reducing the effects of dredged material levees on coastal ... › cv › peterson.mark › docs ›...

Reducing the Effects of Dredged Material Levees onCoastal Marsh Function: Sediment Deposition andNekton Utilization

DENISE J. REED*

Department of Geology and GeophysicsUniversity of New OrleansNew Orleans, Louisiana 70148, USA

MARK S. PETERSON

BRIAN J. LEZINA

Department of Coastal SciencesThe University of Southern Mississippi703 East Beach Drive, Ocean SpringsMississippi 39564, USA

ABSTRACT / Dredged material levees in coastal Louisianaare normally associated with pipeline canals or, more fre-quently, canals dredged through the wetlands to allowaccess to drilling locations for mineral extraction. Thehydrologic impact on marshes behind the levee is of con-cern to coastal resource managers because of the potentialimpact on sediment transport and deposition, and the effecton estuarine organism access to valuable nursery habitat.

This study examined the effects of gaps in dredged materiallevees, compared to continuous levees and natural channelbanks, on these two aspects of marsh function. Field studiesfor sediment deposition were conducted biweekly for a year,and nekton samples were collected in spring and fall. Vari-ation in nekton density among study arears and landscapetypes was great in part because of the inherent samplinggear issues and in part because of differences in charac-teristics among areas. Nekton densities were generallygreater in natural compared to leveed and gapped land-scapes. Differences in landscape type did not explain pat-terns in sediment deposition. The gaps examined appear tobe too restrictive of marsh flooding to provide efficientmovements of floodwaters onto the marsh during moderateflooding events. The ‘‘trapping’’ effect of the leveesincreases sediment deposition during extreme events.Gapping material levees may be an effective method ofpartially restoring upper marsh connection to nekton, but thismethod may work best in lower elevation marshes wherenekton use is greater.

The importance of hydrology for many coastal wet-land functions is well established (e.g., Mitsch andGosselink 2000; Carter 1984), and alterations to tidalexchange, due to diking, transportation embankments,and so on frequently contribute to marsh deterioration(Burdick and others 1997; Hood 2004, Roman andothers 1984). In Louisiana, one of the most commondisruptions of marsh surface hydrology is associatedwith dredged materials levees resulting from excava-tion in past decades (Turner 1987). These levees arenormally associated with pipeline canals or, more fre-quently, canals dredged through the wetlands to allowaccess to drilling locations for oil and gas extraction.The effect of these unnatural, upland, linear, and fre-quently intersecting, features on marsh hydrology hasbeen well documented (e.g., Swenson and Turner

1987; Turner and Cahoon 1988). Studies have alsoexamined the restoration potential of backfilling thecanals with the dredged material even years after theinitial impact (Turner and others 1994; Reed andRozas 1995). These studies show that insufficientmaterial usually remains to infill the canal completelyand, although potentially productive shallow waterhabitats may result, marshes are only regained in theareas from which the dredged material levee is re-moved.

However, the continuing hydrologic impact onmarshes behind the levee is of concern to coastal re-source managers for several reasons. The potential ofthe levees to limit sediment transport to and deposi-tion on the marsh surface may reduce vertical marshaccumulation (DeLaune and others 1990). This couldseverely affect marsh sustainability because of locallyhigh rates of subsidence (Penland and Ramsey 1990;Morton and others 2002). In addition, the levees affectestuarine orgainsm access to the marsh surface, widelyrecoginized as valuable nursery habitat (Beck andothers 2001). The goal of this project was to evaluate afrequently suggested (LCWCRTF 1998) but rarely tes-

KEY WORDS: Coastal Louisiana; Nekton utilization; Sedimentation;Marshes; Dredged material; Hurricane

Published online February 20, 2006.

*Author to whom correspondence should be addressed; email:

[email protected]

Environmental Management Vol. 37, No. 5, pp. 671–685 ª 2006 Springer Science+Business Media, Inc.

DOI: 10.1007/s00267-004-0223-6

ted approach to minimizing the effects of existingdredged material levees on coastal marshes: makinggaps in them.

The rationale for gapping these levees is thatincreasing the hydrologic exchange between the marshsurface and the canal would alleviate some of the det-rimental effects of a continuous levee. The breachingof levees to improve tidal exchange has been used inmany areas as a restoration technique (Roman andothers 2002; Tanner and others 2002; Williams and Orr2002). The dredged material levels in Louisiana fre-quently originated as disposal areas for dredgedmaterial and frequently material was plased irregularly,according to changing environmental conditions ordredge operations. Consequently, many of these leveesare not continuous and include ‘‘gaps.’’ These areareas at approximately marsh elevation, 10–20 m inwidth and devoid of the shrub-scrub vegetation thatcharacterizes the adjacent levee section. This studysought to learn from these existing irregularities in thelevees to inform coastal managers of likely differencesin marsh function associated with creating such ‘‘gaps’’in the future.

Previous evaluations of this technique (Turner andothers 1994) have used theoretical approaches todetermine the value of the resulting restoration. Rec-ognizing the limitations of their approach, Turner andothers suggest that implementation should proceed byremoving small sections of levee and monitoring toevaluate the effects prior to removing more of thebarrier. Given the extensive permitting and contractingprocedures required to actually dredge or fill withincoastal wetlands, preliminary insight into potentialbenefits associated with this restoration approach canbe gained from field studies of these existing gapswithin levees.

The project addressed two working hypotheses:

H1. Marsh surface sediment deposition is greater insalt marshes behind dredged material levees withgaps than in marshes fronted by continuous levees.

H2. Nekton density in salt marshes behind dredgedmaterial levees with gaps is greater than in marshesfronted by continuous levees.

In addition, the study evaluated the function ofmarshes that have not been directly impacted by leveeplacement for comparison with the gapped and con-tinuous levee sites.

This article presents the results of a 1-year field ef-fort focused on addressing these hypotheses and pro-vides guidance for coastal resource managers onfurther steps regarding the management of tidal mar-

shes impacted by levees, dikes or other hydrologicbarriers.

