Patterns of nitrogen mineralization in wetlands of the New Jersey pinelands along a shallow water table gradient.
Abstract: Nitrogen dynamics in wetlands are often assumed to vary with water tables and the resulting patterns of soil moisture, and other soil properties. To test this hypothesis, we have examined N mineralization patterns in two replicate sequences of three forested wetland types located on a shallow water table gradient in the New Jersey Pinelands during a 12 mo sampling period. Each sequence of the three forested wetlands consists of pine wetlands and pine-hardwood swamps with sandy mineral soils and cedar swamps with peat soils along the shallow water table gradient. Although water tables differed between the two mineral-soil wetlands, there were no differences in soil properties between them, including patterns of extractable N or net N mineralization rate. However, peat soils from the cedar swamps had net N mineralization rate 5-10 times higher than the mineral soils from the other two types of wetlands over the sampling period. Although soil moisture was correlated with water table position within wetlands, net N mineralization rate did not vary with water tables, nor did it vary with soil moisture variations within sites. Overall, net N mineralization rate reflects soil type (histosols vs. mineral hydric soil) and organic matter quality (C:N) ratio.
Article Type: Report
Subject: Wetlands (Environmental aspects)
Nitrogen (Environmental aspects)
Fossilization (Research)
Authors: Ehrenfeld, Joan G.
Yu, Shen
Pub Date: 04/01/2012
Publication: Name: The American Midland Naturalist Publisher: University of Notre Dame, Department of Biological Sciences Audience: Academic Format: Magazine/Journal Subject: Biological sciences; Earth sciences Copyright: COPYRIGHT 2012 University of Notre Dame, Department of Biological Sciences ISSN: 0003-0031
Issue: Date: April, 2012 Source Volume: 167 Source Issue: 2
Topic: Event Code: 310 Science & research
Product: Product Code: 2813500 Nitrogen NAICS Code: 32512 Industrial Gas Manufacturing SIC Code: 2813 Industrial gases
Geographic: Geographic Scope: United States Geographic Code: 1USA United States
Accession Number: 287956853

The nitrogen cycle in wetlands is sensitive to many environmental factors. Soil properties have a dominant effect on the various processes that determine the retention or loss of nitrogen (N) from the wetland (Reddy and DeLaune, 2008). On the flat coastal plain of New Jersey, wetlands are located in close proximity to each other along a shallow water table gradient. These wetlands may vary in moisture as well as other soil properties, and thus may vary in nitrogen cycling processes. On the other hand, the degree of difference in nitrogen cycling processes may only reflect changes of soil properties, such as organic matter content and chemical composition, so that small differences in water tables or soil moisture do not affect nitrogen dynamics.

Many studies of N cycling in upland sites have shown that the amount of net N mineralization rate is heavily influenced by moisture, temperature, and the amount and quality of organic matter in the soil, particularly the amount of total N and the C:N ratio of the organic matter (Owen et al., 2003; Parfitt et al., 2005; Knoepp and Vose, 2007). These studies have often relied on gradients of sites arrayed along steep topographic gradients to distinguish different soil conditions. However, wetlands are often found in relatively flat landscapes, where differences in soil properties may be less clear, and where variability in soil properties within sites may be as large as variation among sites.

Despite a large amount of research on wetland nitrogen cycling (extensively reviewed in (Reddy and DeLaune, 2008), comparative studies of N mineralization among different wetland types in the same region are relatively uncommon. Most such studies compare different types of peatland, such as bog and fen peats supporting different types of plant communities and having different amounts of minerotrophic ground water inputs (Updegraff et al., 1995; Bridgham et al., 1998; Aerts et al., 1999; Chapin et al., 2003; Keller et al., 2004), particularly in boreal and arctic ecosystems (Jonasson and Shaver, 1999; Bayley et al., 2005; Fellman and D'Amore, 2007). Conversely, studies of riparian and mineral soil wetlands have compared mineral soils in different types of riparian wetlands or in different locations (Groffman et al., 1992; Verhoeven et al., 2001; Bechtold and Naiman, 2006; Wassen and Olde Venterink, 2006), and have emphasized the role of soil textural differences, geomorphological influences on water flow, and atmospheric deposition as controlling factors differentiating sites. But studies that compare forested wetlands on a drainage catena that include both organic and mineral hydric soils are rare. Fellman and D'Amore (2007) compared peatland soils (bog and forest) with riparian mineral soils in Alaska, and found that net N mineralization rates were faster in riparian soils than either peatland soil. Other studies that have compared organic and mineral wetland soils have concentrated on fluxes other than net N mineralization rates (e.g., Hanson et al., 1994) or have included human-related fluxes such as N deposition, removal of vegetation by mowing, or drainage (Holden et al., 2004; Wassen and Olde Venterink, 2006). In the New Jersey Pinelands, mineral soil and organic soil wetlands occur in close proximity to each other along shallow water table gradients (Ehrenfeld, 1986; Forman, 1998). They receive low levels of N deposition (Dighton et al., 2004; Ayars and Gao, 2007), and are otherwise undisturbed. Previous studies of nitrogen dynamics in these wetlands (Poovarodom et al., 1988; Sangemeswaran, 1995) have reported that nitrate and nitrification are extremely low under most circumstances but can be elevated under disturbed conditions, such as in wetlands affected by stormwater inputs in urban areas (Zhu and Ehrenfeld, 1999).

