sign in sign up
<<
Home , Wetland

Five Decades of Wetland Development of a Constructed Tidal Salt , North Carolina, USA

Aaron Noll, Courtney Mobilian, Christopher Craft

Ecological Restoration, Volume 37, Number 3, September 2019, pp. 163-170 (Article)

Published by University of Wisconsin Press

For additional information about this article https://muse.jhu.edu/article/731195

Access provided at 20 Aug 2019 15:22 GMT from Indiana University Libraries RESEARCH ARTICLE

Five Decades of Wetland Soil Development of a Constructed Tidal , North Carolina, USA Aaron Noll, Courtney Mobilian, and Christopher Craft

ABSTRACT We periodically measured soil properties (bulk density, organic , total , and total ) in a tidal salt marsh constructed in 1970 over the past five decades (1984, 1995, 1998, 2017) to identify trajectories of soil formation over time. Bulk density was greater in surface (0–10 cm) than subsurface (10–30 cm) and decreased with time at a similar rate in both depths. Percent organic carbon (C), nitrogen (N), and total phosphorus (P) increased with time, with in surface soils increasing three times faster than subsurface soils. Surface C and N pools (g/m2, 0–10 cm) exhibited asymptotic trajectories with time, with rapid accumulation in the first two decades that began to equilibrate after 25 years, while subsurface pools of C and N continued to increase in a linear manner. Soil P pools did not exhibit a significant trend related to marsh age. Soil N:P increased in both depths over time, suggesting gradual relaxation of N-limitation of these systems. Our findings can help restoration ecologists identify soil properties, including bulk density, C, N, and C:N suitable for gauging restoration success and better estimate the time frame necessary for recovery of soil-based services.

Keywords: constructed , soil carbon, soil nitrogen, salt marsh restoration

Restoration Recap • • Tidal salt planted with Spartina alterniflora con- • Constructed and restored tidal salt marshes will likely tinue to sequester organic carbon and nitrogen nearly sequester C and N for the foreseeable future assuming five decades after establishment. marshes can accrete sufficient mineral and soil • During the first 25 years, accumulation of C and N was organic matter to keep pace with , and avoid confined to surface (0–10 cm) soils after which C and N human activities such as dredging and boat traffic that enrichment shifted to subsurface (10–30 cm) soils. may increase lateral erosion.

idal marshes are valued for their ability to provide destroyed. However, it may take years for restored wetlands Timportant ecosystem services including shoreline sta- to re-establish some functions of natural systems, particu- bilization, quality improvement, , and larly those functions related to soils (Craft et al. 2003). Tidal (Boesch and Turner 1984, Mcleod marsh and other wetland soils provide environmental con- et al. 2011, Shepard et al. 2011). However, they have been ditions that facilitate biogeochemical and nutrient cycling. subject to widespread degradation and loss stemming from Some unique functions of wetland soils that are distinct coastal development, hydrologic alterations, and sea level from terrestrial soils include high rates of , rise (Pendleton et al. 2012, Craft et al. 2009, Neubauer 2013, nutrient burial from , and accelerated rates Vincent et al. 2013). of organic matter and carbon sequestration. Accumulation Restoration and creation of tidal marshes are used to of organic matter is essential to restoring other ecosystem replace ecosystem services lost when they are degraded or services such as improvement (e.g., denitrifi- cation), providing organic C to support heterotrophic food webs, production of economically valuable (LaSalle Ecological Restoration Vol. 37, No. 3, 2019 et al. 1991, Craft et al. 2003) and building elevation in the ISSN 1522-4740 E-ISSN 1543-4079 ©2019 by the Board of Regents of the University of Wisconsin System.

