In 2016, the US Army Corps of Engineers and project partners implemented a TLP restoration project at a rapidly degrading coastal marsh near Avalon, NJ ( Figure 1 ). Vegetated areas of the marsh were dominated by Spartina alterniflora occupying >90% absolute areal cover of the soil surface, with small amounts of Distichlis spicata and other trace species (e.g., Salicornia Sp.) occupying <5% of the soil surface marsh. The marsh soils consisted of Typic Sulfihemists mapped as the very frequently flooded Appoquinimink-Transquaking-Mispillion complex on 0 to 1 percent slopes (Soil Survey Staff, 2025).

Degradation of the back bay marsh resulted in the interspersion of vegetated areas with expanding unvegetated deepwater pools and shallow water pannes, interior marsh erosion and collapse, and other rapid marsh platform destabilization. This impaired marsh has experienced increasing inundation during which has resulted in dieback of vegetation and weaking of the soil structure via root system decay (Schepers et al., 2017; DeLaune et al., 1994). Following open water area expansion, vegetation can no longer re-establish for a number of reasons. First, the bulk density is reduced past the threshold required for vegetation colonization as the elevation decreases (DeLaune et al., 2016; Day et al., 2011). Second, water depth increased until it prevents re-establishment of vegetation. Finally, wind waves led to expansion of open water by inducing edge erosion (Sapkota and White, 2021). Thus, efforts were initiated to supplement the marsh elevation with dredged sediment to increase bulk density and help stabilize the marsh platform.

TLP was performed via high-pressure spraying of a sediment slurry (~20% sediment, 80% water) in a practice known as “rainbowing,” in which the dredge material is dispersed from a nozzle in an arc-shape, and by moving the dredged sediment outlet pipe into target locations surrounded by coconut fiber roll containment features ( Figure 2 ). There are two assessments of the Avalon, NJ TLP project published in the literature: one assessing biogeochemical properties prior to the sediment placement (Berkowitz et al., 2018) and another assessing biogeochemical properties 0.5 years after the placement (VanZomeren et al., 2018). This study extends the monitoring timeframe, now 6+ years after the TLP project initiation, to investigate the longer-term effects of TLP on wetland soil physical and microbial characteristics. In doing so, the authors of the current study compare data from all three discrete sampling intervals, evaluating the trajectory of changes in soil biogeochemical properties. The same TLP sampling stations were revisited from the two previous studies to ensure that locations were consistent across sampling temporal intervals. Sampling stations included vegetated areas that received TLP and vegetated areas that did not receive TLP to serve as unamended sediment controls ( Figure 1 ).

In August 2023, soil samples (0–5 cm) were collected at two vegetated TLP sites and 1 control site. Three stations were sampled at each site in duplicate using 7 cm diameter acrylic coring tubes for a total of 18 characterization samples (3 sites x 3 stations x duplicates). Samples were placed into labeled Ziploc bags and shipped overnight in a cooler on ice to the laboratory. In addition to the soil characterization samples, quadruplicate intact 20 cm long soil cores were taken from two vegetated TLP sites and one vegetated control site for a total of 12 cores for a N cycling incubation study. Cores were sealed with rubber stoppers and immediately shipped overnight on ice to the laboratory for an assessment of N mineralization and denitrification. For the original TLP, project target elevations ranged from 0.73 to 0.91m NAVD88 (VanZomeren et al., 2018).

Soil characterization followed the same methods utilized by Berkowitz et al. (2018) and VanZomeren et al. (2018) to minimize variance between datasets. At the lab, duplicate soil samples were homogenized, and live root matter with diameter >1 mm was removed. Roots were dried and weighed to determine live root mass on a dry weight basis. The 9 soil samples were analyzed for moisture content, soil dry weight bulk density, weight % organic matter, total C, N and P, extractable nutrients, extractable dissolved organic C and N, microbial biomass N (MBN), and potentially mineralizable N (PMN) rate. Briefly, moisture content was determined from weight difference of wet soil subsamples before and after placing them in a drying oven at 60°C until constant weight (White and Reddy, 2000). Soil bulk density was determined by dividing the total dry weight of the sample by the volume of the core. Soil weight % organic matter (SOM) was determined by loss on ignition (LOI), which consists of ashing dried ground subsamples at 550°C and recording the loss of mass (Andersen, 1976). Total P was determined via the TP ashing-acid digestion method (Andersen, 1976). Total C and total N were determined from dry ground subsamples analyzed on a Costech Analytical Technologies Elemental Combustion System (Berkowitz et al., 2018). Soil extractable nutrients were determined using a 2 M KCl extraction followed by filtration through a 0.45 μm membrane filter and acidifying samples at a pH of <2 (VanZomeren et al., 2018). Samples were analyzed on a SEAL Analytical AQ300 discrete Analyzer (Mequon, WI) using USEPA methods (U.S. EPA, 1993). Dissolved organic C (DOC) and N (DON) were determined from a 2 M K₂SO₄ extraction (White and Reddy, 2000). Microbial biomass N (MBN) was determined via chloroform fumigation-extraction method, in which non-fumigate sample N value is subtracted from fumigate samples N value to determine the amount of N contained within lysed cells (Brookes et al., 1985) with modification by White and Reddy (2000). Extracted liquid samples for DOC, DON and MBN were analyzed on a Shimadzu Scientific Instrument TOC-VCSN (Columbia, MD). Potentially mineralizable nitrogen (PMN) rate was determined from the change of NH₄-N production from soil under anaerobic conditions at 40°C over a period of 10 d (White and Reddy, 2000). Sample were extracted @ day 0, 2, 5, and 10 of the incubation. The 2 M KCL extracted samples were analyzed on a SEAL Analytical AQ300 (U.S. EPA, 1993) and values were plotted over time to calculate PMN rate from linear regression (White and Reddy, 2000).