Few studies have examined whether variation in soil properties at a plot scale (scale of meters) results in variations in net N mineralization rates and, therefore, needs to be taken into account, in addition to the landscape-scale or between-plot variations (scale of 100 s of meters). However, Owen et al. (2003), for example, found high variability within plots along a mountain gradient in Taiwan in all soil properties measured and only few significant relationships between soil properties and N mineralization in plots across the mountain gradient. Wetlands usually array along a shallow water table gradient, but most of biogeochemical processes of nitrogen cycling in wetlands are redox-sensitive. Therefore, wetlands may presumably have high within-site variation of nitrogen cycling, such as nitrogen mineralization, driven by soil properties' variation; and high site variation of redox-sensitive nitrogen cycling may be expected due to groundwater table shifting or soil moisture change. To identify driving factors of nitrogen cycling is of a great importance for understanding nitrogen fates in a wetland ecosystem. Nitrogen cycling in a sequence of wetlands located along a shallow water table gradient in the New Jersey Pinelands was studied to determine how important differences in soil moisture are in determining N dynamics.

We focus on the following questions: (1) how do net mineralization and net nitrification rates vary among different wetland types found in close proximity to each other but at different positions along a shallow water table gradient? (2) how important are within-site variations in soil properties in determining these rates (referred to below as "plot-scale" variation)? (3) how important are between-site variations in soil properties in determining these rates (referred to below as "landscape-scale" variation)?



We sampled six wetlands in two locations within the McDonald's Branch watershed in Brendan T. Byrne State Forest (39[degrees]53'05"N, 74[degrees]30'20"E). The watershed is located in the New Jersey Pinelands, an extensive region of sandy soils and forests dominated by Pinus rigida (Forman, 1998). A set of three adjacent wetland plots was established at each of two sites designated here as M10 and M6, chosen as representative of Pinelands shallow gradient wetlands. The M10 series is at a location with narrow bands of wetlands along the stream, representing upper reaches of the wetland systems; the M6 series represents a lower location in the watershed with broad bands of wetlands bordering the stream. The two sites are approximately 1 km apart. At each site, one plot of each wetland types was located along the gradient, and no more than 50 m from each other. The types included (1) pine wetlands, which are pitch pine (Pinus rigida Mill.) dominated communities at the upland-wetland boundary (here termed as 'PW'), (2) pine-hardwood swamps, which contain mixtures of hardwoods (primarily red maple, Acer rubrum L., black gum (Nyssa sylvatica Marsh.), sweet bay magnolia (Magnolia virginiana L.), and pitch pine (here termed as 'PH'), and (3) cedar swamps, which are dominated by Atlantic white-cedar (Chamaecyparis thyoides (L.) Britton, Sterns & Poggenb.) (here termed as 'CS') in the lowest topographic positions bordering streams and which may contain some of the hardwood tree species mentioned above. Plant community composition has been described by Ehrenfeld and colleagues (1981, 1986, and 1991), Zampella et al. (1992), Forman (1998), Laidig and Zampella (1999), and others. The wetland communities are termed "types" below, and the six sampled wetlands are referred to as "plots."

The PW and PH types are found on hydric mineral soils of the Atsion and Berryland series. The Atsion series (sandy, siliceous, mesic Aeric Alaquods), supporting the PW communities, is poorly drained; the Berryland series (sandy, siliceous, mesic Typic Alaquods), found in the PH stands, is very similar but is classified as very poorly drained (Soil Survey Staff Natural Resources Conservation Service). The white cedar swamps, in contrast, are found on histosols (mapped as Manahawkin muck) that are 0.5-2 m deep (M6 site, 2 m depth, Ml0 site 0.6 m depth).


The study focused on the mineral soil material, since this is the location for most of the plant roots, and the organic horizon was very variable at each plot (ranging from 2 to 12 cm thickness at replicate location within plots). Net N mineralization and nitrification rates were measured on a monthly basis for a period of 12 mo (May 2005 to Apr. 2006) in each plot using a standard buried core method (Robertson et al., 1999). On each sampling date, three samples were taken from just outside the boundary of a 10 m x 10 m plot (used for vegetation measurements in a related study (Laidig et al., 2010) at each of the three plots at each site. Sampling was done on three of the plot boundaries, in order to ensure that any spatial variability in the conditions at each of the sites was fully represented in the study. Thus, there were three replicate samples from each plot per month.

We used a hammer corer with a 20 cm long plastic liner to take two adjacent soil cores of approximately 15 cm length from each sampling spot, first removing the organic horizon. This procedure was done to ensure that all samples were restricted to the mineral material, as the amount of organic matter was quite variable among samples. One core was returned to the lab for immediate analysis; the other was sealed at the bottom with a rubber stopper and covered on the top with a piece of foil, and returned to the soil to incubate for a month. In the two cedar swamps, all samples were taken from saprist peat in the hollows. Samples from Ml0 CS in Jun., Jul., and Oct. 2005 were taken from the hollows with overlying water, which made soil moisture over 100% by volume.

Soils returned to the laboratory were transported back to the laboratory in an iced cooler. Subsamples of the upper 5 cm of the soils were gently mixed and then wet samples were extracted with 2 M KC1 within 24 h, and the extracts analyzed for inorganic nitrogen contents ([NH.sup.4.sub.+]-N and N[O.sub.3] + N[O.sub.2] N) on a Lachat[R] QuickChem[R] FIA+ 8000 autoanalzyer (Milwaukee, WI, USA). Net N mineralization and net nitrification rates are calculated from differences of inorganic nitrogen content between the incubated core and the core immediately analyzed divided by the number of incubation days, expressed as "mg N [m.sup.-3] soil [day.sup.-1]." All nitrogen data are expressed on a volume basis because the very large difference in bulk densities of the mineral and organic soils made comparison on a mass basis misleading. Soil moisture was calculated based on dry weight and adjusted to a volumetric basis using the measured bulk density at each plot.

The following measurements were made on all soil samples: organic horizon thickness (cm) was measured in the field prior to extraction of the cores, soil pH (1:1 in [H.sub.2]O), soil electrical conductivity (EC, 1:1 in [H.sub.2]O) as [mu]S [cm.sup.-1], soil moisture as percent moisture (v/v), and soil organic matter measured as loss-on-ignition, expressed as kg [m.sup.-3] on the basis of oven-dry soil. Bulk density was determined for each plot as the average of three samples taken on one date and used to convert all values calculated on a mass basis to a volumetric basis.

