Use of sediment amendments to rehabilitate sinking coastal swamp forests in Louisiana

Page 1

UL

Ba

b

C

a

ARRAA

KBDDNRST

1

tieiamaas

wofli

j

0h

Ecological Engineering 54 (2013) 183– 191

Contents lists available at SciVerse ScienceDirect

Ecological Engineering

j ourna l ho me page: www.elsev ier .com/ locate /eco leng

se of sediment amendments to rehabilitate sinking coastal swamp forests inouisiana

eth A. Middletona,∗, Ming Jiangb

U.S. Geological Survey, National Wetlands Research Center, 700 Cajundome Boulevard, Lafayette, LA 70506, USAChinese Academy of Science, Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agricultural Ecology, 3195 Weishan Road, Gaoxin District,hangchun 130012, Jilin Province, PR China

r t i c l e i n f o

rticle history:eceived 26 October 2012eceived in revised form 2 January 2013ccepted 16 January 2013vailable online 26 February 2013

eywords:aldcypress swampisturbance dynamicsredge spoil

a b s t r a c t

Coastal wetlands are losing elevation worldwide, so that techniques to increase elevation such as sedimentamendment might benefit these wetlands. This study examined the potential of sediment amendment toraise elevation and support the production and regeneration of vegetation in coastal forests in Louisiana.Before sediment amendment, the vegetation did not differ in these Taxodium distichum–Nyssa aquaticaforests with respect to herbaceous and tree seedling composition, and sapling and tree characteristics.After the application of sediment in January 2007, sediment-amended swamps had higher elevationsand salinity levels than natural swamps. The layer of sediment applied to Treasure Island in Jean LafitteNational Historic Park and Preserve was relatively deep (sediment depth at Site One and Site Two: 0.89and 0.69 m, respectively, six months after application), and may have exceeded an optimal threshold.

yssa aquaticaecruitmentalinityaxodium distichum

Sediment-amended swamp with the highest elevation had some tree mortality and little tree growth ofT. distichum. Also, sediment-amended swamp had higher root biomasses of ruderal species, and lowerspecies richness and cover of herbaceous species. Nevertheless, during controlled water releases during anoil spill emergency in 2010, both sediment-amended and reference forest had higher production levelsthan in other years. While sediment amendment is a compelling management alternative for sinkingcoastal wetlands, optimal thresholds were not determined for these T. distichum–N. aquatica swamps.

l2asptcmp

ogsJ

. Introduction

The eco-geomorphology of coastal wetlands may be altered inhe future because of changes in sea level, storm frequency andntensity, and inputs of freshwater, sediment and nutrients (Dayt al., 2008). To maintain coastal wetlands, the biological and phys-cal processes related to vertical accretion are of critical importancend must equal the level of sea level rise, or eventually the wetlandsay be lost (Cahoon et al., 1995). Worldwide, changes in climate

nd land-use threaten coastal wetlands, and specific managementpproaches to aid coastal wetlands in maintaining elevation areought (Day et al., 2008).

In the northern Gulf Coast of the United States, the majority ofetland loss has been due to coastal sinking because of alterations

f hydrology and sediment inputs, and withdrawal of subsurfaceuid (Turner, 1997). The coastal region of Louisiana is subsid-

ng at a rate of 4–5 ± 2 mm per year−1 due to sediment and water

∗ Corresponding author. Tel.: +1 337 266 618; fax: +1 337 266 8586.E-mail addresses: [email protected] (B.A. Middleton),

[email protected] (M. Jiang).

Er2Ibaaw

925-8574/$ – see front matter. Published by Elsevier B.V.ttp://dx.doi.org/10.1016/j.ecoleng.2013.01.025

Published by Elsevier B.V.

oading (Shinkle and Dokka, 2004) and shifting faults (Dokka,006). Regardless of the cause of coastal subsidence, inundationnd saltwater intrusion are undermining the ability of freshwaterwamp forests dominated by Taxodium distichum–Nyssa aquatica toersist (Conner and Brody, 1989). A higher elevation could improvehe survivorship, production, and regeneration of trees in theseoastal swamps. Sediment amendment from dredging activitiesay offer a potential solution to coastal inundation because the

rocess raises elevation.Dredged sediments have been used to raise the elevations

f coastal salt marshes, brackish marshes, mangroves, and searasses; however, the technique has seldom been used in coastalwamp forests dominated by T. distichum–N. aquatica (Terriordan, Edward Creef and Linda Mathies, U.S. Army Corps ofngineers, personnel communications). Sediment deposition byivers or hurricane activity is a natural process (Turner et al.,006), which sediment amendments resemble in some ways.

n coastal Louisiana, many coastal swamps are waterlogged

ecause of sea level rise and hydrologic alteration (Connernd Brody, 1989), so that tests of vegetation response tortificially raised elevations using sediment amendments arearranted.

Page 2

184 B.A. Middleton, M. Jiang / Ecological Engineering 54 (2013) 183– 191

Fig. 1. Coastal swamps dominated by T. distichum–N. aquatica, which were sediment-amended, or relatively sheltered or unsheltered natural forests (low and high pulse).Study swamps were south of New Orleans in the Barataria Preserve unit of Jean Lafitte NHP&P (lower left). Coordinates for the two sites of each type included sediment-amended: 29.749◦N, −90.145◦W and 29.752◦N, −90.145◦W; low pulse: 29.786◦N, −90.113◦W and 29.7852937◦N, −90.112◦W; high pulse: 29.791◦N, −90.121◦W and29.789◦N, −90.120◦W.

mrueg(22wpCAmtammiaf

iwoafs

2

2

ssFRtKwtil0arR(mwss

Sediment amendments have rejuvenated worldwide saltarshes and mangroves (Lee et al., 2011 and Burchett et al., 1998,

espectively), especially where salt marshes have disappearednder water (e.g., Sabine National Wildlife Refuges; Elsey-Quirkt al., 2009). Constructed marshes can achieve a level of above-round primary production that is equivalent to reference wetlandsCosta-Pierce and Weinstein, 2002; Kentula, 2002; Madrid et al.,012; Shafer and Streever, 2000; Turner and Streever, 2002; Zedler,000; Zedler and Callaway, 1999). In degenerating salt marshesith small to moderate amounts of sediment amendment applied,lant growth often increases (sediment depth: 0–10 and 0–30 cm;roft et al., 2006 and Mendelssohn and Kuhn, 2003, respectively).fter adding sediment (∼2.3 cm), the elevation of one sinking saltarsh increased by 6.2 cm after one year, with most of this addi-

ional elevation due to root biomass (Ford et al., 1999). Even so, certain threshold of sediment could exceed an optimal inter-ediate amount (Slocum et al., 2005), e.g., burial of tree trunksay create low oxygen conditions similar to the effects of flood-

ng (Kozlowski, 1991). Therefore, interest is growing in whether anppropriate amount of sediment amendment can benefit sinkingreshwater swamps.