Twelve intact cores, four replicates each from two different TLP areas, and four replicates from an adjacent non-TLP control area, underwent incubation to determine the ammonium flux rates and nitrate removal capacity of the wetland soil. The NH₄ production rate is linked to supporting plant growth and NO₃ reduction is linked to water quality improvement. Cores were flooded to create a 20 cm water column with salt water adjusted to a salinity of 30 to mimic site conditions using Instant Ocean^®. The cores were placed into a temperature-controlled water bath at 20°C and each core was aerated with room air to ensure aerobic conditions within the water column. The cores were spiked to produce a water column concentration of 1 mg L^-1 NO₃-N. Seven mL water samples were collected each day for a period of 8 days, filtered through a 0.45-μm membrane syringe filter and acidified with H₂SO₄ to a pH <2 (Cheng and White, 2022). Water samples were refrigerated at 4°C until analysis (within 2 weeks of collection) on a SEAL Analytical AQ300 colorimetric analyzer for NO₃-N concentration using method 353.2 (U.S. EPA, 1993). Upon completion of aerobic incubation, the water column was removed and replaced, and the experiment was repeated under anaerobic water column conditions. Cores were sealed with rubber stoppers which contained inlet and outlet glass tubes and constantly bubbled with 99.99% N₂ gas to ensure anaerobic conditions (Bowes et al., 2022). Water samples were collected 2 times each day for a period of 3 days. Both NO₃-N and NH₄-N were measured using a SEAL Analytical AQ300 colorimetric analyzer, methods 353.2 and 350.1, respectively (U.S. EPA, 1993). Values for NO₃-N and NH₄-N concentrations were plotted over time on an areal basis to determine anaerobic and aerobic nitrate reduction rates and anaerobic ammonification rates, respectively.

Statistics were run using R in RStudio (Version 4.3.1; Vienna, Austria). One-way ANOVA testing was used, for which a p-value of <0.05 was considered significant for differences between control and TLP values as well as between time points. Shapiro-Wilk test for normality and Levene’s test for equal variance were used to ensure that ANOVA assumptions were met. A Tukey-HSD post-hoc test was used if assumptions were met, and a Games-Howell post-hoc test was used if Tukey-HSD was violated.

The TLP-treated marsh sites had significantly higher mean bulk density than the control sites at 0.66 g cm^-3 and 0.24 g cm^-3, respectively. The inverse was the case for gravimetric moisture content at 49.1% vs 75.8% for the TLP and control sites, respectively ( Table 1 ). These differences in bulk density and moisture content are expected, even after 6+ years, due to the addition of dense, mineral sediment and reduced soil pore space of the TLP treatment relative to the degraded, organic soil conditions of the control (VanZomeren et al., 2018). The SOM, total C, and total N were significantly lower (~30%) in the TLP soil than the control ( Table 1 ). Soil organic matter (LOI), total C, and total N were all strongly correlated (R>0.9). Total P was not significantly different between TLP and control. Microbial biomass N and PMN rates in the TLP treatment soils were ~1/3 of the control marsh soils. Extractable NH₄-N values were low in both the TLP and control plots, likely a result of ammonium uptake and subsequent porewater depletion during the latter part of growing season when sampling occurred (Feder and White, 2024). Live root mass was significantly higher (3x) in the TLP marsh compared to the control marsh ( Table 1 ). This difference is likely due to variations in soil properties and associate vegetation responses to nutrient availability, which are explored further in the discussion below.

