| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Environmental Science and Engineering, C.B. #7400, Univ. of North Carolina, Chapel Hill, NC 27599-7400 USA
steve_whalen{at}unc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
10–25 µg N2O–N m-2 h-1) within a few days as soils drained. Treatments that did not include N (deionized water, glucose) showed no increase in N2O emission over unfertilized controls. Time-integrated N2O emission was significantly lower for plots amended with swine waste (8.5 mg N2O–N m-2) compared with plots comparably fertilized with NH4–N plus glucose-C (20.8 mg N2O–N m-2), suggesting that some component of the waste adversely affected the microbial N cycling community. The immediate increase in N2O emission following fertilization and accumulation of NO3–N without lag indicated that repeated fertilization throughout the growing season maintained active and responsive nitrifying and denitrifying communities. Percentage fertilizer loss to N2O to the point where fluxes had returned to prefertilization values was low, <1.0%. However, simulated rainfall gave pulsed N2O emission from denitrification of accumulated NO-3–N, indicating that further emissions will occur with an increase in soil moisture.
Abbreviations: WFPS, water-filled pore space • dw, dry weight • HSD, Honestly Significant Difference • NH4, NH4–N • C, glucose-C • NC, NH4–N plus glucose-C • W, liquid swine waste • CT, control • H2O, deionized water • NH4L, 300 mg NH4–N L-1 • NH4H, 600 mg NH4–N L-1 • NO3 L, 300 mg NO3–N L-1 • NO3H, 600 mg NO3–N L-1
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
Modern agriculture is characterized by an exponential increase in the use of N fertilizers (Vitousek et al., 1997). Increased rates of N2O evolution by N-fertilized soils is well documented in field and laboratory studies (e.g., summary by Sahrawat and Keeney, 1986). Thus, the accelerated application of N fertilizers in crop production is regarded as a major reason for enhanced N2O release from soils, and agriculture is presently estimated to contribute 90% of total anthropogenic N2O emissions (Duxbury, 1994).
Field and laboratory studies of N2O emission in agroecosystems focus largely on the effect of mineral N fertilizers. However, contemporary livestock production is characterized by confined growth facilities where collected animal wastes are land-applied as organic fertilizers to the extent that use of organic fertilizers probably exceeds the application of mineral fertilizers on a global basis (Bouwman et al., 1995). Hence, in a recent review of agricultural use of N fertilizers, Sims (1995) cites the urgency for additional data regarding the environmental impacts of organic fertilizers, including enhanced N2O efflux.
The swine production industry in the USA is experiencing a regional shift from the Midwest to the Southeast, where explosive growth of industrial scale swine production facilities has occurred in the past decade (Vansickle, 1997). Fundamental differences exist in regional waste management practices. Swine waste is both stored and land-applied as manures and slurries in the Midwest, while waste material is stored in anaerobic lagoons and the liquid phase is sprayed on soils in the Southeast.
Development of sound management policies that simultaneously conserve N, decrease the amount of N applied to effectively grow crops, and limit N2O emission rely in part on a detailed understanding of the controls and pattern of N2O emission and N cycling dynamics in fertilized soils. Data of this nature are currently lacking for southeastern swine production operations that commonly employ anaerobic lagoon–spray field waste management technology. Consequently, this investigation was aimed at evaluating environmental controls on N2O emission from a spray field at a regionally representative facility. Temporal patterns for N2O efflux and changes in soil physicochemical properties (NO3, NH4, temperature, moisture) were tracked in field plots amended with liquid lagoon effluent, individual chemical constituents of effluent (NH4–N, labile-C), NO3–N, or deionized water.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
|
Experimental
Two experiments were performed during April and early May 1998 while the spray field was planted to winter wheat. Both experiments involved fertilization of 0.7- x 0.7-m (0.5-m2) field plots with liquid swine waste, deionized water, or nutrients (inorganic N, glucose-C). Prior to each experiment, 24 plots were established linearly in groups of three along a transect within one of two adjoining 3- x 40-m sections (120 m2) of the spray field. Plots within a group were separated by 0.3 m. Groups of plots were separated by 1.5 m to ensure that any lateral movement of liquid fertilizer did not contaminate neighboring groups. Nutrient analysis of soil samples collected between plots indicated that <10% of the added fertilizer migrated laterally offsite and that neighboring plots were not affected. A polyvinyl chloride collar (20-cm diam x 11-cm height) was pressed 4 cm into the soil centrally within each 0.5-m2 plot. Nitrous oxide flux determinations (described below) were made within each collar. Groups of collars within a 120-m2 field section that showed no significant differences in mean pre-fertilization N2O emission (statistical methods described below) were selected for experimental manipulation (described below). A rainout shelter extended over the 120-m2 field section and the surrounding soil to exclude precipitation, such that observed changes in N2O emission and soil physicochemical properties would reflect the influence of fertilization alone. Shelters consisted of a permanent frame lattice and removable tarpaulins that were deployed only during actual or anticipated rainfall. The design allowed free circulation of air through the shelters at all times. Plots were successfully covered prior to and during every rain in both experiments.