Experimental Design and Study Area

Field studies addressed the hypotheses by compar-ing marsh surface sediment deposition and nektondensity in interior salt marshes of the Mississippi del-taic plain. Interior marshes are here defined as thoseaway from natural bayou margins or behind dredgedmaterial levees. Three types of salt marsh landscapetypes were examined: marshes fronted by natural bay-ous or channels, marshes fronted by continuousdredged material levees, and marshes fronted by dis-continuous levees with natural or manmade gaps. Ingapped and leveed landscapes, marshes impacted ononly one side by levees were selected so as not toconfound sampling with the problem of semi-impoundment. Distance from channelized open water(e.g., bayou or canal) was the main factor controllingactual sampling locations within these landscapestypes. All measurements were made on the marshsurface, not in ponds or small channels, because veg-etated marsh is the target of most restoration andmitigation activities in Louisiana and the environmentwhere levee effects on hydrology have been most fre-quently documented.

After inspection of aerial photography and a fieldsurvey, three geographic areas (A, B, C) were selectedfor sampling (Figure 1), each including examples ofthe three landscape types described above. The land-scape types were grouped within the areas so as tominimize variation in the basin hydrologic gradientamong landscape types. These areas are within theTerrebonne hydrologic basin of coastal Louisiana andare all located within the salt marsh zone (Chabreckand Linscombe 1988). Within each area and landscapetype, three sampling sites for marsh surface sedimentdeposition were randomly selected, for a total of 27,and small platforms were constituted about 20 m fromthe nearest water body to prevent disturbance of themarsh surface adjacent to sampling sites. Nekton sam-pling sites (10 samples within each landscape type perseason) were randomly chosen within each of the threeareas, and these stations were at least 100 m away fromthe channel and into the marsh.

Methods

Biweekly measurements of marsh surface depositionwere taken using sediment traps (Reed 1992)beginning in October 2001 and ending in October

672 D. J. Reed and others

2002. Sediment traps were deployed in groups of six ateach sampling platform and were renewed every 2weeks. All filter papers used for the traps were pre-weighed prior to deployment, and upon return to thelaboratory, each trap was dried in an oven (60�C for 24hrs) and reweighed. Occasionally the traps were dis-turbed and part of the 63.6 cm2 filter paper was lost. Inthese instances, a transparent grid placed on the filterallowed estimation of the percentage of the area lost.Both pre- and postdeployment weights were adjusted toaccount for the change in area, and data are reported asg/cm2/day based on the length of deployment of eachset of traps.

Experience with this technique in previous studies(e.g., French and others 1995, Hutchinson and others1995) has demonstrated that sediment deposition dataare highly variable, but it responds to both landscape

scale processes as well as local hydrology. Thus, for thisstudy, sediment deposition data were analyzed on thepooled area data using analysis of variance (ANOVA)following log10 transformation, with landscape type asthe main effect. The results of the ANOVA were subjectto a Tukey multiple range test with P < 0.05 designatinga statistically significant difference in sediment depo-sition among landscapes.

In May and September 2002, nekton and physical–chemical data were collected in the upper marshwithin the three landscape types distributed among thethree areas. The order of sampling each area wasdetermined randomly, whereas the order of samplingeach landscape type within area was determined basedon tidal stage. At each area and landscape type, sam-pling commenced about 100 m from the marsh edgeand away from sediment trap sites. Nekton were hap-

Figure 1. Areas and sampling sites within areas based on landscape type. Open circles = natural channel bank, solid cir-cles = continuous levee, striped circles = gaps.

Gapping Levees to Improve Marsh Function 673

hazardly sampled at each site with a 1-m2 · 0.75-m-highaluminum throw trap constructed of a 2.54-cm angle(Jordan and others 1997). The sides of the trap wereconstructed of 3.17-mm ‘‘ace’’ netting. Within each ofthe landscape types, 10 throw trap collections weremade that were generally thrown in a previouslydetermined random compass direction, except whenapproaching sediment traps, a previous sample site, ora small creek or pond. If any of these conditions oc-curred, 180� was added to the compass heading priorto trap deployment. Five paces separated successivethrow trap sites. Once thrown, the trap was quicklypushed into the marsh grass/sediment interface. Sub-sequently, the trap was seated into the sediment withthe use of a flat-bladed metal spade. Once the trap waswell seated, nine physical–chemical variables weremeasured but only salinity (psu) and plant stem density(number/0.25 m2) were retained for comparison (seeReed and others 2004 for details). After plant stemswere cut and cleared, nekton were removed from thetrap with a bar seine (3.17-mm mesh) in five sequentialseine hauls proceeding along all four sides of the trap.Fishes and invertebrates were collected and preservedin 10% buffered formalin. Once in the laboratory, allindividuals were identified to the lowest possible taxon.

Nekton density was compared with a two-way ANOVAwithin season, with area and landscape type as the maineffects. If there was no significant interaction term,(P < 0.05), and a significant F value, means were sepa-rated with a Sidak pairwise comparison test. If a signifi-cant interaction term was noted, we used individual one-way ANOVAs and Sidak pairwise comparisons by areaand landscape type to interpret the trends in the data.Because of the number of ‘‘Zero’’ catches for densityestimates, we transformed density data using log10 (x +0.5) prior to analysis (Underwood 1997). Values are re-ported in non transformed format for clarity. All testswere performed with SPSS (Version 11.5) and signifi-cance was determined when P £ 0.05.

Results

Landscape Characterization

Only plant stem density and salinity were comparedamong area and landscape type because they wereconsidered unlikely to change over the course of thenekton sampling, or appeared to be related to positionalong the estuarine gradient.

There was no significant interaction between areaand landscape type in May for plant stem density, norwas there a significant landscape type difference.However, stem density was significantly differentamong areas. Stem density (0.25 m2) was greater atareas A (mean = 153.7) and B (mean = 153.4) than C(mean = 36.7), with no difference between A and B(Table 1). Plant composition within area C was mainlySpartina alterniflora whereas within areas A and B, itwas dominated by Spartina patens, the difference inplant morphology accounting for the differences instem density. In September, however, there was asignificant interaction effect between landscape typeand area in plant stem density. Density in area A(mean = 581.6) was greater than B (mean = 225.3) orC (mean = 69.4), and area B density was greater thanC (Table 1). Within area A, density in the naturallandscape (mean = 279.7) was less than leveed(mean = 766.4) and gapped (698.6). No differencewas found between gapped and leveed landscapes.Within area C, natural landscape density(mean = 47.0) was lower than gapped (mean = 98.2)(Table 1). There was no difference between the nat-ural and leveed (mean = 63.0) landscapes, but theleveed landscape had marginally lower stem densitythan the gapped. Although a significant differencewas noted in area B, pairwise comparisons were notable to separate the means (Table 1).