The objectives of this project were to determine what effectncreasing elevations of swamps using sediment amendments

ould have on vegetation composition. We also made observations

f tree growth and below-ground production following sedimentmendment. Overall, we explored whether this project supportedorests of diverse species and types (i.e., herbs, shrubs, trees) ininking T. distichum swamps.

fpts

. Materials and methods

.1. Study site and sediment application

The natural experiment included two sediment-amendedwamps on Treasure Island (Appendix Fig. A1) and four referencewamps in long-term research sites, two each along Bayou desamilles and Lower Kenta Canal in Jean Lafitte NHP&P (Fig. 1).eference sites were along canals, which were somewhat shel-ered (Bayou des Familles; low pulse) or less sheltered (Lowerenta Canal; high pulse). The two sediment-amended swampsere positioned off the banks of the waterway, and graded into

he floating marsh and open water of Lake Salvador. Before sed-ment application, the elevations of the sediment-amended andow pulse swamp may have been roughly similar (mean elevation:.093 ± 0.002 and 0.179 ± 0.006 cm, respectively; based on Jiangnd Middleton, 2011), keeping in mind the limitations of the accu-acy of such measurements (uncertainty of Continuously Operatingeference Stations = 1–5 cm; Fig. 2).Sediment-amended swampsSites One and Two; Fig. 1) were treated with sediment amend-

ents starting in January 2007. Overall size of the swamp treatedith sediment was about 24 ha, and the overall amount of cypress

wamp in the national park was about 1573 ha (Dusty Pate, per-onal communication). Sediment applied to swamps was dredged

rom the adjacent waterway of the Bayou Segnette (David Muth,ersonal communication). The exact amount of sediment appliedo these swamps is unknown; however, in June 2007, mean ± S.E.ediment depth was 0.89 ± 0.005 and 0.69 ± 0.004 m at Site One

Page 3

B.A. Middleton, M. Jiang / Ecological E

Fig. 2. Elevation masl in sediment-amended vs. low pulse reference swamps asmeasured with sediment elevation tables (SET) after the growing season in JeanLafitte NHP&P (2006–2011). Different letters indicate significant differences inmeans using contrasts based on the interaction of swamp type × year (F5,12 = 825.8,p < 0.0001; based on Jiang and Middleton, 2011). Two Surface Elevation Table (SET)monuments were established at each of two sites in the sediment-amended and lowpulse swamps in December 2006, before the sediment amendment was appliedto Treasure Island (total of eight SETs; Appendix Fig. 1a). The actual elevationsof sediment in the plots were determined relative to these fixed SET monumentsin January 2010 with a Trimble R8 GNSS dual frequency receiver. Elevation esti-mates were corrected for the sediment-amended and low pulse sites using elevationcontrol points within Continuously Operating Reference Stations (CORS) using theL2L

ayhawwf

dpv

2

a

FdL

ptl

2

swdpw(wmb2−S

2

ahia18re

srus

2

mS(a

SU Gulfnet Real-time Network (Uncertainty = 1–5 cm; Center for Geoinformatics,006), and an established benchmark (Crown Point Industrial, JPW112B, south ofafitte, Louisiana).

nd Site Two, respectively (Fig. 2; Jiang and Middleton, 2011). In theear following sediment-amendment (2007), soil bulk density wasigher and organic matter lower than in low pulse swamps (Jiangnd Middleton, 2011). Note that all Jean Lafitte NHP&P swampsere downstream of the Davis Pond Diversion, a structure fromhich water was released after May 10, 2010 for several months

ollowing an offshore oil spill (Fig. 3; Louisiana DNR, 2012).Dominant tree species in the Jean Lafitte swamps included T.

istichum, N. aquatica, Acer rubrum var. drummondii and Fraxinusrofunda, and with an understory of Sabal minor (personal obser-ation; Appendix Fig. A1a).

.2. Experimental design

Five study plots were chosen within each of the two sediment-mended, low and high pulse swamps (Fig. 1). Each of the five

ig. 3. Discharge rate cf s−1 (cubic feet second−1) based on provisional water gageata from the outfall channel of Davis Pond on Highway 90 (Gauge #DCPBA03l;ouisiana DNR, 2012).

rtst

iioDSGspafhta

2

2

s

ngineering 54 (2013) 183– 191 185

lots was located at stratified random positions at 25 m intervalso the right of a linear transect and marked with a stake (transectength = 125 m).

.2.1. Environmental and elevation assessmentWater depth was measured with a meter stick on the day of

ampling. Elevations of the plots were determined by comparingater depths at SET monuments with water depths at plots. SETata consist of measurements of sediment elevation at each of nineins in four directions from a monument, and these measurementsere taken only at the low pulse and sediment-amended sites

Jiang and Middleton, 2011). To measure salinity (ppt) and pH, poreater was extracted from the soil using a pore water sipper, andeasured with a YSI EC 300® probe. Salinities for Treasure Island

efore 2007 were based on a published study (2.01 ppt; Krauss et al.,009) in a site adjacent to our sediment-amended sites (29.755◦N,90.145◦W; less than 283 and 549 m from Sediment-Amendedites Two and One, respectively; Krauss, personal communication).

.2.2. Vegetation assessmentGround vegetation including tree seedlings and herbs was

ssessed in each of the five plots along the transect in low pulse,igh pulse and sediment-amended sites near the end of the grow-

ng season in 2006 through 2011. Percent cover was visuallyssessed for each species in 1-m2 plots in classes as follows: 1,%; 2, 1–5%; 3, 5–15%; 4, 15–25%; 5, 25–50%; 6, 50–75%; 7, 75–95%;, 95–100% (Daubenmire, 1959). Mid-points of cover classes wereecorded. Numbers of seedlings of woody species were counted inach plot.

