SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 16 May 2007
Published in Soil Sci Soc Am J 71:1058-1065 (2007)
DOI: 10.2136/sssaj2005.00217
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bohlen, P. J.
Right arrow Articles by Gathumbi, S. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bohlen, P. J.
Right arrow Articles by Gathumbi, S. M.
Agricola
Right arrow Articles by Bohlen, P. J.
Right arrow Articles by Gathumbi, S. M.
Related Collections
Right arrow Wetlands and Aquatic Processes
Right arrow Nitrogen
Right arrow Microbial Processes

WETLAND SOILS

Nitrogen Cycling in Seasonal Wetlands in Subtropical Cattle Pastures

Patrick J. Bohlen* and Stanley M. Gathumbi

MacArthur Agro-ecology Research Center, Archbold Biological Station, Lake Placid, FL 33852

* Corresponding author (pbohlen{at}archbold-station.org).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Understanding the effects of agricultural land use on wetland N cycling is critical in areas such as south-central Florida, where widespread agricultural activities intersect with extensive wetland systems. We examined annual net N mineralization (Nmin) (buried-core method) and denitrification rates (acetylene-block method) in 24 small seasonal wetlands on a cattle ranch in this region, 12 each in intensively managed improved pastures (IP) and less intensively managed semi-native pastures (SNP). Wetlands in IP had less detritus, higher N concentrations, lower C/N ratios (0–15 cm), and higher microbial biomass N than did wetlands in SNP. Cumulative annual net Nmin was lower in IP wetlands (98 ± 17 kg N ha–1) than in SNP wetlands (133 ± 18 kg N ha–1). Nitrification was much lower in IP than in SNP wetlands and dominated net Nmin during the dry season (December–June), but was negligible during the flooded period (July–October). Cumulative annual denitrification was lower in IP wetlands (17.7 ± 3.4 kg N ha–1) than in SNP wetlands (34.7 ± 6.3 kg N ha–1). Soil N cycling rates correlated with NO3– and NH4+ concentrations, which correlated with soil C content. Our results show that the more intensive management of improved pastures was associated with declines in wetland soil C content and lower rates of nitrification and denitrification.

Abbreviations: IP, improved pasture • OM, organic matter • SNP, semi-native pasture


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Freshwater wetlands are important sites for transformations of N and other nutrients (Reddy and Patrick, 1975; Bowden, 1987; Johnston, 1991). Nitrogen transformations in wetlands are complex, in part due to the multiple oxidative states of the N molecule (Vymazal, 1995). The major transformations include mineralization of organic N, nitrification and denitrification, NH4 volatilization, and N2 fixation (Reddy and Patrick, 1984). Mineralization of organic N in sediments provides the major source of N to wetland plants and is responsible for the high rates of productivity of many wetland ecosystems.

Understanding factors that influence N mineralization and denitrification in wetlands is critical in agricultural landscapes where wetlands have the potential to buffer losses of N from upland to downstream ecosystems (Whigham and Simpson, 1976; Nixon and Lee, 1986; Johnston, 1991; Mitsch and Gosselink, 1993). Although natural wetlands have the capacity to remove nutrients in the short term, it is unclear how they will function in the long term because of the possibility that they may become nutrient saturated and serve as net sources rather than sinks of nutrients (Knight et al., 1987; Mitsch and Gosselink, 1993). To obtain a better understanding of the function of wetlands in agricultural landscapes, it is necessary to investigate key nutrient cycling processes in these wetlands and to evaluate their response to different management practices and changes in environmental conditions.

Livestock grazing is a major land use surrounding many critical wetland habitats, but information is lacking on the effects of grazing on wetlands or eutrophication of surface and groundwater (Carpenter et al., 1998). A critical region for addressing wetland nutrient dynamics in relation to grazing is South Central Florida, where extensive wetlands systems are interspersed in a landscape dominated by cattle ranchlands in the watershed north of Lake Okeechobee, which is the headwaters for the Florida Everglades (Hiscock et al., 2003). The coverage of wetlands in the region has declined from 25 to 15% during the past several decades due to construction of a drainage network for flood control and the conversion of more land to improved beef cattle pasture (Steinman and Rosen, 2000; Steinman et al., 2001; Hiscock et al., 2003). Although the major concern in the region is over increasing P loads in surface runoff associated with these land use changes, N loads are also increasing in Lake Okeechobee, due mainly to increased N loads from agricultural areas (Kratzer and Brezonik, 2005). The abundant seasonal wetland in these landscapes may be important to regional N dynamics.

