Introduction

In addition to tracing the flow and utilization of organic matter in foodwebs, stable N isotopic ratios (δ15N) have been used as an indicator of the historical nutrient status of lakes (Brenner et al. 1999), estuaries (Orem et al. 1999), and more recently, wetlands (Novak et al. 1999; Wooller et al. 2003). In these studies the δ15N of sedimented/accreted organic matter, either algal-, microbial- or macrophyte-derived, is used to infer changes in nutrient conditions within a given system. Many processes are known to affect the production, composition, and fate of organic matter, and therefore, knowledge of these factors and their effects on organic matter δ15N is essential for successful application of this technique in a given system.

Conceptually, the processes affecting the δ15N of organic matter can be divided into two categories, namely, those determining the isotopic ratio of the living biomass, and those affecting the isotopic ratio of the detrital material during decomposition. In the case of living biomass, δ15N is largely derived from the combination of δ15N of the N source and the physiological mechanisms governing fractionation during plant uptake (Evans 2001; Handley and Raven 1992). In this manner, most macrophytes have δ15N values below that of the available N within a system. As plant N demand increases, less isotopic discrimination during N uptake/assimilation will occur until ultimately, the plant becomes isotopically identical to its N source. This is particularly important during the uptake of NH +4 , and as a result, macrophyte δ15N is beginning to be proven useful at indicating relative N and P limitation in wetlands (Mckee et al. 2002; Clarkson et al. 2005; Inglett and Reddy 2006).

In a recent study of wetland macrophyte δ15N, Inglett and Reddy (2006) observed higher δ15N in plants growing near the discharge of high P drainage waters in the Everglades. In peat-based wetland systems such as the Everglades, the overwhelming majority of soil N is that of plant-derived organic-N, and therefore, it is logical that the isotopic composition of the soil would reflect that of the litter materials contributed by the dominant macrophytes. However, a number of processes can alter organic matter δ15N during decomposition, and thus can greatly affect interpretation of organic matter δ15N patterns (discussed by Lehmann et al. 2002). Among these, immobilization can increase or decrease δ15N by the incorporation of external N (Benner et al. 1991; Fogel and Tuross 1999). During mineralization, loss of high δ15N compounds (e.g., amino acids) can also lower δ15N; however, more frequently, decomposition is thought to increase δ15N of litter material through the loss of isotopically light N (Nadelhoffer and Fry 1988). Also, most mechanisms of N loss (e.g., nitrification, denitrification) are selective for 14N (Bedard-Haughn et al. 2003), therefore, preferential loss of excess 14N from a system over time leaves the residual N pool enriched in 15N (Hogberg 1990, 1991; Martinelli et al. 1999). By this mechanism, alteration of N turnover and retention rates could potentially change the δ15N of the existing soil N reservoir.

Further complicating interpretation of organic matter δ15N is the fact that a major contributor to the isotopic signature of macrophytes in a peat system undoubtedly comes from N mineralized from the peat soil (Nadelhoffer et al. 1996). In this manner, the isotopic changes observed in the plants may be the direct result of isotopic changes in the soil, and vice versa. In nutrient-impacted systems, this results in a circular argument, where it is unknown which isotopic change (in the soil or the plant) precedes the other. Uncertainty in the origin of the derived δ15N signature has important implications regarding the chronological accuracy of δ15N changes in a peat profile. Currently, there are few datasets to refine the importance of either pre- or post-depositional processes in affecting changes to soil δ15N.

In order to better understand the causes of increased wetland δ15N, and to improve the use of δ15N as a predictor of N cycle alteration, it is necessary to assess the relative contribution of pre- and post-depositional mechanisms affecting δ15N in ecosystem components. For this reason, the following study was undertaken to investigate spatial and historical patterns of wetland soil δ15N as they relate to macrophyte δ15N within a well-studied P enrichment gradient of the Florida Everglades ecosystem. Additional experiments were then conducted to test two possible mechanisms for altering N isotopic ratios. Two approaches were taken. In the first approach, δ15N was measured on samples obtained from two previous studies including: (1) a hydroponic growth study of wetland macrophytes grown at both low and high levels of P enrichment (Lorenzen et al. 2001), and (2) a long-term field P dosing experiment in a historically P-limited soil (McCormick and O’Dell 1996; Newman et al. 2004). The second approach utilized a laboratory study to attempt to directly measure the isotopic composition of N mineralized from wetland soils incubated with various levels of added P.