Water table levels were measured in PVC wells on a bi-weekly (growing season) or monthly (winter) basis (Laidig et al., 2010; data courtesy of A. Brown, The Pinelands Commission). Water table levels for the dates on which soil samples were taken were estimated by developing regression equations between the partial-series well data and continuously-recorded water level data at the US Geological Survey's well located at M10C (USGS well MB1) over the period 19 Apr. 2004 to 4Jun. 2006 (data from R. Nicholson, U.S. Geological Survey). All regressions had [r.sup.2] values >0.90 (data from A. Brown, pers. comm.). These regressions were used to estimate water levels on the dates of soil sampling.


Soil variables measured on a monthly basis were compared using repeated measures analyses (PROC GLM, SAS 9.1.3) in which between-subjects factors were the site and the vegetation type, and the within-subjects factor was the month of data collection. Within-subjects (time factor) significance levels were adjusted for departures from sphericity using the Greenhouse-Geissler epsilon correction. Regressions, both simple linear and stepwise multiple regressions, were carried out either in SAS or in SigmaPlot ver. 11.0. Nonparametric Kruskal-Wallis and Mann-Whitney tests were used to test hypotheses about water levels, for which no transformation could normalize the data.

We used Mantel tests to compare matrices of values for the different soil properties with each other. This test compares the structure of distance matrices computed from the raw data for different sets of observations made on the same sample units. In this case, each matrix consisted of the full set of replicate observations within each plot for all plots within each site for each month. Thus, for each soil variable, the matrix consisted of 18 observations (rows; three replicates in each of six plots) for 12 mo (columns). This approach preserves both the spatial and the temporal patterns among the measurements (landscape gradients at the two sites, the monthly samples at each site), and incorporates all the variability observed within as well as between sites. The Mantel test thus compares the structure of each matrix of soil properties with the matrix of N cycling rates and properties. Tests were conducted using Euclidean distances metrics and a Monte Carlo randomization method to test the significance of the standardized Mantel statistic (r), analogous to a correlation coefficient between the distance matrices. The analysis is similar to path analysis, in that it provides measures of relationships among a set of variables but does not require adherence to the assumptions of regression analyses. Analyses were conducted on untransformed data, as preliminary analyses did not show large departures from normality. Mantel tests were conducted in PC-ORD ver. 5.0. Statistical significance was evaluated at [alpha] = 0.05 for all analyses.



Water tables.--Water tables in the six plots followed parallel patterns with levels in the PW and PH plots always below the ground surface, and levels in the cedar swamps at or above the surface for much of the year except late summer through Oct. in 2005 (Fig. 1a; directly observed values). Sites were significantly different from each other in annual median values (Kruskal-Wallis test, H = 135.211, P < 0.0001), as expected.

Soil properties.--Soil properties for variables measured monthly (Table 1) and other descriptors measured once during the study (Table 2) showed that the soils were strongly differentiated by their origin as organic or mineral. For all the variables measured monthly except for organic matter, there was significant variation over time and no significant differences between the two sites (Ml0 and M6). In all cases, the two cedar swamps were significantly different from the two mineral soil wetlands, which were not different from each other. Cedar swamps have lower mean pH, but higher EC, moisture, organic matter, extractable [NH.sub.4.sup.+]-N, and net N mineralization rates than do the PW and PH sites. Static properties (Table 2) showed the expected differences for mineral vs. organic wetland soils, with higher organic carbon and nitrogen and lower bulk density in the organic soils than in the mineral soils (P < 0.001). The C:N ratios in the four mineral soil sites (PW and PH) were significantly high and were much higher than for the organic soils of the cedar swamps (P < 0.001, Table 2).

Organic soils of the two cedar swamps stayed markedly wetter than mineral soils of the PW and PH wetlands throughout the study (Table 1, Fig. lb). Indeed, the four PW and PH plots all had similar levels of soil moisture to each other (Table 1; Fig. lb) despite differences in water tables. Soil moisture was more variable among samples and over time in the mineral-soil wetlands (5-6-fold range of variation in each of the four sites) than the cedar swamp plots (<2-fold range of variation).


Soil moisture varied with water tables in the mineral soil wetlands (Table 3). Simple linear regression relationships between the estimated water levels on the date of sampling and soil moisture were stronger (higher [r.sup.2] values) in the M10 site than the M6 site, for which the water tables could explain only about half of the variation in soil moisture (Table 3). In the cedar swamps, relationships between water tables and soil moisture were notably weaker, with a poor relationship found for one plot (M6 CS) and no relationship found for the other (Ml0 CS; Table 3). For the set of sites together (landscape level), soil moisture is significantly predicted by water table depth (Table 3) but with an [r.sup.2] value of only 0.49, implying that only half the variation in soil moisture across the landscape gradient is explained by water tables.

Water table depth (WT) did not explain variation on a per-plot basis for any other variable (net N mineralization rate, extractable [NH.sub.4.sup.+]-N, or organic matter) but did provide some explanatory power for the set of sites at the landscape scale (Organic matter = -15.77WT + 891.31, [r.sup.2] adj = 0.36, F = 40.96, P < 0.0001; [NH.sub.4.sup.+]-N = -0.03WT + 1.70, [r.sup.2] adj = 0.27, F = 27.35, P < 0.001). However, these relationships all had low explanatory power despite statistical significance. Notably, net N mineralization rates are not significantly explained by water table at either plot or landscape scales.