To increase our understanding of N cycling processes in these wetland systems, we investigated net N mineralization (Nmin) and denitrification rates in isolated wetlands in subtropical cattle pastures during a 1-yr cycle. The goals of this study were to: (i) characterize seasonal patterns and cumulative amounts of net N mineralization and denitrification in isolated wetlands in subtropical cattle pastures; (ii) compare these patterns in intensively managed improved pastures with less intensively managed semi-native pastures; and (iii) examine factors influencing soil characteristics and wetland N cycling processes within the context of surrounding land management.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Site Description and Experimental Design
The isolated freshwater wetlands were located in large experimental pastures at the MacArthur Agro-ecology Research Center (MAERC) at Buck Island Ranch, a 4250-ha cattle ranch operated by Archbold Biological Station near Lake Placid, FL (27°09'N, 81°11'W). The pastures were established in 1998 to investigate the effect of varying cattle stocking densities and pasture type on nutrient loads in surface runoff (Steinman et al., 2003; Capece et al., 2007; Swain et al., 2007). There were 16 experimental pastures, eight each of improved and semi-native pastures. The pastures were managed with a seasonal rotation typical of many ranches in South Florida in which cattle graze more intensively managed improved pastures during the summer and less intensively managed native or semi-native pastures in the winter. The 20.2-ha improved pastures consisted mainly of Bahia grass (Paspalum notatum Flugge, 82% relative cover) followed in abundance by carpetgrass, Axonopus fissifolius (Raddi) Kuhlm. (10% cover), Bermuda grass (Cynodon dactylon L. Pers.), bluestem (Andropogon spp.) and vaseygrass (Paspalum urvillei Steud.) (each <2% cover). Improved pastures were fertilized annually with approximately 56 kg N ha–1, typically applied in March, and were grazed during summer (May–October). These pastures were managed as intensive improved pastures at least since the early 1960s and received annual applications of N, P, and K fertilizer up until 1987, after which they received only N fertilizer (Swain et al., 2007). The 32.4-ha semi-native pastures consisted of a mixture of Bahia grass (42% relative cover) and native grasses such as carpetgrass (16% cover), broomsedge (Andropogon virginicum L., 14% cover), field paspalum (Paspalum laeve Michx.), bluestem, and others. These pastures have never been fertilized and were grazed during winter (November–April).

Twenty-four isolated depressional wetlands, 12 each in improved and semi-native pastures ranging in size from 0.09 to 0.39 ha, were selected for this study (Steinman et al., 2003; Gathumbi et al., 2005). Wetlands in the improved pastures were dominated by the weedy invasive herbaceous species, Juncus effusus L. (soft rush) and Polygonum spp., which are characteristic of more disturbed sites (Cohen et al., 2004). The wetlands in the semi-native pasture were dominated by maidencane (Panicum hemitomon Schult.), with the emergent macrophyte Sagitaria lancifolia L. present in the deeper water areas (Steinman et al., 2003). Improved pasture wetlands also had significant bare ground, whereas the semi-native pasture wetlands had none.

Wetlands occurring in these nearly level (<1% slope) pastures are seasonally flooded during the wet season (June–October). These wetlands receive some surface runoff from the adjacent pastures but they become inundated only when the groundwater table is high. Subsurface drainage is slow due to the poorly drained status of the soils and relatively impermeable subsurface soil horizons. A system of drainage ditches was constructed several decades ago to improve drainage of the sites, and improved pastures are more intensively drained than semi-native pastures. The elevation of the pastures ranges from about 7.9 to 8.5 m above mean sea level and the improved pastures are at a slightly lower elevation (10–15 cm) than the semi-native pastures.

Soils consist of poorly drained fine sands, with parent material consisting of beds of sandy and clayey layers that were transported and deposited by ocean currents during several periods of inundation during the Pleistocene period, forming a series of marine terraces. Soils in the improved pastures are dominated by Felda fine sand (loamy, siliceous, superactive, hyperthermic Arenic Endoaqualfs), and in the semi-native pasture are dominated by Pineda fine sand (loamy, siliceous, active, hyperthermic Arenic Glossaqualfs) with a variable muck layer. Depressional marshes in both pasture types are classified as Tequesta muck (coarse-loamy, siliceous, active, hyperthermic Histic Glossaqualfs). Detailed descriptions of these soil types are given in Gathumbi et al. (2005).

Annual rainfall at the site averages ~1320 mm, with 75% occurring from June–October. Wetland hydroperiods (cumulative days of inundation) are variable (2–10 mo) depending on rainfall and are probably shorter than historical averages due to regionally imposed drainage. The study period included the driest year on record for the region, with a total annual rainfall of only 740 mm in 2000 and a more normal amount of 1300 mm in 2001. Consequently, the wetland hydroperiods were much shorter in 2000 than in 2001 (Fig. 1). Hydroperiods in the wetlands were determined manually by measuring water depth to the nearest centimeter at 45-cm rebar posts driven to ground level near the deepest part of each wetland.