September 2019 ECOLOGICAL RESTORATION 37:3 • 163 face of rising sea level (Nyman et al. 2006, Chmura and Methods Hung 2004, Turner et al. 2002, Hatton et al. 1983). Wetland soil development is a slow but critical process Site Description to restoring wetland functions and re-establishing lost We measured soil properties in a Spartina alterniflora ecosystem services. In addition to the slow pace of soil Loisel tidal salt marsh (34°03'49.15" N, 77°55'40.25" W), development, successful restoration of wetland soils can Snow’s Cut, that was constructed in 1970 on a dredge spoil be challenging due to factors such as altered on the Cape Fear , NC. The U.S. Army Corps and degraded initial substrate, including dredged material of Engineers funded construction of the marsh to dem- that contains little organic matter or nitrogen (Cornell et al. onstrate the use of wetland to stabilize dredge 2007), terrestrial soils where surface (O, A) horizons are spoil material (Woodhouse and Seneca 1974). Marsh con- excavated to achieve target elevations (Craft et al. 2002), struction was accomplished by grading the site to intertidal and agricultural soils with little soil organic matter (Marton elevation, then planting Spartina alternifloraat elevations et al. 2014). In particular, sufficient N and C are needed between mean sea level and mean high water. The marsh to sustain and support biogeochemical is 0.8 hectares in size, occupying a riverine geomorphic processes such as and denitrification in position with a tidal range of approximately 1.2 meters and these N limited systems. Craft et al. (2003) estimated that floodwater ranging from 7–10 parts per thousand 100 g N/m2 and 1000 g C/m2 are needed to achieve full or (Craft et al. 1988). The site is located on the eastern side of complete restoration of these functions and accumulation the island and has served as the focus of numerous studies of such sizable pools may require decades. on ecosystem development of restored tidal marshes (Craft The chronosequence (space for time) approach has been et al. 1988, 1999, 2003, Craft and Sacco 2003, Zheng et al. widely used to assess successional trajectories in restored 2004, Cornell et al. 2007). wetlands and wetland soil, but its explanatory power can be limited by variation due to factors other than ecosystem Sample Collection and Analysis age. For example, it may be difficult to completely match Soils were collected in 1984, 1995, and 1998 and findings hydroperiod (depth, duration, and frequency of inunda- were published in Craft et al. (1988), (1999), and (2003), tion), salinity, vegetation type, or geomorphic position respectively. Soils were sampled again in 2017 using the among restored sites of differing ages; though for tidal identical sampling design, field and lab protocols, and wetlands, tidal hydroperiod and vegetation type (Spartina analyses that were used in previous work at the site. Ten alterniflora) can be readily matched when restoring salt soil cores were collected in total. Five cores were randomly marshes. However, the efficacy of the chronosequence collected from the streamside edge and five cores were col- approach (i.e., different aged sites can be sampled at one lected from the marsh interior. Soil cores were sectioned time) makes it the primary method for evaluating succes- into 0–10 cm and 10–30 cm subsamples in the field, then sion and ecosystem development over time for natural and returned to the lab and air-dried, weighed, ground, and restored (Crocker and Major 1955, Craft et al. sieved through a 2-mm mesh diameter screen. Screening 2003, Ballantine and Schneider 2009, Yang and Guo 2018). removed large roots but included fine roots, detritus and An alternative approach is to sample the same site over soil organic matter. time but this approach is seldom used (Craft et al. 2002, Measured soil properties included bulk density, organic He et al. 2016) because of the long time intervals required carbon, total nitrogen, and total phosphorus. Bulk density between sampling. was calculated by weighing air-dried cores of a known The goal of this study is to evaluate ecosystem develop- volume and applying a moisture correction factor deter- ment of a tidal wetland constructed and planted nearly mined from subsamples oven-dried at 105°C (Blake and five decades ago. We tracked changes in soil properties at Hartge 1986). Total organic carbon (TOC) and nitrogen periodic intervals (1984, 1995, 1998, 2017) since its con- (N) were measured using a Perkin-Elmer CHN analyzer struction in 1970. The work is important because 1) soil (Perkin-Elmer, Norwalk, CT, USA). Total P development is a slow process and decades may pass before was determined using the HNO3-HClO4 soil digestion soils and soil-based functions are completely restored, and method followed by colorimetric analysis (Sommers and 2) nearly all studies of soil change following wetland resto- Nelson 1972). Estuarine sediment (NIST standard no. ration are based on the chronosequence approach and its 1646a) was used as a baseline standard in P digestions, inherent variability in hydrology, salinity, nutrients, land yielding a recovery rate of 94%. An in-house salt marsh use history and other abiotic factors among sites, rather standard was used for C and N, returning recovery rates than the repeated measures approach used here. of 101% and 97% for C and N, respectively. Nutrient (C, N, P) pools (g/m2) in the 0–10 and 10–30 cm depths were calculated using bulk density and C, N, and P concentra- tions. Linear regressions of soil properties versus marsh

