Vincent G. Guida14, 15 and Andrew F. J. Draxler14, 16
Postal Address: 14National Marine Fisheries Serv., 74 Magruder Rd., Highlands, NJ 07732
E-Mail Addresses: 15Vincent.Guida@noaa.gov; 16Andrew.Draxler@noaa.gov
INTRODUCTION
Knowledge of soil biogeochemistry in salt marshes is essential to an understanding of the role of these wetlands in promoting fisheries. Biogeochemistry is the chemistry that mediates interactions among atmosphere, water, minerals, microbiota, and higher organisms in the environment. It is characterized by both linear and cyclic transformations of materials, yielding deposits of end products for some materials, a dynamic steady state for other materials, and the generation of metabolic energy that sustains the structure of the ecosystem.
Organic wetland soils (i.e., hydric histosols) such as those in salt marshes have a great capacity for transformations of forms of carbon, nitrogen, phosphorus, sulfur, and transition metals. These transformations are essential to the productivity and structure of salt marshes, and thereby to the functioning of these marshes as critical fisheries habitat (Boesch and Turner 1984). The same transformations are also critical as exporters of production to adjacent waters (Haines 1979).
Saturation of wetland soils with water impedes the diffusion of gaseous atmospheric oxygen into such soils. Water, through which oxygen diffuses slowly, fills the interstices between solid particles that would otherwise be filled with gas, through which oxygen diffuses rapidly. At the same time, water saturation facilitates the diffusion of nongaseous water-soluble materials. The result is a sharp decline in the reducing-oxidizing (i.e., redox) potential with depth that is characteristic of wetland soils, and one that leads to other characteristics typical of wetland soils (Mitsch and Gosselink 1986).
Some of these other characteristics include anoxia which allows the net accumulation of organic matter from dead plant parts that cannot be converted entirely to carbon dioxide and water without an oxidant, thus allowing the development and maintenance of the characteristically high organic matter content of the histosol. Another characteristic resulting from the abundant organic matter and redox potential gradient is the variety of microenvironments favoring an array of microbiota with extremely diverse metabolic capabilities. These capabilities include the cyclic redox transformations of nitrogen, sulfur, iron, and manganese that are essential to maintaining primary production, and a variety of organic decay mechanisms that allow transfer of energy from refractory lignocellulose plant remains into the trophic web (Howes et al. 1984; Newell 1993).
The aforementioned characteristics are those of established, functioning wetland soils. The re-establishment of productive stands of S. alterniflora has clearly been successful in the replanted salt marshes of the Arthur Kill (Bergen et al. 2000). We have sought to determine if the re-establishment of normal wetland soil characteristics has been as successful. Indeed, we seek to understand to what extent oiling and other urban influences have affected these unseen, but vital biogeochemical features of the Arthur Kill salt marshes.
Previous studies found that created wetlands had lower OC and nitrogen content, but higher redox potential values (measured as Eh) and higher manganese and iron content, than natural wetlands (Craft et al. 1991; Zedler 1993). Other studies have shown that sediment organic content appears to develop slowly in newly created salt marshes, and may take years to reach natural levels (LaSalle et al. 1991; Moy and Levin 1991; Minello and Zimmerman 1992). Likewise, development of saltmarsh nitrogen cycling rates can require years (Thompson et al. 1995). These differences between newly created and established wetlands are related to the low organic matter content of soils used for de novo wetland creation as compared with that of natural marsh soil. Such findings suggest that marshes planted on nonwetland soils cannot immediately duplicate the biogeochemical functions of older, natural systems.
Unlike wetlands that have been "created" though, the Arthur Kill marshes such as the one at Old Place Creek have been replanted on old wetland soils, although ones which have been highly altered in this urban environment. Prior to replanting, those sites were denuded due to the oil spill, and were barren for a few years until replanted. By comparing these replanted wetlands with the denuded areas that were not replanted and with areas that were not denuded, we hoped to investigate what effect these varied histories may have had on the organic matter essential for soil function.
The abundance of soluble inorganic sulfide, primarily in the forms of the dissolved hydrosulfide ion (HS-) and hydrogen sulfide (H2S), is of particular interest because it is an indicator of microbial activity and recent redox history, and because it has a strong influence on primary production. Soluble sulfide is a product of a bacterial respiratory pathway that utilizes the abundant sulfate ion (SO4-2) in saline interstitial water as a source of oxidizing power where other oxidants (including O2 and NO3-) have been depleted. Sulfate reduction is a major form of respiration in salt marshes (Howarth and Giblin 1983), and the sulfide produced is known to inhibit root uptake of nitrogen in S. alterniflora and other wetland plants (Bradley and Morris 1990a; Koch et al. 1990). In the case of S. alterniflora, sulfide inhibition of nitrogen uptake, along with salinity inhibition, are potentially important sources of productivity limitation since growth of this grass appears to be nitrogen-limited in most Gulf and Atlantic Coast marshes (Bradley and Morris 1992).
While measurement of sulfide deals with a narrow set of redox species, measurement of soil Eh integrates the influence of many such chemical species (e.g., O2 , Fe+3, Fe+2, Mn+4, Mn+2, NH4+, NO3-, H2S, HS-, CH4, H2, etc.). Typical depth profiles for Eh in saltmarsh soils have high positive values at the surface, indicating the penetration of hydrospheric and atmospheric O2. This surface layer is underlain by a rapid decline in values within the first few centimeters, indicating exhaustion of O2. Below this rapid decline, there is a continuous slow decline in Eh toward values characterizing sulfide presence, perhaps underlain by still lower values indicating exhaustion of sulfate substrate and supporting only fermentative metabolism (Howes et al. 1981).
Aside from depth in the soil, Eh values are influenced by the presence or absence of vegetation, the type of vegetation, and the seasonal state (i.e., dormancy vs. active growth) of vegetation (de la Cruz et al. 1989). During the growing season, S. alterniflora oxidizes marsh soil (increases Eh) by conducting oxygen downward into its roots and by diffusion to the adjacent soil root zone via specialized aerenchyma tissues. In addition, the grass lowers soil water saturation levels by transferring interstitial water to the atmosphere via evapotranspiration (Howes et al. 1981). Such processes are not active during the dormant season (fall-winter) or where grass has been intentionally burnt or clipped or has died back of other causes; such factors result in lower Eh values that appear to inhibit recolonization in dieback areas where rhizosphere oxidation cannot be re-established simply by reactivation of existing root and rhizome systems (Bertness 1991). On the other hand, low Eh values can also result in the release into interstitial water of dissolved inorganic nutrients essential to plant growth (DeLaune et al. 1981).
The presence of living S. alterniflora, with its seasonal oxidizing influence (i.e., raising Eh values) and release of soluble organic materials into soils, has important consequences for the biogeochemistry of salt marshes. In addition to raising the Eh, the microzones around cordgrass roots: 1) exhibit lower pH; 2) accumulate and immobilize forms of such potentially toxic metals as copper, zinc, and lead, presumably due to the tight binding of such metals to abundant organic ligands (Caçador et al. 1996); 3) oxidize sulfide (Howes et al. 1981); 4) mobilize iron (Kostka and Luther 1995); and 5) facilitate the redox cycling of iron via organic ligand complexes (Luther et al. 1992). These alterations in cordgrass root microzones form a positive feedback on the growth of cordgrass, for which iron is essential and sulfide and excess heavy metals are toxic. Such biogeochemical functions are presumably lost when cordgrass roots die, resulting in greater mobility of heavy metals, higher sulfide levels, and immobilization of iron, making soil less suitable for re-establishment and growth of S. alterniflora.