To assess saplings and shrubs, number and height (cm) ofpecies were recorded in one 10-m2 plots at each site (Plot One,andomly selected). Because seedlings of T. distichum were ratherncommon but also of special interest, the number of T. distichumeedlings were counted within 1 m of the 125 m transect.

.2.3. Swamp productionAnnual root production was assessed using the implanted soil

ass technique (root ingrowth cores; Lund et al., 1970) fromeptember 2008 through September 2011 in all site types. One core7 cm diameter × 30 cm depth) was placed in each plot (5 per site),nd marked with a stake. After one year in September, the core wasemoved from the soil, divided into 3, 10 cm sections, and washedhrough a sieve to remove soil. Collected roots were dried to a con-tant mass at 70 ◦C in a drying oven and weighed without regardo size class.

To measure annual tree growth, dendrometer bands werenstalled on the nearest tree adjacent to each plot along the transectn the low pulse and sediment-amended sites. Dbh was measuredf live T. distichum trees in two sediment-amended swamps inecember 2006 (5 trees in each site), and two low pulse swamps ineptember 2006 (3 trees adjacent to first three plots in each site).rowth was assessed between September 2007 and 2011 in bothite types. Trees were suitable if these were large enough to be aart of the upper canopy of the forest. Bands were installed aboveny tree fluting or buttress, approximately 1.5 m above the soil sur-ace (Keeland and Sharitz, 1993). Bands were placed about 0.5 migher in sediment-amended swamps, but the ground to band dis-ance was somewhat smaller than natural forests after sedimentpplication.

.3. Statistical analysis

.3.1. EnvironmentSET measurements were analyzed using ANOVA comparing

ediment-amended and low pulse swamps by nesting direction

Page 4

186 B.A. Middleton, M. Jiang / Ecological Engineering 54 (2013) 183– 191

Table 1Mean ± S.E. of pH, pore water salinity (ppt), and water depth (cm) based on sample day measurements of plots (∼five per site) along transects in sediment-amended, lowpulse and high pulse swamps in the Barataria Preserve unit of Jean Lafitte National Historical Park and Preserve. Bracketed values show minimum and maximum values.Data not available for 2006–2011 are designated as “na”. The 2006 data were taken prior to sediment application in the sediment-amended swamps on Treasure Island, JeanLafitte NHP&P.

Forest type

Variable Sediment-amended Low pulse High pulse

pH2007–2011 6.5 ± 0.1 [6.3–6.8] 6.5 ± 0.1 [6.3–6.9] 6.6 ± 0.1 [5.9–6.9]Salinity (ppt) 3.3 ± 0.2 [1.7–5.1] 0.6 ± 0.1 [0–1.1] 1.0 ± 0.1 [0.1–2.1]2006a 2.0 1.8 ± 0.1 [1.7–1.9] 2.1 ± <0.1 [2.0–2.1]2007–2011 2.7 ± 0.2 [0.4–5.1] 0.5 ± <0.01 [0.0–1.1] 0.8 ± 0.1 [0.1–2.1]

Water depth (cm) at plots2006 23.5 ± 0.8 [20.0–25.0] 0.0 ± 0.0 [0.0–0.0] 1.3 ± 0.4 [0.0–3.0]

ss et a

ao(2t

2

Nsrpc>nwt(

2

sAoptsrabtemdfitt

orepstdad

bu

3

3

stsW

3

n(tst

aOfpa

viw1tAttcl

3

si

2007–2011 3.0 ± 0.9 [0–25]

a Salinity values prior to sediment amendment on Treasure Island are from Krau

nd pin number within SET number and using the main effectsf swamp type and year, and the interaction of swamp type × yearas reported in the Study Site section; from Jiang and Middleton,011). Environmental measurements reflected conditions taken onhe day of sampling only, so these were not compared statistically.

.3.2. Multivariate analysis of vegetation compositionMultivariate analyses were performed using two-dimensional

MDS for ground vegetation percent cover and for sapling den-ity (site means and means, respectively; per 1-m2 and 10-m2,espectively) using Euclidian dissimilarity matrices after appro-riate transformation for ground vegetation with mean percentover >0.1% per m2 or species of saplings with mean density of0.1 saplings per 10 m2, respectively. Because there were sites witho species in common for the sapling data, a stepacross methodas used (Oksanen, 2012). Data were centered and scaled prior to

he analysis, which was carried out using the Vegan Package in ROksanen, 2012, and R Foundation, 2012).

.3.3. Univariate analysisComparisons of total species richness and cover were made of

ediment-amended, and low and high pulse forests with split-plotNOVA using a repeated measures approach with the main effectsf swamp type, year, and the interaction of swamp type × time withlot and year nested within site, and random effects assigned tohis term (JMP SAS, 2012). A similar analysis was performed forediment and low pulse forests (only) for tree growth (ratio of cur-ent versus previous year), i.e., rate of change per year based onnnual increment data from dendrometer bands. Similarly, annualelow-ground root biomass was compared in these three swampypes using split-plot repeated measures ANOVA with the mainffects of swamp type and depth, year and the interactions of theseain effects. Core depth was nested within plot and site and ran-

om effects assigned to this term. No significant differences wereound in root biomass in the lower two core depths of the rootngrowth cores (10–20 and 20–30 cm; t = −0.001, p = 0.993), so thathe ANOVA was conducted using a combined mean of the lowerwo depths in comparison to the upper depth (upper; 0–10 cm).

Responses of various woody variables were examined using sec-nd order polynomial regression in ANOVA to detect any non-linearesponse over time to sediment-amendment. Woody variablesxamined included total seedling density per 1-m2, seedling coverer m2, total T. distichum seedlings per transect (125 m), totalapling density per 10-m2, average sapling height per 10-m2 (i.e.,

otal sapling height/total sapling density per 1-m2), total shrubensity per 10-m2, and average shrub height per 10-m2. Data wereppropriately transformed to meet assumptions of ANOVA. One-egree-of-freedom contrasts were made of comparisons of interest

as2t

11.3 ± 1.7 [0–71] 14.9 ± 1.7 [0–48]

l. (2009).

ased on the interactions. All statistical analyses were conductedsing JMP SAS (2012).