The total volume of deionized water, swine waste, or liquid nutrient medium added to each fertilized plot in both experiments was 12.5 L, with 0.8 L added to the soil collar and 11.7 L distributed to the remainder of the plot. Liquid was added gradually over a 2-h period with hand-held watering cans to avoid ponding and to ensure ample time for infiltration. Sample cups (130 mL) were positioned along grids within the plot (exclusive of the flux collar) to ensure even application of liquid. Separate addition of liquid to the soil isolated by the collar ensured that soils used for flux determinations received exactly the target application. The total liquid volume applied and the within-plot partition scheme simulated a plot-wide, 2.5-cm application.
Experiment 1 was intended to determine what constituents or combination of constituents of swine waste (liquid alone, N, C) stimulated N2O emission and to evaluate the influence of each treatment on soil N cycling dynamics. Each of the three 0.5-m2 plots within a randomly selected cluster was amended with 2.5 cm of (i) liquid lagoon swine waste that had total N, NH4–N, and dissolved organic C concentrations of 660, 600, and 530 mg L-1, respectively; (ii) NH4Cl (660 mg N L-1); (iii) glucose (530 mg C L-1); (iv) NH4Cl (660 mg N L-1) plus glucose (530 mg C L-1); (v) deionized water; and (vi) no addition (control). Treatments (i) through (vi) were coded W, NH4, C, NC, H2O, and CT, respectively. Chemical constituents of synthesized media were consistent in concentration with the nutrient analysis of the liquid swine waste. The nutrient load to all N-fertilized plots was 165 kg N ha-1.
Experiment 2 was intended to (i) provide information concerning a dose response to fertilization, (ii) give added insight into the relationship between soil N cycling dynamics and N2O emission, and (iii) assess the impact of a simulated rain on N2O emission after fluxes had returned to prefertilization levels. Triplicate, randomly selected plots were amended with 2.5 cm of (i) NH4Cl (600 mg N L-1); (ii) NH4Cl (300 mg N L-1); (iii) KNO3 (600 mg N L-1); (iv) KNO3 (300 mg N L-1); and (v) no addition (control). Treatments (i) and (iii) were designated as ùhighú (150 kg N ha-1), while (ii) and (iv) were considered ùlowú (75 kg N ha-1). Mineral-based, liquid media were used for all treatments in this experiment because swine waste–N is largely NH4 and waste addition would also result in an unwanted organic-C amendment. Treatments (i) through (v) were designated as NH4H, NH4L, NO3H, NO3L, and CT, respectively.
Following nutrient addition, N2O flux determinations were made within 3 or 4 h, daily thereafter for 3 or 4 d and then less frequently (2- or 3-d intervals) until experiments were terminated at 8 or 11 d. The influence of increased soil moisture (simulated rainfall) on N2O emission was evaluated after the flux determination on Day 11 in Exp. 2 by adding 2.5 cm of deionized water as described above and reassessing N2O flux 3 h later.
Supporting physicochemical data were collected in conjunction with N2O flux measurements. A soil temperature profile was measured at a reference station with a multithermistor probe (3-cm intervals) and the mean soil temperature to 20 cm was calculated. A soil core (20-cm length x 2.5-cm diam) collected within each plot was homogenized and analyzed for NO3–N + NO2–N, NH4–N, and moisture content. Replicate
cores showed coefficients of variation (CV) of 11 and 7% when analyzed for NH4–N and NO3–N, respectively. Physicochemical measurements focused on the surface 20 cm soil because >90% of potential nitrifying and denitrifying activity were localized within this soil zone (Whalen et al., 1999).
