Methods

Study Sites

The study took place on two constructed sites of differing ages and a natural reference marsh on Deer Island, MS. The constructed sites, DIMR1 and DIMR2, are adjacent to each other and joined by a containment dike which protects the inner developing marsh platforms from erosion. The method used in the construction of the DIMR sites was the utilization of beneficial-use of dredge material obtained from the maintenance dredging of commercial shipping channels in the nearby area.

The 10+ y constructed site, DIMR1, was constructed in 2004 with dredged sediments sourced from Biloxi Bay channel and is 18-hectare in size. The 10+ y constructed site was first planted in spring 2005 with commercially purchased J. roemerianus (13,440), S. alterniflora (15,400) and S. patens (17,920) within the containment dike. Following extensive dike failure during and after hurricane Katrina (Aug 2005), additional sediments and a more expansive containment feature was constructed during 2010-2011. Further planting of J. roemerianus, S. alterniflora, S. patens, P. amarum, and U. paniculata was periodically completed at the 10+ y constructed site from 2008-2011, with the most substantial planting being of 15,000 J. roemerianus, 5,000 S. alterniflora and 3,000 dune plants in 2008, most of which were subsequently lost due to erosion at the site.

The 2+ y constructed site, DIMR2 is a 16-hectare area constructed from 2015 to 2018 with sediments dredged from multiple sources in Jackson and Harrison Counties in Mississippi. DIMR2 was planted on the eastern third in spring 2016 with commercially purchased J. roemerianus (18,836) and S. alterniflora (18,836) in the interior high and low marsh zones. On the exterior containment dike, S. patens (2,041), Panicum amarum Elliot (bitter panicgrass) (2,041), and Uniola paniculata L. (seaoats) (4,083) were planted. Additional revegetation of the remainder of this site has occurred largely through natural recruitment since then.

The 100+ y reference marsh is approx. 500 m from the 2+ y and 10+ y constructed sites, separated by an upland dune ridge colonized by the longleaf pine (Pinus palustris Mill.) and a variety of shrubs such as Serenoa repens W. Bartram (small saw palmetto), and B. halimifolia. The natural marsh is entirely comprised of a zone mixed with J. roemerianus and S. alterniflora, with mussels and fiddler crab burrows frequently found in the soft, muddy sediment. The natural marsh was selected as a reference site for its proximity to the constructed sites as well as it’s elevation and plant community composition, which makes it an ideal candidate for gauging the progress of the constructed sites towards being comparable to Deer Island’s natural marsh footprint.

Field Sampling and Lab Analyses

Sampling was conducted over six seasons: spring and fall 2017, spring and fall 2018, and spring and fall 2019. In the spring 2017 sampling season, two 100 m long replicate transects were established at each study site with approximately 250 m between starting points. In the fall 2017 sampling season and onward, an additional 100 m replicate transect was added in the middle of the original two transects for additional sampling of community diversity, resulting in 125 m between transects. Starting points at the 2+ y and 10+ y constructed sites were established randomly along the containment dike which lies on the northern part of the site. The starting points were retained across sampling seasons. At each sampling effort, the transects were ran haphazardly from the starting point, generally in the same direction, resulting in similar transect orientation with different ending points across each sampling season.

Sediment characteristics

In the vegetation quadrats, two near surface sediment samples (depth 5-10 cm within the rhizosphere zone) and a deeper sediment samples (20-30 cm, below the rhizosphere zone) were collected from the same hole as the vegetation biomass core. Sampling replication for sediment core samples was two shallow and two deep sediment cores per 1 m²quadrat. The sediment sample was extracted from within the plant rhizosphere zone by pushing the syringe at right angles and parallel to the sediment surface into the adjacent root zone on the side-wall of the hole made after the biomass core had been removed.

Vegetation diversity and cover

In six five seasons, along the two transects, 1 m² quadrats were used to estimate percent cover of species. Quadrats were spaced haphazardly at approximately 20-40 m intervals and placed within representative areas of each marsh zone (low-, mid-, high-marsh, and dunes). The percent cover of the three most abundant species present in the 1 m²quadrat was estimated visually by a minimum of two personnel.

Vegetation biomass

In the selected quadrats, two replicate vegetation biomass cores (15 cm diameter x ~30 cm depth of the rhizosphere) of plant species of interest (S. alterniflora, J. roemerianus, S. patens, D. spicata) were taken for measurements of canopy and rhizosphere biomass. S. alterniflora and J. roemerianus samples were later used in δ¹³ C and δ¹⁵ N stable isotope analysis.

Physical and chemical sediment characteristics

Sediment cores were subsampled for two different analyses, where the measurement derived from each analysis was representative of the sediment collected from the corresponding vegetation biomass core collected from a quadrat. The subsamples were: 1) a 2.5 mL subsample used for sediment organic content via loss on ignition (LOI) and sediment bulk density (g/cm³), and 2) the remainder of the amount of the total collected sediment (approx. 15 mL), used for grain size analysis. Following subsampling, the sediment subsamples were dried at 70° C until moisture was lost from the sediments. All subsamples were weighed after drying. The mass (g) of the dried 2.5 mL subsamples was considered the sediment bulk density (g/cm³). The same subsample was then combusted in a furnace (Thermolyne 62700) at 500° C for 4 h, removing any organic matter within the sample. The combusted subsample was weighed to obtain the ash-free dry weight (AFDW) which was then subtracted from the pre-combustion weight to estimate the loss of organic matter from ignition. The amount of material lost on ignition was compared to the original amount of pre-combustion (inorganic + organic) material to calculate the percent of organic content of the soil sample. The subsamples taken for grain-size analysis were wet-sieved over No. 10 (=2 mm, coarse), No. 35 (= 0.5 mm, fine sand), and No. 230 (= 0.063 mm, very-fine sand) mesh sieves. Following sieving, each grain size was stored in pre-weighed aluminum tins and dried in a drying oven at 70° C until all moisture was lost. The tins containing each grain size were then weighed. This measurement for each grain size was summed and then subtracted from the pre-sieving mass to obtain the amount of silt and clay material lost during the sieving process. The mass of each grain size was then divided by the total amount of the sediment to obtain % coarse material, % fine sand, % very fine sand, and % silt/clay.

Vegetation biomass and stable isotope analysis

Vegetative biomass cores taken from the field were promptly washed to remove sediment and debris from the above-ground material (AGM) and below-ground material (BGM) compartments. Using shears, biomass cores were then separated into species-specific tissue compartments consisting of the AGM portion (stems, leaves, and flowering structures) and the BGM portion (roots). The AGM portion was then separated into alive and dead portions based on color, where green material was considered alive and brown or dark material was considered dead. Following separation, all tissue compartments were placed in previously weighed aluminum tins and allowed to dry in a drying oven at 70º C for a minimum of 3 days to remove all moisture from the tissues. After drying, each tissue compartment was weighed on a balance and the mass (g) was recorded. The BGM, AGM separated as separate live and dead canopy compartments were ground in a Wiley mill to pass a # 40 mesh sieve. Each ground sample was stored in a labelled 20 ml glass scintillation vial. For select tissue samples, a 3-4 mg subsample of the ground material was then packaged into a 4 x 6 mm tin capsule and analyzed by a Thermo Delta V Advantage stable isotope ratio mass spectrometer coupled to a Costech elemental analyzer via a Conflo IV interface, producing measurements of C, N, and δ¹³ C and δ¹⁵ N.