164 • September 2019 ECOLOGICAL RESTORATION 37:3 Figure 1. Linear regressions of (a) Bulk Density, (b) % Organic C, (c) % N, and (d) Total P (μg/g) as a function of marsh age. Each data point represents n = 10. Means and standard errors are plotted. Solid line: 0–10 cm trend, Dashed line: 10–30 cm trend. Time 0 (1970) data represents soil properties (bulk density, %C, %N, total P) of the 10–30 cm depth in 1984. See Methods for details. age were performed in Sigmaplot 13.0 (Systat Software, reasonably estimates values in the 0–10 cm depth where Inc., San Jose, CA) to assess changes in soil properties over most accumulation occurs in the early years after marsh time. Separate regressions were used to analyze change in establishment, but the model may underestimate rates of surface (0–10 cm) versus subsurface (10–30 cm) depths. soil change in the subsurface (10–30 cm) soil during the GIS analysis was used to evaluate changes in area and first 14 years. configuration of the Snow’s Cut dredge spoil island from 1984 to 2018. Maps and calculations of change in island Results area and rates of erosion and accretion were performed using Google Earth Engine API (Gorelick et al. 2017), Soil bulk density decreased in both surface (0–10 cm) and Google Earth Pro (v. 7.3.2), and ArcMap (v. 10.6, Environ- subsurface (10–30 cm) soils as the marsh aged. Although mental Systems Research Institute, Redlands, CA). bulk density was consistently greater in subsurface than Since no soils were collected at the time of marsh estab- surface soils, the rate of decrease over time was similar lishment, rates of nutrient accumulation were based on the between the two depths (Figure 1a). Percent soil organic C assumption that N, P, and organic C pools in the 10–30 cm and N increased with marsh age, with the rate of increase depth in 1984 represent nutrient pools of surface and in both C and N occurring three times more rapidly in subsurface soils in 1970. Since the site was constructed surface than subsurface soils (Figures 1b, 1c). Soil P also on dredge spoil, the 0–10 cm layer following spoil deposi- increased with marsh age (p < 0.10) with rates of increase tion would be reasonably expected to be the same as the nearly three times greater in surface compared to subsur- 10–30 cm depth in 1970. Assuming little has changed, face soils (Figure 1d). what was present in the 10–30 cm depth in 1984 probably Soil C:N (molar ratio) did not significantly vary with was present in the 0–10 cm depth in 1970. Our model marsh age in either surface or subsurface soils (Figure 2a).

September 2019 ECOLOGICAL RESTORATION 37:3 • 165 Figure 2. Linear regressions of (a) C:N Ratio and (b) N:P Ratio as a function of marsh age. Each data point represents n = 10. Means and standard errors are plotted. Solid line: 0–10 cm trend, Dashed line: 10–30 cm trend; Dotted line–Nutrient limitation thresholds (see discussion section for explanation). Time 0 (1970) data represents soil properties (C:N, N:P) of the 10–30 cm depth in 1984. See Methods for details.

Soil C:N also did not differ between surface and subsurface soils. Soil N:P increased over time (p < 0.10), with N:P being somewhat higher in surface soils but showing similar rates of increase in both depths (Figure 2b). Figure 3. Linear regressions of (a) Organic C pool, Organic C and total N pools (g/m2) in the top 30 centi- (b) N pool, and (c) P pool (g/m2) as a function of meters increased approximately three-fold over the 46-year marsh age. Each data point represents n = 10. Means period, increasing from 1770 to 4637 g C/m2 and 75 to and standard errors are plotted. Solid line: 0–10 cm trend, Dashed line: 10–30 cm trend. Time 0 (1970) 264 g N/m2 (Table 1), despite a decrease in soil bulk density. data represents soil C, N, and P pools of the 10–30 cm Soil organic C and N pools both exhibited depth-specific depth in 1984. For the 0–10 cm depth, the pool size successional trajectories (Figures 3a, 3b). Organic C and is half the size of the 10–30 cm pool since its volume N pools in surface soils showed an asymptotic increase (0–10 cm) is half of 10–30 cm. See Methods for details. that initially was rapid but then began to equilibrate after approximately 20–25 years. Organic C and N pools in subsurface soils were slower to initially develop, but after 46 years continue to increase linearly. Phosphorus pools (0–30 cm) did not increase with marsh age, changing from

