2.1 Site description

The Altamaha River is approximately 630 km long and drains the third largest watershed along the Atlantic coast of the U.S. It is the largest undammed river along its mainstem of the Atlantic coast river. The watershed is mostly forested with moderate amounts of agriculture and little urban land (Table 1). Along the length of the watershed, we sampled riparian wetlands, upper and lower alluvial floodplain forests, tidal freshwater forests and marshes, brackish marshes, and salt marshes (Figure 1). One site was sampled for riparian and alluvial floodplain wetlands. For tidal forests and marshes, two sites were sampled. Riparian areas were sampled in the Oconee National Forest south of Athens, Georgia. Upper and lower alluvial floodplain forests were sampled in Riverbend Wildlife Management Area (WMA) southeast of Dublin, Georgia, and Big Hammock WMA, southwest of Glennville, Georgia. Tidal freshwater forests were sampled at Lewis Island, tidal freshwater marshes at Broughton Island, and salt marshes at Rockdedundy Island, all located near Darien Georgia. Latitude, longitude, wetland elevation, and dominant vegetation of the sites are presented in Table 2.

2.2 Soil sampling and lab analysis

For the riparian and alluvial floodplain wetlands, two soil cores, one from the natural levee and one from the interior floodplain, were collected from each site. Riparian and alluvial floodplain wetlands were sampled in 2011, tidal forests in 2008, and tidal marshes in 2005. Core data from tidal forests were published in Craft (2012). Core data from tidal marshes were published in Loomis and Craft (2010). Cores, 8.5 cm diameter by 30 cm deep, were sectioned into 2 cm increments in the field, then transported to the lab where they were air dried and weighed for bulk density. Once dried, increments were ground, passed through a 2 mm mesh sieve, and analyzed for ¹³⁷Cs, ²¹⁰Pb, total C, N, and P.

The bulk density of each depth increment was calculated from the dry weight per unit volume after correcting for the moisture content of an oven-dried (70°C) soil sample (Blake & Hartje, 1986). Before C analysis, subsamples were tested for the presence of carbonates by adding one drop of 0.1 mol L⁻¹ HCl and observing whether effervescence occurred. Samples containing carbonates were treated with 0.1 mol L⁻¹ HCl before C and N analysis. Carbon and N were analyzed using a Perkin-Elmer 2400 CHN analyzer (Perkin-Elmer). Total P was determined colorimetrically after digestion in nitric-perchloric acid (Sommers & Nelson, 1972). Analysis of an-house standard (%C = 6.1, %N = 0.365, n = 20) yielded the average recovery of 103% and 102% for C and N, respectively. For P, analysis of NIST Standard #1646a (estuarine sediment) recovered 88% (n = 18). All analyses were expressed on a dry weight basis by correcting for the moisture content of the soil, determined by weighing 1 g of subsoil before and after drying at 70°C for 24 h.

For radiometric analyses, ground soils were packed into 50 mm diameter by 9 mm Petri dishes and analyzed by gamma spectrometry for ¹³⁷Cs using the 661.6 keV photopeak. Cesium-137 (half-life 30 years) is an impulse marker produced as fallout by aboveground nuclear bomb blast tests (Ritchie & McHenry, 1990). We used the increment with maximum ¹³⁷Cs activity to represent the year 1964, the year of greatest atmospheric deposition.

Rates of C, N, and P accumulation were calculated using ¹³⁷Cs-derived accretion rates, bulk density, and C, N, and P concentration in depth increments above and including the ¹³⁷Cs maximum.

2.3 Statistical analysis

The small sample size precluded statistical analysis of differences in soil characteristics, accretion, and accumulation among riparian, upper and lower alluvial, tidal freshwater forest and marsh, brackish marsh, and salt marsh so wetland means ± SE are presented. Correlation analysis was used to explore associations between soil characteristics, accretion, and accumulation (SAS Institute, 2002) with tests of significance conducted at p ≤ 0.05.