Figure 1
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 1. Average water depth and hydroperiod in wetlands in improved (solid line) and semi-native (dotted line) pastures from July 2000 through March 2002.

 
Sampling Soil to Estimate Nutrient Storage
Soil was sampled at four different depths (detrital layer, 0–15, 15–30, and 30–45 cm) along three, randomly selected transects running from the center to the edge of each wetland. Sampling locations within each transect included one sample taken within a 5-m radius from the center and one sample within 5 m of the edge. Soils taken from the same depth and relative sampling locations from the three transects were pooled together, resulting in one sample from each depth for the center and edge of each wetland (one pooled sample x two locations x four sampling depths = eight samples per wetland). The detrital layer was sampled using a pin block method (Johnson et al., 1991). The pin block was a 15- by 15-cm wooden template with holes in each corner, which was placed on top of the detritus layer and held in place with large (15-cm-long) nails pressed into the soil through the corner of the block. The detrital layer (Oe and Oa horizons) was cut with a serrated knife around the edge of the block and then was undercut at the depth of the mineral soil and removed. The mineral soil beneath the detritus was sampled in three 15-cm increments using a 5-cm diam. soil corer. Bulk density of the mineral soil was determined on a separate set of adjacent 5-cm-diam. cores collected in the same 15-cm increments.

All detrital and soil samples were oven dried (60°C for detritus, 105°C for mineral soil), sieved (4-mm mesh), subsampled, and finely ground for C and N analysis. Separate subsamples were reserved for P fractionation (not reported here) and microbial biomass assays (see below). Only data from the detrital and 0- to 15-cm mineral soil layers are presented to provide estimates of total soil C and N in the surface soil layers where N mineralization and denitrification were measured. More detailed analysis of soil nutrient pools in the wetland sediments are given in Gathumbi et al. (2005).

Soil Sample Analyses
Soil total C and N were analyzed by dry combustion chromatography using a Carlo-Erba NA 1500 C-N analyzer (Haake Buchler Instruments, Saddle Brooke, NJ). Soil microbial biomass C and N were determined using a modified fumigation–incubation method using dry soils rewetted to a consistent water-filled pore space (Franzluebbers et al., 1996). Each dry soil sample was weighed into duplicate beakers and rewetted by adding deionized water to achieve 50% water-filled pore space. The amount of water ({theta}m) added per sample was calculated using a formula modified from Gardner (1986). All the samples were placed into 1.5-L Mason jars and preincubated in the dark at 20 to 25°C for 10 d in the presence of NaOH, to absorb respired CO2, and ~10 mL of deionized water in a scintillation vial to maintain the soil moisture. At 10 d, the NaOH traps were changed and one set of samples from IP and SNP plots were fumigated with ethanol-free chloroform at 20 to 25°C for 24 h. Chloroform vapor was removed by repeated evacuation with a vacuum pump and the samples were inoculated with 0.2 g of unfumigated soil retrieved from the original sample. All samples were reincubated for another 10 d and CO2 evolution measured as before. The CO2 production was determined by HCl titration of NaOH after incubation. Microbial biomass C (MBC) was calculated as follows:

Formula
where CF = CO2–C of the fumigated soil and Cc = CO2–C of the control soil. The constants 1.73 and 0.56 were used to account for the released microbial C adsorbed by soil during fumigation (Horwath et al., 1996). Initial and final soil inorganic N (NO3–N + NH4–N) was determined by extracting 10 g of soil sample with 40 mL of 2 M KCl before and after incubation. Soil C and N concentrations were converted to total nutrients per area by multiplying the nutrient concentration data by the bulk density data for detritus and mineral soil and applying the appropriate conversion for the unit sampling area. Total soil P was determined by ashing 0.5 g of soil at 450°C for 16 h and extracting the ash using aqua regia extractant (3:1 mixture of concentrated HCl/HNO3). The extract was analyzed colorimetrically using the method of Murphy and Riley (1962) on a Technicon Autoanalyzer II (Method 365.1, USEPA, 1993).