166 • September 2019 ECOLOGICAL RESTORATION 37:3 Table 1. Soil organic C, N, and P pools (g/m2, n = 10 cores) in the constructed marsh in 1970, 1984, 1995, 1998, and 2017. Accumulation rate of C, N, and P (g/m2/yr) for 1970–1995 (0–10 cm depth) and 1995–2017 (10–30 cm depth) are in parentheses. Accumulation in the natural reference marsh (1998), seven constructed marshes, and six natural marshes (based on 210Pb dating) are from Craft et al. (2003). Note that the 10–30 cm depth represents two times the volume of the 0–10 cm depth. Since no soils were collected at the time of restoration, the 1970 C, N, and P pools (0–10 cm depth) are based on the assumption that nutrient pools in the 10–30 cm depth in 1984 (divided by 2) represent surface soil C, N, and P pools in 1970. We also assume that C, N, and P pools (10–30 cm depth) did not change during the first 14 years, between 1970 and 1984. See Methods for more details. Pools (g/m2) 0–10 cm Organic C Nitrogen Phosphorus 1970 (590) (25) (12) 1984 1296 ± 89 65 ± 7 15 ± 2 1995 2246 ± 640 (66) 146 ± 62 (3.7) 25 ± 6 (0.3) 1998 2029 ± 480 137 ± 31 36 ± 8 2017 2145 ± 65 118 ± 3 20 ± 2 10–30 cm 1970 (1180) (50) (25) 1894 1180 ± 300 50 ± 6 25 ± 2 1995 1102 ± 220 64 ± 22 21 ± 1 1998 1188 ± 270 80 ± 22 35 ± 4 2017 2492 ± 100 (62) 146 ± 7 (4.6) 24 ± 1 (0) Accumulation (g/m2/yr) This study (0−30 cm) 62−66 3.7−4.6 0−0.3 Seven constructed marshes 18−99 1.3−12.5 0−5 Mean and SE 41 ± 11 4.5 ± 1 0.9 ± 0.7 Natural marsh (1998) 77 4.0 0.5 Six natural marshes 2−105 0.1−15 0−0.5 Mean and SE 38 ± 15 4.9 ± 2.2 0.5 ± 0.3

38 g P/m2 at time zero to 44 g P/m2 after 46 years, nor did by 1.4 ha. Overall, during the past 25 years (1993–2018), they exhibit depth-specific successional trajectories. the island shrank from 8.9 ha to 4.7 ha with nearly all of Soils of the streamside zone developed faster than soils the loss on the northern side (previously discussed) and of the interior. In 2017, streamside zone soils had relatively on the western side (loss of 1.5 ha) where the commercial uniform bulk density (0.30–0.36 g/cm3), organic C (5.2– navigation channel is located. The 0.8 ha restored marsh, 6.5%, 2110–3170 g/m2), and N (0.33–0.35%, 124–202 g/ located on the eastern side of the island where nearly all of m2) in surface and subsurface soils. Whereas surface soils the existing marsh is located, has not succumbed to erosion of the interior had similar bulk density (0.24 g/cm3), C yet, but its northern flank, under current conditions, likely (8.8%) and N (0.50%, 112 g/m2) to surface and subsurface will disappear in the coming 10 to 20 years. Additionally, soils of the streamside zone, subsurface soils of the interior salt marsh vegetation has colonized the southern end of were considerably denser (0.66 g/cm3) and contained less the island where is accreting and new land is formed. C (1.5%, 1810 g/m2) and N (0.7%, 91 g/m2). A created tidal marsh also exhibited more rapid soil Discussion development (decreasing bulk density, increasing C and N) in the streamside than the interior zone that was linked to As marsh succession progresses, the burial of net primary increased hydroperiod (depth, duration, and frequency of production and nutrients in plant tissues contributes to inundation) (Craft et al. 2002). the development of soil organic matter as a reservoir for Over time, the shape and size of the dredge spoil island N and P that is gradually made available to through on which the marsh was established has changed dra- microbial mineralization. Bulk density decreases as the matically due to a combination of sea level rise (see dis- highly porous organic matter is incorporated into the soil. cussion), shipping activities, and island migration. Aerial In our study, bulk density decreased at a rate of 0.16–0.19 g/ photographs (1984–2018) revealed dramatic erosion of cm3/yr (Figure 1a) with little difference in surface versus the north end of the island. Between 1984 and 2018, the subsurface layers. north end eroded approximately 125 m at an annual rate As seen in Figure 1b–c, percent soil organic C and N of 3.7 m/yr. A total of 2.7 ha was lost from the north end follow analogous trajectories. While it is well known that during the 34-year period. During the same time, the south organic C accumulation is linked to buildup of soil organic end of the island extended at a rate of 3.4 m/yr and grew matter, approximately 95% of N in wetland soils exists as