Materials and methods

Site description

Water conservation area 2A is a diked, 547-km2 portion of the northern Everglades ecosystem (Fig. 1). WCA-2A is characterized as a peat-based, freshwater wetland underlain by limestone substrata. In its natural state, WCA-2A is a highly P-limited system; however, inputs of high-P agricultural drainage water along the northeastern perimeter have decreased P limitation in impacted areas. The result of the drainage discharges has been the creation of a P enrichment gradient (Koch and Reddy 1992; Reddy et al. 1998), and a shift toward greater N limitation near the inflows (McCormick et al. 1996; Inglett et al. 2004). The vegetation of WCA-2A has also shifted along the P gradient from interior marshes consisting primarily of sawgrass (Cladium jamaicense Crantz) to nearly monotypic stands of cattail (Typha domingensis Pers.) in the nutrient-affected inflow areas (Childers et al. 2003; Craft and Richardson 1997).

Fig. 1
figure 1

Location of water conservation area 2A, the transect of stations used for spatial characterization of stable N isotopic ratios (δ15N), the mesocosm P-dosing experiment (McCormick et al. 1996; Newman et al. 2004), and sites of soil collection for the bottle mineralization experiment of this study

Full size image

Spatial and historical patterns of soil δ15N

Spatial patterns of δ15N were assessed using samples collected in September 2000 from a transect of ten sites spanning the WCA-2A nutrient gradient (Fig. 1). At each of these sites, samples of soil (0–10 cm) and surface litter were collected at random within an approximately 10-m × 10-m marsh area. Soil samples were taken using a sharpened, thin-wall aluminum tube (10 cm diameter). Samples were placed on ice and returned to the laboratory where soils were screened to remove macroorganic components (e.g., large roots). All samples were dried at 55°C and ground to pass a 2-mm mesh using a Wiley mill. Subsamples of soil and litter were then ball milled for bulk chemical and isotopic analyses.

Historical patterns of δ15N were assessed in vertical peat profiles using samples of intact peat cores from a previous study of WCA-2A (Reddy et al. 1993, Fig. 1). These cores were obtained in 1991 from sites along a similar transect to that used in this current study and correspond to sites of high (1.8 km), medium (3.5 km), and low (8.3 km) levels of P enrichment. The cores were sectioned in 1-cm intervals, screened to remove large root material, dried at 70°C, and ball milled for chemical analysis. For the current study, these samples were again analyzed for total N (TN), total C (TC), and δ15N in February 2002.

Effect of P addition on δ15N of soil N mineralization products

P loading has been shown to increase the overall activity of soil microorganisms in P-limited Everglades soils (Amador and Jones 1993; Reddy et al. 1999). This increased activity is also believed to result in increased NH +4 release from Everglades soils amended with P (Newman et al. 2004; White and Reddy 2000). The determination of N isotopic composition of N mineralized from WCA-2A soils was conducted using surface soils (0–10 cm) collected in February 2003 from two sites representing the high- and low-nutrient extremes of the WCA-2A transect (sites 1 and 10, Fig. 1). Bulk soils from each site were screened to remove large root and rhizome material and homogenized. Three replicate subsamples of soil (200 g wet weight) from each site were placed into 500-ml media bottles to which was added 300 ml of distilled–deionized water containing either 0 (control), 3.5, or 7.0 mg P (added as Na2HPO4). The 3.5 and 7.0 mg P additions were designed to increase the soil P level of the non-impacted site 10 soils by roughly 50 and 100%, respectively. O2 was removed from the bottle solution by bubbling with N2 for 5 min. The closure threads were wrapped in parafilm and the bottles were incubated horizontally in the dark inside a temperature-controlled shaker for 15 days at 25°C. Bottles were shaken horizontally once daily for 1 min at 100 r.p.m.

Three replicate bottles representing each site and receiving no P addition were collected initially to characterize water-extractable soil N. The remaining 18 bottles collected after the 15-day incubation were processed in a similar manner as follows. Each bottle was shaken to create a slurry and poured into replicate 250-ml centrifuge bottles and centrifuged at 6,000 r.p.m. for 10 min. The supernatant from each bottle was filtered (Whatman no. 41) and either acidified (H2SO4) for determination of NH +4 and dissolved total Kjeldahl N (TKN) or left unacidified for determination of δ15N of NH +4 and TN.