In Situ Nitrogen Mineralization Assays
Net N mineralization rates were measured using sequential incubations of buried cores in the field (Hart et al., 1994). Three locations were randomly selected within each wetland and, at each location, two 20-cm sections of polyvinyl chloride (PVC) pipe (5-cm i.d., with small holes drilled at the top to allow atmospheric venting) were gently pressed or hammered into the soil to a depth of 15 cm, leaving 5 cm extending above the soil surface. During dry periods, in which the wetlands were not inundated, the exposed pipes were capped with a PVC cap. During flooded periods the PVC-enclosed soil cores were removed from the soil, placed in water-filled plastic bags, and placed back into the hole from which they were removed. The inserted cores were allowed to incubate in the field for ~4 to 8 wk, depending on environmental conditions. At the start of each successive incubation period, two additional cores (5 by 15 cm) were taken from each location and pooled together to assess the initial concentrations of inorganic N. At the end of each incubation period, the two buried cores at each location were retrieved and combined into a single pooled sample (n = 3 for each wetland on each sample date). Net N mineralization and nitrification rates were determined as the difference between initial and final inorganic N concentrations (Hart et al., 1994).

Denitrification Assays
Denitrification rates were measured nine times between March 2001 and March 2002 using a modified version of the acetylene-block technique (Mosier and Klemedtsson, 1994; Horwath et al., 1998). There are limitations to the acetlyene-based static core method, but it has been used in a wide range of published studies, is relatively simple, and allows a large numbers of samples to be run simultaneously, which is necessary for ecosystem-scale studies (Groffman et al., 1999). On each sample date, a 2-cm-diameter hammer corer fitted with 15-cm-long Plexiglas inserts was used to take two, 10-cm-deep cores consisting of the detrital layer and mineral soil from three locations in each wetland. The cores were sealed immediately on the bottom with no. 9 natural rubber septa and returned to the lab and refrigerated overnight at 4°C. Cores were brought to room temperature, sealed at the top with a rubber septa, and allowed to equilibrate briefly (15 min) before venting the headspace of the cores to the atmosphere and then adding acetylene (to approximately 10 kPa) and mixing the headspace by repeated pumping with a 60-mL syringe. Gas samples from the headspace of the cores were taken after 2 and 6 h and placed in evacuated glass vials fitted with butyl rubber septa. Gas samples, blanks, and standards were analyzed for N2O by a gas chromatograph (Shimadzu Model 14A, Shimadzu Corp., Kyoto, Japan) equipped with an electron capture detector (ECD 63Ni) and a 3-m Poropak Q column. Denitrification rates were estimated from the N2O production between 2 and 6 h.

Statistical Analysis
Soil characteristics in wetland were analyzed with both univariate and multivariate statistical approaches. Except for the ordination analyses described below (PCA and NMS), all analyses were performed using JMP Version 6.0 statistical software (SAS Institute, 2005). Differences between wetlands in improved and semi-native pastures were analyzed with ANOVA. Correlation analysis was used to examine associations among N transformation rates and other soil characteristics. Additionally, soil characteristics of all 24 wetlands were analyzed with principle components analysis (PCA) using PC-ORD, Version 4 (McCune and Mefford, 1999). Soil variables used in the PCA analysis were relativized using a rank adjustment before analysis. The scores of the first principle component axis were treated as a simple univariate response in a nonparametric test (Wilcoxon test) for differences in soil characteristics between wetlands in improved and semi-native pastures, and between wetlands dominated by different plant communities.


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Soil Characteristics
Wetlands in improved pastures had a smaller detritus layer (mean = 6.6 ± 1.05 kg m–2) than did wetlands in semi-native pastures (mean = 10.3 ± 0.72 kg m–2). The concentration of total N in the detritus layer was greater in improved pasture wetlands than in semi-native pasture wetlands, and the C/N ratio of both the detritus layer and upper mineral soil was lower in improved pasture wetlands (Table 1). The C concentration in the upper mineral soil was significantly higher in semi-native pasture wetlands than in improved pasture wetlands. There was significantly more total microbial C and N in the detritus of the semi-native wetlands than in improved pasture wetlands. In the upper mineral soil, however, improved pasture wetlands had slightly more microbial biomass C and significantly more microbial biomass N than semi-native pasture wetlands (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Carbon and N content, C/N ratio and microbial biomass C (MBC) and N (MBN) of the detrital layer and upper 15 cm of mineral soil in wetlands in improved pastures and semi-native pastures (n = 12). Data modified from Gathumbi et al. (2005).