Nitrous Oxide Determinations
Nitrous oxide flux measurements were made by the static chamber technique (Whalen and Reeburgh, 1988). Briefly, open-bottomed, cylindrical polyvinyl chloride covers (20-cm diam x 9-cm height) fitted with a butyl O-ring were inserted onto the permanent soil collars to isolate 0.031 m2 of soil surface and 5.3 L of overlying air. Covers were fitted with a capillary bleed to equalize pressure and an O-seal fitting (Swagelok Co., Solon, OH) equipped with a septum for syringe sampling.
Chamber headspace gases were syringe-sampled on cover emplacement and at 0.25-h intervals thereafter to 0.75 h. Samples were stored prior to analysis by inserting the hypodermic needles of the syringes into Butyl rubber stoppers. Tests showed no significant change in N2O concentration over the <24-h storage interval. Covers were removed from collars between sampling sessions.
Nitrous oxide was measured with a Shimadzu (Columbia, MD) GC-14A 63Ni electron capture detector gas chromatograph. Operating conditions and instrument calibration are described previously (Whalen et al., 1999). The precision of analysis expressed as the CV for 10 replicate injections of standards (301 and 8042 nL L-1 N2O) was <2%. Hourly, area-based N2O–N fluxes were calculated from the time-linear rate of concentration increase in the headspace during chamber deployment. These data were time-integrated using the trapezoidal rule to calculate area-based N2O–N emission over the entire observational period for each experiment.
Soil Physicochemical Measurements
All soil cores collected in conjunction with N2O flux measurements were homogenized and sieved (4-mm mesh). Soil nutrients were extracted (2 M KCl; 10:1 volume/soil wet weight) on a rotary shaker (1 h at 200 rpm). Soil NO3–N + NO2–N (hereafter referred to as NO3–N) and NH4–N concentrations in extracts were determined by the copperized cadmium reduction and the indophenol blue methods, respectively (Keeney and Nelson, 1982), following sample filtration through Whatman no. 42 filter paper. Total N and organic C were determined by dry combustion with a Carlo Erba (Milan, Italy) NA 1500 Elemental Analyzer. Soil moisture was measured gravimetrically (oven-dried at 105°C), and texture was assessed hydrometrically. Percentage water-filled pore space (% WFPS) was calculated as the ratio of volumetric soil water content to total soil porosity. Soil pH was measured potentiometrically on 1:2 soil-deionized water slurries equilibrated for 24 h.
Liquid lagoonal swine effluent was analyzed for NH4–N as described above and for total N by persulfate oxidation (Solorzano and Sharpe, 1980). Dissolved organic C was determined by high temperature catalytic oxidation with a Shimadzu (Columbia, MD) Model 5000 Total Organic Carbon Analyzer.
Statistical
The influence of treatment on N2O emission and soil nutrient concentrations in each experiment was evaluated by Single Factor Analysis of Variance. Multiple comparisons of treatment means was performed by Tukey's Honestly Significant Difference (HSD) Procedure. In both tests, data were log-transformed as necessary to satisfy assumptions of normality and homoscedasticity (Zar, 1984). Pearson correlation analysis was used to relate N2O flux and environmental variables. A significance level of
was used for all tests.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
|
|
|
|
and NC
treatments and between flux and soil NH4 for the NC treatment
. In addition, a significant, negative correlation was observed between N2O emission and soil NO3 for the H2O treatment
.
Experiment 2
Nitrogen fertilization elicited an immediate increase in N2O emission (Fig. 5)
, as in Exp. 1. Nitrous oxide fluxes for NO3H and NO3L treatments peaked at around 4100 µg N m-2 h-1 immediately after nutrient addition and steadily decreased thereafter to Day 11. The maximum observed fluxes represented about a 160-fold increase over prefertilization values of 21 to 27 µg N2O–N m-2 h-1. In contrast to NO3 fertilization, the increase in N2O emission following NH4 fertilization was not as dramatic. The maximum flux of about 590 µg N m-2 h-1 represented about a 12-fold increase over prefertilization values. However, N2O fluxes declined less rapidly for soils fertilized with NH4–N than for soils amended with NO3–N, especially for the NH4H treatment.
|
Time-integration of fluxes to Day 11 (post-watering data excluded) showed that N fertilization increased N2O emission over controls by a factor of about 20 to 40 (Table 1). Nitrous oxide fluxes for the NO3H, NO3L, and NH4H treatments were nearly identical, varying from 75.8 to 88.8 mg N m-2. In contrast, efflux from the NH4L treatment was about half as large and significantly lower, at 38.2 mg N m-2. Collectively, these data point to a dose response for NH+4 fertilization, but not for NO-3 fertilization, at least over the period of measurement. Fractional loss of applied fertilizer to N2O during the measurement period varied from 0.6 to 1.0%.