Nitrogen dynamics.--Extractable [NH.sub.4.sup.+]-N was two to three times higher in the cedar swamps than the mineral soil wetlands throughout the study period. Averaged over all time periods and plots, concentrations of [NH.sub.4.sup.+]-N were slightly higher in the M6 than the Ml0 sites with exceptions of mineral soils in the winter (Table 1; Fig. 2a). There were no significant differences among the four mineral soil plots, but repeated measures analysis showed that there were significant differences over time (Table 1). Maxima were observed in all plots during the winter in the mineral soil plots. In contrast, there was relatively little variation in [NH.sub.4.sup.+]-N concentrations over time in the Ml0 CS plot while the M6 CS plot had the [NH.sub.4.sup.+]-N concentrations peaked in the draught season (Sept. 2005) (Fig. 2a). But, the concentrations of [NH.sub.4.sup.+]-N in the CS plots were not significantly changed over time by statistics.

In almost all samples, extractable N[O.sub.3.sup.-]-N was near or below the detection limit (0.01 mg N [L.sup.-1]) of our method. Thus, net nitrification was rarely observed in any plots during the sampling year (data not shown). Occasional samples showed measureable rates; but when this occulted, it was not observed among the three samples per plot, leading to very high standard deviations of the mean values for each wetland. The lack of nitrification corresponds well to the lack of observed nitrate in the extractable soil samples.

Net N mineralization rates (Fig. 2b) varied with time (Table 1, significant 'time' factor) but did not show clear seasonal patterns. The only seasonal pattern apparent was in the Ml0 CS plot, where higher net N mineralization rates coincided with the drier conditions present in Sept., when water levels fell. There was also a high degree of variability, evident in the standard deviations in Fig. 2b, within each plot for each sampling date. In the mineral soil sites, there was no significant difference between the two wetland types (PW and PH). In the cedar swamps, there were no significant differences either between the two replicate sites due to the high degree of variability.

We examined the relationships between net N mineralization rate and extractable [NH.sub.4.sup.+]-N, and the potentially driving variables of soil moisture and organic matter using Mantel tests to compare the matrices of measurements at each replicate within each plot and over time. This test compares two matrices, each composed of rows of plot values and columns for each month of measurement, for pairs of soil variables. This approach tests the hypothesis that the structure of variation within two matrices is similar to each other, thus testing the spatial and temporal distribution of environmental values over the year of sampling. Mantel coefficients are analogous to regression path analysis coefficients, in that they show the strength of the relationship between the matrices. The result shows that soil moisture, organic matter, and initial extractable [NH.sub.4.sup.+]-N are all closely related to each other (Fig. 3). Soil moisture and organic matter each have a weak correlation with net N mineralization rate, but [NH.sub.4.sup.+]-N has no relationship with net N mineralization rate on this landscape scale. Net N mineralization rate is notably poorly related to the other soil variables.


Our results demonstrate that the two types of wetland on mineral soils were quite similar to each other in soil properties and nitrogen dynamics, despite significant differences in water table levels but are quite different from the cedar swamps. Water tables explain some of the variation in soil properties but leave much variation unexplained. These results did not support our conceptual model that water table levels are the driving variable for all other soil properties. Contrary to expectation, the nitrogen dynamics in all of the plots were only weakly related to within-plot variations in soil moisture and organic matter content, although moisture, organic matter, and extractable [NH.sub.4.sup.+]-N were all closely related to each other (Fig. 3). Soil moisture is a function of water table depth both within most sites and across the landscape gradient. This suggests that factors other than water table control both some soil properties and especially net N mineralization rates.


The patterns of relationship between soil moisture and water tables suggest that in addition to the absolute water table levels, several other factors affect soil moisture. First, the presence of partially confining clay lenses in the profile may disconnect regional water tables and surface soil moisture levels. The M6-PW and -PH plots had notably poor relationships between water tables and soil moisture (Table 3), which may reflect the presence of partially permeable clay lenses at these plots (unpub. obs. courtesy of R. Nicholson, U.S. Geological Survey), which perch the water table periodically after rains. Second, the high water-holding capacity of organic soils in the CS type (Mitsch and Gosselink, 2000; Rabenhorst and Swanson, 2000), may help insulate these stands from the effects of seasonally lowered water tables. Soil moistures in the cedar swamps were saturated or near-saturated (70-100% moisture content by volume) most of the time, with little seasonal variation (Fig. 1), despite the drop in water tables in Aug. through Sept. (Fig. 1), whereas decreases in soil moisture during these months were evident in the mineral soil wetlands. Third, the cedar swamps are also notable for the dense shade cast by the tree and shrub canopy (Little, 1958), and this shading may reduce evaporation from the soil surface and thus contribute to the ability of the soils to retain moisture from both precipitation and ground water discharge sources during water table declines. These results suggest that short-term (1-2 mo) decreases in water tables will have little effect on soil properties or N dynamics of the cedar swamps, in contrast to the mineral-soil wetlands (PH and PW).

Nitrogen dynamics in these wetlands are characterized by the near-absence of nitrification under most conditions. All observations of nitrate concentrations above the detection limit (0.01 mg N [L.sup.-1]) in extracts of soil samples and in pore water samples were isolated occurrences within the set of three replicate samples for any given sampling date. While we cannot determine from our data whether these values represent measurement error or true, albeit rare, nitrate production in the sediments, the high variability and the infrequency of these events suggest that under the conditions of this study, nitrate production does not occur, even during summer drawdowns of the water table. In previous studies of N mineralization in these soils (Poovaradom et al., 1988a, b; Sangemeswaran, 1995), extractable nitrate and nitrification were also seldom observed, and similar results have been found in other bogs and fens (Fellman and D'Amore, 2007). Indeed, in a laboratory study of the Atsion soil that tested a wide range of temperature and moisture conditions, nitrate was almost never observed (Poovaradom et al., 1988b). Low pH is a well-known inhibitory factor for ammonia-oxidizing bacteria (Paul, 2007), and thus the extremely low pH of all soils in this study (<4) is likely to be the main controlling factor preventing nitrification responses to fluctuations in soil moisture. Nitrification has been observed in histosols subject to prolonged but not intermittent drainage (e.g., Hanlon et al., 1997; Prevost et al., 1999; Tarre and Green, 2004). In these studies, however, the drained organic soils are also associated with somewhat higher pH values (>4) and higher mineral content than those studied here. Contrary, in estuarine marshes, nitrification was evident in soils, which pHs were near neutral or alkaline (Bai et al., 2007, 2010). Our previous laboratory studies of the effects of soil moisture on nitrification in soils of these study plots (Yu and Ehrenfeld, 2009) similarly found that prolonged periods (months) of unsaturated conditions resulted in measurable nitrification but that biweekly moisture fluctuations between saturated and unsaturated conditions did not permit nitrification. Thus, both our laboratory and field studies suggest that only with prolonged periods (months) of unsaturated conditions will net nitrification be observed in these very acidic soils.