Long-term effect of P addition on soil δ15N

The effect of long-term P dosing on soil δ15N was assessed using samples obtained from a 5-year P dosing experiment conducted by the South Florida Water Management District (McCormick and O’Dell 1996; Newman et al. 2004). This experiment was conducted at a non-impacted site in the interior of WCA-2A (Fig. 1). At this location, 24 individual plots (1.8 m2) were designated and either enclosed with transparent fiberglass cylinders (1.2 m high × 1.5 m diameter) (21 plots) or left open as controls (three plots). The enclosed plots were dosed weekly with ortho-phosphate (NaH2PO4) at seven rates (0–12.8 g P m−2 year−1) from June 1995 to April 2000. Surface soil samples (0–3 cm) were obtained periodically throughout the dosing period, dried at 85–90°C, and ground for nutrient determination. For this current analysis, only samples of the closed and open controls (0 g P added) and highest loading (12.8 g P m−2 year−1) treatments were used. These archived samples were analyzed for TN, TC, and δ15N in April, 2004, at the University of Florida, Gainesville, Florida.

Effect of P addition on macrophyte δ15N

One test of the effect of P addition on δ15N of macrophytes is to compare plants grown at different levels of P, but the same level of N. It was hypothesized that with P addition, increasing N demand would decrease plant discrimination relative to the N source (as NH +4 ). In this manner, the potential for increasing plant δ15N through altered N uptake would be illustrated. This effect was determined using samples obtained from an experiment by Lorenzen et al. (2001), where specimens of Everglades Cladium and Typha were grown under constant N supply at varying levels of P.

This experiment was carried out at the PhytoNutriTron hydroponic growth facility at the Department of Plant Ecology, University of Aarhus (Lorenzen et al. 1998). In this facility, plants were grown in climate-controlled chambers in solution culture maintained at P levels ranging from 10 μg P l−1 (control) to the highest level (500 μg P l−1), while N (as NH +4 ) was supplied at a constant level (2.4 mg l−1) in all treatments. Nutrient levels were continuously maintained using an automated monitoring and dosing system which ensured that concentrations were maintained within ±10% of the desired set point during experimental periods. At 0, 12 and 18 days of the experiment, from two to four plants from each replicate treatment were harvested at random and divided into the main shoot system (leaves and shoot base), rhizomes and roots. The fractions were washed in distilled water, oven-dried at 80°C, and ground for nutrient analysis. For this current study, only samples collected at 18 days from the control (10 μg P l−1) and high (500 μg P l−1) treatment were used. These stored samples were analyzed for TN, TC, and δ15N in January 2004 at the University of Florida, Gainesville, Florida.

Chemical and isotopic analyses

NH +4 was measured colorimetrically on acidified extract samples using a Technicon autoanalyzer (method 350.1, USEPA 1993). TKN was measured similarly following Kjeldahl digestion (Method 352.1, USEPA 1993). Non-acidified extracts were frozen and prepared for δ15N analysis of TN by lyophilizing 20-ml aliquots and analyzing the resulting powder for TN and δ15N. δ15N of NH +4 was measured using an adapted version of the teflon tape diffusion method (Sorensen and Jensen 1991; Stark and Hart 1996). Briefly, this method utilized MgO (0.2 g per 70 ml solution) to convert sample NH +4 (>75 μg N) to NH3 gas which was then trapped using a glass fiber filter disk (7 mm diameter, Whatman GF-D) acidified with 20 μl of 0.5 M KHSO4. The acidified filter disk was protected by placing it inside a folded piece of commercially-available teflon pipe sealing tape (2.5 cm-wide × ∼4 cm long). The teflon tape packets were press-sealed along the edges and then floated on the surface of the sample solution containing MgO inside in an air-tight vessel (standard 500-ml Mason jar). Jars containing packets were placed into an incubator/shaker (70 r.p.m. at 30°C) and allowed to diffuse for 6 days. The teflon tape packets were then removed from the jars, rinsed with distilled–deionized water, gently blotted, and allowed to dry overnight in a desiccator with a beaker of concentrated H2SO4. After drying, the packets were opened to remove the glass fiber filter disk which was then folded and wrapped in a tin capsule for TN and N isotopic analysis.