There was no significant difference between the NO₃ ^- reduction rates of the TLP cores and the control cores under aerobic ( Figure 3A ) or anaerobic ( Figure 3B ) water column conditions. Nitrate reduction rates under aerobic water column conditions for the control averaged 45.3 ± 9.01 mg m^-2 d^-1 while mean rates for the TLP averaged 53.3 ± 4.49 mg m^-2 d^-1. Nitrate reduction rates under anaerobic water columns were statistically higher, averaging 255 ± 27.7 mg m^-2 d^-1 for the control and 205 ± 15.8 mg m^-2 d^-1 for the TLP. Notably, the N-reduction related water quality improvement potential of these TLP marsh soils was not significantly different than the control marsh, despite the observed differences in soil properties.

The mean anaerobic NH₄ ⁺ flux rate was 44.4 ± 24.14 mg N kg⁻¹ d⁻¹ in both control and TLP marsh soils under anaerobic water column conditions, indicating similar N availability in both sites ( Figure 4 ). This ammonium flux from heterotrophic soil bacterial activity, reflects the soils’ capacity to supply N to the root zone for macrophyte uptake (White and Reddy, 2001).

There was a range of changes of soil parameters and biogeochemical measures across the time points found in this study when compared with the two previous studies covering 6+ years ( Table 2 ). These differences over time are explored in more detail in the following Discussion section.

This study integrated data from the two aforementioned previous studies: Berkowitz et al. (2018), collected in November 2016 (t₀) ( Supplementary Table S1 ), and VanZomeren et al. (2018) collected 0.5 years post-TLP in June 2017 (t₁; Supplementary Table S2 ) with our data 6+ years after project implementation (t₂) to analyze the trajectory of soil properties and wetland functions following project implementation. These data represent the longest investigation of soil N biogeochemical changes in TLP treated wetlands currently available in the literature that the authors are aware.

In November 2015 (t₀), before the initiation of the project, the bulk density of the vegetated TLP marsh soil ( Supplementary Table S1 ; Figure 5A ) was not significantly different from the control. However, by June 2016 (t₁), approximately 0.5 years (t₁) after TLP application, the bulk density in the TLP area was 342% of the control. By August 2023 (t₂), the bulk density remained relatively high, at 272% of the control, which suggests long-term sediment retention related to marsh stability, important as Sapkota and White (2019) found that coastal marsh with higher soil bulk density eroded more slowly. The observed decrease in bulk density over time could reflect root colonization and bioturbation of the TLP sediment, along with increases in soil organic matter. Moisture content increased from 52.4% of the control at t₁ to 64.8% at t₂ ( Figure 5B ), showing an inverse relationship with bulk density.

Although soil organic matter (SOM), total C, and total N as percentages of control increased on average from t₁ to t₂, the changes were not statistically significant ( Figures 6A-C ; Table 2 ). This lack of difference is also likely related to the fact that the density of organic matter remains very low compared to mineral sediment and therefore requires large changes in organic matter content to detect density differences. However, this trend is consistent with the expected gradual accumulation of organic matter in the TLP marsh. Total P increased from 84.5% of the control in t1 to 131% in t₂ ( Table 1 ; Figure 6D ). Again, the mineral fraction will dominate total P in the sediment amended marsh soils. One recent coastal marsh study separated total P into total organic P and total inorganic P to assess the mineral-bound P. When mineral river sediment was deposited in the marsh, the overall total P did not change but the relative distribution of majority total organic P shifted to total inorganic P (Spera et al., 2020).

Extractable ammonium of the TLP sites ( Figure 7A ) did not differ significantly from the controls at any time point. Ammonium levels for all sites were notably low in t₂, likely due to plant uptake. The microbial-mediated process of PMN rose from 22.9% of the control in t₁ to 30.9% in t₂ ( Figure 7B ), while MBN increased from 7.58% to 29.4% of the control over the same period ( Figure 7C ). These increases of PMN and MBN as a % of the control indicate increasing N cycling rates which are typically associated with higher organic matter soils (White and Reddy, 2000). While the soil trajectory of the sediment amended soils appears to be trending closer to the control, the N measures remain well below pre-project site conditions, and available data suggest that the restored wetland soil properties continue to develop along trajectories that differ from both the control and the original site conditions.

Low marsh elevation and bulk density are primary contributors to marsh degradation (Day et al., 2011). The TLP adds mineral material to the marsh soil, increasing elevation and bulk density. The restoration of these two physical factors allows for the re-establishment of vegetation and prevents further ponding and degradation. Additionally, by t₁, the re-vegetation of the TLP area was already taking place with plant pushing up through the dredge material and the restored marsh was visually indistinguishable from naturally vegetated areas in the region by t₂ (Harris et al., 2025).