In May, there was a significant salinity interactionterm between area and landscape type. Salinity wassignificantly higher at the southern area (C,

Table 1. Comparisons of stem density and salinity by month for areas and landscapes

Species Month Area Significance Landscape

Stem density May A = B > C P < 0.01(#/0.25 m2) September A > B > C P < 0.01

A:N < L = GC: N < G; N = L; L < GB:a

Salinity (psu) May C > B = Ab P < 0.01September

A: N < L; N < G; L < G

aOverall difference but Sidak did not separate means values.bIn September, the salinometer broke after collecting data in area A.

N = natural, L = leveed, G = gapped.


mean = 13.9 psu) and decreased towards the north (B,mean = 11.3; mean = 9.5). In September, the sali-nometer broke after collecting data in area A; thus,only landscape comparisons could be made at thatarea. In area A, salinity was significantly lower in nat-ural (mean = 8.0) compared to leveed (mean = 9.2)and gapped (mean = 12.4) landscapes. Salinity at lev-eed landscape types was lower than that at gapped.

Sediment Deposition

The year of biweekly sampling of marsh surfacesediment deposition was designed to encompass the

range of hydrologic and meteorological conditionscommon in southeast Louisiana. There was great vari-ability in the magnitude of the sediment depositionrecorded on the traps among biweekly samplingintervals, even without the influence of Hurricane Lillion the last sampling interval in October 2002 (Fig-ure 2). In general, low rates of sediment depositionwere found in all landscape types during early 2002,and again during the summer and early fall. Slightlyhigher rates were measured in fall 2001 and latespring/early summer 2002. Also early in the data col-lection period, the natural landscapes were experi-

Figure 2. A plot of pooled sediment deposition rates (mean ± SE) for biweekly sampling periods October 2001–October 2002.

Table 2A. Sediment deposition by landscape type pooled by area without the effects of deposition by HurricaneLilli

Treatment Mean g/m2/day SE N ANOVA P Range test

Natural channel 0.00035231 0.00006554 24 ALeveed 0.00020921 0.00005080 24 0.021 ABGaps 0.00015181 0.00003051 24 B

Table 2B. Sediment deposition by landscape type pooled by area for all collected data

Treatment Mean g/m2/day SE N ANOVA P Range test

Natural channel 0.00040942 0.00008490 25 ALeveed 0.00126000 0.00105000 25 0.0483 AGaps 0.00032758 0.00017819 25 A


encing higher rates of sediment deposition than thoseinfluenced entirely or in part by dredged material le-vees, but this difference was less evident throughoutthe remainder of the monitoring period (Figure 2).When data through September 2002 were analyzed,natural channel landscapes had greater sedimentdeposition than gapped, but there was no statisticaldifference (P > 0.05) between the gaps and the leveesor the natural channels and the levees (Table 2A).When Hurricane Lilli storm data were included in theanalysis, the great increase in variability and the veryhigh sediment deposition rates at the levee siteschange the results such that no statistical difference in

sediment deposition among the landscape types can beidentified (Table 2B).

Nekton Sampling

A total of 170 throw trap collections were made (80 inMay and 90 in September). It was not possible to samplethe leveed landscape within area B in May because ofinsufficient water. A total of 1512 individuals were col-lected from eight families of fishes and five families ofinvertebrates. Collections totaled 516 and 996 individu-als for may and September, respectively. In May, 64.9%

of all individuals were collected in natural compared to16.5% at the gapped and 18.6% at the leveed landscape

Table 3. Density (#/m2) of nekton collected by area and landscape type from the 25-27 May 2002 collections

Area A Area B Area C

Families/species Natural Gap Levee Natural Gap Levee Natural Gap Levee

Engraulidae NWAnchoa sp. — 0.1 ± 0.3 — — — — — —

Cyprinodontidae NWAdinia xenica 2.4 ± 1.4 0.9 ± 0.7 0.6 ± 0.7 0.8 ± 1.0 — 2.2 ± 2.1 1.5 ± 1.1 3.8 ± 3.5Cyprinodon variegates 0.4 ± 1.0 03 ± 0.7 — 0.8 ± 1.1 — 1.2 ± 1.5 0.1 ± 0.3 0.2 ± 0.6Lucania parva — — — — — — 0.1 ± 0.3 0.1 = 0.3

Fundulidae NWFundulus jenkinsi — — — — 0.1 ± 0.3 0.1 ± 0.3 — —Fundulus grandis 0.5 ± 1.0 — 0.3 ± 0.7 — — — 0.4 ± 0.7 0.2 ± 0.4 0.5 ± 0.7Fundulus pulvereus 2.8 ± 2.2 0.4 ± 0.5 1.9 ± 1.9 1.8 ± 2.8 0.9 ± 1.3 0.1 ± 0.3 0.2 ± 0.4 0.7 ± 0.8Fundulus similis — — — — — — 0.1 ± 0.3 0.2 ± 0.4

Poeciliidae NWGambusia affinis — — 0.1 ± 0.3 0.2 ± 0.4 — — — —Poecilia latipinna — — 0.1 ± 0.3 0.3 ± 0.5 — — — —

Atherinidae NWMenidia beryllina — — — — — 0.3 ± 0.7 — —

Mugilidae NWMugil cephalus — — — — — 0.7 ± 1.6 — —

Eleotridae NWDormitator maculatus 0.1 ± 0.3 — — — — — — 0.1 ± 03

Palaemonidae NWPalaemonetes pugio 0.1 ± 0.3 — — — — 15.7 ± 9.1 1.3 ± 1.0 0.3 ± 0.5Palaemonetes vulgaris — — — — — 0.6 ± 1.3 — —

Penaeidae NWFarfantepenaeus aztecus — — — — — — 0.1 ± 0.3 —

PortunidaeCallinectes sapidus — — — 0.2 ± 0.4 0.1 ± 0.3 0.2 ± 0.4 — 0.3 ± 0.5

NW

Ocypodidae Uca spp. 0.5 ± 1.1 03 ± 0.5 0.6 ± 0.8 0.2 ± 0.4 0.5 ± 0.8 — 0.7 ± 0.8 0.9 ± 1.8 0.1 ± 0.3

Grapsidae Sesarma spp. 0.1 ± 0.3 — — — — — 0.1 ± 0.3 0.1 ± 0.3 —

NW = no water.