Total N, TC, and δ15N ratios of soil, plant tissue, and glass fiber diffusion disks were determined using a Costech model 4010 elemental analyzer (Costech Analytical Industries, Valencia, Calif.) coupled to a Finnigan MAT DeltaPlusXL mass spectrometer (continuous flow isotope ratio mass spectrometry; Thermo Finnigan, San Jose, Calif.) via a Finnigan Conflo II interface. Ratios of 15N:14N in samples (R sample) are expressed as per mil (‰) differences from the ratio of standard atmospheric N2 (R std) using delta notation (δ) as: δ15Nsample = [(R sample/R std) − 1] × 1,000. Elemental calibration was accomplished using acetanilide (10.4% N, 71.1% C), and measurements were verified using a standard wheat flour (1.85% N, 40.2% C, δ15N = 2.55‰; Iso-Analytical, Cheshire, UK) and (NH4)2SO4 (IAEA-N1, δ15N = 0.4‰; IAEA-N11, δ15N = 3.5‰). Analytical precision for standards during this period was less than ±0.3‰.

Calculation of δ15N in released NH +4 and organic-N

Organic-N was calculated as the simple difference between TN (as TKN) and NH4–N. During the bottle mineralization study it was assumed that there were no N losses (either through denitrification or NH3 volatilization), therefore, net quantities of N released (as NH +4 and organic-N) during the bottle mineralization study were then determined by subtracting the initial amounts of NH4-N or organic-N in water extracts from those present after the 2-week incubation. The δ15N of organic-N was then calculated using values of δ15N of NH +4 and TN, and the elemental concentrations of NH4-N, organic-N and TKN measured in the extracts as follows:

$$ (\delta ^{{15}} {\text{N}})_{{{\text{Org - N}}}} = \frac{{({\text{N}} \times \delta ^{{15}} {\text{N}})_{{{\text{TN}}}} - ({\text{N}} \times \delta ^{{15}} {\text{N}})_{{{\text{NH}}^{{\text{ + }}}_{{\text{4}}} }} }} {{({\text{N}})_{{{\text{Org - N}}}} }} $$

In this calculation, it is important to note that the δ15N of neither NH +4 nor organic-N reflects the isotopic compositions of actual compounds released or mineralized during the study, but reflects the net result of both mineralization, release, and immobilization processes.

Statistical analyses

All statistical tests were conducted using Statistica 6.1 (Statsoft, Tulsa, Okla.). The effect of long-term P addition on soil δ15N in the mesocosm experiment was evaluated using a two-factor ANOVA with main effects of mesocosm treatment [open-control, closed-control, P-loaded (12.8 g P m−2 year−1)], and time (1995, 1996, 1997, 2000). Resulting δ15N of N forms mineralized in the laboratory bottle incubation was assessed using the general linear model multiple ANOVA procedure with fixed effects of soil type (site 1 and site 10), P addition level (0, 3.5, 7.0 mg P), applying the full factorial interaction. All post hoc comparisons of means for significant effects and interaction (α = 0.05) were accomplished using Fisher’s least significant difference test.

Results and discussion

Spatial and historical patterns

It has previously been demonstrated by Inglett and Reddy (2006) that plants of WCA-2A respond to the inputs of agricultural drainage waters with increasing δ15N. Surface litter material and soils on the WCA-2A transect also show increased δ15N (Fig. 2) with values ∼4‰ higher near the inflow sources relative to the interior marsh. Values were also nearly identical for both litter and surface soil at all transect sites (Fig. 2). In peat systems, soils form through accretion of macrophyte biomass, therefore, the similarity between litter and soil is not surprising. However, this contrasts with other studies showing gradual alteration of original plant δ15N during diagenesis (Fogel and Tuross 1999; Melillo et al. 1989). In this case, the similarity between the litter and soil at all WCA-2A transect sites indicates that the immobilization and mineralization processes affecting δ15N are rapid and largely completed during the early phases of litter diagenesis (0.5–1 year). For this reason, the δ15N of peat could serve as an accurate record of conditions at the time of litterfall, and thus, is also a sensitive temporal indicator of nutrient enrichment.