. Results

.1. Environmental comparisons

Water depth ranged from drawn down to shallow flooding inediment-amended, low pulse and high pulse swamps from 2006o 2011 (Table 1). Pore water salinity had a higher range of values inediment-amended than in low and high pulse swamps (Table 1).

ater pH was similar for all swamp types (Table 1).

.2. Ground vegetation characteristics

Herbaceous species richness of sediment-amended swamp didot differ in natural forest and sediment-amended forest in 2006mean species richness = 3.0 ± 0.57; t = 0.4074, p = 0.691; swampype × year interaction: F10, 72 = 2.7, p < 0.05). By 2009–2010,ediment-amended had lower species richness of ground vegeta-ion than natural forest in most years (t = 5.77, p < 0.0001).

Percent cover of ground vegetation did not differ in sediment-mended versus natural forest in 2006 (t = 1.079, p = 0.3086).verall pattern of percent cover of the ground vegetation varied

rom 2007–2009 (forest type × year × year interaction: F2, 12 = 4.0, = 0.037) with mean percent cover decreasing in sediment-mended, but not in natural forest.

Using NMDS analysis, we examined the relationships of groundegetation (including tree seedlings and herbs) to environmentsncluding salinity, year and elevation. Of these factors, only salinity

as significantly related to variation in the model (salinity ∼ axis + axis 2; r2 = 0.5421, p = 0.0009; stress = 0.02872). Some species ofree seedlings were related to higher salinity levels (Salix nigra and. drummondii), and T. distichum and Sabal palmetto were relatedo environments with lower salinity (Fig. 4b [NMDS ground vege-ation]). Similarly, cover of herbaceous species such as Pontederiaordata and Saururus cernuus were related to environments withow salinity (Fig. 4b [NMDS ground vegetation]).

.3. Woody seedling composition

Before sediment amendment in 2006, woody seedling andapling variables did not differ in the three swamp types. After sed-ment amendment, woody seeding variables such as total density

nd cover of seedlings per m2, and total number of T. distichumeedlings along a 125 m transect differed (Appendix 1a–c). In007, seedling and sapling density initially increased so that theotal density of seedlings per m2 and total cover of seedlings per

Page 5

B.A. Middleton, M. Jiang / Ecological Engineering 54 (2013) 183– 191 187

Fig. 4. NMDS graphs of (a) tree species exceeding a mean density of 0.1 saplings per 10-m2, and (b) ground vegetation exceeding a mean percent cover of 0.1% per 1-m2. Species abbreviations include Acer rubrum var. drummondii (Acer), Carex spp. (Carex), Eichhornia crassipes (Eichhornia), Eleocharis palustris (Eleocharis), Hydrocotyle spp.( corda( era (Tr4

mpsiAis

3

t2ts

saiswaaa

3

2taa2

3

n(d

iailhph2G

3

than in low and high pulse swamps (Appendix 3c and Fig. 6;374.5 ± 53.9 g per m2 per year−1 vs. 135.6 ± 26.7 g per m2 peryear−1, respectively; t = −6.65, p < 0.0001). Roots in sediment-amended forests were mostly of the species S. nigra and A. rubrum

Fig. 5. Growth of T. distichum trees based on ratio of diameter (mm per year−1)

Hydrocoytle), Lemna minor (Lemna), Polygonum punctatum (Polygonum), PontederiaSalvinia), Saururus cernuus (Saururus), Taxodium distichum (Taxodium), Triadica sebif.

2 increased in sediment-amended compared to low and highulse swamps (t = 132.6 and −15.3, respectively; p < 0.0001). Sub-equently, seedling density decreased in 2008, and sapling densityn 2010 (i.e., responses fit second order polynomials; p < 0.0001;ppendix 1a–c). Number of seedlings along the transect differed

n high pulse swamps in 2009 from all other years in high pulsewamps, as well as in sediment-amended and low pulse swamps.

.4. Sapling composition

Sapling densities per 10-m2 were the same in all swampypes in 2006, but increased in the sediment-amended swamps in007–2009 (Appendix 2a, t = −3.06, p = 0.003). Sapling heights werealler in sediment-amended swamp than in low and high pulsewamps (Appendix 2b; p = 0.01).

Using NMDS analysis, we examined the relationships of meanpecies densities of saplings with respect to salinity, elevationnd time. Only salinity was significantly related to variationn the model (salinity ∼ axis 1 + axis 2; r2 = 0.4087, p = 0.0080;tress = 0.00406; Fig. 4a [NMDS graph]). Sapling densities of S. nigraere related to higher salinity environments than A. drummondii

nd Triadica sebifera (Fig. 4a [NMDS graph]). Apart from the NMDSnalysis, T. sebifera saplings had very low densities in sediment-mended sites (0.02 ± 0.02 individuals per m2).

.5. Shrub density

Shrub density per 10-m2 was the same in all swamp types from006–2011 (p > 0.01; Appendix 2c). No vines were observed duringhe study. Mean shrub height (cm) was the same in 2006–2007nd 2010–2011 for all swamp types, but was taller in sediment-mended swamps in 2008–2009 (t = −4.695, p < 0.0001; Appendixd).

.6. Tree size and growth

Initial diameter at breast height (dbh) of T. distichum trees didot differ in sediment-amended versus low pulse swamps in 2006Appendix 3a). One of the randomly selected trees with a den-roband died in the sediment-amended site with deeper sediment

iab2s

ta (Pontederia), Sagittaria lancifolia (Sagittaria), Salix nigra (Salix), Salvinia molestaiadica), and Typha glauca (Typha). For additional species information see: Appendix

n 2007 (Site One), and many other dead trees were observedfter sediment application in this site (Appendix Fig. A1c). In 2010n comparison to other years, tree growth as measured by theog ratio of current versus previous year diameter increase wasigher in both sediment-amended and low pulse swamps (t = −9.8,

< 0.0001; Fig. 5, Appendix 3b); also in 2010, tree growth wasigher in sediment-amended than in low pulse swamps (ratio:.8 ± 0.4 vs. 1.5 ±0.1, respectively; t = −7.3, p < 0.0001; Appendix

3b; Fig. 5).