Net N mineralization rates and extractable inorganic N[H.sub.4.sup.+]-N were much higher in the organic soils than in the mineral soils (Table 1, Fig. 2). The N[H.sub.4.sup.+]-N concentrations in the wetlands was highly correlated with the high organic matter content (Fig. 3), but net N mineralization rates were poorly to the organic matter content of the soil, despite nearly 10-fold variations among the replicate samples, and contrary to expectation. Net N mineralization rates were also poorly related to soil moisture contents within plots. Agehara and Warucke (2005) found that nitrogen mineralized from organic materials did not vary with soil moisture in some types of organic materials. Poovaradom et al. (1988b) tested the role of soil moisture in laboratory incubations of organic horizons of the Atsion soil and found that moisture had little effect on rates or on calculated total mineralizable N until the soil was at very low moisture levels (-1.5 MPa). Drury et al. (2003) showed that in mineral agricultural soils there is a range of water contents over which N mineralization does not vary but that it will decrease below some lower limit and above some upper limit. They found that over the range of soils and soil conditions (with and without amendments and compaction), net N[H.sub.4.sup.+]N production was insensitive to moisture levels between 20% and 80% water holding capacity. While these values cannot be extrapolated directly to the sandy mineral soils and organic soils in this study, the concept that N mineralization is insensitive to variations in moisture content over a large range of moisture contents is probably applicable. Fellman and D'Amore (2007), in a comparative study of N mineralization in Alaskan bogs, fens, and riparian wetlands also found that the rates were not well correlated with soil properties other than pH.

Other studies have emphasized the importance of total soil N and the C:N ratio as more important controls on net N mineralization rates than soil moisture. For example, Knoepp and Vose (2007) found that across elevation gradients in the southern Appalachian mountains, moisture and temperature were only important at the extreme ends of the gradients; otherwise, total soil N and C:N ratios were more important determinants of N mineralization. Fernandez et al. (2000) found, in a study of hardwood and softwood forests in Maine, that C:N ratio of the forest floor was the best predictor of net N mineralization rate. Parfitt et al. (2005) also found it was true for pasture soils under different types of management. Our sites have a large difference in C:N ratio between the organic and mineral soils (Table 2), which corresponds to the large differences in net N mineralization rates between these two soil types, regardless of water tables or soil moisture conditions.

The results of the field studies are in line with the results of the fluctuating-moisture incubation experiments conducted in the laboratory (Yu and Ehrenfeld, 2009). In those studies, increases in nitrification and nitrate concentrations during two-week periods of drought (30% WHC) were very small in absolute magnitude, and no differences in net nitrate or net mineral N production were observed. Although precipitation was very limited during Aug. and Sept. of the study period, there were several small rainfall events, and the maximum period without any precipitation was 15 d. The patterns of soil moisture and nitrogen dynamics observed suggest that even small precipitation events are sufficient to prevent extreme reductions in moisture content or increases in nitrification.


In summary, water table differences along the shallow gradient in the Pinelands affect soil moisture, which in turn affects organic matter content and extractable inorganic N. However, net N mineralization rate is poorly related to either water tables or other soil properties. This suggests that other factors affect N availability in these wetlands. The large differences in volumetric measures of net N mineralization rate between the cedar swamps and the PH and PW wetlands (Table 1, Fig. 2), combine with the weak relationships to soil moisture or organic matter, suggests that organic matter quality (C:N ratio; Table 2) and soil type (organic vs. mineral) are more important controlling variables for N availability in these wetlands than is water table directly.

Acknowledgments.--Funding for this study was provided through the Water Supply Fund in accordance with New Jersey Public Law 2001, Chapter 165. We thank Allison Brown, The Pinelands Commission, and Robert Nicholson, U.S. Geological Survey, for assistance in locating plots and for supplying water table data. We also thank Jodi Messina for dedicated assistance with laboratory and field sampling, Kenneth Elgersma for advice on statistical analyses, and Dr. Josh Caplan for comments on an earlier version of the manuscript.


AERTS, R., J. T. A. VERHOEVEN AND D. F. WHIGHAM. 1999. Plant-mediated controls on nutrient cycling in temperate fens and bogs. Ecology, 80:2170-2181.

AGEHARA, S. AND D. D. WARNCKE. 2005. Soil moisture and temperature effects on nitrogen release from organic nitrogen sources. Soil Sci. Soc. Am. J., 69:1844-1855.

AYARAS, J. AND Y. GAO. 2007. Atmospheric nitrogen deposition to the Mullica River-Great Bay estuary. Mar. Environ. Res., 64:590-600.

BAI, J. H., W. DENG, Q. G. WANG, B. S. CUI AND Q. Y. DING. 2007. Spatial distribution of inorganic nitrogen contents of marsh soils in a river floodplain with different flood frequencies from soil-defrozen period. Environ. Monit. Assess., 134:421-428.