Values are presented as mean ± SD (n = 10 collections each).


types. September samples showed a similar pattern with52.8% of the individuals taken in the natural versus17.1% in the gapped and 29.1% in the leveed landscapes.

Collections of those species representing ‡5% of thetotal individuals sampled within a season were subjectto more detailed analysis. In May, the species examinedwere Adinia xenica, Fundulus pulvereus, Cyprinodonvariegatus, and Palaemonetes pugio. In September, inaddition to the above species, Gambusia affinis andPoecilia latipinna were also examined. A clear seasonaleffect in species abundance was evident, and thus datawere not pooled across season for analysis.

May collections. There were significant interactionterms between area and landscape type for densities ofA. xenica, F. pulvereus, and P. pugio (Table 3; Figures 3–7). There was no difference in density of A. xenica innatural landscapes among areas. In leveed landscapes,however, A. xenica density at area C > A. In gappedlandscapes, density in area A was greater than B, but Awas equal to C, and density at C was greater than that atB (Tables 3 and 4). In areas A and B, density was greater

in natural versus gapped or leveed, but density was notdifferent between gapped and leveed (Figure 3). Therewas no difference among landscapes within area C.

For F. pulvereus, there was a significant differencein density in natural landscape among areas, withdensity in A = B and B = C, but density in A > C. Inleveed and gapped landscapes, there were no differ-ences in density across areas (Tables 3 and 4). Withinarea A, F. pulvereus density was greater in naturalversus gapped, but no other differences were found(Figure 4A).

For P. pugio, there was a significant difference indensity in natural landscape types among areas, withdensity in (A = B) < C. Among leveed landscapes, noarea effect was noticed. In gapped landscapes, densitieswere C > (A = B) (Tables 3 and 4). Areas A and B didnot show a landscape effect (Figure 5A). Within areaC, natural landscape, density was greater than at gap-ped and leveed. A plot of the mean density of P. pugioby area and landscape type illustrated that the inter-action was caused by the collection of 99.5% of allindividuals in the natural marsh landscape at area C.

There was no significant interaction term betweenarea and landscape type for C. variegatus, nor was therean area effect. However, density was always greater inthe natural landscape versus the gapped or leveedlandscape regardless of area (Table 3), with no differ-ence between gapped or leveed landscapes.

September collections. Significant interaction termsbetween landscape and area existed for P. latipinna,P. pugio, C. variegatus, F. pulvereus, and G. affinis. ForC. variegatus, density in natural landscapes was (A <B) = C. In leveed landscapes, density in area A = B,with density in C being greater than both (Tables 5and 6). No C. variegatus were collected in area A, andno analysis was performed. However, in area B den-sity in natural landscape was greater than that inleveed or gapped, and at area C, there was higherdensity in the leveed landscape versus the gapped(Figure 6).

There was a significant difference in density ofP. latipinna in natural and leveed landscapes amongareas, with density at area B > (A = C). In gappedlandscapes, there were no area differences noted(Tables 5 and 6). Within areas A and C, there were noP. latipinna density differences among landscapes.However, in area B, density was greater at the naturaland leveed landscapes versus gapped (Figure 7A).

For G. affinis, there was no significant difference indensity in natural landscapes among areas. However, inleveed landscapes, density A > (B = C), and in thegapped landscape type, density in area A > (C > B)

Figure 3. Plot of mean (± 1 SE) landscape type density forAdinia xenica within locations in May. * = no significant dif-ference. Note: Collections were not made at the leveedlandscape type within area B. Circle = natural, square = levee,diamond = gapped landscapes.


Figure 4. Plot of mean (± 1 SE) landscape type density for Fundulus pulvereus within locations for May (A) and September (B).* = no significant difference. Note: Collections were not made at the leveed landscape type within area B.

Figure 5. Plot of mean (± 1 SE) landscape type density Palaemonetes pugio within areas for May (A) and September (B).N/A = no test conducted.


(Tables 5 and 6). Within area A, G. affinis density ingapped and leveed landscapes was greater than densityin natural landscapes. In area B, there was no differ-ence in density among landscapes (Figure 7B), and noindividuals were collected in area C.

For P. pugio, there was a significant difference indensity in natural landscapes, with density in area(A = B) < C. No individuals were collected in any of theleveed landscapes, but in contrast, density in the gap-ped landscapes was not different among areas(Tables 5 and 6). In area C, density at leveed andgapped landscapes was less than density in natural(Figure 5B).

For F. pulvereus, there was a significant difference indensity in natural landscape among areas, with densityin area (A = B) > C. Although there was no area effectin the leveed landscape, gapped landscape density inarea A = (C < B) (Tables 5 and 6). In areas A and B,F. pulvereus density in natural landscapes was greaterthan those in leveed and gapped (Figure 4B). In areaC, there was no difference in density among land-scapes.

Discussion

Both the sediment deposition and nekton denisitydata show complex patterns among landscape typesand changes over season or throughout the year. Thenatural variability of tidal marsh processes and themultiple influences on the two main factors addressedin this study frequency produced more questions thananswers regarding the controls on marsh function.However, the results presented here can provide in-sight into the hypotheses the study was designed toaddress.

Hypothesis 1: Marsh Surface Sediment DepositionIs Greater in Salt Marshes Behind Dredged MaterialLevees with Gaps than in Marshes Fronted byContinuous Levees

The data show that this condition was rarely met.On only two of the sampling periods was meansediment deposition at the gapped landscapes greaterthan the leveed landscapes. The high temporal vari-ability in sediment deposition at any one landscapetype shown in Figure 2 is typical of Louisiana coastalmarshes (e.g., Kuhn and others 1999) However, pre-vious studies that have used this technique to assess theeffect of hydrologic alteration on marsh surface sedi-ment deposition have found significant differencesusing data records of 1 year or less (Reed 1992, Reedand others 1997, Kuhn and others 1999). Because thisstudy finds no such difference between gapped andleveed landscapes, it seems clear that the gaps assessedhere do not result in increased sediment depositionunder nonstorm conditions.