.7. Root annual production

Annual root biomass was higher in sediment-amended swamps

ncrease from the current to previous year using dendrometer bands in sediment-mended vs. low pulse swamps, Jean Lafitte NHP&P, 2008–2011. Dendrometerands were installed in December 2006 in sediment-amended sites, and September006 in low-pulse sites. Growth was assessed until 2011. Different letters indicateignificant differences in means based on the interaction of swamp type × year

Page 6

188 B.A. Middleton, M. Jiang / Ecological E

Fig. 6. Mean annual root biomass (g per m2 per year−1) for sediment-amended,low pulse and high pulse swamp forest of the Barataria Preserve, Jean LafitteNHP&P. Thirty centimeter long root ingrowth cores were inserted into the soilin September, and removed after the following growing season in 2008–2009,2009–2010, 2010–2011. Root biomass did not differ in the 10–20 and 20–30 cmdpe

vsday2

4

dbiipstse2

cddsmsb(2af

aW

Fli

epth, so a mean of these depths was compared to the upper depth (0–10 cm); > 0.05. Different letters indicate significant differences in means based on mainffect comparisons.

ar. drummondii as based on casual visual assessment. Within allwamp types, root biomass was higher near the surface than in

eeper layers of the soil (0–10 vs. 10–30 cm depth; Appendix 3cnd Fig. 6). Root biomass was lower during 2009–2010 than otherears (2009–2010 vs. 2008–2009 and 2010–2011: 122.7 ± 41.4 vs.55.1 ± 33.0, respectively; t = −4.84, p < 0.0001).

sStd

ig. A1. (a) Sediment-amended swamp prior to sediment amendment with platforms useow pulse reference swamp along Bayou de Familles (c) sediment-amended swamp after

nvasion by S. nigra and A. rubrum var. drummondii in 2007 with Dasa Bastlova, U.S. Geolo

ngineering 54 (2013) 183– 191

. Discussion

Specific approaches have been designed to offset coastal degra-ation and land loss by maintaining elevation (Day et al., 2008),ut the success of these techniques have not been fully tested

n various wetland types. Coastal elevation management couldnclude techniques to increase organic matter production to boosteat accumulation, or to apply sediment or sediment-laden diver-ion water (CPRA, 2012). The application of sediment amendmentso coastal wetlands can rehabilitate certain functions by raisingurface elevation, and/or reducing flood stress for vegetation (Fordt al., 1999; Mendelssohn and Kuhn, 2003; Stagg and Mendelssohn,010; Lee et al., 2011).

While sediment application can rehabilitate various types ofoastal wetlands, there may be an upper threshold of optimalepth, so that sediment amendment deeper than this level may beetrimental (Slocum et al., 2005; Stagg and Mendelssohn, 2010). Inalt marshes, sediment amendment exceeding 36 cm can be detri-ental (Stagg and Mendelssohn, 2010). In this study, depths of

ediments in the project area after two years of compression wereelow the upper threshold reported by Stagg and Mendelssohn2010) (31.5 cm above original elevation in Jiang and Middleton,011). Nevertheless, the amount of sediment used in the projectrea of this study may have exceeded an optimal threshold for theseorests.

For any benefit to occur, sediment application ideally would cre-te flooding and salinity conditions amenable to coastal forests.hile other studies indicate that sediment addition can caused

tressed salt marshes to grow more (Mendelssohn and Kuhn, 2003;hrift and Mendelssohn, 2008), we did not observe these condi-ions in the forests of these studies. In particular, some of the T.istichum trees died in Site One (but not Site Two) the year after

d to install SETs, Treasure Island, Jean Lafitte NHP&P (December 2006) including (b)sediment application (June 2007), and (d) sediment-amended swamp after ruderalgical Survey volunteer, shown. Photos by Beth Middleton and Evelyn Anemaet.

Page 7

ical E

shawdi

rtoIcd(m(ihksi

amimebtIss2ps

caisswdtssfdcip(lFmg

itisvsfe

ad

5

scfsteo(mc

A

PstlorE1pm

A

plJs(tp(

Swamp type × year × year 2,12 4.7 0.010**−1

B.A. Middleton, M. Jiang / Ecolog

ediment was applied (2007). Site 1 had lower soil moisture andigher elevation than Site Two after sediment amendment (Jiangnd Middleton, 2011). Nevertheless, when all of these forest typesere exposed to higher amounts of water during hydrologic reme-iation from the oil spill during 2010, trees had higher growth than

n other years.Sediment burial depths beyond the optimal threshold have been

eported for other coastal wetland types. For seagrasses, burialhresholds depend on species, ranging from 2 to 19.5 cm (Halophilavalis and Posidonia australis, respectively; Cubaco et al., 2008).n mangrove swamps, sediment deposition exceeding 10 cm canause mortality (Ellison, 1998). In mangroves, burial toleranceepends on the ability to increase pneumatophore/knee heightEllison, 1998). Burial of knees may be a factor in T. distichum

ortality; knees at Site One were completely buried by sedimentMiddleton, personal observation). The specific causes of mortalityn deeply buried vegetation is unknown; however, certain speciesave appendages with adaptations for improved root aeration (e.g.,nees, mangrove prop roots, adventitious roots: Grosse et al., 1992),o that covering these too deeply with layers of sediment maympede the exchange of gases.

For freshwater species in coastal T. distichum swamps, sedimentpplication might also cause conditions of high salinity and low soiloisture. T. distichum growth is lower in salinity levels exceed-

ng 1.3 ppt (Krauss et al., 2009), and high salinity levels eventuallyay cause T. distichum trees to die or grow more slowly (Hackney

t al., 2007). T. distichum can persist for a time in high salinityy shifting water-use strategy (Krauss and Duberstein, 2010). Inhis study, the stands of sediment-amended forests on Treasuresland had mean pore water salinities of about 3.3 ppt after theediment-application, although measured salinities prior to theediment-application were lower than this (∼2.0 ppt, Krauss et al.,009). While the salinity of the sediment used in the Treasure Islandroject is not known, coastal dredging sometimes utilizes salineediment (Bramley and Rimmer, 1988).