3.1 Soil bulk density and organic C, N, and P concentration

Soil bulk density varied along the waterway, ranging from 0.23 g/cm³ in the tidal freshwater marsh to 0.98 g/cm³ in the upper alluvial floodplain (Table 3). Riparian and floodplain soils were characterized by high bulk density (0.81–0.95 g/cm³) whereas the bulk density of tidal forests and marshes were much lower (0.23–0.40 g/cm³). Conversely, soils of the tidal forest, fresh and brackish marshes contained higher percent organic C (11%–16%) and total N (0.70%–0.95% N) than other wetlands along the continuum (2.1%–4.6% C, 0.17%–0.35% N). Phosphorus concentrations were not as variable as C and N, ranging from 430 µg/g in the upper alluvial floodplain to 710 µg/g in the riparian wetland. The bulk density of salt marshes was intermediate (0.40 g/cm³) to other wetlands.

Carbon:nitrogen (C:N) ranged from 15 to 20 with no clear trend along the waterway (Table 3). Lowest C:N was at the ends of the waterway, in riparian and upper alluvial forest, and in the salt marsh. Nitrogen:phosphorus (N:P), in contrast, exhibited a clear trend along the continuum, increasing from the headwaters to the middle reach of the estuary in brackish and tidal freshwater marsh soils (Table 3). Riparian wetland soils had the lowest N:P (6) that increased downstream to tidal fresh forests (34). N:P of brackish and tidal freshwater marshes also were high (27) but salt marshes had N:P (14) similar to other mineral soil wetlands along the waterway.

Soil C, N, and N:P were strongly negatively correlated with bulk density (Table 4). N:P was strongly and positively correlated with C and N but not with P. C:N was correlated with C but not with N. Accumulation of C and N was positively correlated with accretion rate (r = 0.86, p < 0.0001). C and N accumulation were strongly correlated (Table 4) as were P accumulation and sediment deposition (Table 4).

3.2 Soil accretion and N, P, and organic C accumulation

Soil cores exhibited well-defined ¹³⁷Cs profiles, (Figure 2) with ¹³⁷Cs maxima occurring at depths of 5–15 cm. Accretion rates based on these profiles ranged from a low of 0.6 mm/year in the tidal forest to 6.1 mm/year in the brackish marsh. Accretion exhibited a bi-modal distribution along the continuum with higher rates of accretion in riparian areas (3.9 ± 1.5 mm/year) near the headwaters and in tidal fresh (4.7 ± 0.7 mm/year) and brackish marshes (5.4 ± 1.3 mm/year) near the sea (Figure 3). Accretion rates were lowest in the tidal forest at the head of the estuary (0.9 ± 0.3 mm/year) and in the salt marsh (1.2 ± 0.5 mm/year) at its mouth.

Accumulation of mineral sediment and P exhibited a weaker yet still bimodal distribution with riparian wetlands having much higher rates, 3310 g sediment/m²/year and 2.75 g P/m²/year, as compared to alluvial and tidal wetlands downstream (Figure 4). A secondary peak in sediment and P accumulation was observed in the brackish marsh where 1150 g sediment/m²/year and 0.9 g P/m²/year were deposited. The tidal forest and salt marsh with their low accretion rates also had the lowest sediment (180–380 g/m²/year and P accumulation 0.23–0.37 g/m²/year). Organic C and N burial mirrored accretion with the greatest accumulation in brackish (150 g C/m²/year, 9.5 g N/m²/year) and tidal fresh marshes (102 g C/m²/year, 7.1 g N/m²/year) with a secondary peak in riparian wetlands (86 g C/m²/year, 6.6 g N/m²/year) (Figure 5).

High rates of sediment and P in riparian wetlands reflect greater proximity to inputs of these materials from the terrestrial landscape where most agricultural activities are concentrated (Cooper & Gilliam, 1987; Cooper, Gilliam, Daniels, et al., 1987). The secondary peak in sediment and P accumulation observed in tidal fresh and brackish marshes is attributed to sediment (and P) flocculation in the mixing zone between fresh and saline waters (Loomis & Craft, 2010). For N, the highest accumulation in tidal fresh and brackish marsh is attributed to the high rates of C sequestration in these wetlands where plant productivity and uptake of N as NH₄ and NO₃ is greatest (Więski et al., 2010).