Finally, there is the effect of residual petroleum contamination on biogeochemical functions in Arthur Kill marshes. Early research with crude oil suggested that this substance may not be very toxic to such important microbial processes as the fixation of nitrogen (Thomson and Webb 1984), the reduction of nitrate, manganese, iron, and sulfate, and the production of methane and ammonia (DeLaune et al. 1979). However, subsequent work with specific oil components has shown that naphthalene can inhibit sulfate reduction in saltmarsh sediments while increasing methanogenesis (Keine and Capone 1984). Indeed, there is evidence that fuel oil can increase detrital decay rates (Hershner and Lake 1980), and that low levels of a light hydrocarbon mixture can stimulate CO2 production, methanogenesis, nitrogen fixation, and denitrification in saltmarsh soils (Li et al. 1990).
Because of the key roles played by soil organic matter, sulfide, and redox potential in controlling saltmarsh structure and productivity, we have chosen to measure total and labile (i.e., readily degraded by aerobic microbiota) organic matter in soil, and soluble sulfide and Eh values in interstitial water, as means of characterizing and assessing the comparative biogeochemical condition of Arthur Kill marshes and the possible efficacy of the replanting efforts.
METHODS AND MATERIALS
At each of the six sites, four stations were selected at 0.2 m above mid-tide. The separation between replicate stations at a site ranged between 2 and 20 m (see description of the sampling transect in Chapter II, "Trace Metal Contaminants in Sediments and Ribbed-Mussels"). In order to take into account seasonal variations, sampling regimes were performed during two intervals: September-October 1996 representing fall, and May 1997 representing spring.
Redox Potential
Duplicate vertical profiles of redox potential (measured as Eh) were taken in situ within 10 cm of each other at each station using a Fisher Scientific Model 640 portable pH/millivolt meter. The instrument was calibrated to within 1 mV at 10, 100, and 1000 mV using a Cole-Parmer pH-mV calibrator (#5657-10). A platinum electrode (Thomas Scientific #4096-D20) with a band of platinum (6-mm dia. × 4-mm height) was used as the sample electrode. The reference electrode was a Fisher Scientific (#13-639-62) sleeve-junction calomel electrode. This system was calibrated using three K3Fe(CN)6-K4Fe(CN) 6 solutions of differing redox potentials (ZoBell 1946; Orion Research, Inc. 1983).
Depth in the sediment was measured to the center of the band except for the "zero" reading which was made with the band just immersed in the sediment. Before each profile, the electrode system was equilibrated in water collected in the Arthur Kill adjacent to the site. In the first profile, measurements were made in overlying surface water at 1 cm above the soil surface, at the soil surface (0), 0.5 cm below the surface, and 1-5 cm below the surface at 1-cm intervals, soil penetrability permitting. In the second profile, measurements were made 1-10 cm below the soil surface at 1-cm intervals, soil penetrability permitting. Readings at each depth interval were accepted when the rate of change was <1 mV in 10 s.
Soluble Sulfide
Samples for porewater sulfide determination were obtained with the use of de-ionized water (DIW) equilibration devices, or "peepers" (Hesslein 1976; C. Wigand, U.S. Environmental Protection Agency, 27 Tarzwell Dr., Narragansett, RI, pers. comm.; Figure 25). These devices obtained time-integrated, particulate-free samples of water with the same dissolved chemical composition as interstitial water by allowing an extended period of equilibrations between water outside the peeper and DIW inside the peeper across membranes with molecule-sized pores. Each peeper consisted of a body and two face plates (one on each side) of high-density polyethylene (HPDE) held together with stainless steel screws and washers. Sandwiched between the body and face plates and enclosing five pairs of cylindrical equilibration chambers (each 2.54-cm dia. × 1.9-cm depth, with a capacity of 9.7 ml) were two cellulose dialysis membranes (Spectra/Por1) made from regenerated cellulose and having a molecular weight cutoff of 6000-8000 Daltons.
In order to avoid contamination of subsurface soil and porewater with molecular oxygen from the peepers, the HPDE peeper parts were soaked for 25 days in DIW which had been continuously sparged with N2 gas in order to remove adsorbed O2 prior to assembly. To minimize O2 absorption before insertion in marsh soil, peeper chambers were filled with N2-sparged DIW, and the assembled peepers were wrapped in polyvinylidine chloride film and were transported in N2-sparged water. Peepers used in 1996 had a single set of vertical chambers, while those used in 1997 had two sets, as depicted in Figure 25. As each chamber spanned 2.54 cm in the vertical dimension, samples from the five pairs of samples represent conditions from the following height/depth ranges (+ is above the soil surface and - is below): 1) +3.2 to +0.6 cm (mean = +1.9 cm); 2) -0.6 to -3.2 cm (mean = -1.9 cm); 3) -4.4 to -7.0 cm (mean = -5.7 cm); 4) -8.3 to -10.8 cm (mean = -9.5 cm); and 5) -13.3 to -15.9 cm (mean = -14.6 cm).
Each peeper was driven into the soil with a rubber mallet at one (1996) or two (1997) randomly chosen stations from among the four stations at each site so that four sets of chambers were below the surface of the soil and one above (Figure 25). The peepers were left undisturbed to equilibrate with the porewater for 13-14 days. Upon retrieval, the cellulose membranes covering each chamber were punctured with a Teflon needle (reused after flushing with surface seawater), and samples were withdrawn from the chamber with disposable syringes and transferred to 1.5-ml polypropylene centrifuge tubes for immediate addition of colorimetric reagents and dilution water (i.e., N2-sparged, filtered seawater). Sample dilution by 10- or 100-fold was necessary to overcome the limitations of the analytical method.
Analysis was performed by a micromethod derived from the method of Strickland and Parsons (1972), with reagents calibrated against iodometric titration, and standardization performed using serial dilutions of sulfide-enriched, N2-sparged seawater. Optical absorbance of the samples was determined with a Perkin Elmer Lambda 3B spectrophotometer within 4 hr of collection. Tests of this microprocedure have shown a 2.07±3.04% (i.e., mean ± standard deviation; n = 8) decrease in absorbance of samples after 7 hr of reaction time. Total procedural error for the method was ±10% (n = 26) for sulfide values up to 900 µM. Values between 900 and 3000 µM were ±30%. Values beyond 3000 could not be estimated by our method, and were recorded as ">3000 µM."
Total Organic Carbon
Sediment samples for the determination of TOC were manually collected in 28-mm (internal diameter) butyrate cores, one core per station, to an average depth of 10 cm. Care was taken to prevent disturbance of the sediment surface layer by maintaining the cores upright on ice for the return to the laboratory, where they were frozen at -20°C until processed. Frozen sediment cores were transferred to a 4°C cold room and defrosted. The cored sediment was extruded; the surface layer (to a depth of 1 cm) was collected in individual precleaned glass containers. Due to the vegetated nature of the collection sites, sectioning of the sediment was accomplished with a serrated stainless steel blade. Sectioned sediments were dried overnight at 50°C. All large carbonaceous fragments (>1 mm) were removed; samples were then homogenized with a glass rod. A 100-mg subsample was transferred into an acid-cleaned, uncapped scintillation vial. The acidification technique of Yamamuro and Kayanne (1995) was employed to remove carbonate, while avoiding the dissolution of acid-soluble organic material.