--, H. F. GAO, W. DENG, Z. F. YANG, B. S. CUI AND R. XIAO. 2010. Nitrification potential of marsh soils from two natural saline-alkaline wetlands. Biol. Ferti. Soils., 46:525-529.

BAYLEY, S. E., M. N. THORMAN AND A. R. SZUMIGALSKI. 2005. Nitrogen mineralization and decomposition in western boreal bog and fen peat. Ecoscience, 12:455-465.

BECHOLD, S.J. AND R.J. NAIMAN. 2006. Soil texture and nitrogen mineralization potential across a riparian toposequence in a semi-arid savanna. Soil Biol. Biochem., 38:1325-1318.

BRIDGHAM, S. D., K. UPDEGRAFF AND J. PASTOR. 1998. Carbon, nitrogen and phosphorus mineralization in northern wetlands. Ecology, 79:1545-1561.

CHAPIN, C. T., S. D. BRIDGHAM, J. PASTOR AND K. UPDEGRAFF. 2003. Nitrogen, phosphorus, and carbon mineralization in response to nutrient and lime additions in peatlands. Soil Sci., 168:409-420.

DIGHTON, J., A. R. TUININGA, D. M. GRAY, R. E. HUSKINS AND T. BELTON. 2004. Impacts of atmospheric deposition on New Jersey pine barrens forest soils and communities of ectomycorrhizae. For. Ecol. Manag., 201:133-144.

DRURY, C. F., T. Q. ZHANG AND B. D. KAY. 2003. The non-limiting and least limiting water ranges for soil nitrogen mineralization. Soil Sci. Soc. Am. J., 67:1388-1404.

EHRENFELD, J. G. 1986. Wetlands of the New Jersey Pine Barrens: the role of species composition in community function. Am. Midl. Nat., 115:301-313.

-- AND M. GULICK. 1981. Structure and dynamics of hardwood swamps in the New Jersey Pine Barrens: contrasting patterns in trees and shrubs. Am. J. Bot., 68:500-512.

-- AND J. P. SCHNEIDER. 1991. Chamaecyparis thyoides wetlands and suburbanization: effects of nonpoint source water pollution on hydrology and plant community structure. J. Appl. Ecol., 28:467-490.

FELLMAN, J. B. AND D. V. D'AMORE. 2007. Nitrogen and phosphorus mineralization in three wetland types in southeast Alaska, USA. Wetlands, 27:44-53.

FERNANDEZ, I. K., J. A. SIMMONS AND R. D. BRIGGS. 2000. Indices of forest floor nitrogen status along a climate gradient in Maine, USA. For. Ecol. Manag., 134:177-187.

FORMAN, R. T. T. (ed.). 1998. Pine Barrens: Ecosystem and Landscape. Rutgers University Press, New Brunswick, New Jersey. 601 p.

GROFFMAN, P. M., A. J. GOLD AND R. C. SIMMONS. 1992. Nitrate dynamics in riparian forests - microbial studies. J. Environ. Qual., 21:666-671.

HANLON, E. A., D. G. ANDERSON AND O. A. DIAZ. 1997. Nitrogen mineralization in histosols of the Everglades Agricultural Area. Comm. Soil Sci. Plan., 98:73-87.

HANSON, G. C., P. M. GROFFMAN AND A. J. GOLD. 1994. Denitrification in riparian wetlands receiving high and low groundwater nitrate inputs. J. Environ. Qual., 23:917-922.

HOLDEN, J., P.J. CHAPMAN AND J. C. LABADZ. 2004. Artificial drainage of peatlands: hydrological and hydrochemical process and wetland restoration. Prog. Phys. Geog., 98:95-123.

JONASSON, S. AND G. R. SHAVER. 1999. Within-stand nutrient cycling in arctic and boreal wetlands. Ecology, 80:2139-2150.

KELLER, J. K., J. R. WHITE, S. D. BRIDGHAM AND J. PASTOR. 2004. Climate change effects on carbon and nitrogen mineralization in peatlands through changes in soil quality. Glob. Change Biol., 10:1053-1064.

KNOEPP, J. D. AND J. M. VOSE. 2007. Regulation of nitrogen mineralization and nitrification in Southern Appalachian ecosystems: Separating the relative importance of biotic vs. abiotic controls. Pedobiologia, 51:89-97.

LAIDIG, K. J. AND R. A. ZAMPELLA. 1999. Community attributes of Atlantic white cedar (Chamaecyparis thyoides) swamps in disturbed and undisturbed Pinelands watersheds. Wetlands, 19:35-49.

--, --, A. M. BROWN AND N. A. PROCOPIO. 2010. Development of vegetation models to predict the potential effect of groundwater withdrawals on forested wetlands. Wetlands, 30:489-500.

LITTLE, S. 1958. Silvical characteristics of Atlantic White-Cedar (Chamaeayparis thyoides). In: U.S. Forest Service Northeastern Forest Experiment Station Upper Darby. Pennsylvania. 16 p.

MITSCH, W. AND J. G. GOSSELINK. 2000. Wetlands, 3rd ed. Van Nostrand Reinhold, New York, New York. 910 p.

OWEN, J. S., M. K. WANG, CH. H. WANG, H. B. KING AND H. L. SUN. 2003. Net N mineralization and nitrification rates in a forested ecosystem in northeastern Taiwan. For. Ecol. Manag., 176:519-530.

PARFITT, R. L., G. W. YEATES, D.J. ROSS, A. D. MACKAY AND P.J. BUDDING. 2005. Relationships between soil biota, nitrogen and phosphorus availability, and pasture growth under organic and conventional management. Appl. Soil Ecol., 28:1-13.

PAUL, E. A. (ed.). 2007. Soil Microbiology, Ecology and Biochemistry. 3rd ed. Academic Press, San Diego, California. 532 p.