September 2019 ECOLOGICAL RESTORATION 37:3 • 167 organic N (Craft et al. 1991). Studies in estuarine marshes upstream of the river mouth). In 1910, the tidal range at suggest that while P accumulation in organic matter is a Wilmington was 78 cm (Hackney and Yelverton 1990). significant component in highly organic soils, inorganic From 1936 to 1984, tidal range increased 26 cm, from 104 processes such as sorption and burial as sediment are to 130 cm (Hackney and Yelverton 1990). In 1999, a large generally more important to P accumulation in mineral dredging project was initiated to deepen the channel an soils (Craft 1997). additional 1.23 m. The net effect was an increase in tidal Development of soil properties in constructed marshes flux that continues to bring more into the river. occurs more rapidly in surface soils compared to sub- Over time, these activities led to increased dominance of surface soils. Increases in percent soil organic C, N, and saline tidal marsh and widespread loss of freshwater tidal P occurred in surface soils at a rate approximately three marsh and forest upstream (Hackney and Avery 2015). times greater than was observed in subsurface soils. Greater Today, marshes of the lower Cape Fear River are dominated inputs of roots and macroorganic matter to surface soils by S. alterniflora, the species planted at the restoration site. (Craft et al. 1988) likely explain the rapid increase in Historically, the marshes were mostly fresh. Freshwater organic matter and N in surface soils (McCaffrey and structures soil properties of tidal wetlands such that stocks Thomson 1980, Blum 1993) rather than allochthonous C of C and N are greater in fresh marshes relative to saline entering from tidal inundation. marshes (Craft et al. 2007). Though not significant, after 46 years, soil C:N decreased Deepening and widening of the Cape Fear River chan- in both depths from approximately 25–30 to 20, a value nel invariably led to increased commercial shipping. Since considered to be a threshold for N-limitation (Tisdale et al. 2002, tonnage at the Port of Wilmington increased 150%, 1985). In soils with C:N greater than 20 there is a scarcity from 2 million to over 5 million tons per year (Findley of N relative to C, resulting in immobilization of N by soil et al. 2017). microbes. At C:N less than 20, microbial demands for The combination of dredging activities, increased ship- N are satisfied and excess mineralized N becomes avail- ping, and rising seas will determine if and how much of able for plant uptake on a more consistent basis (Craft the restored marsh will persist in the future. The current et al. 1999). After nearly 50 years, soil C:N of our restored rate of sea level rise along the southeastern (Wilmington) marsh approximates C:N of Atlantic and Gulf natural North Carolina coast is 2.39 ± 0.35 mm/yr (NOAA 2018) S. alterniflora marshes that range from 14–21 (Craft 2001). and is expected to accelerate in the future. During the past Soil N:P increased from 4.5 to more than 13 in both depths 100 years, sea level in the river has risen approximately 30 (p < 0.10). Soils with N:P less than 30 are generally con- cm, further contributing to saltwater intrusion upriver sidered to be N-limited (Verhoeven et al. 1996), whereas (Hackney and Yelverton 1990). Tidal marshes adjust to a soil N:P greater than 30 signifies P-limitation. rising sea level by increasing plant productivity that traps Surface soil C and N pools initially increased faster than more sediment and adds more organic matter to the soil in subsurface soils but began to equilibrate after 25 years (Morris et al. 2002). At the current rate of sea level rise, the to values between 2100–2200 g C/m2 and 120–140 g N/ restored marsh, since its inception, would have accreted m2). However, surface soil C and N pools were 73% and approximately 10 cm of new soil in order to maintain its 78%, respectively, of values measured in a nearby reference elevation relative to the tidal frame. Consequently, today’s salt marsh (2945 g C/m2, 152 g N/m2) (Craft et al. 1999). surface (0–10 cm) soil layer represents mineral sediment Subsurface pools of C and N increased at a slower rate, and and organic matter accumulation that occurred during the after 46 years were 25% and 37%, respectively, (2492 g C/ past 50 years while the subsurface (10–30 cm) soil probably m2; 146 g N/m2) of values measured in the reference marsh represents organic matter (and N) that accumulated in (10,152 g C/m2; 394 g N/m2) (Craft et al. 1999). surface soil initially and was buried over time. The change The large stocks of soil C and N in the reference marsh in C and N in subsurface soils that occurred during the reflect the relic freshwater conditions in the . His- past 25 years may reflect sampling of former surface layers torically, the river was shallow and fresh nearly to its mouth that were buried. (Hackney and Avery 2015). As described by Hackney and Based on the pools and current rate of accumulation Avery, a series of modifications to the river began around of C and N in subsurface soils (Table 1) and pools in 1870 and continue to this day. They include closing one of the nearby reference marsh (10,157 g C/m2, 394 g N/ the two natural inlets, construction of three dams upstream, m2), it would take approximately 124 and 54 years, development of the Intracoastal Waterway, including respectively, for the constructed marsh soil to equal Snow’s Cut in 1929 (North Carolina Division of Parks the C and N pools in the reference marsh. The shorter and Recreation, n.d.), that served as a conduit for saline time frame needed to restore N pools relative to C sup- water from Masonboro Sound to the river, and a number of ports Eugene Odum’s theory of ecosystem development, deepening and widening projects, especially from the 1940s which suggests that limiting nutrients accumulate more to the present. Dredging activities increased not only saline rapidly due to higher rates of biotic retention and uptake water but it increased tidal range at Wilmington (44 km (Odum 1969).