With or without sediment amendment, primary production ofoastal forests could possibly build elevation, giving coastal forestsn innate capacity to keep up with sea level rise and subsidence ass true of some coastal marshes (Langley et al., 2010). In this study,ediment-amended forests had higher root biomass than naturalwamps; however, root material in sediment-amended swampsas almost entirely of recently invaded S. nigra and A. rubrum var.

rummondii, and not related to any increased growth in T. dis-ichum. Similarly, S. nigra readily invades deep layers of sand andilt recently deposited on bars in riverine T. distichum–N. aquaticawamps in the Atchafalaya Basin (Hupp et al., 2008). Root materialrom these shorter lived woody species (S. nigra and A. rubrum var.rummondii) might improve the bulk density and water holdingapacity of soil over time. Even so, southern wetland forests dom-nated by A. rubrum and Nyssa sylvatica sharply declined in rootroduction at sediment deposition rates above 0.3 cm per year−1

Cavalcanti and Lockaby, 2005), so that threshold levels may beess than 0.1–0.4 cm per year−1 in these forests (Jolley et al., 2009).rom the perspective of coastal elevation building, the key issueay be to provide these forests with an amenable environment for

rowth.Species composition of regenerating tree seedlings after sed-

ment addition is important to future forest composition. Inhis study, T. distichum and T. sebifera seedlings were not foundn sediment-amended swamps. Based on our analysis of treeeedling composition and tree growth, S. nigra and A. rubrum

ar. drummondii and not T. distichum are likely to dominateediment-amended swamps in the near future. Despite the fact thatorest composition may change after sediment amendment, furtherxploration of this and other elevation management techniques

ngineering 54 (2013) 183– 191 189

re warranted because these sinking swamps may be otherwiseoomed.

. Conclusion

While sediment application shows some promise to rejuvenateinking coastal wetlands, deep layers of sediment application inoastal swamp forest may shift species the composition in theseorests by changing the composition of herbaceous and woodypecies. The deep layers of sediment used in the project area ofhis study (0.69–0.89 m, six months after application may havexceeded an optimal threshold, which can impair function. Studiesf other coastal wetland types using shallower depths of sediment<35 cm) suggest benefits to the growth of vegetation. Future tests

ight examine the efficacy of shallower sediment amendment inoastal freshwater swamps.

cknowledgements

This study was supported with funding from the U.S. Nationalark Service, POBS61, “Effects of dredge spoil applications onubsiding coastal baldcypress swamps in Jean Lafitte National His-orical Park and Preserve, Louisiana”. The funding agency providedogistical support and access to field sites, but was not involved inther aspects of the study execution, analysis and report prepa-ation. Additional funding came from the U.S. Geological Surveycosystem Program and the National Science Foundation (DEB-049838). Any use of trade, product, or firm names is for descriptiveurposes only and does not imply endorsement by the U.S. Govern-ent.

ppendix 1.

ANOVAs of various woody seedling variables using second orderolynomial regression analysis to compare sediment-amended,

ow pulse and high pulse swamps of the Barataria Preserve unit ofean Lafitte National Historical Park and Preserve, 2006–2011. Dataets analyzed included: (a) total woody seedling density 1 m−2,b) seedling cover in m2 plots, and, (c) total T. distichum seedlingsransect−1 (125 m). Different letters indicate that the mean com-arisons differ from one another based on contrasts in ANOVAp < 0.0001).

Variable df F p Swamp typemean ± S.E.

(a) Total woody seedling density m−2

Swamp type 2,12 31.8 <0.001***Sediment-amended 20.1 ± 3.6a

Low pulse 1.8 ± 0.9b

High pulse 0.2 ± 0.1b

Year 5,20 0.2 0.624Swamp type × year 10,72 0.4 0.651Year × year 1,4 2.8 0.094Swamp type × year × year 2,12 3.2 0.042*

(b) Total woody seedling % cover m−2

Swamp type 2,12 25.5 <0.001***Sediment-amended 3.2 ± 0.9a

Low pulse 0.1 ± <0.1b

High pulse <0.1 ± <0.1b

Year 5,20 4.1 0.044*Swamp type × year 10,72 3.4 0.037*Year × year 1,4 4.2 0.041*

(c) Total T. distichum seedling transectSwamp type 2,3 9.6 <0.001***

Sediment-amended 0.5 ± 0.1a

Low pulse 0.1 ± <0.1b

High pulse 0.4 ± 0.2a

Page 8

1 gical E

A

npJsadhDo

A

had(tBPimA

A

aav

A

f2t

R

B

B

Cmarshes of the southeastern United States: Kaye and Barghoorn revisited. Mar.Geol. 128, 1–9.

Cavalcanti, G.G., Lockaby, B.G., 2005. Effects of sediment deposition on fine root

90 B.A. Middleton, M. Jiang / Ecolo

Year 5,6 0.3 0.555Swamp type × year 10,18 2.2 0.114Year × year 1,1 2.1 0.150Swamp type × year × year 2,3 6.5 0.002**

ppendix 2.

ANOVAs of sapling and shrub variables using second order poly-omial regression analysis to compare sediment-amended, lowulse and high pulse swamps of the Barataria Preserve unit of

ean Lafitte National Historical Park and Preserve, 2006–2011. Dataets analyzed included: (a) total sapling density 10 m−2, (b) aver-ge sapling height 10 m−2 (i.e., total sapling height/total saplingensity 1 m−2), (c) total shrub density 10 m−2, (d) average shrubeight 10 m−2. Note that no vines were observed in the field survey.ifferent letters indicate that the mean comparisons differ fromne another based on contrasts in ANOVA (p < 0.0001).

Variable df F p Swamp typemean ± S.E.

(a) Total sapling density 10 m−2

Swamp type 2,3 6.6 <0.002**Sediment-amended 8.1 ± 3.4a

Low pulse 1.4 ± 0.9b

High pulse 1.5 ± 0.7b

Year 5,6 1.2 0.264Swamp type × year 10,18 0.1 0.875Year × year 1,1 3.4 0.065Swamp type × year × year 2,3 3.1 0.046*

(b) Average sapling height (cm) 10 m−2

Swamp type 2,3 4.3 0.015*Sediment-amended 193.3 ± 1.3a

Low pulse 0.2 ± 0.1b

High pulse 0.3 ± 0.1b

Year 5,6 0.1 0.728Swamp type × year 10,18 0.1 0.890Year × year 1,1 2.0 0.155Swamp type × year × year 2,3 2.0 0.132

(c) Total shrub density 10 m−2

Swamp type 2,3 1.3 0.256Sediment-amended 0.2 ± 0.1a

Low pulse 0.1 ± 0.1a

High pulse 0.2 ± 0.1a

Year 5,6 2.9 0.092Swamp type × year 10,18 1.2 0.310Year × year 1,1 0.2 0.650Swamp type × year × year 2,3 2.0 0.139

(d) Average shrub height (cm) 10 m−2

Swamp type 2,3 6.0 0.003**Sediment-amended 15.0 ± 9.7a

Low pulse 0.3 ± 0.1b

High pulse 0.2 ± 0.1b

Year 5,6 0.1 0.075Swamp type × year 10,18 <0.1 0.995Year × year 1,1 2.9 0.092Swamp type × year × year 2,3 3.4 0.036*

ppendix 3.

