. ...............131...............










'The values were not measured in the study and were taken from published values in the literature. 2Averaged values of leaf, root, rhizome and litter tissue or
across sites to obtain one stable isotope value. 3Average or range of entire study because the authors did not provide the specific values for each site. 4Averaged
values of e ach 2 cm section in the top 10 cm of sediment. 5Average of several species. 6Range taken from a graph. NR = not reported









4-5 Mean 613C Of 6l3C ranges of means for sources in this study and in the literature .........132










(513C
(C:N


Ma ngrove


Water Column

Terrestrial 15139
(C:N


Bacteria ?b13C IC:N


PlIa nkto n ? b13C


Detritus ?b13C
1C:N


Seston 1513C (C:N


Seagrass (513@


Macroalgae ++513C


Figure 4-8. A theoretical diagram of organic carbon sources that may constitute seston and how
they affect seston 613C and C:N. Arrow sizes indicate the possible relative
contributions of each source. I indicates depleted 813C and low C:N, T indicates
enriched 813C and high C:N, and t, indicates mid-range 613C and C:N.









that were the same order of magnitude as our constructed system sediments--MBC was 193 to

715 mg C kg-l in a European seagrass bed (Boschker et al. 2000), 289 to 769 mg C kg-l in a

California salt marsh (Cordova-Kreylos et al. 2006), and 182 mg C kg-l in an Australian seagrass

bed (Moriarty et al. 1985). Note that this conversion equation came from sandy soils, not marine

sediments, so values are not exact but are estimates for comparison purposes.

Organic carbon liability

The magnitude of OC pools is not the only factor that affects C storage, so further data

exploration is needed to assess whether SL 15 stores sediment C as well as other seagrass beds

and mangroves forests. SL 15 sediments must not only have OC pools equal to or greater than

references to function as a significant C store, they must also have their OC stored in long term

pools, where it can be sequestered away from the atmospheric C pool for decades, centuries, and

even millennia. Relative amounts of the OC pool are important because the pool containing the

most OC affects the overall storage abilities of a system. A system with most of its OC in non-

reactive, recalcitrant pools is going to store C longer than a system with most of its OC in active

pools like microbial biomass (Buyanovsky et al. 1994).

The constructed system generally stored more OC in short-term pools than references. In

all sediments except constructed mangrove sediments, ExOC made up less than 1% of the TOC

pool (Fig. 3-9), but the percentage of the TOC pool made up by MBC was greater in constructed

than in reference sediments. In SL 15 mangrove sediments, 53 to 63% of their TOC was MBC,

while in reference sediments 1 1 to 15% of TOC was MBC (Fig. 3-9). This trend was the same in

mangrove surface layers. In SL 15 seagrass 0-10 cm sediments, 24 to 38% of their TOC was

MBC, while in references 17 to 20% of TOC was MBC (Fig. 3-9). SL 15 accreted layers and

reference 0-5 cm depths had similar percentages that ranged from 19 to 27% (Fig. 3-9).

Sediments in this study had more TOC stored as MBC than in other coastal systems, which










Many salt marsh, mangrove, and seagrass ecosystems have been degraded or lost

through disturbances such as dredging channels and developing coastlines for human habitation

(Valiela et al. 2001; Kennish 2002; Zedler 2004). This degradation and loss affects the

biogeochemical functioning of coastal systems including C sequestration. Loss and degradation

of coastal systems therefore affects the global C cycle and may increase the affects of climate

change (Duarte et al. 2005; Bridgham et al. 2006). Globally, 50% of wetlands (freshwater and

coastal) have been lost (Moser 1996). The continuous United States has lost 53% of its wetlands

since the 1780's (Dahl 1990). Since 1989, the United States has had a policy of no net wetland

loss that includes coastal wetlands (Zedler 2004). In Florida, state policy applies this principle to

seagrass systems as well. Therefore when mangrove and seagrass ecosystems are destroyed,

their loss must be mitigated by restoring or creating these systems elsewhere. It is important to

know whether mitigation of coastal ecosystems restores the accumulation and storage capacity of

these important C sinks. Such research can indicate whether mitigation is truly effective and

whether coastal ecosystem restoration can become a policy tool for reducing CO2 emiSsions, as

was suggested by Connor et al. (2001). Functional trajectory studies of constructed systems and

studies comparing constructed systems with natural systems are used to determine the

effectiveness of mitigation in restoring ecosystem functions, like C storage.

Functional traj ectories are used to monitor the development of ecological functions in

constructed ecosystems over time. When constructed systems' functions equal those of

reference systems, the constructed systems are said to be functionally equivalent. Studies that

documented functional traj ectories of OC in restored and constructed salt marshes concluded that

it takes a long time for the restored/constructed marshes to develop SOC pools equal to their

natural counterparts (Simenstad and Thom 1996; Craft 2001; Havens et al. 2002; Morgan and









Water Column Seston 1513C
1C:N



Algal Mat (513C tMC:N












Water Column Seston 1513C

Seagrass (513C Mangrove 15139 C:
C:N (C:N


SL 15 Mangrove


Ref Mangrove


Figure 4-9. Main sources and how they affect surface layer and sediment 813C and C:N. Arrow
sizes indicate the relative contributions of each source. I indicates depleted 813C and
low C:N, 1 indicates enriched 813C and high C:N, and ** indicates mid-range 613C
and C:N.









inundated for longer periods of time (always in the case of seagrass beds), which can create more

highly reducing conditions that slow OM decomposition. More contact with water also means

more contact with, and accumulation of, the dissolved organic carbon (DOC), particulate organic

carbon (POC), and nutrients that water transports. Nutrients and OC stimulate bacterial

production in sediments, nutrients stimulate autotrophic production of OC, and POC settles

becoming part of sediment OM (Gacia et al. 2002).

The second reason seagrass sediments reach OC functional equivalence first is parent

material. In most constructed salt marshes, in the SL 15 mangrove system, and in the Southwest

Florida restored mangroves, the parent material was dredge spoil that is practically devoid of

OM. As previously discussed, dredge spoil was not the only material found in SL 15 seagrass

sediments. There was also OM-rich material originating from old vegetated sediments that were

disturbed during construction, in 5-10 cm depths. At time zero OC is therefore greater in

seagrass sediments. In the New Hampshire seagrass study, the sediment material was not spoil

but a previously vegetated, estuarine "A horizon" that had been devoid of seagrasses for 12 years

(Evans and Short 2005). Like in the 5-10 cm depth of the SL 15 seagrass sediments, it is likely

OC was present before restoration began.

The third reason seagrass sediments reach equivalence before mangrove and salt marsh

sediments is the different OC amounts among the three coastal systems. OC content varies

greatly, even among nearby reference sites (Craft et al. 1999), but generally seagrass sediments

have the lowest OC and mangrove sediments the highest. Reported range in seagrass %OC is

0.15 to 1.3 (Evans and Short 2005; Vichkovitten and Holmer 2005). Reported range in salt

marsh %OC is 1.7 to 13.5 (Moy and Levin 1991; Simenstad and Thom 1996; Zedler and

Calloway 1999; Morgan and Short 2002). Reported range in mangrove %OC is 2.3 to 37










significant amounts of n-alkane-2-ones in tissues of seagrasses and mangroves. In seagrasses,

the most common (82% to 88% of the ketone fraction) n-alkane-2-one was the C25 homologue,

and in mangroves the most common n-alkane-2-ones were the C27-C31 homologues. Gas

chromatograms of sediments in the lower estuary and the shoreward section of the Florida shelf

showed a predominance of seagrass-derived C25 homologues, implying that seagrass was a maj or

SOC source there. In upper estuarine sediments, there was a predominance of higher molecular

weight homologues implying mangroves were the major SOC source. Isotopic measurements of

bulk SOC and n-alkane-2-ones confirmed these conclusions about primary SOC sources because

sediment 813C ValUeS became more enriched (i.e. more like seagrass-derived SOC) as the

samples went from the upper estuary to the Florida shelf. By using biomarkers specific to

vascular plants, however, contributions to SOC from algae and plankton were unknown.

As with the use of bulk stable isotope measurements, there are caveats with the use of lipid

biomarkers. First, this method has not been as extensively studied as the use of bulk isotopes.

Inherent variation of molecular distributions and compound specific isotope signatures within

different tissues of an individual plant, within plants of the same species, across geographical

areas, and across seasons has yet to be documented (Canuel et al. 1997). Also, the more specific

biomarkers may not be applicable to all species of the same plant type. The temperate seagrass,

Z. marina, did not have the predominant C25 n-alkane-2-one homologue that sub-tropical

seagrass species had (Hernandez et al. 2001). Not all major ecosystem components will have

appropriate species-specific biomarkers, so a combination of species-specific and group-specific

biomarkers may have to employed (Mead et al. 2005). Biomarkers confirm the presence of a

certain source in SOC, but they do not necessarily yield relative contributions of sources because

not all sources are represented in each lipid type. Just because the isotopic mixing equation










M~, = Me, *e(-kt) (4-1)

In equation 4-1, Mo is the initial litter mass, Mt is the litter mass at time t, and k is the decay

constant. The decay constant for each species was estimated using nonlinear models in JMP

Version 6 (SAS Institute, Cary, NC).

To investigate whether 613C, TOC, or C:N changed through time in SL 15 sediments and

surface layers, repeated measures ANOVAs were run for both mangrove and seagrass areas. A

spatial power covariance structure was used to account for unequal spacing between time points.

Subjects were the plots on SL 15, and the repeated factor was time. For the 0-5 and 5-10 cm

depths in each system, the ANOVAs were run with depth as a main effect and a time~depth

interaction term. The floc, algal mat, and accreted layers were each run separately in ANOVAs

where time was the only effect. These analyses were run using the mixed procedure in SAS

Version 8 (SAS Institute, Cary, NC).

Comparisons between SL 15 and the reference sites were analyzed using one factorial

ANOVA each for the mangrove and seagrass sediments and one factorial ANOVA each for the

mangrove and seagrass surface layers (algal mat/litter and floc). Sediment ANOVAs consisted

of three fixed factors--site, month, and depth. Surface layer ANOVAs consisted of only the site

and month factors. All two way interactions were tested. SL 15 plot and reference site data

were pooled into two site treatments, SL 15 and reference. Months used in these analyses were

July and November 2006, the sampling dates for which both SL 15 and reference data were

available. For seagrass sediment analysis, SL 15 and reference depths were assigned to 3

categories in order to make comparisons: SL 15 accreted and reference 0-5 cm were depth 1, SL

15 5-10 cm and reference 0-5 cm were depth 2, and SL 15 5-10 cm and reference 10-15 cm were

depth 3. Factorial ANOVAs were run on JMP Version 6 (SAS Institute, Cary, NC).









result in the replacement of fully functioning ecosystems with ineffective surrogates that do not

provide the same functional value (Zedler 2004). Success of most mitigation proj ects is judged

on the survival of macrophytes, not on proper functioning of the ecosystem. With the maj ority

of ecosystem functions are not assessed, the true success of mitigation proj ects is usually

unknown.

One maj or function of coastal ecosystems is C sequestration. The value of this ecosystem

function is increasing with mounting concern about climate change. Anthropogenic release of

greenhouse gases like carbon dioxide (CO2) and methane (CH4) through fossil fuel burning and

deforestation, and livestock production, respectively, is the maj or cause of global climate change

(IPCC 2001). Coastal ecosystems dominated by macrophytes including salt marshes, seagrass

beds, and mangrove forests are high productive habitats that act as sinks for CO2 and therefore

mitigate climate change. Worldwide, salt marshes and mangroves store at least 44.6 Pg C in

their sediments (Chmura et al. 2003). Seagrass beds, which make up only 0. 15% of the global

marine area, account for 15% of the global marine organic C (OC) storage (Hemminga and

Duarte 2000). Global rates of C sequestration in vegetated marine sediments are estimated

between 111 and 216 Tg C y^l ( Duarte et al. 2005). Based on the low estimate, globally

mangroves bury 23.6 Tg C y^l, salt marshes bury 60.4 Tg C y^l, and seagrass bury 27.4 Tg C y^l

(Duarte et al. 2005). In the United States, salt marshes store 400 Tg C and sequester 4.4 Tg C y

1, and mangroves store 61 Tg C and sequester 0.5 Tg C y^l (Bridgham et al. 2006); the C stored

and sequestered by seagrass systems is unknown. Coastal ecosystems also export C to the

oceans where another portion is buried (Duarte et al. 2005).

The capacity of coastal ecosystems to sequester C, like freshwater wetlands, is greater

than the capacity of uplands. These "wetlands" are a natural C sink, while upland systems










of storage in the references' OC pools; 2) OC liability would be greater in sediments of

constructed systems than in reference sediments; 3) SOC sources in constructed systems would

be macroalgae or plankton, while SOC sources in reference systems would be vascular plants,

like mangroves and seagrass. Key findings addressing each objective and the validity of the

hypotheses are presented below.

Objective One: Short Term Trajectories of Sediment Organic Carbon Pools

Contrary to the hypothesis, functional traj ectories were not followed by OC parameters in

the constructed site sediments. Instead of steady increases, SOC parameters either remained

unchanged or increased and decreased throughout the year, driven by seasonal changes in the

water column. The only sediment functional trajectory was followed by the mangrove system's

bulk density, which decreased throughout the year but remained above reference levels.

Functional traj ectories were somewhat followed by surface layers as both microbial biomass C

(MBC) and total OC (TOC) increased. Due to their proximity to OC inputs, it is logical that OC

should increase in the surface layers before they increase in sediments. However, whether

increases in surface layer OC were due to a recovering function or an annual pattern could not be

discerned. For example, the increase in floc MBC and TOC followed the same trend as total

suspended solids, a water quality parameter. Overall, one year was not sufficient time to map

OC functional trajectories in the constructed mangrove and seagrass system. The lack of a

functional traj ectory did not preclude the OC parameters from being functionally equivalent to

reference values.

Objective Two: Comparisons of Sediment Organic Carbon Pools

The hypothesis that SOC pools would be smaller in constructed systems was by and large

correct for mangrove sediments but not for seagrass sediments (Fig. 5-1 and 5-2). Floc and

accreted layers of constructed seagrass sediments reached or exceeded functional equivalence for









containing the most OC affects the overall sequestration abilities of a wetland. A wetland with

most of its OC in the recalcitrant pool is going to sequester C longer than a wetland with most of

its OC in a labile pool like microbial biomass, which has frequent turnover. A study of macro

organic matter (MOM), precursor of SOM, in constructed marshes showed that younger marshes

had more labile MOM than older marshes indicating they were less likely to sequester OC in the

long term (Craft et al. 2003).

Sources of SOC may also be important as they influence OC liability, carbon to nitrogen

ratios (C:N), and the rate at which OC accumulates. Morgan and Short (2002) hypothesized that

the lag time in OC accumulation is because macrophytes must first become established before

they contribute to the SOM pool. Others claim that seston, a mixture of plankton and detritus, is

the main source of SOM so accumulation should occur whether a site has macrophytes or not

(Cammen 1975). Organic matter C:N ratios may also be a significant parameter because the

ratios indicate whether accumulation of C is likely. If the C:N ratio is low, the microbes may be

starved for C, and therefore more likely to decompose OC and respire CO2.

Because these studies have been so broad in scope, they also do not take the time to use

the best methodology for measuring OC. While loss on ignition (LOI) is the easiest way, the

high carbonate content of coastal sediments/soils may interfere with the results (Nieuwenhuize et

al. 1994). Furthermore, LOI is a measure of OM so conversion factors, with their associated

errors, need to be used to convert an OM value to an OC value. Older studies (e.g. Cammen

1975) used the Walkley-Black chromic acid oxidation method to determine OC, which is only

75-90% efficient at obtaining a true OC value (Nieuwenhuize et al. 1994). While errors in the

method are not a significant problem for comparison studies reviewed here, they are a concern if

constructed wetlands are to be used for C emission offsetting. Future work should consider in









km2 (Table 2-1). Together these systems occupy only 0.8% of the global ocean area, but they

contribute 30% of the total ocean C storage (Duarte and Cebrian 1996). This observation

indicates that these systems play a significant role in global sequestration of C.

Coastal ecosystems are better C sinks than terrestrial systems and freshwater wetlands.

Coastal ecosystems and freshwater wetlands can accumulate C indefinitely while terrestrial

systems reach an equilibrium where C fixed equals the amount respired annually (Rabenhorst

2005). Coastal ecosystems and freshwater wetlands' abilities to continually accumulate C are

due to their anoxic sediments where electron acceptors other than Ol must be utilized to

decompose OC. These electron acceptors yield less energy to microbes than Oz, Slowing

decomposition rates (Schlesinger 1997).

Coastal ecosystems are better C sinks than freshwater wetlands, because they release

orders of magnitude less CH4, a pOtent greenhouse gas (Bridgham et al. 1996). CH4 has a higher

radiative forcing capacity than CO2, So its global warming potential (GWP), a measure of its

radiative forcing capacity per one unit mass relative to the radiative forcing capacity of one unit

mass of CO2, iS by definition greater than the GWP of CO2 (IPCC 2001). CH4 TeleaSe OCCUTS

because methanogenesis, the process where CO2 is reduced to CH4 in Order to breakdown

organic matter (OM), is the dominant decomposition pathway in most freshwater systems. In

coastal ecosystems high sulfate levels inhibit methanogenesis as sulfate is a more energetically

efficient electron acceptor than CO2 (CapOne and Kiene 1988).

Rates of Organic Carbon Sequestration

Before any discussion on rates of C accumulation, terminology and resulting caveats must

be addressed. Not all studies use the same terminology when reporting rates of C accumulation

and often studies do not specifically define their rate terminology. Terms used in the literature

that all essentially meant "rate at which OC builds up in soil" were "rate of OC accumulation,"









Stable isotopes of bulk compositions have successfully identified the main SOC sources

in subtropical and tropical coastal ecosystems dominated by mangroves and seagrasses because

potential sources in these ecosystems have a wide range of 613C (Hemminga et al. 1994;

Jennerj ahn and Ittekkot 2002; Gonnocea et al. 2004; Kennedy et al. 2004; Papadimitriou et al.

2005; Smit et al. 2005; Zhou et al. 2006). Mangroves have the most depleted 813C because

Rubisco carboxylase discriminates against isotopically heavy C during C3 photosynthesis

(Hemminga and Mateo 1996; Hemminga and Duarte 2000). Seagrasses have the most enriched

613C, despite C3 characteristics, because of diffusional constraints on C uptake in an aquatic

environment (Hemminga and Mateo 1996). Isotopic signatures of other potential sources such as

plankton and epiphytes generally fall between mangrove and seagrass values (Kennedy et al.

2004; Papadimitriou et al. 2005).

In this study, we determine: 1) significant sources to the SOC in a constructed mangrove

and seagrass system, 2) how sources change over time in a constructed system, and 3) how

sources differ between the constructed system and nearby mangrove and seagrass reference

sediments. We hypothesized that SOC sources in the constructed system will initially be

macroalgae or seston, while SOC sources in the reference systems will be vascular plants like

mangroves and seagrasses.

Methods

Study Site

SL 15 (Fig. 4-1) is a mitigation site located in the subtropical portion of the Indian River

Lagoon (IRL) adj acent to Fort Pierce, Florida. The IRL is a long, shallow, and microtidal water

body that lies in both temperate and subtropical climates. SL 15 is one of many spoil islands

created in the Indian River Lagoon during the construction of the Atlantic Intracoastal

Waterway. These islands sit several meters above sea level and are populated by many exotics,










system with those in more natural, reference systems; 3) to compare the liability of SOC in the

constructed and reference systems; 4) to determine and compare significant sources to the total

SOC pool in the constructed and reference systems.

The hypotheses were: 1) in the short term, storage in the three OC pools studied would

increase in the constructed systems, but would not reach the level of storage in the references'

OC pools; 2) OC liability would be greater in sediments of constructed systems than in reference

sediments; 3) SOC sources in constructed systems would be macroalgae or plankton, while SOC

sources in reference systems would be vascular plants, like mangroves and seagrass.









seagrass sediments therefore do not store C as well as references. OC in both SL 15 and

reference mangrove sediments are a mixture of seston and vascular plants (terrestrial plants in SL

15 and mangrove/seagrass via litter in references). Since terrestrial plants most likely have

decay rates similar to mangroves and slower than most seagrass species, it is possible that the

length of OC storage in SL 15 and reference mangrove sediments are currently similar. The

labile algal mat, however, has the potential of becoming a main source in constructed mangrove

sediments because it caused sediment 813C enrichment throughout the year, which may

ultimately shorten the length of constructed mangrove OC storage.










2003; 16, Cammen 1975; 17, Craft et al. 1993: 18, Craft et al. 1999; 19, Hopkinson 1988: 20, Suzuki et al. 2003; 21,
Romero et al. 1994: 22, Gacia et al. 2002.
'This author reported organic matter accumulation rates, so rates were divided by 2 in order to obtain these
numbers. M1Lodeled sediment profiles instead of measuring them directly. 3Calculated by subtracting the OC in 0-
30 cm from the OC in top 10 cm, divided by the age of the site 4Calculated by subtracting the OC at time 0 from
the OC at time 1, divided by time 1-time 0 5Denotes carbon buried after exportation to the open ocean not in situ.











Table 3-5. Mean (+ SE) concentration (mg kg- ) and storage (g m-2) Of two relatively labile types of organic carbon in SL 15 (n=12)
and reference (n=9) mangrove and seagrass sediments according to depth and month. ExOC=extractable organic carbon
and MBC=microbial biomass carbon.


ExOC
(mg kg' dry soil)


MBC
(mg kg' dry soil)


ExOC
(g m-2


MBC
(gm-2


Month and system Depth


SL 15 Reference SL 15


SL 15
14000 (5000)
1900 (100)
1600 (100)

7200 (1000)
1800 (100)
1800 (200)

16000 (500)

2100 (100)
1340 (80)
1 100 (70)

23000 (3000)

2600 (200)
1400 (100)
1300 (40)


Reference Reference SL 15 Reference


July mangrove



Nov. mangrove


Algal Mat/ litter
0-5 cm
5-10 cm

Algal Mat/ Litter
0-5 cm
5-10 cm


1800 (1000)
130 (30)
86 (7)

750 (200)
76 (5)
80 (4)


8500 (2000)
740 (20)
690 (10)

12000 (3000)
900 (30)
820 (20)

18000 (100)
1700 (100)
840 (40)
800 (40)


24000 (1100)
2300 (100)
990 (30)
1000 (40)


6.7 (2.0)
3.8 (0.3)
2.8 (0.3)

5.7 (1.0)
4.2 (0.5)
2.4 (0.2)

0.33 (0.08)
2.1 (0.1)



0.37 (0.03)
3.4 (0.3)
2.3 (0.1)
2.6 (0.2)


8.8 (3)
6.3 (1)
3.9 (0.5)

4.7 (2)
3.5 (0.2)
3.8 (0.2)


57 (5)
60 (2)
51 (2)


75 (9)
87 (4)
69 (4)

49 (20)
83 (3)
80 (3)

48 (3)

100 (4)
67 (3)
67 (3)

60 (9)

100 (2)
71 (5)
74 (2)


110 (30)
75 (4)
61 (2)


July seagrass


Floc
Accreted
0-5 cm
5-10 cm
10-15 cm

Floc
Accreted
0-5 cm
5-10 cm
10-15 cm


830 (200)
47 (4)
38 (3)

600 (100)
52 (7)
32 (3)

100 (9)
77 (6)
29 (2)
24 (1)


130 (11)
100 (6)
31 (2)
36 (2)


89 (10)

53 (4)
25 (1)
24 (3)

91 (6)

62 (5)
39 (5)
28 (1)


0.26 (0.02) 55 (10)
47 (4)
2.4 (0.1) 64 (2)
1.4 0.1) 59 (2)
1.2 (0.04)


Nov. seagrass


0.24 (0.0)

2.5 (0.1)
2.0 (0.3)
1.7 (0.07)


65 (4)
75 (5)
74 (3)
73 (3)









isotope ratio-monitoring gas chromatography-mass spectroscopy (irm-GCMS) (Canuel et al.

1997).

Canuel et al. (1997) examined the usefulness of isotopic signatures of specific lipid

compounds to identify SOC sources in coastal ecosystems. The study examined isotopic

signatures of bulk organic matter, total lipid extracts, and a whole suite of lipid compounds in

three vascular plants, S. alterniflora, J. roemerianus, and Zostera marina, suspended particulate

matter (SPM), and sediment in North Carolina. Vascular plants had similar molecular

compositions of sterols and fatty acids but differed in hydrocarbon compositions, specifically in

ranges and maxima of n-alkanes and the presence of monosaturated alkenes. SPM had a

different lipid composition than vascular plants; the maj ority (>50%) of SPM' s hydrocarbons

were C25 highly branched isoprenoids (HBI) alkenes. Among different vascular plant lipids there

was a variety of isotopic signatures with an average depletion of 3-5 %o in lipid 813C ValUeS

relative to bulk values. Lipids in Z. mostera, S. alterniflora, and J. roemerianus followed the

same trend in 613C ValUeS as bulk tissues with mean 613C ValUeS tin %o) of -14.8 to -18.9, -18.4 to

-22.6, and -29.0 to -33.8 for lipids and of -10.0, -12.6, and -26.0 for bulk tissues, respectively.

Differences between bulk and lipid signatures demonstrated another reason caution should be

used in analyzing bulk isotopic studies because compounds preserved in SOC may not have the

same signature as bulk plant matter (Canuel et al. 1997). 613C ValUeS of lipids in sediments were

different than those for the same lipids in vascular plants, but similar to SPM lipids. Sediments

had a higher diversity of lipids than vascular plants, small amounts of maj or vascular plant

biomarkers like C21 and C29 (maxima observed in the Z. mostera and S. alterniflora tissues), and

a lot of the major SPM biomarker, C25 HBI. The study concluded vascular plants were only

minor contributors to SOC.










respired to CO2 inStead of being stored in sediments long term. Differences in OC liability were

partially due to differing C limitations and to differences in SOC sources between constructed

and reference systems.

Objective Four: Comparisons of Sediment Organic Carbon Sources

Sources to the SOC pool differed between constructed and reference systems, but not to

the extent that was hypothesized (Fig. 5-1 and 5-2). Ternary diagrams suggested that seston

from the water column was a main SOC source for all systems-constructed and natural. The

true importance of seston, however, was unclear because the low sediment C:N ratios that led to

the conclusion that seston was a main source, can also result from diagenetic transformations. In

mangrove sediments, both systems had lignin-containing higher plants as other main sources--

terrestrial plants in the constructed system and mangroves/seagrass via litter in the reference

systems. The effect of sources on OC storage in mangrove sediments was therefore similar; but

there was an indication that the labile algal mat was becoming an increasingly important source

in constructive sediments, which would shorten OC storage times. Sources in reference seagrass

sediments were unclear. It was apparent though, that a greater amount of SOC was derived from

macroalgae in the constructed system than in the reference system. Litter bag studies

demonstrated that macroalgae generally have the fastest decomposition rates of all aquatic

macrophytes, indicating that the macroalgae-derived OC would not be stored in constructed

sediments for long amounts of time.

Conclusion

Overall, neither mangrove nor seagrass sediments of the constructed system are

functionally equivalent to their respective references in regards to OC storage (Fig. 5-1 and 5-2).

Recovery indices indicate how close various parameters are to equivalence with references.









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

SEDIMENT ORGANIC CARBON POOLS AND SOURCES INT A RECENTLY
CONSTRUCTED MANGROVE AND SEAGRASS ECOSYSTEM

By

Caitlin E. Hicks

December 2007

Chair: K. R. Reddy
Major: Soil and Water Science

Coastal ecosystems are significant natural carbon sinks. If constructed coastal ecosystems

can obtain the same carbon sink capacity as their natural counterparts, then construction and

restoration of these systems has the potential to become a tool for reducing atmospheric CO2. In

this study, sediment organic carbon (OC) of a recently constructed mangrove and seagrass

system in the Indian River Lagoon, Florida was compared with sediment OC of nearby mature,

reference systems. Total OC, extractable OC, and microbial biomass C pools were measured to

compare C storage. Organic C liability in the constructed and reference sites was also measured.

The main sediment OC sources were determined using 13C isotopes and C:N ratios and were

compared among systems. Organic C pools were generally larger in sediments of reference

systems than in sediments of the constructed systems, but differences in pool sizes were much

greater between the constructed and reference mangrove systems. Organic C liability was greater

in the constructed systems indicating their sediments could not store OC for as long as the

references. Seston was a major source of sediment OC in all systems. Other main sources of OC

were higher plant-derived in constructed and reference mangrove and reference seagrass

sediments, but were algal-derived in constructed seagrass sediments. After one year, the C sink

capacity of the constructed systems is less than the capacity of the reference systems, but the



































I I I I I I


SL 15 Mangrove


0


100 %
S-100


75
453 g OC m-2
(Total OC)


Floc
238 g OC m-2





0-5 cm
122 g OC m-2


5-10 cm
93 g OC m-2


Reference Mangrove


0 25


105-159 g OC m-2 1lLiterature value
(Net Accumulation)

50 75 100 %
I 100


S75
22'
(Te
S


I


Litter
978 g OC m-2




0-5 cm
607 g OC m-2

5-10 cm
688 g OC m-2


73 g OC m-2
total OC)


Figure 3-9. Organic carbon (OC) pools in SL 15 and reference mangrove and seagrass sediments.
Beside each box is the total amount of OC in the depths analyzed. OC accumulation
rates were calculated in this study for SL 15 sediments (includes algal mat for SL 15
mangrove) but are literature values for reference sediments (Callaway et al. 1997 for
mangrove and Gonnocea et al. 2004 for seagrass). Boxes show the percentage
distribution of the total OC in each depth and OC pool-1VIBC (dark grey), ExOC
(white), and other (light grey).


120 g OC m-2 1l
(Net Accumulation)

25 I50 75











Table 2-1. Global area of mangrove forests, salt marshes, and seagrass beds.
Global area
System (km2) Data sources
Mangrove forests 200000 1
218000 2
240000 3
Salt marshes 300000 2
400000 4
Seagrass beds 300000 5
600000 6
al, Jennerjahn and Ittekkot 2002; 2, Twilley et al. 1992; 3, Mitsch and Gosselink 2000; 4, Duarte and Cebrian 1996;
5, Suzuki et al. 2003; 6, Hemminga and Duarte 2000.











Table 2-5. Continued
Source Sediment Main How main sources Data
Location Potential sources 8' C Site description 61 C sources determined source
Chale Lagoon, Kenya Seagrass -10.72 -14.8 1 Comparison 12
Mangroves -26.75


Silaqui, Philipines



Pislatan, Philipines


Seston
Seagrass
Epiphytes

Seston
Seagrass
Epiphytes

Seston
Seagrass
Epiphyte

Seston
Seagrass
Epiphyte
POM

Seston
Seagrass

Seston
Seagrass
Epiphyte

Seston
Seagrass


-16.4
-5.8
-9.6

-16.5
-7.5
-10.5


Percent contribution ranges
from mixing equation


Percent contribution ranges
from mixing equation


-14.9


-22.1
-12.4 ,
-17.8


Spain


Iberian Coast

Balearic Islands


2,1 (3?) Percent contribution ranges
1,2 (3?) from mixing equation


-15.8 to
-21.6
-15.8 to
-21.63
-20.07


Fanals Point, Spain


Percent contribution ranges
from mixing equation and
microscopic examinations


Percent contribution ranges
from mixing equation

Percent contribution ranges
from mixing equation


Percent contribution ranges
from mixing equation



Ternary mixing diagram
of 61 C and N:C


-24.7
-12.2
-17
-21.5

-19.6
-8.6

-17.7
-6.0
-8.6

-12.1
-7.6


Can Rhan Lagoon,
Vietnam

Dam Ghia Bay,
Vietnam


Mi Gang II, Vietnam


-18.6


-15.8


-13.2


Seagrasses and mangroves


Fringing mangrove
Lagoon center


Celestun, Mexico


Seston
Seagrass
Mangrove


-22.1
-16.1
-28.62










KENTULA, M. E., R. P. BROOKS, S. E. GWINT, C. C. HOLLAND, A. D. SHERMAN, AND J.
C. SIFNEOS. 1992. An Approach To Decision Making In Wetland Restoration And
Creation. Island Press.

KRISTENSEN, E. 1994. Decomposition Of Macroalgae, Vascular Plants And Sediment Detritus
In Seawater Use Of Stepwise Thermogravimetry. Biogeochemistry 26: 1-24.

LAL, R., J. KIMBLE, E. LEVINTE, AND C. WHITMAN. 1995. World soils and greenhouse
effect: An overview. In R. Lal, J. Kimble, E. Levine, and B. A. Stewart [ed.], Soils and
Global Change. Advances in Soil Science. CRC Press.

LALLIER-VERGES, E., B. P. PERRUSSEL, J. R. DISNAR, AND F. BALTZER. 1998.
Relationships between environmental conditions and the diagenetic evolution of organic
matter derived from higher plants in a modern mangrove swamp system (Guadeloupe,
French West Indies). Org. Geochem. 29: 1663-1686.

LECKIE, S. E., C. E. PRESCOTT, S. J. GRAYSTON, J. D. NEUFELD, AND W. W. MOHN.
2004. Comparison of chloroform fumigation-extraction, phospholipid fatty acid, and DNA
methods to determine microbial biomass in forest humus. Soil Biol. Biochem. 36:
529-532.

LEE, S., AND J. A. FUHRMAN. 1987. Relationships between biovolume and biomass of
naturally derived marine bacterioplankton. Applied and Environmental Microbiology 53:
1298-1303.

LINT, G. H., T. BANKS, AND L. STERNBERG. 1991. Variation in delta-13-C values for the
seagrass Thala;ssia-testudinum and its relations to mangrove carbon. Aquatic Botany 40:
333-341.

LINDAU, C. W., AND L. R. HOSSNER. 1981. Substrate characterization of an experimental
marsh and 3 natural marshes. Soil Sci. Soc. Am. J. 45: 1171-1176.

MACHAS, R., R. SANTOS, AND B. PETERSON. 2006. Elemental and stable isotope
composition of Zostera noltii (Horneman) leaves during the early phases of decay in a
temperate mesotidal lagoon. Estuar. Coast. Shelf Sci. 66: 21-29.

MARCHAND, C., E. LALLIER-VERGES, AND F. BALTZER. 2003. The composition of
sedimentary organic matter in relation to the dynamic features of a mangrove-fringed coast
in French Guiana. Estuar. Coast. Shelf Sci. 56: 119-130.

MATEO, M. A., AND J. ROMERO. 1996. Evaluating seagrass leaf litter decomposition: An
experimental comparison between litter-bag and oxygen-uptake methods. J. Exp. Mar.
Biol. Ecol. 202: 97-106.

MCKEE, K. L., AND P. L. FAULKNER. 2000. Restoration of biogeochemical function in
mangrove forests. Restor. Ecol. 8: 247-259.





Table 2-4. Studies comparing organic carbon

Location Site
Tacoma, Washington' Gog-Le-Hi-Te, Site 1


in restored and constructed coastal marshes to OC in natural reference marshes.
Age Constructed OC Reference OC Depth sampled
(years) (units) (units) (cm) Method" Sourceb
1 3.5 % 3.3 8.7 % 0-2 1 1
2 3.0
3 4.0
6 3.5
1 4.0
2 4.5
3 5.5
6 9.0
1 2.5
2 2.0
3 2.2
6 1.2
1 2.0
2 2.0
3 3.0


Gog-Le-Hi-Te, Site 2




Gog-Le-Hi-Te, Site 3




Gog-Le-Hi-Te, Site 4


Maine, New Hampshire .


Great Bay Estuary


2.0 %
1.5
3.0
2.5


mean = 23 %


77.3 g OC m
184.3
77.3
193.3
77.3
206.4


362.7 g OC m2 0-13


Core Banks, North Carolina


Sound-side marsh, site 1

Sound-side marsh, site 2

Sound-side marsh, site 3


San Diego, California'


San Diego Bay


3.5 %
5.5
7.5
7/0


7.5 11 %


Not reported










samples (Fig. 4-5 through 4-7). All of the samples that fell outside the ternary plots, regardless

of site or depth, did not fit because their N:C ratios were greater than that of the sources.

The maj ority of SL 15 mangrove sediment samples fell near the seston end member. Some

samples fell in the middle of the triangle and others fell close to the terrestrial end member (Fig.

4-5a). Most reference mangrove sediment samples fell between seston and litter end members

(Fig. 4-5b). In terms of 613C, but not in terms of N:C, most SL 15 and reference seagrass

sediment samples were within the range of macroalgal sources (Fig. 4-6). SL 15 seagrass

sediment samples fell far from the seagrass end member (Fig. 4-6a). SL 15 seagrass 0-10 cm

and accreted depths did not differ in their sources. Most reference seagrass samples fell outside

the diagram due to high N:C ratios (Fig. 4-6b). Examining only 613C, reference seagrass

sediments were more enriched than macroalgae and seston but more depleted than seagrass (Fig.

4-3). Reference mangrove litter layer samples from July fell between seston and seagrass end

members but November samples fell in the middle or at the mangrove vertex (Fig. 4-7a).

Reference seagrass floc samples fell between seston and macroalgae end members in July but

outside the diagram in November (Fig. 4-7b). SL 15 seagrass floc fell between seston and

seagrass regardless of sampling data (Fig. 4-7b.)

Discussion

Source Characteristics

613C Of the main potential sources in the studied part of the Indian River Lagoon were

within the range of literature from similar estuarine studies (Table 4-5). Our sources' C:N values

were also within reported literature values of 30 to 99 for mangrove leaves and roots (Lallier-

Verges et al. 1998; Thimdee et al. 2003; Gonnocea et al. 2004; Muzuka and Shunula 2006), of

15 to 21 for seagrass fronds (Thimdee et al. 2003; Gonnocea et al. 2004; Machas et al. 2006), of









(CO3-2, HCO3-1), which inherit the high 813C Of CaCO3 (- 0) (Lin et al. 1991). These carbonate

species then may be utilized by algae as inorganic C sources during daytime photosynthesis.

813C in the literature ranges from -29.4%o to -20.6%o for mangrove sediments (Bouillon et

al. 2003; Thimdee et al. 2003; Gonneea et al. 2004) and from -10.3%o to -26.6%o for seagrass

sediments (Hemminga et al. 1994; Kennedy et al. 2004; Papadimitriou et al. 2005). Sediment

813C in this study are for the most part within literature values. SL 15 and reference mangrove

sediments span the range of literature values from -27.5%o to -19.4%o. SL 15 and reference

seagrass sediments are at the lower end of the literature values with 813C ranging from -23.2%o to

-19.4%o. Differences in 613C among SL 15 and reference sediments and surface layers suggest

their SOC sources differ. Observations of the distribution of primary producers around the sites

also suggest sources differ, even between SL 15 and reference mangrove sites, whose 613 wr

not significantly different.

Source Determination

The ternary diagram indicated that seston was the dominant source for SL 15 mangrove

sediments with some OC being contributed by terrestrial plants and the algal mat (Fig. 4-5a).

Terrestrial sources most likely contributed to SOC before and during construction. During

construction, we observed terrestrial plant parts that were not fully removed by burning and

clearing being mixed into spoil within SL 15's intertidal zone. The algal mat's influence as a

source was supported by 613C enrichment of mangrove sediments over the first year. Mangroves

were not included as a source in the ternary diagrams because SL 15 mangroves were young (>2

years old) and mangrove litter was very sparse. Seston was also a dominant source for reference

mangrove sediments according to the ternary diagram, but in this instance it shared this

designation with the litter layer (Fig. 4-5b). According to Fig. 4-3, mangroves also contributed

to SOC because mean sediment 813C was more depleted than mean seston and litter values.









A portion of the above analyses were performed on data transformed to meet the normality

requirement (see Appendix A for details). Post hoc multiple comparisons were carried out on

significant effects using the Tukey test. Significance was decided using an alpha level of 0.05.

Ternary diagrams (Dittmar et al. 2001; Goni et al. 2003, Gonnocea et al. 2004) were used

to determine the main SOC sources. Because ternary diagrams can only have three end

members, Hield observations and the position of mean sediment 813C relative to mean potential

source 613C On a 613C line (Fig. 4-3) were used to choose the three most likely end members for

each constructed and reference sediment and for the mangrove litter layer and seagrass floc. N:C

of the three end members and sediments were plotted against 813C. N:C ratios are used instead

of C:N ratios because with the larger number in the denominator, they are more statistically

robust (Goni et al. 2003). End members' N:C and 813C were averaged for all sampling dates and

species within that group (e.g.: mangroves), but for plants where multiple parts were measured,

only leaf/frond values were used. The three end members create a triangle that is expanded

according to the standard deviations of the end members to account for natural variability and

analytical error. Sediment samples that fall in the middle of the triangle are assumed to be a

mixture of all three sources, samples that fall along a line connecting two end-members are

considered a mixture of those two sources, and samples that fall around the vertex of an end

member are assumed to have OM from mainly that source. Samples that fall outside of the

expanded triangle have OC contributions from additional sources or have undergone changes

during diagenesis.

Results

Source Characteristics

613C and C:N varied among plant groups. Generally, the lowest 813C and greatest C:N

were found in mangrove leaves and roots and C3 terrestrial plant leaves (Table 4-1). The greatest





PAGE 1

1 SEDIMENT ORGANIC CARBON POOLS AND SOURCES IN A RECENTLY CONSTRUCTED MANGROVE AND SEAGRASS ECOSYSTEM By CAITLIN E HICKS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Caitlin E Hicks

PAGE 3

3 To my parents, Michelle and Richard Hicks, for their wonderful support of all my academic endeavors from preschool onwards

PAGE 4

4 ACKNOWLEDGMENTS I thank the entire Reddy Lab group, all of who helped m e in some way throughout this process. Within the lab group, I especially thank Cory Catts, who endured long, arduous days in the field with me, Angelique Keppler and Meliss a Martin, who answered my countless questions about lab procedures and gradua te school, and Ms. Yu, whose runni ng of the lab is nothing short of miraculous and who was always there to clarify procedures and offer encouragement during long days in the lab. I also thank my advisor, Dr. Reddy, for allowing me to work on a project of my own design. Thanks to my committee, Todd Osborne, Ji m Sickman, and Ted Schuur, who each helped me understand certain concepts or lab procedures Todd Osborne was an integral part of my committee as he mentored me through much of this process. Thanks to Kelly Fischler, Rex Ellis, and Hanna Lee for help in the field and Meghan Brennan for help with statistics. Last but not least, I th ank Alex Pries, who has assisted me in the field, lab, with computer glitches, and in editing portions of this thesis, and who has supported me through both tough times and good times.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................9 ABSTRACT ...................................................................................................................... .............11 CHAP TER 1 INTRODUCTION .................................................................................................................. 13 2 LITERATURE REVIEW .......................................................................................................20 Introduction .................................................................................................................. ...........20 Rates of Organic Carbon Sequestration .................................................................................. 21 Global Rates ....................................................................................................................22 Local Rates ......................................................................................................................23 Measuring Rates of Ca rbon Accumulation .....................................................................24 Comparing Organic Carbon in Restored and Reference Coastal Marshes ............................. 26 Monitoring Constructed Coastal Marshe s Using Functional Trajectories ...................... 28 Functional Trajectory Case Studies ................................................................................. 29 Factors Affecting Functional Equivalence ......................................................................31 Storing Carbon Versus Sinking Carbon .......................................................................... 32 New Directions ................................................................................................................32 Sediment Organic Carbon Source Determination .................................................................. 35 Stable Isotopes ............................................................................................................... ..36 Lipid Biomarker Compounds .......................................................................................... 43 Petrographic Analysis ......................................................................................................48 Nuclear Magnetic Resonance Spectroscopy ...................................................................49 Conclusion .................................................................................................................... ..........51 3 SEDIMENT ORGANIC CARBON STORAGE I N A CONSTRUCTED MANGROVE AND SEAGRASS SYSTEM ................................................................................................. 65 Introduction .................................................................................................................. ...........65 Methods ..................................................................................................................................69 Study Site .........................................................................................................................69 Sediment Sampling ..........................................................................................................70 Laboratory Analyses ........................................................................................................ 71 Statistical Analyses .......................................................................................................... 74 Results .....................................................................................................................................75 Sediment Characteristics ................................................................................................. 75 Trajectory of Constructed System ...................................................................................76

PAGE 6

6 Constructed and Reference Comparisons ........................................................................ 76 Organic Carbon Accumulation Rates .............................................................................. 78 Discussion .................................................................................................................... ...........79 Sediment Characteristics ................................................................................................. 79 Trajectory of Constructed Site .........................................................................................79 Constructed and Reference Equivalence ......................................................................... 81 Organic carbon pools ............................................................................................... 81 Organic carbon lability ............................................................................................. 86 Organic carbon accumulation ................................................................................... 87 Conclusion .................................................................................................................... ...88 4 SOURCES OF SEDIMENT ORGANI C CARB ON IN A CONSTRUCTED MANGROVE AND SEAGRASS SYSTEM ....................................................................... 107 Introduction .................................................................................................................. .........107 Methods ................................................................................................................................109 Study Site .......................................................................................................................109 Litter Bags .....................................................................................................................110 Source Sampling ............................................................................................................111 Sediment Sampling ........................................................................................................ 112 Laboratory Analyses ...................................................................................................... 113 Data Analyses ................................................................................................................114 Results ...................................................................................................................................116 Source Characteristics ...................................................................................................116 Sediments and Surface Layers ....................................................................................... 117 Source Determination .................................................................................................... 118 Discussion .................................................................................................................... .........119 Source Characteristics ...................................................................................................119 Sediments and Surface Layers ....................................................................................... 122 Source Determination .................................................................................................... 123 Conclusion .................................................................................................................... .126 5 SYNTHESIS ..................................................................................................................... ....146 Objective One: Short Term Trajectorie s of Sediment Organic Carbon Pools ...................... 147 Objective Two: Comparisons of Sedim ent Organic Carbon Pools ......................................147 Objective Three: Comparisons of Se dim ent Organic Carbon Lability ................................ 148 Objective Four: Comparisons of Se dim ent Organic Carbon Sources .................................. 149 Conclusion .................................................................................................................... ........149 APPENDIX STATISTICAL TRANSFORMATIONS ............................................................... 155 LIST OF REFERENCES .............................................................................................................157 BIOGRAPHICAL SKETCH .......................................................................................................168

PAGE 7

7 LIST OF TABLES Table page 2-1 Global area of mangrove forests, salt m arshes, and seagrass beds. ...................................52 2-2 Global rates of carbon accumulation in coastal ecosystem sediments .............................. 53 2-3 Rates of carbon accumulation in coastal ecosystem sediments ......................................... 54 2-4 Studies comparing organic carbon in restored and constr ucted coastal m arshes to OC in natural reference marshes .............................................................................................. 58 2-5 Stable isotope values and dominant source conclusions from carbon source determination studies in coastal ecosystems. ..................................................................... 60 3-1 Mean ( SE) bulk density, % shell pi eces, pH, an d Eh (redox potential) of the sediments according to depth and site ............................................................................... 90 3-2 Results of factorial ANOVAs comparing SL 15 and references. ...................................... 91 3-3 Results of the repeated measures ANOVAs for S L 15 mangrove and seagrass sediments...................................................................................................................... ......92 3-4 Mean organic carbon concentrations (%) and storage (g m-2), nitrogen concentrations, and carbon to nitrogen molar ratios of SL 15 and reference mangrove and seagrass sediments according to depth and month .....................................93 3-5 Mean concentration (mg kg-1) and storage (g m-2) of two relatively labile types of organic carbon in SL 15 and referenc e mangrove and seagrass sediments .......................94 3-6 Mean organic carbon lability of orga nic carbon in SL 15 and reference sites accord ing to depth and month. ........................................................................................... 95 3-7 Organic carbon accumulation rates in mangr ove and seagrass system s in this and other studies .......................................................................................................................96 4-1 13C () and C:N ratios for all potential sources of organic carbon to mangrove and seagrass sediments in SL 15 and reference sites ..............................................................128 4-2 Decay constants ( SE) and turnover tim es calculated from a nonlinear regression (exponential decay) of litter bag experiment data. ........................................................... 129 4-3 Results of ANOVAs comparing 13C values in SL 15 and reference mangrove and seagrass sediments and surface layers and of repeated measures ANOVAs of 13C values in SL 15 sediment s and surface layers ..................................................................130 4-4 Mean 13C and C:N ( SE) for sediments and surface layers of SL 15 and reference mangrove and seagrass systems. ......................................................................................131

PAGE 8

8 4-5 Mean 13C or 13C ranges of means for s ources in this study a nd in the literature .........132

PAGE 9

9 LIST OF FIGURES Figure page 1-1 The carbon cycle in seagrass beds. .................................................................................... 18 1-2 The carbon cycle in mangrove forests. .............................................................................. 19 3-1 The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset) ...............97 3-2 Core from SL 15 seagrass system illustra ting the surface layer (floc) and different sediment depths (accreted layer, 0-5 cm, 5-10 cm) ........................................................... 98 3-3 The functional trajectory the bulk dens ity of SL 15 m angrove sediments followed over the first year after construction ..................................................................................99 3-4 The changes in organic carbon parameters ov er the first year afte r construction in SL 15 seagrass and m angrove sediments .............................................................................. 100 3-5 The changes in organic carbon parameters ov er the first year afte r construction in SL 15 seagrass and m angrove surface layers. .......................................................................101 3-6 Comparisons between total organic car bon (TOC) in reference and SL 15 m angrove (top) and seagrass (bottom) sediments .............................................................................102 3-7 Comparisons between extractable organi c carbon (ExOC) in reference and SL 15 m angrove (top) and seagrass (bottom) sediments ............................................................ 103 3-8 Comparisons between microbial biom ass carbon (MBC) in reference and SL 15 m angrove (top) and seagrass (bottom) sediments ............................................................ 104 3-9 Organic carbon (OC) pools in SL 15 and reference m angrove and seagrass sediments...................................................................................................................... ....105 4-1 The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset) .............133 4-2 Core from SL 15 seagrass system illustra ting the surface layer (floc) and different sediment depths (accreted la yer, 0-5 cm, and 5-10 cm) .................................................. 134 4-3 13C averaged over July and November 2006 for SL 15 and reference sediments and surface layers compared to mean 13C of potential sources. ........................................... 135 4-4 Mean 13C of SL 15 sediments and surface layers over the first year after construction .................................................................................................................. ....136 4-5 N:C vs. 13C in ternary m ixing diagrams of three potential OC sources and mangrove sediments. ........................................................................................................ 137

PAGE 10

10 4-6 N:C vs. 13C in ternary m ixing diagrams of thr ee potential OC sources and seagrass sediments...................................................................................................................... ....139 4-7 N:C vs. 13C in ternary m ixing diagrams of th ree potential OC sources and surface layers ........................................................................................................................ ........141 4-8 A theoretical diagram of organic carbon sources that m ay constitute seston and how they affect seston 13C and C:N ....................................................................................... 143 4-9 Main sources and how they affect surface layer an d sediment 13C and C:N ................. 144 5-1 A modified seagrass bed carbon cycle show ing values from this study in constructed (C) and reference (R) systems .......................................................................................... 152 5-2 A modified mangrove fore st carbon cycle showing values from this study in constructed (C) and re ference (R) system s ...................................................................... 153 5-3 Recovery indices of three organic carbon (OC) pools and OC lability p arameters for constructed mangrove and seagrass systems. .................................................................. 154

PAGE 11

11 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science SEDIMENT ORGANIC CARBON POOLS AND SOURCES IN A RECENTLY CONSTRUCTED MANGROVE AND SEAGRASS ECOSYSTEM By Caitlin E. Hicks December 2007 Chair: K. R. Reddy Major: Soil and Water Science Coastal ecosystems are significant natural carbon sinks. If constructe d coastal ecosystems can obtain the same carbon sink capacity as their natural counterparts, then construction and restoration of these systems has the potential to become a tool for reducing atmospheric CO2. In this study, sediment organic carbon (OC) of a recently constructed mangrove and seagrass system in the Indian River Lagoon, Florida was co mpared with sediment OC of nearby mature, reference systems. Total OC, extractable OC, an d microbial biomass C p ools were measured to compare C storage. Organic C lability in the co nstructed and reference sites was also measured. The main sediment OC sources were determined using 13C isotopes and C:N ratios and were compared among systems. Organic C pools were generally larger in sediments of reference systems than in sediments of the constructed systems, but differences in pool sizes were much greater between the constructed and reference mangrove systems. Organic C lability was greater in the constructed systems indicating their sedi ments could not store OC for as long as the references. Seston was a major sour ce of sediment OC in all systems. Other main sources of OC were higher plant-derived in constructed and reference mangrove and reference seagrass sediments, but were algal-derive d in constructed seagrass sedime nts. After one year, the C sink capacity of the constructed systems is less than the capacity of the refe rence systems, but the

PAGE 12

12 constructed seagrass system is functioning more like its reference than the constructed mangrove system. In the long term, however, the potenti al C sink capacity of th e constructed mangrove system is greater.