 
The average soil moisture in the Nmin and denitrification cores was greater in the semi-native pastures (1.09 ± 0.27 and 1.30 ± 0.33 kg kg–1 dry wt., respectively) than in the improved pasture wetlands (0.76 ± 0.31 and 0.90 ± 0.35 kg kg–1 dry wt., respectively) (Student's t-tests, P < 0.05). Soil moisture was lower in the dry season (November–May) than in the wet season (June–October) and reached its lowest level in the Nmin cores in May 2000, after a period of prolonged drought (data not shown). Average organic matter (OM) content across all wetlands was correlated to soil moisture in both Nmin and denitrification samples (r2 = 0.41 and 0.72 for denitrification samples, and 0.60 and 0.74 for Nmin samples under flooded and nonflooded conditions, respectively). Organic matter content in both Nmin and denitrification samples was significantly greater in semi-native pasture wetlands (19.2 ± 4.2 and 23.8 ± 5.6%, respectively) than in improved pasture wetlands (15.7 ± 6.4 and 16.2 ± 6.1%) (Student's t-tests, P < 0.05). The OM content of the Nmin and denitrification samples was similar in samples from improved pasture wetlands, but in the semi-native pasture wetlands the OM content of the denitrification samples was significantly greater than in Nmin samples (one-way ANOVA, P = 0.035).

Nitrogen Mineralization and Inorganic Nitrogen Pools
Net Nmin rates were high in December 2000 following wetland draw down and declined during the course of the dry season, increasing again with the onset of inundation in July 2001 (Fig. 2). In the improved pasture wetlands, Nmin declined linearly during the dry season from a high of around 388 g N ha–1 d–1 at the beginning of the dry season (December 2000), to a low of 53 g ha–1 d–1 in early May 2001, and remained low until July 2001 when wetlands became inundated and rates increased to 526 g N ha–1 d–1. In the semi-native pasture wetlands, net Nmin declined from a high of 550 g N ha–1 d–1 at the beginning of the dry season to a low of 112 g ha–1 d–1 in June 2001, and increased after inundation in July 2001 to 375 g ha–1 d–1. Both net Nmin and cumulative N mineralized were much greater in the semi-native pasture wetlands than in the improved pasture wetlands during the latter half of the dry season (March–May) (Fig. 2). Cumulative annual net Nmin averaged 133 ± 18 kg N ha–1 in the semi-native pasture and 98 ± 17 kg N ha–1 in the improved pasture wetlands. Net nitrification rates were highest at the early part of the dry season (December–February), declined during the course of the dry season (March-June), and were lowest following inundation (July–October) (Fig. 2). Nitrification accounted for about 70% of the total net Nmin in both semi-native and improved pasture wetlands during the dry season (December–June), but did not contribute to net Nmin during the flooded period (July–October) (Fig. 2).


Figure 2
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 2. Net N mineralization rates (top panel) and in situ net nitrification rates (bottom panel) of wetland sediments in semi-native and improved pastures during six sequential buried-core incubations from October 2000 to October 2001. Values are means ± standard errors (n = 12). Asterisks denote statistically significant differences between values on a given date (P < 0.05).

 
Soil NO3 was much greater in the semi-native pasture wetlands than in the improved pasture wetlands during the dry season (December–June) (Fig. 3). Nitrate-N accounted for the majority of the soil inorganic N during the dry season, whereas NH4–N dominated the inorganic N pool following the onset of flooding in July 2001 (Fig. 3). In the semi-native pasture wetlands, NO3–N concentrations increased steadily throughout the dry season and were nearly an order of magnitude higher than in the improved pasture wetlands during the latter half of the dry season (March–June). Soil NH4–N concentrations were nearly the same in semi-native and improved pasture wetlands from October 2000 through March 2001 but were significantly greater in the semi-native pasture wetlands in May though July of 2001 (P < 0.002 on all dates).


Figure 3
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 3. Wetland sediment NO3– and NH4+ concentrations at the beginning of successive buried core incubations in wetlands in improved and semi-native pastures from October 2000 to July 2001.

 
Denitrification
Denitrification rates varied throughout the year but were significantly higher in wetlands in semi-native than in improved pastures on four of the nine sampling dates (Fig. 4). The largest difference between wetlands in the different pasture types occurred at the end of the dry season in June and early July of 2001, when denitrification rates were an order of magnitude greater in semi-native pasture wetlands than in improved pasture wetlands. In the weeks immediately following flooding in early July 2001, denitrification rates declined in semi-native pasture wetlands but increased slightly in the improved pasture wetlands. Rates increased sharply during a temporary draw down that occurred in all semi-native pasture wetlands and 6 of 12 improved pasture wetlands in late Aug. 2001 due to a period of low rainfall (Fig. 4). During that brief period, improved pasture wetlands that were drawn down (average water depth at center ~5 cm) had denitrification rates that were nearly 18 times greater (141.4 ± 59.4 g N ha–1 d–1) than in improved pasture wetlands that were not drawn down (average water depth at center ~30 cm) 7.9 ± 8.0 g N ha–1 d–1). The temporary draw down was followed by a period of prolonged inundation (2–3 mo), during which denitrification rates declined to their lowest levels. Seasonal draw down of the wetlands starting in late November 2001 stimulated an increase in denitrification rates in early December. In the 3 mo following draw down, as the wetlands began to dry out, denitrification rates decreased linearly in the improved pasture wetlands but did not decrease in the semi-native pasture wetlands (March 2002; Fig. 4). Cumulative annual denitrification was nearly two times greater in semi-native pasture wetlands (34.7 ± 6.3 kg N ha–1) than in improved pasture wetlands (17.7 ± 3.4 kg N ha–1), and was significantly greater in semi-native than in improved pasture wetlands on all dates after 6 June 2001 (P < 0.05). Across all 24 wetlands, soil OM content was positively related to mean cumulative denitrification (r2 = 0.35, P = 0.015).