In addition, the sediment deposition data provideinsight into the factors influencing temporal variationsin marsh sedimentary processes. The effect of Hurri-cane Lilli on sediment deposition is marked. Thisreiterates previous studies of storm-related sedimenta-tion in the Terrebonne basin during the passage ofHurricane Andrew. Cahoon and others (1995) mea-sured pronounced increases in marsh surface sedimentdeposition, using the same sediment trap technique,during measurement periods that included storm pas-sage. However, examination of Figure 2 shows thatduring the year of sediment deposition measurements,there are distinct periods of high deposition, in alllandscape types during late Fall 2001 and late Spring2002. This seasonal variation in sediment deposition isgenerally greater in magnitude than the differences insediment deposition among landscape types. Waterlevel data collected at Cocodrie were averaged for thecollection periods and show distinct variation duringthe study period (Figure 8). These data can be used to

Figure 6. Plot of mean (± 1 SE) landscape type density forCyprinodon variegatus within areas for September. N/A = notest conducted.


assess the seasonal patterns in water level at the studysite, despite the distance to the gauge (�20 km). Waterlevels in Louisiana estuaries are strongly influenced bymeteorological conditions (Schroeder and Wiseman1986), and the seasonal variations shown in Figure 8likely reflect regional effects (Marmer 1954) ratherthan influences specific to Cocodrie. Late Fall 2001and late Spring 2002 are both periods of increasedwater levels, especially when compared to early 2001.Water levels remained high during early summer 2002but sediment deposition rates were less. Water levelsincrease dramatically in October 2002 with the passageof Tropical Storm Isidore and Hurricane Lilli.

These seasonal patterns in sediment deposition andat times similar variation in water level, combined withthe analyses of the differences among landscape types(both with and without the effects of Hurricane Lilli),provide important insights into the factors controllingsedimentation across this part of the Terrebonne Basin.Previous studies (Reed 1989) have shown that the pas-sage of winter cold fronts elevates both marsh waterlevel and marsh surface sediment deposition and datafrom Fall 2001 collected during this study are consistentwith such an influence. Sediment availability to thesemarshes can also increase in the Spring as Atchafalaya

River waters penetrate the Terrebonne Basin via a net-work of bayous and canals (Reed and DeLuca 2003).Seasonally increased suspended sediments combinedwith high water levels (Figure 8) would also promotethe combination of sediment availability and opportu-nity for sediment deposition on the marsh surface seenby Reed (1989) as controls on marsh sediment deposi-tion. Hurricanes promote both availability and oppor-tunity by dramatically increasing water level andremobilizing sediments from coastal bays and offshore.However, water levels are so elevated (exceeding 3 mabove Cocodrie datum for a period of 8 hrs during thepeak of Lilli) that basin-scale topography and physiog-raphy becomes more important than the gradientswithin marshes influenced by channels and ponds.Bayous and channels likely become less important asconduits for water and sediments than deep sheetflowand wave action. The data and analysis presented inTables 2A and 2B show a greater increase in sedimentdeposition associated with the storm at the leveed andgapped landscapes than the natural channel land-scapes. During these high magnitude events, the leveesare associated with increased sediment deposition,suggesting that the elevation of the levees, and theirshrub-scrub and tree vegetation, interacts with the

Figure 7. Plot of mean (± 1 SE) landscape density within area for Poecilia latipinna (A) and Gambusia affinis (B) in Septembersamples. * = no significant difference; N/A = no test conducted.


Table 4. Comparison of within-habitat density (no./m2) across areas for May samples having a significant area–landscape interaction

Species Landscape Area Significance

Adinia xenica Natural A = B = C nsLeveeda C > A P < 0.01Gapped A > B; C > B All P < 0.05

A = C nsFundulus pulvereus Natural A > C P < 0.05

B = (A and C) nsLeveeda A = C nsGapped A = B = C ns

Palaemonetes pugio Natural C > (A and B) All P < 0.05A = B ns

Leveeda A = C nsGapped C > (A and B) All P < 0.05

A = B ns

ns = nonsignificant.aNo collections made at area B.

Table 5. Density (no./m2) of nekton collected by area and landscape type from the 4–6 September 2002collectionsa

Area A Area B Area C

Families/species Natural Gap Levee Natural Gap Levee Natural Gap Levee

CyprinodontidaeAdinia xenica 3.2 ± 3.0 1.6 ± 1.4 1.0 ± 0.7 4.3 ± 2.7 1.4 ± 1.7 4.8 ± 2.5 2.3 ± 2.8 2.0 ± 2.0 3.0 ± 2.6Cyprinodon variegatus — — — 7.3 ± 11.0 — 0.1 ± 0.3 0.9 ± 0.9 0.5 ± 1.3 2.2 ± 2.3

FundulidaeFundulus jenkinsi 0.2 ± 0.6 — — — 0.1 ± 0.3 0.4 ± 1.3 0.1 ± 0.3 0.3 ± 0.5 —Fundulus grandis 0.1 ± 0.3 — 0.1 ± 0.3 — — — 0.5 ± 0.5 0.3 ± 0.7 0.5 ± 1.3Fundulus pulvereus 3.7 ± 2.5 0.5 ± 0.5 1.0 ± 0.7 5.2 ± 3.2 1.9 ± 1.9 1.3 ± 1.2 0.7 ± 1.1 0.4 ± 0.7 1.1 ± 1.4Fundulus simills — — — — — — — 0.5 ± 0.7 0.1 ± 0.3

PoeciliidaeGambusia affinis 0.2 ± 0.6 1.7 ± 1.8 3.8 ± 2.9 0.2 ± 0.4 0.2 ± 0.4 0.7 ± 1.3 — — —Poecilia latipinna 1.1 ± 1.2 0.6 ± 0.7 0.3 ± 0.5 8.6 ± 6.9 1.6 ± 23 6.6 ± 6.3 1.7 ± 1.6 1.6 ± 2.2 0.3 ± 0.5