PAGE 13

13 CHAPTER 1 INTRODUCTION Restora tion and construction of coastal ecosyst ems may help mitigate the effects of climate change by reducing atmospheric carbon dioxide (CO2). Global climate change has become a major environmental concern over the past 50 y ears. The anthropogeni c release of greenhouse gases is the major cause of global climate cha nge (IPCC 2001). Atmospheric concentrations of CO2 and methane (CH4), the biggest contributors to climate change, are increased by fossil fuel burning and deforestation, and livestock produc tion, respectively. Highly productive coastal ecosystems including salt marshes, seagrass beds, and mangrove forests are carbon (C) sinks. Salt marshes and mangroves store at least 44.6 Pg C in their sediments (Chmura et al. 2003). Seagrass beds, which make up only 0.15% of globa l marine area, account for 15% of global marine organic C (OC) storage (Hemminga and Duarte 2000). Sediment C storage values are even larger when stores of i norganic C like carbonates are take n into account (Zhu et al. 2002). The C cycle in coastal ecosystems is an open cycle because OC is imported into and exported out of systems by water currents a nd tides. In a seagrass bed, seagrasses and macroalgae (drift and epiphytic) take up CO2 and HCO3 from the water column to produce OC through photosynthesis (Fig. 1-1). When seagrass fronds senesce or break, they (and their associated macroalgae) become litterfall on top of sediments or are exported out of the system. Litterfall OC is either decomposed by microbes or incorporated into the sediment OC (SOC) pool by leaching, bioturbation, or burial. Impor ted OC, which may include terrestrial and mangrove detritus, is trapped by seagrass fronds a nd settles on the sediment. Seston, comprised of plankton, bacteria, and dissolved and particulate OC from in a nd outside the system, is also trapped by seagrass fronds. The fate of trapped OC is the same as litterfall. Seagrass root OC from exudates or dead tissue is immediately part of the SOC pool and can be used by microbes.

PAGE 14

14 SOC can be part of three poolsmicrobial biomass C, labile OC, or recal citrant OC. Generally, microbes consume mainly labile OC, which they respire as CO2 or incorporate into their biomass. When microbes die, their OC becomes pa rt of labile and recalc itrant pools. The more OC is reworked by microbes, the more recalcitran t it becomes. The recalcitrant pool is where OC is sequestered long-term a nd where OC undergoes abiotic condensation into complex humic materials. In mangrove forests, the C cycle is basically the same except for the C sources (Fig. 1-2). The inorganic C source used by mangroves is atmospheric CO2, the litter is mangrove leaves, and imported OC trapped in litter by mangrove roots is seston and seagrass detritus. The capacity of coastal ecosystems to sequest er OC is greater than the capacity of terrestrial ecosystems. Coastal ecosystems are na tural C sinks, while terrest rial systems reach an equilibrium where the net C fixed annually is about zero (Rabenhorst 1995). Constant accumulation of C in coastal ecosystems is due to their anoxic sediments. In these sediments, oxygen is depleted so electron accep tors that are not as efficient must be utilized by microbes to decompose OC. Coastal ecosystems also have a greater OC sequestration capacity than freshwater wetlands because they, unlik e freshwater wetlands, do not use CO2 as a terminal electron acceptor and therefore emit less CH4 (Bridgham et al. 2006). In coastal ecosystems, sulfate is the terminal electron acceptor, and hi gh sulfate levels inhibit methanogenesis (Capone and Kiene 1988). A study of ma ngrove forests did not detect CH4 either dissolved in sediment porewaters or fluxing out of se diments and 51 to 75% of OM oxidation was occurring through sulfate reduction (Alongi et al. 2004). Coastal ecosystems, like salt marshes, mangrove forests, and seagrass beds, may therefore be highly sign ificant C sinks because they accumulate C in sediments without emitting CH4.

PAGE 15

15 Many salt marsh, mangrove, and seagrass ecosystems have been degraded or lost through disturbances such as dredging channels and developing coastlines for human habitation (Valiela et al. 2001; Kennish 2002; Zedler 2004). This degradation and loss affects the biogeochemical functioning of coastal systems including C sequestration. Loss and degradation of coastal systems therefore affects the global C cycle and may increase the affects of climate change (Duarte et al. 2005; Bridgham et al. 2006 ). Globally, 50% of wetlands (freshwater and coastal) have been lost (Moser 1996). The con tinuous United States has lost 53% of its wetlands since the 1780s (Dahl 1990). Since 1989, the United States has had a policy of no net wetland loss that includes coastal wetlands (Z edler 2004). In Florida, state policy applies this principle to seagrass systems as well. Therefore when mangrove and seagrass ecosystems are destroyed, their loss must be mitigated by restoring or creating these systems elsewhere. It is important to know whether mitigation of coastal ecosystems restores the accumulation and storage capacity of these important C sinks. Such research can indi cate whether mitigation is truly effective and whether coastal ecosystem restoration can become a policy tool for reducing CO2 emissions, as was suggested by Connor et al. (2001). Functional trajectory studies of constructed systems and studies comparing constructed systems with natural systems are used to determine the effectiveness of mitigation in restoring ecosystem functions, like C storage. Functional trajectories are used to monitor the development of eco logical functions in constructed ecosystems over time. When cons tructed systems functions equal those of reference systems, the constructed systems are said to be functionally equivalent. Studies that documented functional trajectories of OC in restor ed and constructed salt marshes concluded that it takes a long time for the rest ored/constructed marshes to develop SOC pools equal to their natural counterparts (Simenstad and Thom 1996 ; Craft 2001; Havens et al. 2002; Morgan and

PAGE 16

16 Short 2002; Craft et al. 2003). Functional trajectory studies, w ith one exception (Evans and Short 2005), have been limited to temperate brackish and salt water marshes. There is a need to study functional trajectories in constructed mangrove forests a nd seagrass beds. Functional trajectory studies generally meas ure a suite of ecologi cal functions, so SOC is usually the only variable measured that pertains to ecosystem C storage. Studies that examine multiple OC pools, OC lability, and OC sources ar e needed to more fully understa nd the recovery of C storage functioning in constructed systems. Studies that examine short term changes immediately following construction are also l acking. Short term trajectory studies are important because certain aspects of OC storage may recover quickly. Whether constructed mangrove and seagrass ecosystems provide the same ecological services as their natural counterpa rts with respect to C storage, and whether restoration of these services follow a functional trajectory is curren tly unknown. In this thesis, the trajectories of SOC pools in constructed seagrass and mangrove systems were monitored during the first year after construction completion. SOC pools in the c onstructed systems were also compared with SOC pools in adjacent natural systems. Sediments were the focus of this research because they are the sites of long term C storag e. Variables measured include the amount of OC in three pools (total OC, extractable OC, and microbial biomass C), the lability of SOC, and the C to nitrogen ratios and 13C of sediments and potential SOC sources. The constructed site was a former spoil island called SL 15 in the Indian River Lagoon, FL that was converted to a mangrove and seagrass ecosystem in November 2005. The reference sites were the natura l seagrass beds that surround SL 15 and the nearby mangrove forests that occupy the edges of adj acent spoil islands. The main research objectives were: 1) to determine short term trajectories of SOC pools in a constructed mangrove forest and seagrass bed; 2) to compare SOC pools in the constructed

PAGE 17

17 system with those in more natural, reference systems; 3) to compare the lability of SOC in the constructed and reference systems; 4) to determ ine and compare significant sources to the total SOC pool in the constructed and reference systems. The hypotheses were: 1) in the short term, st orage in the three OC pools studied would increase in the constructed systems, but would not reach the level of storage in the references OC pools; 2) OC lability would be greater in sediments of constructed systems than in reference sediments; 3) SOC sources in constructed syst ems would be macroalgae or plankton, while SOC sources in reference systems would be vasc ular plants, like mangroves and seagrass.

PAGE 18

18 CO2HCO3 Export Extractable OC Recalcitrant OC Labile OC Mangrove/Terrestrial Imports CO2Seagrass and algae Seston Trapped particles Litter Microbial Biomass C Sediment Water Column CO2HCO3 Export Extractable OC Recalcitrant OC Labile OC Mangrove/Terrestrial Imports CO2Seagrass and algae Seston Trapped particles Trapped particles Litter Litter Microbial Biomass C Sediment Water Column Figure 1-1. The carbon cycle in seagrass beds.

PAGE 19

19 CO2 Terrestrial Imports Extractable OC Recalcitrant OC Labile OC CO2Litter Seston and Seagrass Imports Sediment Atmosphere Water column Microbial Biomass C Mangrove CO2 Terrestrial Imports Extractable OC Recalcitrant OC Labile OC CO2Litter Seston and Seagrass Imports Sediment Atmosphere Water column Microbial Biomass C Mangrove Figure 1-2. The carbon cycle in mangrove forests.

PAGE 20

20 CHAPTER 2 LITERATURE REVIEW Introduction In this literature rev iew, rates of organic carbon (OC) accumulation are compiled and compared for three coastal ecosystemssalt mars hes, mangrove forests, and seagrass beds. Studies comparing sediment organic carbon (SOC) pools in restored or constructed salt marshes to SOC pools in natural salt mars hes are then examined. This section does not discuss mangrove forests or seagrass beds because the literature on the functioning of restored or constructed coastal ecosystems is currently limited to salt mars hes. Third, methods for determination of SOC sources are discussed for the three coastal ecosy stems. These coastal ecosystems are dominated by vascular, halophytic macrophytes with mangroves dominated by trees and salt marshes and seagrass beds dominated by grasses and other herbaceous species. Sediments in these systems are C sinks due to their high net primary production, trapping of ma terial from the water column, and O2 limited conditions. These systems are globally distributed. Salt marshes and mangroves occupy non-rocky, sedimentary-driven intertidal zones of the world. Salt marshes predominate in temperate climates, while mangroves predominate in subtropi cal and tropical climates. Salt marshes are generally replaced by mangroves at a latitude of 25 N or S (Mitsch and Gosselink 2000). Seagrass systems are subtidal and are found fr om tropical through temp erate climates where needs such as low light attenuation in the wate r column are met (Hemminga and Duarte 2000). Seagrasses are often found adjacent to their intert idal counterpartssalt marshes or mangroves. Multiple estimates of global area covered by each system differ, but in each system the estimates are within the same orde r of magnitude. According to the average of the estimates, mangrove forests cover 220,000 km2, salt marshes cover 350,000 km2, and seagrass beds cover 450,000

PAGE 21

21 km2 (Table 2-1). Together these systems occ upy only 0.8% of the global ocean area, but they contribute 30% of the total ocean C storage (Duarte and Cebrian 1996). This observation indicates that these systems play a significan t role in global sequestration of C. Coastal ecosystems are better C sinks than te rrestrial systems and freshwater wetlands. Coastal ecosystems and freshwater wetlands can accumulate C indefinitely while terrestrial systems reach an equilibrium where C fixed e quals the amount respired annually (Rabenhorst 2005). Coastal ecosystems and fr eshwater wetlands abilities to continually accumulate C are due to their anoxic sediments where electron acceptors other than O2 must be utilized to decompose OC. These electron acceptor s yield less energy to microbes than O2, slowing decomposition rates (Schlesinger 1997). Coastal ecosystems are better C sinks than freshwater wetlands, because they release orders of magnitude less CH4, a potent greenhouse gas (Bridgham et al. 1996). CH4 has a higher radiative forcing capacity than CO2, so its global warming potential (GWP), a measure of its radiative forcing capacity per one unit mass relative to the radiativ e forcing capacity of one unit mass of CO2, is by definition greater than the GWP of CO2 (IPCC 2001). CH4 release occurs because methanogenesis, the process where CO2 is reduced to CH4 in order to breakdown organic matter (OM), is the dominant decompositi on pathway in most freshwater systems. In coastal ecosystems high sulfate levels inhibit meth anogenesis as sulfate is a more energetically efficient electron acceptor than CO2 (Capone and Kiene 1988). Rates of Organic Carbon Sequestration Before any discussion o n rates of C accumu lation, terminology and resulting caveats must be addressed. Not all studies use the same te rminology when reporting rates of C accumulation and often studies do not specifically define their rate terminology. Terms used in the literature that all essentially meant rate at which OC builds up in soil were rate of OC accumulation,

PAGE 22

22 rate of OC sequestration, rat e of POC burial, rate of refractory accumulation, and organic accumulation rate. Nuances of these terms c ould be gleaned from methodology. Some like rate of C accumulation referred to additions of both labile and refractory OC to the SOC pool (Craft et al. 2003). Others like rate of refractory accumulation re ferred to long-term burial of OC that is unlikely to decompose on a human ti me scale (Cebrian 2002), and others like POC burial were vague (Alongi et al. 2005). Some studies reported OM accumulation, not OC accumulation. Those rates were divided by two to obtain OC accumulation rates. Also it was assumed that C accumulation rates referred to accumulation rates of OC, not total C, because studies that used the term re ported measuring OC. Lastly, fo r indirectly measured rates (modeled or based on mass balance equations) it was not always clear whether rates included amounts of OC from both autochthonous and a llocthonous sources. This review reports all values as C accumulation rates, which refers to the build up of OC in sediments though there may be discrepancies in the lability of accumulating OC. Generally, the longer the timescale of a study, the more likely rates represent long-term burial. Only the term C burial rates definitively refers to long-term storage of refract ory OC. There is a need for future studies to clearly define rate terminology and to be consistent in its use. Global Rates Many scien tists have estimated global rates of C accumulation for coastal ecosystems due to their important role in the global C cycle (Table 2-2). Global rates of C accumulation for these systems are calculated in several ways. The most common way was averaging published accumulation rates for many sites (Duarte and Cebrian 1996; Chmura et al. 2003). Other methods included graphing frequency distributions of published accumula tion rates (Cebrian 2002), scaling up from model-derived rates (Suzuki et al. 2003), or using mass balance equations derived from production and burial estimates (J ennerjahn and Ittekkot 2002). Despite the

PAGE 23

23 different ways of calculating global rates, accumu lation rates for intertidal systems are basically in agreement (Table 2-2). Estimated global a ccumulation rates for mangrove forests range from 92 to 200 g C m-2 yr-1 and rates for salt marshes range from 50 to 175 g C m-2 yr-1 (ignoring the high estimate of Rabenhorst (1995). Rates for se agrass beds are more variable and range from 16.5 to 270 g C m-2 yr-1. Higher variability for seagrass ecosystems is likely because C accumulation rates in seagrass sediments have been studied less than in mangroves and salt marshes. Suzuki et al. (2003) estimated th at seagrasses caused an accumulation of 1.2 g C m-2 yr-1 in deep ocean sediments due to export of th eir primary production to the open ocean and its subsequent burial. While th is review concentrates on in situ accumulation, it is important to recognize that there are other ways these systems contribute to the global C sink. Coastal ecosystems accumulate C at a rates seve ral orders of magnitude greater than rates in terrestrial systems and the open ocean (Table 2-2). C cycling in terrestrial systems should reach a steady-state condition, making them neither a C source nor C sink (Hussein et al. 2004). Disturbances such as fire, however, occur before climax stages causing terrestrial systems to become C sources to the atmosphere. Coastal ecosystem C accumulation rates are greater than open ocean rates because their primary producer s differ. Open ocean phytoplankton have a much lower net primary production (NPP) per unit ar ea, have a greater percentage of their NPP consumed by herbivores, and contain more eas ily decomposed OM than coastal macrophytes ( Duarte and Cebrian 1996; Cebrian 2002). Local Rates Mean global rates of C accum ulation (2-2) we re calculated using rates of C accumulation from numerous local studies (Table 2-3). The majority of C accumulation rates were measured in salt marshes while accumulation rates in seagrass beds were the least measured. Accumulation rates in mangrove fo rests ranged from 33 to 841 g C m-2 yr-1 with the lowest rate

PAGE 24

24 measured in Terminos Lagoon, Mexico (Gonneea et al. 2004) and the highes t rate measured at the low marsh in Jiulonglijang Estuary in China (Alongi et al. 2005). Rates in salt marshes ranged from 2 to 300 g C m-2 yr-1 with the lowest rate measured at a natural site in Dells Creek, North Carolina (Craft et al. 2003) and the highest rate measured behind a continuous canal in Lafourche Parish, Louisiana (Cahoon and Turner 1989). Rates in seagrass beds ranged from 19 to 191 g C m-2 yr-1, a range measured offshore of Cala Culip, Spain (Romero et al. 1994). Comparing the compiled rates for these systems, there were no trends of one system having consistently higher C accumulation rates than the other systems (Table 2-3). The system accumulating C at the highest rates even varied within the same region. For example, in Celestun Lagoon, Mexico, mangroves accumulated more carbon in their sediments than seagrasses, but in Terminos Lagoon, Mexico, the reverse was true (Gonneea et al. 2004). This lack of a trend is supported by Chmuras (2003) review that found no si gnificant differences between C accumulation rates in salt marshes and mangroves. It should be noted, however, that contributions of mangrove forests to C storage on an ecosystem scale may be greater than salt marshes and seagrass beds because large amounts of C are stored for decades in woody biomass of mangrove trees (Twilley et al. 1992). Measuring Rates of Carbon Accumulation Calculating rates of C accum ula tion typically involves three steps. The first step is to measure SOC pools. SOC pools for local rate stud ies were directly measured using either an elemental analyzer after acidification of the sample to get rid of carbonates (Gacia et al. 2002), a TOC analyzer (Brunskill et al 2002; Alongi et al. 2004), or a mass spectrometer (Choi and Wang 2004). SOC pools were indirectly measured by using loss-on-ignition (LOI) values in regression equations describing the relationship between SOM and SOC (Connor et al. 2001). The second step is to age the sediment or measur e rates of sediment accretion. Sediment age and

PAGE 25

25 accretion rates are less straightforward meas urements than SOC pool measurements. Radioisotope dating of cores using either 210Pb and 14C activity or 137Cs and 14C peaks from nuclear bomb fallout were the most commonly used methods to date sediments (e.g. Callaway et al. 1997; Connor et al. 2001; C hoi and Wang 2004; Hussein et al. 2004; Alongi et al. 2005). Other methods of dating sediments were Romero s (1994) use of a shipwreck whose date was known and that had been buried by seagrasses over hundred s of years and Chmu ra et al.s (2001) use of pollen stratig raphy. Short term (1-3 years) accretion rates were measured with feldspar markers ( Cahoon and Turner 1989; Cahoon 1994; Cahoon and Lynch 1997) or sediment traps (Gacia et al. 2002). The third step is to calculate C accumulation ra tes. The amount of OC in a unit of sediment is divided by the age of that se diment unit, or OC in a unit of sediment is multiplied by the rate at which that sediment unit accreted. Sediment ages in restored or constructed systems do not have to be dete rmined because the site ages are known. C accumulation rates can be calculated by the difference between SOC content at the beginning of the restoration process and SOC content at subse quent points after the ini tial restoration, divided by sites age (Cammen 1975; Craft et al. 2003). Calculated rates of C accumulation may be dependent on time scale, which is dependent on the method used. With a half-life of 5730 years, 14C methods are suitable for measuring rates over many millennia, while with a half-life of 22.3 years, 210Pb methods are suitable for measuring rates over a century (Bierman et al. 1998). Bomb fallout methods using 137Cs and 14C peaks can only measure rates over the last 40 years as the peaks ge nerally occur in 1963. Many of the highest rates of C accumulation were measured using the feldspar marker technique, which measures C accumulation rates over a year or two. These rates may be high because surface pools of SOC are relatively labile compared to deeper pools of SOC. Much of the

PAGE 26

26 surface SOC may be mineralized by the time it is buried deeper in the soil profile, where it would be measured if longer term methods were used. Long term rates calculated by 14C dating were slower than rates measured over a decad al (Choi and Wang 2004) or an 100 year time scale (Hussein et al. 2004). Choi and Wang (2004) did not attribute this difference to methodology and speculated that greater C accumulation rates are due to increases in primary production over the last 100 years caused by increased CO2 and nutrients. Comparing Organic Carbon in Rest ored and Reference Coastal Ma rshes Highly productive habitats like coastal ecosystems are C sinks as their high C accumulation rates demonstrate. In these coastal ecosystems atmospheric CO2 becomes stored as OC for long periods of time. Restorati on and construction of coastal ecosystems may therefore help mitigate the effects of climate change by reducing atmospheric CO2 (Connor et al. 2001). If limits are placed on CO2 emissions in the United States, coastal ecosystem restoration and construction may then become a viable option for C offsetting. C offsetting occurs when an industry needs to reduce its net CO2 emissions but can or will not reduce their own emissions, so they invest in projects that reduce emissions elsewhere, such as tree planting. Anthropogenic release of greenhouse gases is the major cause of global climate change (IPCC 2001). CO2 and CH4 are greenhouse gases with the bi ggest effect on climate change due to their concentrations in the atmosphere and radiative forcing capacity (IPCC 2001). Humans increase atmospheric concentrations of CO2 through fossil fuel burning and land us e change and concentrations of CH4 through livestock production. This section of th e review focuses on salt marshes due to the dearth of literature on functiona l trajectories in restored or c onstructed mangrove and seagrass systems. Despite the numerous important ecological functions coastal ecosystems and wetlands provide, which extend well beyond their function as C sinks, many were vi ewed as wastelands

PAGE 27

27 until recently (Broome et al. 1988). These system s were seen as wasted space that could be utilized for agriculture or valuable devel opment. Wetlands, including salt marshes, were summarily destroyed without much thought to the consequences of their destruction through the 1980s. The lower 48 U.S. states lost 53% of its wetlands from the 1870s to the 1980s (Dahl 1990). Globally, it is estimated that 50% of the wetlands have been lost (Moser 1996). When these systems are lost, we lose a sink for anthropogenically-derived CO2. For example, Connor et al. (2001) estimated that if 85% of the coastal marshes in the Bay of Fundy had not been altered for agricultural uses, 3.8 x 10 13 g C could have been stored over the past 160 years. The loss of coastal ecosystems and wetlands therefor e disrupts the global C cycle and may increase the affects of climate change. Since 1989, the U.S. has had a policy of no net wetland loss that includes coastal marshes (Zedler 2004). The policy calls for mitigation if a lternatives to destroying wetlands in the course of development projects are unavailable. This mitigation comes in the form of creating new wetlands onsite or nearby to the lost wetland, restoring an existing, degraded wetland, or buying into wetland mitigation banks (Zedler 2004). Because of coastal marshes importance as C sinks and the widespread replacement of natural marshe s with created marshes, it is important to know whether restored and constructed marshes have OC accumulation rates and storage capacities equivalent to those of natural marshes. Such research can indicate whether marsh creation can become a policy tool for reducing CO2 emissions. Connor et al. (2001) suggested that restoring coastal marshes may help countries reduce their CO2 emissions to the standards set by the Kyoto protocol. Monitoring functional trajectories of constr ucted marshes helps researchers understand if constructed marshes OC storage can equal the storage of natural marshes.

PAGE 28

28 Monitoring Constructed Coastal Mars hes Using Functional Trajectories Functional trajectories are used to track the progress of constructed ecosystems and to compare constructed and natural ecosystems (S imenstad and Thom 1996; Zedler and Callaway 1999; Morgan and Short 2002). Functional trajectory studies often examine a whole suite of ecological attributes (Craft et al. 2003) that act as indicators for more complex ecological functions (Simenstad and Thom 1996). Attributes are measured in the same constructed system over time, or in several different -aged constructed systems in the same region using a space-fortime substitution, to obtain a range of attribute values that can be plotted against time (Kentula 1992). In coastal marshes, OC parameters are often just several of many attributes measured. Data are then fitted to a curve and compared w ith values from natural marshes. The resulting trajectory represents how the attr ibute develops in a restored or constructed marsh over time (Morgan and Short 2002). There ar e two main questions that func tional trajectories studies seek to answer: 1) how long does it take for the attribut e in the restored or constructed marsh to reach functional equivalence (i.e. the mean value of that attribute in a natural marsh); 2) is the mean value of an attribute in a natural marsh the corr ect endpoint for the development of that attribute in the restored or constructed marsh? Not all attributes have the same trajector y, and trajectories of the same attribute may differ across different marshes and depending on the natural reference marsh used. There are many different trajectories that attributes like SOC pools can follow (Kentula et al. 1993; Fig. 3). Some attributes may not even follow a trajecto ry and instead stay relatively constant through time (Zedler and Calloway 1999). Craft et al. (2 003) proposed that differe nt attributes follow one of three trajectories dependi ng on whether they are part of hydrologic, biological, or soil development processes. OC pool formation is part of soil development, which in most cases is the slowest trajectory to reach f unctional equivalence (Craft et al. 2003). If a trajectory fits OC

PAGE 29

29 data well, it can be used to pred ict future levels of OC thereby helping agencies set standards for mitigation project monitoring or determine the am ount of C emission credits a created marsh is worth. In theory, functional traj ectories are a simple way to evaluate the current success and predict the future success of constructed marshes; data, however, do not always fit a smooth line. Often, there is high variability between construc ted marshes (Craft et al. 2003) and between years in the same study (Zedler and Calloway 19 99). The reference marshes used can also influence the predicted success or failure of a constructed marsh as a result of their age, variability (Simenstad and Thom 1996), or stress level. Furthermore, predictions from functional trajectories should be considered with caution becau se they do not take into account disturbances that may alter the trajectory. Functional Trajectory Case Studies The studies reviewed here examine the equi valence in constructed and restored marshes that are one (Morgan and Short 2002) to 42 ye ars old (Craft 2001). Th e first prediction of functional equivalence for SOC in a salt marsh wa s made by Seneca et al. (1976) for one of the first salt marsh creation projects using dredge s poil, which is essentially devoid of OC. They predicted it would take 4 to 25 years for the cons tructed marsh to store as much C as the natural marsh. More recent studies show that it probably takes at least 25 years for OC to reach functional equivalence (Table 2-4). SOC and th e related attribute, sediment organic matter (SOM), seem to be one of the last attributes to reach functional equivalence in marshes after aboveground biomass, sedimentation rates, and diversity of flora and fauna. There is also the possibility that OC will never reach functional equivalence as most studies did not follow marshes for a sufficient duration of time to show equivalence. Most studies on the eastern coast of th e United States found a trend of increasing SOC/SOM over time (Craft 2001; Ha vens et al. 2002; Morgan and S hort 2002; Craft et al. 2003).

PAGE 30

30 A study of different-aged New England salt mars hes found that SOM increased steadily from 2% at a 1-year-old site to 15% at a 15-year-old site (Morgan a nd Short 2002). Studies from the western coast did not find strong directional tr ends of SOM over time. In Tacoma, Washington SOM stayed between 2-4% ove r 5 years (Simenstad and Thom 1996) and in San Diego, California only a slight increase of 3% wa s found over 11 years (Zed ler and Calloway 1999). These differences in trajectories may be more a case of land use than geography. Both of the west coast studies took place in large urban areas, whereas the east coast studies took place in a variety of locales, none as deve loped as Tacoma and San Diego. Only two studies documented tidal marshes that reached functional equivalence with their natural references in terms of SOC. The tidal marshes were 25 (Craft et al. 1999; Craft et al. 2003) and 42 (Craft 2001) years old. Both these ma rshes are located in the southeastern U. S. These marshes achieved functional equivalence possibly because they, or their references, differed from most of the mars hes studied. The 42-year-old ma rsh differed because it was a restored marsh and not a marsh constructed from dredge spoil. Instead, it had been disturbed by a dike that prevented tidal inundation but was removed after only 8 years (Craft 2001). The 25year-old marsh differed because its reference marsh was a relatively new natural salt marsh, which contrasted to other reference marshes that are greater than 2,500 years old (Craft et al. 1999). Because the reference was relatively young, its soils resembled spoil (90% sand) more than a histosol (>10% OM). They were mineral entisols with a high bulk density and low OC content (<1.4%). Reference marshes can determ ine whether or not a constructed marsh reaches functional equivalence because th eir mean attribute values re present functional equivalence finish lines.

PAGE 31

31 Factors Affecting Functional Equivalence The reference m arsh used affects the functiona l equivalence of the constructed marsh. In the last example (Craft et al. 1999), if the SOC pool of the 25 -year-old constructed marsh had been compared to the SOC pool of the 2,500-year-o ld natural marsh with a high OM content, the authors would have concluded that the cons tructed marsh had not yet reached functional equivalence. Many studies choose nearby natural marshes as refere nces without regard to their similarities to constructed marshe s. Studies in urban areas are particularly limited by reference sites as the restored site is often the only la rge area of marsh remaining (Simenstad and Thom 1996). Morgan and Short (2002) solved some of the problems associated with reference site choice when they chose reference sites after comp aring constructed sites to potential reference sites using a principle components analysis (PCA) based on physical attributes like aspect, slope, and size. They used the PCA to choose two we ll-matched reference sites for each constructed site. Because reference marsh is a major f actor in whether a constructed marsh reaches functional equivalence, it should not be chosen arbitrarily. While between-system factors affect whet her a constructed wetland reaches functional equivalence, so do within-system factors like elev ation, depth in the soil pr ofile, and variation in sedimentation rates. Even when a constructed marsh as a whole is far from reaching functional equivalence in terms of SOC, some parts of it may be closer th an others. Several studies found higher SOC at low elevations in constructed marshes (Lindau and Hossner 1981; Craft et al. 2002). Lower elevations are inundated by tides fo r longer periods of time, which leads to more highly reducing conditions that can encourage OC storage. In most soils or sediments, OC naturally decreases with depth, which may hinder the ability of lo wer depths to reach functional equivalence. In one of the only studies to exam ine OC at different depths in the soil profile, upper depths reached functional equivalence quickly while OC values at lower depths did not

PAGE 32

32 increase over 7 years (Havens et al. 2002). Sedimentation of mineral particle s dilutes SOC concentrations. Creek banks often have lower OC concentrations than the interior of marshes (Craft et al. 2002) because they experience greater sedimentation of mineral particles (e.g. Temmerman et al. 2003). Simenstad and Thom (1996) cited sedimentation as a reason why SOM in a restored marsh did not increase with time. Storing Carbon Versus Sinking Carbon Even though most constructed marshes do not yet store the same amount of C as their natural counterparts, they may still be acti ng as C sinks. A few studies examined OC accumulation rates as well as SOC pools and found that OC accumulation rates in constructed marshes are as high as or higher than rates in constructed marshes (Cammen 1975; Craft et al. 2002; Craft 2001). The mean OC accumulation rate of 8 different-aged c onstructed wetlands in North Carolina was 42 g C m-2 yr-1 compared to 43 g C m-2 yr-1 in the reference wetlands, even though the OC pools (g C m-2) in the constructed wetlands were significantly lower (Craft et al. 2003). Additionally, some young marshes have hi gh sedimentation rates (Morgan and Short 2002). Sedimentation may encourage OC accumu lation while reducing SOC concentrations resulting in a reciprocal relationship as was demonstrated in Bay of Fundy marshes (Connor et al. 2001). High sedimentation rates may have prevented SOC pools from increasing in the Tacoma and San Diego constructed marshes while encouraging OC accumulation, which unfortunately was not meas ured in those studies. New Directions The extensive studies on coastal marsh functio nal trajectories have been broad in scope and therefore unable to examine OC dynamics in constructed marshes with sufficient detail. SOC is a conglomeration of pools that include a labile pool, a slowly oxidized pool, a very slowly oxidized pool, and a recalc itrant pool (Eswaran et al. 1995). The pool matters, as the one

PAGE 33

33 containing the most OC affects the overall seque stration abilities of a wetland. A wetland with most of its OC in the re calcitrant pool is going to sequester C longer than a wetland with most of its OC in a labile pool like microbial biomass, which has frequent tur nover. A study of macro organic matter (MOM), precursor of SOM, in constructed marshes showed that younger marshes had more labile MOM than older marshes indicating they were less likely to sequester OC in the long term (Craft et al. 2003). Sources of SOC may also be important as they influence OC labil ity, carbon to nitrogen ratios (C:N), and the rate at which OC accumula tes. Morgan and Short (2002) hypothesized that the lag time in OC accumulation is because macrophytes must first become established before they contribute to the SOM pool. Others claim th at seston, a mixture of pl ankton and detritus, is the main source of SOM so accumulation should occur whether a site has macrophytes or not (Cammen 1975). Organic matter C:N ratios may also be a significant parameter because the ratios indicate whether accumulation of C is likel y. If the C:N ratio is low, the microbes may be starved for C, and therefore more li kely to decompose OC and respire CO2. Because these studies have been so broad in scope, they also do not take the time to use the best methodology for measuring OC. While loss on ignition (LOI) is the easiest way, the high carbonate content of coastal sediments/soils may interfere with the results (Nieuwenhuize et al. 1994). Furthermore, LOI is a measure of OM so conversion factors, with their associated errors, need to be used to c onvert an OM value to an OC va lue. Older studies (e.g. Cammen 1975) used the Walkley-Black chromic acid oxida tion method to determine OC, which is only 75-90% efficient at obtaining a true OC value (Nieuwenhuize et al 1994). While errors in the method are not a significant proble m for comparison studies reviewed here, they are a concern if constructed wetlands are to be used for C emi ssion offsetting. Future work should consider in

PAGE 34

34 situ acidification techniques and subs equent analysis with an elemen tal analyzer as the standard method for OC measurements (Nieuwenhuize et al. 1994). Lastly, research is needed that addresses th e permanency of C stor age in constructed and restored coastal systems. Disturbances like ch anges in nutrient loading, invasive species, and hurricanes can affect C storage. A nutrient lo ading experiment in a North Carolina salt marsh increased microbial respiration and caused a s ubsequent net loss of SO C over 12 years (Morris and Bradley 1999). Alternatively, th e spread of a native, yet invasive, grass species in a natural coastal marsh in France was found to increase SO C storage (Valery et al. 2004). These studies occurred in natural marshes, a nd studies that examine disturban ce effect on SOC in constructed coastal systems are needed because constructe d systems may be less resilient than natural systems. In order for constructe d coastal systems to become a viable option for C offsets, these effects need to be unde rstood and quantified. More thorough studies are needed on the C sink capabilities of restored and constructed coastal ecosystems. Thus far, the vast majority of studies have been carrie d out in salt marshes. Studies are needed in coastal ecosystems like mangrove forests and seagrass beds whose destruction is also routinely mitigated with rest oration and construction. If researchers can prove that constructed systems are effective C sinks by demonstrating that they not only follow trajectories of increasing SOC, but also ha ve OC accumulation rate s similar to natural ecosystems, then constructing coastal ecosystems may become an accepted way to offset CO2 emissions. Institutions such as Climate Neut ral (www.climateneutral.org) and the Chicago Climate Exchange (www.chicagoclimatex.com) c ould use coastal ecosystem construction to offset emissions like they now do with certain forestry and agricultural practices.

PAGE 35

35 Sediment Organic Carbon Source Determination One of the new directions functional trajectory studies could take is exam ining sources of SOC in constructed coastal systems. Determini ng SOC sources is important to the study of OC storage in coastal ecosystems as th e identity of sources is one of the factors that determine OC lability and accumulation rates. Hedges (1992) stat ed that understanding the types of OM that accumulate in marine sediments was one of the key questions that needs to be answered in order to better understand global biogeochemical cycl es. The source determination question most often studied is whether the SOC is of allochthonous (via sedimentation) or autochthonous origin ( Middelburg et al. 1997; Bouillon et al. 2003; Golding et al. 2004). If most SOC is of autochthonous origin, then contributions to SOC from the primary producers needs to be teased apart, but this detailed question is harder to answer and rarely studied (Bouillon et al. 2003). There are many possible SOC sources in seagrass beds, mangroves, and salt marshes. Coastal ecosystems can have allochthonous OC inputs of planktonic origin from the open sea or of terrestrial plant and anthropogenic origin. These system s also have numerous potential autochthonous OC inputs. Within a seagrass bed OC can come from different species of seagrasses, epiphytes, macroalgae, or micro-benthic algae. Seag rass beds can also receive OC inputs from adjacent mangroves (Kennedy et al. 2004; Lin et al. 1991). The complexities of seagrass beds, mangroves, and salt marshes make OC source determination difficult. However, making sense of complex OC sources and their ro le in C accumulation and storage is important for the conservation of coastal ecosystems in the face of increased nutrient loading and sea level rise and the maintenance of their C sink capabilities. Many methods have been used to determine SOC sources; however, no single method has offered a definitive answer among and within syst em types. Some methods were developed for two end-member systemssystems in which there ar e only two distinct sources of OC such as

PAGE 36

36 allochthonous terrestrial plant matter and autoch thonous marine plankton. Such simple systems may be encountered in estuarie s that lack submerged aquatic vegetation (Golding et al. 2004). Sometimes researchers simply group the sources of OC into two end-member groups. For example, in salt marshes the SOC inputs from Spartina can be distinguished from the inputs from all other sources because they differ in their 13C values (Middelburg et al. 1997). These two end-member models are often useful in esti mating the major categories of OC sources (i.e. whether allochthonous or autochthonous), but they cannot fully partition the individual SOC sources. The variety of methods used to determine OC inputs range widely in terms of time and equipment involved. Methods can be as simple as comparing C:N ratios of possible sources with sediment C:N ratios or as complicated as sear ching for a biomarker and then isolating and concentrating that specific compound for isotopi c analysis. The most widely used method involves stable isotopeseither comparing bulk composition of 13C in possible sources and sediments or comparing composition of 13C in lipids found in possible sources and sediments. Lipids and other biomarkers can also be used singly to determine sources. Other methods, which include petrographic analysis and nuclear magnetic resonan ce spectroscopy (NMR), involve comparing relative amounts of differe nt OC structures in the soil. Stable Isotopes In salt m arshes, mangrove forests, and seag rass beds many researchers have tried to determine SOC sources by matching the 13C isotopic signatures of the bulk sediment to the 13C isotopic signatures of the sour ces via strait comparison of th e numbers, mixing models, or diagrams (Table 2-5). In order for stable isotop es to elucidate OC sources, the sources need to have consistently distinct isot opic signatures (Papadimitriou et al. 2005). The various primary producers in coastal systems deve lop distinct isotopic signatur es through their discrimination

PAGE 37

37 against heavy isotopes during carbon uptake a nd fixation. Discrimination against the heavy isotope is highest when the inor ganic C exceeds supply. Generally, C3 plants are lighter ( 13C = -35 to -20 ) than C4 plants ( 13C = -15 to -9 ) in their 13C signatures due to the strong isotopic discrimination of the car boxylase Rubisco, an enzyme that is not found in the C fixation pathway of C4 plants (Hemminga and Mateo 1996; Hemm inga and Duarte 2000). Luckily for C source determination in coastal systems, the principal primary producers of seagrass beds, mangrove forests, and salt marshes all have isotopic signatures distinct fr om the signatures of the less abundant primary producers within these systems. Seagrasses are relatively heavy isotopically with average 13C values of -10 to -11 (Hemminga and Mateo 1996). Mangroves, a C3 plant, have isotopic signatures close to that of many terrestrial primary producers with 13C values around -28 (Jennerjahn and Ittekkot 2002; Kennedy et al. 2004). Spartina species that dominate salt marshes are C4 plants with 13C values around -12 to -13 (Haines 1976; Middelburg et al. 1997). The isot opic signatures of other primary producers such as plankton and epiphytes generally fall below that of seagrasses and Spartina and above that of mangroves (Kennedy et al. 2004; Papadimitriou et al. 2005), but this is not always the case. In order for the bulk stable isotope method to be accurate, the 13C signature of the sources must not change during decomposition, or if they do change, the magnitude of the change needs to be small when compared to in ter-source differences (Pap adimitriou et al. 2005). Changes during decomposition are often small, like the 0.7 difference found between fresh and senescent mangrove leaves in Brazil (Jenne rjahn and Ittekot 2002), but are variable in direction and magnitude depending on the plant (Dai et al. 2005). A study of OC inputs into seagrass ( Posidonia oceanica ) sediments of 22 sites in the northwestern Mediterranean by Papadimitriou et al (2005) is a good example of the potential of

PAGE 38

38 bulk isotopic studies and their inhe rent weaknesses. They measured 13C and 15N isotopic signatures of the top 2cm of fi ne fraction (>63um) sediments and of potential sourcesseston (assumed to represent phytoplankton), aboveand below-ground seagrass tissues, and epiphytes. 13C values of the sediments ranged from -15.8 to -21.5 and average 13C values of seston, epiphytes, below-ground seagrass tissues, and above-ground tissues were -22.1 -17.8 12.1 and -12.6 respectively. No systematic differences in the 15N values of the potential sources were found; most likely because discrimina tion against different N isotopes is not due to physiology of primary producers and because N is often a limiting nutrient. At all sites, SOC was more depleted isotopically than seagrass tissues but less depleted than seston. Using a mixing equation based on one developed by Dauby (1989), they were able to find a range of fractional contribution va lues of each source. seagrass seagrass epiphytes epeiphytes seston seston entseCfCfCfC13 13 13 dim 13 (2-1) In equation 2-1, fi is the unknown proportion of the OC from source i in the SOC pool and 13C is the isotopic signature of source i This equation is used to find the range of f values for each source needed to satisfy the e quation and equal the sediment 13C value. With this model, they were able to determine which sites had seston as the major contributor to SOC and which sites had seagrass as the major contributor to SOC. If this were simply a two end member system involving seagrasses and seston, the elucidation of sources to th e sediments using this model would have been straightforward. But these site s also included epiphytes, and their intermediate 13C signature made it impossible for the model to assign them reasona ble contribution ranges (often the ranges included a 0% contribution). Th us the relative contribution of epiphytes to SOC could not be determined by bulk isotopic methods alone.

PAGE 39

39 A similar method was employed by Kennedy et al. (2004) when they examined SOC sources in seagrass beds, mangroves, and mixed seagrass/mangrove systems in the South China Sea. They measured 13C values of sediments, particles in sediment traps, and potential sources (seagrass leaves, mangrove leaves, epiphytes, and seston). They found consistent and distinct differences in the 13C values of potential sources. The use of the mixing equation gave broad estimates of source contribution, which suggested that seagrasses and mangroves contributed to the SOC in their respective systems, but that seston and epiphytes were probably the dominant sources. As with the study by Papadimitriou et al. (2005), the presence of intermediate signatures (epiphytes and seston) between the two end members (seagrass and mangroves) made determination of contributions fr om sources with intermediate 13C values difficult. In salt marshes, similar problems are enc ountered. While the importance of the main primary producers contribution can be easily elucidated, the contri butions of other sources with less distinct 13C signatures cannot be. In one of the firs t studies that used C isotopes to examine SOC sources, Spartina had the distinctive enriched 13C values (-12.3 to -13.6 ) of C4 plants, but all other vascular plants in cluding species as disparate as Juncus roemerianus and Salicornia virginica had signatures between -22.8 and -26 because they were C3 plants (Haines, 1976). Benthic diatoms in this study were plagued with the same intermediately-valued problem as the previously-discussed epiphytes with 13C values between -16.2 and -17.90/00. Through comparing the primary producer and sedime nt values, Haines concluded sediment 13C values generally reflected values of plants growi ng in the sediments. Sediments beneath C4 plants were slightly more depleted in 13C than the C4 plants, and sediments beneath the C3 plants were slightly more enriched in 13C than the C3 plants. The sediments differences from in situ vegetation may have been due to C3 and C4 plant detritus mixing or input from benthic diatoms.