168 • September 2019 ECOLOGICAL RESTORATION 37:3 We compared C, N, and P accumulation in the nearby Blake G.R. and K.H. Hartge. 1986. Bulk density. Pages 363–375 in natural marsh in 1998 by 210Pb dating (Craft et al. 2003). A. Klute (ed), Methods of Soil Analysis, Part I. Physical and Min- Organic C, N, and P accumulation was 77, 4, and 0.5 g C, eralogical Methods, Agronomy Monograph no. 9, 2nd edition. Madison, WI: American Society of Agronomy. N, P/m2/yr, respectively, and were similar to values mea- Blum, L.K. 1993. Spartina alterniflora root dynamics in a Virginia sured in our constructed marsh (Table 1). In addition to marsh. Marine Progress Series 102:169–178. the nearby reference marsh, we also compared soil C, N, Boesch, D. and R. Turner. 1984. Dependence of fishery species on and P accumulation in our study (Table 1) to accumula- salt marshes: The role of food and refuge. 4:460–468. tion measured in seven additional constructed estuarine Chmura, G.L and G.A. Hung. 2004. Controls on salt marsh accretion: marshes in North Carolina (Craft et al. 2003). That study A test in salt marshes of Eastern Canada. Estuaries 27:70–81. consisted of a chronosequence of restored Spartina marshes Cornell, J.A., C.B. Craft and J.P. Megonigal. 2007. Ecosystem gas exchange across a created salt marsh chronosequence. Wet- ranging from 1–28 years of age, with our study marsh lands 27:240–250. (Snow’s Cut) being the oldest at the time. Results from this Craft, C.B. 1997. Dynamics of nitrogen and phosphorus retention study showed average rates of accumulation of 41 g/m2/ during wetland ecosystem succession. Wetlands Ecology and yr for organic C, with rates ranging from 18–99 g/m2/yr Management 4:177–187. (Table 1). Nitrogen accumulation ranged from 1.3–12.5 g Craft, C.B. 2001. Soil organic carbon, nitrogen, and phosphorus as /m2/yr with a mean of 4.5 g/m2/yr. P accumulation ranged indicators of recovery in restored Spartina marshes. Ecological from 0–5 g/m2/yr with a mean of 0.9 g/m2/yr. As part of Restoration 19:87–91. the Craft et al. (2003) study, C, N, and P accumulation also Craft, C.B. 2007. Freshwater input structures soil properties, vertical accretion, and nutrient accumulation of Georgia and U.S. tidal was measured in six natural reference marshes. Similar to marshes. and Oceanography 52:1220–1230. our constructed marsh and the seven constructed marshes, Craft, C.B. and J.A. Sacco. 2003. Long-term succession of ben - the natural marshes exhibited comparable rates of C, N, thic infauna communities on constructed Spartina alterniflora and P accumulation (Table 1). marshes. Marine Ecology Progress Series 257:45–58. Our current study shows that surface soils (0–10 cm) Craft, C.B., S. Broome and C. Campbell. 2002. Fifteen years of veg- developed more quickly than subsurface (10–30 cm) soils etation and soil development after brackish-water marsh cre- and showed rapid sequestration of C and N in the first 27 ation. 10:248–258. Craft, C.B., S. Broome and E.D. Seneca. 