Application of the Denitrification-Decomposition Model to Predict Carbon Dioxide Emissions under Alternative Straw Retention Methods

Straw retention has been shown to reduce carbon dioxide (CO2) emission from agricultural soils. But it remains a big challenge for models to effectively predict CO2 emission fluxes under different straw retention methods. We used maize season data in the Griffith region, Australia, to test whether the denitrification-decomposition (DNDC) model could simulate annual CO2 emission. We also identified driving factors of CO2 emission by correlation analysis and path analysis. We show that the DNDC model was able to simulate CO2 emission under alternative straw retention scenarios. The correlation coefficients between simulated and observed daily values for treatments of straw burn and straw incorporation were 0.74 and 0.82, respectively, in the straw retention period and 0.72 and 0.83, respectively, in the crop growth period. The results also show that simulated values of annual CO2 emission for straw burn and straw incorporation were 3.45 t C ha−1 y−1 and 2.13 t C ha−1 y−1, respectively. In addition the DNDC model was found to be more suitable in simulating CO2 mission fluxes under straw incorporation. Finally the standard multiple regression describing the relationship between CO2 emissions and factors found that soil mean temperature (SMT), daily mean temperature (T mean), and water-filled pore space (WFPS) were significant.


Introduction
Atmospheric CO 2 concentrations have increased by approximately 35% and are predicted to reach 700 ppmv by the end of 20 century [1]. Soils are the largest carbon pool in terrestrial ecosystem, containing more than two thirds of the total carbon and soil respiration, and contribute an annual atmospheric CO 2 flux 10 times greater than fossil fuel combustion [2]. Therefore, it is crucial to reduce CO 2 emissions from agricultural land.
There are many factors affecting CO 2 emissions, including soil temperature, soil moisture, and soil organic matter (SOM) [3]. Schlesinger and Andrews [4] showed that soil CO 2 emissions increase with soil temperature. La Scala et al. [5] reported that microbial activity, soil respiration, and enzymatic activity increased with rising temperature. Subke et al. [6] suggested that soil moisture was an important nonbiological factor affecting soil CO 2 flux. Li [7] analyzed the sensitivity in predicting CO 2 and N 2 O flux emissions for a series of models and found that the amount of soil organic matter (SOM) was the most important factor. Straw retention has been adopted worldwide to increase crop production. It has been shown to reduce CO 2 but increase crop yield [8]. Li et al. [9] considered that if the straw retention rate increased from 15 to 80% in China, the Chinese agricultural carbon budget should change from negative (emissions from soil of −9.5 × 10 7 t y −1 ) to positive (soil absorbing +8.0×10 7 t y −1 ). While the effects of different straw retention methods on CO 2 emission flux have been studied in many continuous long-term experiments, information on the use of the DNDC model to simulate CO 2 emission under different straw retention scenarios is lack. In particular information is required to predict the total variation of CO 2 emission fluxes and the interrelation of important factors. The objective of this study was to simulate the daily rate of CO 2 emissions in the straw retention and crop growth 2 The Scientific World Journal Mean annual precipitation is 432 mm, and mean maximum and minimum temperatures are 23.0 and 10.3 ∘ C, respectively (measured at the nearest recording station, Leeton). The soil (0-20 cm) is classified as a Typic Natrixeralf and Mundiwa clay loam with clay particle content in 53.11% [10]. The surface soil has a pH of 5.5 (soil : water = 1 : 5) of 0.03 kg carbon (C)/kg and soil bulk density of 1.37 g cm −3 (Table 1). WFPS is calculated as WFPS = (soil gravimetric water content × bulk density)/(1 − (bulk density/particle density)) [10]. Maize was grown at the site on beds (two rows of plants per bed) and irrigated by furrow irrigation.

Experimental Design and Data
Analysis. The field experiment began on 11 May 2010 (day 1) and ended on 10 May 2011 (day 365). There were two maize straw treatments in the field experiment. A randomized block design with three replicates was used in the 12 plots. The maize straw treatments were (1) application of 300 kg N ha −1 , maize straw burnt and left on the field (300N-burn), and (2) application of 300 kg N ha −1 , maize straw mulched (the amount of maize straw was 6750 kg ha −1 ) and incorporated into soil (5 cm) soon after harvest (300N-incorporated The CO 2 fluxes from the soil-plant system were measured on the basis of static transparent chamber and gas chromatography methods [3]. The bottom chambers were empty and the top chambers were connected to an infrared CO 2  [11]. This system measures fluxes integrated over 2 or 3 day intervals and avoids errors associated with manual flux chambers, when measurements are taken only at a particular time period. Each system was powered by a 12 V 120 amphour battery. The battery capacity was supplemented by an 80 W solar panel that extended the period between battery changeover to between 1 and 4 weeks. Soil temperature was measured using type K thermocouples and soil moisture was measured using time domain reflectometry (TDR) (Theta Probes ML2x, Delta-T Devices Ltd., Cambridge, UK). Both sets of sensors were logged by a controller/logger unit. Gas concentrations in the Tedlar sample bags were measured offsite at Aspendale. CO 2 was measured using a Licor 6251 NDIR (Licor, Nebraska, USA). In addition CO 2 concentration was measured directly in the field using a Gascard II CO 2 sensor (Edinburgh Instruments, Edinburgh, UK) incorporated into each chamber controller.  Model accuracy and performance were evaluated by calculating the correlation coefficient and model efficiency (ME) [14]. ME is calculated as