Aliquots, ranging from 10 to 50 mg, were taken from the acidified samples, placed into tin combustion cups, and sealed in preparation for organic carbon and nitrogen analyses by flash combustion in oxygen at 1020°C on a Carlo Erba 1108 elemental analyzer equipped with a 120-position autosampler. Combustion products pass over a catalyst and then over copper to remove excess oxygen and to reduce the nitrogen oxides to elemental nitrogen. Upon separation by gas chromatography, the CO2 and N2 eluent peaks are integrated and reported as organic carbon and nitrogen. Instrument calibration was maintained by performing a series of linear regressions using an acetanilide standard. These standard and additional TOC SRMs (e.g., PACS-1) were placed into the sample rotation for further verification of optimal operation.
Labile Carbon
The 0.1-cm3 sediment surface samples for labile carbon (LC) analysis were obtained by coring marsh sediments using disposable 10-ml plastic syringes with the tapered end cut off. These cores included the upper 1 cm of the soil column. Duplicate samples were taken at each of two randomly chosen stations at each site during May 1997.
The LC content of these samples was estimated via measurement of dissolved oxygen consumption by the microbiota during a 13-day incubation of the sample in a 300-ml BOD (biological oxygen demand) bottle filled with natural seawater and maintained at 20°C (Draxler 1995). Sediment oxygen consumption was compared with oxygen consumption by D-glucose under the same conditions. Therefore, sediment LC, hereafter reported as "µM-C/cm3," signifies micromoles of glucose carbon equivalents per cubic centimeter of sediment.
Data Analysis
Statistical analyses of data were performed with the aid of Jandel Scientific SigmaStat 2.0. Procedures included the use of Kruskal-Wallis, one-way ANOVA and the Mann-Whitney rank sum test for site-by-site comparison of redox potential, sulfide concentrations, and TOC content of soil, with the use of Tukey, Dunn's, or Student-Newman-Keuls pairwise tests, as appropriate, and of Student's t-tests, for seasonal comparison of redox potential.
RESULTS
Redox Potential
A total of 965 measurements of Eh were made for the purpose of creating duplicate depth profiles for redox potential in soil at four stations (designated A through D) at each of the six sites in each of two seasons: fall 1996 and spring 1997 (Figures 26-28). Depths in the soil in these and all subsequent presentations are indicated in terms of negative (-) elevations, with a zero value denoting the soil surface. Most, though not all, sites demonstrated the classic declining gradient in Eh with depth, described in the introduction to this chapter. The highest values were generally at the surface. In most cases, a zone of rapid decline of redox potential was within a depth of 2 cm, with more slowly declining values between -2 and -5 cm, then consistently low values below -5 cm.
Values of Eh for surface water collected adjacent to the sites were +268±37 mV (n = 47) during fall 1996, and +314±23 mV (n = 46) during spring 1997 (Figures 26-28). Since the Eh range of oxygenated pure water is above +350 mV, these low values suggest poor quality in water overlying the marsh during high tide (i.e., little dissolved oxygen and/or an abundance of reductants such as organic matter).
Beneath the soil surface, some profiles within each station diverged from the classic pattern, yielding both extraordinarily low and extraordinarily high Eh values at depth (e.g., Figure 26A). While replanted sites appeared more prone to high subsurface values (Figure 26), this phenomenon was not uniformly demonstrated at all stations nor in all seasons within that treatment group, nor was it exclusively present there (Figure 27 and Figure 28). Of the 40 highest values (+374 to +478 mV), 26 were taken at Old Place Creek - Station A during both fall and spring, and 11 from Saw Mill Creek North - Station B in spring, making these the stations with the most oxic sediments. Paradoxically, Old Place Creek (a replanted site) was also the site where the lowest (i.e., most reduced) values of the study were recorded. Of the 21 lowest values (<-251 mV), 16 came from Old Place Creek - Stations B, C, and D (replanted), making the deeper layers at these stations the most anoxic sediments of the study -- a sharp contrast to the highly oxic character of sediments at Old Place Creek - Station A. Extremely low values were evident at the reference sites (e.g., Figure 27A) as well.
While Eh declined with depth in the soil at most stations, it increased with depth in a few profiles, including those for Old Place Creek - Station A (replanted), Saw Mill Creek North - Station B (replanted), and Tufts Point - Station C (reference) in both fall and spring, and for Con Ed Tower - Stations A and C (unplanted) in spring only. Other stations at those sites did not exhibit this pattern.
Unequal variances and deviations from normality prevented the use of parametric, three-way ANOVA of Eh data to detect significant differences by depth, season, restoration status, and station. Therefore, a separate nonparametric test (i.e., the Kruskal-Wallis, one-way, ANOVA-on-ranks test) for depth, restoration status, and station, and another (i.e., the Mann-Whitney rank sum test) for season, were employed to test the significance of each of these factors. There was a significant difference associated with depth (P <0.001), and indeed, a significant linear correlation with that factor (r2 = 0.433, P <0.01). There were also significant differences by season, by treatment, and by station (all P <0.001). Spring Eh values were significantly higher than fall values. Dunn's pairwise comparison test showed significant differences (P <0.05) between replanted and unplanted sites, and between replanted and reference sites; replanted marshes had significantly higher redox potentials than unplanted or reference marshes.
While the last finding suggests an effect of replanting status upon redox potentials, that pattern was not borne out by examination of Eh values by station. Dunn's pairwise comparison test demonstrated that high redox values for replanted marshes were driven largely by exceptionally high values from Old Place Creek. Despite some very low values, Old Place Creek (replanted) had significantly higher median redox potential than Mill Creek (reference), Con Ed Tower (unplanted), and Saw Mill Creek South (unplanted) sites (all P <0.05), while redox potential values from Saw Mill Creek North, the other replanted site, were only significantly greater than Mill Creek (reference) values (P <0.05). The only other significant comparison was between the two marshes in the reference treatment group, Tufts Point and Mill Creek; Tufts Point had the higher values. Site was more influential than treatment.
The possible effect of replanting on patterns of seasonal shifts in
redox potential (i.e., treatment × seasonal effects) was
also investigated. Seasonal changes in redox values from fall to spring
included cases of significant increases, significant decreases, and no
significant change. Statistical comparison of Eh data
across seasons in the face of strong depth gradients was accomplished
by performing Student's t-tests on mean data for the three depth
zones (i.e., 0 to -2 cm, -3 to -5 cm, and below -5 cm) suggested
by the zonation patterns mentioned previously. The Eh values
in 24 of the 72 site-depth zone combinations (33%) were significantly
greater (more oxidized) in spring 1997 than in fall 1996 (P <0.05).
This phenomenon occurred at all depths and at sites under all treatment
regimes, including Con Ed Tower (unplanted), but not at all stations.