Figure 4
View larger version (17K):
[in this window]
[in a new window]

  		Fig. 4. Denitrification rates in wetlands in semi-native pastures or improved pastures from March 2001 to March 2002. Values are means ± SE (n = 12). Arrows refer to varying hydrologic conditions in the wetlands. Statistically significant differences between values at *P < 0.05; ** statistically significant differences between values at P < 0.01; *** statistically significant differences between values at P < 0.001.

 
Multivariate Analysis of Nitrogen Cycling Rates and Soil Properties
The average net Nmin rates per wetland were positively correlated with nitrification rates and average NH4+ concentration measured at the beginning of the buried core incubations (Table 2). Average nitrification rates were correlated with NH4+ and also NO3– concentrations. Cumulative denitrification was negatively correlated with microbial biomass N, and positively correlated with soil NO3– concentration and average nitrification rates. Both NO3– and NH4+ concentration correlated positively with soil C content. There were no significant correlations between average N transformation rates and soil N or P content or soil C/N ratio in the upper 15 cm of mineral soil.


View this table:
[in this window]
[in a new window]

  		Table 2. Correlation matrix showing the Pearson product-moment correlation values for each pair of wetland soil characteristics in the 0- to 15-cm soil layer (%C, %N, %P, C/N ratio, microbial biomass C [MBC], microbial biomass N [MBN]), average NO3– and NH4+ of initial N mineralization cores, and cumulative N mineralization (Nmin), nitrification, and denitrification in each wetland (n=24).

 
Principle components analysis of soil characteristics showed that the first principle component explained 48.9% and the second principle component explained 15.4% of the cumulative variance (Fig. 5). The variables most strongly associated with the first axis were inorganic N concentrations, N transformation rates, and soil C content. The semi-native pasture wetlands separated clearly from the improved pasture wetlands along Axis 1, except for three improved pasture wetlands, two of which were the only improved pasture wetlands dominated by maidencane (mean = 93% cover) (Fig. 5). The other improved pasture wetland that overlapped with the semi-native pastures along Axis 1 had a greater percentage cover of maidencane (11.4%) than any of the other 10 improved pasture wetlands (mean = 1.6%, range = 0.0–6.7%). Nonparametric tests of PCA scores from the first principle component axis showed that scores for improved and semi-native pasture wetlands were significantly different (two-sample test, Z = –3.38, P = 0.0007). The soil PCA scores for the two improved pasture wetlands dominated by maidencane were significantly different from scores for those with little or no maidencane (two-sample test, Z = –1.83, P = 0.05).


Figure 5
View larger version (9K):
[in this window]
[in a new window]

  		Fig. 5. Results of principle components analysis based on 12 measured soil variables. Each symbol represents a single wetland and the symbol labels refer to wetlands in improved (I) or semi-native (N) pastures. The relative size of the symbols corresponds to average concentration of inorganic soil N measured at the beginning of buried core incubations, which was the variable most strongly associated with the first principle component (eigenvector = –0.72). Asterisks indicate the two improved pasture wetlands that were dominated by maidencane (average 93% cover).

 

   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Seasonal Patterns of Nitrogen Cycling Rates
One of the key controls on N cycling processes in seasonally flooded inland ecosystems is the alternation between terrestrial and flooded conditions (Patrick and Wyatt, 1964; Reddy and Patrick, 1975). Wetlands in our study exhibited both seasonal and interannual variation in flooding, and there was significant variation among wetlands in water depth and the occurrence of draw down events. The transition from terrestrial to flooded conditions had clear effects on N cycling processes, with a shift from a dominance of nitrification during the terrestrial phase when moisture was still adequate, to dominance of ammonification during prolonged flooded periods. Denitrification rates were greatest immediately following draw down events when wetland sediments were wet but exposed to aerobic conditions. The ability of nitrifiers in anoxic environments to become active rapidly when exposed to oxic conditions is an important adaptation of nitrifiers in environments with fluctuating oxic–anoxic conditions, and helps explain the rapid response of nitrifiers and denitrifiers to draw down events following prolonged periods of flooding (Bodelier et al., 1996). This effect was observed in September 2001 when denitrification rates increased sharply during a temporary draw down, and following seasonal draw down at the end of the wet season (Fig. 4).