DNDC Model
where is observed values, is simulated values, is the total number of observations, is the mean of observed values, and is the current observation. ME compares the squared sum of the absolute error with the squared sum of the difference between the observations and their mean value. It compares the ability of the model to reproduce the daily data variability with a much simpler model that is based on the arithmetic mean of the measurements. ME values close to 1 indicate a "near-perfect" fit [15,16].
Five continuous long-term measurement factors were considered for the statistical analysis, namely, daily maximum temperature ( max ), daily minimum temperature, ( min ), daily mean temperature ( mean ), soil mean temperature (SMT), and amount of irrigation. Irrigation was not a daily operation, but it had residual effects on soil water. Therefore, water-filled pore space (WFPS) was used to reflect the water dynamics following irrigation and rainfall events in soil. As a result, five factors which affected the CO 2 emission were selected: max , min , mean , SMT, and WFPS.
Data were analyzed by correlation analysis and path analysis using SPSS 13.0. Path analysis can be used for the analysis of multiple variables and the linear relationship between variables. It was a development of regression analysis [17].

Results
The straw retention period and crop growth period were studied separately because the sources of the CO 2 were different in these two periods [18]. The main sources of CO 2 were straw decomposition, SOM decomposition, and microbial respiration during the straw retention period. Root respiration, SOM decomposition, and microbial respiration are the main sources during the crop growth period [19][20][21].

Simulation of Daily CO 2 Emission during Straw Retention
Period. The simulated and observed values of daily CO 2 emission under different straw retention treatments during straw retention period are shown in Figure 1. The correlation coefficient between simulated and observed values for treatments 300N-burn and 300N-incorporation were 0.74 (Figure 1(a)) and 0.82 (Figure 1(b)), respectively, ( = 74) and the ME values were 0.63 and 0.76.

Daily CO 2 Emission during Crop Growth and Annual CO 2
Emissions. The DNDC model was also used to simulate the daily CO 2 fluxes during crop growth ( Figure 2). The correlation coefficients between the observed and simulated values of CO 2 flux were 0.72 and 0.82 ( = 186) for treatment 300N-burn and 300N-incorporation, respectively ( Figure 2). The corresponding ME values were 0.63 and 0.79. The ME value of 300N-incorporation was higher than that of the 300N-burn. This indicates that the DNDC model was more suitable for simulating CO 2 fluxes for the straw incorporation treatment than the straw burning treatment during the crop growth period.
The observed values of CO 2 emissions during the maize growth season for 300N-Burn and 300N-incorporation were 4.7 t C ha −1 y −1 and 3.5 t C ha −1 y −1 , respectively. The corresponding simulated values were 3.45 t C ha −1 y −1 and 2.13 t C ha −1 y −1 , respectively ( Table 3). The observed values were smaller than the observed values. This is because DNDC model could simulate the CO 2 which is discharged by the microbial activities. The CO 2 which produced though plant  1.0000 * * Correlation is significant at the 0.01 level-2-tailed). * Correlation is significant at the 0.05 level-2-tailed).
root respiration was ignored. This also suggested that the DNDC model might have underestimated the microbial activity and the rate of SOM decomposition [3]. So it is necessary to improve the DNDC model to adjust microbial activity and the rate of SOM decomposition simulation and contain the plant root respiration simulation function.

Sensitive Analysis.
Fixed factors (continuous long-term measurements) were used for the sensitivity analysis. Because the DNDC model was more suitable for the simulation of CO 2 emissions under straw incorporation than the burning of straw, the reason is shown in Sections 3.1 and 3.2, the sensitivity analysis mainly focused on the relationship between different factors and CO 2 emission for the straw incorporation treatment. The CO 2 emissions were significantly correlated with max , min , mean , WFPS, and SMT. The correlation coefficients were 0.5681, 0.5114, 0.5125, 0.5366, and 0.6729, respectively ( < 0.01) (Table 4). Temperature, WFPS, and microbial activity can influence SOM decomposition and associated CO 2 emissions [5,6,19].
Path analysis was used to analyze the relationship among these five factors (Tables 5 and 6). The results showed that SMT, mean , and WFPS were the main controlling factors of CO 2 emissions. The standard multiple regression equation of the CO 2 emission flux was = −34.113 + 0.8067 1 + 0.6392 2 + 0.4014 3 ( = 0.964, < 0.01, = 260), where is CO 2 emission fluxes, 1 is SMT, 2 is mean , and 3 is WFPS. The max and min were not chosen because of large diurnal fluctuation of temperature in the Griffith region. The daily maximum temperature and daily minimum temperature did not reach the optimum temperature for microbial activity.