1988. Nitrogen, phospho- years before leveling off; whereas the subsurface soils had rus and organic carbon pools in natural and transplanted marsh slower rates of accumulation that continued to increase soils. Estuaries 11:272–280. over time. However, subsurface C and N pools were sig- Craft, C.B., S. Broome and E.D. Seneca. 1991. Porewater chemis- nificantly lower compared to a nearby natural marsh, try of natural and created marsh soils. Journal of Experimental suggesting they are still not fully restored even after five and Ecology 152:187–200. decades. The restored marsh continues to sequester C and Craft, C., J. Clough, J. Ehman, S. Joye, R. Park, S. Pennings, et al. N at the current rate of sea level rise though human activi- 2009. Forecasting the effects of accelerated sea-level rise on tidal marsh ecosystem services. Frontiers in Ecology and the Environ- ties, such as dredging and commercial boat traffic, may ment 7:73–78. imperil its long-term persistence. In the larger context, our Craft, C.B., P. Megonigal, S. Broome, J. Stevenson, R. Freese, J. Cor- findings can inform expectations for relative rates of soil nell, et al. 2003. The pace of ecosystem development of con- development in other wetland restoration projects where structed Spartina alterniflora marshes. Ecological Applications nitrogen (and carbon) limit ecosystem development as well 13:1417–1432. as changes in nutrient (C:N, N:P) ratios that determine Craft, C.B., J.M. Reader, J.N. Sacco and S.W. Broome. 1999. Twenty- whether the wetland has achieved equivalence with respect five years of ecosystem development of constructed Spar- to nutrient and biogeochemical cycles. tina alterniflora (Loisel) marshes. Ecological Applications 9:​ 1405–1419. Crocker, R. and J. Major. 1955. Soil development in relation to veg- Ackowledgements etation and surface age at Glacier Bay, Alaska. Journal of Ecol- We thank the Craft boys—Hugh, David, and Patrick—for their ogy 43:427–448. help in collecting samples and transportation to and from the Findley, D. J., J.D. Small, W. Tran, A. Heller, S.A. Bert, S.E. Searcy, field site. We appreciate the help of Elena Solohin who conducted et al. 2014. Economic contribution of the North Carolina Ports. the GIS analysis. The work would not be possible without the Prepared for North Carolina State Ports Authority. ncports.com/ forethought of Steve Broome, Carlton Campbell, Larry Hobbs, wp-content/uploads/2016/07/economic-contribution-north- E.D. Seneca, and W.W. Woodhouse Jr. who undertook the res- carolina-ports.pdf. toration in 1970. Gorelick, N., M. Hancer, M. Dixon, S. Ilyushchenko, D. Thau and R. Moore. 2017. Google earth engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment 202:18–27. References Hackney, C.T. and G.B. Avery. 2015. Tidal wetland community Ballantine, K. and R. Schneider. 2009. Fifty-five years of soil devel- response to varying levels of flooding by saline water.Wetlands opment in restored freshwater depressional wetlands. Ecologi- 35:227–236. cal Applications 19:1467–1480. Hackney, C.T. and G.F. Yelverton. 1990. Effects of human activities and sea level rise on wetland ecosystems in the Cape Fear River