Eight combinations (11%) had more oxidized conditions in the fall, including
Old Place Creek (replanted) - Stations A and B , Con Ed Tower (unplanted)
- Station C, and Saw Mill Creek South (unplanted) - Stations A and C.
In terms of magnitude of significant changes, increases also predominated
over decreases. Seasonal changes in redox potential were significantly
larger where they increased (|
Eh| = 243±148 mV)
from fall to spring than where they decreased (|Eh|
= 79±51 mV) during that interval (P = 0.005). However, with
the largest number of station-depth zone combinations (32 of the 72,
or 44%), no significant seasonal change in redox potential was found.
The stations with the most consistently reducing conditions between -3 and -10 cm despite the change of seasons were Saw Mill Creek South (unplanted) - Stations A and B and Mill Creek (reference) - Stations B and C. Values for soil measurements organized by depth zone, and the results of seasonal statistical comparisons, are summarized by station in Appendix Tables F1-F3. As with EH values and profile shapes, seasonal redox changes did not sort into recognizable patterns by replanting status. Rather, patterns in seasonal change or lack thereof appear to be station specific.
Soluble Sulfide
Soluble sulfide was measured in 120 "peeper" water samples. As with the related Eh measurements, interstitial sulfide concentrations demonstrated distinct and fairly consistent patterns with respect to depth and season. Sulfide increased with depth (P <0.05) and was more abundant in fall 1996 than in spring 1997 at four of the six sites (Figures 29-31) The exceptions were Con Ed Tower (unplanted) and Tufts Point (reference).
At all stations within all sites, soluble sulfide concentrations above the soil surface were <2 µM, and were <80 µM within 3 cm of the sediment interface. With these data removed, site-to-site comparisons of sulfide showed significant differences (P <0.05), but treatment-to-treatment comparisons did not (i.e., Kruskal-Wallis, one-way, ANOVA-on-ranks test). Below the 5-cm depth in fall, Old Place Creek - Station C (replanted) had concentrations exceeding the analytical limit of the analysis employed (>3000 µM). Elsewhere, sulfide concentration approached this level only deeper than 14 cm in fall at Mill Creek - Station C (reference; 2700 µM) and at Saw Mill Creek South - Station D (unplanted; 2650 µM). In spring, values at Old Place Creek - Station A for all depths were <2 µM, while Old Place Creek - Station B contained 1250 µM below 14 cm in depth. Values at Saw Mill Creek North - Station C (also replanted) never exceeded 200 µM in fall, or 100 µM in spring. At the Con Ed Tower (unplanted) and especially Tufts Point (reference) sites, the concentration pattern seen elsewhere of high fall and low spring values was reversed. The most similar pairs of patterns did not share common replanting status: Saw Mill Creek South (unplanted) - Mill Creek (reference) (Figures 30B and 31B), and Con Ed Tower (unplanted) - Saw Mill Creek North (replanted) (Figure 30A and Figure 29B).
Total Organic Carbon
Marked differences among TOC values for surface soils from the Arthur Kill sites showed a closer association with site identity than with replanting treatment status (Figure 32). Comparison of TOC values by site demonstrated significant differences (Kruskal-Wallis, one-way, ANOVA-on-ranks test; P <0.001). Pairwise comparison of sites showed all pairs except Tufts Point:Mill Creek (reference stations) to be significantly different (Student-Newman-Keuls test, P <0.05).
Apparent loss of TOC at Con Ed Tower over the September-to-May interval (Figure 32) was significantly greater (P <0.05) than at any other site. Indeed, most stations at other sites show small increases in TOC from September to May. Site-to-site differences in levels and seasonal patterns of organic matter content obscured any underlying pattern by replanting treatment.
Labile Carbon
The LC content of sediment surface samples from two stations at each site yielded no significant differences by treatment (P = 0.091) or by station (P = 0.152; Kruskal-Wallis, one-way, ANOVA-on-ranks test), although the highest mean values occurred at Con Ed Tower (Figure 33). Mean May LC values correlated weakly, but significantly, with May values for %TOC (r2 = 0.413; n = 12; P = 0.024).
The LC content of tidewater samples from all sites (Saw Mill Creek North and Saw Mill Creek South are combined due to proximity to one another) was remarkably uniform: 160±28 µM-C/cm3 (mean ± standard deviation; n = 5). Standard deviations for replicate values from each site were within 10% of mean values for the same site except in the case of the Saw Mill Creek samples (standard deviation = 41% of mean).
Summary of Results
A brief summary of all of the biogeochemical data arranged by treatment, site, station, and season (Table 12) indicates the wide variation of values for median Eh, median sulfide concentration, and LC concentrations within sites and/or within treatments. Soil surface TOC was somewhat more consistent by site (e.g., low values at Old Place Creek and high values at Con Ed Tower), but not between sites within treatment groups. Expected patterns of treatment-related values were nowhere evident.
Seasonal changes in Eh, median sulfide, and TOC values also illustrated no discernable patterns associated with replanting status.
DISCUSSION
The Arthur Kill provides a locus for the study of an urban gradient system, with unique opportunities to investigate spatio-temporal scales of ecological patterning, the roles of disturbance, and the integral role of humans in the larger ecology of the system (McDonnell and Pickett 1990). Anthropogenic effects include continuous waste discharge into the air and water, episodic pollution events, upland and shoreline alteration, channel dredging, and maritime traffic effects. While the Arthur Kill salt marshes have much in common with salt marshes in other locations, they have a unique character that results from the interaction of such influences with the natural system, as well as from a very deliberate attempt to maintain estuarine wetland systems along a heavily populated, industrialized, and trafficked urban waterway. Extremely high and low soil redox potentials in close proximity, extremely high soil OC levels, and reversals of normal seasonal trends in soil organic carbon were found co-existing with values and trends more typical of systems that are clearly fulfilling wetlands ecological functions, including the provision of fisheries habitat.
Spatio-temporal patterns of porewater redox potential, soluble sulfide, and OC in marsh soils did not correspond with replanting status alone. Statistically significant differences were found for these biogeochemical measures with depth and season. However, these differences were not meaningful for assessment of replanting success because they appeared to owe more to the peculiarities of individual stations than to any common characteristics of replanted, unplanted, and reference marshes, or the particular sites in question (Table 12). Furthermore, quantitative differences among station data within each site were so large, and distributions of values at those stations were so skewed, as to render statistical differences uninterpretable in terms of replanting. No patterns characteristic of replanted, unplanted, or reference marshes were identified, nor were characteristic differences among sites fitting these treatment categories evident. Redox potentials, soluble sulfide and organic levels, depth profile shapes, and seasonal patterns appeared to be mediated by smaller-scale gradients in factors not clearly related to replanting. Our stations and sites were heterogeneous with respect to these factors, likely confounding our efforts to identify replanting-specific effects. Among those likely confounding factors were differences in grain size distribution (see Table 2 for results of grain size analysis), differences in surface and subsurface hydrology, differences in macrobiotic activity, and anthropogenic influences.