Nitrate and NH4–N concentrations for the CT treatment deviated little from prefertilization levels of about 1.5 and 4.0 mg N kg-1 soil throughout the 11-d observational period (data not shown). The time course for change in soil inorganic N concentrations for the NH4H and NH4L treatments (Fig. 6a) was similar to the pattern observed for treatments that included N in Exp. 1 (Fig. 3b). Upon fertilization, NH4–N concentrations immediately (0.12 d) increased to 50 and 19 mg N kg-1 soil for the NH4H and NH4L treatments, respectively. Nitrogen mass balances showed that these concentrations represented 96% (NH4H) and 73% (NH4L) of expected values if the applied N were entirely retained in the 0- to 20-cm soil zone at that time. Thereafter, NH4–N concentrations rapidly declined to prefertilization levels by Day 11. Nitrate concentrations increased steadily to 29 and 12 mg N kg-1 soil by Day 11 in the NH4H and NH4L treatments. Nitrate-fertilized plots (Fig. 6b) showed an initial spike in NO3 concentration to 46 and 22 mg N kg-1 soil for NO3H and NO3L treatments and a gradual decline in concentration to 24 (NO3H) and 7.5 mg N kg-1 soil (NO3L) by Day 11. Soil NH4 remained essentially constant at prefertilization concentrations throughout the entire experiment for both the NO3H and NO3L treatments.
|
|
, NH4L
, and NO3H
treatments. Additional significant, positive correlations were observed between NH4 concentration and N2O flux for the NH4H
and NH4L
treatments and between NO3 concentration and flux for the NO3L
treatment. |
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
The return to prefertilization N2O emission rates within 6 (Fig. 2) to 11 d (Fig. 5) of nutrient addition indicates that the immediate effect of fertilization on flux is short lived. The liquid fertilizers used here stimulated N2O production immediately upon addition by rapidly penetrating the soil surface to contact the microbial community, reducing O2 diffusion, enhancing respiratory activity (NC and W treatments; Exp. 1), and providing N in a form immediately available for denitrification (NO3; Exp. 2) or coupled nitrification–denitrification (treatments including NH4; both experiments). Although N addition initially stimulated emission, moisture assumed increasing importance as experiments progressed. The importance of moisture was clearly demonstrated in Exp. 2 where baseline fluxes at Day 11 were increased 8- to 30-fold by simulated rainfall in all N- treated plots (Fig. 5). The absence of an emission response upon watering control plots that were low in NH4 (4 mg N kg-1) and NO3 (1 mg kg-1) in concert with enhanced emission on watering N-fertilized plots (Fig. 5) indicated that overall N2O production in these soils was controlled by both N availability and soil moisture. Enhanced N2O efflux following rainfall has been commonly reported for agroecosystems and is generally attributed to the dual effects of bringing applied substrate into contact with the soil microbial biomass and reducing soil O2 (Ellis et al., 1998). Liquid swine waste accomplishes the first task upon application. Hence, an N2O flush following rainfall on spray fields will likely be due to the latter effect.
The study site supported a vigorous extant population of nitrifiers, as treatments involving waste or NH4–N addition showed a 15- to 30-fold increase in soil NO3 over prefertilization levels by the termination of these relatively short duration experiments (Figs. 3b and 6a). This conclusion is supported by additional data (not shown) that indicated no change in nitrifier activity in daily monitoring over a 10-d post-fertilization period. Other studies (Nielsen and Revsbech, 1998; Petersen, 1992) report a lag of several days before NO-3 appeared in soils fertilized with liquid cattle waste and attribute this delay to the need for population development. Postfertilization concentrations of NH4 were well below the 300 mg NH4–N kg-1 soil reported by Malhi and McGill (1982) to be inhibitory to nitrification, indicating that this microbial process will be active as long as O2 supply is adequate.
The study site also supported a population of denitrifying bacteria capable of responding rapidly to fertilization. Prefertilization concentrations of soil NO3 (Fig. 3 and 6) were below the 5 mg N kg-1 soil threshold for denitrification (Ryden, 1983). Nitrate fertilization immediately (0.12 h) stimulated N2O emission 7-fold compared with emission following NH4–N addition at similar levels (Fig. 5), pointing at least to the potential for denitrification without a lag upon liquid waste fertilization. Overall, these data indicate that repeated fertilization of this spray field throughout the year is sufficient to continuously maintain active and responsive nitrifying and denitrifying communities.