PAGE 40

40 Inputs to SOC from individual species of C3 plant were unknown because their signatures were not distinct. Middelburg et al. (1997) avoide d problems associated with intermediate and indistinct values by using 13C isotopes for the sole purpose of determining the amount of Spartina-derived SOC in salt marshes in Massachusetts (Great Ma rsh) and the Netherlands (Waarde Marsh). In Great Marsh, high marsh SOC 13C value (-13.4 to -14.5 ) was similar to the Spartina value (12.5 ), but low marsh SOC 13C value (-21 to -19.5 ) was not In Waarde Marsh SOC was 9-12 less than the Spartina value (-12.7 ). They hypothesi zed that depletions of SOC values in Waarde marsh and the low marsh of Gr eat Marsh were due to the input of allocthonous OM such as marine plankton and non-local m acrophytes, but since they did not measure these sources they could not definitively identify which contributed to the depletion. They concluded Great Marsh was a peaty marsh where C accumulation was due to Spartina inputs and Waarde marsh was a mineral marsh where accumulation was due to sedimentation. Problems with intermediate values were enc ountered by all previously discussed studies that tried to comprehensively measure 13C values of all major source s. These studies often used either a variation of the mixing e quation developed by Dauby (1989) or a simple comparison of sources and sediment isotope values. However, other ways to calculate source contributions may partially eliminate problems with intermediate values. Gonneea et al. (2004) used a ternary mixing diagram to elucidate relative source contributions of seagrass, mangroves, and seston to SOC. With a ternary mixing diag ram, all sources were end members as they formed a triangle on a graph of C:N ratios plotted against 13C values. Sediment C:N and 13C values were also plotted on this diagram, whic h had a 10% tolerance interval to account for natural variability in source values. Sediment samples that fell in the middle of the triangle were

PAGE 41

41 assumed to be a mixture of all three sources, sa mples that fell along a line connecting two endmembers were considered a mixture of those two sources, samples that fell near one end member were assumed to have OC predominantly from that source, and samples that fell outside the triangle were assumed to have OC contributions from additional sources. This method is limited to systems with three main sources. Conclusion s of research, like Papadimitriou et al.s (2005) study of seagrass, seston, and epiphytes, could have benefited from this method if had they measured C:N ratios. Comparisons of actual values to values from a predicted model can sometimes help elucidate sources better than a mixing equation based on the actual data. These models are based on biomass of potential sources (C hmura et al. 1987), primary produc tivity (Bull et al. 1999), or %SOC (Middelburg et al. 1997). The problem with these models is that they assume sources contribute to SOM in the same relative proportions as their biomass/productivity. This assumption may not be correct because sources differ in their degrees of lability, in their litterfall, and in the amount of th eir biomass that is exported out of the system. However, models are good approximations, especially in more peat y coastal wetlands where sediments have high OC and sedimentation of allocthonous OC inputs is minimal. For the above methods of calculating SOM sources from isotope values, parameters other than the 13C values, such as biomass or C:N ratios, are needed. These other parameters also support conclusions based on isotopic values alone. Many studies combine C:N ratios with isotopic measurements ( Middelburg et al. 1997; Bouillon et al. 2003; Soto-Jimenez et al. 2003; Thimdee et al. 2003; Gonneea et al. 2004). Corr elations between C:N ratios or %SOC and 13C values are used to assist in determin ing SOC sources. A mild relationship (R2 = 0.26) was found to exist between 13C values and C:N ratios in a Mexican salt marsh where less negative 13C

PAGE 42

42 values corresponded to lower C:N ratios (Soto-Jime nez et al. 2003). In the Mexican marsh lower C:N ratios were thought to be indicative of ma rine producers, specifica lly plankton. A brief review of sediment 13C and C:N values from mangrove lite rature also showed an inverse relationship between the two vari ables (Bouillon et al. 2003). Si milar trends were found when comparing sediment 13C and %SOC values in mangrove s (Bouillon et al. 2003) and salt marshes (Middelburg et al 1997). More depleted 13C values corresponded with higher %SOC in mangroves and more enriched 13C values corresponded with higher %SOC in salt marshes. Generally, the higher the sediment %O C values, the closer the sediment 13C values are to the 13C values of the dominant vege tation (Bouillon et al. 2003). There are other complications with the use of stable isotopes for OC source determination. Problems not already discussed incl ude inherent variati on of isotopic signatur es within different tissues of a single individual (P apadimitriou et al. 2005) and with in a single species (Hemminga and Mateo 1996) across sites (Kennedy et al. 2004), seas ons, and years (Anderson and Fourqurean, 2003; Fourqurean et al 2005). These variations are most pronounced in seagrasses (Thimdee et al. 2003). Such variation may be due to changes in relative uses of dissolved CO2 and HCO3 (sources of inorganic C in water) (Lin et al. 1991), and changes in irradiance, photosynthesis rates, and temp erature (Hemminga and Mateo 1996). Kennedy et al. (2004) found that isotopic signatures of sources such as seagrasses ( 13C = -5.8 to -13.3 ) and seston ( 13C = -9.6 to -22.9 ) varied grea tly among 15 different sites in the South China Sea. The order trend of potential sources 13C signatures (seagrass > ep iphyte > seston > mangrove) remained constant, however. Variation by location means that conclusions based on measurement of SOC 13C values without measuring potential sources may not be valid. SotoJimenez et al. (2003) inappropr iately used isotopes when they only measured sediments

PAGE 43

43 signatures in a Mexican marsh. Average 13C value of sediment was -20.4 and they assumed, based on previously published 13C signatures of sources in temperate estuaries, that dominant SOC sources were plankton and macrophytes. While 13C values of putative sources at each site need to be measured at least once per site, Fourqurean et al. (2005) proposed further determination of source 13C values seasonally and annually. Lipid Biomarker Compounds The use of specific organic com pounds, called biomarkers, to identify SOC sources in coastal ecosystems is becoming more common. Th ese compounds are generally lipids including sterols, fatty acids, and hydrocarbons. The ways these organic compounds are used vary because the compounds vary in their specificitysome can identify groups of organisms such as vascular plants or algae while others may be specific to one genera or species (C anuel et al. 1997). Less specific biomarkers can be used in conjunction with stable isotopes to further differentiate sources from general groups (i .e. vascular plants into C3 and C4 groups). Many studies used biomarkers in concert with bulk stable isotopes (Hernandez et al. 2001 ; Wang et al. 2003) or measured the stable isotopic composition of biomarker compounds in sources and sediment ( Canuel et al. 1997; Bull et al. 1999; Herna ndez et al. 2001; Mead et al. 2005). This method uses a gas chromatography (GC) to determine lipids after a complex and lipid-type specific extraction proce ss. The different lipids are se parated by the GC column due to their different retention time within the co lumn. Different lipids can be identified by comparing their relative retention times on the resulting gas chromatogram with relative retention times on a gas chromat ogram of a known standard. Rela tive amounts of ea ch lipid can also be determined by calculating areas undern eath each peak on the chromatogramlarger areas correspond to a larger amount of that lipid in the sample. Is otopic signatures of these lipids can be determined when the GC is connected to a mass spectrometer in a method known as

PAGE 44

44 isotope ratio-monitoring gas chromatographymass spectroscopy (irm-GCMS) (Canuel et al. 1997). Canuel et al. (1997) examined the usefulness of isotopic signatures of specific lipid compounds to identify SOC sour ces in coastal ecosystems. The study examined isotopic signatures of bulk organic matter, total lipid extr acts, and a whole suite of lipid compounds in three vascular plants, S. alterniflora, J. roemerianus and Zostera marina, suspended particulate matter (SPM), and sediment in North Carolina Vascular plants had similar molecular compositions of sterols and fatty acids but diffe red in hydrocarbon compositions, specifically in ranges and maxima of n -alkanes and the presence of m onosaturated alkenes. SPM had a different lipid composition than vascular plan ts; the majority (>50%) of SPMs hydrocarbons were C25 highly branched isoprenoids (HBI) alkenes. Among different vascular plant lipids there was a variety of isotopic si gnatures with an average de pletion of 3-5 in lipid 13C values relative to bulk values. Lipids in Z. mostera S. alterniflora and J. roemerianus followed the same trend in 13C values as bulk tissues with mean 13C values (in ) of -14.8 to -18.9, -18.4 to -22.6, and -29.0 to -33.8 for lipid s and of -10.0, -12.6, and -26.0 fo r bulk tissues, respectively. Differences between bulk and lipid signatures demonstrated another reason caution should be used in analyzing bulk isotopic studies because compounds preserved in SOC may not have the same signature as bulk plant matter (Canuel et al. 1997). 13C values of lipids in sediments were different than those for the same lipids in vascul ar plants, but similar to SPM lipids. Sediments had a higher diversity of lipids than vascular plants, small amounts of major vascular plant biomarkers like C21 and C29 (maxima observed in the Z. mostera and S. alterniflora tissues), and a lot of the major SPM biomarker, C25 HBI. The study concluded vascular plants were only minor contributors to SOC.

PAGE 45

45 The North Carolina study shows how the mol ecular distribution (div ersity of types and dominant types) of lipids and isotopic signatures of lipid s can be used to determine SOC sources. The method was not without problems, however. Use of compound classes that may not be the most diagnostic for vascular plants may have sk ewed results. Furthermore, this technique is biased toward extractable, not bound, lipids in sedi ments (Canuel et al. 1997) It is important to note that only the top 0.5 cm of sediment was analyzed in this study. Thus, depths where macrophyte roots may contribute to SOC through e xudates or senescent ti ssue were ignored. Not all studies examine a wide range of lipid s. It is common to examine only one lipid class such as n-alkanols (Bull et al. 1999) or n-al kanes (Wang et al. 2003). Studying the 13C values of one type of lipid biomarker can help solve the intermediate-value problem that muddles analysis of SOC sources in bulk isotope studies The use of compounds that are specific to vascular plants (n-alkanes) or to plankton and algae (HBI alkenes; Canuel et al. 1997) for isotope studies can help clarify their contributions to SOC (i.e.: whether a sediment bulk 13C value of 18 is due to an even mix of C4 and C3 plants, only plankton, or a mixture of all three). The nalkanol homologue, C32, was chosen for a study a ddressing contributions of S. alterniflora and Puccinellia maritima to salt marsh SOC in the United Ki ngdom (Bull et al. 1999). By only examining 13C values of an n-alkanol, which pla nkton cannot produce, planktons confounding intermediate 13C value was removed as a factor from the isotopic mixing equation. Using a two-member mixing model based on 13C values of the C32 n-homologue, contributions of S. alterniflora to SOC were calculated. S. alterniflora contributed about 100% of primary biomass to sediments directly beneath S. alterniflora stands and about 50% to sediments beneath P. maritima stands. To fully understand all sources to SOC, this method should be expanded to include group-specific biomarkers for both vasc ular plants and pla nkton. Otherwise when

PAGE 46

46 analyzing only 13C values of a biomarker for one plant group, contributions of plants outside that group remain unknown. Mead et al. (2005) took the examination of group-specific series of homologues a step farther when they used an n-alkane-based proxy, P aq, along with compound-specific stable isotopes to elucidate sources along a gradient of freshwater mars h to estuarine mangrove forests to marine seagrass beds in the Florida Everglades. P aq is calculated from abundances of different n-alkane homologues. 31 29 25 23 25 23CCCCCCPaq (2-2) In equation 2-2, Cx is the amount of the Cx nalkane. Submerged and floating macrophytes like seagrasses contained more abundant mid chain n-alkanes and therefore had a higher P aq than emergent macrophytes and terrestr ial plants like mangroves. This method was able to resolve sources to a greater extent th an studies using only isotopes or only biomarkers in estuaries because sources with similar P aq values were differentiated using n-alkane 13C values and vice versa. Generally, as the gradient went from freshwater marsh to seagrass beds there was a trend of increasing sediment 13C values and increasing P aq values. These trends were further connected to contributions of individual sources through a PCA based on compound specific 13C and Paq. An example using a species-specific biomarker is the study of different homologues of the n-alkane-2-ones lipid series to elucidate SOC sources in the Harney River estuary and the adjacent Florida shelf (Hernandez et al. 2001; Mead et al. 2005). Lipids in this series generally have odd-numbered C chains ranging from 19 to 33 Cs in length. There has been some debate about whether n-alkane-2-ones aris e is sediments directly from pl ant detritus or whether they arise from microbial oxid ation of alkanes. However, Hernandez et al. (2001) were able to find

PAGE 47

47 significant amounts of n-alkane-2-ones in tissues of seagrasses and mangroves. In seagrasses, the most common (82% to 88% of the ket one fraction) n-alkane-2-one was the C25 homologue, and in mangroves the most comm on n-alkane-2-ones were the C27-C31 homologues. Gas chromatograms of sediments in the lower estuary and the shoreward section of the Florida shelf showed a predominance of seagrass-derived C25 homologues, implying that seagrass was a major SOC source there. In upper estuarine sediment s, there was a predominance of higher molecular weight homologues implying mangroves were the ma jor SOC source. Isotopic measurements of bulk SOC and n-alkane-2-ones confirmed these c onclusions about primary SOC sources because sediment 13C values became more enriched (i.e. mo re like seagrass-derived SOC) as the samples went from the upper estuary to the Flor ida shelf. By using biomarkers specific to vascular plants, however, contributions to SOC from algae and plankton were unknown. As with the use of bulk stable isotope measurements, there are caveats with the use of lipid biomarkers. First, this method has not been as extensively studied as the use of bulk isotopes. Inherent variation of molecula r distributions and compound speci fic isotope signatures within different tissues of an individual plant, within plants of the same species, across geographical areas, and across seasons has yet to be documented (Canuel et al. 1997 ). Also, the more specific biomarkers may not be applicable to all species of the same plant type. The temperate seagrass, Z. marina, did not have the predominant C25 n-alkane-2-one homologue that sub-tropical seagrass species had (Hernandez et al. 2001). Not all major ecosystem components will have appropriate species-specific biomarkers, so a co mbination of species-specific and group-specific biomarkers may have to employed (Mead et al. 2005). Biomarkers confirm the presence of a certain source in SOC, but they do not necessarily yield relative contributions of sources because not all sources are represented in each lipid type. Just becau se the isotopic mixing equation

PAGE 48

48 indicates that 50% of a biomarker is from S. alterniflora and the other 50% is from P. maritima does not mean each species contributes to 50% of the bulk SOC. Furthermore, when examining isotopes of group-specific lipids, one must make sure that the lipids being examined have similar abundances in each species. Otherwise, what seems like a greater abundance of one species in SOC might actually be due to a gr eater abundance of that biomarke r in tissues of that species relative to other species. In cases where biomarkers are not likely to be at the same concentration among species, relative abundances of biomarkers within each plant should be included in isotopic mixing e quations (Bull et al. 1999). Petrographic Analysis Petrographic analysis of sedim ents is a le sser-used method to determine SOC sources. Petrographic analysis microscopically examines organic matter for recognizable organic components such as macrophytic tissues, differe ntiated based on their level of decomposition, and algae. Marchland et al. ( 2003) examined SOC sources in ma ngrove forests of various ages using six different categories of plant tissues: Translucent ligno-cellulosic debris (TLC), which exhibited preserved cell wall stru ctures, degraded ligno-cellulosic debris (DLC), which exhibited decaying cell walls, gelified particles (GP), whic h were orange brown gel-like particles produced by cellulose degradation, reddish amorphous orga nic matter (RAOM), in which the cellulose is completely degraded, oxidized opaceous ligno-c ellulosic debris (OLC), which were dark and structureless refractory landderived OM, and grayish amorphous organic matter (GAOM), which were the remains of algae and phytoplankton. This study looked at relative proportions of these various components to understand whethe r SOC sources to mangrove forests were autochthonous algae, mangroves, or allocthonous ri verine detritus. Combining proportions of these OM components with C:N ratios, they f ound that sediments of younger mangrove forests, with their low C:N ratios and high proportion of GAOM, were dominated by algal-derived OM

PAGE 49

49 and that more mature forests, with their hi gher C:N ratios and ligno-cellulosic debris, were dominated by mangrove-derived OM. They also found that the upper sediments of the younger forests and the deeper sediments of the older forests had a lot of OLC, indicating a trapping of allochthonous OM from the river. With this method, known proportions of various OM components can be measured directly, in cont rast to isotopic methods. However, OM components such as TLC and GP cannot be direct ly attributed to any one species of primary producer, but rather to broad clas ses of producers. This study did not take seagrasses into consideration, which may have similar-looking part ially decomposed ligno-cellulosic tissues as mangroves, making it hard to differentiate be tween those two sources using this method. Nuclear Magnetic Resonance Spectroscopy Nuclear m agnetic resonance spectroscopy (NMR), specifically 13C-NMR, is another method that can be used to identify SOC sour ces. However, no known studies document this method in seagrass, mangrove, or salt marsh sedime nts, and therefore an estuarine study is used to illustrate this method. In NMR spectroscopy, a sediment sample is subjected to a magnetic field which causes the nuclei in the 13C atoms to precess as a gyros cope does (Stevenson 1994). A second alternating magnetic field is then a dded, and when the frequency of that second magnetic field matches the frequency of the nuc leis precession, the nuclei of the atom then resonate causing a voltage change that is amplif ied and recorded. A spectrum is produced from this resonance signal. The nuclei resonate at different frequencies depe nding on their chemical environment. Each sample resonates at severa l frequencies, and from the different resonance signals, spectra with several dist inct peaks are produced. Spectr a are plotted using the chemical shiftthe difference between resonance frequencie s of samples and the resonance frequency of a standard, tetramethylsilane (TMS) solution. This chemical shift cal culation is analogous to the calculation of 13C values based on how much the samples values differ from the PDB standard.

PAGE 50

50 Each organic structure, such as an aromatic ring or a carboxyl group, ha s a different resonance subsequent chemical shift; therefore this method allo ws scientists to assign categories of OM to specific chemical shifts (Goldi ng et al. 2004). Using this met hod, types and relative amounts of OM structures in sediments can be elucidated. Golding et al. (2004) used 13C-NMR to study whether SOM was terrestrialor marinederived in upper (fluvial) and lowe r (marine) sections of Australian estuaries. They studied four groups of organic carbon structurescarbonyls, aromatics, O -alkyls, and alkyls. They associated both O -alkyl C and aromatic C with terrestrial pl ant sources because they assumed O alkyl C was from cellulosic ca rbohydrates and aromatic C was from lignin and tannins of vascular plants. The presence of alkyl C indicated marine origins because they associated it with planktonic material. They cautioned, however, that alkyl C may also be present due to microbial decomposition of terrestrial OM. The authors concluded that upper porti ons of estuaries had higher proportions of O -alkyl C and aromatic C, and therefore higher amounts of terrestriallyderived SOC, than lower portions of estuaries. NMR has similar problems as petrographic an alysis because structures being studied cannot be directly assigned to specific primar y producers; the relationship between the producer and the structure must be inferred, and one structur e type can come from se veral producers. This technique may be best suited to situations where sources are grouped into a couple of components such as a study of seagrass/mangrove-derived SOC and planktonic SOC. Despite problems associated with SOC source determination using 13C-NMR, this tool may help scientists better el ucidate roles of plankton and algae, w hose proportions in SOC are difficult to determine via stable isotopes because of their variable and intermediate 13C values. This method also helps scientists und erstand the OC structures not just the OC sources in coastal sediments.

PAGE 51

51 Conclusion This review sought to cover SOC topics relevant to this thesis research in three types of coastal ecosystem ssalt marshes, mangrove forests, and seagrass beds. The original research in this thesis covers OC pools and sources in a constructed mangr ove and seagrass system. This review covered salt marshes in order to add both depth and breadth because C accumulation and constructed ecosystem development are curren tly better understood in salt marshes than in mangrove forests and seagrass beds. This re view discussed C accumulation rates of salt marshes, mangrove forests, and seagrass beds, functi onal trajectories of OC a ttributes in restored and constructed salt marshes, and SOC source de termination methods in salt marshes, mangrove forests, and seagrass beds. The C accumulation section showed that these coastal ecosystems are globally significant C sinks. The functional tr ajectory section showed how OC functions in constructed salt marshes and emphasized the need for further and more in-depth studies of OC in constructed coastal ecosystems. The SOC source determination section showed pros and cons of different SOC determination methods, including bulk stable isotopes, which are utilized in the original thesis research.

PAGE 52

52 Table 2-1. Global area of mangrove forests, salt marshes, and seagrass beds. System Global area (km2) Data sourcea Mangrove forests 200000 1 218000 2 240000 3 Salt marshes 300000 2 400000 4 Seagrass beds 300000 5 600000 6 a1, Jennerjahn and Ittekkot 2002; 2, Twilley et al. 1992; 3, Mitsch and Gosselink 2000; 4, Duarte and Cebrian 1996; 5, Suzuki et al. 2003; 6, Hemminga and Duarte 2000.

PAGE 53

53 Table 2-2. Global rates of carbon accumula tion in coastal ecosystem sediments. System Average type Global areal rate (g C m-2 yr-1) Global rate (Tg C yr-1) Data sourcea Mangrove forests Mean1 92 20 1 Mean 100 22 2 Estimate2 200 44 3 Mode1 115 25 4 Salt marshes Mean 50-5000 17.5-1750 5 Mean 100 35 1 Mean 175 61 2 Mode 115 25 4 Intertidal3 Mean 210 120 6 Seagrass beds Estimate4 1.2 0.54 7 Mean 133 60 2 Estimate 270 122 8, 9 Mode 36.5 16.5 4 Open ocean Mean 0.22 170 2 Terrestrial systems5 Tundra 0.2-2.4 10 Temperate forest 0.7-10 10 Tropical rainforest 2.3 10 Temperate grassland 2.2 10 Temperate desert 0.8 10 a1, Twilley et al. 1992; 2, Duarte and Cebrian 1996; 3, Jenne rjahn and Ittekkot 2002; 4, Cebrian 2002; 5, Rabenhorst 1995; 6, Chmura et al. 2003; 7, Suzuki et al. 2003; 8, Duarte and Chiscano 1999; 9, Hemminga and Duarte 2000; 10, Schlesinger 1990. 1Both the mean and mode numbers were derived from compiling numbers from published studies. 2The estimates were either scaled up from a single study or derive d from a rough back-of-the -envelope calculation. 3Number includes contribution of both mangrove forests and salt marshes. 4Estimate is of amount being exported and subsequently buried in th e open ocean sediments, not in situ accumulation. 5These numbers represent long term accumulation rates measured since the end of the last ice age.

PAGE 54

54 Table 2-3. Rates of carbon accumulation in coasta l ecosystem sediments and the methods used to calculate the time co mponent of the rates. Location Site C accumulation (g C m-2 yr-1) Time scale Method Data sourcea Mangrove forests Herbert River Estuary, Australia 180 Century 210Pb profiles 1 Jiulonglijang Estuary, China High intertidal 168 Century 210Pb profiles 2 Mid intertidal 204 Century 210Pb profiles 2 Low intertidal 841 Century 210Pb profiles 2 Florida Keys, Florida Rhizophora mangle 1591 30 Year 137Cs profiles 3 Avicennia germinans 1051 30 Year 137Cs profiles 3 Rookery Bay, Florida Fringe 2281 Annual Feldspar marker 4 Basin 3281 Annual Feldspar marker 4 Exposed island 2911 Annual Feldspar marker 4 Sheltered island 1911 Annual Feldspar marker 4 Matang Forest Preserve, Malaysia 150 8,000 Year Estimate 5 5-yr-old stand 101 Century 210Pb profiles 6 18-yr-old stand 110 Century 210Pb profiles 6 85-yr-old stand 127 Century 210Pb profiles 6 Celestun Lagoon, Mexico 55-70 Century 210Pb profiles 7 Chelem Lagoon, Mexico 67-104 Century 210Pb profiles 7 Terminos Lagoon, Mexico 33 Century 210Pb profiles 7 Sawi Bay, Thailand 184-281 Decadal 137Cs and 210Pb profiles 8 Brackish Marshes Cameron Parish, Louisiana Natural waterway 7001 Annual Feldspar marker 9 Restricted canal 351 Annual Feldspar marker 9 Restricted natural waterway 301 Annual Feldspar marker 9 Fina La Terre, Louisiana Unmanaged 751 Annual Feldspar marker 10 Managed 101 Annual Feldspar marker 10 Rockefeller Refuge, Louisiana Unmanaged 3351 Annual Feldspar marker 10

PAGE 55

55 Table 2-3. Continued Location Site C accumulation (g C m-2 yr-1) Time scale Method Data sourcea Salt marshes Upper Bay of Fundy, Canada Low marsh 39 30 Year 137Cs profiles 11 High marsh 194 30 Year 137Cs profiles 11 Outer Bay of Fundy, Canada Low marsh 76 30 Year 137Cs profiles 11 High marsh 188 30 Year 137Cs profiles 11 St Marks NWR, Florida Low marsh 117 12 Year 14C bomb uptake2 12 Mid marsh 101 12 Year 14C bomb uptake2 12 High marsh 65 12 Year 14C bomb uptake2 12 Low marsh 25 400-600 Year 14C profiles2 12 Mid marsh 22 400-600 Year 14C profiles2 12 High marsh 20 400-600 Year 14C profiles2 12 Lafourche Parish, Louisiana Continuous canal 3001 Annual Feldspar marker 10 Discontinuous canal 2001 Annual Feldspar marker 10 Natural waterway 6501 Annual Feldspar marker 10 Cedar Creek, Maryland 89 150 Year 210Pb profiles2 13 18.5 Millennia 14C profiles2 13 Hell Hook, Maryland 78 150 Year 210Pb profiles2 13 39.8 Millennia 14C profiles2 13 Barnstable, Massachusetts 96 NA Modeled 14 Biloxi Bay, Mississippi 180 Decadal 137Cs profiles 3 Waarde Marsh, Netherlands 105 NA Modeled 14 Consultant, North Carolina 3-yr-old constructed 39 3 Year OC / Time2 15 Natural reference 35-51 Decadal 137Cs and 210Pb profiles 15 Drum Inlet, North Carolina Bare spoil 80 16 Month OC / Time3 16 Spoil planted with Spartina 87 16 Month OC / Time3 16 Fertilized spoil with S p artina 96.8 16 Month OC / Time3 16 Dills Creek, N orth Carolina 13-yr-old constructed 62 13 Year OC / Time3 15

PAGE 56

56 Table 2-3. Continued. Location Site C accumulation (g C m-2 yr-1) Time Scale Method Data Sourcea DOT, North Carolina 1 yr-old constructed 99 1 Year OC / Time3 15 Natural reference 30-36 Decadal 137Cs and 210Pb profiles 15 Jacob's Creek, North Carolina Irregularly-flooded streamside 146 Decadal 137Cs profiles 17 Irregularly-flooded backmarsh 107 Decadal 137Cs profiles 17 Marine Lab, North Carolina 26-yr-old constructed 34 26 Year OC / Time3 15 Natural reference 15 Decadal 210Pb profiles 15 Oregon Inlet, North Carolina Regularly-flooded streamside 58.9 Decadal 137Cs profiles 17 Regularly-flooded backmarsh 21.3 Decadal 137Cs profiles 17 Pine Knoll Shores, North Carolina 21-yr-old constructed 125 11 Year OC / Time4 18 Natural reference 115 11 Year OC / Time4 18 Port, North Carolina 8-yr-old constructed 27 8 Year OC / Time3 15 Natural reference 28-32 Decadal 137Cs and 210Pb profiles 15 Snow's cut, North Carolina 25-yr-old constructed 99 11 Year OC / Time4 18 Natural reference 159 11 Year OC / Time4 18 Swansboro, North Carolina 11-yr-old constructed 18 11 Year OC / Time3 15 Natural reference 105-115 Decadal 137Cs and 210Pb profiles 15 Aransas NWR, Texas 1671 Decadal 137Cs profiles 3 San Bernard NWR, Texas 2071 Decadal 137Cs profiles 3 United States Average 83 NA Compiled 19 Seagrass Beds Aburatsubo Bay, Japan 1.2-1.55 NA Modeled 20 Celestun Lagoon, Mexico 40 Century 210Pb profiles 7 Terminos Lagoon, Mexico 53-65 Century 210Pb profiles 7 Cala Culip, Spain 19-191 600 Year Shipwreck 21 Fanals Point, Spain 182 Annual Sediment trap 22 a1, Brunskill et al. 2002; 2, Alongi et al. 2005; 3, Callaway et al. 1997; 4, Cahoon and Lynch 1997; 5, Ong 1993; 6, Alongi et al. 2004; 7, Gonneea et al. 2004; 8, Alongi et al. 2001; 9, Cahoon and Turner 1989; 10, Cahoon 1994; 11, Connor et al. 2001; 12, Choi and Wang 2004; 13, Hussein et al. 2004; 14, Middelburg et al. 1997; 15, Craft et al.

PAGE 57

572003; 16, Cammen 1975; 17, Craft et al. 1993; 18, Craft et al. 1999; 19, Hopkinson 1988; 20, Suzuki et al. 2003; 21, Romero et al. 1994; 22 Gacia et al. 2002. 1This author reported organic matter accumulation rates, so rates were divided by 2 in order to obtain these numbers. 2Modeled sediment profiles instead of measuring them directly. 3Calculated by subtracting the OC in 030 cm from the OC in top 10 cm, divided by the age of the site 4Calculated by subtracting the OC at time 0 from the OC at time 1, divided by time 1-time 0 5Denotes carbon buried after exportation to the open ocean not in situ.

PAGE 58

58Table 2-4. Studies comparing organi c carbon in restored and constructed coastal mars hes to OC in natural reference marshes. Location Site Age (years) Constructed OC (units) Reference OC (units) Depth sampled (cm) Methoda Sourceb Tacoma, Washington1 Gog-Le-Hi-Te, Site 1 1 3.5 % 3.3 8.7 % 0-2 1 1 2 3.0 3 4.0 6 3.5 Gog-Le-Hi-Te, Site 2 1 4.0 2 4.5 3 5.5 6 9.0 Gog-Le-Hi-Te, Site 3 1 2.5 2 2.0 3 2.2 6 1.2 Gog-Le-Hi-Te, Site 4 1 2.0 2 2.0 3 3.0 Maine, New Hampshire1,2 Great Bay Estuary 1 2.0 % mean = 23 % 0-5 1 2 2 1.5 3 3.0 6 2.5 14 16 Core Banks, North Carolina Sound-side marsh, site 1 0 77.3 g OC m-2362.7 g OC m-20-13 2 3 1.3 184.3 Sound-side marsh, site 2 0 77.3 1.3 193.3 Sound-side marsh, site 3 0 77.3 1.3 206.4 San Diego, California1 San Diego Bay 2 3.5 % 7.5 11 % Not reported 1 4 4 5.5 8 7.5 11 7/0

PAGE 59

59Table 2-4. Continued Location Site Age (years) Constructed OC (units) Reference OC (units) Depth sampled (cm) Methoda Sourceb Georgia Sappelo Island 42 1264 g C m-2 1372 g C m-2 0-10 1 5 North Carolina Pamlico River Estuary 5 886 kmol C ha-110270 kmol C ha-1 0-30 3 6 15 1866 Virginia Gloucester Point 5 95 g C m-2 129 163 g C m-2 0-2 1 7 12 120 5 50 146 174 14-16 12 53 North Carolina2 DOT 1 400 g C m-2 3800 g C m-2 0-30 1 8 Consultant 3 600 4600 Port 8 900 2000 Swansboro 11 1000 4600 Dills Creek 13 1800 4900 Pine Knoll 24 1200 1000 Marine Lab 26 2900 5100 Snows Cut 28 2900 10000 a1, loss-on-ignition; 2, Walkley-Black oxidation; 3, CHN analyzer. b1, Simenstad and Thom 1996; 2, Morgan and Short 2002; 3, Cammen 1975; 4, Zedler and Calloway 1999; 5, Craft 2001; 6, Craft et al. 2002; 7, Havens et al. 2002; 8, Craft et al. 2003 1Signifies study measured organic matter (OM) only, not organic carbon (OC). 2Signifies study did not measure the same wetland overtime but instead used a space-for-time substitution.

PAGE 60

60Table 2-5. Stable Isotope values and dominant source conclusions from carbon source determination studies in coastal ecosystems Location Potential sources Source 13C Site description Sediment 13C Main sourcesa How main sources determined Data sourceb Mangrove forest Eastern Brazil Seston Spartina Mangroves -21 to -221-26.82 -27 Mangroves Riverine Shelf Slope -26.9 -23.8 -21.3 -20.5 4 4 2 2 Comparison 1 Gazi Bay, Kenya Mangroves -28.25 Rhizophora mucronata -25.3 4 Comparison 2 Gazi Bay, Kenya Mangroves -24.12 Ceriops tagal -22.7 4 Comparison 2 Southeast Asia Seston Mangroves -20.5 to -233-27 to -293 Coringa Wildlife Sanctuary, India Galle, Sri Lanka Pampala, Sri Lanka -29.4 to -20.63 2 4 4 Compared to curve of 2 source mixing model 3 Mangroves and salt marsh Chiricahueto, Mexico NR Marsh -20.4 2 Compared to literature values 4 Salt Marsh Florida NR Spartina alterniflora Juncus roemerianus -16.9 -23.9 Comparison 5 Sapelo Island, Georgia Diatoms S. alterniflora S. virginica D. spicata S. virginicus J. roemerianus B. frutescens -17.0 -12.9 -26.0 -13.1 -13.3 -22.8 -26.0 Bare creekbank Tall Spartina Short Spartina S. virginica high marsh Sand flat Mixed vegetation -18.9 -16.0 -17.9 -21.6 -22.6 -19.3 NR NR 5 NR NR NR Comparison 6 Barataria Bay, Louisiana S. Alterniflora -12.1 to -13.6 Marsh -16.2 5 Compared to predicted values based on producer biomass 7

PAGE 61

61Table 2-5. Continued Location Potential sources Source 13C Site description Sediment 13C Main sourcesa How main sources determined Data sourceb Plum Island, Massachusetts S. Alterniflora T. latifolia -13.3 -25.3 Mid marsh Upper marsh Mudflat -18.94 -22.814 -19.394 both both Both Comparison and distributions of long chain n-Alkanes 8 Waarde Marsh, Netherlands Spartina Allochthonous OM -12.72 -25.5 Marsh -22 to -24.6 6 Compared to curve of 2 source mixing model 9 Cape Lookout Bight, North Carolina Seston Seagrass Spartina J. roemerianus -18.4 -10.0 -12.6 -26.0 Fall Spring -17.8 -20.3 2 2 Comparison with lipid distributions and lipid 13C 10 Dorset, United Kingdom Spartina anglica Pucinella maritima -12.1 -26.9 S. anglica P. maritima Mudflat -17.6 -21.4 -20.4 5 5 (50%) 5 (40%) Mixing model using compound specific 13C 11 Barnstable, Massachusetts Spartina Allochthonous OM -12.51 -25.51 High marsh Low marsh -13.4 to -14.5 -21 to -19.5 5 6 Compared to curve of 2 source mixing model 9 Seagrass beds Gazi Bay, Kenya Seagrass Mangroves Sediment Traps -19.7 -26.75 -23.3 Closest to mangroves -22.9 4,1 Comparison 12 Gazi Bay, Kenya Seagrass Mangroves POM -18.3 -26.75 -22.5 Closer to mangroves -20.6 4,1 Comparison 2 Gazi Bay, Kenya Seagrass Mangroves POM -15.8 -26.75 -19.2 Farther from mangroves -18.5 1 Comparison 2 Gazi Bay, Kenya Seagrass Mangroves POM -10.70 -26.75 -13.7 Farthest from mangroves -15.14 1 Comparison 2

PAGE 62

62Table 2-5. Continued Location Potential sources Source 13C Site description Sediment 13C Main sourcesa How main sources determined Data sourceb Chale Lagoon, Kenya Seagrass Mangroves -10.72 -26.75 -14.8 1 Comparison 12 Silaqui, Philipines Seston Seagrass Epiphytes -16.4 -5.8 -9.6 -10.3 3 Percent contribution ranges from mixing equation 12 Pislatan, Philipines Seston Seagrass Epiphytes -16.5 -7.5 -10.5 -14.9 2 Percent contribution ranges from mixing equation 12 Spain Seston Seagrass Epiphyte -22.13 -12.42,3 -17.83 Iberian Coast Balearic Islands -15.8 to -21.63 -15.8 to -21.63 2,1 (3?) 1,2 (3?) Percent contribution ranges from mixing equation 13 Fanals Point, Spain Seston Seagrass Epiphyte POM -24.7 -12.2 -17 -21.5 -20.07 2 Percent contribution ranges from mixing equation and microscopic examinations 14 Can Rhan Lagoon, Vietnam Seston Seagrass -19.6 -8.6 -18.6 NR Percent contribution ranges from mixing equation 12 Dam Ghia Bay, Vietnam Seston Seagrass Epiphyte -17.7 -6.0 -8.6 -15.8 2 Percent contribution ranges from mixing equation 12 Mi Giang II, Vietnam Seston Seagrass -12.1 -7.6 -13.2 NR Percent contribution ranges from mixing equation 12 Seagrasses and mangroves Celestun, Mexico Seston Seagrass Mangrove -22.1 -16.12 -28.62 Fringing mangrove Lagoon center -24 -20 4,2 1,2 Ternary mixing diagram of 13C and N:C 15

PAGE 63

63Table 2-5. Continued Location Potential sources Source 13C Site description Sediment 13C Main sourcesa How main sources determined Data sourceb Chelem, Mexico Seston Seagrass Mangrove -22.1 -15.42 -27.12 Fringe mangrove Seagrass bed -23.1 to -26.16 -17.2 to -22.46 1,2 4,2 Ternary mixing diagram of 13C and N:C 15 Terminos, Mexico Seston Seagrass Mangrove -25.3 -11.9 -28.62 Fringe mangrove Seagrass bed -26 -16 4,2 1,2 Ternary mixing diagram of 13C and N:C 15 Santa Barbara, Philipines Seston Seagrass Epiphyte Mangrove -19.0 -10.9 -12.9 -28.6 Seagrass bed -22.7 4 Percent contribution ranges from mixing equation 12 Buenavista, Philipines Seston Seagrass Epiphyte Mangrove -17.7 -11.7 -13.1 -28.1 Seagrass bed -15.7 1 Percent contribution ranges from mixing equation 12 Umalagan, Philipines Seston Seagrass Epiphyte Mangrove -27.6 -12.3 -22.9 -28.4 Seagrass bed -26.6 2 or 4 Percent contribution ranges from mixing equation 12 Khung Krabaen Bay, Thailand Seston Seagrass Macroalgae Mangrove Shrimp feed -20.62 -10.5 -15.65 -28.82,5 -22.5 Canals Mangroves Inner bay Mouth of bay Offshore -26.5 -26.3 -15.1 -19.2 -17.5 4 4 1,7 2 2 Comparison 16 Ghia Luan, Vietnam Seston Seagrass Mangrove -21.6 -13.3 -27.9 Seagrass -24.6 4 Percent contribution ranges from mixing equation 12 aThe numbers in the main sources column signify the following: 1, seagrass; 2, seston; 3, epiphytes; 4, mangroves; 5, Spartina; 6, other; 7, macroalgae. b1, Jennerjahn and Ittekkot 2002; 2, Hemminga et al. 1994; 3, Bouillon et al. 2003; 4, Soto-Jimenez et al. 2003; 5, Johnson and Calder 1973; 6, Haines 1976; 7, Chmura et al. 1987 ; 8, Wang et al. 2003; 9, Middelburg et al. 1997; 10, Canuel et al. 1997; 11, Bull et al. 1999; 12, Kennedy et al. 2004; 13, Papadimitriou et al. 2005; 14, Gacia et al. 2002; 15, Gonneea et al. 2004; 16, Thimdee et al. 2003.

PAGE 64

641The values were not measured in the study and were taken from published values in the literature. 2Averaged values of leaf, root, rhizome and litter tissue or across sites to obtain one stable isotope value. 3Average or range of entire study b ecause the authors did not provide th e specific values for each site. 4Averaged values of e ach 2 cm section in the top 10 cm of sediment. 5Average of several species. 6Range taken from a graph. NR = not reported

PAGE 65

65 CHAPTER 3 SEDIMENT ORGANIC CARBON STORAGE I N A CONSTRUCTED MANGROVE AND SEAGRASS SYSTEM Introduction Coastal ecosystem s such as salt marshes, mangrove forests, and seagrass beds are being degraded and lost worldwide as a result of the eutrophication, sedimentati on, and destruction that accompany coastal development for human habitation, agriculture, and aquaculture (Valiela et al. 2001; Kennish 2002; Zedler 2004). In the United St ates development and infilling are the main causes of coastal ecosystem loss (Dahl 2000). In the last two decades, humans have caused the loss of 18% of the known worldwide area of seag rass beds (Green and Short 2003), and in the last five decades, have caused the loss of about 35% of the worlds mangrove forests (Valiela et al. 2001; Alongi 2002). In the United States, about 50% of salt marshes have been lost historically (Kennish 2001) a nd 25% of mangrove forests have been lost since the 1950s (Bridgham et al. 2006). United States seagrass beds had a relatively constant area between 1986 and 1997 in, what is to our knowledge, the onl y nationwide seagrass i nventory (Dahl 2000). Smaller scale studies, however, have demonstrated local declines in the extent of seagrasses (Zieman et al. 1999; Short et al. 2006). When coastal systems are lost, we lose not only wildlife habitat, storm surge protection, and economically-im portant fish and shellfish nurseries, but also biogeochemical functions like phosphorus retenti on, denitrification, and carbon (C) sequestration (Alongi 2002; Duarte 2002; Zedler and Kercher 2005). The United States has a policy of no net wetla nd loss that includes coastal wetlands as part of the Clean Water Act (Zedler 2004; Zedler and Kercher 2005). Florida policy applies this nonet-loss principle to seagrass beds as well (Florida Admini strative Code, Chapter 18-21). Destruction of mangrove and seagrass ecosystem s in Florida requires compensatory mitigation via restoration of an existing ecosystem or construction of a new ecosystem. Mitigation can

PAGE 66

66 result in the replacement of fully functioning eco systems with ineffective surrogates that do not provide the same functional value (Zedler 2004). Success of most mitigation projects is judged on the survival of macrophytes, not on proper func tioning of the ecosystem. With the majority of ecosystem functions are not assessed, the true success of mitigation projects is usually unknown. One major function of coastal ecosystems is C sequestration. The value of this ecosystem function is increasing with mounting concern ab out climate change. Anthropogenic release of greenhouse gases like carbon dioxide (CO2) and methane (CH4) through fossil fuel burning and deforestation, and livestock produc tion, respectively, is the major ca use of global climate change (IPCC 2001). Coastal ecosystems dominated by macrophytes including salt marshes, seagrass beds, and mangrove forests are high producti ve habitats that act as sinks for CO2 and therefore mitigate climate change. Worldwide, salt mars hes and mangroves store at least 44.6 Pg C in their sediments (Chmura et al. 2003). Seagra ss beds, which make up only 0.15% of the global marine area, account for 15% of the global ma rine organic C (OC) storage (Hemminga and Duarte 2000). Global rates of C sequestration in vegetated marine se diments are estimated between 111 and 216 Tg C y-1 ( Duarte et al. 2005). Based on the low estimate, globally mangroves bury 23.6 Tg C y-1, salt marshes bury 60.4 Tg C y-1, and seagrass bury 27.4 Tg C y-1 (Duarte et al. 2005). In the United States, salt ma rshes store 400 Tg C and sequester 4.4 Tg C y1, and mangroves store 61 Tg C and sequester 0.5 Tg C y-1 (Bridgham et al. 2006); the C stored and sequestered by seagrass systems is unknown. Coastal ecosystems also export C to the oceans where another portion is buried (Duarte et al. 2005). The capacity of coastal ecosystems to sequest er C, like freshwater wetlands, is greater than the capacity of uplands. These wetlands are a natura l C sink, while upland systems

PAGE 67

67 eventually reach an equilibrium where amount of C fixed equals the amount respired annually, if disturbances like fire do not cause a loss of C first (Rabenhorst 1995). Constant accumulation of C in wet ecosystems is due to their anaerobi c sediments where alternate electron acceptors, which are not as efficient as oxygen, must be uti lized to decompose C. The capacity of coastal ecosystems to sequester C is also greater than that of freshwater wetlands (Bridgham et al. 2006). Bridgham et al. (2006) found that estuarine wetla nds sequestered C 10 times faster on an areal basis than other wetlands. These high rates are due to estuarine wetlands high sedimentation rates, high percent soil C, and burial due to sea level rise (Connor et al 2001; Bridgham et al. 2006). Coastal ecosystems have another advantag e over freshwater wetlands. They have lower rates of methanogenesis, so the C they store is not being converted to CH4, a more potent greenhouse gas than CO2. In the United States, freshwat er mineral wetlands emit 2.4 Tg CH4 y-1 while salt marshes and mangroves emit only 0.027 and 0.004 Tg CH4 y-1, respectively (Bridgham et al. 2006). When these coastal ecosystems are impacted, a portion of the biospheres C storage and sequestration capacity is lost, which may ex acerbate climate change by causing more CO2 to be in the atmosphere than would be if these sy stems were intact. Lo ss of vegetated coastal ecosystems has caused at least a 25% decrease in their global C sequestration capacity (Duarte et al. 2005). Bridgham et al. (2006) estimated that losses of salt marshes and mangroves in the conterminous United States have caused a net flux of 402 Tg C y-1 into the atmosphere. The upside is that restoration and constructi on of coastal systems may help mitigate the effects of climate change by increasing C seque stration. For example, if all dyked salt marshes in Canada were restored, an additional 2.4 to 3.6 x 1011 g C y-1 would be sequestered, which would contribute 5% to Canadas CO2 emissions reduction identified in the Kyoto Protocol

PAGE 68

68 (Connor et al. 2001). It is therefore important to know if restoration and construction of coastal systems returns the C accumulation and storage cap acity of these C sinks. Such research can indicate whether mitigation is effective and if coastal wetland restoration can become a policy tool for reducing CO2 emissions as was suggested by Connor et al. (2001). Studies that focus on functional trajectories of OC in restored/constructed syst ems and compare OC between restored/constructed and natural sy stems help answer these questions. Functional trajectories are used to track the progress of constructed systems over time and to compare constructed and reference sy stems (Simenstad and Thom 1996; Zedler and Callaway 1999; Morgan and Short 2002). Functional trajectory st udies examine many ecological attributes that act as indicators of more complex ecosystem functions (Simenstad and Thom 1996; Craft et al. 2003) Attributes reach functional equivalence when they have a value similar to the reference. Functions can follow linear, asym ptotic, and sigmoidal trajectories (Kentula et al. 1993) or no trajectory at all (Zedler and Calloway 1999). Craft et al. (2003) proposed that different attributes follo w one of three trajector ies depending on whether they are part of hydrologic, biological, or soil development processes. OC pool formation is a soil development process, and soil development pr ocesses generally follow the longest trajectory before reaching functional equivalence (Craft et al. 2003). There have been many studies documenting functional trajectories of sediment OC (SOC) or orga nic matter (OM) in restored and constructed tidal marshes (Simenstad a nd Thom 1996; Craft 2001; Havens et al. 2002; Morgan and Short 2002; Craft et al. 2003) but, to our knowledg e, only one in seagrass beds (Evans and Short 2005) and onl y a comparison study in mangrove forests (McKee and Faulkner 2000).