ANOVA of production data including log (a) diameter at breasteight (initial dbh in 2007), (b) tree growth ratio of T. distichums the ratio of annual diameter increase measured with den-rometer bands (mm year−1; 2007–2011), (c) annual root biomassg m−3 year−1; 2008–2011). Dendrometer bands were placed onrees in sediment-amended and low pulse swamps (only) of thearataria Preserve unit of Jean Lafitte National Historical Park and

reserve. For tree growth (mm), there was a swamp type × yearnteraction (F1, 4 = 5.9, P = 0.0108). Different letters indicate that the

ean comparisons differ from one another based on contrasts inNOVA (p < 0.0001).

C

C

ngineering 54 (2013) 183– 191

Variable df F p Mean ± S.E.

(a) dbh (initial, 2007)Swamp type 1,8 <0.1 0.932

Sediment-amended 36.0 ± 2.9a

Low flow 36.5 ± 6.5a

(b) Tree growth (mm year−1)Swamp type 1,10 12.4 0.001***

Year 3,12 22.9 <0.001***Swamp type × year 3,32 2.7 <0.001***

(c) root biomass (g m−3 year−1)Swamp type 2,12 22.9 <0.001***

Sediment-amended 374.5 ± 53.9a

Low pulse 121.2 ± 28.7b

High pulse 150.0 ± 45.3b

Depth 1,8 5.7 0.018*0–10 cm 299.3 ± 47.9a

10–30 cm 123.6 ± 17.8b

Year 2,16 11.9 <0.001***2008–2009 272.7 ± 52.7a

2009–2010 122.7 ± 41.4b

2010–2011 238.1 ± 40.6a

Swamp type × depth 2,24 2.8 0.066Swamp type × year 4,36 2.0 0.096Swamp type × depth × year 4,72 1.4 0.239

ppendix 4.

Abbreviations and life history types of species used in the NMDSnalyses (Fig. 4). In these analyses, each species of tree sapling had

density exceeding 0.1 saplings per 10 m2 (Fig. 4a) and groundegetation had a cover exceeding 0.1% m−2 (Figs. 4b).

Species Abbreviation Type

Acer rubrum var. drummondii Acer TreeCarex spp. Carex GraminoidEichhornia crassipes Eichhornia Floating herb (exotic)Eleocharis palustris Eleocharis GraminoidHydrocotyle spp. Hydrocoytle Floating leaved herbLemna minor Lemna Floating herbPolygonum punctatum Polygonum HerbPontederia cordata Pontederia HerbSagittaria lancifolia Sagittaria HerbSalix nigra Salix TreeSalvinia molesta Salvinia Floating herb (exotic)Saururus cernuus Saururus HerbTaxodium distichum Taxodium TreeTriadica sebifera Triadica Tree (exotic)Typha glauca Typha Emergent

ppendix E. Supplementary data

Supplementary data associated with this article can beound, in the online version, at http://dx.doi.org/10.1016/j.ecoleng.013.01.025. These data include Google maps of the most impor-ant areas described in this article.

eferences

ramley, R.G.V., Rimmer, D.L., 1988. Dredged materials – problems associated withtheir use on land. J. Soil Sci. 39, 469–482.

urchett, M.D., Pulkownik, A., Grant, A., MacFarlane, G., 1998. Rehabilitationof a saline wetland, Olympics 2000 site, Sydney (Australia) – I: manage-ment strategies based on ecological needs assessment. Mar. Pollut. Bull. 37,515–525.

ahoon, D., Reed, D., Day, J., 1995. Estimating shallow subsidence in microtidal salt

dynamics in riparian forests. Soil Sci. Soc. Am. J. 69, 729–737.enter for Geoinformatics, 2006. LSU Center 4 Geoinformatics, Louisiana State Uni-

versity, Baton Rouge, Louisiana, http://c4g.lsu.edu/joomla/onner, W.H., Brody, M., 1989. Rising water levels and the future of southeastern

Louisiana swamp forests. Estuaries 12, 318–323.

Page 9

ical E

C

C

C

D

D

D

E

E

F

G

H

H

J

J

JK

K

K

K

K

L

L

C

L

L

M

M

O

R

S

S

S

S

S

T

T

T

B.A. Middleton, M. Jiang / Ecolog

osta-Pierce, B.A., Weinstein, M.P., 2002. Use of dredge material for coastal restora-tion. Ecol. Eng. 19, 181–186.

roft, A.L., Leonard, L.A., Alphin, T.D., Cahoon, L.B., Posey, M.H., 2006. The effectsof thin layer sand renourishment on tidal marsh processes: Masonboro Island,North Carolina. Estuar. Coasts 29, 737–750.

ubaco, S., Santos, R., Duarte, C.M., 2008. The impact of sediment burial and erosionon seagrasses: a review. Estuar. Coast. Shelf Sci. 7, 354–366.

aubenmire, R., 1959. A canopy-coverage method of vegetational analysis. North-west Sci. 33, 43–64.

ay, J.W., Christian, R.R., Boesch, D.M., Yánez-Arancibia, A., Morris, J., Twilley, R.R.,Naylor, L., Schaffner, L., Stevenson, C., 2008. Consequences of climate change onthe ecogeomorphology of coastal wetlands. Estuar. Coasts 31, 477–491.

okka, R.K., 2006. Modern-day tectonic subsidence in coastal Louisiana. Geology 34,281–294.

llison, J.C., 1998. Impacts of sediment burial on mangroves. Mar. Poll. Bull. 37,420–426.

lsey-Quirk, T., Middleton, B., Proffitt, C.E., 2009. Seed dispersal and seedling emer-gence in a created and natural salt marsh on the Gulf of Mexico coast insouthwest Louisiana, USA. Restor. Ecol. 17, 422–432.