AtherinidaeMenidia beryllina — — — — — — 0.2 ± 0.4 — —

MugilidaeMugil cephalus — — — — — — 0.2 ± 0.6 0.3 ± 0.7 —

EleotridaeDormitator muculatus — — — 0.1 ± 0.3 — 0.1 ± 0.3 — — —

SciaenidaeCynoscion nebulosus — — — — — — 0.2 ± 0.4 — —

PalaemonidaePalaemonetes pugio — — — — — — 8.4 ± 7.3 0.1 ± 0.3 —

PenaeidaeLitopenaeus setiferus — — — — — 0.8 ± 1.6 — —

PortunidaeCallinectes sapidus 0.1 ± 0.3 — — — — — 2.9 ± 2.0 1.0 ± 1.3 0.2 ± 0.4

OcypodidaeUca spp. 0.2 ± 0.4 0.1 ± 0.3 0.5 ± 0.9 — 0.3 ± 0.5 0.9 ± 0.9 0.1 ± 0.3 — —

GrapsidaeSesarma spp. 0.1 ± 0.3 — — — — — — — —

Values are presented as mean ± SD (n = 10 collections each).


storm sheetflows, slowing the flows and consequentlyincreasing sediment deposition. During regular tidalinundation and the moderate increases in water levelassociated with meteorological fronts, the levees act toblock flows and sediment transport from the canals andchannels to the marsh surface. In contrast, whensheetflow across the surface of the marsh complex isdominant under hurricane conditions, levees act as‘‘roughness’’ elements to baffle flow and enhance sed-iment deposition.

The specific effect of tropical storms and hurricanesis impossible to predict, but the results of this studysupport the contention that the elevation of Louisianacoastal marshes isolated from riverine sediments ismaintained by fronts and storms (Reed 2002). More-over, the insights gained here into the effects oflandscape-scale marsh topography on sedimentationpatterns suggests that higher magnitude–lower fre-quency events such as hurricanes can result in sedi-ment deposition in areas not as influenced by

Table 6. Comparison of within-habitat density (no./m2) across areas for September samples having a significantarea–landscape interactiona

Species Landscape Area Significance

Poecilia latipinna Natural B > (A and C) All P < 0.05A = C ns

Leveed B > (A and C) All P < 0.05A = C ns

Gapped A = B = C ns

Palaemonetes pugio Natural C > (A and B) All P < 0.05A = B ns

Leveed —Gapped A = B = C ns

Cyprinodon variegatus Natural B > A P < 0.05C > (A and B) ns

Leveed C > (A and B) All P < 0.05A = B ns

Gapped A = B = C ns

Fundulus pulvereus Natural C < (A and B) All P < 0.05A = B ns

Leveed A = B = C nsGapped B > C P < 0.05

A = (B and C) ns

Gambusia affinis Natural A = B = C nsLeveed A > (B and C) All P < 0.05

B = C nsGapped A > B All P < 0.05

A > CC > B

ans = nonsignificant; - = equal densities.

Table 7. Frequency of occurrence of species density (no./m2) by month when gapped landscape types werecompared to levee and natural typesa

May September

A. xenica F. pulvereus P. pugio P. latipinna F. pulvereus P. Pugio G. affinis C. variegatus

G > N 25G < N 40 20 50 16.7 40 50 25G = N 20 40 33.3 20 25 25G > LG < L 20 16.7G = L 40 20 50 33.3 40 50 50 50Number of possible comparisons 5 5 2 6 2 5 4 4

aG = gapped, L = levee, and N = natural landscape types. Data extracted from Figures 3–7.


moderate magnitude–higher frequency events such ascold fronts. Whether such sedimentation contributesto the sustainability of marshes influenced by dredgedmaterial levees depends on whether the resultant in-creased or maintained marsh elevation can counteractpotential ponding and waterlogging effects (e.g.,Swenson and Turner 1987) induced by the levees effecton marsh drainage.

Hypothesis 2: Nekton Density in Salt MarshesBehind Dredged Material Levees with Gaps isGreater Than in Marshes Fronted by Continuouslevees

Many of the nekton species collected in this studyare marsh residents, and are by definition more toler-ant of variability in abiotic conditions, and extensivelyuse marsh edge surface during all or part of their lifehistory (Rozas and Reed 1993, Fulling and others 1999,Rozas and Zimmerman 2000). In general, the nektondensities seemed more dependent on local area char-acteristics than on the landscape types examined inthis study. The slight difference in vegetation amongthe study areas appeared to influence nekton densitiesmore than landscape type within an area.

The nekton density data are more complex than thesediment data, but for most of the species/area/seasonsampling shown in Figures 3–7 where differences wereidentified, greater densities of nekton were found inthe natural landscapes compared to the leveed orgapped. This pattern is illustrated for four of theindividual species shown (Fundulus pulvereus, Palae-

monetes pugio, Cyprinodon variegatus, and Poecilialatipinna) in Tables 3 and 5. Only Gambusia affinisduring the September sampling showed higher densi-ties in the leveed and gapped landscapes compared tonatural landscapes. This may be a consequence of thevery small size of the species, its livebearing reproduc-tive life history (Brown-Peterson and Peterson 1990),and its being more able than the other species to sur-vive on the marsh surface in small ponds. These attri-butes are likely especially important for individuals whoare stranded on leveed landscapes at low tide due toinefficient surface drainage.

In contrast, Adinia xenica was fairly evenly distrib-uted across all areas and landscape types. This distri-bution pattern is typical of a number of marshresident nekton and reflects its association with avariety of subhabitats within Louisiana coastal marshesand its physiological capabilities to do well in harshenvironments. For example, Forman (1968) notedthat A. xenica, C. variegatus, and F. pulvereus are widelydistributed in Louisiana marshes and are common ina variety of coastal landscape types, including pools inthe marsh surface. Cyprinodontids, in general, arewidely tolerant of low oxygen, fluctuating salinity, andextremes in water temperature (Griffith 1974, Cregoand Peterson 1997). A number of cyprinodontid andpoecilid species, including many of those reported inthis study, are also adapted for aerial respiration(Kramer and Mehegan 1981, Halpin and Martin1999).