POOVARADOM, S., R. L. TATE AND R. A. BLOOM. 1988a. Nitrogen mineralization rates of the acidic, xeric soils of the New Jersey Pinelands: field studies. Soil Sci., 145:257-263.

--, --, AND --. 1988b. Nitrogen mineralization rates of the acidic, xeric soils of the New Jersey Pinelands: laboratory studies. Soil Sci., 145:337-344.

PREVOST, M., A. P. PLAMANDON AND P. BELLEAU. 1999. Effects of drainage of a forested peatland on water quality and quantity. J. Hydrol., 214:130-143.

RABENHORST, M. C. AND D. SWANSON. 2000. Histosols. In: M. E. Sumner (ed.). Handbook of Soil Science. CRC Press, Boca Raton, Florida. E183-E209.

REDDY, K. R. AND R. D. DELAUNE. 2008. Biogeochemistry of Wetlands Science and Application. CRC Press, Boca Raton, Florida. 774 p.

ROBERTSON, G. P., D. WEDIN, P. M. GROFFMAN, J. M. BLAIR, E. A. HOLLAND, K. NADELHOFFER, C. S. BLEDSOE and P. SOLLINS (eds.). Standard Soil Methods for Long-term Ecological Research. Oxford University Press, New York, New York, 258-271.

SANGEMESWARAN, S. 1995. Nitrogen dynamics in three types of wetlands of the New Jersey Pinelands. M. S. Thesis, Graduate Program in Ecology and Evolution, Rutgers University, New Brunswick, New Jersey.

Soil Survey Staff Natural Resources Conservation Service. Official Soil Series Descriptions [Online Version]. U.S. Department of Agriculture, Lincoln, Nebraska).

TARRE, S. AND M. GREEN. 2004. High-rate nitrification at low pH in suspended- and attached-biomass reactors. Appl. Environ. Microb., 70:6481-6487.

UPDEGRAFF, K., S. D. BRIDGAM, J. PASTOR AND C. A. JOHNSTON. 1995. Environmental and substrate controls over carbon and nitrogen mineralization in northern wetlands. Ecol. Appl., 5:151-163.

VERHOEVEN, J. T. A., D. F. WHIGHAM, R. VAN LOGTESTIJN AND J. O'NEILL JR. 2001. A comparative study of nitrogen and phosphorus cycling in tidal and non-tidal riverine wetlands. Wetlands, 21:210-222.

WASSEN, M. J. AND H. OLDE VENTERINK. 2006. Comparison of nitrogen and phosphorus fluxes in some European fens and floodplains. Appl. Veg. Sci., 9:213-222.

YU, S. AND J. EHRENFELD. 2009. The effects of changes in soil moisture on nitrogen cycling in acid wetland types of the New Jersey Pinelands (USA). Soil Biol. Biochem., 41:2394-2405.

ZAMPELLA, R. A., G. MOORE AND R. E. GOOD. 1992. Gradient analysis of pitch pine (Pinus rigida Mill) lowland communities in the New Jersey Pinelands. Bull. Torrey Bot. Club, 119:253-261.



(1) Corresponding author: e-mail:; Telephone & FAX: +86-592-6190778


Department of Ecology, Evolution, and Natural Resources, SEBS, Rutgers University, New Brunswick,

New Jersey 08901



Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China and Department of Ecology,

Evolution, and Natural Resources, SEBS, Rutgers University, New Brunswick, New Jersey 08901
TABLE 1.--Mean values of soil variables measured each month at
three replicate locations within each site. Statistics are the F
values and significance levels  of two-way repeated measures
analyses of variance testing vegetation types and months. Tests
are based on n = 12 observations, one per month, for each
variable and site. Significance levels: * P < 0.05; ** P < 0.01;
*** P < 0.001; and ns means "not significant." Only significant
factors are listed. Significance  is based on between-subject
(sites, types, interactions) and within-subject (time,
interactions repeated measures analyses of variance, with p level
Site corrected  using the Greenhouse-Geisser epsilon correction
for departures from sphericity). Comparisons of vegetation type
means based on tests of least square means

                     Type                  pH

M10            Pine wetland (PW)   3.83 [+ or -] 0.02
               Pine hardwood       3.67 [+ or -] 0.03
               swamp (PH)
               Cedar swamp (CS)    3.65 [+ or -] 0.02

M6             Pine wetland (PW)   3.82 [+ or -] 0.03
               Pine hardwood       3.82 [+ or -] 0.02
                 swamp (PH)
               Cedar swamp (CS)    3.62 [+ or -] 0.02

Statistical    Type (df = 2)          CS < PH = PW
result                                   12.9 ***
(F value)
               Time (df = 11)            13.7 ***
               Time X Type                 ns
                 (df = 11)

                     Type             [cm.sup.-1])

M10            Pine wetland (PW)   49.6 [+ or -] 3.38
               Pine hardwood       64.0 [+ or -] 3.88
               swamp (PH)
               Cedar swamp (CS)    75.1 [+ or -] 2.00

M6             Pine wetland (PW)   48.4 [+ or -] 3.02
               Pine hardwood       51.5 [+ or -] 2.30
                 swamp (PH)
               Cedar swamp (CS)    85.8 [+ or -] 2.35

Statistical    Type (df = 2)         CS > PH = PW
result                                 36.7 ***
(F value)
               Time (df = 11)           5.34 ***
               Time X Type               ns
                 (df = 11)

                                        Moisture %
                                          (g 100
                     Type              [cm.sup.-3])

M10            Pine wetland (PW)   17.1 [+ or -] 1.47
               Pine hardwood       22.8 [+ or -] 1.65
               swamp (PH)
               Cedar swamp (CS)    86.5 [+ or -] 5.48

M6             Pine wetland (PW)   13.8 [+ or -] 1.00
               Pine hardwood       15.9 [+ or -] 1.12
                 swamp (PH)
               Cedar swamp (CS)    73.8 [+ or -] 2.00