The discussion of Daily CO 2 Emission during Straw
Retention Period. The observed and simulated CO 2 emission decreased with time regardless of straw retention treatment (Figure 1). This was because the main source of CO 2 emission was straw decomposition in the straw retention period [20]. The straw decomposition rate decreased with time. The simulated and observed CO 2 fluxes in the straw burnt treatment were higher than those under straw incorporation. The reason for the difference may be related to soil structural differences, particularly the reduced accessibility of N by plant roots in the burned treatment [22]. This difference may have contributed to the greater N 2 O emissions from soil in our straw burn treatments. Perhaps the process was the same in the CO 2 experiments. The change in soil structural in the straw burn treatment may allow microbes to obtain more mineral nutrition after straw burn and promote the formation of granular structure. Ruser et al. [23] investigated the impact of compaction on soils from a row cropping system and found that CO 2 production has a sensitive influence on the soil compaction. The other reason is in agreement The Scientific World Journal 5  with the finding by Beer et al. [24] that more greenhouse gas was evolved from straw burn treatments monitored using automatic chambers over the whole season. Therefore the straw burn treatment discharged more CO 2 than the straw incorporation treatment, so that the CO 2 emissions were lower in the straw incorporation than in the straw burn [9]. The correlation coefficient between simulated and observed values and ME values implies that the DNDC model can be used to simulate daily CO 2 fluxes in the straw retention period. The ME value for 300N-incorporation was higher than for 300N-burn, suggesting that the DNDC model was more suitable to simulate CO 2 emission under straw incorporation than under the burning of straw.

The Discussion of Daily CO 2 Emission during Crop Growth
and Annual CO 2 Emissions. Both the observed and simulated values for the treatment 300N-burn were higher than those for the 300N-incorporation. The result was the same as the straw retention method. Straw decomposition rate varies with the depth of incorporation [25]. It has been shown that the straw decomposition rate during the 32 weeks of study was the highest at the 5 cm soil depth (decomposed > 65%), followed by the 15 cm soil depth (62%), the lowest for the straw materials left on the soil surface (50%) [7]. Under the 300N-burn treatment, the maize straw was burnt and left on the field, and the main products of maize straw combustion were CO 2 and plant ash. CO 2 was emitted to the atmosphere directly. Plant ash was the main driver of microbial activity. Part of the plant ash was used by microbes in the straw retention period and the other part was used in the crop growth period. So increasing soil organic carbon associated with straw incorporation [26] would drive decreased CO 2 emission. On the other hand, the base of plant ash contains substantial mineral nutrients, which could be used by the microbes directly for energy [27,28]. Wakelin et al. [29] found that stubble burnt and incorporation led to totally dissimilar soil microbial populations. The straw, which was incorporated into the soil, might require a long organic matter decomposition process [7,30]. The soil organic matter decomposition process is slow. Therefore, the soil microbes in the straw incorporated treatment might obtain less energy than in the soil burn treatment, so that straw incorporation discharges CO 2 slower than straw burn.
The correlation coefficient between simulated and observed values and ME values indicates that the DNDC model was more suitable for simulating CO 2 fluxes for the straw incorporation treatment than the straw burning treatment during the crop growth period.

The
Discussion of Sensitive Analysis. SMT, mean , and WFPS showed a direct positive effect (the corresponding direct path coefficients were 0.8067, 0.6392, and 0.4014, resp.) and indirect positive effect (the total indirect path coefficients were 1.1016, 1.2499, and 1.0915, resp.) on CO 2 emissions (Table 6). This indicates that SMT, mean , and WFPS could directly and indirectly affect the microbial activity and decomposition of straw and organic matter in soil and control CO 2 emissions. The total indirect path coefficients of SMT, mean , and WFPS were greater than direct path coefficients. SMT, mean , and WFPS could influence the microbes in decomposing the organic matter and straw to release CO 2 [9]. The WFPS included irrigation and rainfall events in soil. So irrigation and rainfall events into soil may mainly indirectly affect CO 2 emissions.

Conclusions
The DNDC model can be used to simulate CO 2 emissions under different straw retention practices in the Griffith region, NSW, Australia. The results showed that the simulation values and trends were very close to the measured values of daily CO 2 fluxes, CO 2 annual emissions, and emission 6 The Scientific World Journal factors for all straw retention methods. The model accuracy for the 300N-incorporated treatment was higher than that for the 300N-burn treatment in both the straw retention period and the crop growth period. This implies that the DNDC model is more appropriate for simulation of CO 2 emissions under straw incorporation treatment.
CO 2 emissions were positively correlated with max , min , mean , WFPS, and SMT. The path analysis showed the standard multiple regression equation of the CO 2 emission was = −34.113 + 0.8067 1 + 0.6392 2 + 0.4014 3 . The SMT, mean , and WFPS were the main factors influencing CO 2 emission under different straw retention methods. Management of these practices will help mitigate CO 2 emissions in cropping systems