Conclusions

Variation in nekton density among areas and land-scape types was great, in part, due to the inherentsampling gear issues (Rozas and Minello 1997), and, inpart, because of differences in characteristics amongareas. Differences in salinity and plant stem densitybetween area C and the other two areas suggest thatexamination of marsh function as attempted in thisstudy is extremely sensitive to specific site characteris-tics. Our data indicate that nekton densities weregenerally greater in natural compared to leveed andgapped landscapes. This was especially the case in areaC that was dominated by S. alterniflora, with lower plantdensity and higher salinity. Additionally, water depthin area C was always greater than areas A and B, sug-gesting that elevation was lower and thus nekton mayhave had greater access to the marsh surface (Reedand others 2004). Comparison of gapped to natural orleveed nekton densities (Table 7) indicated that thefrequency of G < N (16.7–50%), G = N (20–40%),G < L (16.7–50%) or G = L (20–50%) suggest that

Figure 8. Plot of mean water level for each sediment depo-sition collection period as measured at the tide gauge atLUMCON in Cocodrie, Louisiana.


gapping material levees were generally equal to orgreat than leveed or natural landscapes and may be aneffective method of partially restoring access to theupper marsh by nekton, but that this method maywork best in S. alterniflora-domimted marshes whereelevation is lower, tidal inundation greater, and nektonuse greater. More research is needed to support thistrend.

Although differences among areas were less impor-tant in explaining patterns of sediment deposition, thestudy provided less conclusive insights into the effectsof gapping levees on marsh surface sedimentation.Rather, this study confirmed previous work that iden-tified the need for both enhanced sediment availabilityin the water column with effective marsh flooding toincrease sediment deposition. The gaps examined hereappear to be too restrictive to provide efficient move-ment of flood waters onto the marsh during moderateflooding events. Should managers wish to begin pilotimplementation of this restoration technique, anexperimental approach incorporating gaps greaterthan 20 m, together with monitoring similar to thatconducted for this study, would better assess its utilityin improving marsh function. Such efforts should bemonitored over a number of years to more clearlyidentify differences associated with levee gaps from thenatural spatial and temporal variability characteristic ofcoastal marshes. However, any pilot gapping projectsmust also consider the findings of this study that dur-ing high magnitude events, the ‘‘trapping’’ effect ofthe levees increases sediment deposition and that ex-treme flooding dominates the effect of the gaps onmarsh hydrology.

A possible benefit of levee gapping that was notexamined here is the potential for an improvement inmarsh soil drainage as the ponding and waterloggingeffect of the levees is possibly ameliorated. Furtherstudies would be required to elucidate whether gap-ping of levees can reduce soil waterlogging and possi-bly benefit vegetation in areas affected by dredgedmaterial placement.

AcknowledgmentsWe thank C. M. Woodley, G. L. Waggy, M. L.

Partyka, D. Bond, L. Dancer, and J. Anderson forassistance in field and lab activities. The LouisianaUniversities Marine Consortium provided water leveldata for Cocodrie, and Ms. Chandra Molinere isthanked for her assistance. Dr. Nina De Luca con-ducted statistical analysis of the sediment depositiondata.

Literature Cited

Beck, M. W., K. L. Heck Jr., K. Able, D. Childers, D. Eggleston,B. M. Gillanders, B. Halpern, C. Hays, K. Hoshino, T. Mi-nello, R. Orth, P. Sheridan, and M. Weinstein. 2001. Theidentification, conservation, and management of estuarineand marine nurseries for fish and invertebrates. Bioscience51:633–641.

Brown-Peterson, N. J., and M. S. Peterson. 1990. Comparativelife history of female mosquitofish, Gambusia affinis, in tidalfreshwater and oligohaline habitats. Environmental Biology ofFishes 27:330–341.

Burdick, D. M., M. Dionne, R. M. Boumanns, and F. T. Short.1997. Ecological responses to tidal restorations of twonorthern New England salt marshes. Wetlands Ecology andManagement 4:129–144.

Cahoon, D. R., D. J. Reed, J. W. Day, G. D. Steyer, R. M.Boumans, J. C. Lynch, D. McNally, and N. Latif. 1995. Theinfluence of hurricane Andrew on sediment distribution inLouisiana coastal marshes. Journal of Coastal Research SI21:280–294.

Carter, V 1984. An overview of the hydrologic concerns re-lated to wetlands in the United States. Canadian Journal ofBotany 64:364–374.

Chabreck, R. H., and R. G. Linscombe. 1988. Louisianacoastal marsh vegetative type map 1988. LouisianaDepartment of Wildlife and Fisheries, Baton Rouge,Louisiana, 10 maps.

Crego, G. J., and M. S. Peterson. 1997. Salinity tolerance offour-ecologically distinct species of Fundulus (Pisces:Fundulidae) from the northern Gulf of Mexico. Gulf ofMexico Science 15:45–49.

DeLaune, R. D., W. H. Patrick, and N. Breemen. 1990. Pro-cesses governing marsh formation in a rapidly subsidingcoastal environment. Catena 17:277–288.

Forman, W. W. 1968. Notes on the ecology of six species ofcyprinodontid fishes from Grand Terre, Louisiana. Pro-ceedings of the Louisiana Academy of Sciences 31:39–40.

French, J. R., T. Spencer, A. Murray, and N. S. Arnold. 1995.Geostatistical analyses of sediment deposition in two smalltidal wetlands, Norfolk, U.K. Journal of Coastal Research11:308–321.

Fulling, G. L., M. S. Peterson, and G. J. Crego. 1999. Com-parison of Breder traps and seines used to sample marshnekton. Estuaries 22:224–230.

Griffith, R. W. 1974. Environment and salinity tolerance inthe genus Fundulus. Copeia 1974:319–331.

Halpin, P. M., and K. L. M. Martin. 1999. Aerial respiration inthe salt marsh fish Fundulus heteroclitus (Fundulidae). Copeia1999:743–748.

Hood, G. W. 2004. Indirect environmental effects of dikes onestuaries and tidal channels: thinking outside of the dike forhabitat restoration and monitoring. Estuaries 27:273–282.

Hutchinson, S. E., F. Sklar, and C. Roberts. 1995. Short termsediment dynamics in a southeastern U.S.A. Spartina marsh.Journal of Coastal Research 11:370–380.