Fig. 2
figure 2

Patterns of δ15N of surface litter materials and soils (0–10 cm) along the nutrient gradient of water conservation area 2A

Full size image

Analysis of the sectioned cores displays the peat isotopic record in WCA-2A (Fig. 3). In the case of the non-impacted core, it is clear that the peat of WCA-2A is very stable throughout the past century. In other studies of wetland peat it has been observed that there is a steady maturation of peat with increasing age, and this reflects a steady enrichment of δ15N with depth in the core profile (e.g., Novak et al. 1999). Such downcore increases in δ15N are presumed to result from N loss processes (N mineralization, denitrification, etc.); however, the results presented in Fig. 3, demonstrate that like the elemental C and N composition, the δ15N of bulk peat N remains remarkably constant (at ∼2.5‰) throughout the profile of the non-impacted WCA-2A core (8.3 km). Currently, there is no explanation for this lack of change with depth in these Everglades soils; however, the constant δ15N in WCA-2A could represent a general lack of microbial activity (and N cycling) as a result of P limitation in these soils.

Fig. 3
figure 3

Vertical patterns of C N, P, δ15N, and dominant pollen composition of peat soils from impacted (1.8 km from inflows), moderately impacted (3.5 km), and non-impacted (8.1 km) WCA-2A sites. The 1964 date (arrow) was determined using the 137Cs peak method (Reddy et al. 1993) and is shown for the impacted core only. **Data for percent of pollen are summarized from Bartow et al. (1996) for a core taken at 1.4 km from the WCA-2A inflow. TC Total C, TP total P, TN total N

Full size image

In contrast to the non-impacted site, the isotopic profile of the 1.9 km site clearly shows the effect of nutrient enrichment in a manner similar to the isotopic changes of the transect surface components (Fig. 3). These changes included a gradual enrichment of P beginning about 13 cm in the profile, and concurrent with the P increase, a gradual alteration of the δ15N peat signature toward the surface. Dating of the core sections using the 137Cs profile method (Reddy et al. 1993) demonstrated that P enrichment and the changes in δ15N began to occur around or shortly after 1964. In a historical context, records indicate that discharges into WCA-2A began in the late 1950s, and 1960–1963 marked the completion of the WCA-2A inflow control structures just north of the impacted site of this study (Bartow et al. 1996; Light and Dineen 1994).

Potential mechanisms of δ15N increase

Many mechanisms are known which can increase or decrease system δ15N and therefore could be used to explain the downcore δ15N patterns. Perhaps the most obvious of these is the presence of an isotopically enriched N source on the WCA-2A drainage water inflows. This possibility was discussed by Inglett and Reddy (2006) with the conclusion that the δ15N of inflow N (4‰) was only slightly higher than that of the ecosystem components of the WCA-2A interior. For this reason, Inglett and Reddy (2006) concluded that this mechanism alone could not be used to explain the increased δ15N of the live macrophytes.

Alternatively, increased utilization of N derived from N2 fixation (having a δ15N = 0‰) may explain lowered δ15N values. Isotopic composition of N2-fixing algal biomass has been shown to decrease with increases in nitrogenase activity (indicated by C2H2 reduction) (Inglett et al. 2004; Rejmakova et al. 2004). In the Everglades, this mechanism could be quite significant due to both an abundance of N2-fixing cyanobacterial periphyton and overall high rates of nitrogenase activity in many of these systems (Inglett et al. 2004). Despite this potential effect, however, it has also been shown that P loading increases overall rates of nitrogenase activity in both the periphyton and detrital ecosystem components of WCA-2A (Inglett et al. 2004). Increased nitrogenase activity should result in lowered δ15N in P-impacted Everglades areas, therefore it is unclear how this mechanism can be used to explain the increased δ15N in P-affected WCA-2A soils.

Other mechanisms affecting δ15N include the enhancement of previously discussed microbial processes (N mineralization, denitrification, etc.). In this manner, P additions may lead to increased microbial activity, N mineralization, and subsequent loss of isotopically light N from the system. Increased respiration, N mineralization, and rates of potential denitrification have been noted in Everglades soils amended with P (Amador and Jones 1993; Newman et al. 2004; White and Reddy 2000, 2003). If sufficiently increased, these processes are known to discriminate against 15N, and thus, N loss following P enrichment of Everglades soils could increase the δ15N of the residual soil N. A strong correlation between soil δ15N and total P (TP) lends support to this hypothesis (Fig. 4), indicating that δ15N becomes enriched at a rate proportional to the amount of P in the system.