Examining the trajectories of the percentage difference of biogeochemical properties reveals that most soil characteristics linked to microbial activity in the TLP soil are moving towards conditions observed at control sites. However, some properties, including bulk density and organic matter content, are likely to follow distinct patterns due to the addition of mineral sediments, which instantaneously impact soil structure and composition (Berkowitz et al., 2022) ( Figures 5A , 6A ). The organic matter and total C and N each display slight, but non-significant average increases over time. In properties related to N cycling, such as total N, PMN and MBN, there is a similar, non-significant increase over time, suggesting these characteristics may ultimately converge toward control marsh levels over long time scales ( Figures 6A–D ; Table 2 ). As plants re-establish and soil organic material accumulates, further increases in organic matter and N are anticipated over time (Craft et al., 2003). Longer-term data are needed to accurately define the progression of these changes and the slope of the restoration trajectory curve. Additional data is also required to inform the application of control (or ‘reference’) conditions as milestones for determining restoration project success, or if alternative process-based measures of ecological function and service delivery yield more desirable outcomes.

Raising marsh elevation reduces the duration of soil inundation, leading to increased oxidation at the soil surface. In contrast, prolonged flooding creates more anaerobic conditions in surface soil (Reddy and DeLaune, 2008). This shift toward oxidation with higher elevation may initially accelerate the decomposition of newly deposited detrital material, potentially lowering C storage from detritus. Increased decomposition from higher elevation will affect soil properties such as PMN, MBN, TN, and TC. However, shorter inundation periods also reduce flooding stress on plants. Snedden et al. (2015) found that increased elevation led to higher plant primary proactivity and biomass, with notable growth observed at elevations up to 19 cm above mean high water—the maximum elevation in their study. Other studies extending the soil surface elevation up to 30 and 40 cm showed that intermediate elevations best supported plant growth (Kirwan and Guntenspergen, 2012; Morris et al., 2013). Thus, balancing plant productivity with decomposition is crucial to optimizing ecological outcomes when implementing marsh restoration projects using TLP, where elevation gradient and target tolerances are often smaller than other coastal restoration/creation approaches.

As mentioned previously, the live root density in the TLP marsh soil is 3x higher than the vegetated control per unit volume of soil. One explanation for this difference is that the plants responded to a lower pool of N availability and increased root biomass to compensate. In the 2023 (t₂) TLP marsh dataset, the soil contained approximately 1/3 of the TN, 1/3 of the MBN, and 1/3 of the PMN rate of the control, which are inversely proportional to the 3x increased root matter found in the TLP soils. This result suggests that the plants in the TLP marsh increased root mass in order to compensate for lower nutrient availability (Deegan et al., 2012). Payne et al. (2021) also reported higher belowground biomass in TLP marshes, approximately 50% greater root density, 3 years after the TLP project. Ford et al. (1999) reported that root biomass had reached or exceeded that of the control just 1 year after a TLP in another TLP project. Another possible explanation for the increased root growth is the absence of prior root occupation in the soil. Roots and rhizomes naturally expand in the subsurface, invading any viable nearby soil (Cahoon et al., 2021; Ford et al., 1999). Consequently, the unused soil space following TLP allows for the rapid expansion of root mass into previously unoccupied areas, as there is minimal belowground vegetation. This enables plant roots to quickly establish themselves in the dredged sediment, effectively filling the unclaimed “territory”.

Whether root growth is driven by physical or nutrient dynamics or a combination of factors, the increase in plant roots and rhizomes has the potential to enhance marsh sediment retention and resilience (Stagg and Mendelssohn, 2011). Increased root biomass strengthens soil shear resistance and reduces erosion (Graham and Mendelssohn, 2014; Sasser et al., 2018). The TLP is likely to contribute to coastal plant community resilience by providing new soil for roots and rhizomes to colonize. Consequently, the combined effects of increased root biomass and the greater bulk density of TLP-amended soils suggest that TLP enhances the marsh’s capacity for storm protection and erosion resilience.

Thin-layer placement has the potential to be an effective method to build elevation capital and C storage strategy to enhance biogeochemical nutrient cycling. In addition to burying existing organic rich soil (Bartolucci and Fulweiler, 2024), TLP has been shown to stimulate increased growth of belowground biomass. Notably, primary production in marsh plants is typically higher in belowground vs aboveground biomass. Stagg et al. (2017) reported that belowground biomass in coastal marshes can be up to five times greater than aboveground biomass. Further research is needed on the longer-term biogeochemical cycling capacity and trajectory of sites restored using TLP and other techniques.