Influence of Grain Size Distribution
One possible confounding factor that could explain some of the variation in biogeochemical characteristics is difference in grain size distribution (see Table 2 for results of grain size analysis). Osgood and Zieman (1993) and Osgood et al. (1995) found that sandy marsh sediments associated with newly-developed natural marshes in Virginia had higher redox potentials, lower interstitial sulfide concentrations, and lower organic content than older, siltier sites nearby. Similarly, it appears that the upper layers of sediment at Old Place Creek and probably also Con Ed Tower have been maintained in relatively "young" condition by exposure to strong currents and by wave action associated with wakes from large vessels. Air enters the interstices among the grains of rigid (incompressible) sandy deposits as water drains or evaporates away during low tide, promoting penetration of oxic conditions to the extent allowed by soil column drying. Soils composed largely of silt and clay compress (collapse) as they lose water, leaving no air-filled interstices. Such soils continue to have low permeability to oxygen during subaerial exposure despite water loss, allowing anaerobic conditions to persist during low tide. Soils of intermediate grain size composition exhibit partial compression. Indeed, compressibility has been found to be linearly correlated to silt-clay content (Bradley and Morris 1990b).
Sediments from Arthur Kill marshes span the gamut of textures represented in the compressibility vs. silt-clay content regression of Bradley and Morris (1990b). The high energy sediments at Old Place Creek (replanted) - Station A fall at the totally incompressible end of the relationship (compressibility = 0%). Assuming a rapid rate of lateral drainage for this site as a result of proximity to a porous creekbank (Howes and Goehringer 1994), the presence of high redox potential well beneath the surface, the low organic content, and the lack of sulfide at this site are not surprising, even if the actual values of Eh (up to +450 mV) are beyond the maximum values generally reported for saltmarsh soils (e.g., Howes et al. 1986; de la Cruz et al. 1989; Craft et al. 1991; Osgood and Zieman 1993; Osgood et al. 1995; Thompson et al. 1995; Ewing et al. 1997; Madureira et al. 1997).
All stations at Saw Mill Creek North, Saw Mill Creek South, Tufts Point, and Mill Creek, with the exception of Mill Creek - Station D, fit near the totally compressible end (compressibility = 89%) of the regression, which should result in low Eh and sulfide values beneath the surface layers. While this is true in many cases, paradoxically high Eh and low sulfide values at some of these sites (e.g., Saw Mill Creek North - Station C in fall and Saw Mill Creek North - Station B in spring) must be the result of factors not associated with grain size distribution and its influence on compressibility and porosity. Predicted compressibility values for the remaining Old Place Creek and Mill Creek stations lie toward the incompressible end (Old Place Creek - Station B = 20%, Old Place Creek - Station C = 9%, Old Place Creek - Station D = 11%, and Mill Creek - Station D = 39%), and might thus be expected to have intermediate values of Eh, sulfide, and organic content. This is true of most OC content values at Old Place Creek (Station C in spring excepted), but not for Mill Creek - Station D, and not true regarding Eh or sulfide. Exceptionally low redox values (Old Place Creek - Stations B and C) and high sulfide levels (Old Place Creek - Station C) at these stations must again be attributed to factors other than grain size distribution.
While grain size determinations on Con Ed Tower sediments were not possible (see Chapter II, "Trace Metal Contaminants in Sediments and Ribbed-Mussels"), that site's exposed location along the Arthur Kill navigational channel suggests sandy/gravelly sediments similar to Old Place Creek - Station A, which fit with high redox potentials and low sulfide concentrations at Con Ed Tower - Station C in fall and Con Ed Tower - Stations A and D in spring. Although the TOC and LC methods used here did not allow us to distinguish between petroleum hydrocarbons and "natural" (e.g., algal, root/rhizome, detrital, and microbiological) organic matter, extraordinarily high OC levels at Con Ed Tower undoubtedly resulted from high levels of residual petroleum hydrocarbons (see Chapter III, "Petroleum Hydrocarbons in Sediments and Ribbed-Mussels"), visible as a tarry crust.
Subsurface Hydrology
Air can enter saltmarsh soil as a result of water removal by means of lateral subsurface drainage, which is a dominant mechanism along creekbanks, and by evapotranspiration as mediated by S. alterniflora and other vascular plants (Howes et al. 1986). Indeed, lateral drainage is the more rapid process where conditions permit, and its rate increases with proximity to the nearest creekbank (Howes and Goehringer 1994).
A pattern of exceptionally high Eh values (> +350 mV) that increase with depth regardless of season was observed at two of the eight replanted stations: Old Place Creek - Station A and Saw Mill Creek North - Station B. This type of redox profile has not been reported from natural marshes, but has occurred in marsh soil that had been experimentally drained of interstitial water for an extended period (Portnoy and Valiela 1997). Old Place Creek - Station A and Saw Mill Creek North - Station B, unlike other stations at those sites, were evidently subject to very rapid drainage and air entry during low tide. We believe that these processes occurred because these stations were closer to creekbanks than the other stations, despite all of the stations being at the same intertidal elevation. In the case of Old Place Creek - Station A, rapid drainage was facilitated by a very coarse grain size distribution (i.e., complete incompressibility). In the case of Saw Mill Creek North - Station B, it appears that drainage was promoted by the heavy riddling of the adjacent bank by fiddler crab burrows.
Surface Hydrology
Care was taken to ensure that each station was located at the same tidal height (i.e., 0.2 m above mean sea level) so as to eliminate possible variations in biogeochemistry stemming from differing frequency and duration of tidal flooding that attend small differences in elevation (Cahoon and Reed 1995). However, our six stations were subject to differing wave and current regimes, and to differing surface water quality. Old Place Creek and Con Ed Tower were subject to the greatest wave and current energies (wind-, tide-, and vessel-driven), as indicated by the low silt/clay content of Old Place Creek sediments. Differences in silt/clay content among stations within Old Place Creek suggest hydrological differences on a scale of a few meters or less at that site, resulting in differences in deposition of detrital material that contributes to TOC and LC. Sediments at more sheltered sites (e.g., Saw Mill Creek North and Saw Mill Creek South) have more uniform sediment textures and relatively less variable TOC values. Less uniformity at Tufts Point and Mill Creek suggest at least occasional episodes of higher energy, despite sheltered locations.
During 1998, there were also north-to-south gradients in average water quality measures in the Arthur Kill, including dissolved oxygen, inorganic nutrients, fecal coliform counts, and degree of water column stratification (NYCDEP 1998). Gradients in these or other unmeasured water quality parameters could result in biogeochemical differences among sites. Considering the high level of LC in the water adjacent to these sites, tidal inundation may dominate biogeochemical processes in this urban marsh complex, with the greatest effect being exerted in the north.
Macrobiotic Activity
Leaving aside those sites with heterogeneous sediment texture, the remaining sites -- based on their porewater Eh and soluble sulfide values -- demonstrated large differences among stations within site. Two elements of site macrobiota may be influencing these differences: saltmarsh cordgrass and burrowing crabs. At all vegetated sites except Con Ed Tower, the redox and sulfide profile differences may be attributable to highly localized variations in soil aeration associated with variations in density of S. alterniflora roots and rhizomes (Luther and Church 1988; Madureira et al. 1997), despite station-to-station similarity in above-ground biomass (C. Alderson et al., Salt Marsh Restoration Team, Natural Resources Group, New York City Parks, 200 Nevada Ave., Staten Island, NY, unpubl. data). Soil aeration via diffusion from S. alterniflora roots may explain our frequent observation of higher redox values in spring as compared with fall in irregular, narrow depth bands (regions of maximum root density) at vegetated sites (see Figures 26-28), and as also reported elsewhere (Howes et al. 1981; de la Cruz et al. 1989).