Fig. 4
figure 4

Correlation of δ15N with total P content in peat core samples from the impacted (1.8 km from inflows), moderately impacted (3.5 km), and non-impacted (8.1 km) WCA-2A sites

Full size image

Effect of P on δ15N of soil N mineralization products

If NH4- and organic-N released following P addition is sufficiently light isotopically, the residual soil N pool may become measurably enriched in 15N. A direct measurement of water-extractable N predicts that available NH +4 present in the low P soils of WCA-2A (Table 1, site 10) maintains a δ15N signature approximately 7‰ higher than the bulk soil N, while that of water-extractable TN is approximately 1.5‰ higher. If these N forms were directly lost through flux from the soil, the resulting soil would actually become more depleted. For this reason, it is unlikely that loss of this N source would result in increased soil δ15N. Alternatively, it can also be assumed that P additions resulting in increased soil microbial biomass, activity, and net N mineralization (White and Reddy 2000) could significantly impact the δ15N of available N forms.

Table 1 Selected physical, nutrient, and isotopic characteristics (mean ± SD, n = 3) of soils used in bottle mineralization study. wt. Weight
Full size table

Results of the bottle mineralization study demonstrate the effects of added P on N mineralization processes in the contrasting soils from sites 1 and 10 (Fig. 5). In the highly P-enriched soils of site 1, P addition at the 7.0 mg level reduced the amount of total water-extractable N (Fig. 5). The opposite effect of added P was observed in the soils of site 10, with increased amounts of total water-extractable N in both the 3.5 and 7.0 mg P treatments relative to control after the 2-week incubation (Fig. 5). The increased production of total water-extractable N in P amended soils of site 10 was not the result of increases in NH +4 which showed either no change in the 3.5 mg P treatment or decreased levels in the 7.0 mg P treatment relative to the non-amended control (Fig. 5). In contrast, water-extractable organic-N was higher than that of the control by 45 and 58% in the site 10 soils amended with 3.5 and 7.0 mg P, respectively (Fig. 5).

Fig. 5
figure 5

Net amounts (μg N g dry weight soil−1) and stable N isotopic signature (15N) of N forms released from nutrient-impacted (Site 1) and non-impacted (Site 10) WCA-2A soils following 15-day bottle incubations with 0 mg (control), 3.5 mg, and 7.0 mg of added P. Points represent the means ± SD of three replicate bottles

Full size image

The δ15N of the net mineralized NH +4 ranged between 7.0 and 7.7‰ and was not significantly affected by either site or added P level (Fig. 5). There was also no effect of added P on the δ15N of organic N compounds released during the 2-week bottle incubation of the site 1 soils. In contrast, there was a progressive effect of added P on the δ15N of water-extractable organic N in the soils of site 10, with δ15N increasing relative to the control by 1.8 and 3.0‰ in the 3.5 and 7.0 mg P treatments, respectively.

Based on the measurements of water-extractable N before and after the bottle incubation, it is clear that extractable organic N is more isotopically depleted than NH +4 at both WCA-2A sites. At both sites 1 and 10, extractable organic N was more isotopically depleted than the bulk soil N by ∼1 to 2‰ (Fig. 5). It is also evident, that addition of P to the low-P soils of site 10 resulted in increased release of these organic N compounds. Continued loss of compounds with δ15N lower than the bulk soil could result in increased δ15N of the residual site 10 soil; however, the large quantity of such compounds required to enact a measurable change on the total soil N argues against this possibility. Moreover, the addition of P appeared to reverse the process in the site 10 soils by increasing the δ15N of these extractable organic N to levels even above δ15N of the bulk soil N.

The consistency in the δ15N of NH +4 regardless of site or P level and effect of P to increase the δ15N of organic-N released from site 10 soils, indicate that soil mineralization may not affect the δ15N of soil through a loss of isotopically light N. However, the bottle study may not have been an adequate test for this mechanism. For example, the negative effect of P on N mineralization in the 7.0 mg P treatment indicates that significant N immobilization was present. In this case, any mineralized NH +4 would have been quickly reassimilated by the greatly increased microbial population at that high P level. Undoubtedly, the high production of NH +4 in the Newman et al. (2004) study occurred at much lower added P levels than even those of the 3.5 mg P treatment in the bottle incubation. For this reason, it is possible that both the 3.5 and 7.0 mg P treatments resulted in a higher N demand than that experienced during the mesocosm dosing study, and thus, the increased δ15N of extractable organic N could represent the incorporation of high δ15N NH +4 .

