1
Application of the DNDC model to predict emissions of N2O from Irish
2
agriculture.
3
4
5
6
7
8
9
10
11
12
M. Abdalla1, M. Wattenbach2, P. Smith2, P. Ambus3, M. Jones1 and M. Williams1
1
Department of Botany, School of Natural Sciences, Trinity College Dublin, Dublin 2, Ireland
2
School of Biological Sciences, University of Aberdeen, Cruickshank Building, St. Machar Drive,
Aberdeen, AB24 3UU, UK.
3
Riso Research Centre, Technical University of Denmark, Frederikborgvej 399, DK-4000,
Roskilde
Key words: Nitrous oxide, DNDC model, arable, pasture
13
14
ABSTRACT
15
16
A mechanistic model that describes N fluxes from the soil, DeNitrification
17
DeComposition (DNDC), was tested against seasonal and annual data sets of nitrous
18
oxide flux from a spring barley field and a cut and grazed pasture at the Teagasc Oak
19
Park Research Centre, Co. Carlow, Ireland. In the case of the arable field, predicted
20
fluxes of N2O agreed well with measured fluxes for medium to high fertilizer input
21
values (70 to 160 kg N ha-1) but described poorly measured fluxes from zero fertilizer
22
treatments. In terms of cumulative flux values, the relative deviation of the predicted
23
fluxes from the measured values was a maximum of 6% for the highest N fertilizer inputs
24
but increased to 30% for the medium N and more than 100% for the zero N fertilizer
25
treatments. A linear correlation of predicted against measured flux values for all fertilizer
26
treatments (r2 = 0.85) was produced, the equation of which underestimated the seasonal
27
flux by 24%. Incorporation of literature values from a range of different studies on arable
28
and pasture land did not significantly affect the regression slope. DNDC describe poorly
29
measured fluxes of N2O from reduced tillage plots of spring barley. Predicted cumulative
30
fluxes of N2O on plots disc ploughed to 10cm, underestimated measured values by up to
31
55%.
32
1
1
For the cut and grazed pasture the relative deviations of predicted to measured fluxes
2
were 150 and 360% for fertilized and unfertilized plots. This poor model fit is considered
3
due to DNDC overestimating the effect of initial soil organic carbon (SOC) on N2O flux,
4
as confirmed by a sensitivity analysis of the model. As the arable and grassland soils
5
differed only in SOC content, reducing SOC to the arable field value significantly
6
improved the fit of the model to measured data such that the relative deviations decreased
7
to 9 and 5% respectively. Sensitivity analysis highlighted air temperature as the main
8
determinant of N2O flux, an increase in mean daily air temperature of 1.5oC resulting in
9
almost 90% increase in the annual cumulative flux. Using the Hadley Centre Global
10
Climate Model data (HCM3) and the IPCC emission scenarios A2 and B2, DNDC
11
predicted increases in N2O fluxes of approximately 30% (B2) and 60% (A2) from the
12
spring barley field and approximately 20% (A2 and B2) from the cut and grazed pasture
13
by the end of this centaury (2061-2090).
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
2
1
INTRODUCTION
2
3
National inventories of N2O fluxes from agricultural soils, as required by signatory
4
countries to the United Nations Framework Convention of Climate Change (UNFCC),
5
are in the main derived from the use of the default IPCC Tier 1 method, where 1.25% of
6
applied inorganic nitrogen to agricultural soils is assumed to be released to the
7
atmosphere as nitrous oxide-N (Bouwman, 1996; IPCC, 1997; 2000). This standard
8
reporting procedure has advantages in collating annual inventories but may mask
9
significant variations in emission factors (EFs) on a regional scale (Schmid et al., 2001;
10
Laegreid and Aastveit, 2002). For instance in Ireland, published EFs derived from field
11
measurements of N2O using either eddy covariance or static chamber methods vary from
12
3.4% for Cork grassland and 0.7 to 4.9% of the applied N fertilizer for the Wexford
13
grassland depending on soil type, land management, climate and year (Hsieh et al., 2005;
14
Hyde et al., 2005; Flechard et al., 2007).
15
16
Given the considerable expense of establishing and maintaining relevant flux
17
measurement sites, the use of simulation models to estimate N2O fluxes from agricultural
18
soils using soil and climate data has obvious benefits. Modelling also allows easy
19
interpretation of the complex links between soil physical, chemical and microbial
20
processes that underpin nitrification, denitrification and decomposition. Models can
21
simulate the processes responsible for production, consumption and transport of N2O in
22
both the long and short term, and also on a spatial scale (Williams et al., 1992).
23
24
Simulation models range from simple empirical relationships based on statistical analyses
25
to complex mechanistic models that consider all factors affecting N2O production in the
26
soil (Li et al., 1992; Frolking et al., 1998; Stenger et al., 1999; Freibauer 2003; Roelandt
27
et al., 2005; Jinguo et al., 2006). Variations in soil moisture, soil temperature, carbon and
28
nitrogen substrate for microbial nitrification and denitrification are critical to the
29
determination of N2O emissions (Leffelaar and Wessel, 1988; Tanji, 1982; Frissel and
30
Van Veen, 1981; Batlach and Tiedje, 1981; Cho et al., 1979). One widely used
31
mechanistic model is DeNitrification DeComposition (DNDC) developed to assess N2O,
3
1
NO, N2 and CO2 emissions from agricultural soils (Li et al., 1992a, 1994; Li 2000). The
2
rainfall driven process-based model DNDC (Li et al., 1992) was originally written for
3
USA conditions. It has been used for simulation at a regional scale for the United States
4
(Li et al., 1996) and China (Li et al., 2001). Advantages of DNDC are that it has been
5
extensively tested and has shown reasonable agreement between measured and modelled
6
results for many different ecosystems such as grassland (Brown et al., 2001; Hsieh et al.,
7
2005; Saggar et al., 2007), cropland (Li, 2003; Cai et al., 2003, Yeluripati et al., 2006;
8
Pathak et al., 2006; Tang et al., 2006) and forest (Li, 2000; Stange et al., 2000; Kesik et
9
al., 2006). The model has reasonable data requirement and is suitable for simulation at
10
appropriate temporal and spatial scales.
11
The DNDC model contains 4 main sub-models (Li et al., 1992; Li, 2000); the soil climate
12
sub-model calculates hourly and daily soil temperature and moisture fluxes in one
13
dimension, the crop growth sub-model simulates crop biomass accumulation and
14
partitioning, the decomposition sub-model calculates decomposition, nitrification, NH3
15
volatilization and CO2 production whilst the denitrification sub-model tracks the
16
sequential biochemical reduction from nitrate (NO3) to NO2-, NO, N2O and N2 based on
17
soil redox potential and dissolved organic carbon.
18
19
This paper presents a field evaluation of DNDC for an Irish sandy loam soil under both
20
arable and grassland crops with different fertilizer and tillage regimes. Results are
21
discussed in terms of the suitability of this model for estimating annual and seasonal
22
fluxes of N2O from Irish agriculture. In addition, DNDC is used to estimate future N2O
23
fluxes from Irish agriculture due to climate change using climate data generated by the
24
Hadley Centre Global Climate models (HadCM3; Sweeney and Fealy, 2003).
25
26
27
28
29
30
4
1
MATERIALS AND METHODS
2
3
Experiments
4
Measurements of N2O flux were carried out for a spring barley field from April–August for
5
two consecutive seasons (2004/05), and for a cut and grazed pasture from October 2003 to
6
November 2004. Both fields were located at the Oak Park Research Centre, Carlow,
7
Ireland (52o86′ N, 6o54′ W). The arable field was seeded with spring barley (cv. Tavern) at
8
a density of 140 kg ha-1 and managed under two different tillage regimes; conventional
9
tillage where inversion ploughing to a depth of 22 cm was carried out in March, five weeks
10
prior to planting, and reduced tillage to a depth of 15 cm which was carried out in
11
September of the year before. The field was sprayed with weed killer (Roundup Sting) at
12
4.0L ha-1, three times per season, once pre- and twice post-planting.
13
The cut and grazed pasture has been permanent grassland for at least the past eighty years
14
and was ploughed and reseeded in October 2001 with perennial ryegrass (Lolium perenne
15
L., cv Cashel) at a density of 13.5 kg ha-1 and white clover (Trifolium repens L., cv Aran)
16
at a density of 3.4 kg ha-1. Daily minimum and maximum air temperature (oC) and
17
rainfall in (mm) were recorded at the Teagasc Research Centre Weather Station (Met
18
Eireann). Initial soil properties and climate factors of both sites are summarized in
19
Table 1.
20
21
For the arable field in 2004, three rates of N-fertilization 140 (N1), 70 (N2) and 0 (N3) kg
22
N ha-1, were applied once on the 27th of April, whereas in 2005, two fertilizer applications
23
took place on the 12th of April 106 (N1), 53 (N2) and 0 (N3) kg N ha-1, and on the10th of
24
May 53 (N1), 26 (N2) and 0 (N3) kg N ha-1. The total amount of N-fertilization applied in
25
2005 was therefore 159 (N1), 79 (N2) and 0 (N3) kg N ha-1. For the cut and grazed pasture,
26
nitrogen fertilizer was applied at a total rate of 200 kg N ha-1 y-1 divided in to two
27
applications of 128 and 72 kg N ha-1 on the 2nd of April and the 27th of May respectively.
28
Separate areas of the field were kept unfertilized as control plots. Fertilizer was applied in
29
the form of Calcium Ammonium Nitrate (CAN). Animal grazing was from July to
30
November 2003 and from July to November 2004 with a stocking rate of 2 cattle ha-1.
5
1
Field N2O fluxes
2
Nitrous oxide fluxes were measured from 24 replicated chambers at the arable field and 7
3
replicated chambers at the cut and grazed pasture, using the methodology of Smith et al.,
4
(1995). Measurements were taken every week except for times of fertilizer application
5
where sampling was increased to 2 times per week. Samples were taken using a 60 ml
6
gas-tight syringe after flushing of the syringe to ensure adequate mixing of air within the
7
chamber. All 60 ml of the sample was then injected into a 3ml gas-tight vial with a vent
8
needle inserted into the top, and stored until analysis. Gas samples were measured within
9
one month of collection using a gas chromatograph (Shimadzu GC 14B, Kyoto, Japan)
10
with electron capture detection.
11
12
DNDC model
13
In this study the DNDC model (version 8.9; http://www.dndc.sr.unh.edu/) was tested for
14
both the arable field and the cut and grazed pasture. All field management variables,
15
including grain yield, fertilizer application and tillage system (where reduced tillage was
16
defined as disk or chisel ploughing to 10cm) were input into the model. Soil properties
17
and climate input data are summarized in Table 1. For the arable field model testing was
18
possible only for the growth period of the crop, whilst for the cut and grazed pasture 12
19
months of data were used. The model testing was carried out by (1) comparing the
20
measured and modelled temporal pattern of weekly N2O flux values, (2) comparing the
21
measured and modelled cumulative N2O fluxes (using weekly values), and (3) comparing
22
the measured and modelled emission factors.
23
24
The relative deviation (y) of the modelled flux from measured flux values was calculated
25
by the following equation:
26
27
Y = (XS – XO)/XO x 100,
28
29
where XO and XS are the measured and modelled fluxes respectively. Annual and
30
seasonal cumulative flux for DNDC outputs were calculated as the sum of simulated
31
daily fluxes (Cai et al., 2003). EFs for the modelled data were calculated by subtracting
6
1
cumulative DNDC flux data for unfertilized soils from that of the fertilized soils and
2
dividing by the N fertilizer input corrected for ammonia volatilization (10%). Sensitivity
3
analysis was carried out by varying a single determinant factor whilst keeping other
4
factors constant for one annual cycle of the model. Determinant factors tested are listed in
5
Table 4.
6
7
Simulation of future N2O flux
8
Climate change impact on N2O fluxes from the spring barley and the cut and grazed
9
pasture was studied using climate data generated from the Hadley Centre Global Climate
10
Model (HadCM3; Sweeney and Fealy, 2003). A baseline climate period (1961-1990) and
11
two future climate scenarios 2055 (2041-2070) and 2075 (2061-2090) were investigated
12
along with the IPCC emission scenarios A2 and B2 (Nakicenovic et al., 2000; IPCC,
13
2007). Data generation was provided by the Department of Geography, National
14
University of Ireland, Maynooth (Sweeney and Fealy, 2003). Elevations in CO2 were
15
assumed by 2055 to be 581 ppmv and by 2075 to be 700 ppmv compared with a baseline
16
concentration of
17
managements for both the spring barley and the cut and grazed pasture were assumed to
18
be the same management as in 2004 for all scenarios (Table 1).
365 ppmv CO2 compatible with the IS95a (IPCC, 1995). Field
19
20
21
22
23
24
25
26
27
28
29
30
31
7
1
RESULTS AND DISCUSSION
2
3
Results presented in this paper assess the reliability of the DNDC model for estimating
4
N2O fluxes from both a spring barley field and a cut and grazed pasture by validating
5
model output with flux measurements collected on a weekly basis for up to two years.
6
Several management practices were examined, including conventional tillage, reduced
7
tillage and variable rates of N-fertilizer application. Climate and soil input variables for
8
DNDC are illustrated in Table 1. Field data measurements were used for all of the
9
variables listed except for atmospheric CO2, rainfall N, clay fraction and depth of the soil
10
water retention layer. Here default values were used. Collectively DNDC was better at
11
predicting N2O fluxes for high inputs of N fertilizer (>140 kg N ha-1) than for zero or low
12
N input treatments (0 to 70 kg N ha-1). In addition the model appeared to be unduly
13
sensitive to the influence of soil organic carbon. DNDC predicted a significant increases
14
of approximately 20 to 60% in future N2O fluxes from Irish cereal and grassland fields,
15
by the end of this centaury.
16
17
Arable field
18
Measurements of N2O flux were limited to the growth period of the barley crop hence
19
annual estimates of flux were not produced. Figures 1 to 3 relate to a comparison of the
20
modelled and measured fluxes for 2004/2005 as either daily values (Figures 1 to 2), or
21
cumulative flux (Figure 3). In general the temporal pattern of N2O flux was different
22
between modelled and measured data, DNDC extending the influence of added fertilizer
23
over a wider time period and producing smaller peaks. This is more pronounced for the
24
higher fertilizer treatments in 2004 than 2005 (Figures 1A, 1C and 2A) and can be clearly
25
seen in the cumulative flux plots (Figures 3A and 3B). This discrepancy between the
26
years maybe related to DNDC overestimating the water filled pore space (WFPS) in 2004
27
as opposed to 2005, WFPS being a critical determinant of N2O flux at the time of
28
fertilizer application (Keller and Reiners, 1994; Ruser et al., 1998; Dobbie and Smith,
29
2001). This is illustrated in Figure 4A where modelled WFPS values were consistently
30
higher than measured values in 2004, with maximum differences of 25 to 30% being
8
1
recorded. In comparison, modelled values for 2005 approximated to measured values
2
with maximum differences of only 13 to 16%.
3
The tillage options provided by DNDC do not allow the reduced, non-inversion tillage
4
used in our study to be fully described. In contrast to the conventional tillage plots,
5
DNDC significantly underestimated the N2O flux from the reduced tillage plots for the
6
medium and higher fertilizer treatments by up to 55% (Figures 3B and 3D). This may not
7
be critical for modeling N2O fluxes from Irish agriculture as reduced cultivation and
8
direct drilling of cereal crops represents at most only 10% of arable land, < 40,000 ha
9
(Fortune et al., 2003; ECAF, 2004).
10
11
Cumulative fluxes from sowing to harvest are given in Table 2. Modelled fluxes for the
12
high fertilizer inputs agreed with field measured values, giving the smallest relative
13
deviations from field data of -1 and -6%. These deviations increase significantly as
14
fertilizer input is reduced. The largest % deviation, and hence the worst fit was obtained
15
for the zero fertilizer treatments, with relative deviations of -35 to more than 5000%
16
calculated. Clearly DNDC is best suited for medium to high N input treatments and does
17
not account for negative flux values that can occur in low to zero N input treatments
18
where the soil acts as a sink for N2O (Ryden, 1981; Clayton et al., 1997). Similar DNDC
19
results for high and medium N fertilizer inputs have been reported for rice fields by
20
Zheng et al., 1999 (381 kg N ha-1; 8% deviation), for maize fields by Crill et al., 2000
21
(181 kg N ha-1; 3.5% deviation), for grass by Hsieh et al., 2005 (337 kg Nha-1; 33%
22
deviation) and for barley fields by Flessa et al., 1995 (50 kg N ha-1; 36% deviation).
23
However, these observations are not consistent in the literature. In contrast to our results
24
far better agreements between modelled and measured flux values have been obtained for
25
low to zero N inputs by Li, (1992), Mosier et al., (1996), Terry et al., (1981) and Crill et
26
al., (2000).
27
28
The wide range of CAN input values provided by this study allowed a linear regression of
29
modelled vs measured cumulative fluxes underlining the suitability of DNDC for
30
predicting N2O flux. This is illustrated in Figure 5, where observed and modelled data
31
from Table 2 have been plotted. The regression (y = 0.78x - 6.5) accounts for 85% of the
9
1
variation in the data, the predicted y values underestimating measured values by 24%.
2
Similar data cited by De Vries et al., (2005), from a range of published studies on
3
grasslands and cereal systems, is also presented in Figure 5. Data from our study fits well
4
within this group and improves the slope of the regression to y = 1.1x + 0.35, (r2 = 0.76).
5
6
Cut and grazed pasture
7
Our results suggest that DNDC is unduly sensitive to initial soil organic carbon content.
8
Measured and modelled cumulative fluxes of N2O from the cut and grazed pasture are
9
shown in Table 3 (annual) and Figure 6 (weekly) and highlight the poor fit of the model
10
where high relative deviation values were calculated. The only major difference between
11
the arable and the cut and grazed pasture soils is that the latter has significantly higher
12
organic carbon content (0.038 as opposed to 0.019 kg C kg-1 dwt). Changing the initial
13
soil organic C content for the model to the lower, arable soil value greatly improved the
14
fit of the model to the observed values (Figure 6). Using these new values the annual N2O
15
flux for the fertilized plots is 2797 g N2O-N ha-1 (a relative deviation of 9%) and for the
16
control plots is 1110 g N2O-N ha-1 (a relative deviation of 5%) as shown in Table 3.
17
This would question the present algorithms in the model describing the effect of soil
18
organic carbon on N2O flux. The model is very sensitive to SOC; a 20% increase in SOC
19
corresponds to a 62% increase in N2O flux (see below). Similar over-estimations of the
20
effects of initial SOC by DNDC have also been reported by Li et al., (1992a), Brown et
21
al., (2002) and Hsieh et al., (2005).
22
23
Sensitivity analysis
24
Given the good fit of the model to the conventional tillage data, the sensitivity of the
25
model outputs for the arable field to changes in soil characteristics, fertilizer N and
26
climate were also investigated. The following scenarios were chosen:
27
(1) Changes in bulk density
28
(2) Changes in initial SOC
29
(3) Changes in fertilizer use
30
(4) Changes in rainfall and air temperature.
31
10
1
The model appears highly sensitive to changes in bulk density and as mentioned
2
previously, SOC. Increasing the bulk density of the soil from 1.4 to 1.8 g cm-1, an
3
increase of 29%, resulted in a more than equivalent increase in both the apparent rate of
4
denitrification (53%) and the predicted N2O flux (89%), these increases presumably due
5
to more substrate N being made available through increased mineralization (Table 4).
6
Thus according to DNDC, any management treatment that increases the bulk density of
7
the soil, such as reduced tillage, would also significantly increase N2O flux as has been
8
observed by Aulakh et al., (1984); Baggs et al., (2003) and Six et al., (2004). Reduced
9
tillage is also associated with increases in SOC. By increasing the baseline SOC value by
10
20% increases N2O flux by 85%. Hence for at least two associated aspects of reduced
11
tillage, N2O flux has been predicted to increase significantly questioning the use of this
12
management technique as a means of lowering total greenhouse gas emissions from the
13
soil (Six et al., 2004; Li et al., 2005).
14
15
Model outputs were also highly sensitive to changes in fertilizer type, with a switch from
16
the principle form of N fertilizer used in cereal production in Ireland, CAN, to urea or
17
ammonium sulphate fertilizers resulting in predicted increases in N2O flux of 76 and 81%
18
respectively. Model outputs however, proved the most sensitive to changes in air
19
temperature. Here an increase of 1.5oC in the daily average air temperature resulted in a
20
89% increase in N2O flux and a 73% increase in the rate of soil denitrification. In contrast,
21
changes in rainfall of ± 20% resulted in changes in N2O flux of the order of ± 26%.
22
23
For the arable field, emission factors for the modelled data ranged from 0.3 to 0.6% of the
24
fertilizer N applied, whereas measured EFs ranged from 0.4 to 0.7% of the fertilizer N
25
applied. Modelled and measured EFs are comparable, but are both significantly lower
26
than the IPCC default value of 1.25%. However, literature EF values for cereal crops are
27
extremely variable, ranging from 0.2 to 8% (Eichner, 1990; Kaiser et al., 1998; Smith et
28
al., 1998, Dobbie et al., 1999) and are dependent upon temperature, moisture and soil
29
type (Flechard et al., 2007).
30
31
11
1
Simulation of future N2O flux
2
Figures 7 and 8 illustrate the DNDC predicted fluxes of N2O from both the barley field
3
(conventional tillage only) and the cut and grazed pasture for emission scenarios A2 and
4
B2 using data generated by the Hadley Centre Global Climate Model. A baseline climate
5
period (1961-1990) and two future climate scenarios for 2055 (2041-2070) and 2075
6
(2061-2090) were investigated.
7
Future temperatures are expected to increase especially during the spring and summer
8
periods of crop growth and fertilizer application. ICARUS (2006) predicts the July mean
9
temperature to increase by up to 2.5oC by the end of this century which will influence soil
10
denitrification and consequently N2O flux (Addiscott, 1983; Scott et al., 1986;
11
Beauchamp et al., 1989; Flessa et al., 2002). Wetter winters are also predicted, increasing
12
by as much as 11% by the end of the century (ICARUS, 2006). Besides displacement of
13
N2O by soil water, as the WFPS increase, the diffusion of oxygen into soil aggregates
14
will decrease stimulating denitrification (Dobbie and Smith, 2001). These increases in
15
temperature and rainfall effects will result in seasonal increases in N2O flux as clearly
16
seen in Figure 7.
17
In all cases DNDC simulates three specific peaks in N2O flux throughout the year, the
18
magnitude of these peaks being greatest for the cut and grazed pasture. The first peak
19
from day 50 to 75 is primarily due to seasonal rainfall, as is the third peak from day 225
20
to 350, the second peak however, from day 100 to 150 relates to fertilizer application. A
21
major difference between the two fields is that the third peak for the spring barley field
22
also coincides with crop residue incorporation resulting in a more spiked appearance. For
23
both crops however, DNDC simulated an increase in N2O emissions with each climate
24
scenario due to increasing CO2, temperature and rainfall variability. This increase is
25
particularly prominent for each seasonal peak in the spring barley field, but for the cut
26
and grazed pasture seems primarily associated with the third peak (Figures 7 and 8).
27
Annual cumulative fluxes derived from the modelled outputs are summarised in Table 5,
28
and illustrate a significantly greater flux of N2O-N from the cut and grazed pasture due to
29
higher N fertilizer application rate in addition to organic N inputs from grazing cattle.
30
However the modelled baseline value of approximately 15 kg N2O-N ha-1 y-1 is almost 5
12
1
times higher than the measured annual flux for 2004 (Table 3), even assuming the same
2
initial SOC value as the cereal field. Major seasonal differences between the modelled
3
and measured flux values appear to centre on the first and third seasonal peaks, none of
4
which were seen to occur for the grassland field in 2004 (data not shown). Accepting this
5
limitation on model outputs there would appear to be no significant difference between
6
the emission scenarios A2 and B2 with regard to both grassland and cereal fluxes of N2O
7
by the year 2075. Here fluxes are predicted to increase by approximately 20% for
8
grassland sites to 18 kg N2O-N ha-1 and by approximately 30 to 60% for the cereal sites
9
to 6 kg N2O-N ha-1 y-1 (Table 5).
10
11
CONCLUSIONS
12
13
In its present format DNDC is only suitable for medium to high N input systems, the
14
accuracy of the prediction being highly dependant on the level of fertilizer application,
15
with high fertilizer inputs producing low relative deviations between modelled and
16
measured fluxes of the order of 1 to 6% for the arable field under conventional tillage.
17
Prediction of N2O fluxes from reduced tillage plots however was poor with DNDC
18
consistently underestimating measured field values. Here relative deviations ranged from
19
-20 to -93%. One major disadvantage of the model was the limited choice of tillage input
20
options available, none describing the reduced tillage treatment used in this study.
21
Prediction of N2O fluxes from the cut and grazed grassland was also poor with model
22
outputs significantly overestimating measured field values giving relative deviations of
23
150 to 360%. From the sensitivity analysis we tentatively suggest that DNDC
24
overestimates the effect of SOC on mineralization and denitrification. By reducing the
25
SOC input values to those of the cereal field we could significantly improve the fit of the
26
model, reducing relative deviation scores to approximately 5 -10%.
27
28
Accepting the limitations of the model we used DNDC to predict future increases in N2O
29
flux due to climate change for our cereal and grassland fields in Ireland using the Hadley
30
Centre Global Climate Model data and the IPCC emission scenarios A2 and B2. Both
31
fields resulted in significant increases in N2O flux by the year 2075, grassland flux
13
1
increasing by 19 to 22% and arable flux increasing by 31 to 59%. In actual terms the
2
predicted flux for 2075 is significantly higher for grassland fields (18 kg N2O-N ha-1 y-1)
3
than for the cereal fields (6 kg N2O-N ha-1 y-1) with little difference being observed
4
between the A2 and B2 scenarios.
5
6
ACKNOWLEDGEMENTS
7
8
This work was funded by the EU sixth framework program (contract EVK2-CT2001-
9
00105, Greengrass Project Europe) and Irish EPA project No: 2001-CD-C1M1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
14
1
REFERENCES
2
3
Addiscott, T.M., 1983. Kinetics and temperature relationships of mineralization and
4
nitrification in Rothamsted soils with differing histories. Soil Science 34, 343-
5
353.
6
Aulakh, M.S., Rennie D.A., Paul, E.A., 1984. The influence of plant residues on
7
denitrification rates on conventional and zero-tilled soils. Soil Science Society of
8
America Journal 48, 790-794.
9
Baggs, E.M., Stevenson, M., Pihlatie, M., Regar, A., Cook, H., Cadish, G. 2003. Nitrous
10
oxide emissions following application of residues and fertilizer under zero and
11
conventional tillage. Plant and Soil 254, 361-370.
12
Batlach, R.M. and Tiedje, J.M., 1981. Kinetic explanation for accumulation of nitrite,
13
nitric oxide and nitrous oxide during bacterial denitrification. Applied and
14
Environmental Microbiology 42, 1074-1084.
15
16
17
18
19
20
21
Beauchamp E.G., Trevors J.T., Paul J.W., 1989. Carbon sources for bacterial
denitrification. Advances in Soil Sciences 10, 113-142.
Bouwman, A.F., 1996. Direct emissions of nitrous oxide from agricultural soils. Nutrient
Cycling in Agro-ecosystems 46, 53-70.
Brown, L., Jarvis, S.C., Headon, D., 2001. A farm scale bases for predicting nitrous oxide
emissions from dairy farms. Nutrient Cycling in Agro-ecosystems 60, 149-158.
Brown, L., Syed, B., Javris, S.C., Sneath, R.W., Philips, V.R., Goudling, K.W.T., Li, C.,
22
2002. Development and application of mechanistic model to estimate emission of
23
nitrous oxide from UK agriculture. Atmospheric Environment 36, 917-928.
24
Cai, Z., Swamoto, T., Li, C., Kang, G., Boonjawat, J., Mosier, A., Wassmann, R.,
15
1
Tsuruta, H., 2003. Field validation of the DNDC-model for greenhouse gas
2
emissions in East Asian cropping systems. Global Biogeochemical Cycles 17, No.
3
4, 1107.
4
Cho, C.M., Sakdinan, L., Chang, C., 1979. Denitrification intensity and capacity of three
5
irrigated Alberta soils. Soil Science Society of America Journal 43, 949-950.
6
Clayton, H., McTaggart, I.P., Parker, J., Swan, L., Smith, K.A., 1997. Nitrous oxide
7
emissions from fertilized grassland: A 2-year study of the effects of N fertilizer
8
form and environmental conditions. Biology and Fertility of Soils 25, 252-260.
9
Crill, P., Keller, M., Weitz, A., Grauel, B., Veldkamp, E., 2000. Intensive field
10
measurements of nitrous oxide emissions from a tropical agricultural soil. Global
11
Biogeochemical Cycles 14, 85-96.
12
De Vries, W., Kros, J., Kuikman, P.J., Velthof, G.L., Voogd, J.C.H., Weiggers, H.J.J.,
13
Butterbach-Bahl, Denier Van der Gon, H.A.C., Van Amstel, A.R., 2005. Use of
14
measurements and models to improve the national IPCC based measurements of
15
soil emissions of nitrous oxide. Environmental Sciences 2, 217-233.
16
Dobbie, K. E., McTaggart, I.P., Smith K. A., 1999. Nitrous oxide emissions from
17
intensive agricultural systems: variations between crops and seasons; key driving
18
variables; and mean emission factors. Journal of Geophysical Research 104,
19
26891-26899. Eaton, I.J. and Patriquin, D.J., 1989. Denitrification in low bush
20
blue berry soils. Canadian Journal of Soil Science 69, 303-312.
21
Dobbie, K. E. and Smith K. A., 2001. The effects of temperature, water filled pore space
22
and land use on N2O emissions from imperfectly drained gleysol. European
23
journal of Soil Science, 52, 667-673.
24
25
26
Eichner, M.J., 1990. Nitrous oxide emissions from fertilized soils: summary of available
data. Journal of Environment Quality 19, 272-280.
ECAF, 2004. European Conservation Agriculture Federation, 2004. http:// www.ecaf.org
16
1
Flechard, C., Ambus, P., Skiba, U., Rees, R.M., Hensen, A., Van den Pol, A., Soussana,
2
J.F., Jones, M., Clifton-Brwon, J., Raschi, A., Horvath, L., Van Amstel, A.,Neftel,
3
A., Jocher, M., Ammann, C., Fuhrer, J., Calanca, P., Thalman, E., Pilegaard, K.,
4
Di Marco, C., Campbell, C., Nemitz, E., Hargreaves, K.J., Levy, P., Ball, B.,
5
Jones, S., Van de Bulk, W.C.M., Groot, T., Blom, M., Gunnink, H., Kasper, G.,
6
Allard, V., Cellier, P., Laville, P., Henault, C., Bizouard, F., Jolivot, D., Abdalla,
7
M., Williams, M., Baronti, S., Berretti, F., Grosz, B., Dominques, R., 2007.
8
Effects of climate and management intensity on nitrous oxide emissions in
9
grassland systems across Europe. Agriculture, Ecosystems and Environment 121.
10
11
135-152.
Flessa, H., Dorsh, P., Beese, F., 1995. Seasonal variation of N2O and CH4 fluxes
12
in differently managed arable soils in southern Germany. Journal of Geophysical
13
Research 100, 23115-23124.
14
Flessa, H, Ruser, R., 2002. N2O and CH4 fluxes in potato fields: automated
15
measurement, management effects and temporal variation. Geoderma 105, 307-
16
325.
17
18
19
Fortune, T., Kennedy, T., Mitchell, B., Dunne, B., 2003. Reduced cultivation- Agronomic
and Environmental Aspects. www.teagasc.ie/publications/2003.
Freibauer, A. and Kaltschmitt, 2003. Regionalized inventory of biogenic greenhouse gas
20
emissions from European agriculture. European Journal of Agronomy 19, 135-
21
160.
22
23
24
Frissel, M. and Van Veen, J.A., 1981.Simulation of nitrogen behaviour of soil plant
systems (eds). PUDOC, Wageningen. P. 277.
Frolking, S.E., Mosier, A.R., Ojima, D.S., Li, C., Parton, W.J., Potter, C.S., Priesack, E.,
25
Stenger, R., Haberbosch, C., Dorsch, P., Flessa, H., Smith, K.A., 1998.
26
Comparisons of N2O emissions from soils at three temperate agricultural sites:
27
simulations of year-round measurements by four models. Nutrients Cycling in
28
Agro-ecosystems 52, 77-105.
17
1
Hsieh, C.I., Leahy, P., Kiely, G., Li, C., 2005. The effect of future climate perturbations
2
on N2O emissions from a fertilized humid grassland. Nutrient Cycling in Agro-
3
ecosystems 73, 15-23.
4
Hyde, B., Ryan, M., Hawkins, M., Connolly, J., Carton, O.T., 2005. Nitrous oxide
5
emissions. End of Project Report, RMIS 4974. Teagasc,
6
Research Centre, Co. Wexford.
Johnstown Castle
7
ICARUS, 2006. Irish Climate Analysis and Research Unit. www.icarus.nuim.ie
8
IPCC, 1995. Greenhouse gases and aerosols. Climate Change. Cambridge University
9
10
11
Press, Cambridge, pp. 1-41.
IPCC, 1997. Revised 1996 IPCC guidelines for national greenhouse gas inventories.
IPCC/OECD/IEA, IPCC, Geneva, Switzerland.
12
IPCC, 2000. Good practice guidance and uncertainty management in national greenhouse
13
gas inventories. In: Liep, A., Dammgen, U., Kuikman, P., Van Amstel, A., 2005.
14
The quality of European (EU-15) greenhouse gas inventories from agriculture.
15
Environmental Sciences 2, 177-192.
16
17
18
IPCC, 2007.Climate Impacts, Adaptation and Vlunerability. IPCC WGII Fourth
Assessment Report, Geneva, Switzerland.
Jinguo Y., Zheng N., Chenli W., 2006. Vegetation NPP distribution based on MODIS
19
data and CASA model–A case study of northern Hebei province. Chinese
20
Geographical Science 16, 334-341.
21
Kaiser, E.A., Kohrs, K., Kucke, M., Schnug, E., Heinemeyer, O., Munch, J.C., 1998.
22
Nitrous oxide release from arable soil: Importance of fertilization, crops and
23
temporal variation. Soil and Biochemistry 30, 1553-1563.
24
Keller, M. and Reiners, W.A. 1994. Soil-atmosphere exchange of nitrous oxide, nitric
18
1
oxide, and methane under secondary succession of pasture to forest in the Atlantic
2
lowlands of Cost Rica. Global Biogeochemical Cycles 8, 399-409.
3
Keisk, M., Ambus, P., Baritz, R., Bruggemann, N., Butterbach-bahl, K., Damm, M.,
4
Duyzer, J., Hovarth, L., Kiese, R., Kitzler, B., Leip, A., Li, C., Pihlatie, M.,
5
Pilegaard, K., Seufert, G., Simpson, D., Smaiatek, G., Skiba, U., Vesala, T.,
6
Zechmeisster-Boltenstern, S., 2006. Inventory of N2O and NO emissions from
7
European forest soils. Biogeosciences. Physiological Research111, GO2018.
8
Laegreid M. and Aastveit A.H., 2002. Nitrous oxide emissions from field applied
9
fertilizers. In: Petersen S.O and Olesen J.E. (Eds.). Greenhouse gas inventories for
10
agric. In the Nordic Countries, DIAS Report. Plant production no, 81. Danish
11
Institute of Agriculture. Sciences, pp. 122-134.
12
13
14
15
16
17
18
Leffelaar, P.A. and Wessel, W.W., 1988. Denitrification in a homogenous closed system:
experiment and simulation. Soil Science 146, 335-349.
Li, C., 2000. Modelling trace gas emissions from agricultural ecosystems. Nutrients
Cycling in Agro-ecosystems 58, 259-276.
Li, C., 2003. Greenhouse gas emissions from croplands of China. Quaternary
Sciences 23, 493-503.
Li, C., Frolking, S., Frolking, T.A., 1992. A model of nitrous oxide evolution from soil
19
driven by rainfall events.1. Model structure and sensitivity. Geophysical Research
20
97, 9759-9776.
21
22
23
Li, C., Frolking, S., Harris, R., 1994. Modelling carbon biogeochemistry in agricultural
soils. Global Biogeochemical Cycles 8, 237-254.
Li, C., Narayanan, V., Harriss, R., 1996. Model estimate of N2O emissions from
24
agricultural lands in the United States. Global Bio-geophysical Cycles
25
10, 297-306.
19
1
Li, C., Frolking, S., Butterbach-Bahl, K., 2005. Carbon sequestration in arable soils is
2
likely to increase nitrous oxide emissions, offsetting reductions in climate
3
radiative forcing. Climate Change 72, 321-338.
4
Nakicenovic, N., and Intergoervmental Panel on Climate Change. Special Report on
5
Emissions Scenarios. Pp 599 S., Cambridge University Press, Cambridge, UK,
6
2000.
7
Pathak, H., Li, C., Wassmann, H., Ladha, J.K., 2006. Simulation of Nitrogen balance in
8
Rice-wheat systems of the Indoigangetic plains. Soil Science Society of America
9
Journal 70, 1612-1622.
10
Roelandt, C., Van Wesemael, B., and Rounsevell, M., 2005. Estimating annual N2O
11
emissions from agricultural soils in temperate climate. Global Change Biology 11,
12
1701-1711.
13
Ruser, R., Flessa, H., Schilling, R., Steidl, H., Beese, F., 1998. Soil compaction
14
and fertilization effects on nitrous oxide and methane fluxes in potato fields. Soil
15
Science Society of America Journal 62, 1587-1595.
16
17
18
Ryden, J.C., 1981. Nitrous oxide exchange between a grassland soil and atmosphere.
Nature (London) 292, 235-237.
Saggar, S., Giltrap, D.L., Li, C., Tate, K.R., 2007. Modelling nitrous oxide
19
emissions from grazed grassland in New Zealand. Journal of Agriculture
20
Ecosystems and Environment 119, 205-216.
21
Schmid, M., Neftel, A., Riedo, M., Fuhrer, J., 2001. Process-based modelling of nitrous
22
oxide emissions from different nitrogen sources in mown grassland. Nutrient
23
Cycling of Agro-ecosystem 60, 177-187.
24
Six, J., Ogle, S., Breidt, F.J., Contant, R.T., Mosier, A.R., Paustian, K., 2004. The
25
potential to mitigate global warming with no-tillage management is only realized
26
when practised in the long term. Global Change Biology 10, 155-160.
20
1
Smith, K.A., Clayton, H., McTaggart, I.P., 1995. The measurement of nitrous oxide
2
emissions from soil by using chambers. Philosophical Transactions of the Royal
3
Society of London, Series a 351, 327-337.
4
Smith, K.A., McTaggart, I.P., Dobbie, K.E., Konen, F., 1998. Emissions of N2O from
5
Scottish agricultural soils, as a function of fertilizer N. Nutrient Cycling in Agro-
6
ecosystems 52, 123-130.
7
Stange, F., Butterbach-Bahl, K., Papen, H., Zechmeister-Boltenstern, S., Li, C., Aber, J.,
8
2000. A process oriented model of N2O and NO emissions from forest soils:
9
Sensitivity analysis and validation. Journal Geophsical Research 105, 4385-4398.
10
Stenger R., Priesack E., Barkle G., Sperr C., 1999. A tool for simulating nitrogen and
11
carbon dynamics in the soil–plant –atmosphere systems. NZ land treatment
12
collective. Proceedings Technical Session 20: Modelling of land treatment
13
systems. M. Tomer, M. Robinson, G. Gielen (Eds). New Plymouth, 14-15
14
October 1999; Pp 19-28.
15
Sweeney, J and Fealy, R., 2003. Establishing Reference Climate Scenarios. In: J.
16
Sweeney (Ed.), Climate Change Scenarios and Impacts for Ireland. ERTDI report
17
Series No. 15. Environmental Protection Agency, Johnstown Castle, Wexford, pp.
18
5-32.
19
20
Tang, H., Qiu, J., Van Ranst, E., Li, C., 2006. Estimation of soil organic carbon storage in
cropland of China based on DNDC model. Goederma 134, 200-206.
21
Tanji, K.K., 1982. Modelling of the soil nitrogen cycle. Agronomy 22, 721-772.
22
Terry, R.E., III, R.L.T. & Duxbury, J.M. 1981. The effect of flooding on nitrous
23
24
25
oxide emissions from and organic soil. Soil Science, 132: 228-232.
Willams, E.J., Hutchinson, G.L., Feshsenfeld, F.C., 1992. NOx and N2O emissions from
soil. Global Biogeochemical Cycles 6, 351-388.
21
1
Yeluripati, J.B., Li, C., Frolking, S., Nayak, D.R., Adhya, D.K., 2006. Field validation of
2
DNDC model for methane and nitrous oxide emissions from rice-based
3
production systems of India. Nutrients Cycling in Agro-ecosystems 74, 157-174.
4
Zheng, X.H., Wang, M.X., Wang, Y.S., Shen, R.X., Li, J., Heyer, J., Kögge, M.,
5
Papen H., Jin, J.S., Li, L.T., 1999. Characters of greenhouse gas (N2O, NO, CH4)
6
emissions from croplands of Southeast China, World Resource Review 11, 229–
7
246
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
22
1
2
3
4
TABLES
Table 1: DNDC model input data for both the spring barley and the pasture fields
Climate data
Latitude (degree)
Yearly maximum of average
Daily temperature (oC)
Yearly minimum of average
Daily temperature (oC)
Yearly accumulated precipitation
(mm).
N concentration in rainfall (mg Nl-1)
Atmospheric CO2 concentrations (ppm)
Soil properties (0-10 cm depth)
Vegetation type
Soil texture
Bulk density (g cm-3)
Clay fraction
Soil pH
Initial organic C content at surface soil
(kg Ckg-1).
Harvest
Soil tillage
WFPS at field capacity
WFPS at wilting point
Depth of water-retention layer (cm)
Slope (%)
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Spring barley field
52o86′ N
13
Pasture field
52o86′ N
13
4.0
4.0
792
792
0.001*
380*
0.001*
380*
Barley crop
Sandy loam
1.4
0.19*
7
0.019
Moist pasture
Sandy loam
1.0
0.34*
7.3
0.038
Grain harvest, mulch/till
Conventional and reduced
0.68
0.12
100*
0.0
Grazing/ cutting
None
0.87
0.09
100*
0.0
*
Default values
23
1
2
3
Table 2: Observed and modelled seasonal N2O emissions from the arable conventional
and reduced tillage plots.
2004 season
Conventional
tillage
Treatment
140 kg N ha-1
70 kg N ha-1
0 kg N ha-1
Reduced tillage 140 kg N ha-1
70 kg N ha-1
0 kg N ha-1
2005 season
Conventional
159 kg N ha-1
tillage
79 kg N ha-1
0 kg N ha-1
Reduced tillage 159 kg N ha-1
79 kg N ha-1
0 kg N ha-1
4
5
6
7
8
Seasonal emissions (g N2O-N ha-1)
Relative
Observation Model
Difference deviation (%)
-1
788
780
-8
269
2
978
494
87
350
110
590
220
30
+81
+108
-388
-274
-57
1053
993
-60
563
170
1058
567
135
450
110
793
320
10
-113
-60
-265
-247
-125
30
5400
-40
-55
-66
-6
-20
-35
-25
-44
-93
Table 3: Observed and modelled annual N2O emissions from the cut and grazed pasture
(2004).
Treatment
Before adjusting SOC
200 kg N ha-1
0 kg N ha-1
After adjusting SOC
200 kg N ha-1
0 kg N ha-1
Seasonal emissions (g N2O-N ha-1)
Observation Model
Difference
Relative Deviation (%)
2573
1054
6613
3970
4040
2926
157
360
2573
1054
2797
1110
224
56
9
5
9
10
11
12
13
14
15
16
17
18
19
24
1
2
3
Table 4: Sensitivity of DNDC to changes in soil characteristics, management and climate
for the spring barley field (conventional tillage, 2004).
Scenario
*Baseline
Bulk density (g cm-1)
1
1.6
1.8
Initial soil organic
carbon
+20%
-20%
Fertilizer type
Urea
Ammonium sulphate
Rainfall
+20%
-20%
Air temperature
+20%
-20%
4
5
6
7
8
9
10
11
Mineralization
(kg N ha-1y-1)
257.4
Annual N2O flux (kg N Denitrification
ha-1y-1)
(kg N ha-1y-1)
1.4
4
194
290.8
324.2
0.67
2.11
2.65
1.67
4.33
6.13
305.8
211.1
2.59
0.69
6.1
1.74
257.4
257.4
2.46
2.54
4.81
4.9
267.1
244.5
1.76
1.41
4.51
2.98
269.6
243.2
2.65
0.93
6.92
2.34
*Baseline scenario: Bulk density 1.4gcm-3, SOC 0.0194 kg C kg-1, fertilizer applied and timing (140kg N/ha
CAN, on the 27th of April), annual average max. and min. air temperature 13.7 and 4.8 oC and average
daily precipitation 2.2cm and soil tillage to 22cm depth carried in March five weeks before planting.
Table 5: DNDC future simulated annual cumulative N2O flux values for the grassland
and arable fields under emission scenarios A2 and B2.
Time Period
Cumulative Flux
Increase from (1961-1990)-base
line value (%)
-1
(Kg N2O-N ha )
Grassland
A2
B2
A2
B2
1961-1990
14.8
14.7
2041-2070
16.6
15.8
12.2
7.8
2061-2090
18
17.4
21.6
18.7
1961-1990
4.0
3.9
2041-2070
5.3
4.0
33.7
3.61
2061-2090
6.3
5.1
58.6
31.4
Barley
25
1
2
3
4
FIGURES
N2O flux (gN 2O-N ha-1 d-1)
60
60
A
50
50
40
40
30
30
20
20
10
10
0
0
-10
-10
90
110
130
150
170
190
210
N2O flux (gN 2O-N ha -1 d-1)
130
150
170
190
210
230
110
130
150
170
190
210
230
110
130
150
170
190
210
D
20
20
10
10
0
0
-10
-10
110
130
150
170
190
210
230
30
N2O flux (gN 2O-N ha-1 d-1)
110
30
C
90
90
30
F
E
20
20
10
10
0
0
-10
-10
90
110
130
150
170
Time (days after the 1
5
6
7
8
90
230
30
B
st
190
210
230
of January)
90
Time (days after the 1
st
of January)
Figure 1: Comparison of model-simulated (○) and field measured N2O (●) flux from the
high (upper), medium (bottom) and low (lower) fertilized conventional tillage in 2004
(A,C,E) and 2005 (B,D,F). Arrows show time of fertilizer application.
26
230
1
2
3
N2O flux (gN 2O-N ha -1 d-1)
60
60
A
50
50
40
40
30
30
20
20
10
10
0
0
-10
-10
90
110
130
150
170
190
210
90
230
N2O flux (gN 2O-N ha-1 d-1)
30
130
150
170
190
210
230
110
130
150
170
190
210
230
110
130
150
170
190
210
D
20
20
10
10
0
0
-10
-10
110
130
150
170
190
210
230
90
30
N2O flux (gN 2O-N ha-1 d-1)
110
30
C
90
30
F
E
20
20
10
10
0
0
-10
-10
90
110
130
150
170
Time (days after the 1
4
5
6
7
8
9
10
11
12
13
B
st
190
210
230
90
of January)
Time (days after the 1
st
230
of January)
Figure 2: Comparison of model-simulated (○) and field measured N2O (●) flux from the
high (upper), medium (bottom) and low (lower) fertilized reduced tillage in 2004 (A, C, E)
and 2005 (B, D, F). Arrows show time of fertilizer application.
27
(gN2O-N ha -1)
Cumulative N 2O flux
1000
900
800
700
600
500
400
300
200
100
0
-100
90
(gN2O-N ha -1)
Cumulative N 2O flux
1200
1000
900
800
700
600
500
400
300
200
100
0
-100
A
110
130
150
170
190
210
110
130
150
170
190
210
230
130
150
170
190
210
230
1200
D
1000
1000
800
800
600
600
400
400
200
200
0
0
110
130
150
170
190
210
230
Time (days from 1st January)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
90
230
C
90
B
90
110
Time (days from 1st January)
Figure 3: Comparisons of cumulative model-simulated (open symbol) and field measured
(solid symbol) N2O fluxes from the high (), medium (■) and low (▲) fertilized plots in
2004 and 2005 for conventional (A and C) and reduced (B and D) tillage system.
28
80
A
70
WFPS (%)
60
50
40
30
20
10
0
14-Mar 03-Apr 23-Apr 13-May 02-Jun 22-Jun 12-Jul 01-Aug 21-Aug
80
B
70
60
WFPS (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
50
40
30
20
10
0
14-Mar 03-Apr 23-Apr 13-May 02-Jun 22-Jun 12-Jul 01-Aug 21-Aug
Figure 4: Comparison between the measured (●) and modelled (○) WFPS from CN1
treatment in 2004 (A) and 2005 (B). Arrows indicate time of N fertilizer application
29
Cumulative N2O flux (gN2O-N ha-1)
7000
A
6000
5000
4000
3000
2000
1000
0
250
300
350
400
300
350
400
450
500
550
600
650
700
7000
Cumulative N2O flux (gN2O-N ha-1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
B
6000
5000
4000
3000
2000
1000
0
250
450
500
Time (days from the 1
st
550
600
650
700
January 2003)
Figure 5: Comparison between the measured (●) and modelled cumulative N2O from the
fertilized (A) and control (B) pasture plots before (○) and after (∆) adjusting soil organic
carbon.
36
37
38
39
30
1
3
4
Modeled N2O flux (kgN2O-N ha-1)
2
1.25
A
1.00
0.75
0.50
0.25
0.00
5
0.00
12.5
7
10.0
8
9
10
11
Modeled N2O flux (kgN2O-N ha-1)
6
13
14
15
16
0.50
2.5
5.0
0.75
1.00
1.25
7.5
10.0
12.5
B
7.5
5.0
2.5
0.0
0.0
12
0.25
Measured N2O flux (kgN2O-N ha-1)
Figure 6: (A) Correlation between the model-simulated and field measured N2O fluxes
for the arable field. y = 0.78x -6.5 (r2 = 0.85). (B) Correlation between the modelsimulated and field measured N2O fluxes from our arable (●), pasture (∆) and other
literature DNDC studies (○). y = 1.1x + 0.35, (r2 = 0.76).
31
85
A
80
75
70
-1 -1
N2O fluxes (gN 2O-N ha d )
65
60
55
50
45
40
35
30
25
20
15
10
5
0
0
25
50
75
100
125
50
75
100
125
150
175
200
225
250
275
300
325
350
375
400
1
65
B
60
55
-1 -1
N2O fluxes (gN 2O-N ha d )
50
45
40
35
30
25
20
15
10
5
0
0
25
150
175
200
225
250
275
300
325
350
375
Julian days
2
3
4
5
Figure 7: DNDC simulated N2O flux from the barley field soil at baseline climate; 19611990 (●), 2055 (○) and 2075 (∆) for the emission scenarios HCM3-A2 (A) and HCM3B2 (B).
32
400
A
320
300
280
260
N2 O fluxes (gN 2 O-N ha-1 d-1 )
240
220
200
180
160
140
120
100
80
60
40
20
0
0
25
50
75
100
125
150
175
200
225
250
275
300
325
350
375
400
225
250
275
300
325
350
375
400
Julian days
1
320
B
300
280
N2 O fluxes (gN 2 O-N ha-1 d-1 )
260
240
220
200
180
160
140
120
100
80
60
40
20
0
0
2
3
4
5
25
50
75
100
125
150
175
200
Figure 8: DNDC simulated N2O flux from the cut and grazed pasture soil at baseline
climate; 1961-1990 (●), 2055 (○) and 2075 (∆) for the emission scenarios HCM3-A2 (A)
and HCM3-B2 (B).
33