Other studies have shown that NH4 is the main end product of N mineralization under waterlogged conditions and that nitrification increases as groundwater levels drop (Hefting et al., 2004). The highest denitrification rates tend to occur under conditions of intermediate waterlogging, when the presence of aerobic and anaerobic hot spots allow nitrification and denitrification to coexist (Groffman, 1994; McClain et al., 2003; Hefting et al., 2004). The water table level can override other controls on N cycling processes such as soil texture, geomorphic context, N input, or vegetation, and was clearly important in determining seasonal patterns of N cycling activity in our study.

Differences between Pasture Types in Seasonal Nitrogen Cycling Activity
Greater N cycling activity in the semi-native pasture wetlands compared with the improved pasture wetlands in the dry season may have been due to the greater amounts of plant litter, soil OM, and soil moisture, which apparently created ideal conditions for high net nitrification rates (Fig. 4). Soil moisture levels remained relatively high throughout the dry season in semi-native pasture wetlands, but dropped off steadily from March to June in the improved pasture wetlands (data not shown), where declining soil moisture probably limited denitrification and Nmin.

Greater annual denitrification flux from the semi-native pasture wetlands than from the improved pasture wetlands was due mainly to differences occurring in the dry season, when rates were highest, and not in the wet season, when flooded conditions inhibited NO3 production and denitrification. It is possible that the acetylene block method did not adequately measure denitrification rates at soil microsites such as the thin upper layer of soil or around the oxidized surface of plant roots (Reddy and Graetz, 1988); however, periods of prolonged flooding in fluctuating environments lead to decreasing redox conditions that decrease NO3 production, even around plant roots (Bodelier et al., 1996). Accordingly, the contribution of flooded periods to overall denitrification flux in our study was negligible relative to fluxes that occurred immediately after draw down, during the dry season when surface soils were still moist, or in recently flooded soils when NO3 was still available (Fig. 4).

Wetland Characteristics Affecting Nitrogen Transformation Rates
Soil characteristics that correlated best with overall N transformation rates in wetland soils were the concentration of inorganic soil N forms (Table 2). Soil inorganic N concentrations, N transformation rates, and soil C were all associated with the first axis in the principle components analysis of wetland soil properties, indicating the interrelatedness of these soil properties (Fig. 5). The positive correlation of soil NH4+ and NO3– concentrations with soil total C in the upper 15 cm of soil indicates that total soil C was a more important influence on soil N transformation rates than soil N content or the soil C/N ratio. Another indication of the importance of soil C is the positive relationship between the soil OM content and the average denitrification rate, a relationship that has been reported in other studies (Delaune et al., 1998; Davis et al., 2004). Thus, factors that influence the distribution and amount of soil C in these systems had an important influence on N transformation rates and probably explain much of the difference between improved and semi-native pasture wetlands.

Two factors that influence soil C in seasonal wetlands and that are susceptible to impact by ranch management activities are soil disturbance and vegetation type. Livestock grazing does not disturb the soil as much as tillage, but some studies have reported that grazing reduced soil C in grasslands (Bauer et al., 1987; Johnston et al., 1971). Other studies have reported that grazing increased soil C (Smoliak et al., 1972; Franzluebbers et al., 2000). The impacts of grazing and trampling by livestock on flooded soils is not as well studied as in upland systems, but are likely to be significant because of the large amounts of C in wetland soils and the high degree of physical disturbance and mixing that occurs under flooded conditions. The direct impacts of livestock on soil due to trampling, incorporation of plant litter, and other physical effects is complicated by longer term changes in the plant community, which can alter soil C by affecting total plant productivity or the quality or decomposability of plant C inputs (Frank et al., 1995; Wardle, 2002). Our study was not designed specifically to examine the role of soil disturbance or vegetation on soil OM or N cycling rates, but the importance of these two factors can be inferred from the difference in soil properties between wetlands in improved and semi-native pastures, and between wetlands with or without maidencane.