The same kind of fall-spring redox increases at Con Ed Tower are not explainable in terms of cordgrass aeration; there was no cordgrass. We suspect that subsurface drainage of presumably incompressible, sandy sediments there was more evident during our spring visit than during our fall one, thus mimicking the redox behavior of vegetated sites.
Some sites (i.e., Saw Mill Creek North and Tufts Point) were heavily populated by mixed populations of two fiddler crab species: the Atlantic marsh fiddler, Uca pugnax, and the redjointed fiddler, U. minax. Fiddler crabs contribute substantially to increasing Eh values and decreasing soluble sulfide along inner-marsh-to-creekbank gradients (Gardner et al. 1988). This effect is due to the increase in surface area (Katz 1980), and hence gas exchange (Montague 1981), promoted directly by the presence of the burrows, and indirectly by the increase in production of cordgrass (Bertness 1985) with its attendant increase in soil aeration potential. Close to the creekbank, we suspect an added effect due to a propensity for burrows to facilitate lateral drainage. These burrow effects are probably the cause of extremely variable Eh values at Saw Mill Creek North and Tufts Point, and of low soluble sulfide values at all Saw Mill Creek North stations and at Tufts Point - Station D in fall. Those low fall values at Tufts Point - Station D created an apparent reversal of the expected seasonal pattern of high fall - low spring values. Results might have been different had another Tufts Point station been utilized in September-October 1996. Lower densities of fiddler crabs at Saw Mill Creek South resulted in much less variable values of redox and soluble sulfide.
Anthropogenic Influences
Another confounding factor was continued anthropogenic impacts. The Arthur Kill marshes are receiving substantial and unequal organic matter subsidies (including petroleum hydrocarbons) as a result of their location on an urban waterway. This may be the cause of extremely low redox and high TOC values at Old Place Creek - Station C. Other authors have not reported Eh below -350 mV in saltmarsh soils (e.g., Patrick and DeLaune 1977; Howes et al. 1981; Armstrong et al. 1985; de la Cruz et al. 1989; Bertness 1991; Osgood and Zieman 1993; Ewing et al. 1997; Madureira et al. 1997). Given the unique history of the site, extremely low redox values at Old Place Creek suggest pockets of very reduced organic material, probably petroleum of patchy spatial distribution. Indeed, the occasional odor of volatile petroleum components from soil at that site, and a single exceptional value for TOC at Old Place Creek - Station C in spring (i.e., 8% among values ranging from 0.1% to 2.5%, and well above the maximum value predicted from grain sizes), support the existence of such pockets. We speculate that pockets of volatile hydrocarbons persisted at Old Place Creek because spill remnants there were buried soon after deposition, quite possibly by the replanting process itself. Burial prevented weathering, so volatile components persisted. By contrast, spilled petroleum left exposed at Con Ed Tower was heavily weathered, leaving only surficial tarry deposits that did not produce low redox values despite exceedingly high TOCs.
The very high TOC values (approaching 50%) at Con Ed Tower exceed the values (~25-40%) in even very peaty unaltered marsh soil in Massachusetts (Portnoy and Giblin 1997), and far exceed those reported from other saltmarsh soils (Williams et al. 1994). The refractory nature of the weathered petroleum that accounts for these high values at Con Ed Tower is evident in the comparison of LC at Con Ed Tower stations with that at other stations with far lower TOC values. While TOC values are as much as five times higher at Con Ed Tower than at other sites, LC values are only marginally higher than at most other sites, and are statistically indistinguishable from all of the other sites as a whole. Apparently, the excessive OC at Con Ed Tower cannot be metabolized readily by the microbiota, even under aerobic circumstances.
We found no evidence for persistent derangement of biogeochemical metabolic processes resulting from the 1990 oiling of the Arthur Kill marshes. Heterotrophic bacterial activity, and presumably biogeochemical function, in salt marshes can become highly disturbed by oiling (Vacelet et al. 1985). In particular, polycyclic aromatic hydrocarbons (PAHs) are known to inhibit sulfate reduction while stimulating methanogenesis, probably via elimination of substrate competition between sulfate-reducing and methanogenic bacteria (Keine and Capone 1984). If acute effects like these persisted as chronic conditions in the Arthur Kill marshes, especially at Old Place Creek and Con Ed Tower, the result should be low Eh values associated with fermentative methane production but without sulfide production. Where such disturbances were expected to be minimal (i.e., reference marshes), a loose correlation would therefore be expected between Eh and soluble sulfide, since sulfate reduction is a major metabolic process in sulfate-rich marine waters in contact with organic matter in the absence of molecular oxygen. Any disturbance in the Eh:sulfide relationship in other treatment groups or at other sites, as evidenced by significantly different regressions, would indicate an important shift in biogeochemical function. The relationships between soluble sulfide concentrations and mean Eh in the Arthur Kill salt marshes were logarithmic, as anticipated from the Nernst equation (Figure 34). All regressions by treatment were significant (P <0.003).
Analysis of covariance with these three regressions demonstrated no significant difference (P = 0.52) in the relationship when data were plotted by treatment. Neither did the six regressions calculated by station demonstrate any significant difference (P = 0.68). Thus, no persistent disturbance was evident in the Eh:sulfide relationship associated with replanting status or stations, providing no evidence for inhibition of sulfate reduction. Indeed, the higher soluble sulfide levels in the Arthur Kill marshes (3+ mM) were toward the high end of values reported for saltmarsh soil interstitial water, but within values (approaching 6 mM) reported for Massachusetts (Teal and Howes 1996). Sufficient data were not available for analysis of relationships grouped by individual stations.
While we found no metabolic disturbance attributable to oiling, the gross composition of sediments at some unplanted and replanted sites was measurably different from that of all of the other sites. Extraordinarily high TOC values resulted from petroleum residues. The OC at Con Ed Tower, and possibly at Old Place Creek - Station C, presumably residual petroleum at least in part, was not subject to the same deposition, sorting, retention, and loss processes as elsewhere. When plotted against percent silt/clay content, most %TOC values fell below the line defined by the equation, %TOC = 0.20 × %silt/clay, which is roughly consistent with the positions of most Old Place Creek and spring Mill Creek values (Figure 35). The data from Old Place Creek and Mill Creek (spring only) were chosen for regression to represent a "maximum biogenic TOC limit," because they were from stations exhibiting high TOC over a wide range of grain size distributions, and from sites that were not visibly tarry. The Old Place Creek - Station C spring value, which has more than twice the predicted TOC content, is an exception to this scheme. The TOC values for Con Ed Tower were also beyond predictions. While Con Ed Tower values were not plotted due to the inability to perform grain size analysis, the %TOC for Con Ed Tower - Stations A, B, and D in fall and Station A in spring (35-49%) was 10-24% beyond the predicted maximum of 25% for pure silt/clay sediments. The TOC values (8-16%) for the remaining Con Ed Tower station-date combinations suggest either substantial silt content, or sandy sediments with values exceeding predictions.