The rapid return of NH4–N to prefertilization levels and the static NO3 concentrations at 8 or 11 d for plots amended with NH4–N or swine waste (Fig. 3b and 6a) indicated that volatilization, immobilization, plant uptake, and nitrification had exported, consumed, or transformed all of the added nutrient. The loss of 4 to 27% of added N in waste and NH4–N treatments within hours of fertilization is consistent with the reported 69% loss of N via NH3 volatilization within 24 h of liquid swine waste application to a Georgia spray field (Sharpe and Harper, 1997). Liquid swine waste has low mineralization potential because most of the total N is NH4–N, a form that is rapidly processed. Thus, liquid swine waste contrasts sharply with manures and slurries in terms of N cycling dynamics. Manures and slurries have medium-term effects (i.e., several months) on soil nutrient status and microbial activity due to gradual decomposition of organics (Beauchamp, 1997; Ellis et al., 1998). High concentrations of NO3 in NH4- and waste-treated plots at the termination of experiments when N2O fluxes had returned to baseline levels point to the potential for additional N2O production from denitrification uncoupled from nitrification (Nielsen and Revsbech, 1998) if soil moisture conditions become favorable.
Experiment 2 (Table 1) indicated a dose response for N2O emission in spray field soils as the NH4H treatment (150 kg N ha-1) resulted in a time-integrated N2O efflux 2.1-fold higher than the NH4L treatment (75 kg N ha-1). Dose effects have also been reported for N2O emission from manure-amended agricultural soils (e.g., Lessard et al., 1996). No dose response for N2O flux was observed for NO3-amended soils, suggesting that N2O production from denitrification was already maximum at the low fertilization rate. Limmer and Steele (1982) found denitrification potential to be independent of NO3 concentration above 25 mg NO3–N kg-1 soil in a range of soils. Soil NO3–N levels following the NO3L treatment (Fig. 6b) indicate that maximum denitrification rates were achieved at a somewhat lower concentration here. Comparable time-integrated N2O fluxes for the NO3L, NO3H, and NH4H treatments (Table 1) strongly suggested that most N2O production in response to spray field fertilization results from denitrification.
The best predictor of N2O flux was % WFPS, as a significant positive correlation was found between these two variables for five of the seven treatments that involved N addition in this study. This is in accord with the observation made previously that % WFPS immediately after fertilization was likely optimum for N2O production by both nitrification and denitrification. The lack of a consistent correlation between N2O flux and soil nutrients underscores the fact that microbial activity is frequently localized in microsites and measurements taken on bulked soil samples may not adequately characterize conditions at these microsites (Clayton et al., 1994).
Although the N load and liquid volume of the NH4, NC, and W treatments in Exp. 1 were similar, the time-integrated N2O flux was significantly lower for the W treatment than for the NC treatment (Table 1). Soil NH4–N concentrations were statistically similar on all dates for all treatments involving N fertilization in Exp. 1 (Fig. 3b). Hence, a higher immediate postfetilization N loss to volatilization from the W treatment cannot explain the lower time-integrated N2O emission. Dendrooven et al. (1998) postulated that specific compounds in pig slurry such as antibiotics inhibited nitrification in soil cores. Likewise, it is possible that antibiotics or dietary supplements routinely added to swine feed [(Zn, Cu, and Se; Hatfield et al., (1993)] negatively affect the microbial N cycling community when concentrated in excretory matter and applied to soils. Qualitatively, data for temporal changes in soil NO3–N support this postulate as concentrations were significantly higher for the NH4 and NC treatments than for the W treatment at the end of the experiment (Fig. 3b).
Fractional loss of applied fertilizer as N2O in both experiments (range 0.05–1.0%) was somewhat less than the average of 1.25% reported by Bouwman (1994) following an analysis of 87 agroecosystems amended with mineral and organic fertilizers. However, data reported here undoubtedly underestimated total loss to N2O in response to fertilization because they did not account for additional episodic N2O production from residual NO3–N in response to precipitation. Thus, these experiments, which purposely excluded rainfall, simply provided an indication of the time course for N2O flux to peak and return to the background level in the absence of precipitation.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 10, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Wulf, M. Maeting, and J. Clemens Application Technique and Slurry Co-Fermentation Effects on Ammonia, Nitrous Oxide, and Methane Emissions after Spreading: II. Greenhouse Gas Emissions J. Environ. Qual., November 1, 2002; 31(6): 1795 - 1801. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