It could also be asserted that another problem in the bottle incubation was inhibition of nitrification and denitrification. Under field conditions, the process of NH +4 production is not isolated, and it may also be possible that O2 production in the water column is sufficient to allow nitrification to proceed into the surface soil. Nitrification and denitrification are both highly fractionating processes (Mariotti et al. 1981, 1988) and if present, should result in significant 15N increases in the soil NH +4 pool. The maintenance of anaerobic conditions in this bottle technique would have prevented nitrification, and thus, did not allow sufficient enrichment in 15N to occur. It is therefore possible that N mineralization, coupled with nitrification and denitrification could result in the overall enrichment of surface and/or porewater N under field conditions.

Long-term P-dosing effects on soil 15N

The test of potential δ15N changes as a result of P additions under field conditions lies in the isotopic analysis of samples archived from the Newman et al (2004) mesocosm dosing experiment. Because that experiment was conducted using P additions in a field setting, processes such as hydrologic transport, or coupled nitrification/denitrification would not have been excessively hindered. The experiment of Newman et al. (2004) resulted in significantly increased soil TP from ∼400 to greater than 600 mg P kg−1 soil in the highest dosing rate (12.8 g P m−2 year−1) (Table 2). The increased levels of soil TP were detected after only 1 year of dosing; however, within the first few weeks after dosing, measurements of surface and porewaters indicated a significant production of NH +4 with levels up to 5 mg N l−1 higher than controls (Newman et al. 2004). There were strong gradients in concentration between the pore and floodwater suggesting a significant increase in overall N flux rates from the soil.

Table 2 Values of selected nutrient parameters (total N, total C) and C:N ratios observed in soil samples (0–3 cm) from the WCA-2A long-term mesocosm P-dosing experiment (Newman et al. 2004). Values represent the mean ± SD of three replicate mesocosms except where noted
Full size table

Despite the evidence of significant enhancement of N cycling following P loading, analysis of soil samples from the experiment of Newman et al. (2004) reveal no significant change in the overall δ15N of bulk N in the 0–3 cm soil samples from the highest P-loaded mesocosms even after 5 years of loading (Fig. 6). The slight fluctuation which was measured in these soils was small (<2‰) and was consistent between dosed and control mesocosms indicating that seasonal or environmental conditions, or sampling protocols, had a greater effect on δ15N in this system.

Fig. 6
figure 6

Patterns of stable N isotopes (δ15N) of natural WCA-2A soils and those dosed with P (12 g P m−2 year−1) in the mesocosm study by Newman et al. (2004). Points represent the mean ± SD of samples obtained from three replicate mesocosms

Full size image

P dosing effects on macrophyte 15N

Macrophyte δ15N signatures may be passed down to the soil via decomposition of litter. To test this macrophyte control of soil δ15N, samples of leaf and root tissues from the hydroponic growth experiment of Lorenzen et al. (2001) were analyzed for δ15N. Because the experiment provided a constant supply of N at varying levels of P, the magnitude of potential reduction in plant discrimination during N uptake could be assessed. The results depicted in Fig. 7 demonstrate that as P level (and N demand) increases in these macrophytes, δ15N also increases. As stated previously, increased δ15N is likely the result of reduced discrimination during N uptake and assimilation by which plant N becomes more isotopically similar to the δ15N of the N source (Evans 2001). By this mechanism, Cladium and Typha grown at high P levels increased the δ15N of leaves and roots by as much as 12‰ relative to those grown at low P levels.

Fig. 7
figure 7

Stable N isotopic ratios (δ15N) of native WCA-2A macrophytes (Typha and Cladium) grown in solution culture of low (10 μg l−1) and high (100 μg l−1) ambient P concentrations in the experiment by Lorenzen et al. (2001). *P < 0.05 (significantly different means; Fisher’s LSD test)

Full size image

Similar shifts in δ15N have been seen in macrophytes of other wetlands including mangroves (Mckee et al. 2002) and New Zealand bog species (Clarkson et al. 2005). In these cases, the increased macrophyte δ15N occurred over a corresponding shift from P to N limitation similar to that in WCA-2A. McKee et al. (2002) also found that P addition increased the δ15N of mangroves growing in P-limited sites and postulated that the increased δ15N was not the result of a change in the δ15N of the N source, but rather reflected a decreased fractionation by the mangroves as N demand increased. Clarkson et al. (2005) further enhanced this mechanistic model based on plant N demand by showing that mycorrhizal status was not responsible for observed δ15N patterns of their sampled bog species.