Hydrocarbon contamination of salt marshes decreases naturally over time, as demonstrated by long-term monitoring of the Ile Grande marsh damaged by the Amoco Cadiz spill in 1978 (Mille et al. 1998). Recent laboratory investigations have indeed suggested that bacterial community structure in oil-contaminated saltmarsh soil returns toward that of uncontaminated soil as oil components are degraded (Bachoon 1999), suggesting a concomitant restoration of biogeochemical function. Paradoxically, the urban environs of the Arthur Kill marshes may have aided this recovery of soil microbiological function. Fertilization of oil-contaminated saltmarsh soil with inorganic nitrogen has been shown to accelerate the bacterial metabolism of alkane and PAH fractions (Jackson and Pardue 1999), as well as to directly stimulate growth of S. alterniflora (Lin and Mendelssohn 1998). High nutrient levels associated with urban discharges may well have aided these processes in this case. Average dissolved inorganic nitrogen (i.e., ammonium + nitrate + nitrite) for the waters of the northern end of the Arthur Kill during the summer of 1998 was in the range of 72-89 µg-at/L, and total phosphorus for the same area and time was in the range of 7-10 µg-at/L (NYCDEP 1998).
CONCLUSIONS
Had confounding factors not been active, we still might have had difficulty detecting clear differences in marsh biogeochemistry attributable to the replanting efforts. We believe that previous investigators readily found such distinctions because the soils in their restoration sites were not originally marsh soils with previous exposure to regular tidal inundation (e.g., Craft et al. 1991; Thompson et al. 1995). At our Arthur Kill sites, by contrast, restoration was attempted by replanting S. alterniflora in formerly vegetated marsh soil which had been, and continued to be, exposed to regular tidal inundation (i.e., without hydrological regime alteration).
While it may require years for nonmarsh soils to attain the organic content and other biogeochemical characteristics of natural marsh soil, we propose that it also requires years for marsh soil to lose its organic content, corresponding redox and sulfide profiles, and perhaps other biogeochemical properties if the hydrological regime remains unaltered. Our data suggest that Arthur Kill soils have retained their biogeochemical characteristics for several years despite oiling damage and subsequent periods of barrenness. This, in part, may explain why replanting has been extraordinarily successful in re-establishing vegetation in oil-damaged salt marshes in this location.
REFERENCES CITED
Armstrong, W.; Wright, E.J.; Lythe, S.; Gaynard, T.J. 1985. Plant zonation and the effects of spring neap tidal cycle on soil aeration in a Humber salt marsh. J. Ecol. 73:323-339.
Bachoon, D. 1999. Changes in salt marsh sediment microbial community structure in response to petroleum contamination. Diss. Abstr. Int. B Sci. Eng. 60:473.
Bergen, A.; Alderson, C.; Bergfors, R.; Aquila, C.; Matsil, M.A. 2000. Restoration of a Spartina alterniflora salt marsh following a fuel oil spill, New York City, NY. Wetlands Ecol. Manag. 8:185-195.
Bertness, M.D. 1985. Fiddler crab regulation of Spartina alterniflora production on a New England salt marsh. Ecology 66:1042-1055.
Bertness, M.D. 1991. Zonation of Spartina patens and Spartina alterniflora in a New England salt marsh. Ecology 72:138-148.
Boesch, D.F.; Turner, R.E. 1984. Dependence of fishery species on salt marshes: the role of food and refuge. Estuaries 7:460-468.
Bradley, P.M.; Morris, J.T. 1990a. Influence of oxygen and sulfide concentration on nitrogen uptake kinetics in Spartina alterniflora. Ecology 71:282-287.
Bradley, P.M.; Morris, J.T. 1990b. Physical characteristics of salt marsh sediments: ecological implications. Mar. Ecol. Prog. Ser. 61:245-252.
Bradley, P.M.; Morris, J.T. 1992. Effect of salinity on the critical nitrogen concentration of Spartina alterniflora Loisel. Aquat. Bot. 43:149-161.
Caçador, I.; Vale, C.; Catarino, F. 1996. Accumulation of Zn, Pb, Cu, Cr and Ni in sediments between roots of the Tagus estuary salt marshes, Portugal. Estuarine Coastal Shelf Sci. 42:393-403.
Cahoon, D.R.; Reed, D.J. 1995. Relationships among marsh surface topography, hydroperiod, and soil accretion in a deteriorating Louisiana salt marsh. J. Coastal Res. 11:357-369.
Craft, C.B.; Seneca, E.D.; Broome, S.W. 1991. Porewater chemistry of natural and created marsh soils. J. Exp. Mar. Biol. Ecol. 152:187-200.
de la Cruz, A.; Hackney, C.T.; Bhardwaj, N. 1989. Temporal and spatial patterns of redox potential (Eh) in three tidal marsh communities. Wetlands 9:181-190.
DeLaune, R.D.; Patrick, W.H., Jr.; Buresh, R.J. 1979. Effect of crude oil on a Louisiana Spartina alterniflora marsh. Environ. Pollut. 20:21-31.
DeLaune, R.D.; Reddy, C.N.; Patrick, W.H., Jr. 1981. Effect of pH and redox potential on concentration of dissolved nutrients in an estuarine sediment. J. Environ. Qual. 10:276-279.
Draxler, A.F.J. 1995. Changes in sediment biogeochemistry resulting from cessation of sewage sludge dumping in the New York Bight. In: Studholme, A.L.; O'Reilly, J.E.; Ingham, M.C., eds. Effects of the cessation of sewage sludge dumping at the 12-mile site: 12-Mile Dumpsite Symposium, Long Branch, New Jersey, June 1991. NOAA Tech. Rep. NMFS 124:133-143.
Ewing, K.; McKee, K.L.; Mendelssohn, I.A. 1997. A field comparison of sublethal stress in the salt-marsh grass Spartina patens. Estuaries 20:48-65.
Gardner, L.R.; Wolaver, T.G.; Mitchell, M. 1988. Spatial variations in the sulfur chemistry of salt marsh sediments at North Inlet, South Carolina. J. Mar. Res. 46:815-836.
Haines, E.B. 1979. Interactions between Georgia salt marshes and coastal waters: a changing paradigm. In: Livingston, R.J, ed. Ecological processes in coastal and marine systems. New York, NY: Plenum Press; p. 35-46.
Hershner, C.; Lake, J. 1980. Effects of chronic oil pollution on a salt-marsh community. Mar. Biol. (Berl.) 56:163-173.
Hesslein, R.H. 1976. An in situ sampler for close interval porewater studies. Limnol. Oceanogr. 21:912-914.
Howarth, R.W.; Giblin, A. 1983. Sulfate reduction in the salt marshes of Sapelo Island, Georgia. Limnol. Oceanogr. 28:70-82.
Howes, B.L.; Dacey, J.W.H.; Goehringer, D.D. 1986. Factors controlling the growth of Spartina alterniflora: feedbacks between above-ground production, sediment oxidation, nitrogen and salinity. J. Ecol. 74:881-898.
Howes, B.L.; Goehringer, D.D. 1994. Porewater drainage and dissolved organic carbon and nutrient losses through the intertidal creekbanks of a New England salt marsh. Mar. Ecol. Prog. Ser. 114:289-301.