September 2019 ECOLOGICAL RESTORATION 37:3 • 169 estuary, North Carolina, USA. Pages 55–61 in D.F. Whigham from conversion and degradation of vegetated coastal ecosys- (ed), Wetland Ecology and Management: Case Studies. The Neth- tems. Plos ONE 7:e43542. erlands: Kluwer Academic Publishers. Shepard, C.C., C.M. Crain and M.W. Beck. 2011. The Protective Role Hatton, R.S., R.D. DeLaune and W.H. Patrick. 1983. Sedimentation, of Coastal Marshes: A Systematic Review and Meta-analysis.​ accretion, and subsidence in marshes of Barataria Basin, Loui- Plos ONE 6:1–11. siana. Limnology and Oceanography 28:494–502. Sommers, L.E. and D.W. Nelson 1972. Determination of total phos- He, Y., S. Widney, M. Ruan, E. Herbert, X. Li and C. Craft. 2016. phorus in soil: A rapid perchloric acid digestion procedure. Soil Accumulation of soil carbon drives denitrification potential Science Society of America 36:902–904. and lab-incubated greenhouse production along a chronose- Tisdale, S.L., W.L. Nelson, and J.D. Beaton. 1985. and quence of salt marsh development. Estuarine, Coastal and Shelf . , NY: Macmillan. Science 172:72–80. Turner, R.E., E.M. Swenson and C.S. Milan. 2002. Organic and LaSalle, M. W., M.C. Landin and J.G. Sims. 1991. Evaluation of the inorganic contributions to vertical accretion in salt marsh sed- flora and fauna of a Spartina alternifloramarsh established on iments. Pages 583–595 in M.P. Weinstein and D.A. Kreeger (eds), dredged material in Winyah Bay, South Carolina. Wetlands 11:​ Concepts and Controversies in Tidal Marsh Ecology. Dordrecht, 191–208. Nether­lands: Springer. Marton, J.M., M.S. Fennessy and C.B. Craft. 2014. USDA con- Verhoeven, J.T., A.W. Koerselman and F.M. Meuleman. 1996. Nitro- servation practices increase carbon storage and water quality gen or phosphorus limited growth in herbaceous wet vegetation: improvement functions: An example from Ohio. Restoration relations with atmospheric inputs and management regimes. Ecology 22:117–124. Trends in Ecology and Evolution 11:494–497. McCaffrey, R.J. and J. Thomson. 1980. A record of the accumulation Vincent, R.E., D.M. Burdick and M. Dionne. 2013. Ditching and of sediment trace metals in a Connecticut salt marsh. Advances ditch-plugging in New England salt marshes: Effects on hydrol- in Geophysics 22:165–236. ogy, elevation, and soil characteristics. Estuaries and Mcleod, E., G.L. Chmura, S. Bouillon, R. Salm, M. Bjork, C.M. 36:610–625. Duarte, et al. 2011. A blueprint for : toward an Woodhouse, Jr, W.W. and E.D. Seneca. 1974. Propagation of Spartina improved understanding of the role of vegetated coastal habi- alterniflora for Substrate Stabilization and Salt Marsh Develop- tats in sequestering CO2. Frontiers in Ecology and the Environ- ment. Coastal Engineering Research Center. Prepared for U.S. ment 10:552–560. Army Corps of Engineers. Morris, J.T., P.V. Sundareshwar, C.T. Nietch, B. Kjerfve and D.R. Yang, R.M., and W.W. Guo. 2018. Exotic Spartina alterniflora Cahoon. 2002. Response of coastal wetlands to rising sea level. enhances the soil functions of a coastal ecosystem. Ecology 83:2869–2877. Society of America Journal 82:901–909. Neubauer, S.C. 2013. Ecosystem responses of a tidal Zheng, L., R.J. Stevenson, and C.B. Craft. 2004. Changes in ben- experiencing saltwater intrusion and altered hydrology. Estuar- thic algal attributes during salt marsh restoration. Wetlands 24:​ ies and Coasts 36:491–507. 309–323. National Oceanic and Atmospheric Administration (NOAA). 2018. Sea level trends: Monthly data for station 8658120 (Wilmington, North Carolina). tidesandcurrents.noaa.gov/sltrends/sltrends_ station.shtml?id=8658120. North Carolina Division of Parks and Recreation. Fort Fisher State Aaron Noll, School of Public and Environmental Affairs, Recreation Area: History. n.d. www.ncparks.gov/fort-fisher- Indiana University; Bloomington, Indiana, USA. state-recreation-area/history. Nyman, J.A., R.J. Walters, R.D. DeLaune and W.H. Patrick. 2006. Courtney Mobilian, School of Public and Environmental Marsh vertical accretion via vegetative growth. Estuarine, Coastal Affairs, Indiana University; Bloomington, Indiana, USA. and Shelf Science 69:370–380. Odum, E.P. 1969. The strategy of ecosystem development. Science Christopher Craft (corresponding author), School of Public 164:262–270. and Environmental Affairs, Indiana University; Bloomington, Pendleton, L., D.C. Donato, B.C. Murray, S. Crooks, W.A. Jenkins, Indiana, USA, [email protected] S. Sifleet, et al. 2012. Estimating global “blue carbon” emission

170 • September 2019 ECOLOGICAL RESTORATION 37:3