The 12‰ increase in the δ15N of plants grown at high P levels in this study is similar to the 13‰ increases observed by Clarkson et al. (2005), and is of a sufficient magnitude to explain the 8‰ enrichment of Typha in areas near the WCA-2A inflows (Inglett and Reddy 2006). Unfortunately, no samples were available from the intermediate P levels in the Lorenzen et al. (2001) hydroponic experiment, therefore, it is unclear if there are differences between δ15N of Cladium and Typha at conditions of moderate N demand similar to those observed on the WCA-2A transect by Inglett and Reddy (2006). Also, we were unable to determine the isotopic composition of the N source [(NH4)2SO4] used in the Lorenzen et al. (2001) study, therefore it is impossible to determine the exact isotopic discrimination by the two plants at high and low P levels. The consistency between Cladium and Typha in the high-P treatment may indicate the N source used in the Lorenzen et al. (2001) study was approximately 12‰. Using the measured δ15N of water-extractable NH +4 at the impacted site 1 (Table 1), it is then reasonable to assume that at a maximum reduction in physiological discrimination, the biomass δ15N of Typha and Cladium near the WCA-2A inflows would be approximately 7–8‰. This value is close to the 8‰ value recorded for δ15N of live Typha leaves measured near the inflows (Inglett and Reddy 2006).

Conclusion

Many studies attempt to use δ15N as an indicator of N cycling to identify things such as varying N sources, N pollution, or a relative increase/decrease in processes such as N2 fixation or denitrification. In this study, live macrophytes, litter, and soils of a peat-based wetland system increased in δ15N in response to P eutrophication. The observed changes in δ15N were not explained by the more accepted mechanisms such as N pollution, lowered rates of N2 fixation, or increased denitrification. For this reason, two potential mechanisms explaining P-induced changes in macrophyte and soil δ15N were tested. The analysis of plants grown under constant N supply and varying P levels demonstrated that as P levels increase, the δ15N of Everglades Cladium and Typha also increases (up to 12‰), likely as a result of the plants increased demand for N (Lorenzen et al. 2001). This reduction of isotopic discrimination during N uptake is consistent with general isotopic theory and may justify the biomass δ15N observed for Typha (8‰) and Cladium (6‰) near the WCA-2A inflows.

In contrast to results for the macrophyte process, direct P addition to WCA-2A soils resulted in contradictory conclusions in bottle mineralization assays, and no observable change in soil δ15N in a long-term field-level P dosing experiment. Based on these results, it is possible to conclude that δ15N changes observed in the WCA-2A peat soils are derived from macrophyte isotopic changes which directly result from elevated P levels in the system. Despite the plausibility of the results, however, it is still important to view these results with caution as additional processes affecting macrophyte δ15N (e.g., mycorrhizal status and potential fractionating effects) remain untested in the WCA-2A system.

Combined with results of other studies such as McKee et al. (2002) and Clarkson et al. (2005), these results from the Everglades seem to indicate that macrophyte physiological processes can be a dominant process in determining soil and ecosystem δ15N, especially in cases involving shifts in N limitation. This finding has important implications concerning N cycling in similar wetlands where P loading impacts the overall storage of N within the system. In this, δ15N demonstrates that the majority of N taken up by plants is likely retained within the plant litter, and what is lost via leaching, mineralization, etc. is likely reincorporated into litter as it undergoes decomposition. The availability of P (and potentially other plant-limiting nutrients) is of key importance during this process as a determinant of the demand for N by macrophytes regulating production of biomass and microbial communities regulating N retention during peat accretion.

The results of this study also serve as another demonstration of the caution which must be exercised when interpreting observed isotopic signals. Reliance upon a single, predetermined process to explain δ15N patterns should be avoided unless experimental validation can be accomplished and the effects of competing processes eliminated. In this work, the role of nutrient limitation was particularly important as P availability affected the δ15N of ecosystem components. The resulting ∼12‰ effect which macrophytes can exert on the overall δ15N of a system is significant when compared with the typical ranges observed in global δ15N patterns (Amundson et al. 2003). For this reason, other studies should also consider the potential role of non-nitrogenous nutrients (such as P and K) in affecting N demand and thus altering δ15N through organic matter production.