The major difference in the surface soil profile of wetlands in improved and semi-native pastures was the thicker layer of surface detritus and greater amounts of OM in the upper mineral soil of wetlands in semi-native pastures (Table 1). This OM not only provided a potential C source for heterotrophic microorganisms but also, in combination with a thick litter layer, enhanced retention of soil moisture during the dry season, buffering and enhancing N cycling activities. The thinner litter layer and lower soil OM content in improved pasture wetlands were associated with significantly altered seasonal patterns of N cycling activity, and with lower average N cycling activities relative to semi-native pasture wetlands, and were most likely due to the more intense ranching activities in the improved pastures. The more intensive ranching activities in improved pasture included higher long-term cattle stocking densities in improved pastures (~0.08 ha animal unit month [AUM]–1,) than in semi-native pastures (~0.15 ha AUM–1) in semi-native pastures. Differences in timing of grazing in improved and semi-native pastures also has consequences for soil disturbance and plant communities. Cattle graze improved pastures during the summer growing season, when they remove large amounts of actively growing plant tissues and can significantly alter plant community composition via selective grazing. Trampling impacts are greater in summer due to flooded soil conditions. By contrast, cattle graze semi-native pastures in the winter when they have less impact on removal of new growth and the composition of the plant community, and when drier soils are less susceptible to intense mixing due to trampling. The combined effect of more intense cattle trampling and greater plant biomass removal in wetlands of improved pastures could have contributed to reduced C sequestration in wetland soils.

Although we did not directly measure the influence of cattle on soil OM in this study, we examined their effects in a related study by putting 16-m2 grazing exclosures in five improved pasture wetlands in other areas of the ranch and measuring soil characteristics inside and outside the exclosures through time (P. Bohlen, personal communication, 2006). After 3 yr, the amount of soil OM in the upper 15 cm of mineral soil was significantly greater within the exclosures (6.7 ± 5.0 kg m2) than in adjacent paired plots outside the exclosures (2.5 ± 0.9 kg m2) (paired t-test, P = 0.05). This strong short-term response clearly indicates that cattle can significantly reduce OM content in seasonal wetlands and supports our conclusion that cattle activity contributed to the lower soil C content in improved pasture wetlands relative to semi-native pasture wetlands.

Another major difference between wetlands in the different pasture types was the greater dominance of maidencane in the semi-native pasture wetlands. Wetlands dominated by maidencane, which included all 12 semi-native pasture wetlands and two improved pasture wetlands, had significantly different soil characteristics than the 10 improved pasture wetlands that lacked maidencane. Cattle preferentially graze maidencane and can eliminate it from wetlands where grazing pressure is high throughout the growing season (Soil Conservation Service, 1987). In the grazing exclosure study described above, the cover of maidencane increased from an average of 2.1% cover outside the exclosures to 32.1% inside the exclosures after three growing seasons (cf. 2.4% maidencane cover in 10 of 12 improved pasture wetlands in the current study). The only two improved pasture wetlands dominated by maidencane in our study were in the control pastures that lacked cattle since 1996 (4–5 yr before this study). Thus, wetlands where cattle have the largest grazing and trampling impacts are the same wetlands where they have the greatest potential to alter plant communities. To tease apart the interrelated effects of grazing, trampling, and plant communities would require long-term field experiments with cattle grazing treatments applied to randomly selected wetlands in different pasture types.

In conclusion, seasonal fluctuation in water levels and differences between pasture types influenced soil N cycling processes and soil OM in seasonal wetlands embedded within the ranch landscape. Fluctuating water levels drove temporal patterns in N transformation rates, but the influence of pasture type on wetland N cycling activities was more complex. Our findings suggest that intensive summer grazing of wetlands in improved pastures reduced soil N cycling by lowering soil OM relative to less intensive winter grazing practices of wetlands in semi-native or unimproved areas. This effect occurred despite the higher N fertilizer inputs to improved pastures and lower soil C/N ratios in improved pasture wetlands.


   ACKNOWLEDGMENTS
 
This work was supported by funds from USDA-CSREES Competitive Grant no. 00-35101-9282. We thank Mike McMillian, Lourdes Rojas, Wilhelmina Tsang, Julie Golod, Nicola Clegg, Christy Edwards, and Anna Knipps for assistance in the field and laboratory work. This paper is contribution no. 101 to the MacArthur Agro-ecology Research Center.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication July 7, 2005.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bohlen, P. J.
Right arrow Articles by Gathumbi, S. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bohlen, P. J.
Right arrow Articles by Gathumbi, S. M.
Agricola
Right arrow Articles by Bohlen, P. J.
Right arrow Articles by Gathumbi, S. M.
Related Collections
Right arrow Wetlands and Aquatic Processes
Right arrow Nitrogen
Right arrow Microbial Processes

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome