HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Grant, R. F. Articles by Puurveen, H. Search for Related Content PubMed Articles by Grant, R. F. Articles by Puurveen, H. Agricola Articles by Grant, R. F. Articles by Puurveen, H. Related Collections Soil Models Nitrogen Soil Biochemistry

Soil Chemistry

Modeling the Effects of Fertilizer Application Rate on Nitrous Oxide Emissions

^a Dep. of Renewable Resources, Univ. of Alberta, Edmonton, AB, Canada T6G 2E3
^b Agric. and Agri-Food Canada Research Branch, Ottawa, ON, Canada K1A 0C6
^c Conservation and Development, Alberta Agric. and Rural Development, Edmonton, AB, Canada T6H 5T6
^d Crop Diversification Centre North, Alberta Agric. and Rural Development, Edmonton, AB, Canada T6H 5Z2

^* Corresponding author (robert.grant{at}afhe.ualberta.ca)

	ABSTRACT

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
The attribution of N₂O emission factors to N inputs from chemical fertilizers requires an understanding of how those inputs affect the biological processes from which these emissions are generated. We propose a detailed model of soil N transformations as part of the ecosystem model ecosys for use in attributing N₂O emission factors to fertilizer use. In this model, the key biological processes—mineralization, immobilization, nitrification, denitrification, root, and mycorrhizal uptake—controlling the generation of N₂O were coupled with the key physical processes—convection, diffusion, volatilization, dissolution—controlling the transport of the gaseous reactants and products of these biological processes. Physical processes controlling gaseous transport and solubility caused large temporal variation in the generation and emission of N₂O in the model. This variation limited the suitability of discontinuous surface flux chambers measurements used to test modeled N₂O emissions. Continuous flux measurements using micrometeorological techniques were better suited to the temporal scales at which variation in N₂O emission occurred and at which model testing needed to be conducted. In a temperate, humid climate, modeled N₂O emissions rose nonlinearly with fertilizer application rate once this rate exceeded the crop and soil uptake capacities for added N. These capacities were partly determined by history of fertilizer use, so that the relationship between N₂O emissions and current N inputs depended on earlier N inputs. A scheme is proposed in which N₂O emission factors rise nonlinearly with fertilizer N inputs that exceed crop plus soil N uptake capacities.

Abbreviations: DOC, dissolved organic carbon • SOC, soil organic C • SON, soil organic N • WFPS, water-filled pore space

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
CURRENT IPCC METHODOLOGY for assessing N₂O emissions from agriculture relies on a constant emission factor (ratio of increase in N₂O emission attributed to an increase in fertilizer application, assigned a value of 1.25 ± 1.0%) for all N inputs from chemical fertilizer or organic amendments based on a general review of field measurements (Mosier et al., 1998). The magnitude of this emission factor is highly uncertain, and values derived from field experiments are frequently found to be smaller. Laegreid and Aastveid (2002) derived an emission factor of about 0.8% for field-applied fertilizer from a review that included more recent measurements of N₂O emissions than those from which the IPCC value was derived. Bouwman et al. (2002b) used a regression model fitted to a range of field measurements to derive a fertilizer-induced emission factor of 0.9%. However emission factors estimated from field measurements are variable (Helgason et al., 2005), reflecting the large spatial and temporal variability in the N₂O fluxes on which estimates are based. Only a small part of this variability may be attributed to N inputs (Laegreid and Aastveid, 2002), with much of the remainder being attributed to climate, soil quality, and land management. Because of this variability, more site-specific emission factors for different N inputs may be used in place of the IPCC value in national inventories of N₂O emissions if these factors are deemed scientifically valid following peer review. However valid methodology for deriving more site-specific emission factors is under development.

Site-specific conditions that most affect N₂O emissions appear to be those that determine the length of time that soil water contents remain higher than a threshold value (0.6 of water-filled pore space [WFPS]) above which N₂O is generated. Thus N₂O emission factors have been found to rise with rainfall, poor soil drainage (de Klein et al., 2003) that may be caused by high clay content (Bouwman et al., 2002a), and lower topographic position (Corre et al., 1999). These emission factors may also vary with N input rate. Kachanoski et al. (2003) found that even under drier conditions, N₂O emissions rose exponentially with rates of spring-applied urea once rates exceeded maximum economic values of 10 g N m^–2 for irrigated wheat in southern Alberta. Similarly Izaurralde et al. (2004) reported significant increases in N₂O emissions when fertilizer rates exceeded recommended values of 4.5 g N m^–2 for a wheat–fallow rotation in southern Saskatchewan. In a broad survey of N₂O emission measurements, Bouwman et al. (2002a) found that N₂O emissions rose little with N fertilizer at low application rates, but rose sharply at application rates greater than 10 g N m^–2. Some of this rise may be attributed to the larger fraction of denitrification products released as N₂O under higher mineral N concentrations (Chantigny et al., 1998). On the other hand, Hénault et al. (1998) reported that N₂O emission rose linearly with fertilizer rate for optimally and excessively fertilized rapeseed in France. These contrasting results might be reconciled by the finding of Kaiser et al. (1998) that if larger N fertilizer applications caused substantial increases in crop growth, consequent increases in transpiration could lower WFPS during summer, thereby offsetting the effects of larger N inputs on N₂O emissions.

Much of the uncertainty in N₂O emission factors for N inputs is caused by uncertainty in the measurement of N₂O fluxes, particularly in the temporal and spatial scales at which measurements need to be made. Many emission factors are estimated from measurement periods of weeks or months following fertilization. However ignoring emissions during the rest of the year can cause substantial underestimation of annual fluxes (Bouwman et al., 2002a). Repeated application of N inputs such as manure over several years may cause accumulation of NO₃^– and organic matter that can gradually raise N₂O emissions above those expected from applications during the current year (Chang et al., 1998). This finding indicates that emission factors for soils with a long history of N inputs may be underestimated from measurements of N₂O fluxes in soils with a short history of N inputs. Therefore emission factors need to be estimated from at least a full year of measurements at sites with a known history of soil amendments. Also many emission factors are estimated from measurements taken near midday from once to a few times per week or month, even though diurnal variation in fluxes can be as much as an order of magnitude (e.g., Williams et al., 1999). Low measurement frequency can introduce bias into temporally aggregated estimates of N₂O emission (Thornton et al., 1996; Bouwman et al., 2002a; Flessa et al., 2002; Pattey et al., 2005a). Replicated measurements of N₂O fluxes used in these estimates are highly variable because of their sensitivity to small differences in WFPS, especially in warmer soils, further raising uncertainty in emission factors derived from them.

The complexity of climate, soil, and management controls on N₂O emissions, and uncertainty in their measurement, have led to the use of mathematical models to account for site-specific effects on emission factors used in regional or national inventories. Sozanska et al. (2002) developed a multiple regression model from field measurements under diverse site conditions to assess N₂O emissions from the UK. They found that N input was the most significant single predictor of N₂O emissions, which in their model rose linearly with inputs, while WFPS accounted for some additional variation in emissions. Bouwman et al. (2002b) proposed a multiple regression model fitted to field measurements in which N₂O emissions rose exponentially with N application rate. Schmid et al. (2001) proposed a more process-based model in which N₂O was generated from nitrification and denitrification reactions that were first-order functions of NH₄⁺ concentration and CO₂ production, respectively. In their model, the fraction of nitrification partitioned to N₂O rose with temperature and WFPS to capture the sensitivity of N₂O emissions to soil temperature and water content. This model gave linear increases in N₂O emissions with increasing rates of NH₄NO₃ application. The sensitivity of N₂O emissions to added N in the process model DNDC (Li et al., 1992) rose with soluble C concentrations arising from SOC decomposition, suggesting that emission factors for N inputs would rise with SOC. This model result is consistent with the findings of Bouwman et al. (2002a) who noted that soils with higher SOC tended to have higher N₂O emissions. However, Sozanska et al. (2002) found that field measurements of N₂O emissions were unrelated to SOC. These apparently contrasting findings may indicate that emissions are mostly driven by decomposition of recent plant residue, which may be a small and not necessarily constant fraction of SOC.

Models of N₂O emissions may be used to account for site-specific effects of N fertilizer application on N₂O emission factors. However these models must be capable of simulating the large temporal variation consistently measured during the brief emission events when most N₂O release occurs. Such simulation requires the coupled modeling of the microbiological processes that generate N₂O and the physical controls on transfer of the gaseous reactants and products of these processes. This coupling should be accomplished at a time scale consistent with that of temporal variation in N₂O fluxes, preferably hourly or less. Such coupling is accomplished in the terrestrial ecosystem model ecosys (Grant, 2001), where key processes of mineralization (Grant et al., 1993a, 1993b), nitrification (Grant, 1994, 1995), and denitrification (Grant, 1991) are coupled with exchange and transport of gasses in aqueous and gaseous states (Grant, 1993a, 1993b; Grant and Roulet, 2002) to simulate N₂O fluxes from soils under different climates and land managements (Grant et al., 1993c, 1993d; Grant and Pattey, 1999, 2003). We propose to use this model to examine the basis for deriving N₂O emission factors that account for different rates of fertilizer application in agricultural fields under different climates.

	MODEL DEVELOPMENT

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
General Overview
The hypotheses for N₂O transformations are described below, and in further detail in Grant and Pattey (2003), with reference to equations and definitions listed in the Appendix and values in Table 1. These hypotheses are part of a model of soil C, N, and P transformations (Grant et al., 1993a, 1993b), which is in turn part of the ecosystem model ecosys (Grant, 2001). The model functions in one, two, or three dimensions by representing all state and rate variables according to their west to east (x), north to south (y), and vertical (z) positions within a complex landscape. The landscape descriptors are omitted from variables in Eq. [1] to [36] of the Appendix for clarity.

View this table: [in this window] [in a new window]   Table 1. Properties of the Brandon Clay Loam at the Greenbelt Farm and Malmo silt loam at the Ellerslie Research Station used to initialize ecosys.

Oxidation of Dissolved Organic Carbon and Reduction of Oxygen by Heterotrophs
Constraints on heterotrophic oxidation of DOC imposed by O₂ uptake are solved in four steps:

(1)DOC oxidation under non-limiting O₂ is calculated from active biomass and DOC concentration ([A2]);

(2)O₂ reduction under non-limiting O₂ is calculated from (1) using a set respiratory quotient ([A3]);

(3)O₂ reduction under ambient O₂ is calculated from radial O₂ diffusion through water films of thickness determined by soil water potential ([A4a]) coupled with active uptake at heterotroph surfaces driven by (2) ([A4b]). O₂ diffusion and active uptake is population-specific, allowing the development of more anaerobic conditions at microbial surfaces associated with more biologically active substrates. O₂ uptake by heterotrophs also accounts for competition with O₂ uptake by nitrifiers, roots and mycorrhizae;

(4)DOC oxidation under ambient O₂ is calculated from (2) and (3) ([A5]). The energy yield of DOC oxidation drives the uptake of additional DOC for construction of microbial biomass M_i,h according to construction energy costs of each heterotrophic population (Eq. [7] to [13] in Grant and Pattey, 2003). Energy costs of denitrifiers are slightly larger than those of obligate heterotrophs, placing denitrifiers at a competitive disadvantage for growth and hence DOC oxidation if electron acceptors other than O₂ are not used.

Oxidation of Dissolved Organic Carbon and Reduction of Nitrate, Nitrite, and Nitrous Oxide by Denitrifiers
Constraints imposed by NO₃^– availability on DOC oxidation by denitrifiers are solved in five steps:

(1)NO₃^– reduction under non-limiting NO₃^– is calculated from a fraction of electrons demanded by DOC oxidation but not accepted by O₂ because of diffusion limitations ([A6]);

(2)NO₃^– reduction under ambient NO₃^– is calculated from (1) [([A7]);

(3)NO₂^– reduction under ambient NO₂^– is calculated from demand for electrons not met by NO₃^– in (2) [([A8]);

(4)N₂O reduction under ambient N₂O is calculated from demand for electrons not met by NO₂^– in (3) [([A9]);

(5)Additional DOC oxidation enabled by NO_x reduction in (2), (3), and (4) is added to that enabled by O₂ reduction from [A5], the energy yield of which drives additional DOC uptake for construction of M_i,n. This additional uptake offsets the disadvantage incurred by the larger construction energy costs of denitrifiers.

Oxidation of Ammonia and Reduction of Oxygen by Nitrifiers
Constraints on nitrifier oxidation of NH₃ imposed by O₂ uptake are solved in four steps:

(1)substrate (NH₃) oxidation under non-limiting O₂ is calculated from active biomass and from NH₃ and CO₂ concentrations ([A11]);

(2)O₂ reduction under non-limiting O₂ is calculated from (1) using set respiratory quotients ([A12]);

(3)O₂ reduction under ambient O₂ is calculated from radial O₂ diffusion through water films of thickness determined by soil water potential ([A13a]) coupled with active uptake at nitrifier surfaces driven by (2) ([A13b]). O₂ uptake by nitrifiers also accounts for competition with O₂ uptake by heterotrophic DOC oxidizers, roots and mycorrhizae;

(4)NH₃ oxidation under ambient O₂ is calculated from (2) and (3) ([A14]). The energy yield of NH₃ oxidation drives the fixation of CO₂ for construction of microbial biomass M_i,n according to construction energy costs of each nitrifier population (Eq. [32] to [34] in Grant and Pattey, 2003).

Oxidation of Nitrite and Reduction of Oxygen by Nitrifiers
Constraints on nitrifier oxidation of NO₂^– imposed by O₂ uptake ([A15] to [A18]) are solved in the same way as are those of NH₃ ([A11] to [A14]). The energy yield of NO₂^– oxidation drives the fixation of CO₂ for construction of microbial biomass M_i,o according to construction energy costs of each nitrifier population.

Oxidation of Ammonia and Reduction of Nitrite by Nitrifiers
Constraints on nitrifier oxidation imposed by NO₂^– availability are solved in three steps:

(1)NO₂^– reduction under non-limiting NO₂^– is calculated from a fraction of electrons demanded by NH₃ oxidation but not accepted by O₂ because of diffusion limitations ([A19]),

(2)NO₂^– reduction under ambient NO₂^– and CO₂ is calculated from Step (1) [([A20]), competing for NO₂^– with [A18],

(3)Additional NH₃ oxidation enabled by NO₂^– reduction in (2) [A21] is added to that enabled by O₂ reduction from [A14]. The energy yield from this oxidation drives the fixation of additional CO₂ for construction of M_i,n.

Uptake of Ammonium and Reduction of Oxygen by Roots and Mycorrhizae

(1)NH₄⁺ uptake by roots and mycorrhizae under non-limiting O₂ is calculated from mass flow and radial diffusion between adjacent roots and mycorrhizae ([A22a]) coupled with active uptake at root and mycorrhizal surfaces ([A22b]),

(2)O₂ reduction is calculated from (1) plus potential oxidation of root and mycorrhizal nonstructural C under non-limiting O₂ using set respiratory quotients ([A23]),

(3)O₂ reduction under ambient O₂ is calculated from mass flow and radial diffusion between adjacent roots and mycorrhizae ([A24a]) coupled with active uptake at root and mycorrhizal surfaces driven by (2) ([A24b]). O₂ uptake by roots and mycorrhizae also accounts for competition with O₂ uptake by heterotrophic DOC oxidizers, and nitrifiers,

(4)Oxidation of root and mycorrhizal nonstructural C under ambient O₂ is calculated from (2) and (3) ([A25]),

(5)NH₄⁺ uptake by roots and mycorrhizae under ambient O₂ is calculated from (1), (2), and (3) ([A26]).

Cation Exchange and Ion Pairing of Ammonium
A Gapon selectivity coefficient is used to solve cation exchange of NH₄⁺ vs. Ca²⁺ ([A27]) and a solubility product is used to equilibrate soluble NH₄⁺ and NH₃ ([A28]).

Volatilization of Ammonium
Exchange of NH₃ between aqueous and gaseous states is driven by disequilibrium between aqueous and gaseous concentrations according to a temperature-dependent solubility coefficient, and is constrained by an air-water interfacial area that depends on air-filled porosity ([A29]).

Soil Transport and Surface-Atmosphere Exchange of Gaseous Substrates and Products
Exchange of all modeled gases ( = CH₄, O₂, CO₂, N₂, N₂O, NH₃, and H₂) between aqueous and gaseous states is calculated in [A30] as described for NH₃ in [A29]. These gases undergo convective-dispersive transport through soil in aqueous ([A31]) and gaseous ([A33]) states driven by soil water flux and gas concentration gradients. Water fluxes are the product of hydraulic conductances and water potential differences, calculated from the sum of matric, osmotic and gravitational components (Eq. [35]–[43] in Grant and Pattey, 2003). Dispersive transport is controlled by aqueous dispersion ([A32]) and gaseous diffusion ([A34]) coefficients calculated from water- and gas-filled porosity. Exchange of all gases between the atmosphere and both aqueous ([A35]) and gaseous ([A36]) states at the soil surface are driven by atmosphere-surface gas concentration differences. This exchange is controlled by boundary layer conductance above the soil surface, calculated from wind speed and from structure of vegetation and surface litter, and by aqueous dispersion ([A32]) and gaseous diffusion ([A34]) coefficients below the soil surface, calculated from surface water- and gas-filled porosity.

	FIELD EXPERIMENTS

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
Greenbelt Research Farm, Ottawa, ON
The N₂O emissions used to test model hypotheses were measured in a temperate, humid climate at the Greenbelt Research Farm in Ottawa, ON, Canada (45°18'N, 75°44'W, mean annual temperature 6.0°C, mean annual precipitation 944 mm). A drained 26-ha field of Brandon clay loam (fine-clayey, mixed, non-acid, mesic, Typic Humaquept; Table 1) was maintained in a maize-soybean rotation under conventional tillage for several years before 2000. Following a soybean crop in 1999, this field was disc-plowed on 4 May 2000, harrowed on 17 May 2000, cultivated on 21 May 2000, fertilized with N, P and K at 1.35, 3.0, and 3.7 g m^–2 respectively and planted with maize (Zea mays L. cv. Pioneer 3905) at 6.67 plants m^–2 on 29 May 2000. Sections of the field were fertilized with another 15.7 g N m^–2 (18 ha) or 9.4 g N m^–2 (8 ha) as anhydrous NH₃ injected at 15 cm on 5 Jul. 2000, and harvested on 6 Nov. 2000. Surrounding fields were also planted with maize.

Two eddy covariance measurement systems were used to record fluxes over the sections of the field fertilized with NH₃ at 15.7 and 9.4 g N m^–2. Fluxes of momentum, CO₂, and of sensible and latent heat were calculated every 30 min from air temperature, three-dimensional wind velocity (sonic anemometer DAT-310, Kaijo-Denki Company, Tokyo), and from H₂O and CO₂ molar fractions (infrared gas analyzer LI-6262, LI-COR Inc., Lincoln, NE) recorded at 20 Hz in fast response, absolute mode at a height of 2 m above the canopy (Pattey et al., 1996; Pattey et al., 2001). Fluxes of N₂O were measured from 11 March to 11 April and from 6 July to 29 August 2000 by the flux-gradient technique using two towers each having two inlets separated by 0.50 m, which were maintained in the inertial sublayer. The N₂O flux gradients were measured for 30 min averaging periods alternately between the areas fertilized with 15.7 and 9.4 g N m^–2 using a tunable diode laser (TDL) trace gas analyzer (Campbell Scientific, Logan, UT). The N₂O fluxes were calculated as described in Grant and Pattey (1999) and screened according to the following criteria: the absolute pressure difference between the two air intakes of the gradient should be lower than 0.01 kPa (0.1 mbar), the pressure in the sampling cell should be higher than 6.4 kPa (64 mbar), [N₂O] standard deviation should be smaller than 3.2 x 10^–3 µmol mol^–1 (3.2 ppb) and the ratio of the [N₂O] standard deviation over the [N₂O] gradient should be lower than 3.2. Night time data associated with windy conditions (i.e., u_* > 0.09 m s^–1) were kept following the procedure in Pattey et al. (2002), which allows the retaining of an acceptable proportion of calm periods within predominantly windy nights. The N₂O flux resolution was estimated to be about 0.018 mg N m^–2 h^–1 (Pattey et al., 2005b). Half-hourly averaged data for solar radiation, air temperature, humidity, wind speed, and precipitation were recorded continuously at the experimental site from September 1997.

Ellerslie Research Station, Edmonton, AB
Model results were also compared with field measurements in a boreal subhumid climate at the Ellerslie Research Station near Edmonton, AB, Canada (53°19'N, 113°34'W, mean annual temperature 2.4°C, mean annual precipitation 483 mm) during years with different precipitation. These comparisons were performed to determine if N₂O emissions were likely to be smaller in a cooler, drier climate than that at Ottawa. A fertilizer trial was conducted in 5 by 3 m field plots on a Malmo silt loam (Typic Cryoboroll– Table 1) under conventionally tilled wheat (Triticum aestivum L.) (chisel cultivation and harrowing in fall and again in spring) following several years under unfertilized wheat. The trial was first planted in 1984 and maintained under fall (mid-October) or spring (mid-May) applications of urea at 0, 9, or 18 g N m^–2 yr^–1. The N₂O emissions were measured weekly between 1100 and 1400 h MDT, with additional measurements after rainfall, in three replicates of each fertilizer treatment in each trial during 2001 to 2003 using vented plexiglass chambers (0.60 m length x 0.15 m width x 0.10 m height). Gas samples of 20 mL were withdrawn by syringe from the chamber 30 min. after placement on a collar installed to a depth of 5 cm in each field plot. Samples were injected into 10-mL exetainers from which N₂O concentrations were measured on a gas chromatograph (Varian Canada Inc., Mississauga, ON) with a ⁶³Ni electron capture detector.

	MODEL EXPERIMENT

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
Greenbelt Research Farm, Ottawa, ON
Ecosys was initialized with the physical properties of the Brandon clay loam (Table 1) and the biological properties of maize and soybean. The model was then run through ten cycles of a conventionally tilled and fertilized maize–soybean rotation under repeated sequences of hourly surface boundary conditions for solar radiation, air temperature, humidity, wind speed, and precipitation recorded at the Greenbelt Farm during 1998 and 1999. This part of the model run allowed soil microbial activity (Eq. [A1] to [A26]) to equilibrate under conditions similar to those that existed at the experimental site before 2000. The model run was then continued under hourly surface boundary conditions recorded during 2000 with simulated tillage, fertilizing, planting, and harvesting practices corresponding to those conducted at the field site. During the model run, all biological (Eq. [A1]–[A26]) and physical (Eq. [A27]–[A36]) processes were solved on time steps of 1 h and 2 min respectively, with surface boundary conditions assumed constant during each hour. Surface N₂O fluxes from the model (Eq. [A35] and [A36]) were compared with those measured by TDL over the sections of the field to which anhydrous NH₃ had been applied at 15.7 and 9.4 g N m^–2.

To examine the relationship between modeled annual N₂O emissions and fertilizer rates at Ottawa, ecosys was run from 1 July 2000 to 30 April 2001 under the site conditions of the Greenbelt Farm with anhydrous NH₃ from 0 to 30 g N m^–2 injected to a depth of 0.15 m on 5 July 2000. Model output for cumulative emissions over this period were then related to application rate.

Ellerslie Research Station, Edmonton, AB
The model was initialized with the physical properties of the Malmo silt loam (Table 1) and the biological properties of wheat. Model runs were started with conventionally tilled, unfertilized continuous wheat under a 5-yr sequence of hourly surface boundary conditions for solar radiation, air temperature, humidity, wind speed, and precipitation recorded at the Ellerslie Research Station from 1999 to 2003. This part of the model run allowed soil microbial activity (Eq. [A1] to [A26]) to equilibrate under conditions similar to those that existed at the experimental site before the start of fertilization in 1984. The model run then continued through four more cycles of the 5-yr weather sequence with conventionally tilled continuous wheat fertilized at times and rates corresponding to those in the field trial.

Modeled and measured results from this trial were used to examine the relationship between fertilizer rates and N₂O emissions during a dry year (2001) and a year with near-normal precipitation (2003).

	RESULTS

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
Measured vs. Modeled Nitrous Oxide Fluxes Following Different Rates of Anhydrous Ammonia
The weather at Ottawa during 2000 was humid, with frequent precipitation (Fig. 1a ) that kept soil water content above field capacity (Table 1) until late July (Fig. 1b). Measured and modeled N₂O fluxes rose gradually after fertilization with 15.7 or 9.4 g N m^–2 of anhydrous NH₃ on DOY 187 (Fig. 2a ) following rainfall on DOY 191 (Fig. 1a) and warming on DOY 194 to DOY 195 (Fig. 1c) at the Greenbelt Research Farm. Elevated fluxes exceeding 0.25 mg N m^–2 h^–1 were maintained in the model and in the field from DOY 194 to DOY 209 (15.7 g N m^–2) (Fig. 2a,b,c) or to DOY 201 (9.4 g N m^–2) (Fig. 2a,b). In the model, anhydrous NH₃ was added directly to aqueous NH₃ (NH_3s) at the depth of injection in a two-dimensional banded zone that then expanded horizontally and vertically by diffusion. Within both the band and non-band zones, [NH_3s] could be oxidized by nitrifiers ([A14], [A21]), paired with H⁺ to form NH₄⁺ (Eq. [A28]), or volatilized ([A29]) and released to the atmosphere ([A30] to [A36]). Large [NH_3s] in the band drove rapid potential oxidation by nitrifiers (X'_NH3i,n in [A11]), which accelerated nitrifier O₂ demand (R'_O2i,n in [A12]), especially under higher soil temperatures (f_t in [A11]). Oxygen uptake by nitrifiers (R_O2i,n in [A13]) failed to meet this demand whenever nitrifier aqueous O₂ concentrations ([O_2mi,n] in [A13b]) declined with respect to the Michaelis–Menten constant for O₂ (K_O2n in [A13b]). Such declines occurred whenever soil aqueous O₂ concentrations ([O_2s] in [A13a]) could not be maintained by dissolution of gaseous O₂ (V_O2 in [A30]). Dissolution slowed when low air-filled porosity _g that occurred near the soil surface after rainfall (Fig. 1a), reduced gaseous diffusivity (D_gO2 in [A34]). This diffusivity in turn reduced surface O₂ gas exchange (Q'_gO2 in [A35]) and soil transport (Q_gO2 in [A33]) needed to maintain [O_{2 g}] in the soil profile. Dissolution also slowed when the air–water interface (a_w in [A30]) became smaller with smaller _g (Kim et al. 1999). Dissolution slowed further when higher soil temperatures reduced aqueous solubility of O₂ (f_tsO2 in [A30]), and hence the value of [O_2s] in equilibrium with [O_2g] that sustained O₂ uptake by microbial ([A4], [A13], [A17]), root, and mycorrhizal ([A24]) populations.

View larger version (46K): [in this window] [in a new window]   Fig. 1. (A) Precipitation measured, (B) 10-cm soil water content () modeled, and (C) 10-cm soil temperature measured (symbols) and modeled (line) at Ottawa during 2000. Vertical lines indicate period of N2O emissions shown in Fig. 2.

View larger version (40K): [in this window] [in a new window]   Fig. 2. N2O fluxes measured (symbols) and modeled (lines) at Ottawa during (a) DOY 186– 195, (b) DOY 196–205, and (c) DOY 206–215 in 2000 following anhydrous NH3 applications of 15.5 and 9.4 g N m–2 on DOY 187.

Diurnal rises in N₂O fluxes modeled with diurnal rises in soil temperature were attributed to the combined effects of more rapid O₂ demand (R'_O2i,h in [A3] from f_t in [A2]) and slower O₂ supply (V_O2 in [A30]) caused by decreasing aqueous solubility (f_tsO2 in [A30]) and hence aqueous O₂ concentrations ([O_2s] in [A4], [A13], [A17], and [A24]). These changes in demand and supply of O₂ raised the demand for electron acceptors unmet by O₂ from nitrifiers (R'_O2i,n- R_O2i,n in [A19]) and denitrifiers (R'_O2i,d- R_O2i,d in [A6]), particularly when high soil water contents reduced O₂ gaseous diffusivity (D_gO2 in [A34]), air-water interfacial area ([A30]), and hence gaseous transfer ([A33]). Large diurnal rises in modeled fluxes were also attributed to accelerated volatilization of aqueous N₂O (V_N2O in [A30]) caused by decreasing aqueous solubility during diurnal soil warming (f_tsN2O in [A30]), as proposed by Blackmer et al. (1982). The processes driving rises in N₂O emissions modeled during diurnal soil warming were reversed during diurnal soil cooling, causing later declines in N₂O emissions.

Measured and modeled N₂O fluxes declined gradually after DOY 209 (15.7 g N m^–2) (Fig. 2c) or DOY 201 (9.4 g N m^–2) (Fig. 2b). In the model, the rate of NO₂^– reduction (R_{NO2i, n} in [A20]) declined when the rate of NH₃ oxidation (X_{NH3i, n} in [A14]) was slowed by declining [NH_3s] ([A11]) once most fertilizer NH₃ had been oxidized. The later decline in NH₃ oxidation in the 15.7 vs. 9.4 g N m^–2 treatment allowed substantial N₂O emissions to be modeled for several days after rainfall on DOY 203 to DOY 204. Only small emissions were measured or modeled in both fertilizer treatments after DOY 210. Total N₂O emissions between DOY 187 and 241 (6 July to 29 August 2000) estimated from gap-filled field measurements and calculated in the model were 0.170 and 0.214 g N m^–2 respectively after an anhydrous NH₃ application of 15.7 g N m^–2, but only 0.045 and 0.092 g N m^–2 respectively after one of 9.4 g N m^–2.

Comparison of N₂O fluxes in the model with those measured continuously using micrometeorological techniques was complicated by the effects of changing wind speed and direction on source areas of N₂O detected by the TDL. Diurnal variation in N₂O fluxes modeled at Ottawa appeared similar to that measured during DOY 185 to DOY 205 (Fig. 2a, 2b), but smaller than that during DOY 205 to DOY 215 (Fig. 2c). However the detection of diurnal trends in the measured fluxes was constrained by the frequent rejection of nighttime values because low wind speeds caused inadequate friction velocity for accurate measurements.

Modeled Nitrous Oxide Emissions vs. Application Rate of Anhydrous Ammonia
Nitrous oxide emissions modeled from 1 Jul. 2000 to 30 Apr. 2001 rose with anhydrous NH₃ application from 0.04 g N₂O-N m^–2 at 0 g NH₃–N m^–2 to 0.59 g N₂O-N m^–2 at 30 g NH₃–N m^–2 (Table 2). Most of these emissions were modeled within the first 30 d of NH₃ application on 5 Jul. 2000. Emissions rose nonlinearly with the amount of NH₃ applied, so that the emission factor for anhydrous NH₃ rose from 0.1% with an application of 3 g NH₃–N m^–2 to 1.8% with applications of 30 g NH₃– N m^–2 (Table 2). This rise occurred because N additions from larger applications of NH₃ exceeded the N uptake capacities of the crop and soil (collectively the ecosystem N uptake capacity), as apparent in the greater concentrations of mineral N remaining in the soil 1 yr after application (Table 2). Residual mineral N increased with NH₃ application because (1) N uptake ([A22]) during the period of rapid nitrification after fertilizer application did not increase with applications exceeding 6 g NH₃–N m^–2, as evident from N removal in maize grain (Table 2), (2) N immobilization in soil organic matter ([A1]) rose less with NH₃ as application rates increased, as evident from smaller rises in soil organic N (SON) (Table 2), (3) N losses by volatilization ([A29]) and emission ([A33], [A36]) rose with NH₃ application but accounted for only 12% of the NH₃ added, (4) N gains from non-symbiotic N₂ fixation declined with NH₃ application, but rates were small (Table 2). Greater concentrations of mineral N drove more rapid oxidation of NH_3s ([A11]) and uptake of O₂ ([A12]), and hence more rapid reduction of NO₂^– ([A20]) during nitrification, and more rapid reduction of NO₃^– ([A7]) and NO₂^– ([A8]) during denitrification, over longer periods of time.

View larger version (43K): [in this window] [in a new window]   Fig. 3. N2O fluxes measured (symbols) and modeled (lines) at Ellerslie during June and early July in (a) 2001 and (b) 2003 following spring urea applications of 9 and 18 g N m–2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
Nitrous Oxide Emissions vs. Fertilizer Application Rate
Model results indicated that N₂O emissions rose nonlinearly with fertilizer application rates when these rates caused mineral N availability to exceed ecosystem (crop + soil) N uptake capacity as indicated by declining gains in maize N yield and SON, and by rises in residual mineral N with further NH₃ application (Table 2). This nonlinear rise caused emission factors to increase with application rates. These findings are consistent with those from a global survey of N₂O emission experiments by Bouwman et al. (2002a) in which emissions rose little with application rates below 10 g N m^–2 but rose more rapidly with higher rates. Kachanoski et al. (2003) also found that N₂O emissions rose exponentially with fertilizer application rates, particularly when rates exceeded the maximum economic rate of 10 g N m^–2 above which crop yields did not rise substantially.

The crop N uptake capacity modeled at Ottawa was exceeded at NH₃ application rates as low as 6 g N m^–2 because the field site had been maintained under an intensively managed cereal-legume rotation for several years before 2000, during which large amounts of N had been applied as fertilizer to cereal crops or had been fixed symbiotically by legume crops. The consequent accumulation of low C/N residue during the 20-yr model spinup under this rotation allowed mineralization of SON ([A1a]) to meet most of the crop N uptake requirement ([A26]) in the model during 2000, so that fertilizer additions contributed mostly to increases in mineral N rather than grain N (Table 2). Maize yields recorded from this field are large without fertilizer, and increase little with fertilizer use (Pattey et al., 2001). Rates of fertilizer use that caused crop N uptake capacity to be exceeded would, if sustained over time, cause greater accumulation of SON and mineral N, and subsequently more rapid mineralization and nitrification, and therefore larger N₂O emissions. An example is shown in Fig. 4 of a longer-term trend in N₂O emissions modeled with sustained fertilizer application under variable precipitation during model spinup at Ellerslie. Applications of 18 g N m^–2 yr^–1 caused crop + soil N uptake capacity to be exceeded after 6 yr so that SON stopped rising (Fig. 4b), causing rapid accumulation of mineral N and gradual rises in N₂O emissions thereafter (Fig. 4c).

View larger version (28K): [in this window] [in a new window]   Fig. 4. Changes in (b) soil organic N (SON), (c) mineral N (NH4+ + NO3–), and (d) N2O emissions modeled under (a) variable precipitation during 25 yr of urea application at 9 and 18 g N m–2 yr–1.

	APPENDIX

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 
Mineralization and Immobilization of Ammonium by All Microbial Populations

INH4i,n,j = Mi,m,j,CCNj –Mi,m,j,N (INH4i,n,j < 0) A1a INH4i,n,j = (Mi,m,j,CCNj –Mi,m,j,N)[NH4+]/([NH4+] + KNH4m) (INH4i,n,j > 0) A1b INO3i,n,j = (Mi,m,j,CCNj –(Mi,m,j,N + INH4i,n,j) [NO3–]/([NO3–] + KNO3m) (INO3i,n,j > 0) A1b

Oxidation of DOC and Reduction of Oxygen by Heterotrophs

X'DOCi,h = {X'DOC Mi,h,a [DOCi]/([DOCi]) + KXh} ft A2 R'O2i,h = RQC X'DOCi,h A3 RO2i,h = 4nMi,h,aDsO2 ([O2s] –[O2mi,h])[rmrw/(rw – rm)] A4a = R'O2i,h[O2mi,h]/([O2mi,h] + KO2h) A4b XDOCi,h = X'DOCi,h RO2i,h/R'O2i,h A5

Oxidation of DOC and Reduction of Nitrate, Nitrite, and Nitrous Oxide by Denitrifiers

R'NO3i,d = ENOx fe (R'O2i,d –RO2i,d) A6 RNO3i,d = R'NO3i,d [NO3–]/([NO3–] + KNO3d) A7 RNO2i,d = (R'NO3i,d –RNO3i,d)[NO2–]/([NO2–] + KNO2d) A8 RN2Oi,d = 2 (R'NO3i,d –RNO3i,d –RNO2i,d)[N2O]/([N2O] + KN2Od) A9 XDOCi,d = XDOCi,d (from [A5]) + FNOx (RNO3i,d + RNO2i,d) + FN2O RN2Oi,d A10

Oxidation of Ammonia and Reduction of Oxygen by Nitrifiers

X'NH3i,n = X'NH3 Mi,n,a {[NH3s]/([NH3s] + KNH3n)} {[CO2s]/([CO2s] + KCO2)} ft A11 R'O2i,n = RQNH3 X'NH3i,n + RQC X'Ci,n A12 RO2i,n = 4 n Mi,n,a DsO2 [rm rw/(rw –rm)] ([O2s] – [O2mi,n]) A13a = R'O2i,n [O2mi,n]/([O2mi,n] + KO2n) A13b XNH3i,n = X'NH3i,n RO2i,n /R'O2i,n A14

Oxidation of Nitrite and Reduction of Oxygen by Nitrifiers

X'NO2i,o = X'NO2 Mi,o,a {[NO2–]/([NO2–] + KNO2o)} {[CO2s]/([CO2s] + KCO2)} ft A15 R'O2i,o = RQNO2 X'NO2i,o + RQC X'Ci,o A16 RO2i,o = 4 n Mi,o,a DsO2 [rm rw/(rw – rm)] ([O2s] – [O2mi,o]) A17a = R'O2i,o [O2mi,o]/([O2mi,o] + KO2o) [A17b] XNO2i,o = X'NO2i,o RO2i,o /R'O2i,o [A18]

Oxidation of Ammonia and Reduction of Nitrite by Nitrifiers

R'NO2i,n = ENOx fe (R'O2i,n –RO2i,n) A19 RNO2i,n = R'NO2i,n {[NO2–]/([NO2–] + KNO2n)}{[CO2s]/([CO2s] + KCO2)} A20 XNH3i,n = XNH3i,n (from [A14]) + 0.33 RNO2i,n A21

Uptake of Ammonium and Reduction of Oxygen by Roots and Mycorrhizae

U'NH4s,z = {Uws,z[NH4+] + 2Ls,z De NH4 ([NH4+] – [NH4+]s,z)/ln(ds,z/rs,z)} also NO3–, H2PO4– A22a = U'NH4s,z As,z ([NH4+]s,z –[NH4+mn]s,z)/([NH4+]s,z –[NH4+mn]s,z + KNH4 [A22b] R'O2s,z = RQNH4 U'NH4s,z + RQC X'Cs,z [A23] RO2 s,z = 2Ls,z DsO2 ([O2s] – [O2ms,z])/ln(rw/rs,z)} [A24a] = R'O2s,z [O2ms,z]/([O2ms,z] + KO2s,z) [A24b] XCs,z = X'Cs,z RO2s,z/R'O2s,z [A25] UNH4s,z = U'NH4s,z RO2s,z/R'O2s,z [A26]

Cation Exchange and Ion Pairing of Ammonium

X-Ca + 2 [NH4+] 2X-NH4 + [Ca2+] [A27] [NH4+] [NH3s] + [H+] [A28

Volatilization of Ammonium

VNH3 = aw DtNH3 (S'NH3 ftsNH3 NH3g] –[NH3s]) [A29]

Soil Transport and Atmosphere- Surface Exchange of Gaseous Substrates and Products

V = aw Dt (S' fts g] –[s]) [A30] Qs = Qw [s] + 2Ds ([s]z –[s]z+1)/(Zz + Zz+1) [A31] Ds = |Qw| + D's fta s 0.5(wz + wz+1) [A32] Qg = –Qw [g] + 2Dg ([g]z –[g]z+1)/(Zz + Zz+1) [A33] Dg = D'g ftg 0.5(gz + gz + 1) /pß [A34] Q's = ga ([a] –{2[s]1Ds/Z1 + ga [a]}/{2 Ds S' fts/Z1 + ga}) [A35] Q'g = ga ([a] –{2[g]1Dg/Z1 + ga [a]}/{2 Dg/Z1 + ga}) [A36]

	GLOSSARY

TOP ABSTRACT INTRODUCTION MODEL DEVELOPMENT FIELD EXPERIMENTS MODEL EXPERIMENT RESULTS DISCUSSION APPENDIX GLOSSARY REFERENCES

 

Subscript Definition a active component of Mi,m (Grant et al., 1993a, 1993b) d heterotrophic denitrifier population (subset of h) h heterotrophic community (subset of m) i substrate-microbe complex (Grant et al., 1993a, 1993b) j kinetic components of Mi,m m all microbial communities n autotrophic ammonia oxidizer population (subset of m) o autotrophic nitrite oxidizer population (subset of m) s plant population (e.g. species) z root population (roots or mycorrhizae) Variable Definition Units Equation Notes As,z root surface area m2 m–2 [A22] = rs,z2 Ls,z aw air-water interfacial area m2 m–2 [A29], [A30] (Skopp, 1985) sensitivity of Dg to g – [A34] 2 (Millington and Quirk, 1960) ß sensitivity of Dg to p – [A34] 0.67 (Millington and Quirk, 1960) [Ca2+] concentration of Ca2+ in soil solution mol m–3 [A27] CNj maximum ratio of Mi,m,j,N to Mi,m,j,C maintained by Mi,m,j g N g C–1 [A1] 0.22 and 0.13 for j= labile and resistant [CO2s] CO2 concentration in soil solution g C m–3 [A11], [A15], [A20] De NH4 effective dispersivity-diffusivity of NH4 during root uptake m2 h–1 [A22] D'g gaseous diffus