PAGE 69

69 Given the limited scope of these studies, many questions remain unexplored. First, the majority of studies on restored coastal system s have been performed in temperate salt and brackish marshes. Second, these studies only measured SOC or sediment OM as one of a suite of variables and did not deeply examine various SOC pools or character istics. Third, these studies only examined long term trends and not short term changes that may occur immediately following construction of an ecosystem. Whether constructed mangrove and seagrass ecosystems provide the same ecological services as their natural counterpa rts with respect to the C sink, and if the restoration of this service follows a functional trajectory is curr ently unknown. In this study, OC storage in a constructed seagrass and mangrove system in the Indian River Lagoon, FL was examined and its OC storage functioning was compared with the f unctioning of adjacent mature systems. Specific objectives were to: 1) determine whether extrac table OC, microbial biomass C, total OC pools, and OC lability follow a short term trajectory in sediments of a construc ted mangrove forest and seagrass bed and 2) evaluate whether the constr ucted system has reached functional equivalence by comparing SOC between constructed and natural systems. We hypothesize d that, in the short term, SOC storage would increase in the constructed system but would not reach the level of SOC storage in natural systems. Methods Study Site SL 15 (Fig. 3-1) is a m itigation site located in the Indian River Lagoon (IRL) adjacent to Fort Pierce, Florida. It is one of many spoil islands created in the IRL during the construction of the Atlantic Intracoastal Waterw ay that sit several meters above sea level. These islands are populated by many exotics, such as Australian Pine ( Casuvina casuvina ) and Brazilian Pepper ( Shinus terebinthifolius ), in their interiors and by na tive red, black, and white mangroves

PAGE 70

70 ( Rhizophora mangle Avicennia germinans and Laguncularia racemosa ) around their margins. To mitigate destruction of a nearby mangrove forest and seagrass bed, seagrass and mangrove systems were created on SL 15. These systems were created by burning and removing interior vegetation and removing dredge spoil to create several different elevations. The seagrass bed, which remains submerged during low tide, is at the lowest elevation, th e mangrove forest, which is exposed at low tide, is at the middle elevation, and a maritime forest occurs above sea level at the highest elevation. The mangrove fringe of SL 15 was left intact except for a few flushing channels. Between the constructed seagrass and mangrove systems a thin Spartina alterniflora buffer was planted. The mangrove forest was planted with R. mangle and maritime forests were planted with Coccoloba uvifera Borrichia frutescens Rapanea guinensis Conocarpus erectus and Distichlis spicata but seagrasses were left to colonize naturally. Natural systems near SL 15 include its original mangrove forest fringe, surrounding seagrass beds, and mangrove fringes of adjacent spoil islands, which are at least 40 years old. Sediment Sampling Four, 2 m x 2 m plots were established in the mangrove forest and in the seagrass bed on SL 15 (Fig. 3-1). Three, 7 cm in diameter sediment cores from each of these plots were retrieved in November 2005, January (mangrove only), February (seagrass only), May, July, and November 2006. Cores were taken from different areas of the plots each time to ensure an area was not re-sampled. For referenc es, three randomly-selected plots were established in natural mangrove forests and seagrass beds within 1 km of SL 15. These plots were sampled in July and November 2006 using the same procedure as for SL 15 plots. Sediment co res were sectioned in the field and stored in plastic bags on ice for tran sport and then in a 4C re frigerator. SL 15 cores were initially divided into 0-5 cm and 5-10 cm sediment depths. In subsequent samplings, material had accumulated on top of the seagrass bed, which was collected and analyzed

PAGE 71

71 separately from the original sediment depths as an accreted layer. Surface layersfloc from seagrass systems, algal mats from the SL 15 mangrove system, and litter layers from the reference mangrove systemwere collected from each core and were composited for each plot. Differences in color and texture were used to se parate accreted and surfac e layers from original depths except for floc, which was the fraction of the accreted layer that poured off (Fig. 3-2). Average heights of accreted and surface layers were measured for bulk density calculations. One core per plot was retrieved in September 2006 and brought intact to the laboratory for pH and Eh (redox potential) measurements. Laboratory Analyses Sectioned sedim ents and surface layers were weighed to determine bulk density. Rocks, roots, and detritus were removed from the samp le before homogenization, and the volume and weight of large rocks were taken into acc ount when calculating bulk density. After homogenization of each sample, a subsample was weighed to determine moisture content and the remaining sample was split into two parts. On e part (wet sample) was stored in airtight containers at 4C and the other was freeze-dried for 48 hours. Moisture content was determined after subsamples were dried at 105C for 24 hours. Intact cores from September 2006 were incuba ted upright in tanks filled with 25 ppt saltwater made with Instant Ocean (Marineland Labs, Moorpark, CA). Platinum electrodes were inserted into each core at 2.5 cm, at 7.5 cm, at 12.5 cm (reference seagrass only), and halfway through the accreted layer (SL 15 seagrass only). Platinum electrodes stabilized for 24 hours, and then Eh was measured using an Accume t AP71 handheld meter and an Accumet calomel reference electrode. Eh values were corrected re lative to a standard hydrogen electrode. Cores were then sectioned into 0-5 cm, 5-10 cm, a nd 10-15 cm or accreted depths as previously described. pH was measured on 5 g of each de pth using a Fisher Accumet AR50 pH meter.

PAGE 72

72 Total OC (TOC) and total nitrogen (TN) we re measured on freeze-dried sediment and surface layer samples. Freeze-dried samples we re composited by plot and sieved through a 1 mm mesh screen to remove large shell pieces and carbonate rock, which were weighed so their mass could be accounted for in calculations. Samp les were then ball-milled to a fine powder in stainless steel canisters. Inorganic C (IC) in was removed from sample s via vapor acidification (Hedges and Stern 1983; Harris et al. 2001). Samples were wei ghed out into 9 x 5 mm or 10 x 10 mm silver capsules (Thermo Scientific, Waltham MA and CE Elantech, Lakewood, NJ), moistened with deionized water, and placed in an ai rtight container with a beaker of concentrated HCl (12 M ) for 24 hours before being dried at 60C fo r 24 hours. Samples were then rolled and analyzed for OC on an elemental analyzer (ECS 4010, Costech Analytical Technologies, Valencia, CA). Peach leaves (NIST 1547) were used for calibration standards, and sucrose and an internal soil standard were used for quality control. Tests were run on sand samples with various carbonate percentages and total weights to assess the effi cacy of vapor acidification and to determine the maximum sample mass that still ensured complete removal of IC. Furthermore, concurrent measures of 13C were used to confirm complete re moval of IC, and if incomplete IC removal was suspected, samples were rerun at a lo wer total mass. Unacidified samples were run separately in tin capsules (Cos tech) on a Flash EA 1112 series elemental analyzer (Thermo Scientific, Waltham, MA) for TN. Acetinilide was used for ca libration standards, and peach leaves (NIST 1547) and an internal soil st andard were used fo r quality control. Extractable organic C (ExOC) and microbia l biomass C (MBC) were measured using a modified fumigation-extraction procedure (Vance et al. 1987; Joergensen and Mueller 1995). Approximately 5 g of moist sample was weighed out in duplicate for sediment, algal mat, and litter samples and 10 g was weighe d out for floc samples. One set of samples was immediately

PAGE 73

73 extracted with 25 mL of 0.5 M K2SO4 for an hour and then filtered through a Whatman 42 filter. The second set was fumigated in an ethanol fr ee-chloroform atmosphere for 24 hours before being extracted as above. Extracts were dilu ted, acidified, and run for OC on a Shimadzu TOC5050A (Shimadzu North America, Columbia, MD). OC in non-fumigated samples was ExOC. The difference between OC in fumigated and non-f umigated samples, multiplied by a correction factor of 2.22 (Wu et al. 1990; Joergensen 1996; Jenkinson et al. 2004), was MBC. Sediment oxygen demand (SOD; APHA 1992), normalized to TOC, was used as a measure of OC lability. SOD was measured by mixing 10 mg of wet sample with about 300 mL of oxidized, salt water in dark biological oxygen demand (BOD) bot tles. The salt water was created by dissolving Instant Ocean Sea Salt (Mar ineland Labs) into deionized water until the solution reached 25 ppt. Dissolved oxygen (DO) content of the wate r was measured initially and after 24 hours by a Fisher Accume t AR40 DO meter. Measurements were taken after the water and sample in each BOD bottle were thoroughly mi xed on stir plates for 30 minutes. While abiotic and chemotrophic reactions can cause decr eases in DO, these reactions most likely did not cause a significant O2 reduction during this experiment because samples were already exposed to O2 during processing. Furthermore, NH4 + levels in the samples were low (unpublished data) and pH did not change durin g incubation, which would have indicated oxidation of sulfide in the sa mples,. The majority of O2 depletion was therefore assumed to be due to biological, heterotrophic oxidation of OC. OC accumulation rates (g OC m-2 y-1) in SL 15 were calculated using equation 3-1 (Cammen 1975; Craft et al. 1999). system a i f on accumulatiA OCOCOC OC (3-1)

PAGE 74

74 In equation 3-1, OCf is the final amount of TOC (g OC m-2) in the top 0-10 cm, OCi is the initial amount of TOC in the top 0-10 cm, OCa is the amount of TOC in the accreting layer, and Asystem is the age of the system in year s. Without dating sediments using 137Cs, 210Pb, or 14C profiles, OC accumulation rates in reference systems could not be calculated. Statistical Analyses Repeated measures analysis of variance (ANOV As) were run to investigate if parameters in SL 15 sediments and surface layers followed a functional trajectory over time. ANOVAs were run with a spatial power covariance struct ure to account for the unequal spacing between time points. Subjects were SL 15 plots and the re peated factor was month. The 0-5 and 5-10 cm depths were run together in each system in AN OVAs with depth as a main effect. Floc, algal mat, and accreted layers were run separately in ANOVAs. Replicate cores had to be averaged for each plot and month so the data fit the structure required for repeated measures analysis. A parameter followed a trajectory if its repeated measures ANOVA had a significant time effect and it demonstrated an increasing or decreasing (i n the case of bulk density) trend over time. Analyses were run using the mixed procedure in SAS Version 8 (SAS Institute, Cary, NC). Comparisons between SL 15 and reference sites were analyzed using one factorial ANOVA each for the mangrove and seagrass sedi ments and one factorial ANOVA each for the mangrove and seagrass surface layers. Sedime nt ANOVAs consisted of three fixed factors site, month, and depth. Surface layer ANOVAs cons isted of site and month factors. All two way interactions were tested. Plot data we re pooled into two site treatments, SL 15 and reference. July and November 2006 were the months. For seagrass sediment analysis, SL 15 and reference depths were assigned to 3 categories in order to make comparisons: SL 15 accreted and reference 0-5 cm were depth 1, SL 15 5-10 cm and reference 0-5 cm were depth 2, and SL 15 5-10 cm and reference 10-15 cm were depth 3. Factorial ANOVAs only compare the same

PAGE 75

75 depths across different sites and not di fferent depths across different sites (i.e.: it compares SL 15 mangrove 0-5 cm to reference 0-5 cm but not SL 15 mangrove 0-5 cm to reference 5-10 cm), so one-way ANOVAs were also run when site*depth interactions of the factorial ANOVAs did not reveal all interesting trends. Data were averag ed by each site and depth over July and November samplings for these one-way ANOVAs. Factorial and one-way ANOVAs were run on JMP Version 6 (SAS Institute, Cary, NC). For all analyses most data were transformed to meet the normality requirement (see Appendix A for details). Post hoc multiple comp arisons were carried out on significant effects using the Tukey test. Significance was d ecided using an alpha level of 0.05. Results Sediment Characteristics SL 15 sediments (0-5 cm and 5-10 cm) had higher bulk densities than reference sediments (Table 1; site effect, p<0.0001, Table 2) as did the SL 15 mangrove algal mat. The seagrass accreted layer had a bulk density similar to the 0-5 cm depth of the seagrass reference. In seagrass sediments, bulk densities were grea test in the lowest depths, but in mangrove sediments were greater in 0-5 cm depths (T able 1; depth effect, p<0.026, Table 2). SL 15 seagrass sediments had orders of magnitude mo re shell fragments than reference sediments, while SL 15 and reference mangrove sediments ha d similar amounts of shell fragments (Table 1). pH in SL-15 seagrass and reference sediments and in SL-15 mangrove sediments ranged from 8.0 8.3. Reference mangrove sediments had a pH of 7.5 (Table 1). Redox potentials in the upper sediment depths were similar between SL-15 and reference s ites, but were more negative in the lower depths of the reference sediments (Table 1).

PAGE 76

76 Trajectory of Constructed System Parameters measured in SL 15 sediments di d not follow a trajectory over time, except for mangrove sediment bulk density, which signifi cantly decreased with time (month effect, p<0.0001, Table 3, Fig. 3-3). OC parameter values seemed to shift randomly when there were significant monthly changes as fo r MBC in all sediments, and ExOC and TOC in seagrass 0-10 cm sediments (month effect, p<0.021, Table 3, Fig. 3-4). OC parameters followed a pattern in seagrass sediments in which low values occurred in February and July while high values occurred in May and November (Fig. 3-4). TN and C:N also changed without direction when they did change significantly (month effect, p<0.041, Table 3). There were no significant changes in lability for either mangrove or seag rass sediments. Significant differences between depths were few. In mangrove sediments, 0-5 cm depths had greater bul k density and TN, and in all sediments, 0-5 cm depths had greate r lability (depth effect, p<0.031, Table 3). SL 15 surface layers (mangrove algal mat a nd seagrass floc) followed a trajectory of significantly increasing MBC (p <0.0051, Table 3, Fig. 3-5). Extractable OC, TOC, and TN significantly changed with time in floc, with TOC and TN generally increasing (p<0.050, Table 3, Fig. 3-5). C:N significantly ch anged without a trend in floc (p<0.043, Table 3). Lability of OC in the mangrove algal mat significantly increase d with time, while lability in seagrass floc significantly decreased with time (p<0.052, Table 2). Constructed and Reference Comparisons TOC was significantly higher in reference th an in SL 15 mangrove and seagrass systems on both a concentration and storage basis, except in seagrass floc where TOC was similar between sites (site effect, p<0.0005, Table 2; Table 4). TOC differences between sites were greatest in mangrove sediments (Fig. 3-6). On a concentration basis in seagrass sediments, sites had similar TOC in depth one, but had different TOC in depths two and three (site x depth

PAGE 77

77 interaction, p=0.018, Table 2; Fig. 36b). On a storage basis, all layers had lower TOC in SL 15 so there was not a significant in teraction, but a Tukey revealed la yers one and three had similar TOC across sites (Fig. 3-6b; one-way ANOVA, df=5, p<0.0001). In seagrass sediments, TOC was greatest in depth one (depth effect, p>0.013, Table 2; Table 4). TN was significantly higher in reference th an in SL 15 mangrove and seagrass systems (site effect, p<0.0001, Table 2; Tabl e 4), except in seagrass floc wh ere month affected which site had higher TN (site x month interaction, p=0.041, Table 2; Table 4). C:N was significantly higher in the sediments and surface layers of mangrove references but was similar in the sediments and surface layers of seagrass site s (site effect, p<0.0097, Table 2; Table 4). In mangrove sediments, ExOC was significan tly higher in references but, in seagrass sediments, was significantly higher in SL 15 (site effect, p>0.0013, Table 2, Table 5). ExOC (storage basis) of the 0-5 depth in SL 15s ma ngrove system was similar to reference depths while SL 15s 5-10 depth had significantly lower ExOC (site x depth interaction, p=0.058, Table 2; Fig. 3-7a). In the seagrass systems, ExOC (c oncentration basis) was similar in depths two and three across sites while depth one in SL 15 had greater ExOC th an depth one in the reference (site x depth interaction, p<0.0001, Ta ble 2; Fig. 3-7b). On a stor age basis, however, ExOC of depths two and three in SL 15 were higher than the references, but depth one had similar ExOC across sites (site x depth in teraction, p=0.0017, Table 2; Fig. 3-7b). Upper depths had significantly more ExOC in both mangrove a nd seagrass sediments (depth effect, p<0.0038, Table 2; Table 5). Surface layer ExOC did not significantly differ except for seagrass floc where ExOC was greater on a concentration basis in SL 15 (site effect, p= 0.020, Table 2; Table 5). MBC was significantly higher in reference site s for mangrove and seagrass sediments on a concentration and storage basis (site effect, p<0.0001, Table 2; Tabl es 5; Fig. 3-8). In mangrove

PAGE 78

78 sediments, SL 15 0-5 cm depths had similar MBC to reference 5-10 cm depths on a storage basis (Fig. 3-8a; one-way ANOVA, df=3, p<0.0001). On a concentration basis, depths two and three of SL 15 seagrass sediments had significantly lower MBC than those depths in reference sediments, while depth one MBC was similar across sites (Fig. 3-8b; one-way ANOVA, df=5, p<0.0001). On a storage basis, depths two and three had similar MBC across sites, but depth one had significantly lower MBC in SL 15 (site x depth interaction, p<0.0001, Table 2; Fig. 3-8b). MBC was significantly greater in November than in July for both mangrove and seagrass sediments (month effect, p<0.0009 Table 2; Table 5). MBC was significantly greater in upper depths of both mangrove and seagrass sediments (depth effect, p<0.0066, Table 2; Table 5 and 6). Surface layers had similar MBC to resp ective references (Table 2; Table 5). SL 15 systems had significantly greater OC labili ty than reference systems in all sediments and surface layers except for floc (site effect, p<0.013, Table 2; Table 6). Only depth one in seagrass sediments was similar across sites (site x depth interaction, p<0. 0001, Table 2; Table 6). In mangrove sediments, the 0-5 cm depth had sign ificantly greater lability than the 5-10 cm depth while in seagrass sediments, depth two ha d the greatest lability (depth effect, p<0.0027, Table 2; Table 6). In mangrove surface layers, lability of the SL 15 algal mat increased while lability of reference litter d ecreased from July to November (site x month interaction, p<0.0001, Table 2; Table 6). Organic Carbon Accumulation Rates OC accumulation rates in SL 15 sediments were between 168 to 231 g OC m-2 y-1 in the seagrass sediments, but were between -119 to -148 g OC m-2 y-1 in the mangrove sediments. When algal mat accumulations were added to mangrove sediments accumulations, rates ranged from 29 to 236 g OC m-2 y-1. Floc OC accumulations were not added to seagrass sediments due to the transient nature of floc, whic h is easily swept away by currents.

PAGE 79

79 Discussion Sediment Characteristics SL 15 and reference sediments are physically different from one another because SL 15 sediments parent material is dredge spoil, as is apparent from their high amount of shell fragments (Table 1). Furthermore SL 15 sedi ments were compacted during construction. SL 15s accreted layer differs from other SL 15 sedi ments because it is a layer of post-construction deposition and was not compacted by equipment. In comparisons of constructed and reference salt marshes and mangrove forests, bulk density wa s almost always greater in constructed sites (Craft et al. 1999; McKee and Faulkner 2000; Craft et al. 2002). Redox potentials of all sites were negative implying anaerobic conditions and a slow rate of decompos ition. Redox potentials in this study are generally more negative than those found in other mangrove (McKee 1993, Otero et al. 2006) and seagrass sedi ments (Terrardos et al. 1999), and sediment pH in this study are generally higher than in other mangrove se diments (McKee and Faulkn er 2000; Otero et al. 2006) but similar to other seagrass sediment s (Burdige and Zimmerman 2002; Daby 2003). C:N ratios only differed between mangrove c onstructed and reference sediments (Table 4). Lower C:N ratios in the ma ngrove SL 15 sediments are due to their very low TOC. The rest of the C:N ratios are the same between SL 15 and reference sites due to similar proportions of C and N despite SL 15 having lower amounts of C and N overall. In this study, C:N ratios could therefore not be used as the ultimate metric of restoration success as was suggested for salt marshes by Craft (2001). Trajectory of Constructed Site In SL 15 sedim ents, only mangrove bulk densit y followed a functional trajectory in which it decreased within 2 months of construction completion but remained higher than the reference values (Fig. 3-3). This initial decrease may have occurred as these sediments decompressed,

PAGE 80

80 aided by water movement into interstitial space s, once compaction-causing construction ceased and tides could access the site. The seagrass sect ion of SL 15 was completed a month before the rest of SL 15. Seagrass sediments therefore decompressed earlier and may have experienced a similar bulk density decrease before sampling began. OC parameters in SL 15 sediments did not follow a trajectory, although OC pools in SL 15 seagrass sediments seemed to follow a pattern (Fig. 3-4). External seasonal factors, not ecosystem development, were likely the force driv ing these patterns. A review of physical and chemical water column data in IRL from Nove mber 2005 to November 20 06 revealed potential correlations that could explain the pattern (SFWMD 2007; station IRL 36). Lows in OC parameters corresponded with lows in salinity, hi ghs in total Kjeldahl nitrogen, and the lowest (February) and highest (July) wa ter temperatures of the year. Nitrogen probably did not cause these trends because N levels in the IRL are not high enough to be toxic to bacteria, but temperature or salinity may have. If the overl ying water affected OC parameters in seagrass sediments, it explains why mangrove sediments, which are only in contact with water at high tide, experienced the pattern to a much lesser extent. In both SL 15 surface layers, TOC and MBC followed a trajectory where they increased over time (Fig. 3-5). As a surface layer, seagrass floc is more likely to respond to water column changes than sediments. Floc TOC and MBC, however, followed a different pattern than seagrass sediments and IRL salinity and temperatur e. Floc OC pool increases match increases in IRL total suspended solids from February to November 2006 (SFWMD 2007; station IRL 36). Since the floc is mostly water (95%), it is likely that its solids are correlated to water column solids, which include OC substrate and microbes. Algal mat MBC increases are likely due to the algal mats maturation as it became larger and denser throughout the year (per sonal observation).

PAGE 81

81 Overall, during the first year following c onstruction, with the exception of the mangrove algal mat, OC changes in SL 15 are due to s easonality and water quality. These seasonalitycaused changes are large and may obscure any changes due to increasing functions. High interannual variability that mask directional changes has been obs erved in a restored California salt marsh (Zedler and Callaway 1999). SL 15 ch anges were greatest in ExOC and MBC, pools with fast turnover rates. One year may not be ample time to observe changes in more stable OC pools like TOC. Constructed and Reference Equivalence Organic carbon pools A lack of trajectories does not preclude OC on SL 15 from being functionally equivalent to reference OC. Examining depths separatel y, 0-5 cm SL 15 mangrove sediments approached functional equivalence on a storage basis for ExOC and MBC (Fig. 3-7 and 38). Most depths of SL 15 seagrass sediments were at or exceeded functional equivalence for all OC pools on a storage basis (Fig. 3-7 and 3-8). The reason for this equivalence was bulk density. Because bulk density of SL 15 0-10 cm sediment s is greater than reference sedi ments, when OC concentrations are multiplied by bulk density in order to be reported on a storage basis, the resulting parameters in SL 15 are often greater than or equal to th e resulting parameters in reference sediments. Accreted layers were an exception because their bu lk densities were the same as the references and their heights were usually le ss than the references 5 cm. TOC equivalence did not occur on a concentr ation or a storage ba sis in the mangrove sediments but occurred for accreted and 0-5 cm de pths in seagrass sediments. Accreted layers reached equivalence because the material accumula ting from the water column is likely the same material being trapped by seagrasses in reference sediments. It is odd, at first, that 5-10 cm depths reached equivalence before 0-5 cm depths because inputs of OC to SL 15 sediments were

PAGE 82

82 most likely coming the water column and benthic vegetation, which in the first year did not include deeply rooting plants. Th e 5-10 cm depth, however, was not completely dredge spoil. It contained mangrove clay from pr e-construction mangrove areas and a buried A horizon from the seagrass bed that occupied the site before spoil island creation (Fis chler 2006). These other sediments were exposed and mixed with dredge spoil during construction and had more OC than dredge spoil due to their origins in vegetated systems. In the surface layers, the seagrass floc r eached or exceeded equivalence in terms of all OC pools. SL 15 floc may have exceeded refe rence values due to its position inside the mangrove fringe of SL 15. In the subtidal portion of SL 15 there were areas of slower tidal flow that caused settling of water column material (Fischler 2006), which would include OC. The algal mat reached equivalence in ExOC and MB C but not TOC. Lower TOC in the algal mat than in the litter layer is because the litter laye r consisted of higher plant material like mangrove leaves and seagrass that contain more recalcitrant C than algae (Kristensen 1994). Surface layers are first to receive input s that contribute directly and indirect ly to OC pools, such as of light, water column nutrients, and detritu s. Therefore, it is not surprising that most of their OC parameters would reach equivalence within the fi rst year. Upper depths reached OC functional equivalence quickly while lower depths failed to increase over 7 years in a constructed Virginia salt marsh (Havens et al. 2002). The majority of studies that test functional trajectories of TOC or organic matter (OM) in restored and created salt marshes do not see OC r each functional equivalence. In studies that ranged from oneto 42-year-old marshes, only two reached equivalen ce with their natural wetland references in terms of SOC (Simenst ad and Thom 1996, Zedler and Calloway 1999; Craft 2001, Havens et al. 2002, Morgan and Short 2002, Craft et al. 2003). They were 25 (Craft

PAGE 83

83 et al. 2003) and 42 (Craft 2001) years old. These au thors concluded that it takes a long time for restored salt marshes to develop SOC pool equi valence and acknowledged that such equivalence may never be reached. Predictions from salt marsh studies may be valuable for understanding trajectories of constructed mangrove forests because both are inter tidal systems that take a long time to reach equivalence. Sediment OM in a 6-year and 14 -year-old mangrove forest s in Southwest Florida remained at 18 to 32% of reference forest va lues (McKee and Faulkner 2000). SL 15 mangrove sediment TOC was well below that of references. The lack of a TOC trajectory for mangrove sediments contrasts to findings in salt marsh studi es and indicate that not reaching equivalence is a possibility. In all salt marsh studies except one (Simenstad and Thom 1996) a trajectory of increasing OC/OM was documented (Zedler and Ca lloway 1999; Craft 2001; Craft et al. 2002; Havens et al. 2002; Morgan and Short 2002; Craf t et al. 2003). Even a young constructed salt marsh in North Carolina increased its sediment TOC by over 100% in 1.3 years (Cammen 1975). Predictions from salt marshes studies do not work for constructed and restored seagrass beds. OM content of restored sediments was hi gher than one reference and lower than another throughout the first 8 years in the only other known study of seagrass functional trajectories (on the New Hampshire coast, Evans and Short 2005). In SL 15 seagrass sediments, TOC was functionally equivalent in 2 out of 3 depths within a year. There are several reasons why OC in seag rass sediments reach functional equivalence before OC in mangrove forests and in salt marshe s. The first reason is elevation. In several studies of restored and constr ucted salt marshes, soil development was correlated to marsh elevation so that OC/OM was higher at lower elevations (Lindau a nd Hossner 1981; Moy and Levin 1991; Craft et al. 2002). OC equivalence occurs faster at lo wer elevations because they are

PAGE 84

84 inundated for longer periods of time (always in the case of seagrass beds), which can create more highly reducing conditions that slow OM decomp osition. More contact with water also means more contact with, and accumulation of, the disso lved organic carbon (DOC), particulate organic carbon (POC), and nutrients that water transpor ts. Nutrients and OC stimulate bacterial production in sediments, nutrients stimulate autotrophic production of OC, and POC settles becoming part of sediment OM (Gacia et al. 2002). The second reason seagrass sediments reach OC functional equivalence first is parent material. In most constructed salt marshes, in the SL 15 mangrove system and in the Southwest Florida restored mangroves, the parent material wa s dredge spoil that is practically devoid of OM. As previously discussed, dredge spoil was not the only material found in SL 15 seagrass sediments. There was also OM-rich material orig inating from old vegetated sediments that were disturbed during construction, in 5-10 cm depths. At time zero OC is therefore greater in seagrass sediments. In the New Hampshire se agrass study, the sediment material was not spoil but a previously vegetated, estuar ine A horizon that had been devoid of seagrasses for 12 years (Evans and Short 2005). Like in the 5-10 cm depth of the SL 15 seagrass sediments, it is likely OC was present before restoration began. The third reason seagrass sediments reach equivalence before mangrove and salt marsh sediments is the different OC amounts among the three coastal systems. OC content varies greatly, even among nearby reference sites (Craft et al. 1999), but generally seagrass sediments have the lowest OC and mangrove sediments th e highest. Reported range in seagrass %OC is 0.15 to 1.3 (Evans and Short 2005; Vichkovitten and Holmer 2005). Reported range in salt marsh %OC is 1.7 to 13.5 (Moy and Levin 1991; Simenstad and Thom 1996; Zedler and Calloway 1999; Morgan and Short 2002). Re ported range in mangrove %OC is 2.3 to 37

PAGE 85

85 (McKee and Faulkner 2000; Alongi et al. 2001; Jennerjohn and Ittekkot 2002; Alongi et al. 2004; Bouillon et al. 2004; Otero et al. 2006). The functional equivalen ce bar is therefore lowest for seagrass sediments, which was true in this study where reference sediments mean %OC was 1.4 in mangroves and only 0.74 in seagrass. Lower than reported %OC values in this studys reference mangrove sediments are likel y due to their position around spoil islands mangrove reference sites, just as SL 15, began development in dredge spoil. No known functional trajectory studies have measured OC pools with short turnover times like ExOC and MBC. These OC pools were the only pools to approach equivalence in mangrove sediments. Because these pools are more active (Buyanovsky et al 1994; Rochette and Gregorich 1998), they are likely to develop faster in sediment s. Constructed and reference sediments in this study had MBC th at was about equal to greater than MBC in a North Sea tidal flat, a Brazilian mangrove forest, and an arctic salt marsh (Joergensen and Mueller 1995; Otero et al. 2006; Buckeridge and Jeffe ries 2007). Those other studies are the only known to measure MBC via fumigation extraction in a marine environment. MBC measured by fumigationextraction has been found to correlate well with MBC measured by phospholipids fatty acid (PLFA) analysis but not by DNA analysis or subs trate-induced respiratio n (Bailey et al. 2002; Leckie et al. 2004). A relationship between fumigated and extracted C and total PLFA concentrations has been deve loped by Bailey et al. (2002). 2.46)(4.2 total flushPLFA CFE (3-2) In equation 3-2, CFE is the unc orrected flush of OC (ug C g-1 soil) resulting from fumigation and PLFAtotal (nmol g-1 soil) is the total amount PLFA extracted from th e soil. Multiplying the results by the 2.22 CFE to MBC correction factor, MBC from this study was compared to MBC in studies that used the PLFA method. Converted measurements of PLFA yielded MBC values

PAGE 86

86 that were the same order of magnitude as our constructed system sedimentsMBC was 193 to 715 mg C kg-1 in a European seagrass bed (Bos chker et al. 2000), 289 to 769 mg C kg-1 in a California salt marsh (Cordova-Kreyl os et al. 2006), and 182 mg C kg-1 in an Australian seagrass bed (Moriarty et al. 1985). Note that this c onversion equation came from sandy soils, not marine sediments, so values are not exact but are estimates for comparison purposes. Organic carbon lability The magnitude of OC pools is not the only fact or that affects C stor age, so further data exploration is needed to assess whether SL 15 st ores sediment C as well as other seagrass beds and mangroves forests. SL 15 sediments must not only have OC pools equal to or greater than references to function as a signifi cant C store, they must also ha ve their OC stored in long term pools, where it can be sequestered away from the atmospheric C pool for decades, centuries, and even millennia. Relative amounts of the OC pool are important because the pool containing the most OC affects the overall storag e abilities of a system. A system with most of its OC in nonreactive, recalcitrant pools is going to store C lo nger than a system with most of its OC in active pools like microbial biomass (Buyanovsky et al. 1994). The constructed system generally stored more OC in short-term pools than references. In all sediments except constructed mangrove sedime nts, ExOC made up less than 1% of the TOC pool (Fig. 3-9), but the percentage of the TOC pool made up by MBC was gr eater in constructed than in reference sediments. In SL 15 mangr ove sediments, 53 to 63% of their TOC was MBC, while in reference sediments 11 to 15% of TOC wa s MBC (Fig. 3-9). This trend was the same in mangrove surface layers. In SL 15 seagrass 0-10 cm sediments, 24 to 38% of their TOC was MBC, while in references 17 to 20% of TOC wa s MBC (Fig. 3-9). SL 15 accreted layers and reference 0-5 cm depths had similar percenta ges that ranged from 19 to 27% (Fig. 3-9). Sediments in this study had more TOC stored as MBC than in other coastal systems, which

PAGE 87

87 generally had less than 10% of their MBC as TOC (Boschker et al. 2000 ; Bouillon et al. 2004; Cordova-Kreylos et al. 2006). OC limitation is a possible reason for high microbial biomass. The low C:N ratios of constructed and referenc e sediments suggest a C limitation (Sterner and Elser 2002). When microbes are C limited they te nd to sequester C in th eir cells instead of respiring C for energy (Anderson 2003). This mechanism is supported by another study with high MBC percentages (23 to 50% of TOC), as its sediments also had low TOC (<1.0%) (Joergensen and Mueller 1995). Constructed sediments do not store OC as well as reference sediments because the lability of SL 15 OC was greater than references at al l depths except for seagrass floc and accreted layers. Lability is a proxy for the decomposability of OCthe greater the lability, the faster OC is decomposed releasing C back to the atmosphere. It is therefore unlikel y that labile OC would be stored in sediments for l ong periods of time. One study of macro organic matter (MOM), a precursor of sediment OM, in constructed mars hes showed that younger marshes had more labile MOM than older marshes indicating they were less likely to sequester OC in the long term (Craft et al. 2003) Organic carbon accumulation Rates of OC accumulation are another fact or that determines how well constructed systems function as OC stores. Pool sizes m easure how much C systems are keeping from the atmosphere, lability indicates how long C is likely sequestered, and accumulation rates measure how much C is being actively taken from the atmosphere (via plant production). Salt marsh studies found equal and even greater OC accumu lation rates in constr ucted marshes (Cammen 1975; Craft et al. 1999; Craft et al. 2003). In this study, OC accumulation rates in constructed seagrass sediments were similar to those in ot her studies, but rates of constructed mangrove sediments were much lower than other studies un less the algal mat was included (Table 7, Fig. 3-

PAGE 88

88 9). Negative rates in mangrove sediments were due to a decrease bulk density throughout the year while TOC concentrations remained constant but if the algal mat becomes more permanent (i.e. buried) its OC will more than compensate for negative rates. Positive rates in seagrass sediments were driven by the accreting layer. It is unknown whether the accreted layer in seagrass sediments of SL 15 will continue to accumul ate material at the same rates as in the first year. Continued accumulation depends on how much the accreted layer formation was due to a physical response to an uneven benthic surface after construction and how much was due to macroalgae and seagrasses trapping particles from the water column. Conclusion Mangrove sediments are farther fr om being equivalent C stores than seagrass sediments. Mangrove sediments have only begun to reach equivalence in active pools (ExOC and MBC) and contain a relatively small am ount of TOC, while seagrass sedi ments have equivalent TOC at most depths (Fig. 3-9). The di fference between constructed and re ference OC lability is also much greater in mangrove than in seagrass se diments, and OC accumulation rates in mangrove sediments are negative (if the algal mat is excluded). However, if constructed mangrove sediments do begin to follow a functional trajecto ry, their potential OC storage is greater than constructed seagrass sediments because mangrov e reference sediments have larger TOC pools, less OC stored as MBC (Fig. 3-9), and lower OC lability than seagrass reference sediments. Overall, due to potential OC limitations, low TOC values for their ecosystem type, and nitrogen eutrophication (Morris and Brad ley 1999; Sigua and Tweedle 2003) IRL coastal ecosystems are probably not as effective at storing C as their counterparts elsewhere. The C storage capabilities of coastal ecosystem s make them a great contender for use as C offsets. One year is not enough time to discer n whether these systems will become significant C stores. More studies should i nvestigate constructed coastal ecosystems as potential C sinks by

PAGE 89

89 measuring functional trajectories, OC lability and OC accumulation rates. If constructed systems are similar to natural systems, then constructing coas tal ecosystems may become an accepted way to offset CO2 emissions, which would encourage more restoration.

PAGE 90

90Table 3-1. Mean ( SE) bulk density, % shell pieces, pH, and Eh (redox potential) of the sediments according to depth and site. The bulk density and % shell data were averaged over the July and November 2006 sampling periods (n=24 for SL 15 and n=18 for references). The pH and Eh data were measured in September 2006 (n=3). System Depth Bulk density (g cm-3) Shells >1mm (%) pH Eh (mV) SL 15 Reference SL 15 Reference SL 15 Reference SL 15 Reference Mangrove Algal mat/ Litter 1.03 (0.2) 0.41 (0.1) 0 0 NA NA NA NA 0-5 cm 1.62 (0.03) 0.95 (0.04) 24 (2) 21 (5) 8.3 (0.04) 7.6 (0.07) -71 (90) -130 (40) 5-10 cm 1.48 (0.03) 0.93 (0.05) 30 (2) 14 (2) 8.2 (0.03) 7.5 (0.03) -5.7 (100) -160 (8) Seagrass Floc 0.32 (0.04) 0.54 (0.09) 0 0 NA NA NA NA Accreted 0.91 (0.05) 6.0 (1) 8.2 (0.04) -98 (50) 0-5 cm 1.51 (0.04) 0.89 (0.04) 20 (3) 0.33 (0.1) 8.3 (0.04) 8.0 (0.18) -230 (4) -150 (30) 5-10 cm 1.48 (0.04) 1.03 (0.03) 19 (5) 0.49 (0.2) 8.2 (0.03) 8.3 (0.07) -180 (60) -240 (8) 10-15 cm 1.20 (0.02) 0.48 (0.1) 8.3 (0.14) -320 (40)

PAGE 91

91Table 3-2. Results of factorial ANOVAs comparing SL 15 and re ferences. Sediments and surface layers of the mangrove and seagrass systems were each run individually. BD=Bulk Density, ExOC=Extractable or ganic carbon, MBC=Microbial biomass carbon, TOC=Total organic carbon, an d TN=Total nitrogen. Concentration (conc) parameters are reported in mg kg-1 dry soil, and storage parame ters are reported in g m-2. ANOVA Effect BD TOC (conc) (storage) TN (conc) C:N ExOC (conc) (storage) MBC (conc) (storage) Lability Sediment Mangrove Site *** *** *** *** ** *** *** ** *** Month NS NS NS NS NS NS NS *** *** NS Depth NS NS NS NS ** ** ** *** ** Site*Month NS NS NS NS NS NS NS NS Site*Depth NS NS NS NS NS NS NS NS Month*Depth NS NS NS NS NS NS NS NS NS NS Seagrass Site *** ** ** *** NS *** *** *** *** *** Month NS NS NS ** NS *** *** *** *** NS Depth *** *** *** *** NS *** *** *** *** *** Site*Month NS NS NS NS NS NS NS NS ** NS Site*Depth *** NS NS *** ** NS *** *** Month*Depth NS NS NS NS NS NS NS NS NS NS Surface layers Mangrove algae/litter Site *** *** ** ** NS NS NS NS ** Month NS NS NS NS NS NS NS NS NS Site*Month NS NS NS NS NS NS NS NS Seagrass floc Site NS NS NS NS NS NS NS NS Month NS NS NS NS NS NS ** NS NS Site*Month NS NS NS ** NS NS NS NS NS For significance NS=not significant, p = or <0.05, **p < 0.01, ***p < 0.0001. Please see Appendix A for a table listing how these data were transformed prior to running the factorial ANOVA.

PAGE 92

92Table 3-3. Results of the repeated measures ANOVAs for SL 15 mangrove and seagrass sediments (0-5 cm, 5-10 cm, and seagrass accreted) and surface layers (algal mat and floc). ANOVA Effect BD (g cm-3) ExOC (mg kg-1) MBC (mg kg-1) TOC (%) TN (%) C:N (molar ratio) Lability (mg O2 g-1OC hr-1) Mangrove 0-10 Month *** NS *** NS NS Depth ** NS NS NS ** NS Algal mat Month NS NS ** NS NS Seagrass 0-10 Month NS *** *** ** NS NS NS Depth NS NS NS NS NS NS Accreted Month NS NS NS NS NS Floc Month *** * For significance NS=not significant, p <0.05, **p < 0.01, ***p < 0.0001. Please see Appendix A for a table listing how these data were transformed prior to running the repeated measures ANOVA.

PAGE 93

93Table 3-4. Mean ( SE) organic carbon concentrations (%) and storage (g m-2), nitrogen concentrations, and carbon to nitrogen molar ratios of SL 15 (n=4) and reference (n=3) mangrove and seagrass sediments according to depth and month. TOC=total organic carbon and TN =total nitrogen. Month and system Depth TOC (%) TOC (g m-2) TN (%) C:N (molar ratio) SL 15 Reference SL 15 Reference SL 15 Reference SL 15 Reference July mangrove Algal mat/ litter 2.49 (0.7) 11.9 (3) 170 (20) 670 (50) 0.26 (0.07) 1.1 (0.3) 8.2 (.22) 9.3 (0.29) 0-5 cm 0.13 (0.04) 1.3 (0.1) 110 (30) 610 (50) 0.018 (0.005) 0.096 (0.02) 5.0 (0.99) 9.6 (2.5) 5-10 cm 0.11 (0.03) 1.4 (0.4) 77 (20) 620 (90) 0.010 (0.003) 0.11 (0.03) 8.0 (1.7) 9.2 (1.3) Nov. mangrove Algal mat/ litter 3.27 (0.7) 18.0 (3) 310 (60) 1300 (400) 0.37 (0.1) 0.79 (0.1) 8.1 (0.92) 21 (5.4) 0-5 cm 0.17 (0.02) 1.3 (0.3) 140 (10) 600 (90) 0.024 (0.004) 0.12 (0.02) 5.5 (0.33) 7.0 (1.1) 5-10 cm 0.14 (0.03) 1.7 (0.5) 110 (30) 760 (200) 0.013 (0.001) 0.14 (0.03) 6.2 (0.77) 8.8 (0.9) July seagrass Floc 1.8 (0.4) 2.7 (1) 60 (20) 84 (40) 0.20 (0.06) 0.39 (0.1) 8.2 (0.95) 5.7 (0.55) Accreted 0.9 (0.01) 260 (50) 0.096 (0.01) 7.6 (1.1) 0-5 cm 2260 (0.07) 0.91 (0.6) 170 (40) 440 (20) 0.023 (0.008) 0.12 (0.01) 6.9 (0.27) 6.7 (0.16) 5-10 cm 0.27 (0.1) 0.65 (0.8) 200 (80) 330 (20) 0.027 (0.01) 0.084 (0.01) 7.0 (0.98) 6.6 (0.19) 10-15 cm 0.63 (1.0) 370 (40) 0.077 (0.01) 7.0 (0.21) Nov. seagrass Floc 5.0 (1) 3.7 (0.3) 130 (30) 96 (10) 0.62 (0.1) 0.30 (0.04) 6.8 (0.08) 10.8 (0.65) Accreted 1.0 (0.1) 340 (40) 0.13 (0.02) 6.1 (0.38) 0-5 cm 0.25 (0.05) 0.97 (0.1) 180 (30) 390 (30) 0.027 (0.006) 0.15 (0.02) 6.7 (0.76) 5.4 (0.22) 5-10 cm 0.41 (0.1)) 0.65 (0.6) 280 (60) 340 (10) 0.041 (0.01) 0.10 (0.01) 7.3 (1.1) 5.6 (0.35) 10-15 cm 0.62 (0.3) 370 (1) 0.089 (0.009) 6.1 (0.36)

PAGE 94

94Table 3-5. Mean ( SE) concentration (mg kg-1) and storage (g m-2) of two relatively labile types of organic carbon in SL 15 (n=12) and reference (n=9) mangrove and seagrass sediments according to depth and month. ExOC=extractable organic carbon and MBC=microbial biomass carbon. Month and system Depth ExOC (mg kg-1 dry soil) MBC (mg kg-1 dry soil) ExOC (g m-2) MBC (g m-2) SL 15 Reference SL 15 SL 15 Re ference Reference SL 15 Reference July mangrove Algal Mat/ litter 830 (200) 1800 (1000) 8500 (2000) 14000 (5000) 6.7 (2.0) 8.8 (3) 57 (5) 75 (9) 0-5 cm 47 (4) 130 (30) 740 (20) 1900 (100) 3.8 (0.3) 6.3 (1) 60 (2) 87 (4) 5-10 cm 38 (3) 86 (7) 690 (10) 1600 (100) 2.8 (0.3) 3.9 (0.5) 51 (2) 69 (4) Nov. mangrove Algal Mat/ Litter 600 (100) 750 (200) 12000 (3000) 7200 (1000) 5.7 (1.0) 4.7 (2) 110 (30) 49 (20) 0-5 cm 52 (7) 76 (5) 900 (30) 1800 (100) 4.2 (0.5) 3.5 (0.2) 75 (4) 83 (3) 5-10 cm 32 (3) 80 (4) 820 (20) 1800 (200) 2.4 (0.2) 3.8 (0.2) 61 (2) 80 (3) July seagrass Floc 100 (9) 89 (10) 18000 (100) 16000 (500) 0.33 (0.08) 0.26 (0.02) 55 (10) 48 (3) Accreted 77 (6) 1700 (100) 2.1 (0.1) 47 (4) 0-5 cm 29 (2) 53 (4) 840 (40) 2100 (100) 2.2 (0.1) 2.4 (0.1) 64 (2) 100 (4) 5-10 cm 24 (1) 25 (1) 800 (40) 1340 (80) 1.8 (0.1) 1.4 (0.1) 59 (2) 67 (3) 10-15 cm 24 (3) 1100 (70) 1.2 (0.04) 67 (3) Nov. seagrass Floc 130 (11) 91 (6) 24000 (1100) 23000 (3000) 0.37 (0.03) 0.24 (0.0) 65 (4) 60 (9) Accreted 100 (6) 2300 (100) 3.4 (0.3) 75 (5) 0-5 cm 31 (2) 62 (5) 990 (30) 2600 (200) 2.3 (0.1) 2.5 (0.1) 74 (3) 100 (2) 5-10 cm 36 (2) 39 (5) 1000 (40) 1400 (100) 2.6 (0.2) 2.0 (0.3) 73 (3) 71 (5) 10-15 cm 28 (1) 1300 (40) 1.7 (0.07) 74 (2)

PAGE 95

95 Table 3-6. Mean ( SE) organic carbon lability of organic carbon in SL 15 (n=4) and reference (n=3) sites according to depth and month. System and month Depth Lability (mg O2 g-1OC hr-1) SL 15 Reference July mangrove Algal mat/ litter 1120 (300) 760 (290) 0-5 cm 1480 (580) 520 (170) 5-10 cm 360 (120) 332 (130) Nov. mangrove Algal mat/ litter 2230 (540) 320 (100) 0-5 cm 1210 (250) 355 (21) 5-10 cm 572 (350) 198 (48) July seagrass Floc 631 (73) 782 (320) Accreted 387 (45) 0-5 cm 1170 (120) 527 (34) 5-10 cm 856 (120) 677 (48) 10-15 cm 469 (29) Nov. seagrass Floc 333 (79) 421 (120) Accreted 555 (52) 0-5 cm 1280 (110) 492 (25) 5-10 cm 800 (69) 625 (38) 10-15 cm 626 (40)

PAGE 96

96 Table 3-7. Organic carbon accumulation rates in mangrove and seagrass systems in this and other studies. System Rate (g OC m-2 y-1) Location and remarks Sourcea Seagrass 195 Florida, USA This study 40-65 Mexico 1 19-191 Spain 2 182 Spain 3 Mangrove -189 Florida, USA; sediment of 1-year old planted system This study 120 Florida, USA; above system with algal mat included This study 180 Australia 4 168-841 China 5 105-159 Florida, USA 6 191-3281 Florida, USA 7 101-127 Malaysia 8 33-104 Mexico 1 184-281 Thailand 9 a 1, Gonneea et al. 2004; 2, Romero et al. 1994; 3, Gacia et al 2002; 4, Brunskill et al. 2002; 5, Alongi et al. 2005; 6, Callaway et al. 1997; 7, Cahoon and Lynch 1997; 8, Alongi et al. 2004; 9, Alongi et al. 2001 1This author reported organic matter accumulation rates, so rates were divided by 2 to obtain these numbers.

PAGE 97

97 Figure 3-1. The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset). SL 15 is the large island in the center. Circles are mangrove system plots and squares are seagrass system plots. Symbols outside of SL 15 are the reference sites, which have one plot each.

PAGE 98

98 Figure 3-2. Core from SL 15 seagrass system illustrating the surface laye r (floc) and different sediment depths (accreted layer, 0-5 cm, 5-10 cm). Note the difference in color between the accreted layer and 0-5 cm depth. settledfloc accreted layer 0-5 cm layer 5-10 cm layer

PAGE 99

99 Figure 3-3. The functional trajectory the bulk density of SL 15 mangrove sediments followed over the first year after c onstruction. The symbols are the mean values for each sampling date (n=12) and error bars are SE.

PAGE 100

100 Figure 3-4. The changes in organi c carbon parameters over the first year after construction in SL 15 seagrass and mangrove sediments. Th e symbols are the mean values for each sampling date (n=12 for ExOC and MBC and n=4 for TOC) and error bars are SE.

PAGE 101

101 Figure 3-5. The changes in organi c carbon parameters over the first year after construction in SL 15 seagrass and mangrove surface layers. Th e symbols are the mean values for each sampling date (n=4) and error bars are SE.

PAGE 102

102 Figure 3-6. Comparisons between total organic carbon (TOC) in reference and SL 15 mangrove (top) and seagrass (bottom) sediments. The bars are mean TOC averaged over month (July and November 2006) for each depth of sediment (n=4 for SL 15 and n=3 for reference). Error bars are SE. Depths in the seagrass systems are as follows: 1= SL 15 accreted and reference 0-5, 2= SL 15 0-5 and reference 5-10, 3= SL 15 5-10 and reference 10-15. An asterisk indicates a si gnificant site effect (Table 3-5). Capital letters are results of a Tukey test performed after a significa nt site x depth interaction, and lowercase letters are results of a Tuke y performed after an unsignificant site x depth interaction, but a signi ficant one-way ANOVA. Bars that share letters are not significantly different. B A

PAGE 103

103 Figure 3-7. Comparisons between extractable or ganic carbon (ExOC) in reference and SL 15 mangrove (top) and seagrass (bottom) sediments. The bars are mean ExOC averaged over month (July and November 2006) for each depth of sediment (n=12 for SL 15 and n=9 for reference). Error bars are SE Depths in the seagrass systems are as follows: 1= SL 15 accreted and reference 05, 2= SL 15 0-5 and reference 5-10, 3= SL 15 5-10 and reference 10-15. An asterisk indicates a significant site effect (Table 3-5). Capital letters are results of a Tuke y test performed after a significant site x depth interaction, and lower case letters are results of a Tukey performed after an unsignificant site x depth interaction, but a significant one way ANOVA. Bars that share letters are not significantly different. B A

PAGE 104

104 Figure 3-8.Comparisons between microbial bi omass carbon (MBC) in reference and SL 15 mangrove (top) and seagrass (bottom) sediments. The bars are mean MBC averaged over month (July and November 2006) for each depth of sediment (n=12 for SL 15 and n=9 for reference). Error bars are SE Depths in the seagrass systems are as follows: 1= SL 15 accreted and reference 05, 2= SL 15 0-5 and reference 5-10, 3= SL 15 5-10 and reference 10-15. An asterisk indicates a significant site effect (Table 3-5). Capital letters are results of a Tuke y test performed after a significant site x depth interaction, and lower case letters are results of a Tukey performed after an unsignificant site x depth interaction, but a significant one way ANOVA. Bars that share letters are not significantly different. A B

PAGE 105

105 120 g OC m-2y-1(Net Accumulation) 453 g OC m-2(Total OC) Floc238 g OC m-20-5 cm122 g OC m-25-10 cm93 g OC m-2 SL 15 Mangrove% 120 g OC m-2y-1(Net Accumulation) 453 g OC m-2(Total OC) 120 g OC m-2y-1(Net Accumulation) 453 g OC m-2(Total OC) Floc238 g OC m-20-5 cm122 g OC m-25-10 cm93 g OC m-2 SL 15 Mangrove% 105-159 g OC m-2y-1Literature value(Net Accumulation) 2273 g OC m-2(Total OC) Litter978 g OC m-20-5 cm607 g OC m-25-10 cm688 g OC m-2 Reference Mangrove% 105-159 g OC m-2y-1Literature value(Net Accumulation) 2273 g OC m-2(Total OC) 105-159 g OC m-2y-1Literature value(Net Accumulation) 2273 g OC m-2(Total OC) Litter978 g OC m-20-5 cm607 g OC m-25-10 cm688 g OC m-2 Reference Mangrove% Figure 3-9. Organic carbon (OC) pools in SL 15 and reference mangrove and seagrass sediments. Beside each box is the total amount of OC in the depths analyzed. OC accumulation rates were calculated in this study for SL 15 sediments (includes algal mat for SL 15 mangrove) but are literature values for re ference sediments (Ca llaway et al. 1997 for mangrove and Gonnoeea et al. 2004 for seagrass). Boxes show the percentage distribution of the total OC in each depth and OC poolMBC (dark grey), ExOC (white), and other (light grey).

PAGE 106

106 195 g OC m-2y-1 (Net Accumulation) 814 g OC m-2(Total OC) Floc97 g OC m-2Accreted layer300 g OC m-25-10 cm240 g OC m-20-5 cm177 g OC m-2SL 15 Seagrass% 195 g OC m-2y-1 (Net Accumulation) 814 g OC m-2(Total OC) 195 g OC m-2y-1 (Net Accumulation) 814 g OC m-2(Total OC) Floc97 g OC m-2Accreted layer300 g OC m-25-10 cm240 g OC m-20-5 cm177 g OC m-2SL 15 Seagrass% 40-65 g OC m-2y-1Literature value(Net Accumulation) 1207 g OC m-2(Total OC) Reference SeagrassFloc90 g OC m-20-5 cm414 g OC m-210-15 cm371 g OC m-25-10 cm332 g OC m-2% 40-65 g OC m-2y-1Literature value(Net Accumulation) 1207 g OC m-2(Total OC) 40-65 g OC m-2y-1Literature value(Net Accumulation) 1207 g OC m-2(Total OC) Reference SeagrassFloc90 g OC m-20-5 cm414 g OC m-210-15 cm371 g OC m-25-10 cm332 g OC m-2% Figure 3-9. Continued

PAGE 107

107 CHAPTER 4 SOURCES OF SEDIMENT ORGANIC CARB ON IN A C ONSTRUCTED MANGROVE AND SEAGRASS SYSTEM Introduction Sediments can accumulate organic carbon (OC) from in situ vegetation, drift macroalgae, plankton, and water column terrest rialand marine-derived detrit us. Understanding sources of OC in soils and sediments is important to our un derstanding of local and global C cycles (Hedges 1992). The source of OC influences the quality and stability of OC in sedi ments. OC sources, like temperature and oxygen availability, affect decomposition rates (Chapin et al. 2002), which in turn affect OC sequestration. Certai n ecosystems, like macrophyte-dominated coastal systems, accumulate and store large amounts of OC in their sediments. These salt marshes, mangrove forests, and seag rass beds are sinks for CO2 and therefore mitigate climate change by keeping C out of the atmosphere. Worldwide, salt marshes and mangroves store at least 44.6 Pg C in their sediments (Chmura et al. 2003), equivalent to 2% of the global soil C pool (Lal et al. 1995). Seagrass beds, which make up only 0.15% of global marine area, account for 15% of the global marine OC storage (Hemminga and Duarte 2000). Determining the vegetation that are the main OC sources to coastal sediments helps researchers predict how changing environmental conditions may affect the future of these significant C stores. Coastal ecosystems are experiencing great losses worldwide (Valie la et al. 2001; Alongi 2002; Green and Short, 2003). The loss of vege tated coastal ecosystems has caused at least a 25% decrease in their global C sequestration capacity (Duarte et al. 2005). Constructing coastal ecosystems may restore a portion of the lost C sink (Connor et al. 2001). Knowing OC sources of constructed coastal systems can indicate whether these constructed systems can become effective at storing OC. For example, a constructed mangrove system whose principle sedimentary OC (SOC) source is relatively labile macroalgae will not store as much C for as

PAGE 108

108 long amount of time as a well-established mangr ove system whose main OC sources are the more recalcitrant leaves and roots of mangroves. There are a myriad of methods researchers utilize to determine OC sources. The most widely used method measures bulk stable isotopes (usually 13C and 15N) in possible sources and sediments. Bulk analyses measure isotopic sign atures of entire OC pools in sediments or of whole plant parts. Sources are then determined by a simple comparison of source and sediment isotopic signatures (Haines 1976; Hemminga et al. 1994; Jennerja hn and Ittekkot 2002; Thimdee et al. 2003) or by mixing models (Dauby 1989; Kennedy et al. 2004; Papadimitriou et al. 2005; Zhou et al. 2006;). Other parameters are used with isotopic signature s to determine sources using ternary diagrams of N:C ratios plotted against 13C (Gonnoeea et al. 2004; Miserocchi et al. 2007) or more complex mixing models using 13C and biomass or %OC as parameters (Chmura et al. 1987; Middelburg et al. 1997; Bouillon et al. 2003;). Sources must have consistently distinct stable isot opic signatures for this method to be useful (Papadimitriou et al. 2005). Lipids are also used as biomarkers to determine OC sources (Wang et al. 2003). The lipids, generally sterols, fatty acids, or hydrocarbons, vary in specificity as some can identify groups of organisms such as vascular plants or algae while others may be specific to one genera or species (Canuel et al. 1997). Finer resoluti on of sources is possibl e when the isotopic signatures of lipids are measured in compound speci fic stable isotope anal yses (Canuel et al. 1997; Bull et al. 1999; Hernandez et al. 2001; Mead et al. 2005) Some lesser-used methods involve comparing relative amounts of certain OC st ructures in the soil, either visually as in petrographic analysis (Lallier-V erges et al. 1998; Marchland et al. 2003) or chemically as in nuclear magnetic resonance spec troscopy (Golding et al. 2004).

PAGE 109

109 Stable isotopes of bulk compositions have su ccessfully identified the main SOC sources in subtropical and tropical coastal ecosystems dominated by mangroves and seagrasses because potential sources in these ecosystems have a wide range of 13C (Hemminga et al. 1994; Jennerjahn and Ittekkot 2002; Gonnoeea et al. 20 04; Kennedy et al. 2004; Papadimitriou et al. 2005; Smit et al. 2005; Zhou et al. 2006). Mangroves have the most depleted 13C because Rubisco carboxylase discriminates agai nst isotopically heavy C during C3 photosynthesis (Hemminga and Mateo 1996; Hemminga and Duarte 2000). Seagrasses have the most enriched 13C, despite C3 characteristics, because of diffusional constraints on C uptake in an aquatic environment (Hemminga and Mateo 1996). Isotopic si gnatures of other potential sources such as plankton and epiphytes generally fall between mangrove and seagrass values (Kennedy et al. 2004; Papadimitriou et al. 2005). In this study, we determine: 1) significant sources to the SOC in a constructed mangrove and seagrass system, 2) how sources change over time in a constructed system, and 3) how sources differ between the constructed system and nearby mangrove and seagrass reference sediments. We hypothesized that SOC sources in the constructed syst em will initially be macroalgae or seston, while SOC sources in the re ference systems will be vascular plants like mangroves and seagrasses. Methods Study Site SL 15 (Fig. 4-1) is a mitigation site located in the subtropical portion of the Indian River Lagoon (IRL) adjacent to Fort Pier ce, Florida. The IRL is a long, shallow, and microtidal water body that lies in both temperate an d subtropical climates. SL 15 is one of many spoil islands created in the Indian River Lagoon during the construction of the Atlantic Intracoastal Waterway. These islands sit se veral meters above sea level a nd are populated by many exotics,

PAGE 110

110 such as Australian Pine ( Casuvina casuvina) and Brazilian Pepper ( Shinus terebinthifolius ), in their interiors and by native red, black, and white mangroves (Rhizophora mangle Avicennia germinans and Laguncularia racemosa ) around their edges. To mitigate destruction of a nearby mangrove forest and seagrass bed, seagrass a nd mangrove systems were created on SL 15. These systems were created by burning and rem oving interior vegetation and removing dredge spoil down to several different elevations. Th e seagrass bed, which remains submerged during low tide, is at the lowest eleva tion, the mangrove forest, which is exposed at low tide, is at the middle elevation, and at the hi ghest elevation, above sea level, is a maritime forest. The mangrove fringe of SL 15 was left intact excep t for a few flushing channels. In between the constructed seagrass and mangrove systems a thin Spartina alterniflora buffer was planted. The mangrove forest was planted with R. mangle and maritime forests were planted with Coccoloba uvifera Borrichia frutescens Rapanea guinensis Conocarpus erectus and Distichlis spicata but seagrasses were left to colonize naturally. Natural systems near SL 15 include its original mangrove forest fringe, surrounding seagrass be ds, and mangrove fringes of adjacent spoil islands, which are at least 40 years old. Litter Bags Plant material from Syringodeum filiforme Thalassia testudinum Halodule beaudettei Acanthophora spicifera, Sargassum spp, A. germinans and R. mangle were collected in July 2006. Living seagrass fronds were taken from th e beds around SL 15, which is similar to the material ripped off by wind and wave events (Moore and Fairweather 20 06). Clumps of live macroalgae were taken from the subtidal areas in and around SL 15. Yellow mangrove leaves, the kind about to fall, were taken from trees on the edge of SL 15 and surrounding islands. Plant material was transported back to the laboratory and rinsed. Epiphytes were removed from seagrass fronds and macroalgae. Plant material wa s then air dried for several weeks before being

PAGE 111

111 weighed by species into 2-3 g (only 1 g for A. spicifera ) allotments and placed intact into 13 cm x 13 cm litter bags of nylon mesh with 0.5 x 0.25 mm holes. This litter bag study is a common garden study where we investigated the relative decomposition rates of the potential sources to SOC, so all litter bags were placed in the same area of the SL 15 mangrove system. On Septem ber 8, 2006, litter bags were placed on the sediment surface, pinned down with metal stakes, and overlaid with large wire mesh to prevent them from washing away. Three litter bags fr om each plant species were randomly collected at 2, 4, 8, 16, and 32 weeks. Sediment and algae were rinsed from the litter bags in the laboratory before the bags were air dried for several week s. Once dry, the bags were opened, and plant material in each bag was weighed. Source Sampling Plants were sampled on SL 15 in January July, October, and November 2006 and at reference sites in July and November 2006. Sa mpled plants included a ll potential sources to SOC found in and around SL 15 and reference site s and fell into 3 main groups: Subtidal, which include seagrasses ( S. filiforme T. testudinum H. beaudettei, Halophila johnsonii ) and macroalgae (epiphytes on seagrasses, Acanthophora spicifera Caulerpa sertulariodes, Sargassum spp., Ulva spp., Chaetomorpha linum Rosenviga intricata Hypnea cervicornis Gracilaria tikvahiae, and Enteromorpha spp.); Intertidal, which included mangroves (Avicennia germinans Rhizophora mangle Laguncularia racemosa ), Sueda linearis and Spartina alterniflora ; and Terrestrial ( Schinus terebenthifolias, Casuarina equisetifolia Coccoloba uvifera and Triplasis purpurea ). Not all plants were collect ed at all sampling dates because some plants, particularly species of macroalgae were not present throughout the year. Vascular plant samples were a composite of 3-5 live, hea lthy leaves or fronds fr om greater than three individuals collected across the sampling area (i .e. SL 15 or reference sites). Macroalgae

PAGE 112

112 samples were composites of different clumps co llected from across the sampling area. Epiphyte samples were composites of algal material scrape d from seagrass fronds in the laboratory. Roots of seagrasses and mangroves were taken from se diment cores for analysis; they were not identified to species. Roots of A. germinans R. mangle and S. alterniflora were collected in the field as well. At the laboratory, seagrass fronds were scraped clean, and seagrasses, roots, and macroalgae were rinsed. All plants were dried at 60C for three days before being initial ground on a Wiley mill (if necessary) and then ground to a fine powder using a ball mill. Seston was collected in May, September, October, and November 2006 and February 2007. For each seston sample, 500 mL of water was collected from the middle of the water column in the subtidal area of SL 15. Three samples each were taken on a flood and an ebb tide except in February 2007, where onl y ebb tide samples were collected Water samples were kept on ice and transported to the laboratory wh ere they were filtered through precombusted Whatman GF/F glass fiber filters. Blanks of 500 mL of deionized water were also filtered for each sampling event. Filters were then freeze-dried for 24 hours. Sediment Sampling Four, 2 m x 2 m plots were established in th e mangrove forest and in the seagrass bed on SL 15 (Fig. 4-1). Three, 7 cm in diameter sedime nt cores from each of these plots were retrieved in November 2005, January (mangrove only), February (seagrass only), May, July, and November 2006. Cores were taken from different areas of the plots each time to ensure an area was not re-sampled. For references, three randoml y-selected plots were established in natural mangrove forests and seagrass beds within 1 km of SL 15. These plots were sampled in July and November 2006 using the same procedure as for SL 15 plots. Sediment co res were sectioned in the field and stored in plastic bags on ice for tran sport and then in a 4C refrigerator. SL 15 cores were initially divided into 0-5 cm and 5-10 cm sediment depths. In subsequent samplings,

PAGE 113

113 material had accumulated on top of the seagrass section, which was collected and analyzed separately from the original sediment depths as an accreted layer. Surface layersfloc from seagrass systems, algal mats from the SL 15 mangrove system, and litter layers from the reference mangrove systemwere collected from each core and were composited by plot. Differences in color and texture were used to se parate accreted and surfac e layers from original depths except for floc, which was the fraction of the accreted layer that pour ed off (Fig. 4-2). Laboratory Analyses Rocks, roots, and detritus were removed from each sample prior to homogenization. Samples were then freeze-dried for 48 hours. Freeze-dried sediment samples were composited by plot and sieved through a 1 mm mesh to rem ove large shell pieces and carbonate rock, which were weighed so their mass could be accounted fo r in calculations. Sediment and surface layer samples were then ball-milled to a fine powder in stainless steel canisters. TOC, TN, and 13C were measured in sediment, surface layers, seston filters, and plant samples. TOC and TN were used to calculate C: N ratios on a molar basis. Inorganic carbon (IC) was removed from sediment, surface layer, and seston samples via vapor acidification (Hedges and Stern 1983; Harris et al. 2001; Gonneea et al. 2004). Sediment and surface layer samples were weighed out into 9 x 5 mm or 10 x 10 mm silver capsules (Thermo Scientific, Waltham MA and CE Elantech, Lakewood, NJ), which were arranged in plastic well plates and moistened with deionized water before acidification. Th ree holes (7 mm in diameter) were cut from each seston filter with a hole punch and arranged in plastic well plates. The filled well plates were then placed in a glass desiccator with a beaker of concentrated HCl (12 M ) for 24 hours before being dried at 60C for another 24 hours. Sest on filter samples were then put into 10 x 10 mm silver capsules. Plant samples were weighed in to 9 x 5 mm tin capsules (Costech Analytical Technologies, Valencia, CA). All samples were combusted on an elemental analyzer (ECS

PAGE 114

114 4010, Costech) in line with an isotope ratio mass spectrometer (ThermoFinnigan MAT Delta Plus XL, Thermo Scientific, Waltham MA) for %OC and 13C. Plants were analyzed for %TN simultaneously. Peach leaves (NIST 1547) were us ed for EA calibration, with sucrose and an internal soil standard used as check standards. Sucrose and Peach leaves were used as internal standards for mass spectrometry measurements. C isotopes were reported in per mil notation based on deviations from the Pee Dee Dolomite standard. Tests were run on sand samples with various carbonate percentages and total weights to assess the efficacy of the vapor acidification method and determine the maximum sample mass th at still ensured complete removal of IC. Furthermore, 13C values were used to confirm complete removal of IC because the presence of carbonate greatly raised 13C values. If incomplete IC removal was suspected, samples were rerun at a lower total mass. The 13C of filter blanks were account ed for in the calculation of seston 13C. Unacidified sediment, surface layer, and seston samples were run separately in tin capsules (Costech) on an elemental analyzer (Flash EA 1112 Series, Thermo Scientific, Waltham, MA) for TN. Acetini lide was used for calibration st andards, while peach leaves (NIST 1547) and an internal soil standard were used for quality control. Data Analyses Individual plant 13C and C:N were averaged across sites and sampling dates. Values of certain species were also averaged together into plant groups of seagrass, macroalgae, mangroves, or C3 terrestrial. Differences in 13C between sampling date and tide phase (ebb or flood) were tested on seston samples using one -way analyses of variance (ANOVAs) in JMP Version 6. For all sources, C:N ratios are reported, even though N: C ratios are used in graphs, so data can be easily compared across studies. L itter mass loss for each species was modeled using a first-order exponential decay curve.

PAGE 115

115 )( 0*kt teMM (4-1) In equation 4-1, M0 is the initial litter mass, Mt is the litter mass at time t, and k is the decay constant. The decay constant for each species was estimated using nonlinear models in JMP Version 6 (SAS Institute, Cary, NC). To investigate whether 13C, TOC, or C:N changed through time in SL 15 sediments and surface layers, repeated measures ANOVAs were run for both mangrove and seagrass areas. A spatial power covariance structure was used to account for unequal spacing between time points. Subjects were the plots on SL 15, and the repeat ed factor was time. For the 0-5 and 5-10 cm depths in each system, the ANOVAs were run w ith depth as a main effect and a time*depth interaction term. The floc, algal mat, and accre ted layers were each r un separately in ANOVAs where time was the only effect. These analyses were run using the mixed procedure in SAS Version 8 (SAS Institute, Cary, NC). Comparisons between SL 15 and the referen ce sites were analyzed using one factorial ANOVA each for the mangrove and seagrass sedi ments and one factorial ANOVA each for the mangrove and seagrass surface layers (algal ma t/litter and floc). Sediment ANOVAs consisted of three fixed factorssite, month, and depth. Surface layer ANOVAs consisted of only the site and month factors. All two way interactions we re tested. SL 15 plot and reference site data were pooled into two site treatments, SL 15 and re ference. Months used in these analyses were July and November 2006, the sampling dates fo r which both SL 15 and reference data were available. For seagrass sediment analysis, SL 15 and reference depths were assigned to 3 categories in order to make comparisons: SL 15 accr eted and reference 0-5 cm were depth 1, SL 15 5-10 cm and reference 0-5 cm were depth 2, and SL 15 5-10 cm and reference 10-15 cm were depth 3. Factorial ANOVAs were run on JMP Ve rsion 6 (SAS Institute, Cary, NC).

PAGE 116

116 A portion of the above analyses were performe d on data transformed to meet the normality requirement (see Appendix A for details). Post hoc multiple comparisons were carried out on significant effects using the Tukey test. Significance was decided usi ng an alpha level of 0.05. Ternary diagrams (Dittmar et al. 2001; Goni et al. 2003, Gonnoeea et al. 2004) were used to determine the main SOC sources. Because ternary diagrams can only have three end members, field observations and the position of mean sediment 13C relative to mean potential source 13C on a 13C line (Fig. 4-3) were used to choose the three most likely end members for each constructed and reference sediment and for the mangrove litter layer and seagrass floc. N:C of the three end members and se diments were plotted against 13C. N:C ratios are used instead of C:N ratios because with the larger number in the denominator, they are more statistically robust (Goni et al. 2003). End members N:C and 13C were averaged for all sampling dates and species within that group (e.g.: mangroves), but for plants where multiple parts were measured, only leaf/frond values were used. The three en d members create a triangle that is expanded according to the standard deviations of the e nd members to account for natural variability and analytical error. Sediment samples that fall in the middle of the triangle are assumed to be a mixture of all three sources, samples that fa ll along a line connecting two end-members are considered a mixture of those two sources, and samples that fa ll around the vertex of an end member are assumed to have OM from mainly th at source. Samples that fall outside of the expanded triangle have OC contributions from additional sources or have undergone changes during diagenesis. Results Source Characteristics 13C and C:N varied among plant groups. Generally, the lowest 13C and greatest C:N were found in mangrove leaves and roots and C3 terrestrial plant leaves (Table 4-1). The greatest

PAGE 117

117 13C and a relatively low C:N we re found in seagrass fronds. S. alterniflora had a low 13C and high C:N. Seston had low 13C and low C:N. Seston samples had greater 13C in fall than in winter (ANOVA, df=4, p=0.0002) but did not differ between ebb a nd flood tides (ANOVA, df=1, p=0.54). Compared to varia tion among plant groups, variation of 13C within plant groups was usually low with mangrove tissues of all species varying by less th an 2.5 and seagrass tissues (except H. johnsonii) by less than 3.4. The exception was the macroalgae group, whose 13C varied by 15. Macroalgae had high vari ability with C:N ratios as well. Plant tissue type influenced C:N ratios with greater C:N in roots than in leaves for both mangroves and seagrasses. Plants also differed in their decay rates, even within groups (Table 4-2). The greatest decay constants, and fastest rates of decay, were for a macroalgae ( A. spicifera ) and a seagrass ( S. filiforme). The slowest decay rates were for a seagrass ( H. beaudettei) and a mangrove ( R. mangle ). Sediments and Surface Layers 13C of SL 15 sediments and surface layers, w ith the exception of the 0-10 cm seagrass sediments, changed significantly over time (mont h effect, p<0.034, Table 4-3, Fig. 4-4). Mangrove sediments and seagrass accreted layers had 13C that increased towards the mean 13C of their respective references over the first year after construction (Fig. 4-4). The mangrove algal mats 13C also increased but moved away from refe rence values (Fig. 4-4). Most of the layers did not have changing %OC or C:N ratios throughout the year C:N ratios changed significantly without dire ction in mangrove sediments and seagrass floc (Chapter 3). TOC significantly changed in seagrass 0-10 cm sedime nts and floc, but only wi th direction in floc, where it increased over ti me (Chapter 3). The 13C, TOC, or C:N valu es did not differ among sediment depths (Table 4-3, Chapter 3).

PAGE 118

118 SL 15 seagrass sediments had lower 13C than reference seagrass sediments (site effect, p<0.0001, Tables 4-3 and 4-4), but SL 15 mangrove sediments had 13C similar to reference mangrove sediments (p=0.40, Tables 4-3 and 4-4). SL 15 floc was more depleted than reference floc in July but more enriched than referen ce floc in November (month x site interaction, p<0.0001, Table 4-3 and 4-4). SL 15 algal mat was more enriched than re ference litter in both months, but the difference was gr eater in November (month x site interaction, p<0.0001, Table 4-3 and 4-4). TOC (%) was generally lower in SL 15 sediments than references with the exception of the SL 15 seagrass accreted laye r and floc, which had similar TOC to the references 0-5 cm depth and floc respectively (Chapter 3). C:N ratios were similar in seagrass sediments and floc but were lower in SL 15 mangrove sediments and surface layers than in respective mangrove references (Chapter 3). Source Determination Putting source (plants and seston) and sediment 13C data together indicates potential sources to the various sediments and surface layers (Fig. 4-3). Using observations from the field and Fig. 4-3, the three ternar y diagram end members for SL 15 mangrove sediment were seston, algal mat, and terrestrial plants (Fig. 4-5a). Seston, litter, and mangrove s were the end members for reference mangrove sediments (Fig. 4-5b). Seston, seagrass, and macroalgae were the end members for SL 15 and reference seagrass sedime nts and floc (Fig. 4-6 and 4-7b). Seston, seagrass, and mangroves were the end members for reference mangrove litter (Fig. 4-7a). SL 15 algal mats did not need a diagram because they are their own source as primary producers. Ternary diagrams explained 74% of SL 15 mang rove sediment samples, 92% of reference mangrove sediment samples, 71% of SL 15 seagrass sediment samples, 33% of reference seagrass sediment samples, 100% of reference ma ngrove litter samples, and 89% of seagrass floc

PAGE 119

119 samples (Fig. 4-5 through 4-7). All of the samples that fell outside the te rnary plots, regardless of site or depth, did not fit b ecause their N:C ratios were greate r than that of the sources. The majority of SL 15 mangrove sediment samp les fell near the seston end member. Some samples fell in the middle of the triangle and others fell close to the terrestrial end member (Fig. 4-5a). Most reference mangrove sediment sa mples fell between seston and litter end members (Fig. 4-5b). In terms of 13C, but not in terms of N:C, mo st SL 15 and reference seagrass sediment samples were within the range of m acroalgal sources (Fig. 4-6). SL 15 seagrass sediment samples fell far from the seagrass end member (Fig. 4-6a). SL 15 seagrass 0-10 cm and accreted depths did not differ in their sources. Most reference seagrass samples fell outside the diagram due to high N:C ratios (Fig. 4-6b). Examining only 13C, reference seagrass sediments were more enriched than macroalgae a nd seston but more depleted than seagrass (Fig. 4-3). Reference mangrove litter layer samples from July fell between seston and seagrass end members but November samples fell in the middl e or at the mangrove vertex (Fig. 4-7a). Reference seagrass floc samples fell between seston and macroalgae end members in July but outside the diagram in November (Fig. 4-7b). SL 15 seagrass floc fell between seston and seagrass regardless of sampling data (Fig. 4-7b.) Discussion Source Characteristics 13C of the main potential sources in the st udied part of the I ndian River Lagoon were within the range of literature from similar estuarine studies (Table 4-5). Our sources C:N values were also within reported literature values of 30 to 99 for mangrove leav es and roots (LallierVerges et al. 1998; Thimdee et al. 2003; G onnoeea et al. 2004; Muzuka and Shunula 2006), of 15 to 21 for seagrass fronds (Thimdee et al. 2003 ; Gonnoeea et al. 2004; Machas et al. 2006), of

PAGE 120

120 5.8 to 9.3 for seston (Gonnoeea et al. 2004; Zhou et al. 2006), and of 7 to 30 for macroalgae (Kristensen 1994; Thimdee et al. 2003). 13C of plants vary within different tissues (Viz zini et al. 2003; Papadi mitriou et al. 2005;), within a single species (Hemminga and Mateo 1 996), across sites (Kennedy et al. 2004), seasons (Vizzini et al. 2003), and year s (Anderson and Fourqurean 2003; Fourqurean et al. 2005). Variations are most pronounced in seagrasses (Thimdee et al. 2003) and macroalgae. In submerged vegetation variation is due to the relative uses of dissolved CO2 and bicarbonate, the source of inorganic C in the water, temperature, irradiance, and subsequent photosynthesis rates (Lin et al. 1991; Hemmi nga and Mateo 1996). Seston 13C can also vary temporally, spatially, and between ebb and flood tides (Hemminga et al. 1994). These variations in source 13C make it necessary to measure all potential sources 13C for each study area, instead of relying on literature values, and ideally, measure significant sources across tissues, sites, and seasons. 13C variations within individual s ources and plant groups in this st udy were generally smaller than differences among main sources, so the variat ions most likely do not affect our source determinations. Furthermore, where 13C did overlap among main sources their C:N ratios set them apart, as with seston and mangroves, or they were not both end members for the same ternary diagram. There is some concern about whether 13C of plant tissues changes during diagenesis because large changes in 13C could lead to misleading source determinations. Studies that measured fresh and senescent mangrove leaves and seagrass found small (generally >1) differences (Thimdee et al. 2003; Gonnoeea et al. 2004). Decomposition studies found significant but minor (0.55 to 2) changes in seagrass, mangrove, and macroalgae 13C during diagenesis (Fenton and Ritz 1988; Fourqurean a nd Schrlau 2003), but others found no significant

PAGE 121

121 changes (Machas et al. 2006). Where 13C did change in decomposition studies of multiple species, the initial differences in 13C between species were still clear. Unfortunately, we did not measure changes in 13C of our plant tissues during decomposition. Given the small magnitude of changes found in other studies and the large differences in 13C between groups of potential sources, diagenetic changes in 13C are unlikely to cause misidentification of the main SOC sources in this study. Changes in C:N during decomposition also occur and can be greater in magnitude than 13C changes (Fourqurean and Schrlau 2003) Studies of mangrove, seagrass, and macroalgal decomposition have found decrea ses and increases in C:N ratios that were dependent upon species or tissue (Twilley et al. 1986; Bourgues et al. 1996; Fourqurean and Schrlau 2003); others found no change in C:N ratios (Machas et al. 2006). Decay constants of seagrasses on SL 15 were wi thin literature values, which ranged from 0.002 to 0.12 day-1 (Mateo and Romero 1996; Machas et al. 2006; Moore and Fairweather 2006). T. testudinum had a greater decay constant and therefore faster deco mposition in this study than in Florida bay (Fourqurean and Schrlau 2003). Mangrove decay constants were also within literature values that ranged from 0.0048 to 0.022 day1 (Fourqurean and Schrlau 2003; AkeCastillo et al. 2006; Ramo s e Silva et al. 2006). R. mangle s decay constant in this study fell on the low end of R. mangle reported values. Estimated macroa lgae decay constants ranged widely from 1 to 0.014 (Foreman and Smith 1984; Mews et al. 2006). The decay constant of Sargassum spp. in our study was at the low e nd of the range, probably because Sargassum has more structural components than most other macroa lgae. Surprisingly, differences among decay constants in this study did not fall along plant groups. We expected mangroves to have the lowest decay constants and macroalgae to have the highest with seagrass falling in between (Kristensen 1994; Bourgues et al. 1996; Fourqurean and Schrlau 2003). However, S. filiforme

PAGE 122

122 decomposed as fast as the macroalgae and T. testudinum s decomposition was at the rate of A. germinans These results indicate that in terms of decomposition, species identity matters more than the group to which a species belongs. For source determination, these results specify which species of an end member group are more likel y to contribute to SOC because the slower a species decomposes, the better chance its OC will be buried in sediments. Sediments and Surface Layers Changes in SL 15 SOC 13C over the course of a year indicate new SOC sources are adding to the sediment TOC pool or, without a ch ange in TOC, decomposition of old source OC while new source OC accumulates. These chan ges were greater in upper sections of both mangrove (0-5 cm) and seagrass (accreted layer) sediments because the inputs of new OC reach the top of sediments first. Bioturbation then brings new OC inputs deeper into the profile. Bioturbating organisms were observed in mangrove but not in seagrass sediments, which may explain why 13C of deeper mangrove sediments changed over time but deeper seagrass sediments did not. Surface layers had the greatest 13C changes through the year. All changes were positive so that the new SOC sources to SL 15 after construction must be more enriched than old OC sources. Old OC sources were relatively depleted in 13C as they were most likely the terrestrial plants that i nhabited SL 15 pre-construction. In mangrove sediments, the new source was most likely the alga l mat and in seagrass accreted layers and floc the new sources were macroalgae or seagrass (Fig. 4-3). Enrichment of algal mat 13C is due to changing inorganic C sources, as unlike other sediments and layers, the algal mat is its own producer of OC. As the algal mats grow, so doe s their influence on the biogeochemistry of their environment. Photosynthesis and respiration within the algal mat changes the pH of the water around it (Kayombo et al. 2002). At night respiration decreases the pH, which can cause CaCO3 in the sediment below the algal mat to dissolve. CaCO3 dissolves into vari ous carbonate species

PAGE 123

123 (CO3 -2, HCO3 -1), which inherit the high 13C of CaCO3 ( 0) (Lin et al. 1991). These carbonate species then may be utilized by algae as inorga nic C sources during daytime photosynthesis. 13C in the literature ranges from -29.4 to -20.6 for mangrove sediments (Bouillon et al. 2003; Thimdee et al. 2003; Gonneea et al. 2004) and from -10.3 to -26.6 for seagrass sediments (Hemminga et al. 1994 ; Kennedy et al. 2004; Papadimitr iou et al. 2005). Sediment 13C in this study are for the most part within l iterature values. SL 15 and reference mangrove sediments span the range of literature values from -27.5 to -19.4. SL 15 and reference seagrass sediments are at the lower end of the literature values with 13C ranging from -23.2 to -19.4. Differences in 13C among SL 15 and reference sediments and surface layers suggest their SOC sources differ. Observations of the distribution of primary producers around the sites also suggest sources differ, even between SL 15 and reference mangrove sites, whose 13C were not significantly different. Source Determination The ternary diagram indicated that seston was the dominant source for SL 15 mangrove sediments with some OC being contributed by terr estrial plants and the algal mat (Fig. 4-5a). Terrestrial sources most likely contributed to SOC before and during construction. During construction, we observed terrestrial plant parts that were no t fully removed by burning and clearing being mixed into spoil within SL 15s in tertidal zone. The algal mats influence as a source was supported by 13C enrichment of mangrove sedime nts over the first year. Mangroves were not included as a source in the ternary diagrams because SL 15 mangroves were young (>2 years old) and mangrove litter was very sparse. Seston was also a dominant source for reference mangrove sediments according to the ternary diag ram, but in this instance it shared this designation with the litter layer (F ig. 4-5b). According to Fig. 4-3, mangroves also contributed to SOC because mean sediment 13C was more depleted than m ean seston and litter values.

PAGE 124

124 Seston and macroalgae were the dominant OC sources in SL 15 seagrass sediments according to the ternary diagram (Fig. 4-6a). Se agrasses, which had colonized most of SL 15 at generally low densities by July 2006 (Fischler 2007), were not yet important SOC sources. Seston and macroalgae as main SOC sources were further supported by observationsdrift macroalgae was frequently found buried in the a ccreted layer section of cores throughout the study where it seemed to trap part icles from the water, driving accretion. High N:C (low C:N) ratios of seagrass reference sediments interfer ed with determining sources via the ternary diagram (Fig. 4-6b). Samples outside of the diagram can indicate an unknown source of SOC, but that is unlikely as almost all plants encoun tered were measured and none had high N:C (low C:N) ratios (Table 4-1). According to 13C only, seston and seagrass are probably both sources because reference seagrass SOC 13C falls in the middle of those end members. The contribution of macroalgae is unknown though due to its intermediate 13C. Mangroves were not chosen as a potential source for seagrass sediments as mangrove litter was observed in frequently on seagrass sediments. Therefore mangroves influence to SOC was believed to be mediated through seston. Sources to the reference mangrove litter layer change with season as the ternary diagram indicates that seagrass and seston are the dominant sources in July but mangroves are the dominant sources in November (Fig. 4-7a). C onclusions from the ternary diagram match field observations. The litter layer was primarily seag rass wrack in July but was primarily partiallydecomposed mangrove leaves in November. Since the litter layer is one of the main sources to reference mangrove SOC, seagrass and mangroves are therefore also sources to mangrove SOC through the litter. Sources to seagrass floc va ried seasonally for refere nces but not SL 15. Seston was a dominant floc OC source for all se asons and sites according to the ternary mixing

PAGE 125

125 diagram (Fig. 4-7b). Macroalgae was also a do minant source for November reference floc samples. Ternary diagrams indicated that seston is a dominant source to almost all sediments and surface layers regardless of site. Seston is not the only source, however, because all sediments and surface layers are more enriched in 13C than seston (Fig. 4-3). A review of source determination studies in mangrove and seagra ss sediments found that seston was a dominant source at 47% of sites (Chapter 2). OC in sediments of young mangrove forests were dominated by algal and seston sources, just as the constructed sediments were in this study (Marchland et al. 2003; Alongi et al. 2004). In sediments with low %OC, as in this study (Chapter 3), the dominant macrophytes such as ma ngroves or seagrass seemed less likely to be significant OC sources (Gonnoeea et al. 2004; Ke nnedy et al. 2004). Middelburg et al. (1997) showed a significant relationship of decreasing 13C (more depleted than in situ macrophyte 13C) with decreasing %SOC. These trends may be because wh en seston settles onto sediments, it does so with inorganic particles, which dilute SOC, lowering the %OC. Seston comes from a variety of sources as it is made up of phyt oplankton, zooplankton, bacteria, and detritus (Fig. 4-8). Its high 13C in this study is indicative of a mangrove or terrestrial origin. Its high N:C (low C:N), how ever, indicates a mixture of phytoplanktom, which have C:N ratios from 7.7 to 10.1 (Holligan et al. 1984), and bacterioplankton, which have C:N ratios from 2.6 to 4.3 (Lee and Fuhrman 1987). Ot her estuarine studies si milarly had seston with low 13C and low C:N ratios (Hemminga et al. 1994; Cifuentes et al. 1996; Zhou et al. 2006). Cifuentes et al. (1996) demonstrat ed that bacteria in the water column were likely immobilizing N in the process of decomposing terrestrial-d erived organic matter, which could lead to incorporation of that nitrogen into organic matter during humification and a lower C:N ratio.

PAGE 126

126 Low C:N ratios of many sediment samples are co ncerning because it may have lead us to overstate the importance of seston as a source because it is the onl y source with equally low C:N ratios. Just as microbial activity likely lowere d C:N ratios in seston (C ifuentes et al. 1996), it could lower C:N ratios in sediment s. Decreasing sediment C:N ratios during diagenesis has also occurred in other source determination st udies (Thimdee et al. 2003; Gonnoeea et al. 2004; Kennedy et al. 2004). Changes due to bacteria are likely because a high percentage of TOC in these sediments is microbial biomass (11 to 63% ; Chapter 3). Decreases in source C:N ratios during decomposition may also explain the relativ ely low C:N ratios of the sediments compared with living source material. Unfortunately, C:N ratios during decomposition were not measured in this study. Results of studies that measured decomposition in similar systems were equivocal (see source characteristics secti on). Another reason for low C:N ratios is the eutrophication of the IRL, which has greatly increased the availabi lity of inorganic N sources (Sigua and Tweedale 2003). Due to the influence of f actors other than source identity in determining sediment C:N ratios, caution is emphasized in interp reting ternary diagram results. Conclusion In all sediments, seston was a dominant source and diagenesis of organic matter within sediments lowered sediment C:N ratios (Fig. 4-9). Because the other main sources differed between SL 15 and reference sediments (Fig. 49), their abilities to sequester SOC probably differ too. The litter bag decomposition study su ggests which SOC sources are likely to be sequestered in sediments the longest This information allows us to predict how OC storage will differ in sediments of SL 15 and reference sites. Since seston is an OC source for all sediments, the fact that its decomposition was not measured should not greatly affect these predictions. Because fast-decaying macroalgae OC dominates in SL 15 seagrass sediments, they are unlikely to store OC for as long as reference seagrass se diments. A year after construction, SL 15s

PAGE 127

127 seagrass sediments therefore do not store C as well as references. OC in both SL 15 and reference mangrove sediments are a mixture of seston and vascular plants (terrestrial plants in SL 15 and mangrove/seagrass via litter in references). Since terrest rial plants most likely have decay rates similar to mangroves and slower than most seagrass species, it is possible that the length of OC storage in SL 15 and reference ma ngrove sediments are currently similar. The labile algal mat, however, has the potential of becoming a main source in constructed mangrove sediments because it caused sediment 13C enrichment throughout the year, which may ultimately shorten the length of constructed mangrove OC storage.

PAGE 128

128 Table 4-1. 13C () and C:N ratios for all potential s ources of organic carbon to mangrove and seagrass sediments in SL 15 and referen ce sites averaged over various collection times and plant parts (unless otherwise noted). Values in parentheses are SE; where no standard error is listed, the valu e is for a single composite sample. Location Species 13C () C:N Subtidal Seagrass -11.54 (0.84) Leaves -10.95 (1.3) 14 (0.52) Roots -12.42 (0.78) 31 (2.5) Syringodium filiforme -9.23 (0.86) 14 (0.5) Thalassia testudinum -9.83 (1.2) 13 (0.9) Halodule beaudettei -12.62 11 Halophila johnsonii -20.07 16 Epiphytes -16.20 (1.9) 11 (0.6) Macroalgae -21.00 (0.98) 18 (1.6) Acanthophora spicifera -17.11 (0.78) 12 (0.3) Caulerpa sertulariodes -18.36 14 Sargassum spp. -17.48 (0.29) 28 (0.9) Daysa baillouviana -32.06 15 Ulva spp. -20.64 14 Chaetomorpha linum -25.29 25 SL subtidal macroalgae -21.75 (0.86) 16 (2) Rosenviga intricata -20.94 (1.1) 16 (0.8) Hypnea cervicornis -19.69 (1.2) 21 (0.1) Gracilaria tikvahiae -22.50 (1.4) 11 (0.4) Enteromorpha spp. -25.24 19 Seston -26.29 (0.47) 6.5 (0.2) May 2006 -27.30 (0.44) 7 (0.5) September 2006 -25.78 (2.2) 6 (0.5) October 2006 -24.23 (0.21) 6 (0.2) November 2006 -24.58 (1.3) 6.5 (0.7) February 2007 -30.10 (0.84) 7.5 (1) Intertidal Spartina alterniflora -12.94 (0.30) Leaves -13.37 (0.27) 27 (2) Roots -12.30 (0.04) 42 (11) Sueda linearis -29.22 14 Mangrove -26.95 (0.24) Leaves -27.27 (0.32) 27 (1.4) Roots -26.18 (0.20) 58 (3.7) Avicennia germinans -27.46 (0.53) Leaves -27.77 (0.49) 22 (1.7) Roots -25.26 55 Rhizophora mangle -27.17 (0.36) Leaves -27.32 (0.47) 31 (1.7) Roots -26.34 (0.30) 55 (9) Laguncularia racemosa -26.31 (0.84) 26 (3.0) Terrestrial C3 terrestrial -27.54 (0.44) 33 (4) Schinus terebenthifolius (leaves) -28.30 31 Casuarina equisetifolia (needles) -26.33 34 Coccoloba uvifera (leaves) -27.77 (0.32) 32 (11) Borrichia frutescens -27.86 (0.60) 21 (0.9) Distichlis spicata -13.67 (0.86) 27 (2.6)

PAGE 129

129 Table 4-2. Decay constants ( SE) and turnove r times calculated from a nonlinear regression (exponential decay) of litter bag experiment data. Species k (day-1) Turnover time (days) Haludule beaudettei 0.0049 (0.0006) 203 Thalassia testudinum 0.0099 (0.001) 101 Syringodium filiforme 0.046 (0.006) 22 Acanthophora spicifera 0.070 (0.006) 14 Sargassum spp. 0.019 (0.001) 53 Avicennia germinans 0.0093 (0.0004) 109 Rhizophora mangle 0.0047 (0.0004) 213

PAGE 130

130 Table 4-3. Results of ANOVAs comparing 13C values in SL 15 and reference mangrove and seagrass sediments and surface layers (right) and of repeated measures ANOVAs of 13C values in SL 15 sediments and surface layers (left). ANOVA Effect 13C ANOVA Effect 13C Repeated measures Sediment comparison Mangrove 0-10 Month Mangrove Site NS Depth NS Month NS Month*Depth NS Depth NS Mangrove algal mat Month ** Site*Month NS Seagrass 0-10 Month NS Site*Depth NS Depth NS Month*Depth NS Month*Depth NS Seagrass Site *** Seagrass accreted Month Month NS Seagrass floc Month Depth NS Site*Month NS Site*Depth NS Month*Depth NS Surface Comparison Mangrove algal mat/litter Site *** Month NS Site*Month *** Seagrass floc Site ** Month *** Site*Month *** For significance NS=not significant, p <0.05, **p < 0.01, ***p < 0.0001. Please see Appendix A for a table listing how these data were transformed prior to running the 3-way ANOVA.

PAGE 131

131 Table 4-4. Mean 13C and C:N ( SE) for sediments and surface layers of SL 15 and reference mangrove and seagrass systems. System and month Depth 13C C:N (molar) SL 15 Reference SL 15 Reference July mangrove Algal Mat/ Litter -15.72 (0.46) -18.21 (0.63) 8.2 (0.2) 9.3 (0.3) 0-5 cm -23.57 (1.03) -24.12 (0.91) 5.0 (1) 9.6 (3) 5-10 cm -23.70 (0.50) -22.17 (1.40) 8.0 (2) 9.2 (1) Nov. mangrove Algal Mat/ Litter -11.96 (0.47) -24.43 (1.56) 8.1 (0.9) 21 (5) 0-5 cm -21.96 (0.18) -23.42 (0.70) 5.5 (0.3) 7.0 (1.1) 5-10 cm -24.11 (0.46) -24.10 (0.31) 6.2 (0.8) 8.8 (0.9) July seagrass Floc -21.05 (0.14) -18.93 (0.27) 8.2 (1) 5.7 (0.6) Accreted -21.00 (0.44) 7.6 (1) 0-5 cm -21.03 (0.35) -18.97 (0.29) 6.9 (0.3) 6.7 (0.2) 5-10 cm -20.93 (0.60) -19.11 (0.20) 7.0 (1) 6.6 (0.2) 10-15 cm -18.79 (0.33) 1.0 (0.2) Nov. seagrass Floc -19.50 (0.15) -24.46 (0.56) 6.8 (0.1) 10.8 (0.6) Accreted -20.06 (0.20) 6.1 (0.4) 0-5 cm -21.55 (0.41) -17.87 (0.21) 6.7 (0.8) 5.4 (0.2) 5-10 cm -20.80 (0.48) -18.80 (0.46) 7.3 (1) 5.6 (0.4) 10-15 cm -19.10 (0.48) 6.1 (0.4)

PAGE 132

132 Table 4-5. Mean 13C or 13C ranges of means for s ources in this study and in the literature. Means are averaged across and ranges are acros s plant parts, species, and sites for this study and where applicable in the literature The sources listed here are the main SOC sources that were used in this studys ternary diagrams. 13C () 13C () Source This study Other studies Data sourceaSeagrass Mean: -11.54 -10.0 1 Range: -20.07 to -9.23 -19.7 to -10.7 2 -13.3 to -5.8 3 -12.4 4 -12.2 5 -16.1 to -11.9 6 -10.5 7 -23 to -3, -10 (mode) 8 -10.4 to -7.2 9 -14.6 to -8.8 10 -12.7 to -11.4 11 Mangrove Mean: -26.95 -27.0 12 Range: -27.77 to -25.26 -28.3 to -24.1 2 -29.0 to -27.0 13 -28.4 to -27.9 3 -28.8 7 -28.2 14 -27.9 15 -30.1 to -28.3 16 -29.7 to -25.9 17 Macroalgae Mean: -21.00 -31.7 to -16.6 11 Range: -32.06 to -17.11 -21.5 to -15.0 18 -26.0 to -20.9 15 -15.61 7 Seston Mean: -26.29 -22.0 to -21.0 12 Range: -30.10 to -24.23 -23.0 to -20.5 13 -18.4 1 -23.3 to -13.7 b 2 -27.6 to -12.1 3 -22.1 4 -24.7 5 -25.32 to -22.06 b 6 -20.6 7 -22.6 b 19 -28.1 to -20.8 b 20 -26.4 b 21 Terrestrial (C3) Mean: -27.54 -28 to -25 22 Range: -28.30 to -26.33 -30 to -25 23 -26 24 a1, Canuel et al. 1997; 2, Hemminga et al. 1994; 3, Kennedy et al. 2004; 4, Papadimitriou et al. 2005; 5, Gacia et al. 2002; 6, Gonnoeea et al. 2004; 7, Thimdee et al. 2003; 8, Hemminga and Mateo 1996; 9, Anderson and Fourqurean 2003; 10, Vizzini et al. 2003; 11, Smit et al. 2005; 12, Jennerjahn and Ittekkot 2002; 13, Bouillon et al. 2003; 14, Bouillon et al. 2004; 15, Bouillon et al. 2002; 16, Lallier-Verges et al. 1998; 17, Muzuka and Shunula 2006; 18, Fenton and Ritz 1988; 19, Zhou et al. 2006; 20, Dittmar et al. 2001; 21, Cifuentes et al 1996; 22, Miserocchi et al. 2007; 23, Kang et al. 2007; 24, Ogrinc et al. 2005 bCalled particulate organic matter (POM) or suspended particulate matter (SPM) by the authors

PAGE 133

133 Figure 4-1. The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset). SL 15 is the large island in the center. Circles are mangrove system plots and squares are seagrass system plots. Symbols outside of SL 15 are the reference sites, which have one plot each.

PAGE 134

134 Figure 4-2. Core from SL 15 seagrass system illustrating the surface laye r (floc) and different sediment depths (accreted layer, 0-5 cm, and 5-10 cm). Note the difference in color between the accreted layer and 0-5 cm depth. settledfloc accreted layer 0-5 cm layer 5-10 cm layer

PAGE 135

135 Figure 4-3. 13C averaged over July and November 2006 for SL 15 and reference sediments and surface layers compared to mean 13C of potential sources. terrestrial mangroves macroalgae Spartina seagrass epiphytes 5-10 0-5 5-10 0-5 litter algal mat floc 5-15 0-5 floc accreted 0-5 5-10 seston Potential Source Ref Seagrass SL 15 Man g rove Ref Mangrove SL 15 Seagrass 13C

PAGE 136

136 Figure 4-4. Mean 13C of SL 15 sediments and surface layers over the first year after construction. Error bars are SE. Reference lines are 13C averaged over depth (for sediment s) and month (except for reference floc) for the respective reference systems.

PAGE 137

137 SL 15 Mangrove SedimentsC -30-28-26-24-22-20-18-16-14-12-10N:C 0.00 0.05 0.10 0.15 0.20 0.25 0.30 algal mat terrestrial seston Figure 4-5. N:C vs. 13C in ternary mixing diagrams of three potential OC sources and mangrove sediments. Circles are the m ean end member values and boxes are standard deviation of N:C and 13C. Triangles are mangrove sediment values for SL 15 (A) and the reference (B). A

PAGE 138

138 Reference Mangrove Sediments 13 C -30-28-26-24-22-20-18-16N:C 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 seston mangrove litter Figure 4-5. Continued B

PAGE 139

139 SL 15 Seagrass Sediments13C -30-25-20-15-10-5N:C 0.00 0.05 0.10 0.15 0.20 0.25 0.30 seston macroalgae seagrass Figure 4-6. N:C vs. 13C in ternary mixing diagrams of three potential OC sources and seagrass sediments. Circles are the mean end member values and boxes are standard deviation of N:C and 13C. Filled triangles are 0-10 cm sediment values for SL 15 (A) and 0-15 cm sediment values for referen ce (B). Open triangles are accreted layer values for SL 15 (A). A

PAGE 140

140 Reference Seagrass Sediments13C -30-25-20-15-10-5N:C 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 seston macroalgae seagrass Figure 4-6. Continued B

PAGE 141

141 Reference Mangrove Litter 13C -30-25-20-15-10-5N:C 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 seston mangrove seagrass Figure 4-7. N:C vs. 13C in ternary mixing diagrams of th ree potential OC sources and surface layers. Circles are the mean end member values and boxes are standard deviation of N:C and 13C. Filled triangles are reference lit ter layer values (A) and SL 15 floc values (B). Open triangles ar e reference floc values (B). A November July

PAGE 142

142 Seagrass Floc13C -30-25-20-15-10-5N:C 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 seston macroalgae seagrass Figure 4-7. Continued B July November

PAGE 143

143 Water Column Seston 13C C:N Detritus ? 13C Seagrass 13CC:N Plankton ? 13CC:N Bacteria ? 13C C:N Terrestrial 13CC:N Mangrove 13CC:N Diagenesis C:N ?Macroalgae 13CC:N ? Water Column Seston 13C C:N Detritus ? 13C Seagrass 13CC:N Seagrass 13CC:N Plankton ? 13CC:N Bacteria ? 13C C:N Terrestrial 13CC:N Terrestrial 13CC:N Mangrove 13CC:N Diagenesis C:N ?Macroalgae 13CC:N ? Figure 4-8. A theoretical diagram of organic carbon sources that may constitute seston and how they affect seston 13C and C:N. Arrow sizes indicate the possible relative contributions of each source. indicates depleted 13C and low C:N, indicates enriched 13C and high C:N, and indicates mid-range 13C and C:N.

PAGE 144

144 SL 15 Mangrove Ref Mangrove Figure 4-9. Main sources and how they affect surface layer and sediment 13C and C:N. Arrow sizes indicate the relative contributions of each source. indicates depleted 13C and low C:N, indicates enriched 13C and high C:N, and indicates mid-range 13C and C:N. Water Column Seston 13C Terrestrial 13CC:NC:N Algal Mat13C C:N Sediment13C C:N Diagenesis ? 13C C:N Water Column Seston 13C Terrestrial 13CC:N Terrestrial 13CC:NC:N Algal Mat13C C:N Sediment13C C:N Sediment13C C:N Diagenesis ? 13C C:N Water Column Seston 13CC:N Litter13C C:N Sediment13C C:N Diagenesis ? 13C C:N Seagrass 13CC:N Mangrove 13CC:N Water Column Seston 13CC:N Litter13C C:N Sediment13C C:N Sediment13C C:N Diagenesis ? 13C C:N Seagrass 13CC:N Seagrass 13CC:N Mangrove 13CC:N

PAGE 145

145 SL 15 Seagrass Ref Seagrass Figure 4-9. Continued. Water Column Seston 13CC:N Floc13C C:N Sediment13C C:N Diagenesis ? 13C C:N Seagrass 13CC:N Macroalgae 13CC:N Water Column Seston 13CC:N Floc13C C:N Sediment13C C:N Diagenesis ? 13C C:N Seagrass 13CC:N Seagrass 13CC:N Macroalgae 13CC:N Water Column Seston 13CC:N Floc13C C:N Sediment13C C:N Diagenesis ?13C C:N Macroalgae 13CC:N Seagrass 13CC:N Water Column Seston 13CC:N Floc13C C:N Sediment13C C:N Diagenesis ?13C C:N Macroalgae 13CC:N Seagrass 13CC:N Seagrass 13CC:N

PAGE 146

146 CHAPTER 5 SYNTHESIS Coastal e cosystems including salt marshes, seagrass beds, and mangrove forests are more effective carbon (C) sinks than terr estrial systems and freshwater wetlands (Chapter 2). These ecosystems store large amounts of OC and actively accumulate OC at high rates; about 44.6 Pg C is stored and about 120 Tg C y-1 accumulates in salt marsh and mangrove sediments globally (Jennerjahn and Ittekkot 2002; Ch mura et al. 2003). Many co astal ecosystems have been degraded or lost due to anth ropogenic disturbances (Valiela et al. 2001; Kennish 2002; Zedler 2004). Destruction of coastal eco systems increases atmospheric CO2 concentrations because their organic C (OC) stores are often mineralized as a result a nd their future OC sequestration capacity is lost (Duarte et al 2005; Bridgham et al. 2006). Gene rally in the United States, the destruction of coastal ecosystems must be mitigated by restoring or creating coastal ecosystems elsewhere. Whether mitigation of seagrass and mangrove systems restores the C sink capacity is currently not well-studied. In the present study, functional trajectories of sediment OC (SOC) parameters in a constructed mangrove and seagrass system in th e Indian River Lagoon, Florida were measured. Sediment OC (SOC) parameters in constructed systems were also compared to mature reference systems to indicate if constructed sediments had reached functional equiva lence in terms of OC storage. The objectives of this study were: 1) to determine s hort term trajectories of SOC pools in a constructed mangrove forest and seagrass be d; 2) to compare SOC pools in the constructed system with those in reference systems; 3) to co mpare the lability of SOC in the constructed and reference systems; 4) to determine and compare significant sources to th e total SOC pool in the constructed and reference systems. The hypothese s were: 1) in the short term, storage in the three OC pools studied would increa se in the constructed systems, but would not reach the level

PAGE 147

147 of storage in the references OC pools; 2) OC lability would be greater in sediments of constructed systems than in reference sediments; 3) SOC sources in c onstructed systems would be macroalgae or plankton, while SOC sources in reference systems would be vascular plants, like mangroves and seagrass. Key findings addre ssing each objective and the validity of the hypotheses are presented below. Objective One: Short Term Trajectories of Sediment Organic Carbon Pools Contrary to the hypothesis, func tional trajectories were not fo llowed by OC parameters in the constructed site sediments. Instead of st eady increases, SOC parameters either remained unchanged or increased and decr eased throughout the ye ar, driven by seasonal changes in the water column. The only sediment functional tr ajectory was followed by the mangrove systems bulk density, which decreased throughout the y ear but remained above reference levels. Functional trajectories were somewhat followed by surface layers as both microbial biomass C (MBC) and total OC (TOC) increased. Due to their proximity to OC inputs, it is logical that OC should increase in the surface layers before th ey increase in sediments. However, whether increases in surface layer OC were due to a reco vering function or an annual pattern could not be discerned. For example, the increase in floc MBC and TOC followed the same trend as total suspended solids, a water quality parameter. Ov erall, one year was not sufficient time to map OC functional trajectories in the constructed mangrove and seagrass system. The lack of a functional trajectory did not prec lude the OC parameters from be ing functionally equivalent to reference values. Objective Two: Comparisons of Sediment Organic Carbon Pools The hypothesis that SOC pools would be smalle r in constructed systems was by and large correct for mangrove sediments but not for seag rass sediments (Fig. 5-1 and 5-2). Floc and accreted layers of constructed seagrass sediment s reached or exceeded functional equivalence for

PAGE 148

148 all three OC poolsTOC, Extractable OC (ExO C), and MBC. The 0-10 cm depths, also reached equivalence for ExOC pools. On a storag e (areal) basis, equivalence was also reached by mangrove 0-5 cm depths for ExOC and MBC a nd seagrass 0-10 cm depths for MBC. This equivalence was only on a storage basis because it was driven by greater bulk densities in the constructed sediments. Seagrass sediments reached SOC pool equivalence more than mangrove sediments due to their constant inundation, parent material, and lo wer equivalence goal (reference seagrass sediments had less TOC and Ex OC than reference mangrove sediments). SOC pool sizes were not the only factors that indicated if cons tructed systems had attained functionally equivalent OC storageinformati on about OC accumulation ra tes and OC lability was also needed. OC accumulation rates in the constructed mangrove and seagrass systems were similar to literature values if accumulation in surface layers was considered. It was unknown if the constructed systems could sust ain these accumulation rates over the long term or if the rates reflected an immediate response in SOC after cons truction. Larger prop ortions of the TOC pool were MBC in constructed 0-10 cm sediments indi cating greater SOC lability and therefore less SOC storage in constructed systems. Objective Three: Comparisons of Se diment Organic Carbon Lability Generally, constructed systems SOC was more labile than reference system SOC for both mangrove and seagrass sediments (Fig. 5-1 a nd 5-2). Lability was only similar between constructed and reference systems in the upper portion of seagrass sediments and in seagrass floc. These results confirm hypothesis two. Grea ter lability of OC in the constructed system indicates that the constructed sy stem does not function as well as reference systems in terms of OC storage. Even when SOC pool sizes are simila r to references, as in the 5-10 cm depth of constructed seagrass sediments, greater OC labili ty indicates that OC storage is not functionally equivalent. The more labile OC is, the mo re likely it will be mineralized by microbes and

PAGE 149

149 respired to CO2 instead of being stored in sediments long term. Differences in OC lability were partially due to differing C limitations and to di fferences in SOC sources between constructed and reference systems. Objective Four: Comparisons of Se diment Organic Carbon Sources Sources to the SOC pool differed between c onstructed and reference systems, but not to the extent that was hypothesized (Fig. 5-1 and 52). Ternary diagrams suggested that seston from the water column was a main SOC source for all systemsconstructed and natural. The true importance of seston, however, was unclear b ecause the low sediment C: N ratios that led to the conclusion that seston was a main source, can al so result from diagenetic transformations. In mangrove sediments, both systems had lignin-cont aining higher plants as other main sources terrestrial plants in the constructed system a nd mangroves/seagrass via litter in the reference systems. The effect of sources on OC storage in mangrove sediments was therefore similar; but there was an indication that th e labile algal mat was becoming an increasingly important source in constructive sediments, which would shorten OC storage times. Sources in reference seagrass sediments were unclear. It was apparent though, that a greater amount of SOC was derived from macroalgae in the constructed system than in the reference system. Litter bag studies demonstrated that macroalgae generally have the fastest decomposition rates of all aquatic macrophytes, indicating that the m acroalgae-derived OC would not be stored in constructed sediments for long amounts of time. Conclusion Overall, neither mangrove nor seagrass sediments of the constructed system are functionally equivalent to their re spective references in regards to OC storage (Fig. 5-1 and 5-2). Recovery indices indicate how close various parameters are to equivalence with references.

PAGE 150

150 reference dconstructeXXRI log (5-1) In Equation 5-1, RI is the recovery index, Xconstructed is the value of the parameter in the constructed system, and Xreference is the value of the parameter in the reference system. RIs equal to zero indicate equivalence, less than zero indica te that equivalence has not been reached, and greater than zero indicate equivalence has been surpassed. The constructed seagrass system is closer to equivalence than the constructed mangrove system (Fig.5-3). Upper depths of constructed seagrass sediments had similar SOC pool s and lability to upper depths of reference seagrass sediments, causing the seagrass sediments to be closer to equivalence. As the mangroves and seagrasses within the constructed sy stems mature, it is likely that their SOC will become less algae-derived, leading to lower OC lability and better OC storage. Dominance of seston as a source in all systems means that a sw itch in less significant sources may take time to register in the sediments. Ultimately, if the OC pools and lability in reference systems are any indication (Fig. 5-1 and 5-2) and if functional trajectories are followed in the future, the constructed mangrove system will become more ef fective at OC storage than the constructed seagrass system. This study adds to the body of research on f unctional trajectories. It is one of two known studies on seagrass trajectories and the only known study of ma ngrove trajectories. More importantly, it is the first known study to examine the trajectory of OC storage in depth. If constructed and restored coastal ecosystems store and accumulate OC as well as their established counterparts, corporations and governments could construct coastal system s in exchange for C credits. This action can replace lost system s and restore many of the ecologically important functions these systems provide. Conclusions based on this study are limited because it only followed constructed systems during the first year of recovery. The constructed mangrove and

PAGE 151

151 seagrass system in this study accumulated SOC at rates similar to rates in mature systems over the first year. If these rates are sustained and mo re OC is stored in long-term, recalcitrant pools, then the constructed systems will be effective C sinks. Longer term studies are needed to fully assess the effectiveness of constructing coastal ecosystems for OC storage.

PAGE 152

152 Extractable OC 2.8 g m-2(C) 2.0 g m-2(R) Other OC 555 g m-2 (C) 910 g m-2 (R) 0.047 g C g-1OC hr-1 (C) 0.036 g C g-1OC hr-1(R) Seagrass (R) and Algae (C) Seston (C, R) Trapped particles Litter Decomposition k=0.0049 0.046 d-1 (Seagrass) k=0.019 0.070 d-1(Algae) Microbial Biomass C 256 g m-2 (C) 295 g m-2 (R) Sediment Water Column Extractable OC 2.8 g m-2(C) 2.0 g m-2(R) Other OC 555 g m-2 (C) 910 g m-2 (R) 0.047 g C g-1OC hr-1 (C) 0.036 g C g-1OC hr-1(R) Seagrass (R) and Algae (C) Seston (C, R) Trapped particles Trapped particles Litter Decomposition k=0.0049 0.046 d-1 (Seagrass) k=0.019 0.070 d-1(Algae) Microbial Biomass C 256 g m-2 (C) 295 g m-2 (R) Sediment Water Column Figure 5-1. A modified seagrass bed carbon cycle showing values from this study in constructed (C) and reference (R) system s. Organic carbon (OC) pools are the sum of sediment and surface layer means (July and November data). Rates of microbial carbon respiration are the mean of all depths (sediment and surface la yers) in July and November, adjusted from an O2 uptake rate to a carbon release rate by an assumed 1 O2 to 6 C molar ratio. Bolded words are th e main contributors to sediment OC pools.

PAGE 153

153 Terrestrial (C) Litter Seagrass (R) Seston (C, R) Sediment Atmosphere Water column Mangrove (R) Litter Decomposition k=0.0049 0.046 d-1 (Seagrass) k=0.0047 0.0093 d-1(Mangrove) 0.075 g C g-1OC hr-1 (C) 0.026 g C g-1OC hr-1(R) Other OC 453 g m-2 (C) 2270 g m-2 (R) Microbial Biomass C 206 g m-2 (C) 221 g m-2 (R) Extractable OC 12.8 g m-2(C) 15.5 g m-2(R) Algal Mat (C) Terrestrial (C) Litter Seagrass (R) Seston (C, R) Sediment Atmosphere Water column Mangrove (R) Litter Decomposition k=0.0049 0.046 d-1 (Seagrass) k=0.0047 0.0093 d-1(Mangrove) 0.075 g C g-1OC hr-1 (C) 0.026 g C g-1OC hr-1(R) Other OC 453 g m-2 (C) 2270 g m-2 (R) Microbial Biomass C 206 g m-2 (C) 221 g m-2 (R) Extractable OC 12.8 g m-2(C) 15.5 g m-2(R) Algal Mat (C) Figure 5-2. A modified mangrove forest carbon cycle showing values from this study in constructed (C) and reference (R) systems. Organic carbon (OC) pools are the sum of sediment and surface layer means (July and November data). Rates of microbial carbon respiration are the mean of all depths (sediment and surface layers) in July and November, adjusted from an O2 uptake rate to a carbon release rate by an assumed 1 O2 to 6 C molar ratio. Bolded words are th e main contributors to sediment OC pools.

PAGE 154

154 -1.00 -0.50 0.00 0.50 1.00 TOC (S)MBC (S)ExOC (S)Lability Mangrove Seagrass Figure 5-3. Recovery indices of three organic carbon (OC) pools and OC lability parameters for constructed mangrove and seagrass systems. For each system, OC pools are summed for all sediment depths and surface layers a nd lability is the average of all depths and surface layers. TOC=total organic carbon, MBC=microbial biomass carbon, ExOC=extractable organic carbon, and (S) indicates that these OC pool parameters were calculated on a storage basis.

PAGE 155

155 APPENDIX A STATISTICAL TRANSFORMATIONS Table A-1. How data were transform ed to meet the normality assumption prior to running ANOVAs. For parameters, TOC=total organic carbon, TN= total nitrogen, C:N= carbon to nitrogen ratio, ExOC=extractable organic ca rbon, and MBC=microbial biomass carbon. For transformations, NT= not transformed, Sqrt=square root, and a C (as in X-C) indicates that a constant was subtracted or adde d to a parameter before it was tr ansformed via square root, log, arcsine, etc. Chapter ANOVA Depths Parameter Transformation 3 Mangrove factorial Sediments 0-10 Bulk density NT TOC (conc) Log 10 (Arcsin) TOC (storage) Log 10 TN (conc) Log 10 (Arcsin) C:N Sqrt ExOC (conc) Log e ExOC (storage) Log e MBC (conc) Sqrt (MBC-C) MBC (storage) Log e Lability Log 10 Seagrass factorial Sediments 1-3 Bulk density NT TOC (conc) NT TOC (storage) NT TN (conc) NT C:N Log e ExOC (conc) Log e (ExOC-C) ExOC (storage) Log 10 MBC (conc) Sqrt (MBC-C) MBC (storage) NT Lability Log 10 Mangrove factorial Surface layers Bulk density Sqrt TOC (conc) Log 10 (Arcsin) TOC (storage) Log 10 TN (conc) Log 10 (Arcsin) C:N Log e (C:N-C) ExOC (conc) Log e (ExOC-C) ExOC (storage) Log 10 MBC (conc) Sqrt MBC (storage) Sqrt Lability Log e Seagrass factorial Surface layers Bulk density NT TOC (conc) Arcsin (sqrt) TOC (storage) NT TN (conc) Sqrt C:N Sqrt ExOC (conc) Sqrt ExOC (storage) Sqrt MBC (conc) Sqrt MBC (storage) Log e Lability NT Mangrove repeated Measures Sediments 0-10 Bulk density Log e TOC (conc) Sqrt ExOC (conc) NT

PAGE 156

156 Table A-1. Continued Chapter ANOVA Depths Parameter Transformation 3 TN (conc) Sqrt C:N Log e Lability Sqrt Algal mat Bulk density NT TOC (conc) NT ExOC (conc) NT MBC (conc) Sqrt (MBC-C) TN (conc) Log e (TN-C) C:N Log e Lability Log e Seagrass repeated Measures Sediments 0-10 Bulk density Log e TOC (conc) Sqrt ExOC (conc) NT MBC (conc) NT TN (conc) Sqrt C:N Sqrt Lability Sediments accreted Bulk density NT TOC (conc) Log e ExOC (conc) Sqrt (ExOC-C) MBC (conc) Sqrt TN (conc) Log e C:N Log e (C:N-C) Lability NT Floc Bulk density Log 10 TOC (conc) NT ExOC (conc) Sqrt (ExOC-C) MBC (conc) NT TN (conc) NT C:N Log e (C:N-C) Lability NT 4 Mangrove factorial Sediments 0-10 13c NT Seagrass factorial Sediments 1-3 13c NT Mangrove factorial Surface layers 13c NT Seagrass factorial Surface layers 13c Log 10 ( 13C*-1) Mangrove repeated measures Sediments 0-10 13c NT Algal mat 13c NT Seagrass repeated measures Sediments 0-10 13c NT Sediments accreted 13c NT Floc 13c NT

PAGE 157

157 LIST OF REFERENCES ABER, J. D. AND J. M. MELILLO. 2001. Terrest rial Ecosystems, 2nd ed. Harcourt Academic Press. AKE-CASTILLO, J. A., G. VAZQUEZ, AND J. LOPEZ-PORTILLO. 2006. Litterfall and decomposition of Rhizophora mangle L. in a coastal lagoon in the southern Gulf of Mexico. Hydrobiologia 559: 101-111. ALONGI, D. M. 2002. Present stat e and future of the world's mangrove forests. Environmental Conservation 29: 331-349. ------, J. PFITZNER, L. A. TROTT, F. TI RENDI, P. DIXON, AND D. W. KLUMPP. 2005. Rapid sediment accumulation and microbial mi neralization in forests of the mangrove Kandelia candel in the Jiulongjiang Estuary, Chin a. Estuar. Coast. Shelf Sci. 63: 605-618. ------, AND OTHERS. 2004. Sediment accumulation and organic material flux in a managed mangrove ecosystem: estimates of land-ocean-atmosphere exchange in peninsular Malaysia. Mar. Geol. 208: 383-402. ------, AND OTHERS. 2001. Organic carbon accumulation and metabolic pathways in sediments of mangrove forests in southern Thailand. Mar. Geol. 179: 85-103. ANDERSON, W. T., AND J. W. FOURQUREAN. 2003. In traand interannual variability in seagrass carbon and nitrogen stable isotopes from south Florida, a preliminary study. Org. Geochem. 34: 185-194. APHA. 1992. Biological Oxygen Demand. Standard methods for the examination of water and wastewater. American Pub lic Health Association. BAILEY, V. L., A. D. PEACOCK, J. L. SMITH, AND H. BOLTON. 2002. Relationships between soil microbial biomass determined by chloroform fumigation-extraction, substrate-induced respiration, and phospholipid fatty acid an alysis. Soil Biol. Biochem. 34: 1385-1389. BIERMAN, P. R. AND OTHERS. 1998. Erosion, Weathering, and Sedimentation, p. 647-678. In C. Kendall and J. J. McDonnell [eds.], Is otope Tracers in Catchment Hydrology. Elsevier. BOSCHKER, H. T. S., A. WIELEMAKER, B. E. M. SC HAUB, AND M. HOLMER. 2000. Limited coupling of macrophyte production and bacterial carbon cycling in the sediments of Zostera spp. meadows. Mar. Ecol.-Prog. Ser. 203: 181-189. BOUILLON, S., F. DAHDOUHGUEBAS, A. RAO, N. KOE DAM, AND F. DEHAIRS. 2003. Sources of organic carbon in mangrove sedi ments: variability and possible ecological implications. Hydrobiologia 495: 33-39.

PAGE 158

158 ------, N. KOEDAM, A. V. RAMAN, AND F. DEHAIRS. 2002. Primary producers sustaining macro-invertebrate communities in intertidal mangrove forests. Oecologia 130: 441-448. ------, T. MOENS, I. OVERMEER, N. KOEDAM, AND F. DEHAI RS. 2004. Resource utilization patterns of epifauna from mangrove forests with contrasting inputs of local versus imported organic matte r. Mar. Ecol.-Prog. Ser. 278: 77-88. BOURGUES, S., I. AUBY, R. DEWIT, AND P. J. LABOURG. 1996. Differential anaerobic decomposition of seagrass ( Zostera noltii ) and macroalgal ( Monostroma obscurum ) biomass from Arcachon Bay (France). Hydrobiologia 329: 121-131. BRIDGHAM, S. D., J. P. MEGONIGAL, J. K. KELLER, N. B. BLISS, AND C. TRETTIN. 2006. The carbon balance of North American wetlands. Wetlands 26: 889-916. BROOME, S. W., E. D. SENECA, AND W. W. WOODHOUSE 1988. Tidal salt-marsh restoration. Aquatic Botany 32: 1-22. BRUNSKILL, G. J., I. ZAGORSKIS, AND J. PFITZNER. 2002. Carbon burial rates in sediments and a carbon mass balance for the He rbert River region of the Great Barrier Reef continental shelf, North Queensland, Australia. Estuar. Coast. Shelf Sci. 54: 677-700. BUCKERIDGE, K. M., AND R. L. JEFFERIES. 2007. Vegetation loss alters soil nitrogen dynamics in an Arctic salt marsh. Journal Of Ecology 95: 283-293. BULL, I. D. AND OTHERS. 1999. Es timating the contribution of Spartina anglica biomass to salt-marsh sediments using compound specific stable carbon isotope measurements. Org. Geochem. 30: 477-483. BURDIGE, D. J., AND R. C. ZIMMERMAN. 2002. Impact of sea grass density on carbonate dissolution in Bahamian se diments. Limnol. Oceanogr. 47: 1751-1763. BUYANONOVSKY, G. A., M. ASLAM, AND G. H. WAGNER. 1994. Ca rbon turnover in soil physical fractions. Soil Sci. Soc. Am. J. 58: 1167-1173. CAHOON, D. R. 1994. Recent accretion in 2 managed marsh impoundments in coastal Louisiana. Ecol. Appl. 4: 166-176. ------, AND J. C. LYNCH. 1997. Vertical Accr etion and shallow subsidence in a mangrove forest of southwestern Florida, U.S.A. Mangroves and Salt Marshes 1: 173-186. ------, AND R. E. TURNER. 1989. Accretion and canal impacts in a rapidly subsiding wetland.2. Feldspar marker horizon technique. Estuaries 12: 260-268. CALLAWAY, J. C., R. D. DELA UNE, AND W. H. PATRICK. 1997. Sediment accretion rates from four coastal wetlands along the Gulf of Mexico. J. Coast. Res. 13: 181-191. CAMMEN, L. M. 1975. Accumulation rate and tur nover time of organic carbon in a salt-marsh sediment. Limnol. Oceanogr. 20: 1012-1014.

PAGE 159

159 CANUEL, E. A., K. H. FREEMAN, AND S. G. WAKEHAM. 1997. Isotopic compositions of lipid biomarker compounds in estuarine plan ts and surface sediments. Limnol. Oceanogr. 42: 1570-1583. CAPONE, D. G., AND R. P. KI ENE. 1988. Comparison of microbi al dynamics in marine and fresh-water sediments contrasts in anaer obic carbon catabolism. Limnol. Oceanogr. 33: 725-749. CEBRIAN, J. 2002. Variability a nd control of carbon consumption, export, and accumulation in marine communities. Limnol. Oceanogr. 47: 11-22. ------, M. F. PEDERSEN, K. D. KROEGER, AND I. VALIELA. 2000. Fate of production of the seagrass Cymodocea nodosa in different stages of meadow formation. Mar. Ecol.-Prog. Ser. 204: 119-130. CHAPIN, F. S., P. A. MATSON, AND H. A. MOONEY. 2002. Principles of Terrestrial Ecosystem Ecology. Springer Science. CHMURA, G. L., P. AHARON, R. A. SOCKI, AND R. ABERNETHY. 1987. An inventory of C-13 abundances in coastal wetlands of Loui siana, USA Vegetation And Sediments. Oecologia 74: 264-271. ------, S. C. ANISFELD, D. R. CAHOON, AND J. C. LYNCH. 2003. Global carbon sequestration in tidal, saline wetl and soils. Glob. Biogeochem. Cycle 17:1111. ------, L. L. HELMER, C. B. BEECHER, AND E. M. SUNDERLAND. 2001. Historical rates of salt marsh accretion on the outer Bay of Fundy. Canadian Journal Of Earth Sciences 38: 1081-1092. CHOI, Y. H., AND Y. WA NG. 2004. Dynamics of carbon sequestration in a coastal wetland using radiocarbon measurements Glob. Biogeochem. Cycle 18. CIFUENTES, L. A., R. B. COFFIN, L. SO LORZANO, W. CARDENAS, J. ESPINOZA, AND R. R. TWILLEY. 1996. Isotopic and elemental variations of carbon and nitrogen in a mangrove estuary. Estuar. Coast. Shelf Sci. 43: 781-800. CONNOR, R. F., G. L. CHMURA, AND C. B. BEECHER. 2001. Carbon accumulation in Bay of Fundy salt marshes: Implications for restoration of reclaimed marshes. Glob. Biogeochem. Cycle 15: 943-954. CORDOVA-KREYLOS, A. L. AND OTHERS. 2006. Diversity, composition, and geographical distribution of microbial communities in Ca lifornia salt marsh sediments. Applied And Environmental Microbiology 72: 3357-3366. CRAFT, C. 2001. Soil organic carbo n, nitrogen, and phosphorus as indicators of recovery in restored Spartina marshes. Ecological Restoration 19: 87-91.

PAGE 160

160 ------, S. BROOME, AND C. CAMPBELL. 2002. Fifteen years of vegetation and soil development after brackish-water marsh creation. Restor. Ecol. 10: 248-258. ------, AND OTHERS. 2003. The pace of ecosystem development of constructed Spartina alterniflora marshes. Ecol. Appl. 13: 1417-1432. ------, READER, J. N. SACCO, AND S. W. BR OOME. 1999. Twenty-five years of ecosystem development of constructed Spartina alterniflora (Loisel) marshes. Ecol. Appl. 9: 1405-1419. ------, E. D. SENECA, AND S. W. BROOME. 1993. Vertical accretion in microtidal regularly and irregularly flooded estuarine marshes. Estuar. Coast. Shelf Sci. 37: 371-386. DABY, D. 2003. Some quantitative aspects of seagrass ecology in a coas tal lagoon of Mauritius. Marine Biology 142: 193-203. DAHL, T. E. 1990. Wetland losses in the United States 1780's to 1980's. F. a. W. S. U. S. Department of the Interioir, Washington, D.C. [ed.]. DAI, J. H., M. Y. SUN, R. A. CULP, AND J. E. NOAKES. 2005. Changes in chemical and isotopic signatures of plant materials during degradation: Implication for assessing various organic inputs in estu arine systems. Geophysical Research Letters 32. DAUBY, P. 1989. The stable carbon isotope ratios in benthic food webs of the Gulf of Calvi, Corsica. Cont. Shelf Res. 9: 181-195. DITTMAR, T., R. J. LARA, AND G. KATTNER 2001. River or mangrove? Tracing major organic matter sources in tropical Brazilian coastal waters. Mar. Chem. 73: 253-271. DUARTE, C. M. 2002. The future of seagra ss meadows. Environmental Conservation 29: 192-206. ------, AND J. CEBRIAN. 1996. The fate of mari ne autotrophic producti on. Limnol. Oceanogr. 41: 1758-1766. ------, AND C. L. CHISCANO. 1999. Seagrass biom ass and production: a reassessment. Aquat. Bot. 65: 159-174. ------, J. J. MIDDELBURG, AND N. CARACO. 2005. Major role of marine vegetation on the oceanic carbon cycle. Biogeosciences 2: 1-8. EVANS, N. T., AND F. T. SHORT. 2005. Functi onal trajectory models for assessment of transplanted eelgrass, Zostera marina L., in the Great Bay Estuary, New Hampshire. Estuaries 28: 936-947. FENTON, G. E., AND D. A. RITZ. 1988. Changes in carbon and hydrogen stable isotope ratios of macroalgae and seagrass during decom position. Estuar. Coast. Shelf Sci. 26: 429-436.

PAGE 161

161 FISCHLER, K. C. 2006. Observations and Ch aracterizations of Subaqueous Soils and Seagrasses in a Recently Constructed Habita t in the Indian River Lagoon, Florida. University of Florida. FOURQUREAN, J. W., AND J. E. SCHRLAU. 2003. Changes in nutrient content and stable isotope ratios of C and N during decomposition of seagrasses and mangrove leaves along a nutrient availability gradient in Florida Bay, USA. Chemistry and Ecology 19: 373-390. ------, S. P. ESCORCIA, W. T. ANDERSON, AND J. C. ZIEMAN. 2005. Spatial and seasonal variability in elemental content, delta C-13, and delta N-15 of Thalassia testudinum from South Florida and its implications for ecosystem studies. Estuaries 28: 447-461. GACIA, E., C. M. DUARTE, AND J. J. MIDD ELBURG. 2002. Carbon a nd nutrient deposition in a Mediterranean seagrass ( Posidonia oceanica ) meadow. Limnol. Oceanogr. 47: 23-32. GOLDING, C. J., R. J. SMERNI K, AND G. F. BIRCH. 2004. Characterisation of sedimentary organic matter from three south-eastern Aust ralian estuaries using solid-state C-13-NMR techniques. Marine and Freshwater Research 55: 285-293. GONI, M. A., M. J. TEIXEIRA, AND D. W. PERKEY. 2003. Sources and distribution of organic matter in a river-dominated estuary (Winyah Bay, SC, USA). Estuar. Coast. Shelf Sci. 57: 1023-1048. GONNEEA, M. E., A. PAYTAN, AND J. A. HERRERA-SILVEIRA. 2004. Tracing organic matter sources and carbon burial in mangrove sediments over the past 160 years. Estuar. Coast. Shelf Sci. 61: 211-227. HAINES, E. B. 1976. Stable carbon isotope ratios in biota, soils and tidal water of a Georgia salt-marsh. Estuarine And Coastal Marine Science 4: 609-616. HARRIS, D., W. R. HORWATH, AND C. VAN KESSEL. 2001. Ac id fumigation of soils to remove carbonates prior to total organic car bon or carbon-13 isotopic analysis. Soil Sci. Soc. Am. J. 65: 1853-1856. HAVENS, K. J., L. M. VARNELL, AND B. D. WATTS. 2002. Matura tion of a constructed tidal marsh relative to two natural reference tidal marshes over 12 years. Ecol. Eng. 18: 305-315. HEDGES, J. I. 1992. Global biogeochemical cy cles Progress Aa problems. Mar. Chem. 39: 67-93. ------, AND J. H. STERN. 1984. Carbon And Nitr ogen Determinations Of Carbonate-Containing Solids. Limnol. Oceanogr. 29: 657-663. HEMMINGA, M. A., AND C. M. DUARTE 2000. Seagrass Ecology. Cambridge. ------, AND M. A. MATEO. 1996. Stable carbon isotope s in seagrasses: Variab ility in ratios and use in ecological studies. Mar. Ecol.-Prog. Ser. 140: 285-298.

PAGE 162

162 ------, F. J. SLIM, J. KAZUNGU, G. M. GANSSEN, J. NIEUWE NHUIZE, AND N. M. KRUYT. 1994. Carbon Outwelling From A Ma ngrove Forest With Adjacent Seagrass Beds And Coral-Reefs (Gazi Bay, Kenya). Mar. Ecol.-Prog. Ser. 106: 291-301. HERNANDEZ, M. E., R. MEAD, M. C. PE RALBA, AND R. JAFFE 2001. Origin and transport of n-alkane-2-ones in a subtropica l estuary: potential biomarkers for seagrassderived organic matter. Org. Geochem. 32: 21-32. HOPKINSON, C. S. 1988. Patterns of organic carbon exchange between coastal ecosystems: The mass balance approach in salt marsh ecosystems, p. 122-154. In B. O. JAnsson [ed.], Coastal-Offshore Ecosystem Interactions. Springer-Verlag. HUSSEIN, A. H., M. C. RABENHORST, A ND M. L. TUCKER. 2004. Modeling of carbon sequestration in coastal marsh soils. Soil Sci. Soc. Am. J. 68: 1786-1795. IPCC. 2001. Third assessment report--Climate Ch ange 2001. World Meteorlogical Organization and United Nations. JENNERJAHN, T. C., AND V. ITTEKKOT. 2002. Relevance of mangroves for the production and deposition of organic matter along tropical continental margins. Naturwissenschaften 89: 23-30. JOERGENSEN, R. G., AND T. MUELLER. 1995. Estim ation of the microbial biomass in tidal flat sediment by fumigation-extractio n. Helgolander Meeresuntersuchungen 49: 213-221. JOHNSON, R. W., AND J. A. CA LDER. 1973. Early diagenesis of fatty-acids and hydrocarbons in a salt-marsh environment. Geochim. Cosmochim. Acta 37: 1943-1955. KANG, C. K., E. J. CHOY, S. K. PAIK, H. J. PARK, K. S. LEE, AND S. AN. 2007. Contributions of primary organic matter sour ces to macroinvertebrate production in an intertidal salt marsh (Scirpus triqueter) ecosystem. Mar. Ecol.-Prog. Ser. 334: 131-143. KAYOMBO, S., T. S. A. MBWETTE, A. W. MAYO, J. H. Y. KATIMA, AND S. E. JORGENSEN. 2002. Diurnal cycles of variation of physical-che mical parameters in waste stabilization ponds. Ecological Engineering 18: 287-291. KENNEDY, H., E. GACIA, D. P. KENNEDY, S. PAPADIMITRIOU, AND C. M. DUARTE. 2004. Organic carbon sources to SE Asian coas tal sediments. Estuar. Coast. Shelf Sci. 60: 59-68. KENNISH, M. J. 2001. Coastal salt marsh syst ems in the US: A review of anthropogenic impacts. J. Coast. Res. 17: 731-748. ------. 2002. Environmental threats and environm ental future of estuaries. Environmental Conservation 29: 78-107.

PAGE 163

163 KENTULA, M. E., R. P. BROOKS, S. E. GWIN C. C. HOLLAND, A. D. SHERMAN, AND J. C. SIFNEOS. 1992. An Approach To D ecision Making In Wetla nd Restoration And Creation. Island Press. KRISTENSEN, E. 1994. Decomposition Of Macroalg ae, Vascular Plants And Sediment Detritus In Seawater Use Of Stepwise Thermogravimetry. Biogeochemistry 26: 1-24. LAL, R., J. KIMBLE, E. LEVINE, AND C. WHITMAN. 1995. World soils and greenhouse effect: An overview. In R. Lal, J. Kimble, E. Levine, a nd B. A. Stewart [ed.], Soils and Global Change. Advances in Soil Science. CRC Press. LALLIER-VERGES, E., B. P. PERRUSSEL, J. R. DISNAR, AND F. BALTZER. 1998. Relationships between environmental conditions and the diagenetic evolution of organic matter derived from higher plants in a m odern mangrove swamp system (Guadeloupe, French West Indies). Org. Geochem. 29: 1663-1686. LECKIE, S. E., C. E. PRESCOTT, S. J. GR AYSTON, J. D. NEUFEL D, AND W. W. MOHN. 2004. Comparison of chloroform fumigation-extraction, phospholipid fatty acid, and DNA methods to determine microbial biomass in forest humus. Soil Biol. Biochem. 36: 529-532. LEE, S., AND J. A. FUHRMAN. 1987. Relatio nships between biovolume and biomass of naturally derived marine bacterioplankton. Applied and Environmental Microbiology 53: 1298-1303. LIN, G. H., T. BANKS, AND L. STERNBERG. 1991. Variation in delta-13-C values for the seagrass Thalassia-testudinum and its relations to mangrove carbon. Aquatic Botany 40: 333-341. LINDAU, C. W., AND L. R. HOSS NER. 1981. Substrate characteri zation of an experimental marsh and 3 natural marshes. Soil Sci. Soc. Am. J. 45: 1171-1176. MACHAS, R., R. SANTOS, AND B. PETERSON. 2006. Elemental and stable isotope composition of Zostera noltii (Horneman) leaves during the early phases of decay in a temperate mesotidal lagoon. Estuar. Coast. Shelf Sci. 66: 21-29. MARCHAND, C., E. LALLIER-VERGES, AND F. BALTZER. 2003. The composition of sedimentary organic matter in relation to th e dynamic features of a mangrove-fringed coast in French Guiana. Estuar. Coast. Shelf Sci. 56: 119-130. MATEO, M. A., AND J. ROMERO. 1996. Evaluati ng seagrass leaf litter decomposition: An experimental comparison between litter-bag and oxygen-uptake methods. J. Exp. Mar. Biol. Ecol. 202: 97-106. MCKEE, K. L., AND P. L. FAULKNER. 2000. Re storation of biogeochemical function in mangrove forests. Restor. Ecol. 8: 247-259.

PAGE 164

164 MEAD, R., Y. P. XU, J. CHONG, AND R. JA FFE. 2005. Sediment and soil organic matter source assessment as revealed by the molecular distribution and carbon isotopic composition of n-alkanes. Org. Geochem. 36: 363-370. MEWS, M., M. ZIMMER, AND D. E. JELINSKI. 2006. Species-specific decomposition rates of beach-cast wrack in Barkley Sound, British Co lumbia, Canada. Mar. Ecol.-Prog. Ser. 328: 155-160. MIDDELBURG, J. J., J. NIEUWENHUIZE, R. K. LUBBERTS, AND O. VAN DE PLASSCHE. 1997. Organic carbon isotope systematics of coastal marshes. Estuar. Coast. Shelf Sci. 45: 681-687. MISEROCCHI, S., L. LANGONE AND T. TESI. 2007. Content and isotopic composition of organic carbon within a flood la yer in the Po River prodelta (Adriatic Sea). Cont. Shelf Res. 27: 338-358. MITSCH, W. J., AND J. G. GOSSELINK. 2000. Wetlands, 3rd ed. John Wiley and Sons. MOORE, T. N., AND P. G. FA IRWEATHER. 2006. Decay of multiple species of seagrass detritus is dominated by species identity, with an important influence of mixing litters. Oikos 114: 329-337. MORGAN, P. A., AND F. T. SHOR T. 2002. Using functional trajec tories to track constructed salt marsh development in the Great Bay Estu ary, Maine/New Hampshire, USA. Restor. Ecol. 10: 461-473. MORIARTY, D. J. W. AND OTHERS. 1985. Microb ial biomass and productivity in seagrass beds. Geomicrobiology Journal 4: 21-51. MORRIS, J. T., AND P. M. BRADLEY. 1999. Effect s of nutrient loading on the carbon balance of coastal wetland sediments. Limnol. Oceanogr. 44: 699-702. MOSER, M., C. PRENTICE, AND S. FRAZIER. 1996. A global overview of wetland loss and degradation. RAMSAR, 6th Meeting of the Conference of Contracting Parties. MOY, L. D., AND L. A. LEVIN. 1991. Are spartina marshes a replaceable resource a functional-approach to evaluation of marsh creation efforts. Estuaries 14: 1-16. MUZUKA, A. N. N., AND J. P. SHUNULA. 2006. Stable isot ope compositions of organic carbon and nitrogen of two mangrove stands along the Tanzanian coastal zone. Estuar. Coast. Shelf Sci. 66: 447-458. NIEUWENHUIZE, J., Y. E. M. MAAS, AND J. J. MIDDELBURG. 1994. Rapid analysis of organic-carbon and nitrogen in particulate materials. Mar. Chem. 45: 217-224. ONG, J. E. 1993. Mangroves--a carbon source and sink. Chemosphere 27: 1097-1107.