Jordan, F., S. Coyne, and J. C. Trexler. 1997. Sampling fishesin vegetated habitats: effects of habitat structure on sam-


pling characteristics of the 1-m2 throw trap. Transactions ofthe American Fisheries Society 126:1012–1020.

Kramer, D. L., and J. P. Mehegan. 1981. Aquatic surface res-piration, an adaptive response to hypoxia in the guppy,Poecilia reticulate (Pisces, Poeciliidae). Environmental Biologyof Fishes 6:299–313.

Kuhn, N. L., I. A. Mendelssohn, and D. J. Reed. 1999. Alteredhydrology effects on Louisiana salt marsh function. Wet-lands 19:617–626.

Louisiana Coastal Wetlands Conservation and RestorationTask Force (LCWCRTF) and the Wetlands Conservationand Restoration Authority. 1998. Coast 2050: toward asustainable coastal Louisiana. Louisiana Department ofNatural Resources, baton Rouge, Louisiana, 161 pp.

Marmer H. A. 1954. Tides and sea level in the Gulf of Mexico.In: P. S. Galtsoff (ed.) Gulf of Mexico, its origin, waters andmarine life. Fishery bulletin 89. US Fish and Wildlife Ser-vice, Washington, DC, 604 pp.

Mitsch, W. J., and J. G. Gosselink. 2000. Wetlands. 3rd ed.John Wiley, New York.

Morton, R. A., N. A. Buster, and M. D. Krohn. 2002. Sub-surface controls on historical subsidence rates and associ-ated wetland loss in Southcentral Louisiana. Gulf CoastAssociation of Geological Societies Transactions 52:767–778.

Penland, S., and K. E. Ramsey. 1990. Relative sea-level rise inLouisiana and the Gulf of Mexio:1908–1988. Journal ofCoastal Research 6:323–342.

Reed, D. J. 1989. Patterns of sediment deposition in subsidingcoastal salt marshes, Terrebonne Bay, Louisiana: the role ofwinter storms. Estuaries 12:222–227.

Reed, D. J. 1992. The effect of weirs on sediment depositionin Louisiana coastal marshes. Environmental Management16:55–65.

Reed, D. J. 2002. Sea level rise and coastal marsh sustain-ability: geological and ecological factors in the MississippiDelta Plain. Geomorphology 48:233–243.

Reed, D. J., and L. P. Rozas. 1995. Restoring Louisiana coastalhabitats through infilling existing pipeline canals: a geo-morphological evaluation. Wetlands 15:149–158.

Reed, D. J., and N. De Luca. 2003. Managing to promotesediment deposition in subsiding Louisiana coastal mar-shes: the Brady Canal project. Proceedings Coastal Sedi-ments �03, May 18–23, 2003. Clearwater Beach, Florida, pp1–12.

Reed, D. J., N. Luca, and A. L. Foote. 1997. Effect of hydro-logic management on marsh surface sediment depositionin coastal Louisiana. Estuaries 20:301–311.

Reed, D. J., M. S. Peterson, and B. J. Lezina. 2004. Can gaps inmaterial levees improve marsh function in coastal Louisi-ana? Final report submitted to U.S. Army Corps of Engi-neers Engineer Research and Development Center,

Vicksburg, Mississippi 39180. Contract no. DACW 42-01-P-0280, 46 pp.

Roman, C. T., W. A. Niering, and R. S. Warren. 1984. Saltmarsh vegetation change in response to tidal restrictions.Environmental Management 8:141–149.

Roman, C. T., K. B. Raposa, S. C. Adamowicz, M.-J. James-Pirri, and J. G. Catena. 2002. Quantifying nekton and veg-etation response to tidal restoration of a New England saltmarsh. Restoration Ecology 10:450–460.

Rozas, L. P., and D. J. Reed. 1993. Nekton use of marsh-surface habitats in Louisiana (USA) deltaic salt marshesundergoing Submergence. Marine Ecology Progress Series96:147–157.

Rozas, L. P., and T. J. Minello. 1997. Estimating densities ofsmall fishes and decapod crustaceans in shallow estuarinehabitats: a review of sampling design with focus on gearselection. Estuaries 20:199–213.

Rozas, L. P., and R. J. Zimmerman. 2000. Small-scale patternsof nekton use among marsh and adjacent shallow non-vegetated areas of the Galveston Bay estuary, Texas (USA).Marine Ecology Progress Series 193:217–239.

Schroeder, W., and W. J. Wiseman. 1986. Low-frequencyshelf-estuarine exchange processes in Mobile Bay andother estuarine systems of the Northern Gulf of Mexico.Pages 355–367 in D. A. Wolfe (eds.), Estuarin variability.Academic Press, Orlanda, Florida.

Swenson, E. M., and R. E. Turner. 1987. Spoil banks: effectson a coastal marsh water-level regime. Estuarine, Coastal andShelf Science 24:599–609.

Tanner, C. D., J. R. Cordell, J. Rubey, and L. Tear. 2002.Restoration of freshwater intertidal habitat functions atSpencer Island, Everett, Washington. Restoration Ecology10:564–576.

Turner, R. E. 1987. Relationship between canal and leveedensity and coastal land loss in Louisiana. US Departmentof Interior, Fish and Wildlife Service, National WetlandsResearch Center. Biological Report 85 (14).

Turner, R. E., and D. R. Cahoon. 1988. Causes of wetland lossin the coastal central Gulf of Mexico, volume II: Technicalnarrative. U.S. Department of the Interior, Minerals Man-agement Service, Gulf of Mexico OCS Regional Office, NewOrleans, Louisiana.

Turner, R. E., J. M. Lee, and C. Neill. 1994. Backfilling canalsto restore wetlands: empirical results in coastal Louisiana.Wetlands Ecology and Management 3:63–78.

Underwood, A. J. 1997. Experiments in ecology. Their logicaldesign and interpretation using analysis of variance. Cam-bridge University Press, Cambridge, 504 pp.

Williams, P. B., and M. K. Orr. 2002. Physical evolution ofrestored breached levee salt marshes in the San FranciscoBay estuary. Restoration Ecology 10:527–542.


reducing the effects of dredged material levees on coastal ... › cv › peterson.mark › docs ›...

Documents