ord, M.A., Cahoon, D.R., Lynch, J.C., 1999. Restoring marsh elevation in a rapidlysubsiding salt marsh by thin-layer deposition of dredged material. Ecol. Eng. 12,189–205.

rosse, W., Frye, J., Lattermann, S., 1992. Root aeration in wetland trees by pressur-ized gas transport. Tree Phys. 10, 285–295.

ackney, C.T., Avery, G.B., Leonard, L.A., Posey, M., Alphin, T., 2007. Biological, chem-ical, and physical characteristics of tidal freshwater swamp forest of the LowerCape Fear River/Estuary, North Carolina. In: Conner, W.H., Doyle, T.W., Krauss,K.W. (Eds.), Ecology of Tidal Freshwater Forested Wetlands of the SoutheasternUnited States. Springer, Netherlands, pp. 183–221.

upp, C.R., Demas, C.R., Kroes, D.E., Day, R.H., Doyle, T.W., 2008. Recent sedimen-tation patterns within the central Atchafalaya Basin, Louisiana. Wetlands 28,125–140.

iang, M., Middleton, B.A., 2011. Soil characteristics of sediment-amendedcoastal baldcypress (Taxodium distichum) swamps in Louisiana. Wetlands 31,725–744.

olley, R.L., Lockaby, B.G., Cavalcanti, G.G., 2009. Productivity of ephemeral head-water riparian forests impacted by sedimentation in the southeastern UnitedStates coastal plain. J. Environ. Qual. 38, 965–979.

MP SAS, 2012. Statistical Analysis System, v. 10.0.2. Cary, North Carolina.eeland, B.D., Sharitz, R.R., 1993. Accuracy of tree growth measurements using

dendrometer bands. Can. J. Forest. Res. 23, 2454–2457.entula, M.E., 2002. Wetland Restoration and Creation, National Water Summary

on Wetland Resources, U.S. Geological Survey.

ozlowski, T.T., 1991. Soil aeration and growth of forest trees. Scand. J. Forest Res.

1, 113–123.rauss, K.W., Duberstein, J.A., Doyle, T.W., Conner, W.H., Day, R.H., Inabinette, L.W.,

Whitbeck, J.L., 2009. Site condition, structure, and growth of baldcypress alongtidal/non-tidal salinity gradients. Wetlands 29, 505–519.

Z

Z

ngineering 54 (2013) 183– 191 191

rauss, K.W., Duberstein, J.A., 2010. Sapflow and water use of freshwater wet-land trees exposed to saltwater incursion in a tidally influenced South Carolinawatershed. Can. J. Forest. Res. 40, 525–535.

angley, J.A., McKee, K.L., Cahoon, D.R., Cherry, J.A., Megonigal, J.P., 2010. ElevatedCO2 stimulates marsh elevation gain, counterbalancing sea-level rise. PNAS 106,6182–6186.

ee, I., Park, S., Ryu, S., Kobayashi, N., 2011. Ecological restoration index for evalua-tion of artificial salt marsh. J. Coast. Res. 27, 959–965.

oastal Protection and Restoration Authority of Louisiana (CPRA), 2012. Louisiana’sComprehensive Master Plan for a Sustainable Coast. Coastal Protection andRestoration Authority of Louisiana, Baton Rouge, LA.

ouisiana Department of Natural Resources (LDNR), 2012. Strategic Online NaturalResources Information System, Baton Rouge, Louisiana, http://sonris.com

und, Z.F., Pearson, R.W., Buchanan, G.A., 1970. An implanted soil mass techniqueto study herbicide effects on root growth. Weed Sci. 18, 279–281.

adrid, E.N., Quigg, A., Armitage, A.R., 2012. Marsh construction techniques influ-ence net plant carbon capture by emergent and submerged vegetation in abrackish marsh in the northwestern Gulf of Mexico. Ecol. Eng. 43, 54–63.

endelssohn, I.A., Kuhn, N.L., 2003. Sediment subsidy: effects on soil–plantresponses in a rapidly submerging coastal salt marsh. Ecol. Eng. 21, 115–128.

ksanen, J., 2012. Multivariate Analysis of Ecological Communities in R: Vegan Tuto-rial, www.cc.oulu.fi/jarioksa/opetus/metodi/vegantutor.pdf

Foundation, 2012. R. Version 2.15.1. The R Foundation for Statistical Computing,www.r-project.org

hafer, D.J., Streever, W.J., 2000. A comparison of 28 natural and dredged mate-rial salt marshes in Texas with an emphasis on geomorphological variables.Wetlands Ecol. Manage. 8, 353–366.

hinkle, K.D., Dokka, R.K., 2004. Rates of Vertical Displacement at Benchmarks in theLower Mississippi Valley and the Northern Gulf Coast. NOAA Technical ReportNOS/NGS 50. NOAA, Silver Springs, MD.

hrift, A.M., Mendelssohn, I.A., 2008. Salt marsh restoration with sediment-slurryamendments following a drought-induced large-scale disturbance. Wetlands28, 1071–1085.

tagg, C.L., Mendelssohn, I.A., 2010. Restoring ecological function to a submergedsalt marsh. Restor. Ecol. 18, 10–17.

locum, M.G., Mendelssohn, I.A., Kuhn, N.L., 2005. Effects of sediment slurry enrich-ment on salt marsh rehabilitation: plant and soil responses over seven years.Estuaries 28, 519–528.

urner, R.E., 1997. Wetland loss in the northern Gulf of Mexico: multiple workinghypotheses. Estuaries 20, 1–13.

urner, R.E., Streever, B., 2002. Approaches to Coastal Wetland Restoration: North-ern Gulf of Mexico. SPB Academic Publishing, Hague, The Netherlands.

urner, R.E., Baustian, J.J., Swenson, E.M., Spicer, J.S., 2006. Wetland sedimentation

from Hurricanes Katrina and Rita. Science 314, 449–452.

edler, J.B., 2000. Progress in wetland restoration ecology. Trends Ecol. Evol. 15,402–407.

edler, J.B., Callaway, J.C., 1999. Tracking wetland restoration: do mitigation sitesfollow desired trajectories? Restor. Ecol. 7, 69–73.