PAGE 165

165 OTERO, X. L., T. O. FERREIRA, P. VI DAL-TORRADO, AND F. MACIAS. 2006. Spatial variation in pore water geoche mistry in a mangrove system (Pai Matos island, CananeiaBrazil). Applied Geochemistry 21: 2171-2186. PAPADIMITRIOU, S., H. KENN EDY, D. P. KENNEDY, C. M. DUARTE, AND N. MARBA. 2005. Sources of organic matter in seagrass-colon ized sediments: A stable isotope study of the silt and clay fraction from Posidonia oceanica meadows in the western Mediterranean. Org. Geochem. 36: 949-961. RABENHORST, M. C. 1995. Carbon storage in tidal marsh soils, p. 93-104. In R. Lal, J. Kimble, E. Levine and B. A. Stewart [eds.] Soils and Global Change. Advances in Soil Science. CRC Lewis. RAMOS E SILVA, C. A., A. P. DA SILVA, AND S. R. DE OLIVEIRA. 2006. Concentration, stock and transport rate of heavy metals in a tropical red mangrove, Natal, Brazil. Mar. Chem. 99: 2-11. ROCHETTE, P., AND E. G. GREGORICH. 1998. Dynamics of soil microbial biomass C, soluble organic C and CO2 evolution after th ree years of manure application. Canadian Journal of Soil Science 78: 283-290. ROMERO, J., M. PEREZ, M. A. MATEO, AND E. SALA. 1994. The belowground organs of the Mediterranean seagrass Posidonia-oceanica as a biogeochemical sink. Aquatic Botany 47: 13-19. SCHLESINGER, W. H. 1990. Evidence from chr onosequence studies for a low carbon-storage potential of soils. Nature 348: 232-234. ------. 1997. Biogeochemistry: An Analysis of Global Change. Academic Press. SENECA, E. D. S. W. B., W. W. WOODHO USE, L. M. CAMMEN, J. T. LYON. 1976. Establishing Spartina alterniflora marsh in North Carolina. Environmental Conservation 3: 185-188. SFWMD. 2005-2006. Environmental Database (D BHYDRO). South Florida Water Management District. URL: www.sfwmd.gov/org/ema/dbhydro/. Date accessed (July 2007). SHORT, F. T., E. W. KOCH, J. C. CREED K. M. MAGALHAES, E. FERNANDEZ, AND J. L. GAECKLE. 2006. SeagrassNet monitoring across the Americas: case studies of seagrass decline. Marine Ecol ogy-an Evolutionary Perspective 27: 277-289. SIGUA, G. C., AND W. A. TW EEDALE. 2003. Watershed scale assessment of nitrogen and phosphorus loadings in the Indian River Lagoon basin, Florida. Journal of Environmental Management 67: 363-372. SIMENSTAD, C. A., AND R. M. THOM. 1996. Functional equivalency trajectories of the restored Gog-Le-Hi-Te estuarine wetland. Ecol. Appl. 6: 38-56.

PAGE 166

166 SMIT, A. J., A. BREARLEY, G. A. HYNDES, P. S. LAVERY AND D. I. WALKER. 2005. Carbon and nitrogen stable isotope analysis of an Amphibolis griffithii seagrass bed. Estuar. Coast. Shelf Sci. 65: 545-556. SMITH, B. D., AND R. E. FO REMAN. 1984. An assessment of seaweed decomposition within a southern Strait of Georgia s eaweed community. Marine Biology 84: 197-205. SOTO-JIMENEZ, M. F., F. PAEZ-OSUNA, A ND A. C. RUIZ-FERNANDEZ. 2003. Organic matter and nutrients in an altered subtropi cal marsh system, Chiricahueto, NW Mexico. Environ. Geol. 43: 913-921. STEVENSON, F. J. 1994. Humus Chemistry: Genesis, Composition, Reactions, 2nd ed. John Wiley and Sons, Inc. SUZUKI, Y., M. FUJII, B. E. CASARETO, A. FURUTA, A ND Y. ISHIKAWA. 2003. CO2 sequestration and fate of organic matters within seagrass ( Zostera marina) ecosystem. J. Chem. Eng. Jpn. 36: 417-427. TEMMERMAN, S., G. GOVERS, S. WARTEL, AND P. MEIRE. 2003. Spatial and temporal factors controlling short-term sedimentation in a salt and freshwater tidal marsh, Scheldt estuary, Belgium, SW Netherlands. Ea rth Surface Proce sses And Landforms 28: 739-755. TERRADOS, J. AND OTHERS. 1999. Are seagrass growth and survival constrained by the reducing conditions of the sediment? Aquatic Botany 65: 175-197. THIMDEE, W., G. DEEIN, C. SANGRUNGRUANG, J. NISHIOKA, AND K. MATSUNAGA. 2003. Sources and fate of organic matter in Khung Krabaen Bay (Thailand) as traced by delta C-13 and C/N atomic ratios. Wetlands 23: 729-738. TWILLEY, R. R., R. H. CHEN AND T. HARGIS. 1992. Carbon sinks in mangroves and their implications to carbon budget of tropical coastal ecosystems. Water Air Soil Pollut. 64: 265-288. ------, G. EJDUNG, P. ROMARE, AND W. M. KEMP. 1986. A comparative-study of decomposition, oxygen-consumption and nutrien t release for select ed aquatic plants occurring in an estuarine environment. Oikos 47: 190-198. VALERY, L., V. BOUCHARD, AND J. C. LEFEUV RE. 2004. Impact of the invasive native species Elymus athericus on carbon pools in a salt marsh. Wetlands 24: 268-276. VALIELA, I., J. L. BOWEN, AND J. K. YORK. 2001. Mangrove forests: One of the world's threatened major tropical environments. Bioscience 51: 807-815. VANCE, E. D., P. C. BROOKES, AND D. S. JENKINSON. 1987. An extraction method for measuring soil microbial biomass-C. Soil Biology & Biochemistry 19: 703-707. VICHKOVITTEN, T., AND M. HOLMER. 2005. Disso lved and particulate organic matter in contrasting Zostera marina (eelgrass) sediments. J. Exp. Mar. Biol. Ecol. 316: 183-201.

PAGE 167

167 VIZZINI, S., G. SARA, M. A. MATEO, AND A. MAZZOLA. 2003. delta C-13 and delta N-15 variability in Posidonia oceanica associated with seasonality and plant fraction. Aquatic Botany 76: 195-202. WANG, X. C., R. F. CHEN, AND A. BERRY. 2 003. Sources and preservation of organic matter in Plum Island salt marsh sediments (MA, USA): Long-chain n-alkane s and stable carbon isotope compositions. Estuar. Coast. Shelf Sci. 58: 917-928. WU, J., R. G. JOERGENSEN, B. POMMERE NING, R. CHAUSSOD, AND P. C. BROOKES. 1990. Measurement of soil microbial biomass c by fumigation extraction an automated procedure. Soil Biology & Biochemistry 22: 1167-1169. ZEDLER, J. B. 2004. Compensating for wetla nd losses in the United States. Ibis 146: 92-100. ------, AND J. C. CALLAWAY. 1999. Tracking wetland restoration: Do mitigation sites follow desired trajectories? Restor. Ecol. 7: 69-73. ------, AND S. KERCHER. 2005. Wetland resources: Status, trends, ecosystem services, and restorability. Annual Review Of Environment And Resources 30: 39-74. ZHOU, J., Y. WU, J. ZHANG, Q. KANG, AND Z. LIU. 2006. Carbon and nitrogen composition and stable isotope as potential indicators of source and fate of organic matter in the salt marsh of the Changjiang Estuary, China. Chemosphere 65: 310-317. ZHU, Z. B., R. C. ALLE R, AND J. MAK. 2002. Stable carbon isotope cycling in mobile coastal muds of Amapa, Brazil. Cont. Shelf Res. 22: 2065-2079. ZIEMAN, J. C., J. W. FOURQUREAN, AND T. A. FRANKOVICH. 1999. Seagrass die-off in Florida Bay: Long-term trends in abu ndance and growth of turtle grass, Thalassia testudinum Estuaries 22: 460-470.

PAGE 168

168 BIOGRAPHICAL SKETCH Caitlin Hicks was born in Fram ingham, Massachusetts and grew up in Western Massachusetts. She spent her summer vacations on Cape Cod where she learned to love nature among the Capes beaches, salt marshes, and fo rests. She attended Middlebury College in Vermont where she enjoyed views of the Green Mountains and Adirondacks every day. She pursued a joint major in biology and environmen tal studies, graduating in May 2004. During college, she spent a challenging but immensely rewa rding semester at the Ecosystem Center of the Woods Hole Marine Biol ogy Laboratory where she came to appreciate the cycling of elements within ecosystems. Upon graduation, Ca itlin worked as a res earch assistant for Dr. Steward Pickett at the Institute of Ecosystem Studi es in Millbrook, New York. While at that job, she ran a laboratory, worked on GIS projects, an d enjoyed arduous field work in Kruger National Park, South Africa. In August 2005, seeking a ch ange from the northeast climate, biota, and culture she had grown up with, Caitlin moved to Gainesville, Flor ida to begin her masters work with Dr. Reddy in the University of Florida s Soil and Water Science Department. For her masters work, Caitlin was given the freedom to pick her own project, so she chose to work in a coastal environment, like the ones she loved as a child, and on carbon, her favorite element. Caitlin is now beginning her PhD in the Botany De partment at the University of Florida working with Dr. Ted Schuur. She has switched field site s but not elements and will now study carbon in Alaska.







SOC can be part of three pools--microbial biomass C, labile OC, or recalcitrant OC. Generally,

microbes consume mainly labile OC, which they respire as CO2 Or incorporate into their

biomass. When microbes die, their OC becomes part of labile and recalcitrant pools. The more

OC is reworked by microbes, the more recalcitrant it becomes. The recalcitrant pool is where

OC is sequestered long-term and where OC undergoes abiotic condensation into complex humic

materials. In mangrove forests, the C cycle is basically the same except for the C sources (Fig.

1-2). The inorganic C source used by mangroves is atmospheric CO2, the litter is mangrove

leaves, and imported OC trapped in litter by mangrove roots is seston and seagrass detritus.

The capacity of coastal ecosystems to sequester OC is greater than the capacity of

terrestrial ecosystems. Coastal ecosystems are natural C sinks, while terrestrial systems reach an

equilibrium where the net C fixed annually is about zero (Rabenhorst 1995). Constant

accumulation of C in coastal ecosystems is due to their anoxic sediments. In these sediments,

oxygen is depleted so electron acceptors that are not as efficient must be utilized by microbes to

decompose OC. Coastal ecosystems also have a greater OC sequestration capacity than

freshwater wetlands because they, unlike freshwater wetlands, do not use CO2 aS a terminal

electron acceptor and therefore emit less CH4 (Bridgham et al. 2006). In coastal ecosystems,

sulfate is the terminal electron acceptor, and high sulfate levels inhibit methanogenesis (Capone

and Kiene 1988). A study of mangrove forests did not detect CH4 either dissolved in sediment

porewaters or fluxing out of sediments and 51 to 75% of OM oxidation was occurring through

sulfate reduction (Alongi et al. 2004). Coastal ecosystems, like salt marshes, mangrove forests,

and seagrass beds, may therefore be highly significant C sinks because they accumulate C in

sediments without emitting CH4.









Discussion


Sediment Characteristics

SL 15 and reference sediments are physically different from one another because SL 15

sediments' parent material is dredge spoil, as is apparent from their high amount of shell

fragments (Table 1). Furthermore SL 15 sediments were compacted during construction. SL

15's accreted layer differs from other SL 15 sediments because it is a layer of post-construction

deposition and was not compacted by equipment. In comparisons of constructed and reference

salt marshes and mangrove forests, bulk density was almost always greater in constructed sites

(Craft et al. 1999; McKee and Faulkner 2000; Craft et al. 2002). Redox potentials of all sites

were negative implying anaerobic conditions and a slow rate of decomposition. Redox potentials

in this study are generally more negative than those found in other mangrove (McKee 1993,

Otero et al. 2006) and seagrass sediments (Terrardos et al. 1999), and sediment pH in this study

are generally higher than in other mangrove sediments (McKee and Faulkner 2000; Otero et al.

2006) but similar to other seagrass sediments (Burdige and Zimmerman 2002; Daby 2003).

C:N ratios only differed between mangrove constructed and reference sediments (Table

4). Lower C:N ratios in the mangrove SL 15 sediments are due to their very low TOC. The rest

of the C:N ratios are the same between SL 15 and reference sites due to similar proportions of C

and N despite SL 15 having lower amounts of C and N overall. In this study, C:N ratios could

therefore not be used as the ultimate metric of restoration success as was suggested for salt

marshes by Craft (2001).

Trajectory of Constructed Site

In SL 15 sediments, only mangrove bulk density followed a functional trajectory in which

it decreased within 2 months of construction completion but remained higher than the reference

values (Fig. 3-3). This initial decrease may have occurred as these sediments decompressed,









eventually reach an equilibrium where amount of C Eixed equals the amount respired annually, if

disturbances like fire do not cause a loss of C first (Rabenhorst 1995). Constant accumulation of

C in wet ecosystems is due to their anaerobic sediments where alternate electron acceptors,

which are not as efficient as oxygen, must be utilized to decompose C. The capacity of coastal

ecosystems to sequester C is also greater than that of freshwater wetlands (Bridgham et al. 2006).

Bridgham et al. (2006) found that estuarine wetlands sequestered C 10 times faster on an areal

basis than other wetlands. These high rates are due to estuarine wetlands' high sedimentation

rates, high percent soil C, and burial due to sea level rise (Connor et al. 2001; Bridgham et al.

2006). Coastal ecosystems have another advantage over freshwater wetlands. They have lower

rates of methanogenesis, so the C they store is not being converted to CH4, a mOre potent

greenhouse gas than CO2. In the United States, freshwater mineral wetlands emit 2.4 Tg CH4 y-1

while salt marshes and mangroves emit only 0.027 and 0.004 Tg CH4 y-1, TOSpectively (Bridgham

et al. 2006).

When these coastal ecosystems are impacted, a portion of the biosphere' s C storage and

sequestration capacity is lost, which may exacerbate climate change by causing more CO2 to be

in the atmosphere than would be if these systems were intact. Loss of vegetated coastal

ecosystems has caused at least a 25% decrease in their global C sequestration capacity (Duarte et

al. 2005). Bridgham et al. (2006) estimated that losses of salt marshes and mangroves in the

conterminous United States have caused a net flux of 402 Tg C y^l into the atmosphere.

The upside is that restoration and construction of coastal systems may help mitigate the

effects of climate change by increasing C sequestration. For example, if all dyked salt marshes

in Canada were restored, an additional 2.4 to 3.6 x 1011 g C y-l would be sequestered, which

would contribute 5% to Canada's CO2 emiSsions reduction identified in the Kyoto Protocol










weighed by species into 2-3 g (only 1 g for A. spicifera) allotments and placed intact into 13 cm

x 13 cm litter bags of nylon mesh with 0.5 x 0.25 mm holes.

This litter bag study is a "common garden" study where we investigated the relative

decomposition rates of the potential sources to SOC, so all litter bags were placed in the same

area of the SL 15 mangrove system. On September 8, 2006, litter bags were placed on the

sediment surface, pinned down with metal stakes, and overlaid with large wire mesh to prevent

them from washing away. Three litter bags from each plant species were randomly collected at

2, 4, 8, 16, and 32 weeks. Sediment and algae were rinsed from the litter bags in the laboratory

before the bags were air dried for several weeks. Once dry, the bags were opened, and plant

material in each bag was weighed.

Source Sampling

Plants were sampled on SL 15 in January, July, October, and November 2006 and at

reference sites in July and November 2006. Sampled plants included all potential sources to

SOC found in and around SL 15 and reference sites and fell into 3 main groups: Subtidal, which

include seagrasses (S. fihiforme, T. testudinum, H. beaudettei, Halophila johnsonii) and

macroal gae epiphytess on seagrasses, Acanthophora spicifera, Caulerpa sertulariodes,

Salrga~ssum spp., Ulva spp., Chaetomorpha linum, Rosenviga intricate, Hypnea cervicornis,

Gracilaria tikvahiae, and Enteromorpha spp.); Intertidal, which included mangroves (Avicennia

germinans, Rhizophora mangle, Laguncularia racemosa), Sueda linearis, and Spartina

alterniflora; and Terrestrial (Schinus terebenthifolia~s, Casuarina equisetifolia, Coccoloba

uvifera, and Triplasis purpurea). Not all plants were collected at all sampling dates because

some plants, particularly species of macroalgae were not present throughout the year. Vascular

plant samples were a composite of 3-5 live, healthy leaves or fronds from greater than three

individuals collected across the sampling area (i.e. SL 15 or reference sites). Macroalgae











looo r-


Concentration (%)


Storage (g m )


S0-5 cm
11 5-10 cm


800


600-


400-


200-


1.5 t


1.0 t


0.5 t


&


500
1
2
3 400


300


0.9 t-


ABC
BC


0.6 t


SL 15


0.3 t


Reference


Reference SL 15


Figure 3-6. Comparisons between total organic carbon (TOC) in reference and SL 15 mangrove
(top) and seagrass (bottom) sediments. The bars are mean TOC averaged over month
(July and November 2006) for each depth of sediment (n=4 for SL 15 and n=3 for
reference). Error bars are + SE. Depths in the seagrass systems are as follows: 1= SL
15 accreted and reference 0-5, 2= SL 15 0-5 and reference 5-10, 3= SL 15 5-10 and
reference 10-15. An asterisk indicates a significant site effect (Table 3-5). Capital
letters are results of a Tukey test performed after a significant site x depth interaction,
and lowercase letters are results of a Tukey performed after an unsignifieant site x
depth interaction, but a significant one-way ANOVA. Bars that share letters are not
significantly different.












Constructed and Reference Comparisons .....__.....___ ..........._ ............7
Organic Carbon Accumulation Rates .....__.....___ ..........._ ............7
Discussion .............. ...._ ...............79...
Sediment Characteristics .............. ...............79....

Traj ectory of Constructed Site. ........._.__............ ...............79...
Constructed and Reference Equivalence ......__....._.__._ ............. ............8
Organic carbon pools .............. ...............81....
Organic carbon liability ................. ...............86..._._ ._......
Organic carbon accumulation............... ..............8
Conclusion ........._.__....... .__ ...............88....


4 SOURCES OF SEDIMENT ORGANIC CARBON IN A CONSTRUCTED
MANGROVE AND SEAGRAS S SY STEM .............. ...............107....


Introducti on ................. ...............107....... ......
M ethod s .............. ...............109....

Study Site............... ...............109.
Litter Bags ................. ...............110........ ......
Source Sampling............... ...............11
Sediment Sampling ........_................. ...............112......
Laboratory Analyses ........_................. ..........._..........11
Data Analyses ........_................. ..........._..........11
Re sults..........._...._ ........ ...............116.....
Source Characteristics ..........._..._ ...............116......._ ......
Sediments and Surface Layers ....._._ ................. ...............117 ....
Source Determination ........_................. ..........._..........11
Discussion ............... .. ......._._ ...............119......
Source Characteristics ..........._..._ ...............119......._ ......
Sediments and Surface Layers ....._._ ................. ...............122 ....
Source Determination ........_................. ..........._..........12
Conclusion ........_................. ........_._ .........12


5 SY NT HESIS ......._ ................. ...............146......


Obj ective One: Short Term Traj ectories of Sediment Organic Carbon Pools. ................... ..147
Obj ective Two: Comparisons of Sediment Organic Carbon Pools ................. ................. 147
Obj ective Three: Comparisons of Sediment Organic Carbon Lability ............... .... ........._..148
Obj ective Four: Comparisons of Sediment Organic Carbon Sources ................. ...............149
Conclusion ................. ...............149.............



APPENDIX STATISTICAL TRANSFORMATIONS .............. ...............155....


LIST OF REFERENCES ............. ............ ...............157...


BIOGRAPHICAL SKETCH ................. ...............168.............









Each organic structure, such as an aromatic ring or a carboxyl group, has a different resonance

subsequent chemical shift; therefore this method allows scientists to assign categories of OM to

specific chemical shifts (Golding et al. 2004). Using this method, types and relative amounts of

OM structures in sediments can be elucidated.

Golding et al. (2004) used 13C-NMR to study whether SOM was terrestrial- or marine-

derived in upper (fluvial) and lower (marine) sections of Australian estuaries. They studied four

groups of organic carbon structures-carbonyls, aromatics, O-alkyls, and alkyls. They

associated both O-alkyl C and aromatic C with terrestrial plant sources because they assumed O-

alkyl C was from cellulosic carbohydrates and aromatic C was from lignin and tannins of

vascular plants. The presence of alkyl C indicated marine origins because they associated it with

planktonic material. They cautioned, however, that alkyl C may also be present due to microbial

decomposition of terrestrial OM. The authors concluded that upper portions of estuaries had

higher proportions of O-alkyl C and aromatic C, and therefore higher amounts of terrestrially-

derived SOC, than lower portions of estuaries.

NMR has similar problems as petrographic analysis because structures being studied

cannot be directly assigned to specific primary producers; the relationship between the producer

and the structure must be inferred, and one structure type can come from several producers. This

technique may be best suited to situations where sources are grouped into a couple of

components such as a study of seagrass/mangrove-derived SOC and planktonic SOC. Despite

problems associated with SOC source determination using 13C-NMR, this tool may help

scientists better elucidate roles of plankton and algae, whose proportions in SOC are difficult to

determine via stable isotopes because of their variable and intermediate 813C ValUeS. This method

also helps scientists understand the OC structures, not just the OC sources in coastal sediments.









Short 2002; Craft et al. 2003). Functional trajectory studies, with one exception (Evans and

Short 2005), have been limited to temperate brackish and salt water marshes. There is a need to

study functional trajectories in constructed mangrove forests and seagrass beds. Functional

traj ectory studies generally measure a suite of ecological functions, so SOC is usually the only

variable measured that pertains to ecosystem C storage. Studies that examine multiple OC pools,

OC liability, and OC sources are needed to more fully understand the recovery of C storage

functioning in constructed systems. Studies that examine short term changes immediately

following construction are also lacking. Short term traj ectory studies are important because

certain aspects of OC storage may recover quickly.

Whether constructed mangrove and seagrass ecosystems provide the same ecological

services as their natural counterparts with respect to C storage, and whether restoration of these

services follow a functional traj ectory is currently unknown. In this thesis, the traj ectories of

SOC pools in constructed seagrass and mangrove systems were monitored during the first year

after construction completion. SOC pools in the constructed systems were also compared with

SOC pools in adjacent natural systems. Sediments were the focus of this research because they

are the sites of long term C storage. Variables measured include the amount of OC in three pools

(total OC, extractable OC, and microbial biomass C), the liability of SOC, and the C to nitrogen

ratios and 813C Of sediments and potential SOC sources. The constructed site was a former spoil

island called SL 15 in the Indian River Lagoon, FL that was converted to a mangrove and

seagrass ecosystem in November 2005. The reference sites were the natural seagrass beds that

surround SL 15 and the nearby mangrove forests that occupy the edges of adj acent spoil islands.

The main research obj ectives were: 1) to determine short term traj ectories of SOC pools in

a constructed mangrove forest and seagrass bed; 2) to compare SOC pools in the constructed









et al. 2003) and 42 (Craft 2001) years old. These authors concluded that it takes a long time for

restored salt marshes to develop SOC pool equivalence and acknowledged that such equivalence

may never be reached.

Predictions from salt marsh studies may be valuable for understanding traj ectories of

constructed mangrove forests because both are intertidal systems that take a long time to reach

equivalence. Sediment OM in a 6-year and 14-year-old mangrove forests in Southwest Florida

remained at 18 to 32% of reference forest values (McKee and Faulkner 2000). SL 15 mangrove

sediment TOC was well below that of references. The lack of a TOC trajectory for mangrove

sediments contrasts to findings in salt marsh studies and indicate that not reaching equivalence is

a possibility. In all salt marsh studies except one (Simenstad and Thom 1996) a trajectory of

increasing OC/OM was documented (Zedler and Calloway 1999; Craft 2001; Craft et al. 2002;

Havens et al. 2002; Morgan and Short 2002; Craft et al. 2003). Even a young constructed salt

marsh in North Carolina increased its sediment TOC by over 100% in 1.3 years (Cammen 1975).

Predictions from salt marshes studies do not work for constructed and restored seagrass

beds. OM content of restored sediments was higher than one reference and lower than another

throughout the first 8 years in the only other known study of seagrass functional traj ectories (on

the New Hampshire coast, Evans and Short 2005). In SL 15 seagrass sediments, TOC was

functionally equivalent in 2 out of 3 depths within a year.

There are several reasons why OC in seagrass sediments reach functional equivalence

before OC in mangrove forests and in salt marshes. The first reason is elevation. In several

studies of restored and constructed salt marshes, soil development was correlated to marsh

elevation so that OC/OM was higher at lower elevations (Lindau and Hossner 1981; Moy and

Levin 1991; Craft et al. 2002). OC equivalence occurs faster at lower elevations because they are











Table 4-3. Results of ANOVAs comparing 613C ValUeS in SL 15 and reference mangrove and
seagrass sediments and surface layers (right) and of repeated measures ANOVAs of
813C ValUeS in SL 15 sediments and surface layers (left).
ANOVA Effect 813C ANOVA Effect 813


Repeated measures
Mangrove 0-10


Mangrove algal mat
Seagrass 0-10


Seagrass accreted
Seagrass floc


Sediment comparison
Mangrove





Seagrass






Surface Comparison
Mangrove algal
mat/litter


Seagrass floc


Month
Depth
Month*Depth
Month
Month
Depth
Month*Depth
Month
Month


Site
Month
Depth
Site*Month
Site*"Depth
Month*Depth
Site
Month
Depth
Site*Month
Site*"Depth
Month*Depth

Site
Month
Site*Month

Site
Month
Site*Month


For significance NS=not significant, p <0.05, **p < 0.01, ***p < 0.0001. Please see Appendix A for a table
listing how these data were transformed prior to running the 3-way ANOVA.









4010, Costech) in line with an isotope ratio mass spectrometer (ThermoFinnigan MAT Delta

Plus XL, Thermo Scientifie, Waltham MA) for %OC and 813C. Plants were analyzed for %TN

simultaneously. Peach leaves (NIST 1547) were used for EA calibration, with sucrose and an

internal soil standard used as check standards. Sucrose and Peach leaves were used as internal

standards for mass spectrometry measurements. C isotopes were reported in per mil notation

based on deviations from the Pee Dee Dolomite standard. Tests were run on sand samples with

various carbonate percentages and total weights to assess the efficacy of the vapor acidiaication

method and determine the maximum sample mass that still ensured complete removal of IC.

Furthermore, 613C ValUeS were used to confirm complete removal oflIC because the presence of

carbonate greatly raised 813C ValUeS. If incomplete IC removal was suspected, samples were

rerun at a lower total mass. The 613C Of filter blanks were accounted for in the calculation of

seston 613C. Unacidified sediment, surface layer, and seston samples were run separately in tin

capsules (Costech) on an elemental analyzer (Flash EA 1112 Series, Thermo Scientific,

Waltham, MA) for TN. Acetinilide was used for calibration standards, while peach leaves

(NIST 1547) and an internal soil standard were used for quality control.

Data Analyses

Individual plant 813C and C:N were averaged across sites and sampling dates. Values of

certain species were also averaged together into plant groups of seagrass, macroalgae,

mangroves, or C3 terrestrial. Differences in 613C between sampling date and tide phase (ebb or

flood) were tested on seston samples using one-way analyses of variance (ANOVAs) in JMP

Version 6. For all sources, C:N ratios are reported, even though N:C ratios are used in graphs, so

data can be easily compared across studies. Litter mass loss for each species was modeled using

a first-order exponential decay curve.









signatures in a Mexican marsh. Average 613C Value Of sediment was -20.4 %o and they assumed,

based on previously published 813C Signatures of sources in temperate estuaries, that dominant

SOC sources were plankton and macrophytes. While 613C ValUeS of putative sources at each site

need to be measured at least once per site, Fourqurean et al. (2005) proposed further

determination of source 613C ValUeS seasonally and annually.

Lipid Biomarker Compounds

The use of specific organic compounds, called biomarkers, to identify SOC sources in

coastal ecosystems is becoming more common. These compounds are generally lipids including

sterols, fatty acids, and hydrocarbons. The ways these organic compounds are used vary because

the compounds vary in their specificity--some can identify groups of organisms such as vascular

plants or algae while others may be specific to one genera or species (Canuel et al. 1997). Less

specific biomarkers can be used in conjunction with stable isotopes to further differentiate

sources from general groups (i.e. vascular plants into C3 and C4 grOups). Many studies used

biomarkers in concert with bulk stable isotopes (Hernandez et al. 2001; Wang et al. 2003) or

measured the stable isotopic composition of biomarker compounds in sources and sediment (

Canuel et al. 1997; Bull et al. 1999; Hernandez et al. 2001; Mead et al. 2005).

This method uses a gas chromatography (GC) to determine lipids after a complex and

lipid-type specific extraction process. The different lipids are separated by the GC column due

to their different retention time within the column. Different lipids can be identified by

comparing their relative retention times on the resulting gas chromatogram with relative

retention times on a gas chromatogram of a known standard. Relative amounts of each lipid can

also be determined by calculating areas underneath each peak on the chromatogram--larger

areas correspond to a larger amount of that lipid in the sample. Isotopic signatures of these lipids

can be determined when the GC is connected to a mass spectrometer in a method known as









CHAPTER 1
INTTRODUCTION

Restoration and construction of coastal ecosystems may help mitigate the effects of climate

change by reducing atmospheric carbon dioxide (CO2). Global climate change has become a

maj or environmental concern over the past 50 years. The anthropogenic release of greenhouse

gases is the major cause of global climate change (IPCC 2001). Atmospheric concentrations of

CO2 and methane (CH4), the biggest contributors to climate change, are increased by fossil fuel

burning and deforestation, and livestock production, respectively. Highly productive coastal

ecosystems including salt marshes, seagrass beds, and mangrove forests are carbon (C) sinks.

Salt marshes and mangroves store at least 44.6 Pg C in their sediments (Chmura et al. 2003).

Seagrass beds, which make up only 0. 15% of global marine area, account for 15% of global

marine organic C (OC) storage (Hemminga and Duarte 2000). Sediment C storage values are

even larger when stores of inorganic C like carbonates are taken into account (Zhu et al. 2002).

The C cycle in coastal ecosystems is an open cycle because OC is imported into and

exported out of systems by water currents and tides. In a seagrass bed, seagrasses and

macroalgae (drift and epiphytic) take up CO2 and HCO3~ fTOm the water column to produce OC

through photosynthesis (Fig. 1-1). When seagrass fronds senesce or break, they (and their

associated macroalgae) become litterfall on top of sediments or are exported out of the system.

Litterfall OC is either decomposed by microbes or incorporated into the sediment OC (SOC)

pool by leaching, bioturbation, or burial. Imported OC, which may include terrestrial and

mangrove detritus, is trapped by seagrass fronds and settles on the sediment. Seston, comprised

of plankton, bacteria, and dissolved and particulate OC from in and outside the system, is also

trapped by seagrass fronds. The fate of trapped OC is the same as litterfall. Seagrass root OC

from exudates or dead tissue is immediately part of the SOC pool and can be used by microbes.









CHAPTER 4
SOURCES OF SEDIMENT ORGANIC CARBON INT A CONSTRUCTED MANGROVE AND
SEAGRASS SYSTEM

Introduction

Sediments can accumulate organic carbon (OC) from in situ vegetation, drift macroalgae,

plankton, and water column terrestrial- and marine-derived detritus. Understanding sources of

OC in soils and sediments is important to our understanding of local and global C cycles (Hedges

1992). The source of OC influences the quality and stability of OC in sediments. OC sources,

like temperature and oxygen availability, affect decomposition rates (Chapin et al. 2002), which

in turn affect OC sequestration. Certain ecosystems, like macrophyte-dominated coastal

systems, accumulate and store large amounts of OC in their sediments. These salt marshes,

mangrove forests, and seagrass beds are sinks for CO2 and therefore mitigate climate change by

keeping C out of the atmosphere. Worldwide, salt marshes and mangroves store at least 44.6 Pg

C in their sediments (Chmura et al. 2003), equivalent to 2% of the global soil C pool (Lal et al.

1995). Seagrass beds, which make up only 0. 15% of global marine area, account for 15% of the

global marine OC storage (Hemminga and Duarte 2000). Determining the vegetation that are the

main OC sources to coastal sediments helps researchers predict how changing environmental

conditions may affect the future of these significant C stores.

Coastal ecosystems are experiencing great losses worldwide (Valiela et al. 2001; Alongi

2002; Green and Short, 2003). The loss of vegetated coastal ecosystems has caused at least a

25% decrease in their global C sequestration capacity (Duarte et al. 2005). Constructing coastal

ecosystems may restore a portion of the lost C sink (Connor et al. 2001). Knowing OC sources

of constructed coastal systems can indicate whether these constructed systems can become

effective at storing OC. For example, a constructed mangrove system whose principle

sedimentary OC (SOC) source is relatively labile macroalgae will not store as much C for as










depths across different sites and not different depths across different sites (i. e.: it compares SL 15

mangrove 0-5 cm to reference 0-5 cm but not SL 15 mangrove 0-5 cm to reference 5-10 cm), so

one-way ANOVAs were also run when site~depth interactions of the factorial ANOVAs did not

reveal all interesting trends. Data were averaged by each site and depth over July and November

samplings for these one-way ANOVAs. Factorial and one-way ANOVAs were run on JMP

Version 6 (SAS Institute, Cary, NC).

For all analyses most data were transformed to meet the normality requirement (see

Appendix A for details). Post hoc multiple comparisons were carried out on significant effects

using the Tukey test. Significance was decided using an alpha level of 0.05.

Results

Sediment Characteristics

SL 15 sediments (0-5 cm and 5-10 cm) had higher bulk densities than reference

sediments (Table 1; site effect, p<0.0001, Table 2) as did the SL 15 mangrove algal mat. The

seagrass accreted layer had a bulk density similar to the 0-5 cm depth of the seagrass reference.

In seagrass sediments, bulk densities were greatest in the lowest depths, but in mangrove

sediments were greater in 0-5 cm depths (Table 1; depth effect, p<0.026, Table 2). SL 15

seagrass sediments had orders of magnitude more shell fragments than reference sediments,

while SL 15 and reference mangrove sediments had similar amounts of shell fragments (Table

1). pH in SL-15 seagrass and reference sediments and in SL-15 mangrove sediments ranged

from 8.0 8.3. Reference mangrove sediments had a pH of 7.5 (Table 1). Redox potentials in

the upper sediment depths were similar between SL-15 and reference sites, but were more

negative in the lower depths of the reference sediments (Table 1).









Inputs to SOC from individual species of C3 plant were unknown because their signatures were

not distinct.

Middelburg et al. (1997) avoided problems associated with intermediate and indistinct

values by using 613C iSotopes for the sole purpose of determining the amount of Spartina-derived

SOC in salt marshes in Massachusetts (Great Marsh) and the Netherlands (Waarde Marsh). In

Great Marsh, high marsh SOC 813C value (-13.4 to -14.5 %o) was similar to the Spartina value (-

12.5 %o), but low marsh SOC 813C value (-21 to -19.5 %o) was not. In Waarde Marsh SOC was

9-12 %o less than the Spartina value (-12.7 %o). They hypothesized that depletions of SOC

values in Waarde marsh and the low marsh of Great Marsh were due to the input of allocthonous

OM such as marine plankton and non-local macrophytes, but since they did not measure these

sources they could not definitively identify which contributed to the depletion. They concluded

Great Marsh was a peaty marsh where C accumulation was due to Spartina inputs and Waarde

marsh was a mineral marsh where accumulation was due to sedimentation.

Problems with intermediate values were encountered by all previously discussed studies

that tried to comprehensively measure 613C ValUeS of all maj or sources. These studies often

used either a variation of the mixing equation developed by Dauby (1989) or a simple

comparison of sources and sediment isotope values. However, other ways to calculate source

contributions may partially eliminate problems with intermediate values. Gonneea et al. (2004)

used a ternary mixing diagram to elucidate relative source contributions of seagrass, mangroves,

and seston to SOC. With a ternary mixing diagram, all sources were end members as they

formed a triangle on a graph of C:N ratios plotted against 813C ValUeS. Sediment C:N and 813C

values were also plotted on this diagram, which had a 10% tolerance interval to account for

natural variability in source values. Sediment samples that fell in the middle of the triangle were










constructed seagrass system is functioning more like its reference than the constructed mangrove

system. In the long term, however, the potential C sink capacity of the constructed mangrove

system is greater.









most likely coming the water column and benthic vegetation, which in the first year did not

include deeply rooting plants. The 5-10 cm depth, however, was not completely dredge spoil. It

contained mangrove clay from pre-construction mangrove areas and a buried "A horizon" from

the seagrass bed that occupied the site before spoil island creation (Fischler 2006). These other

sediments were exposed and mixed with dredge spoil during construction and had more OC than

dredge spoil due to their origins in vegetated systems.

In the surface layers, the seagrass floc reached or exceeded equivalence in terms of all

OC pools. SL 15 floc may have exceeded reference values due to its position inside the

mangrove fringe of SL 15. In the subtidal portion of SL 15 there were areas of slower tidal flow

that caused settling of water column material (Fischler 2006), which would include OC. The

algal mat reached equivalence in ExOC and MBC but not TOC. Lower TOC in the algal mat

than in the litter layer is because the litter layer consisted of higher plant material like mangrove

leaves and seagrass that contain more recalcitrant C than algae (Kristensen 1994). Surface layers

are first to receive inputs that contribute directly and indirectly to OC pools, such as of light,

water column nutrients, and detritus. Therefore, it is not surprising that most of their OC

parameters would reach equivalence within the first year. Upper depths reached OC functional

equivalence quickly while lower depths failed to increase over 7 years in a constructed Virginia

salt marsh (Havens et al. 2002).

The maj ority of studies that test functional traj ectories of TOC or organic matter (OM) in

restored and created salt marshes do not see OC reach functional equivalence. In studies that

ranged from one- to 42-year-old marshes, only two reached equivalence with their natural

wetland references in terms of SOC (Simenstad and Thom 1996, Zedler and Calloway 1999;

Craft 2001, Havens et al. 2002, Morgan and Short 2002, Craft et al. 2003). They were 25 (Craft










In equation 3-1, OCfis the final amount of TOC (g OC m-2) in the top 0-10 cm, OC, is the initial

amount of TOC in the top 0-10 cm, OC, is the amount of TOC in the accreting layer, and Asuste;;

is the age of the system in years. Without dating sediments using 137CS, 210Pb, or 14C prOflles,

OC accumulation rates in reference systems could not be calculated.

Statistical Analyses

Repeated measures analysis of variance (ANOVAs) were run to investigate if parameters

in SL 15 sediments and surface layers followed a functional trajectory over time. ANOVAs

were run with a spatial power covariance structure to account for the unequal spacing between

time points. Subjects were SL 15 plots and the repeated factor was month. The 0-5 and 5-10 cm

depths were run together in each system in ANOVAs with depth as a main effect. Floc, algal

mat, and accreted layers were run separately in ANOVAs. Replicate cores had to be averaged

for each plot and month so the data fit the structure required for repeated measures analysis. A

parameter followed a traj ectory if its repeated measures ANOVA had a significant time effect

and it demonstrated an increasing or decreasing (in the case of bulk density) trend over time.

Analyses were run using the mixed procedure in SAS Version 8 (SAS Institute, Cary, NC).

Comparisons between SL 15 and reference sites were analyzed using one factorial

ANOVA each for the mangrove and seagrass sediments and one factorial ANOVA each for the

mangrove and seagrass surface layers. Sediment ANOVAs consisted of three fixed factors--

site, month, and depth. Surface layer ANOVAs consisted of site and month factors. All two

way interactions were tested. Plot data were pooled into two site treatments, SL 15 and

reference. July and November 2006 were the months. For seagrass sediment analysis, SL 15

and reference depths were assigned to 3 categories in order to make comparisons: SL 15 accreted

and reference 0-5 cm were depth 1, SL 15 5-10 cm and reference 0-5 cm were depth 2, and SL

15 5-10 cm and reference 10-15 cm were depth 3. Factorial ANOVAs only compare the same










------, N. KOEDAM, A. V. RAMAN, AND F. DEHAIRS. 2002. Primary producers sustaining
macro-invertebrate communities in intertidal mangrove forests. Oecologia 130: 441-448.

------, T. MOENS, I. OVERMEER, N. KOEDAM, AND F. DEHAIRS. 2004. Resource
utilization patterns of epifauna from mangrove forests with contrasting inputs of local
versus imported organic matter. Mar. Ecol.-Prog. Ser. 278: 77-88.

BOURGUES, S., I. AUBY, R. DEWIT, AND P. J. LABOURG. 1996. Differential anaerobic
decomposition of seagrass (Zostera noltii) and macroalgal (M~onostroma obscurum)
biomass from Arcachon Bay (France). Hydrobiologia 329: 121-131.

BRIDGHAM, S. D., J. P. MEGONIGAL, J. K. KELLER, N. B. BLISS, AND C. TRETTIN.
2006. The carbon balance of North American wetlands. Wetlands 26: 889-916.

BROOME, S. W., E. D. SENECA, AND W. W. WOODHOUSE. 1988. Tidal salt-marsh
restoration. Aquatic Botany 32: 1-22.

BRUNSKILL, G. J., I. ZAGORSKIS, AND J. PFITZNER. 2002. Carbon burial rates in
sediments and a carbon mass balance for the Herbert River region of the Great Barrier
Reef continental shelf, North Queensland, Australia. Estuar. Coast. Shelf Sci. 54: 677-700.

BUCKERIDGE, K. M., AND R. L. JEFFERIES. 2007. Vegetation loss alters soil nitrogen
dynamics in an Arctic salt marsh. Journal Of Ecology 95: 283-293.

BULL, I. D. AND OTHERS. 1999. Estimating the contribution of Spartina anglica biomass to
salt-marsh sediments using compound specific stable carbon isotope measurements. Org.
Geochem. 30: 477-483.

BURDIGE, D. J., AND R. C. ZIMMERMAN. 2002. Impact of sea grass density on carbonate
dissolution in Bahamian sediments. Limnol. Oceanogr. 47: 1751-1763.

BUYANONOVSKY, G. A., M. ASLAM, AND G. H. WAGNER. 1994. Carbon turnover in soil
physical fractions. Soil Sci. Soc. Am. J. 58: 1167-1173.

CAHOON, D. R. 1994. Recent accretion in 2 managed marsh impoundments in coastal
Louisiana. Ecol. Appl. 4: 166-176.

------, AND J. C. LYNCH. 1997. Vertical Accretion and shallow subsidence in a mangrove
forest of southwestern Florida, U.S.A. Mangroves and Salt Marshes 1: 173-186.

------, AND R. E. TURNER. 1989. Accretion and canal impacts in a rapidly subsiding wetland.2.
Feldspar marker horizon technique. Estuaries 12: 260-268.

CALLAWAY, J. C., R. D. DELAUNE, AND W. H. PATRICK. 1997. Sediment accretion rates
from four coastal wetlands along the Gulf of Mexico. J. Coast. Res. 13: 181-191.

CAMMEN, L. M. 1975. Accumulation rate and turnover time of organic carbon in a salt-marsh
sediment. Limnol. Oceanogr. 20: 1012-1014.









Seston and macroalgae were the dominant OC sources in SL 15 seagrass sediments

according to the ternary diagram (Fig. 4-6a). Seagrasses, which had colonized most of SL 15 at

generally low densities by July 2006 (Fischler 2007), were not yet important SOC sources.

Seston and macroalgae as main SOC sources were further supported by observations--drift

macroalgae was frequently found buried in the accreted layer section of cores throughout the

study where it seemed to trap particles from the water, driving accretion. High N:C (low C:N)

ratios of seagrass reference sediments interfered with determining sources via the ternary

diagram (Fig. 4-6b). Samples outside of the diagram can indicate an unknown source of SOC,

but that is unlikely as almost all plants encountered were measured and none had high N:C (low

C:N) ratios (Table 4-1). According to 813C Only, seston and seagrass are probably both sources

because reference seagrass SOC 813C falls in the middle of those end members. The contribution

of macroalgae is unknown though due to its intermediate 813C. Mangroves were not chosen as a

potential source for seagrass sediments as mangrove litter was observed infrequently on seagrass

sediments. Therefore mangrove's influence to SOC was believed to be mediated through seston.

Sources to the reference mangrove litter layer change with season as the ternary diagram

indicates that seagrass and seston are the dominant sources in July but mangroves are the

dominant sources in November (Fig. 4-7a). Conclusions from the ternary diagram match field

observations. The litter layer was primarily seagrass wrack in July but was primarily partially-

decomposed mangrove leaves in November. Since the litter layer is one of the main sources to

reference mangrove SOC, seagrass and mangroves are therefore also sources to mangrove SOC

through the litter. Sources to seagrass floc varied seasonally for references but not SL 15.

Seston was a dominant floc OC source for all seasons and sites according to the ternary mixing









BIOGRAPHICAL SKETCH

Caitlin Hicks was born in Framingham, Massachusetts and grew up in Western

Massachusetts. She spent her summer vacations on Cape Cod where she learned to love nature

among the Cape's beaches, salt marshes, and forests. She attended Middlebury College in

Vermont where she enjoyed views of the Green Mountains and Adirondacks every day. She

pursued a j oint maj or in biology and environmental studies, graduating in May 2004. During

college, she spent a challenging but immensely rewarding semester at the Ecosystem Center of

the Woods Hole Marine Biology Laboratory where she came to appreciate the cycling of

elements within ecosystems. Upon graduation, Caitlin worked as a research assistant for Dr.

Steward Pickett at the Institute of Ecosystem Studies in Millbrook, New York. While at that j ob,

she ran a laboratory, worked on GIS proj ects, and enjoyed arduous field work in Kruger National

Park, South Africa. In August 2005, seeking a change from the northeast climate, biota, and

culture she had grown up with, Caitlin moved to Gainesville, Florida to begin her masters work

with Dr. Reddy in the University of Florida' s Soil and Water Science Department. For her

masters work, Caitlin was given the freedom to pick her own proj ect, so she chose to work in a

coastal environment, like the ones she loved as a child, and on carbon, her favorite element.

Caitlin is now beginning her PhD in the Botany Department at the University of Florida working

with Dr. Ted Schuur. She has switched field sites but not elements and will now study carbon in

Alaska.