Howes, B.L.; Howarth, R.W.; Teal, J.M.; Valiela, I. 1981. Oxidation-reduction potentials in a salt marsh: spatial patterns and interactions with primary production. Limnol. Oceanogr. 26:350-360.
Howes, B.L.; Dacey, J.W.H.; King, G.M. 1984. Carbon flow through oxygen and sulfate reduction pathways in salt marsh sediments. Limnol. Oceanogr. 29:1037-1051.
Jackson, W.A.; Pardue, J.H. 1999. Potential enhancement of biodegradation of crude oil in Louisiana salt marshes using nutrient amendments. Water Air Soil Pollut. 109:343-355.
Katz, L.C. 1980. Effects of burrowing by the fiddler crab, Uca pugnax (Smith). Estuarine Coastal Mar. Sci. 11:233-237.
Keine, R.P.; Capone, D.G. 1984. Effects of organic pollutants on methanogenesis, sulfate reduction and carbon dioxide evolution in salt marsh sediments. Mar. Environ. Res. 13:141-160.
Koch, M.S.; Mendelssohn, I.A.; McKee, K.L. 1990. Mechanism for the hydrogen sulfide-induced growth limitation in wetland macrophytes. Limnol. Oceanogr. 35:399-408.
Kostka, J.E.; Luther, G.W., III. 1995. Seasonal cycling of Fe in saltmarsh sediments. Biogeochemistry 29:159-181.
LaSalle, M.W.; Landin, M.C.; Sims, J.G. 1991. Evaluation of the flora and fauna of a Spartina alterniflora marsh established on dredged material in Winyah Bay, South Carolina. Wetlands 11:191-208.
Li, Y.; Morris, J.T.; Yoch, D.C. 1990. Chronic low level hydrocarbon amendments stimulate plant growth and microbial activity in salt-marsh microcosms. J. Appl. Ecol. 27:159-171.
Lin, Q.; Mendelssohn, I.A. 1998. The combined effects of phytoremediation and biostimulation in enhancing restoration and oil degradation of petroleum contaminated wetlands. Ecol. Eng. 10:263-274.
Luther, G.W., III; Church, T.M. 1988. Seasonal cycling of sulfur and iron in porewaters of a Delaware salt marsh. Mar. Chem. 23:295-309.
Luther, G.W., III; Kostka, J.E.; Church, T.M.; Sulzberger, B.; Stumm W. 1992. Seasonal iron cycling in the salt-marsh sedimentary environment: the importance of ligand complexes with Fe(II) and Fe(III) in the dissolution of Fe(III) minerals and pyrite, respectively. Mar. Chem. 40:81-103.
Madureira, M.J.; Vale, C.; Simões Gonçalves, M.L. 1997. Effect of plants on sulphur geochemistry in the Tagus salt-marshes sediments. Mar. Chem. 58:27-37.
McDonnell, M.J.; Pickett, S.T.A. 1990. Ecosystem structure and function along urban-rural gradients: an unexploited opportunity for ecology. Ecology 7:1232-1237.
Mille, G.; Munoz, D.; Jacquot, F.; Rivet, L.; Bertrand, J.-C. 1998. The Amoco Cadiz oil spill: evolution of petroleum hydrocarbons in the Ile Grande salt marshes (Brittany) after a 13-year period. Estuarine Coastal Shelf Sci. 47:547-559.
Minello, T.J.; Zimmerman, R.J. 1992. Utilization of natural and transplanted Texas salt marshes by fish and decapod crustaceans. Mar. Ecol. Prog. Ser. 90:273-285.
Mitsch, W.J.; Gosselink, J.G. 1986. Wetlands. New York, NY: Van Nostrand Reinhold; 539 p.
Montague, C.L. 1981. The influence of some larger animals on subsurface metabolic processes in intertidal sediments. Estuaries 4:289.
Moy, L.D.; Levin, L.A. 1991. Are Spartina marshes a replaceable resource? A functional approach to evaluation of marsh creation efforts. Estuaries 14:1-16.
Newell, S.Y. 1993. Decomposition of shoots of a salt-marsh grass: methodology and dynamics of microbial assemblages. Adv. Microb. Ecol. 13:301-326.
NYCDEP [New York City Department of Environmental Protection]. 1998. New York Harbor Water Quality Survey. 1998. New York, NY: New York City Department of Environmental Protection; 74 p.
Orion Research, Inc. 1983. Instruction manual platinum redox electrodes. Cambridge, MA: Orion Research.
Osgood, D.T.; Santos, M.C.V.; Zieman, J.C. 1995. Sediment physico-chemistry associated with natural marsh development on a storm-deposited sand flat. Mar. Ecol. Prog. Ser. 120:271-283.
Osgood, D.T.; Zieman, J.C. 1993. Spatial and temporal patterns of substrate physicochemical parameters in different-aged barrier island marshes. Estuarine Coastal Shelf Sci. 37:421-436.
Patrick, W.H., Jr.; DeLaune, R.D. 1977. Chemical and biological redox systems affecting nutrient availability in coastal wetlands. Geosci. Man 18:131-137.
Portnoy, J.W.; Giblin, A.E. 1997. Effect of historic tidal restriction on salt marsh sediment chemistry. Biogeochemistry 36:275-303.
Portnoy, J.W.; Valiela, I. 1997. Short-term effects of salinity reduction and drainage on salt-marsh biogeochemical cycling and Spartina (cordgrass) production. Estuaries 20:569-578.
Strickland, J.D.H.; Parsons, T.R. 1972. A practical handbook of seawater analysis. Bull. Fish. Res. Board Can. 167.
Teal, J.M.; Howes, B.L. 1996. Interannual variability of a salt-marsh ecosystem. Limnol. Oceanogr. 41:802-809.
Thompson, S.P.; Paerl, H.W.; Go, M.C. 1995. Seasonal patterns of nitrification and denitrification in a natural and a restored salt marsh. Estuaries 18:399-408.
Thomson, A.D.; Webb, K.L. 1984. The effect of chronic oil pollution on salt-marsh nitrogen fixation (acetylene reduction). Estuaries 7:2-11.
Vacelet, E.; Mille, G.; Le Campion-Alsumard, T.; Plante-Cuny, M.-R. 1985. Evolution des populations bacteriennes dans les sediments des marais maritimes de l'Ile Grande pollues par l'Amoco Cadiz. Int. Rev. Gesamt. Hydrobiol. 70:815-827.
Williams, T.P.; Bubb, J.M.; Lester, J.N. 1994. Metal accumulation within salt marsh environments: a review. Mar. Pollut. Bull. 28:277-290.
Yamamuro, M.; Kayanne, J. 1995. Rapid direct determination of organic carbon and nitrogen in carbonate-bearing sediments with a Yanaco MT-5 CHN analyzer. Limnol. Oceanogr. 40:1001-1005.
Zedler, J.B. 1993. Canopy architecture of natural and planted cordgrass marshes: selecting habitat evaluation criteria. Ecol. Appl. 3:123-138.
ZoBell, C.E. 1946. Studies of redox potential of marine sediments. Bull. Am. Assoc. Petrol. Geol. 30:477-513.