Statistical    Type (df = 2)           CS > PH = PW
result                                  136.4 ***
(F value)
               Time (df = 11)            3.64 *
               Time X Type                ns
                 (df = 11)

                                      Organic matter
                     Type             (kg [m.sup.-3])

M10            Pine wetland (PW)     203.8 [+ or -] 27.2
               Pine hardwood         288.3 [+ or -] 35.7
               swamp (PH)
               Cedar swamp (CS)    1,178.5 [+ or -] 65.1

M6             Pine wetland (PW)     212.2 [+ or -] 18.3
               Pine hardwood        198.01 [+ or -] 17.3
                 swamp (PH)
               Cedar swamp (CS)    1,281.6 [+ or -] 37.4

Statistical    Type (df = 2)           CS > PH = PW
result                                   94.4 ***
(F value)
               Time (df = 11)               ns
               Time X Type                  ns
                 (df = 11)

                     Type           (g N [m.sup.-3])

M10            Pine wetland (PW)   0.47 [+ or -] 0.07
               Pine hardwood       0.55 [+ or -] 0.08
               swamp (PH)
               Cedar swamp (CS)    1.85 [+ or -] 0.14

M6             Pine wetland (PW)   0.54 [+ or -] 0.05
               Pine hardwood       0.44 [+ or -] 0.04
                 swamp (PH)
               Cedar swamp (CS)    2.71 [+ or -] 0.18

Statistical    Type (df = 2)          CS > PH = PW
result                                  43.3 ***
(F value)
               Time (df = 11)            3.12 *
               Time X Type                 ns
                 (df = 11)

                                          Net N
                                     (mg N [m.sup.-3]
                     Type                  day)

M10            Pine wetland (PW)    11.6 [+ or -] 4.04
               Pine hardwood        13.0 [+ or -] 7.38
               swamp (PH)
               Cedar swamp (CS)    108.5 [+ or -] 26.0

M6             Pine wetland (PW)   22.1 [+ or -] 13.26
               Pine hardwood        11.4 [+ or -] 3.61
                 swamp (PH)
               Cedar swamp (CS)    137.8 [+ or -] 24.70

Statistical    Type (df = 2)           CS > PH = PW
result                                   26.9 ***
(F value)
               Time (df = 11)            9.15 ***
               Time X Type               6.37 ***
                 (df = 11)

TABLE 2.--Soil properties measured once during the study. n = 3
for each measurement and data  were present as mean [+ or -]
standard error (SE). Statistics are the F values and significance
levels of two-way  analyses of variance testing site and
vegetation types. Tests are based on n = 3 replicates for each
variable and site. Significance levels: * P < 0.05; ** P < 0.01;
*** P < 0.001; and ns means "not  significant." Only significant
factors are listed. Comparisons of sit and vegetation type means
based on  tests of least square means

Site              Type                %C                  %N

M10           Pine wetland    4.3 [+ or -] 0.74    0.1 [+ or -] 0.02
              Pine hardwood   3.5 [+ or -] 0.49    0.1 [+ or -] 0.01
              Cedar swamp     13.3 [+ or -] 3.11   0.5 [+ or -] 0.13

M6            Pine wetland    4.0 [+ or -] 0.26    0.1 [+ or -] 0.01
              Pine hardwood   6.6 [+ or -] 0.47    0.2 [+ or -] 0.01
              Cedar swamp     26.3 [+ or -] 3.31   1.1 [+ or -] 0.12

Statistical   Site (df = 1)        M6 > M10            M6 > M10
result                             11.4 **              13.6 **
(F value)
               Type (df=2)       CS > PW = PH        CS > PW = PH
                                   42.6 ***            59.8 ***
               Site X Type          6.6 *               8.3 **
                (df = 2)

Site              Type           C:N (molar)          Bulk density

M10           Pine wetland    41.0 [+ or -] 1.38   0.91 [+ or -] 0.13
              Pine hardwood   42.2 [+ or -] 1.55   0.62 [+ or -] 0.10
              Cedar swamp     26.8 [+ or -] 1.29   0.15 [+ or -] 0.02

M6            Pine wetland    37.3 [+ or -] 0.88   1.03 [+ or -] 0.20
              Pine hardwood   37.0 [+ or -] 4.36   1.01 [+ or -] 0.18
              Cedar swamp     25.6 [+ or -] 0.65   0.21 [+ or -] 0.01

Statistical   Site (df = 1)           ns                   ns
(F value)
               Type (df=2)       PW = PH > CS         PW = PH > CS
                                   26.5 ***             22.0 ***
               Site X Type            ns                   ns
                (df = 2)

TABLE 3.--Linear regression between ground water table levels
(WT, cm below the surface of hollows)  and volumetric soil
moisture (SM), based on estimated water level on the day of soil
sampling. Soil  moisture is a mean of the three replicates for
each month analysis

Site     Type                  Equation                 sub.adj]

M10      Pine wetland          SM = -0.270WT + 27.78      0.61
         Pine hardwood swamp   SM = -0.420WT + 34.13      0.81
         Cedar swamp                     ns                ns

M6       Pine wetland          SM == -0.253WT + 20.77     0.52
         Pine hardwood swamp   SM = -0.367WT + 23.44      0.55
         Cedar swamp           SM = -0.389WT + 77.15      0.35

All                            SM = -0.94WT + 64.79       0.49

Site     Type                  F (P)

M10      Pine wetland          15.97 (0.0025)
         Pine hardwood swamp   40.87 (<0.001)
         Cedar swamp           ns

M6       Pine wetland          11.68 (0.0068)
         Pine hardwood swamp   11.82 (0.0064)
         Cedar swamp            7.45 (0.016)

All                            69.44 (0.001)
Gale Copyright: Copyright 2012 Gale, Cengage Learning. All rights reserved.