Trajectory of Constructed System

Parameters measured in SL 15 sediments did not follow a traj ectory over time, except for

mangrove sediment bulk density, which significantly decreased with time (month effect,

p<0.0001, Table 3, Fig. 3-3). OC parameter values seemed to shift randomly when there were

significant monthly changes as for MBC in all sediments, and ExOC and TOC in seagrass 0-10

cm sediments (month effect, p<0.021, Table 3, Fig. 3-4). OC parameters followed a pattern in

seagrass sediments in which low values occurred in February and July while high values

occurred in May and November (Fig. 3-4). TN and C:N also changed without direction when

they did change significantly (month effect, p<0.041, Table 3). There were no significant

changes in liability for either mangrove or seagrass sediments. Significant differences between

depths were few. In mangrove sediments, 0-5 cm depths had greater bulk density and TN, and in

all sediments, 0-5 cm depths had greater liability (depth effect, p<0.031, Table 3).

SL 15 surface layers (mangrove algal mat and seagrass floc) followed a traj ectory of

significantly increasing MBC (p<0.0051, Table 3, Fig. 3-5). Extractable OC, TOC, and TN

significantly changed with time in floc, with TOC and TN generally increasing (p<0.050, Table

3, Fig. 3-5). C:N significantly changed without a trend in floc (p<0.043, Table 3). Lability of

OC in the mangrove algal mat significantly increased with time, while liability in seagrass floc

significantly decreased with time (p<0.052, Table 2).

Constructed and Reference Comparisons

TOC was significantly higher in reference than in SL 15 mangrove and seagrass systems

on both a concentration and storage basis, except in seagrass floc where TOC was similar

between sites (site effect, p<0.0005, Table 2; Table 4). TOC differences between sites were

greatest in mangrove sediments (Fig. 3-6). On a concentration basis in seagrass sediments, sites

had similar TOC in depth one, but had different TOC in depths two and three (site x depth










Table 4-2. Decay constants (+ SE) and turnover times calculated from a nonlinear regression
(exponential decay) of litter bag experiment data.
k Tumover time
Species (day ') (days)
Haludule beaudettei 0.0049 (0.0006) 203
Thalassia testudinum 0.0099 (0.001) 101
Syringodium filiforme 0.046 (0.006) 22
Acanthophora spicifera 0.070 (0.006) 14
Sargassum spp. 0.019 (0.001) 53
Avicennia germinans 0.0093 (0.0004) 109
Rhizophora mangle 0.0047 (0.0004) 213







































Figure 1-2. The carbon cycle in mangrove forests.





























O 2007 Caitlin E Hicks











Table 2-2. Global rates of carbon accumulation in coastal ecosystem sediments.
Global areal rate Global rate
System Average type (g C m yr ) (Tg C yr ) Data sources
Mangrove forests Mean' 92 20 1
Mean 100 22 2
Estimated 200 44 3
Mode' 115 25 4

Salt marshes Mean 50-5000 17.5-1750 5
Mean 100 35 1
Mean 175 61 2
Mode 115 25 4

Intertidal3 Mean 210 120 6

Seagrass beds Estimate4 1.2 0.54 7
Mean 133 60 2
Estimate 270 122 8, 9
Mode 36.5 16.5 4

Open ocean Mean 0.22 170 2

Terrestrial systems
Tundra 0.2-2.4 10
Temperate forest 0.7-10 10
Tropical rainforest 2.3 10
Temperate grassland 2.2 10
Temperate desert 0.8 10
al. Twilley et al. 1992: 2, Duarte and Cebrian 1996: 3, Jennerjahn and Ittekkot 2002; 4, Cebrian 2002; 5, Rabenhorst
1995: 6, Chmura et al. 2003; 7, Suzuki et al. 2003; 8, Duarte and Chiscano 1999; 9, Hemminga and Duarte 2000; 10,
Schlesinger 1990.
1Both the mean and mode numbers were derived from compiling numbers from published studies. 2The estimates
were either scaled up from a single study or derived from a rough "back-of-the-envelope" calculation. 3Number
includes contribution of both mangrove forests and salt marshes. 4Estimate is of amount being exported and
subsequently buried in the open ocean sediments, not in situ accumulation. 5These numbers represent long term
accumulation rates measured since the end of the last ice age.










9). Negative rates in mangrove sediments were due to a decrease bulk density throughout the

year while TOC concentrations remained constant, but if the algal mat becomes more permanent

(i.e. buried) its OC will more than compensate for negative rates. Positive rates in seagrass

sediments were driven by the accreting layer. It is unknown whether the accreted layer in

seagrass sediments of SL 15 will continue to accumulate material at the same rates as in the first

year. Continued accumulation depends on how much the accreted layer formation was due to a

physical response to an uneven benthic surface after construction and how much was due to

macroalgae and seagrasses trapping particles from the water column.

Conclusion

Mangrove sediments are farther from being equivalent C stores than seagrass sediments.

Mangrove sediments have only begun to reach equivalence in active pools (ExOC and MBC)

and contain a relatively small amount of TOC, while seagrass sediments have equivalent TOC at

most depths (Fig. 3-9). The difference between constructed and reference OC liability is also

much greater in mangrove than in seagrass sediments, and OC accumulation rates in mangrove

sediments are negative (if the algal mat is excluded). However, if constructed mangrove

sediments do begin to follow a functional traj ectory, their potential OC storage is greater than

constructed seagrass sediments because mangrove reference sediments have larger TOC pools,

less OC stored as MBC (Fig. 3-9), and lower OC liability than seagrass reference sediments.

Overall, due to potential OC limitations, low TOC values for their ecosystem type, and nitrogen

eutrophication (Morris and Bradley 1999; Sigua and Tweedle 2003) IRL coastal ecosystems are

probably not as effective at storing C as their counterparts elsewhere.

The C storage capabilities of coastal ecosystems make them a great contender for use as C

offsets. One year is not enough time to discern whether these systems will become significant C

stores. More studies should investigate constructed coastal ecosystems as potential C sinks by










RI = log(Xconstucte /X~,erecen) (5-1)

In Equation 5-1, RI is the recovery index, Xconstructed is the value of the parameter in the

constructed system, and Xeference is the value of the parameter in the reference system. RI's

equal to zero indicate equivalence, less than zero indicate that equivalence has not been reached,

and greater than zero indicate equivalence has been surpassed. The constructed seagrass system

is closer to equivalence than the constructed mangrove system (Fig.5-3). Upper depths of

constructed seagrass sediments had similar SOC pools and liability to upper depths of reference

seagrass sediments, causing the seagrass sediments to be closer to equivalence. As the

mangroves and seagrasses within the constructed systems mature, it is likely that their SOC will

become less algae-derived, leading to lower OC liability and better OC storage. Dominance of

seston as a source in all systems means that a switch in less significant sources may take time to

register in the sediments. Ultimately, if the OC pools and liability in reference systems are any

indication (Fig. 5-1 and 5-2) and if functional traj ectories are followed in the future, the

constructed mangrove system will become more effective at OC storage than the constructed

seagrass system.

This study adds to the body of research on functional traj ectories. It is one of two known

studies on seagrass trajectories and the only known study of mangrove trajectories. More

importantly, it is the first known study to examine the traj ectory of OC storage in depth. If

constructed and restored coastal ecosystems store and accumulate OC as well as their established

counterparts, corporations and governments could construct coastal systems in exchange for C

credits. This action can replace lost systems and restore many of the ecologically important

functions these systems provide. Conclusions based on this study are limited because it only

followed constructed systems during the first year of recovery. The constructed mangrove and










------, F. J. SLIM, J. KAZUNGU, G. M. GANSSEN, J. NIEUWENHUIZE, AND N. M.
KRUYT. 1994. Carbon Outwelling From A Mangrove Forest With Adj acent Seagrass
Beds And Coral-Reefs (Gazi Bay, Kenya). Mar. Ecol.-Prog. Ser. 106: 291-301.

HERNANDEZ, M. E., R. MEAD, M. C. PERALBA, AND R. JAFFE. 2001. Origin and
transport of n-alkane-2-ones in a subtropical estuary: potential biomarkers for seagrass-
derived organic matter. Org. Geochem. 32: 21-32.

HOPKINSON, C. S. 1988. Patterns of organic carbon exchange between coastal ecosystems:
The mass balance approach in salt marsh ecosystems, p. 122-154. In B. O. JAnsson [ed.],
Coastal-Offshore Ecosystem Interactions. Springer-Verlag.

HUSSEIN, A. H., M. C. RABENHORST, AND M. L. TUCKER. 2004. Modeling of carbon
sequestration in coastal marsh soils. Soil Sci. Soc. Am. J. 68: 1786-1795.

IPCC. 2001. Third assessment report--Climate Change 2001. World Meteorlogical Organization
and United Nations.

JENNERJAHN, T. C., AND V. ITTEKKOT. 2002. Relevance of mangroves for the production
and deposition of organic matter along tropical continental margins. Naturwissenschaften
89: 23-30.

JOERGENSEN, R. G., AND T. MUELLER. 1995. Estimation of the microbial biomass in tidal
flat sediment by fumigation-extraction. Helgolander Meeresuntersuchungen 49: 213-221.

JOHNSON, R. W., AND J. A. CALDER. 1973. Early diagenesis of fatty-acids and hydrocarbons
in a salt-marsh environment. Geochim. Cosmochim. Acta 37: 1943-1955.

KANG, C. K., E. J. CHOY, S. K. PAIK, H. J. PARK, K. S. LEE, AND S. AN. 2007.
Contributions of primary organic matter sources to macroinvertebrate production in an
intertidal salt marsh (Scirpus triqueter) ecosystem. Mar. Ecol.-Prog. Ser. 334: 131-143.

KAYOMBO, S., T. S. A. MBWETTE, A. W. MAYO, J. H. Y. KATIMA, AND S. E.
JORGENSEN. 2002. Diurnal cycles of variation of physical-chemical parameters in waste
stabilization ponds. Ecological Engineering 18: 287-291.

KENNEDY, H., E. GACIA, D. P. KENNEDY, S. PAPADIMITRIOU, AND C. M. DUARTE.
2004. Organic carbon sources to SE Asian coastal sediments. Estuar. Coast. Shelf Sci. 60:
59-68.

KENNISH, M. J. 2001. Coastal salt marsh systems in the US: A review of anthropogenic
impacts. J. Coast. Res. 17: 731-748.

------. 2002. Environmental threats and environmental future of estuaries. Environmental
Conservation 29: 78-107.









Monitoring Constructed Coastal Marshes Using Functional Trajectories

Functional traj ectories are used to track the progress of constructed ecosystems and to

compare constructed and natural ecosystems (Simenstad and Thom 1996; Zedler and Callaway

1999; Morgan and Short 2002). Functional trajectory studies often examine a whole suite of

"ecological attributes" (Craft et al. 2003) that act as indicators for more complex ecological

functions (Simenstad and Thom 1996). Attributes are measured in the same constructed system

over time, or in several different-aged constructed systems in the same region using a space-for-

time substitution, to obtain a range of attribute values that can be plotted against time (Kentula

1992). In coastal marshes, OC parameters are often just several of many attributes measured.

Data are then fitted to a curve and compared with values from natural marshes. The resulting

traj ectory represents how the attribute develops in a restored or constructed marsh over time

(Morgan and Short 2002). There are two main questions that functional trajectories studies seek

to answer: 1) how long does it take for the attribute in the restored or constructed marsh to reach

functional equivalence (i.e. the mean value of that attribute in a natural marsh); 2) is the mean

value of an attribute in a natural marsh the correct endpoint for the development of that attribute

in the restored or constructed marsh?

Not all attributes have the same traj ectory, and traj ectories of the same attribute may

differ across different marshes and depending on the natural reference marsh used. There are

many different traj ectories that attributes like SOC pools can follow (Kentula et al. 1993; Fig. 3).

Some attributes may not even follow a trajectory and instead stay relatively constant through

time (Zedler and Calloway 1999). Craft et al. (2003) proposed that different attributes follow

one of three traj ectories depending on whether they are part of hydrologic, biological, or soil

development processes. OC pool formation is part of soil development, which in most cases is

the slowest traj ectory to reach functional equivalence (Craft et al. 2003). If a traj ectory fits OC









surface SOC may be mineralized by the time it is buried deeper in the soil profile, where it

would be measured if longer term methods were used. Long term rates calculated by 14C dating

were slower than rates measured over a decadal (Choi and Wang 2004) or an 100 year time scale

(Hussein et al. 2004). Choi and Wang (2004) did not attribute this difference to methodology

and speculated that greater C accumulation rates are due to increases in primary production over

the last 100 years caused by increased CO2 and nutrients.

Comparing Organic Carbon in Restored and Reference Coastal Marshes

Highly productive habitats like coastal ecosystems are C sinks as their high C

accumulation rates demonstrate. In these coastal ecosystems atmospheric CO2 becomes stored

as OC for long periods of time. Restoration and construction of coastal ecosystems may

therefore help mitigate the effects of climate change by reducing atmospheric CO2 (COnnor et al.

2001). If limits are placed on CO2 emiSsions in the United States, coastal ecosystem restoration

and construction may then become a viable option for C offsetting. C offsetting occurs when an

industry needs to reduce its net CO2 emiSsions but can or will not reduce their own emissions, so

they invest in projects that reduce emissions elsewhere, such as tree planting. Anthropogenic

release of greenhouse gases is the maj or cause of global climate change (IPCC 2001). CO2 and

CH4 are greenhouse gases with the biggest effect on climate change due to their concentrations

in the atmosphere and radiative forcing capacity (IPCC 2001). Humans increase atmospheric

concentrations of CO2 through fossil fuel burning and land use change and concentrations of CH4

through livestock production. This section of the review focuses on salt marshes due to the

dearth of literature on functional traj ectories in restored or constructed mangrove and seagrass

sy stem s.

Despite the numerous important ecological functions coastal ecosystems and wetlands

provide, which extend well beyond their function as C sinks, many were viewed as wastelands










4-6 N:C vs. 813C in ternary mixing diagrams of three potential OC sources and seagrass
sediments ........._ ...... .___ ...............139....

4-7 N:C vs. 813C in ternary mixing diagrams of three potential OC sources and surface
layers ........._ ...... ...............141....

4-8 A theoretical diagram of organic carbon sources that may constitute seston and how
they affect seston 613C and C:N............... ...............143..

4-9 Main sources and how they affect surface layer and sediment 813C and C:N .................1 44

5-1 A modified seagrass bed carbon cycle showing values from this study in constructed
(C) and reference (R) systems............... ...............152

5-2 A modified mangrove forest carbon cycle showing values from this study in
constructed (C) and reference (R) systems ................ ...............153..............

5-3 Recovery indices of three organic carbon (OC) pools and OC liability parameters for
constructed mangrove and seagrass systems. ............. ...............154....









values corresponded to lower C:N ratios (Soto-Jimenez et al. 2003). In the Mexican marsh lower

C:N ratios were thought to be indicative of marine producers, specifically plankton. A brief

review of sediment 813C and C:N values from mangrove literature also showed an inverse

relationship between the two variables (Bouillon et al. 2003). Similar trends were found when

comparing sediment 813C and %SOC values in mangroves (Bouillon et al. 2003) and salt

marshes (Middelburg et al. 1997). More depleted 813C ValUeS corresponded with higher %SOC

in mangroves and more enriched 813C ValUeS corresponded with higher %SOC in salt marshes.

Generally, the higher the sediment %OC values, the closer the sediment 813C ValUeS are to the

613C ValUeS of the dominant vegetation (Bouillon et al. 2003).

There are other complications with the use of stable isotopes for OC source determination.

Problems not already discussed include inherent variation of isotopic signatures within different

tissues of a single individual (Papadimitriou et al. 2005) and within a single species (Hemminga

and Mateo 1996) across sites (Kennedy et al. 2004), seasons, and years (Anderson and

Fourqurean, 2003; Fourqurean et al. 2005). These variations are most pronounced in seagrasses

(Thimdee et al. 2003). Such variation may be due to changes in relative uses of dissolved CO2

and HCO3~ (Sources of inorganic C in water) (Lin et al. 1991), and changes in irradiance,

photosynthesis rates, and temperature (Hemminga and Mateo 1996). Kennedy et al. (2004)

found that isotopic signatures of sources such as seagrasses (613C = -5.8 to -13.3 %o) and seston

(613C = -9.6 to -22.9 %o) varied greatly among 15 different sites in the South China Sea. The

order trend of potential sources' 613C Signatures (seagrass > epiphyte > seston > mangrove)

remained constant, however. Variation by location means that conclusions based on

measurement of SOC 813C ValUeS without measuring potential sources may not be valid. Soto-

Jimenez et al. (2003) inappropriately used isotopes when they only measured sediments









until recently (Broome et al. 1988). These systems were seen as wasted space that could be

utilized for agriculture or valuable development. Wetlands, including salt marshes, were

summarily destroyed without much thought to the consequences of their destruction through the

1980's. The lower 48 U.S. states lost 53% of its wetlands from the 1870's to the 1980's (Dahl

1990). Globally, it is estimated that 50% of the wetlands have been lost (Moser 1996). When

these systems are lost, we lose a sink for anthropogenically-derived CO2. For example, Connor

et al. (2001) estimated that if 85% of the coastal marshes in the Bay of Fundy had not been

altered for agricultural uses, 3.8 x 10613 g C COuld have been stored over the past 160 years. The

loss of coastal ecosystems and wetlands therefore disrupts the global C cycle and may increase

the affects of climate change.

Since 1989, the U.S. has had a policy of no net wetland loss that includes coastal marshes

(Zedler 2004). The policy calls for mitigation if alternatives to destroying wetlands in the course

of development proj ects are unavailable. This mitigation comes in the form of creating new

wetlands onsite or nearby to the lost wetland, restoring an existing, degraded wetland, or buying

into wetland mitigation banks (Zedler 2004). Because of coastal marshes' importance as C sinks

and the widespread replacement of natural marshes with created marshes, it is important to know

whether restored and constructed marshes have OC accumulation rates and storage capacities

equivalent to those of natural marshes. Such research can indicate whether marsh creation can

become a policy tool for reducing CO2 emiSsions. Connor et al. (2001) suggested that restoring

coastal marshes may help countries reduce their CO2 emiSsions to the standards set by the Kyoto

protocol. Monitoring functional trajectories of constructed marshes helps researchers understand

if constructed marshes' OC storage can equal the storage of natural marshes.










SMIT, A. J., A. BREARLEY, G. A. HYNDES, P. S. LAVERY, AND D. I. WALKER. 2005.
Carbon and nitrogen stable isotope analysis of an Amphibolis griffithii seagrass bed.
Estuar. Coast. Shelf Sci. 65: 545-556.

SMITH, B. D., AND R. E. FOREMAN. 1984. An assessment of seaweed decomposition within
a southern Strait of Georgia seaweed community. Marine Biology 84: 197-205.

SOTO-JIMENEZ, M. F., F. PAEZ-OSUNA, AND A. C. RUIZ-FERNANDEZ. 2003. Organic
matter and nutrients in an altered subtropical marsh system, Chiricahueto, NW Mexico.
Environ. Geol. 43: 913-921.

STEVENSON, F. J. 1994. Humus Chemistry: Genesis, Composition, Reactions, 2nd ed. John
Wiley and Sons, Inc.

SUZUKI, Y., M. FUJII, B. E. CASARETO, A. FURUTA, AND Y. ISHIKAWA. 2003. CO2
sequestration and fate of organic matters within seagrass (Zostera marina) ecosystem. J.
Chem. Eng. Jpn. 36: 417-427.

TEMMERMAN, S., G. GOVERS, S. WARTEL, AND P. MEIRE. 2003. Spatial and temporal
factors controlling short-term sedimentation in a salt and freshwater tidal marsh, Scheldt
estuary, Belgium, SW Netherlands. Earth Surface Processes And Landforms 28: 739-755.

TERRADOS, J. AND OTHERS. 1999. Are seagrass growth and survival constrained by the
reducing conditions of the sediment? Aquatic Botany 65: 175-197.

THIMDEE, W., G. DEEIN, C. SANGRUNGRUANG, J. NISHIOKA, AND K. MATSUNAGA.
2003. Sources and fate of organic matter in Khung Krabaen Bay (Thailand) as traced by
delta C-13 and C/N atomic ratios. Wetlands 23: 729-738.

TWILLEY, R. R., R. H. CHEN, AND T. HARGIS. 1992. Carbon sinks in mangroves and their
implications to carbon budget of tropical coastal ecosystems. Water Air Soil Pollut. 64:
265-288.

------, G. EJDUNG, P. ROMARE, AND W. M. KEMP. 1986. A comparative-study of
decomposition, oxygen-consumption and nutrient release for selected aquatic plants
occurring in an estuarine environment. Oikos 47: 190-198.

VALERY, L., V. BOUCHARD, AND J. C. LEFEUVRE. 2004. Impact of the invasive native
species Elymus athericus on carbon pools in a salt marsh. Wetlands 24: 268-276.

VALIELA, I., J. L. BOWEN, AND J. K. YORK. 2001. Mangrove forests: One of the world's
threatened major tropical environments. Bioscience 51: 807-815.

VANCE, E. D., P. C. BROOKES, AND D. S. JENKINSON. 1987. An extraction method for
measuring soil microbial biomass-C. Soil Biology & Biochemistry 19: 703-707.

VICHKOVITTEN, T., AND M. HOLMER. 2005. Dissolved and particulate organic matter in
contrasting Zostera marina (eelgrass) sediments. J. Exp. Mar. Biol. Ecol. 316: 183-201.









0.24

0.22

0.20

0.18

0.16


Seagrass Floc


July'



-A


0.14 -

0.12 -


t _


0.10-

0.08-

0.06 -

0.04-

0.02-
-30


_~_____.....


-25


-20


-15


-10


813C


Figure 4-7. Continued









Conclusion

This review sought to cover SOC topics relevant to this thesis research in three types of

coastal ecosystems--salt marshes, mangrove forests, and seagrass beds. The original research in

this thesis covers OC pools and sources in a constructed mangrove and seagrass system. This

review covered salt marshes in order to add both depth and breadth because C accumulation and

constructed ecosystem development are currently better understood in salt marshes than in

mangrove forests and seagrass beds. This review discussed C accumulation rates of salt

marshes, mangrove forests, and seagrass beds, functional traj ectories of OC attributes in restored

and constructed salt marshes, and SOC source determination methods in salt marshes, mangrove

forests, and seagrass beds. The C accumulation section showed that these coastal ecosystems are

globally significant C sinks. The functional trajectory section showed how OC functions in

constructed salt marshes and emphasized the need for further and more in-depth studies of OC in

constructed coastal ecosystems. The SOC source determination section showed pros and cons of

different SOC determination methods, including bulk stable isotopes, which are utilized in the

original thesis research.










seagrass system in this study accumulated SOC at rates similar to rates in mature systems over

the first year. If these rates are sustained and more OC is stored in long-term, recalcitrant pools,

then the constructed systems will be effective C sinks. Longer term studies are needed to fully

assess the effectiveness of constructing coastal ecosystems for OC storage.









data well, it can be used to predict future levels of OC thereby helping agencies set standards for

mitigation proj ect monitoring or determine the amount of C emission credits a created marsh is

worth. In theory, functional trajectories are a simple way to evaluate the current success and

predict the future success of constructed marshes; data, however, do not always fit a smooth line.

Often, there is high variability between constructed marshes (Craft et al. 2003) and between

years in the same study (Zedler and Calloway 1999). The reference marshes used can also

influence the predicted success or failure of a constructed marsh as a result of their age,

variability (Simenstad and Thom 1996), or stress level. Furthermore, predictions from functional

traj ectories should be considered with caution because they do not take into account disturbances

that may alter the traj ectory.

Functional Trajectory Case Studies

The studies reviewed here examine the equivalence in constructed and restored marshes

that are one (Morgan and Short 2002) to 42 years old (Craft 2001). The first prediction of

functional equivalence for SOC in a salt marsh was made by Seneca et al. (1976) for one of the

first salt marsh creation proj ects using dredge spoil, which is essentially devoid of OC. They

predicted it would take 4 to 25 years for the constructed marsh to store as much C as the natural

marsh. More recent studies show that it probably takes at least 25 years for OC to reach

functional equivalence (Table 2-4). SOC and the related attribute, sediment organic matter

(SOM), seem to be one of the last attributes to reach functional equivalence in marshes after

aboveground biomass, sedimentation rates, and diversity of flora and fauna. There is also the

possibility that OC will never reach functional equivalence as most studies did not follow

marshes for a sufficient duration of time to show equivalence.

Most studies on the eastern coast of the United States found a trend of increasing

SOC/SOM over time (Craft 2001; Havens et al. 2002; Morgan and Short 2002; Craft et al. 2003).









decomposed as fast as the macroalgae and T. testudinum's decomposition was at the rate ofA.

germinans. These results indicate that in terms of decomposition, species identity matters more

than the group to which a species belongs. For source determination, these results specify which

species of an end member group are more likely to contribute to SOC because the slower a

species decomposes, the better chance its OC will be buried in sediments.

Sediments and Surface Layers

Changes in SL 15 SOC 813C Over the course of a year indicate new SOC sources are

adding to the sediment TOC pool or, without a change in TOC, decomposition of old source OC

while new source OC accumulates. These changes were greater in upper sections of both

mangrove (0-5 cm) and seagrass accretedd layer) sediments because the inputs of new OC reach

the top of sediments first. Bioturbation then brings new OC inputs deeper into the profile.

Bioturbating organisms were observed in mangrove, but not in seagrass sediments, which may

explain why 613C Of deeper mangrove sediments changed over time but deeper seagrass

sediments did not. Surface layers had the greatest 813C changes through the year.

All changes were positive so that the new SOC sources to SL 15 after construction must be

more enriched than old OC sources. Old OC sources were relatively depleted in 613C as they

were most likely the terrestrial plants that inhabited SL 15 pre-construction. In mangrove

sediments, the new source was most likely the algal mat and in seagrass accreted layers and floc

the new sources were macroalgae or seagrass (Fig. 4-3). Enrichment of algal mat 813C iS due to

changing inorganic C sources, as unlike other sediments and layers, the algal mat is its own

producer of OC. As the algal mats grow, so does their influence on the biogeochemistry of their

environment. Photosynthesis and respiration within the algal mat changes the pH of the water

around it (Kayombo et al. 2002). At night respiration decreases the pH, which can cause CaCO3

in the sediment below the algal mat to dissolve. CaCO3 dissolves into various carbonate species










0.20 -

0.18 -

0.16 -

0.14 -

0.12 -

0.10 -

0.08 -

0.06 -

0.04 -

0.02 -


Reference Mang rove Litter


mang rove


-30


-25


-20 -15


-10


813C

Figure 4-7. N:C vs. 813C in ternary mixing diagrams of three potential OC sources and surface
layers. Circles are the mean end member values and boxes are + standard deviation
of N:C and 613C. Filled triangles are reference litter layer values (A) and SL 15 floc
values (B). Open triangles are reference floc values (B).












Table 2-5. Stable Isotope values and dominant source conclusions from carbon source determination studies in coastal ecosystems.
Source Sediment Main How main sources Data
Location Potential sources 8' C Site description 61 C sources determined source
Mangrove forest


-21 to -22
-26.8
-27


Eastern Brazil


Seston
Spartina
Mangroves


Mangroves

Mangroves

Seston
Mangroves


Mangroves
Riverine
Shelf
Slope


-26.9
-23.8
-21.3
-20.5

-25.3

-22.7


Comparison


Gazi Bay, Kenya

Gazi Bay, Kenya

Southeast Asia


-28.25

-24.12


Rhizophora mucronata


Comparison


Ceriops tagal


4 Comparison


-20.5 to -23
-27 to -29


Coringa Wildlife
Sanctuary, India
Galle, Sri Lanka
Pampala, Sri Lanka


-29.4 to
-20.6


Compared to curve of
2 source mixing model


Mangroves and salt marsh

Chiricahueto, Mexico NR


Marsh


-20.4


Compared to
literature values


Comparison


Salt Marsh


Florida


Spartina alterniflora
Juncus roemerianus

Bare creekbank
Tall Spartina
Short Spartina
S. virginica high marsh
Sand flat
Mixed vegetation


-16.9
-23.9

-18.9
-16.0
-17.9
-21.6
-22.6
-19.3


-16.2


Sapelo Island, Georgia


Diatoms
S. alterniflora
S. virginica
D. spicata
S. virginicus
J. roemerianus
B. ~frutescens

S. 41terniflora


Comparison


-17.0
-12.9
-26.0
-13.1
-13.3
-22.8
-26.0


Barataria Bay,
Louisiana


-12.1 to -13.6


Marsh


Compared to predicted
values based on producer
biomass









Factors Affecting Functional Equivalence

The reference marsh used affects the functional equivalence of the constructed marsh. In

the last example (Craft et al. 1999), if the SOC pool of the 25-year-old constructed marsh had

been compared to the SOC pool of the 2,500-year-old natural marsh with a high OM content, the

authors would have concluded that the constructed marsh had not yet reached functional

equivalence. Many studies choose nearby natural marshes as references without regard to their

similarities to constructed marshes. Studies in urban areas are particularly limited by reference

sites as the restored site is often the only large area of marsh remaining (Simenstad and Thom

1996). Morgan and Short (2002) solved some of the problems associated with reference site

choice when they chose reference sites after comparing constructed sites to potential reference

sites using a principle components analysis (PCA) based on physical attributes like aspect, slope,

and size. They used the PCA to choose two well-matched reference sites for each constructed

site. Because reference marsh is a maj or factor in whether a constructed marsh reaches

functional equivalence, it should not be chosen arbitrarily.

While between-system factors affect whether a constructed wetland reaches functional

equivalence, so do within-system factors like elevation, depth in the soil profile, and variation in

sedimentation rates. Even when a constructed marsh as a whole is far from reaching functional

equivalence in terms of SOC, some parts of it may be closer than others. Several studies found

higher SOC at low elevations in constructed marshes (Lindau and Hossner 1981; Craft et al.

2002). Lower elevations are inundated by tides for longer periods of time, which leads to more

highly reducing conditions that can encourage OC storage. In most soils or sediments, OC

naturally decreases with depth, which may hinder the ability of lower depths to reach functional

equivalence. In one of the only studies to examine OC at different depths in the soil profile,

upper depths reached functional equivalence quickly while OC values at lower depths did not





For Pier


N


O


350 Meters


87.5 175
I I I I I


Figure 3-1. The study area in the Indian River Lagoon, next to Fort Pierce, Florida (inset). SL 15
is the large island in the center. Circles are mangrove system plots and squares are
seagrass system plots. Symbols outside of SL 15 are the reference sites, which have
one plot each.


SL 15


Indian
RFiver
Lagoon










samples were composites of different clumps collected from across the sampling area. Epiphyte

samples were composites of algal material scraped from seagrass fronds in the laboratory. Roots

of seagrasses and mangroves were taken from sediment cores for analysis; they were not

identified to species. Roots ofA. germinans, R. mangle, and S. alterniflora were collected in the

Hield as well. At the laboratory, seagrass fronds were scraped clean, and seagrasses, roots, and

macroalgae were rinsed. All plants were dried at 60oC for three days before being initial ground

on a Wiley mill (if necessary) and then ground to a Eine powder using a ball mill.

Seston was collected in May, September, October, and November 2006 and February

2007. For each seston sample, 500 mL of water was collected from the middle of the water

column in the subtidal area of SL 15. Three samples each were taken on a flood and an ebb tide

except in February 2007, where only ebb tide samples were collected. Water samples were kept

on ice and transported to the laboratory where they were filtered through precombusted

Whatman GF/F glass fiber filters. Blanks of 500 mL of deionized water were also filtered for

each sampling event. Filters were then freeze-dried for 24 hours.

Sediment Sampling

Four, 2 m x 2 m plots were established in the mangrove forest and in the seagrass bed on

SL 15 (Fig. 4-1). Three, 7 cm in diameter sediment cores from each of these plots were retrieved

in November 2005, January (mangrove only), February (seagrass only), May, July, and

November 2006. Cores were taken from different areas of the plots each time to ensure an area

was not re-sampled. For references, three randomly-selected plots were established in natural

mangrove forests and seagrass beds within 1 km of SL 15. These plots were sampled in July and

November 2006 using the same procedure as for SL 15 plots. Sediment cores were sectioned in

the field and stored in plastic bags on ice for transport and then in a 4oC refrigerator. SL 15 cores

were initially divided into 0-5 cm and 5-10 cm sediment depths. In subsequent samplings,









changes (Machas et al. 2006). Where 613C did change in decomposition studies of multiple

species, the initial differences in 613C between species were still clear. Unfortunately, we did not

measure changes in 613C Of our plant tissues during decomposition. Given the small magnitude

of changes found in other studies, and the large differences in 613C between groups of potential

sources, diagenetic changes in 613C are unlikely to cause misidentification of the main SOC

sources in this study. Changes in C:N during decomposition also occur and can be greater in

magnitude than 613C changes (Fourqurean and Schrlau 2003). Studies of mangrove, seagrass,

and macroalgal decomposition have found decreases and increases in C:N ratios that were

dependent upon species or tissue (Twilley et al. 1986; Bourgues et al. 1996; Fourqurean and

Schrlau 2003); others found no change in C:N ratios (Machas et al. 2006).

Decay constants of seagrasses on SL 15 were within literature values, which ranged from

0.002 to 0.12 day-l (Mateo and Romero 1996; Machas et al. 2006; Moore and Fairweather 2006).

T. testudinunt had a greater decay constant and therefore faster decomposition in this study than

in Florida bay (Fourqurean and Schrlau 2003). Mangrove decay constants were also within

literature values that ranged from 0.0048 to 0.022 dayl (Fourqurean and Schrlau 2003; Ake-

Castillo et al. 2006; Ramos e Silva et al. 2006). R. nzangle's decay constant in this study fell on

the low end of R. mangle reported values. Estimated macroalgae decay constants ranged widely

from 1 to 0.014 (Foreman and Smith 1984; Mews et al. 2006). The decay constant of Sarga~ssun

spp. in our study was at the low end of the range, probably because Salrga~ssunt has more

structural components than most other macroalgae. Surprisingly, differences among decay

constants in this study did not fall along plant groups. We expected mangroves to have the

lowest decay constants and macroalgae to have the highest with seagrass falling in between

(Kristensen 1994; Bourgues et al. 1996; Fourqurean and Schrlau 2003). However, S. filifornze









Sediment Organic Carbon Source Determination

One of the new directions functional traj ectory studies could take is examining sources of

SOC in constructed coastal systems. Determining SOC sources is important to the study of OC

storage in coastal ecosystems as the identity of sources is one of the factors that determine OC

liability and accumulation rates. Hedges (1992) stated that understanding the types of OM that

accumulate in marine sediments was one of the key questions that needs to be answered in order

to better understand global biogeochemical cycles. The source determination question most

often studied is whether the SOC is of allochthonous (via sedimentation) or autochthonous origin

( Middelburg et al. 1997; Bouillon et al. 2003; Golding et al. 2004). If most SOC is of

autochthonous origin, then contributions to SOC from the primary producers needs to be teased

apart, but this detailed question is harder to answer and rarely studied (Bouillon et al. 2003).

There are many possible SOC sources in seagrass beds, mangroves, and salt marshes. Coastal

ecosystems can have allochthonous OC inputs of planktonic origin from the open sea or of

terrestrial plant and anthropogenic origin. These systems also have numerous potential

autochthonous OC inputs. Within a seagrass bed OC can come from different species of

seagrasses, epiphytes, macroalgae, or micro-benthic algae. Seagrass beds can also receive OC

inputs from adjacent mangroves (Kennedy et al. 2004; Lin et al. 1991). The complexities of

seagrass beds, mangroves, and salt marshes make OC source determination difficult. However,

making sense of complex OC sources and their role in C accumulation and storage is important

for the conservation of coastal ecosystems in the face of increased nutrient loading and sea level

rise and the maintenance of their C sink capabilities.

Many methods have been used to determine SOC sources; however, no single method has

offered a definitive answer among and within system types. Some methods were developed for

two end-member systems--systems in which there are only two distinct sources of OC such as











Table 3-3. Results of the repeated measures ANOVAs for SL 15 mangrove and seagrass sediments (0-5 cm, 5-10 cm, and seagrass
accreted) and surface layers (algal mat and floc).
BD ExOC MBC TOC TN C:N Lability
ANOVA Effect (g cm-3) (mg kg ') (mg kg ') (%) (%/) (molar ratio) (mg 02 g-10C lu- )
Mangrove
0-10 Month *** NS *** NS NS
Depth ** NS NS NS ** NS *
Algal mat Month NS NS ** NS NS *
Seagrass
0-10 Month NS *** *** ** NS NS NS
Depth NS NS NS NS NS NS *
Accreted Month NS NS NS NS NS
Floc Month *** *


For significance NS=not significant, p <0.05, **p < 0.01, ***p < 0.0001. Please see Appendix A for a table listing how these data were transformed prior to
running the repeated measures ANOVA.










separately from the original sediment depths as an accreted layer. Surface layers--floc from

seagrass systems, algal mats from the SL 15 mangrove system, and litter layers from the

reference mangrove system--were collected from each core and were composite for each plot.

Differences in color and texture were used to separate accreted and surface layers from original

depths except for floc, which was the fraction of the accreted layer that poured off (Fig. 3-2).

Average heights of accreted and surface layers were measured for bulk density calculations. One

core per plot was retrieved in September 2006 and brought intact to the laboratory for pH and Eh

redoxx potential) measurements.

Laboratory Analyses

Sectioned sediments and surface layers were weighed to determine bulk density. Rocks,

roots, and detritus were removed from the sample before homogenization, and the volume and

weight of large rocks were taken into account when calculating bulk density. After

homogenization of each sample, a subsample was weighed to determine moisture content and the

remaining sample was split into two parts. One part (wet sample) was stored in airtight

containers at 4oC and the other was freeze-dried for 48 hours. Moisture content was determined

after subsamples were dried at 105oC for 24 hours.

Intact cores from September 2006 were incubated upright in tanks filled with 25 ppt

saltwater made with Instant Ocean (Marineland Labs, Moorpark, CA). Platinum electrodes were

inserted into each core at 2.5 cm, at 7.5 cm, at 12.5 cm (reference seagrass only), and halfway

through the accreted layer (SL 15 seagrass only). Platinum electrodes stabilized for 24 hours,

and then Eh was measured using an Accumet AP71 handheld meter and an Accumet calomel

reference electrode. Eh values were corrected relative to a standard hydrogen electrode. Cores

were then sectioned into 0-5 cm, 5-10 cm, and 10-15 cm or accreted depths as previously

described. pH was measured on 5 g of each depth using a Fisher Accumet AR50 pH meter.









A similar method was employed by Kennedy et al. (2004) when they examined SOC

sources in seagrass beds, mangroves, and mixed seagrass/mangrove systems in the South China

Sea. They measured 813C ValUeS of sediments, particles in sediment traps, and potential sources

(seagrass leaves, mangrove leaves, epiphytes, and seston). They found consistent and distinct

differences in the 613C ValUeS of potential sources. The use of the mixing equation gave broad

estimates of source contribution, which suggested that seagrasses and mangroves contributed to

the SOC in their respective systems, but that seston and epiphytes were probably the dominant

sources. As with the study by Papadimitriou et al. (2005), the presence of intermediate

signatures epiphytess and seston) between the two end members (seagrass and mangroves) made

determination of contributions from sources with intermediate 813C ValUeS difficult.

In salt marshes, similar problems are encountered. While the importance of the main

primary producer' s contribution can be easily elucidated, the contributions of other sources with

less distinct 813C Signatures cannot be. In one of the first studies that used C isotopes to examine

SOC sources, Spartina had the distinctive enriched 813C ValUeS (-12.3 to -13.6 %o) of C4 plants,

but all other vascular plants including species as disparate as Juncus roemerianus and Salicornia

virginica had signatures between -22.8 and -26 %o because they were C3 plants (Haines, 1976).

Benthic diatoms in this study were plagued with the same intermediately-valued problem as the

previously-di scussed epiphytes with 813C ValUeS between -16.2 and -17.90/oo. Through

comparing the primary producer and sediment values, Haines concluded sediment 813C ValUeS

generally reflected values of plants growing in the sediments. Sediments beneath C4 plants were

slightly more depleted in 613C than the C4 plants, and sediments beneath the C3 plants were

slightly more enriched in 613C than the C3 plants. The sediments' differences from in situ

vegetation may have been due to C3 and C4 plant detritus mixing or input from benthic diatoms.










LIST OF REFERENCES


ABER, J. D., AND J. M. MELILLO. 2001. Terrestrial Ecosystems, 2nd ed. Harcourt Academic
Press.

AKE-CASTILLO, J. A., G. VAZQUEZ, AND J. LOPEZ-PORTELLO. 2006. Litterfall and
decomposition of Rhizophora mangle L. in a coastal lagoon in the southern Gulf of
Mexico. Hydrobiologia 559: 101-111.

ALONGI, D. M. 2002. Present state and future of the world's mangrove forests. Environmental
Conservation 29: 331-349.

------, J. PFITZNER, L. A. TROTT, F. TIRENDI, P. DIXON, AND D. W. KLUMPP. 2005.
Rapid sediment accumulation and microbial mineralization in forests of the mangrove
Kandelia candel in the Jiulongjiang Estuary, China. Estuar. Coast. Shelf Sci. 63: 605-618.

------, AND OTHERS. 2004. Sediment accumulation and organic material flux in a managed
mangrove ecosystem: estimates of land-ocean-atmosphere exchange in peninsular
Malaysia. Mar. Geol. 208: 383-402.

------, AND OTHERS. 2001. Organic carbon accumulation and metabolic pathways in sediments
of mangrove forests in southern Thailand. Mar. Geol. 179: 85-103.

ANDERSON, W. T., AND J. W. FOURQUREAN. 2003. Intra- and interannual variability in
seagrass carbon and nitrogen stable isotopes from south Florida, a preliminary study. Org.
Geochem. 34: 185-194.

APHA. 1992. Biological Oxygen Demand. Standard methods for the examination of water and
wastewater. American Public Health Association.

BAILEY, V. L., A. D. PEACOCK, J. L. SMITH, AND H. BOLTON. 2002. Relationships
between soil microbial biomass determined by chloroform fumigation-extracti on,
substrate-induced respiration, and phospholipid fatty acid analysis. Soil Biol. Biochem. 34:
1385-1389.

BIERMAN, P. R. AND OTHERS. 1998. Erosion, Weathering, and Sedimentation, p. 647-678.
Dr C. Kendall and J. J. McDonnell [eds.], Isotope Tracers in Catchment Hydrology.
Elsevier.

BOSCHKER, H. T. S., A. WIELEMAKER, B. E. M. SCHAUB, AND M. HOLMER. 2000.
Limited coupling of macrophyte production and bacterial carbon cycling in the sediments
of Zostera spp. meadows. Mar. Ecol.-Prog. Ser. 203: 181-189.

BOUILLON, S., F. DAHDOUH-GUEBAS, A. RAO, N. KOEDAM, AND F. DEHAIRS. 2003.
Sources of organic carbon in mangrove sediments: variability and possible ecological
implications. Hydrobiologia 495: 33-39.










(Rhizophora mangle, Avicennia germinans, and Laguncularia racemosa) around their margins.

To mitigate destruction of a nearby mangrove forest and seagrass bed, seagrass and mangrove

systems were created on SL 15. These systems were created by burning and removing interior

vegetation and removing dredge spoil to create several different elevations. The seagrass bed,

which remains submerged during low tide, is at the lowest elevation, the mangrove forest, which

is exposed at low tide, is at the middle elevation, and a maritime forest occurs above sea level at

the highest elevation. The mangrove fringe of SL 15 was left intact except for a few flushing

channels. Between the constructed seagrass and mangrove systems a thin Spartina alterniflora

buffer was planted. The mangrove forest was planted with R. mangle, and maritime forests were

planted with Coccoloba uvifera, Borrichia frutescens, Rapanea guinensis, Conocarpus erectus,

and Distichlis spicata, but seagrasses were left to colonize naturally. Natural systems near SL 15

include its original mangrove forest fringe, surrounding seagrass beds, and mangrove fringes of

adj acent spoil islands, which are at least 40 years old.

Sediment Sampling

Four, 2 m x 2 m plots were established in the mangrove forest and in the seagrass bed on

SL 15 (Fig. 3-1). Three, 7 cm in diameter sediment cores from each of these plots were retrieved

in November 2005, January (mangrove only), February (seagrass only), May, July, and

November 2006. Cores were taken from different areas of the plots each time to ensure an area

was not re-sampled. For references, three randomly-selected plots were established in natural

mangrove forests and seagrass beds within 1 km of SL 15. These plots were sampled in July and

November 2006 using the same procedure as for SL 15 plots. Sediment cores were sectioned in

the field and stored in plastic bags on ice for transport and then in a 4oC refrigerator. SL 15 cores

were initially divided into 0-5 cm and 5-10 cm sediment depths. In subsequent samplings,

material had accumulated on top of the seagrass bed, which was collected and analyzed





SL 15 Seagrass




Floc
97 OC m-2
Accreted layer
300 g OC m-2



0-5 cm
177 g OC m-2

5-10 cm
240 g OC m-2


100 %
--100



S75
814 g OC m-2
(Total OC)




-25


Reference Seagrass


0 25


40-65 g OC m-2 1-Literaturevalue
(Net Accumulation)


75 100 %


V 50


Floc
90 g OC m-2
0-5 cm
414 g OC m-2


5-10 cm
332 g OC m-2


10-15 cm
371 g OC m-2


- 75
1207 g OC m-2
(Total OC)
- so



- 25


Figure 3-9. Continued


195 g OC m-2 1l
(Net Accumulation)

25 I50 75









interaction, p=0.018, Table 2; Fig. 3-6b). On a storage basis, all layers had lower TOC in SL 15

so there was not a significant interaction, but a Tukey revealed layers one and three had similar

TOC across sites (Fig. 3-6b; one-way ANOVA, df=5, p<0.0001). In seagrass sediments, TOC

was greatest in depth one (depth effect, p>0.013, Table 2; Table 4).

TN was significantly higher in reference than in SL 15 mangrove and seagrass systems

(site effect, p<0.0001, Table 2; Table 4), except in seagrass floc where month affected which site

had higher TN (site x month interaction, p=0.041, Table 2; Table 4). C:N was significantly

higher in the sediments and surface layers of mangrove references but was similar in the

sediments and surface layers of seagrass sites (site effect, p<0.0097, Table 2; Table 4).

In mangrove sediments, ExOC was significantly higher in references but, in seagrass

sediments, was significantly higher in SL 15 (site effect, p>0.0013, Table 2, Table 5). ExOC

(storage basis) of the 0-5 depth in SL 15's mangrove system was similar to reference depths

while SL 15's 5-10 depth had significantly lower ExOC (site x depth interaction, p=0.058, Table

2; Fig. 3-7a). In the seagrass systems, ExOC (concentration basis) was similar in depths two and

three across sites while depth one in SL 15 had greater ExOC than depth one in the reference

(site x depth interaction, p<0.0001, Table 2; Fig. 3-7b). On a storage basis, however, ExOC of

depths two and three in SL 15 were higher than the references, but depth one had similar ExOC

across sites (site x depth interaction, p=0.0017, Table 2; Fig. 3-7b). Upper depths had

significantly more ExOC in both mangrove and seagrass sediments (depth effect, p<0.003 8,

Table 2; Table 5). Surface layer ExOC did not significantly differ except for seagrass floc where

ExOC was greater on a concentration basis in SL 15 (site effect, p=0.020, Table 2; Table 5).

1VBC was significantly higher in reference sites for mangrove and seagrass sediments on a

concentration and storage basis (site effect, p<0.0001, Table 2; Tables 5; Fig. 3-8). In mangrove





































Figure 5-2. A modified mangrove forest carbon cycle showing values from this study in
constructed (C) and reference (R) systems. Organic carbon (OC) pools are the sum of
sediment and surface layer means (July and November data). Rates of microbial
carbon respiration are the mean of all depths (sediment and surface layers) in July and
November, adjusted from an Oz uptake rate to a carbon release rate by an assumed 1
02 to 6 C molar ratio. Bolded words are the main contributors to sediment OC pools.